RADIONUCLIDES IN THE ENVIRONMENT International Conference on Isotopes in Environmental Studies: Aquatic Forum 2004, 25–29 October, Monaco
RADIOACTIVITY IN THE ENVIRONMENT A companion series to the Journal of Environmental Radioactivity Series Editor M.S. Baxter Ampfield House Clachan Seil Argyll, Scotland, UK Volume 1: Plutonium in the Environment (A. Kudo, Editor) Volume 2: Interactions of Microorganisms with Radionuclides (F.R. Livens and M. Keith-Roach, Editors) Volume 3: Radioactive Fallout after Nuclear Explosions and Accidents (Yu.A. Izrael, Author) Volume 4: Modelling Radioactivity in the Environment (E.M. Scott, Editor) Volume 5: Sedimentary Processes: Quantification Using Radionuclides (J. Carroll and I. Lerche, Authors) Volume 6: Marine Radioactivity (H.D. Livingston, Editor) Volume 7: The Natural Radiation Environment VII (J.P. McLaughlin, S.E. Simopoulos and F. Steinhäusler, Editors) Volume 8: Radionuclides in the Environment (P.P. Povinec and J.A. Sanchez-Cabeza, Editors)
RADIONUCLIDES IN THE ENVIRONMENT International Conference on Isotopes in Environmental Studies: Aquatic Forum 2004, 25–29 October, Monaco
Editors P.P. Povinec Comenius University, Bratislava, Slovakia and IAEA-MEL, Monaco
J.A. Sanchez-Cabeza IAEA-MEL, Monaco
AMSTERDAM – BOSTON – HEIDELBERG – LONDON – NEW YORK – PARIS SAN DIEGO – SAN FRANCISCO – SINGAPORE – SYDNEY – TOKYO
ELSEVIER B.V. Sara Burgerhartstraat 25 P.O. Box 211, 1000 AE Amsterdam The Netherlands
ELSEVIER Inc. 525 B Street, Suite 1900 San Diego, CA 92101-4495 USA
ELSEVIER Ltd The Boulevard, Langford Lane Kidlington, Oxford OX5 1GB UK
ELSEVIER Ltd 84 Theobalds Road London WC1X 8RR UK
© 2006 Elsevier Ltd. All rights reserved. This work is protected under copyright by Elsevier, and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. Permissions may be sought directly from Elsevier’s Rights Department in Oxford, UK: phone: (+44) 1865 843830; fax: (+44) 1865 853333; e-mail:
[email protected]. Requests may also be completed on-line via the Elsevier homepage (http://www.elsevier.com/locate/permissions). In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (+1) (978) 7508400; fax: (+1) (978) 7504744; and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London WlP 0LP, UK; phone: (+44) 20 7631 5555; fax: (+44) 20 7631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of Elsevier Science is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier’s Rights Department, at the phone, fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made.
First edition 2006 Library of Congress Cataloging in Publication Data A catalog record from the Library of Congress. British Library Cataloguing in Publication Data A catalogue record from the British Library. ISBN-13: 978-0-08-044909-8 ISBN-10: 0-08-044909-3 ISSN: 1569-4860 The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed in The Netherlands.
v
Preface
Radioactive and stable isotopes have been applied as tracers for better understanding of environmental processes for about fifty years, contributing to diverse areas such as atmospheric transport, ocean circulation, groundwater hydrology, sedimentation processes and climate change. However, research has always been limited by the techniques available for sampling and analysis. Recently we have seen important achievements, such as the use of robotic systems based on remotely operating vehicles (ROV) and autonomous underwater vehicles (AUV) to sample the ocean environment and the use of satellite information for the optimisation of sampling programmes. In the field of analytical technologies we have moved from simple radiochemical methods and gas counters to robotic radiochemical technologies and sophisticated detectors working on line with powerful computers, often situated deep underground to protect them against the cosmic radiation background. The philosophy of analysis of long-lived radionuclides has changed from the concept of counting decays to the counting of atoms using highly sensitive mass spectrometers working either with low energy ions, such as inductively coupled mass spectrometry (ICPMS), thermal ionisation mass spectrometry (TIMS) and resonance ionisation mass spectrometry (RIMS), or with ions accelerated to hundreds of MeV in accelerator mass spectrometers (AMS). These new developments in sampling and analytical techniques have been accompanied by changes in the philosophy and organisation of research, as institutional and national investigations have been replaced by global international projects such as World Ocean Circulation Experiment (WOCE), Joint Global Ocean Flux Study (JGOFS), Climate Variability and Predictability Study (CLIVAR), Past Global Changes (PAGES), Worldwide Marine Radioactivity Studies (WOMARS), Global Marine Biochemistry of Trace Elements and Isotopes (GEOTRACES), Southern Hemisphere Ocean Tracer Study (SHOTS), etc. These and other topics for the better understanding of key processes in the aquatic environment, responsible for its future development and its protection, were at the forefront of the IAEA’s International Conference on Isotopes in Environmental Studies – AQUATIC FORUM 2004 convened in Monaco from 25 to 29 October 2004, which was the most important gathering of the year of isotope environmental scientists. The conference was organised by the IAEA’s Marine Environment Laboratory in cooperation with the Intergovernmental Oceanographic Commission of UNESCO, the International Hydrological Programme of UNESCO, the Commission Internationale pour l’Exploration Scientifique de la Mer Mediterranée, and the Abdus Salam International Centre for Theoretical Physics. The conference was hosted by
vi
Preface
the Principality of Monaco. Over 320 experts from 60 IAEA Member States and 6 international organisations delivered 185 oral presentations in 6 plenary and 31 parallel sessions and made 130 poster presentations. The conference reviewed the present state of the art isotopic methods for investigation of the aquatic environment. Four workshops were held simultaneously: • ATOMS-Med Workshop – development of a project proposal for oceanographic investigations in the Eastern Mediterranean. • El Niño – Research Coordination Meeting of the new IAEA Coordinated Research Project investigating climate change using isotopic records in the marine environment. • CELLAR Workshop – Collaboration of European Low-Level Underground Laboratories. • GSI Workshop on Groundwater–Seawater Interactions in coastal zones, organised in cooperation with the IAPSO Commission on Groundwater–Seawater Interactions. The main conference highlights, which included the latest developments in the field, were: (i) new information on behaviour, transport and distribution of isotopes in the aquatic environment; (ii) recent climate change records using isotopic tracers in the environment (tree rings, corals, sediments); (iii) global oceanic studies by WOCE, WOMARS, SHOTS and GEOTRACERS projects; (iv) impacts of groundwater–seawater interactions on coastal zones; (v) groundwater dynamics and modelling, important for management of freshwater sources; (vi) new trends in radioecological investigations, concentrating on the protection of marine biota against radioactive contamination; (vii) transfers in analytical technologies from bulk analyses to particle and compound specific analyses of environmental samples; (viii) development of new isotopic techniques, such as AMS, RIMS and ICPMS, and their applications in environmental studies; (ix) new trends in radiometrics underground techniques; (x) new in situ radiometrics technologies, and many other exciting topics which were presented and discussed during the Conference. The contents of the book are assembled from selected papers presented during the Conference in plenary and parallel oral sessions. The Proceedings published by the IAEA contains the other papers presented in both oral and poster sessions which could not be included here for space reasons. In our opinion, the proceedings constitute an important contribution to environmental isotopic research. Finally, the Scientific Secretary of the Conference (P.P. Povinec), would like to thank colleagues on the Scientific Committee, those of sponsoring organisations, session chairmen, speakers in oral and poster session, colleagues at IAEA for their help in preparation and organisation of the Conference, and in general all the participants for their contributions to the success of the Conference. Special thanks are due to the scientific reviewers who donated their time and expertise to assure the high scientific quality of the papers presented in this book, namely to T. Altzitzoglou, R. Aravena, D. Arnold, G. Barci-Funel, G. Barrocu, D.L. Biddulph, E. Boaretto, A. Bode, R. Bojanowski, D. Boust, W.C. Burnett, J. Carroll, J. Dean, S. de Mora, J. de Oliveira, R. Delfanti, E. Duran, V. Egorov, V.I. Ferronsky, H. Florou, S.W. Fowler, T. Gäfvert, R. García-Tenorio, J.R. Gat, J.M. Godoy, S. Grabowska, E. Güngör, T. Hamilton, I. Harms, S. Hauser, G. Heusser, K. Hirose, A. Hogg, E. Holm, T. Honda, G.H. Hong, M. Horvat, N. Horvatinˇci´c, X. Hou, M. Hult, Y. Igarashi, Y. Ikeuchi, M. Iosjpe, T. Ito, C. Jeandel, G. Jia, D. Jones, G. Kanish, R.M. Key, W.E. Kieser, M. Köhler, K. Komura, E. Kontar, M. Korun, A. Kryshev, R. Ladygiene, M. Laubenstein, G. Lazorenko, M.-H. Lee,
Preface
vii
S.-H. Lee, J. Mattila, R.L. Michel, J.W. Mietelski, J. Miralles, M. Nakano, S. Nielsen, B. Obeli´c, C. Papucci, H. Pettersson, J. Paatero, T.-S. Park, A.P. McNichol, G. Polykarpov, W. Plastino, A. Priller, C.R. Quètel, G. Raisbeck, P. Rajec, J.-L. Rays, N. Reguigui, P. Roos, K. Rozanski, B. Salbu, M. Schubert, M. Schwaiger, F.W. Schwartz, S. Shima, K. Sihra, T. Stieglitz, A. Stolarz, R. Szymczak, Y. Tateda, H. Thebault, P. Theodorsson, I. Tolosa, O. Togawa, C. Tsabaris, D. Tsumune, S. Valulovsky, J. Vives i Battle, O. Voitsekhovitch, L. Vöröss, M. Warnau and E. Wyse. In publishing this book and the IAEA Proceedings we aim to make the use of isotopes more widespread in the environmental disciplines and to further stimulate work in this exciting field. P.P. Povinec, J.A. Sanchez-Cabeza, Editors IAEA-MEL Monaco
This page intentionally left blank
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
v
1. Environmental isotope tracers . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
1. Application of accelerator mass spectrometry to environmental and paleoclimate studies at the University of Arizona by A.J. Timothy Jull, George S. Burr, J. Warren Beck, Gregory W.L. Hodgins, Dana L. Biddulph, John Gann, Arthur L. Hatheway, Todd E. Lange and Nathaniel A. Lifton . . . . . . . .
3
2. Discriminating biogenic and anthropogenic chlorinated organic compounds using multi-isotope analyses of individual compounds by Kazushi Aranami, Steven J. Rowland and James W. Readman . . . . . . . . . . . . . . . . . .
24
3. Shift in stable water isotopes in precipitation in the Andean Amazon: Implications of deforestation or greenhouse impacts? by A. HendersonSellers and K. McGuffie . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
39
2. Oceanic radionuclide tracers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
51
4. Southern Hemisphere Ocean Tracer Study (SHOTS): An overview and preliminary results by M. Aoyama, M. Fukasawa, K. Hirose, R.F.C. Mantoura, P.P. Povinec, C.S. Kim and K. Komura . . . . . . . . . . . . . . . . . . . .
53
5. Plutonium isotopes in seawater of the North Pacific: Effects of close-in fallout by K. Hirose, M. Aoyama, C.S. Kim, C.K. Kim and P.P. Povinec . . .
67
6. Distribution of anthropogenic radionuclides in the water column off Rokkasho, Japan by Shigeki Shima, Shin-ichi Gasa, Ken-ichi Iseda, Tomoharu Nakayama and Hisao Kawamura . . . . . . . . . . . . . . . . . . . . . . . .
83
ix
x
Contents
7.
Artificial radionuclides in the Yellow Sea: Inputs and redistribution by G.H. Hong, C.S. Chung, S.-H. Lee, S.H. Kim, M. Baskaran, H.M. Lee, Y.I. Kim, D.B. Yang and C.K. Kim . . . . . . . . . . . . . . . . . . . . . . .
3. Radionuclides in the European seas . . . . . . . . . . . . . . . . . . . . . . . . . . 8.
9.
10.
96 135
Distribution of anthropogenic radionuclides in the water column of the south-western Mediterranean Sea by S.-H. Lee, F.R. Mantoura, P.P. Povinec, J.A. Sanchez-Cabeza, J.-L. Pontis, A. Mahjoub, A. Noureddine, M. Boulahdid, L. Chouba, M. Samaali and N. Reguigui . . . . . . . . . . .
137
Distribution of anthropogenic radionuclides in Moroccan coastal waters and sediments by M. Benmansour, A. Laissaoui, S. Benbrahim, M. Ibn Majah, A. Chafik and P.P. Povinec . . . . . . . . . . . . . . . . . . . . . . . . . . .
148
137 Cs
in seawater and sediment along the Algerian coast by A. Noureddine, M. Menacer, R. Boudjenoun, M. Benkrid, M. Boulahdid, M. Kadi-hanifi, S.-H. Lee and P.P. Povinec . . . . . . . . . . . . . . . . . . . . . . . . . . .
156
11. Physical and chemical characteristics of 137 Cs in the Baltic Sea by Galina Lujanien˙e, K˛estutis Jokšas, Beata Šilobritien˙e and Rasa Mork¯unien˙e . . .
165
4. Radioecological studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
181
12. Comparison of the MARINA II dispersion model with CSERAM for estimating concentrations of radionuclides in UK waters by Kamaljit Sihra, Antony Bexon and John Aldridge . . . . . . . . . . . . . . . . . . . . . . .
183
13. Assessment of the discharge of NORM to the North Sea from produced water by the Norwegian oil and gas industry by T. Gäfvert, I. Færevik and A.L. Rudjord . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
193
14. Uranium mining and ore processing in Ukraine – radioecological effects on the Dnipro River water ecosystem and human health by O. Voitsekhovitch, Y. Soroka and T. Lavrova . . . . . . . . . . . . . . . . . . . . . . . . . . . .
206
and 241 Am distributions in an alpine wetland, Boréon (France) by Maïa Schertz, Hervé Michel, Geneviève Barci-Funel and Vittorio Barci . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
215
16. Concentrations and characteristics of uranium isotopes in drinking waters collected in Italy and the Balkan regions and their radiological impact on the public by Guogang Jia, Giancarlo Torri, Umberto Sansone, Piera Innocenzi, Silvia Rosamilia, Antonio Di Lullo and Stefania Gaudino . . . . . . . . . .
223
15.
90 Sr, 137 Cs, 238 Pu, 239/240 Pu
Contents
17. The radiological evaluation of uranium, radium and radon in metallic and thermo-metallic springs in Ikaria Island, the eastern Aegean Sea, Greece by H. Florou, K. Kehagia, A. Savidou and G. Trabidou . . . . . . . . . . . . . 5. Isotope biomonitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xi
235 243
18. Bioaccumulation of radiocaesium in Arctic seals from Northeast Greenland by JoLynn Carroll, Kristina Rissanen and Tore Haug . . . . . . . . . . . . .
245
19. Anthropogenic radionuclides in biota samples from the Caspian Sea by J. Gastaud, B. Oregioni, S.V. Pagava, M.K. Pham and P.P. Povinec . . . . .
255
20.
21.
210 Po
in fish, algae, mussel and beach sediment samples collected along the Turkish coast of the Black Sea by Nurdan Güngör, Emin Güngör, B. Gül Göktepe and Güler Köksal . . . . . . . . . . . . . . . . . . . . . . . . . . .
265
210 Po in mussels (Mytilus galloprovincialis) and sediments along the Turkish
coast of the Aegean Sea by Aysun U˘gur, Güngör Yener, Sayhan Topcuo˘glu, U˘gur Sunlu, Serpil Aközcan and Banu Özden . . . . . . . . . . . . . . . . .
272
22. Stable nitrogen isotopes reveal weak dependence of trophic position of planktivorous fish on individual size: A consequence of omnivorism and mobility by Antonio Bode, Pablo Carrera and Carmela Porteiro . . . . . . . .
281
6. Isotope hydrology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
295
23. Radiocarbon loss from DIC in vadose water flow above the Judea Aquifer, Israel by Israel Carmi, Mariana Stiller, Joel Kronfeld, Yoseph Yechieli, Miriam Bar-Matthews, Avner Ayalon and Elisabetta Boaretto . . . . . . . . . . . .
297
24. Stable water isotopes as tools for basin-scale water cycle: Diagnosis of the Murray–Darling by A. Henderson-Sellers, P. Airey, K. McGuffie and D.J.M. Stone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
307
25. Isotopic characteristics of the Sava River basin in Slovenia by Nives Ogrinc, Tjaša Kanduˇc and Janja Vaupotiˇc . . . . . . . . . . . . . . . . . . . . . . . .
317
26.
222 Rn
as a tracer for the estimation of infiltration of surface waters into aquifers by M. Schubert, K. Knoeller, H.-C. Treutler, H. Weiss and J. Dehnert . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
326
27. Monitoring of geochemical and geophysical parameters in the Gran Sasso aquifer by Wolfango Plastino . . . . . . . . . . . . . . . . . . . . . . . . . .
335
xii
Contents
7. Groundwater–seawater interactions . . . . . . . . . . . . . . . . . . . . . . . . . .
343
28. Coastal water exchange rate studies at the southeastern Brazilian margin using Ra isotopes as tracers by Joselene de Oliveira, Mathew Charette, Mathew Allen, Elisabete de Santis Braga and Valdenir Veronese Furtado . . . . . .
345
29. Submarine groundwater discharge in the southeastern Mediterranean (Israel) by Y. Weinstein, G. Less, U. Kafri and B. Herut . . . . . . . . . . . . . . .
360
30. Submarine groundwater discharge investigations in Sicilian and Brazilian coastal waters using an underwater gamma-ray spectrometer by Pavel P. Povinec, Isabelle Levy-Palomo, Jean-Francois Comanducci, Joselene de Oliveira, Benjamino Oregioni and Agata M.G. Privitera . . . .
373
31. Isotope hydrochemical investigation of saline intrusion in the coastal aquifer of Karachi, Pakistan by A. Mashiatullah, R.M. Qureshi, M.A. Tasneem, T. Javed, C.B. Gaye, E. Ahmad and N. Ahmad . . . . . . . . . . . . . . . .
382
8. Coastal radionuclide studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
395
32. Temporal variations and behaviour of 90 Sr and 137 Cs in precipitation, river water and seawater in Japan by Yoshihiro Ikeuchi . . . . . . . . . . . . . .
397
33. Isotope fractionations and radiocarbon ages of beach rock samples collected from the Nansei Islands, southwest of Japan by Kunio Omoto . . . . . . .
406
34. A model of recent sedimentation in the Cananeia–Iguape estuary, Brazil by R.T. Saito, R.C.L. Figueira, M.G. Tessler and I.I.L. Cunha . . . . . . . . .
419
9. Modelling of environmental processes . . . . . . . . . . . . . . . . . . . . . . . .
431
35. Simulation of the advection–diffusion–scavenging processes for 137 Cs and 239,240 Pu in the Japan Sea by Masanao Nakano . . . . . . . . . . . . . . . .
433
36. A biokinetic model for the uptake and depuration of radioiodine by the edible periwinkle Littorina littorea by J. Vives i Batlle, R.C. Wilson, P. McDonald and T.G. Parker . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
449
37. Environmental modelling: Modified approach for compartmental models by M. Iosjpe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
463
38. Assessment of 137 Cs outspread in the Lithuanian part of the Baltic Sea by L. Davuliene, N. Tarasiuk, N. Spirkauskaite, G. Trinkunas and L. Valkunas
477
Contents
10. Radiometrics techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xiii
493
39. Low-level germanium gamma-ray spectrometry at the µBq/kg level and future developments towards higher sensitivity by G. Heusser, M. Laubenstein and H. Neder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
495
40. Depth profiles of environmental neutron fluxes in water and lead by Y. Hamajima and K. Komura . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
511
41. Radiocarbon measurement by liquid scintillation spectrometry at the Gran Sasso National Laboratory by Wolfango Plastino and Lauri Kaihola
520
42. Monte Carlo simulation of the muon-induced background of an antiCompton gamma-ray spectrometer placed in a surface and underground laboratory by Pavol Vojtyla and Pavel P. Povinec . . . . . . . . . . . . . . . . .
529
43. IAEA-MEL’s underground counting laboratory – The design and main characteristics by P.P. Povinec, J.-F. Comanducci, I. Levy-Palomo and F. Avaullee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
538
44. Levels of airborne radionuclides at Hegura Island, Japan by K. Komura, N. Muguntha Manikandan, Y. Yamaguchi, M. Inoue, T. Abe and Y. Murata
554
45. The use of sodium iodide detectors to locate buried radioactive particles in the seabed off Dounreay nuclear facility, Scotland by J. Toole, S.C. Innes, M. Liddiard, J. Cassidy and S. Russ . . . . . . . . . . . . . . . . . . . . . .
562
11. Mass spectrometry techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . .
579
46. Isotope selective ultratrace analysis of plutonium by resonance ionisation mass spectrometry by Stefan Bürger, Razvan Aurel Buda, Horst Geckeis, Gerhard Huber, Jens Volker Kratz, Peter Kunz, Christoph Lierse von Gostomski, Gerd Passler, Ariane Remmert and Norbert Trautmann . . . . . . .
581
47. Two 60-year records of 129 I from coral skeletons in the South Pacific Ocean by D.L. Biddulph, J.W. Beck, G.S. Burr and D.J. Donahue . . . . . . . . .
592
48. Factors influencing the determination of ultra low levels of Pu-isotopes by sector field ICP-MS by Per Roos . . . . . . . . . . . . . . . . . . . . . . . .
599
12. Management of data quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49. Recent IAEA reference materials and intercomparison exercises for radionuclides in the marine environment by M.K. Pham, J. Gastaud, J. La Rosa, S.-H. Lee, I. Levy-Palomo, B. Oregioni and P.P. Povinec . . . . . . . . . .
615
617
xiv
Contents
50. Towards quality excellence in radioanalytical laboratories at STUK, Finland by Tarja K. Ikäheimonen, Seppo Klemola and Pia Vesterbacka . . . . . . .
629
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
641
Keyword Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
643
1. Environmental isotope tracers
This page intentionally left blank
3
Application of accelerator mass spectrometry to environmental and paleoclimate studies at the University of Arizona A.J. Timothy Jull* , George S. Burr, J. Warren Beck, Gregory W.L. Hodgins, Dana L. Biddulph, John Gann, Arthur L. Hatheway, Todd E. Lange, Nathaniel A. Lifton University of Arizona, Tucson, AZ 85721, USA Abstract A wide range of climatic, geologic and archaeological records can be characterized by measuring their 14 C and 10 Be concentrations, using accelerator mass spectrometry (AMS). These records are found not only in the traditional sampling sites such as lake sediments and ice cores, but also in diverse natural records. The purpose of this paper is to highlight some selected applications of AMS at the University of Arizona, including sample preparation, applications of AMS radiocarbon dating to learning about climatic changes in the past, modern 14 C studies, and 10 Be and 129 I measurements. Keywords: Paleoclimate studies, Accelerator mass spectrometry, Radiocarbon dating, 14 C, 10 Be, 129 I
1. Introduction A firm chronology is an important key to the understanding of past climatic changes and their relationship to other events. To be able to correlate distinct climatic features, we must be able to correlate independently-dated events. Hence, the improvement in the radiocarbon calibration curve over the last 26,400 yr has allowed us to cross-correlate fluctuations in the 14 C curve with climatic fluctuations in such things as ice-core records. This capability has improved attempts to cross-correlate different climatic events observed in one record with other proxy records. This extension of the calibration curve used tree rings to about 11,500 calibrated years and beyond that used corals and varved marine sediments. Other newer but perhaps less-reliable records can take us back to the limits of radiocarbon dating, using lake sediments and speleothem records. * Corresponding author. Address: NSF Arizona AMS Laboratory, University of Arizona, Physics Building, 1118 East Fourth St, Tucson, AZ 85721, USA; phone: (+1) 520 621 6816; fax: (+1) 520 621 9619; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08001-0
© 2006 Elsevier Ltd. All rights reserved.
4
A.J.T. Jull et al.
We should also note that the same event might have a different manifestation or have a phase lag in different regions. Hence, it is also important that the underlying geochronology is sound. This is particularly true during the Glacial/Interglacial transition, which is of great interest due to the scale of climatic change at that time. During the Holocene, we also observe appreciable climatic fluctuations. These are less well understood, but may be associated with solar forcing (Damon and Sonnett, 1991). Other periodicities in the Holocene climatic record can often be related to solar fluctuations, the most obvious being in the medieval warm period and the Maunder minimum, periods associated with colder weather in Europe. There is a variety of literature on this subject. In recent years, millennial-scale periodicities (e.g. Bond et al., 1997; Alley et al., 2001) are recognized in a number of records, including varved lake sediments, loess deposits, marine records and forest-fire records. In this paper, we will highlight some climatic signals, which can be well dated using the small-sample capabilities of accelerator mass spectrometry (AMS). These signals can be seen not only in the climate record but affect the extinction of mega fauna as well as archeological events. We will also discuss the use of AMS for 10 Be and 129 I studies.
2. Improvements in the radiocarbon calibration The radiocarbon “calibration curve” was originally established by studying the changes in 14 C content of known-age tree rings. The first trees studied were Bristlecone Pines from the White Mountains of California. Tree ring 14 C measurements from living and dead wood from these long-lived trees were cross-correlated to establish a 14 C chronology longer than the life of a single tree (see Fig. 1). Dendrochronology, the science of correlating and cross-referencing variations in the widths of tree rings has been used to produce even longer chronologies using German and Irish Oak tree records, incorporating subfossil wood recovered from buried logs in river sediments and peat bogs. Currently, the continuous tree-ring sequence extends back 11,500 years. We know that the production rate of radiocarbon in the atmosphere changes with time and there are also changes in the amount of CO2 exchange with the oceans. Both these effects lead to changes in the “apparent” age, which we term the radiocarbon age of the sample. The radiocarbon age is defined simply from the amount of 14 C relative to “modern carbon”, defined as 1950 AD wood, where the industrial effect of reduced 14 C has been removed, effectively this is 1850 AD wood age-corrected back to 1950 AD. Using these assumptions, the radiocarbon age is easily computed as 14 C sample
Radiocarbon age = −8033 ln 14
Cmodern
.
There have been continuing improvements in the length of the radiocarbon time-scale. The improvement in the radiocarbon calibration curve over the last 26,000 yr has allowed us to cross-correlate fluctuations in the 14 C curve directly with those in the ice-core record. This capability has improved attempts to cross-correlate different climatic events observed in one record with other proxy records. This extension of the calibration curve used tree rings to about 11,500 calibrated years and beyond that used corals and varved marine sediments. There
Application of accelerator mass spectrometry
5
Fig. 1. Ancient Bristlecone pines from the White Mountains, California.
are also newer but as yet less well-established records which should take us back to the limits of radiocarbon dating, using lake sediments and speleothem records. An example of a section of the calibration curve is given in Fig. 2. The fluctuations in this curve demonstrate changes in the 14 C production rate and/or changes in the terrestrial carbon cycle have occurred. Intriguingly, it has often been noted that reversals (negative excursions to younger apparent age) are associated with cold climatic events and lowered carbon content of the atmosphere. This suggests a possibility of linkages to climate driven by carbon-cycle changes, such as might be produced by ocean circulation changes. They could also, of course, be due to a higher production rate. Excursions to higher 14 C age are likely due to reductions in production rate or increases in the ventilation of 14 C-depleted carbon from the deep ocean. Conversely, excursions to lower 14 C age are likely due to reductions in production rate or attenuated ocean ventilation rates. Periods of constant 14 C over an extended period of time can also be explained by reductions in production rate. These differences can be understood as the number of 14 C atoms in a given column of atmosphere, which can be approximated for times short compared to the half-life as N14 = P14 t,
6
A.J.T. Jull et al.
Fig. 2. Section of the tree-ring calibration curve. The radiocarbon age is given on the vertical axis and the calendar age on the horizontal axis.
where P is a complex function of production rate over the period the 14 C remains in the atmosphere (1–2 yr) and t is the (short) residence time of 14 C in the atmosphere. This approximation is valid since the residence time is very short compared to the mean life of 8,267 yr. Our modelling results suggest that the major features of this record cannot be produced with solar variability or terrestrial magnetic field modulation alone, but also require significant changes in the carbon cycle. This is also true of the substantial decline observed between 26–11 ka BP (from ∼700h to ∼100h), which is considered too large to be solely a result of reduced production from increased shielding by the Earth’s magnetic field. Superimposed on these large-scale trends are millennial and sub-millennial variations that apparently coincide with abrupt shifts in climate as recorded in the Greenland ice cores. Coral records offer the highest possible resolution for radiocarbon calibration beyond the limit of the tree-ring chronologies, because of their relatively fast growth rate. However, the existing radiocarbon calibration contains few continuously banded examples from corals. Two exceptions are a Diploastrea coral head from Vanuatu which lived during the Younger Dryas (Burr et al., 1998) and a Goniastrea coral from Papua New Guinea which lived approximately 13,000 years ago, according to U–Th dating. This subannual record (Fig. 3) is the highest resolution radiocarbon calibration record from that time period. Speleothems offer similar advantages to coral, with a time resolution of decades or better. One such record – obtained from a stalagmite recovered from a cave in the Bahamas – provides a nearly continuous record of atmospheric 14 C from 45 to 11 ka (Beck et al., 2001;
Application of accelerator mass spectrometry
7
Fig. 3. Radiocarbon age versus coral band number for a Goniastrea coral collected at 47 m depth, Huon Peninsula, Papua New Guinea. One band equals 6 months.
Richards and Beck, 2001). This record (Fig. 4), derived using TIMS U, Th and Pa measurements and AMS 14 C ages reveals highly elevated and extremely variable 14 C between 45 and 33 ka BP which appear to be correlative with peaks in cosmogenic 36 Cl and 10 Be isotopes (Baumgartner et al., 1998) observed in polar ice cores. 3. Accelerator Mass Spectrometry 3.1. AMS radionuclide methods Radionuclide measurement using Accelerator Mass Spectrometry (AMS) differs from the decay counting methods in that the amount of the radionuclide in the sample is measured directly, rather than waiting for individual radioactive decay events. This makes the technique 1,000 to 10,000 times more sensitive than decay counting for 14 C, for example. In the case of radionuclides such as 10 Be, where the low-energy β-rays were always difficult to count, AMS has improved the situation much further. This sensitivity is achieved by accelerating sample atoms as ions to high energies using a particle accelerator, and using nuclear particle detection techniques. A diagram of the 3 MV AMS machine at the University of Arizona is shown in Fig. 5. Today, an external precision of about ±0.35% in 14 C content, or ±30 years in uncalibrated radiocarbon age is possible on a single 0.5-milligram-sized sample target in 20 minutes of measurement time. Samples as small as 100 micrograms or less have been successfully dated to about ±80 years BP and even smaller samples (∼10–20 ng) have been measured for special experiments. With longer counting times or when multiple targets are measured, we can reduce the single target error to about 0.2%, or better than ±20 years in radiocarbon age. In the case of longer-lived radionuclides such as 26 Al, 10 Be, 36 Cl, 41 Ca and 129 I, which were very difficult to measure using radiometric techniques, AMS has made measurements
8
A.J.T. Jull et al.
(A)
(B) Fig. 4. (A) Radiocarbon calibration curve and variations in atmospheric 14 C based on stalagmite GB-89-24-1 from 45 to 11 ka, and comparison with several other 14 C records. (B) 14 C record for GB-89-24-1 from 26 to 10 ka.
Application of accelerator mass spectrometry
9
Fig. 5. Photograph of the high-energy section of the 3 MV National Electrostatics AMS machine at the University of Arizona.
of small amount of these radionuclides routine (Tuniz et al., 1998; Fifield, 1999; Jull et al., 2005). 3.1.1. AMS at Arizona The AMS systems at Arizona consist of the following basic components and sequence of events. Figure 5 shows a photograph of the high-energy end of our 3 MV NEC system. There are many good descriptions of the methodology of accelerator mass spectrometry (Tuniz et al., 1998; Fifield, 1999; Jull et al., 2003; Kutschera, 2005), so we refer the reader to these publications for a detailed discussion of the operation of an AMS system. 3.1.2. Conventions and definitions The radiocarbon age is determined from the ratio of 14 C/13 C or 14 C/12 C compared to “modern”, which is defined as the expected level in 1950 AD. 14 C ages are normalized to a constant value of δ 13 C, so any isotopic fractionation effects are removed from the age calculation. Due to industrial and nuclear effects, the 1950 AD value is computed from the corrected value of 1850 AD wood. 14 C age = −8033 ln(F ), as above, where F =
14 C/13 C sample 14 C/13 C modern
is the ratios of 14 C/13 C in the sample and modern carbon, respectively. In discussion of the calibration of radiocarbon age and other estimates of the “true age”, we often use 14 C (see Stuiver and Polach, 1977), which can be defined as the divergence of the
10
A.J.T. Jull et al.
production rate of 14 C at the time of production, compared to present-day production. Hence, 14 C = (Fm − 1)e−(cal BP/8267) × 1000h, where Fm is the 14 C activity, expressed as fraction of modern carbon (Donahue et al., 1990b), and cal BP is the “expected true age” of the sample. If the “expected true age” is defined by another dating method, whether it be U–Th measurements, varve counting, dendrochonology, or correlation of a sediment record with ice-core records, it might be subject to other systematic errors. Hence, records based on different estimates of age or with potential reservoir effects need to be taken with caution. Beyond this time we still observe large fluctuations in the 14 C record. We do not know the cause of these fluctuations, however large changes in cosmic-ray flux, supernova events and large-scale changes in ocean circulation have been hypothesized (cf. Beck et al., 2001). Intriguingly, some of these 14 C “spikes” are also observed in the records of cosmogenic 10 Be in ice and marine sediments (Raisbeck et al., 1990; McHargue et al., 1995, 2000), and 129 I in a speleothem (see Biddulph et al., 2006, p. 612, this volume). 3.2. Chemical preparation of samples and stable-isotope analysis A very important factor in AMS measurements is proper sample preparation. It is important that the sample is clean and all contaminants have been removed. Some basic procedures for AMS radiocarbon samples at Arizona were summarized by Jull et al. (2004a): 1. Acid-base-acid method for charcoal, wood, cellulose, plant material, animal tissue: After physical inspection, samples are cleaned with 1 N HCl acid, 0.1% NaOH and 1 N HCl (acid-base-acid (ABA) pretreatment), washed with distilled water, dried, and combusted at 900◦ C with CuO. Hatté et al. (2001) discussed some modifications and potential problems with the acid-base-acid method. 2. Carbonates: In general, samples are etched with 100% H3 PO4 to remove 50–85% of the carbonate, dried and hydrolyzed with H3 PO4 as discussed by Burr et al. (1998). 3. Selective combustion for sediments. After cleaning in 1 N HCl and drying, the sample is combusted at 400◦ C in ∼0.3 atm. oxygen gas (McGeehin et al., 2001). 4. Oxidative acid cleaning for old charcoal: We have constructed a new line to clean charcoal samples using the oxidative acid method of Bird et al. (1999). This is of particular interest for charcoal samples >20 kyr. We have also applied this method to sediments from archaeological settings. 5. Textiles, parchment, canvas, art works and artifacts: The samples are given the ABA pretreatment and after washing and drying, they are Soxhlet extracted with hexane, then ethanol and finally methanol. After washing in distilled water, and drying, they are combusted at 900◦ C with CuO. Bruhn et al. (2001) proposed a more complex, 6-stage cleaning procedure for art works. Of course, for some specialized samples, other methods need to be developed for these applications. After cleaning and combustion, samples are converted to graphite using a modification of the method of Slota et al. (1987), using an Fe catalyst and Zn as the reducing agent for the reaction 2CO → CO2 + C.
Application of accelerator mass spectrometry
11
Finally, the graphite powder is pressed into a target holder of Arizona design, which is now widely used in all NEC ion sources. The measurement of 14 C follows the procedures and calculations described in detail by Donahue et al. (1990a, 1990b). Further discussion on the question of interlab differences in calculations were highlighted by McGeehin et al. (2001).
4. Sample processing improvements 4.1. Automated sample pretreatment We have introduced a number of automated or semi-automated techniques in our laboratory. We have constructed a continuous flow apparatus for the routine chemical treatment (acid– base–acid) of samples. This device allows us to process samples with successive chemical reagents using computer control. We have set up a Carlo Erba CHN analyzer and we have interfaced the CHN analyzer with a stable isotope ratio mass spectrometer. Among other things, this system now allows us to measure elemental CN ratios and δ 15 N of bones, as indicators of bone preservation and dietary input. These instruments will be linked to an automated CO2 collection system. The objective is to provide the capability for automated sample combustion, elemental analysis, stable isotope analysis, carbon dioxide purification and sample collection. 4.2. Specialized sediment sample pretreatment We have developed a new method for sediment dating which combines physical and chemical pretreatments, followed by stepped combustion at different temperatures (McGeehin et al., 2001, 2004). Inherited clay- or silt-bound carbon can be an important contaminant for conventional bulk sediment analysis. In some sediments, the humin fraction (material resistant to the acid and base pretreatment) combusting at 400◦ C (“low temperature fraction”) appears promising for dating sediment. In some sediments, the humin fraction could contain pollen, macrofossils, wood or charcoal. We are applying this method to a wide range of sedimentary environments. For low organic concentration sediments from the last glacial maximum age, the low temperature fraction can be 2–4 ka younger than the bulk fraction (Fig. 6). Results from this work assist in understanding the relationship between archaeological records and paleoclimatic records for particular sites. We plan to refine the stepped combustion method by improving our understanding of temperature oxidation thresholds and by characterizing the organic materials in each temperature fraction. Another sediment dating technique which we are studying is designed to remove contaminants left after a conventional acid–base–acid pretreatment. Preliminary results using a dichromate-wet oxidation (1 M dichromate plus 1 M sulfuric acid) method (ABOX), initially developed as a way to purify old charcoals (Bird et al., 1999) show significant reductions in the amount of total organic carbon in sediments. The remaining “oxidation-resistant” components are likely to be composed of charcoal or soot, as opposed to other forms of organic material in soils (Bird et al., 1999).
12
A.J.T. Jull et al.
Fig. 6. Deviation of the radiocarbon ages measured in various humin and humic acid fractions, compared to the age of plant macrofossil remains in the same sediment stratum (from McGeehin et al., 2001).
4.3. Compound-specific studies and bone chemistry We have also begun several new techniques and strategies in our laboratory for extracting carbon from highly contaminated materials that are difficult or impossible to obtain accurate radiocarbon dates using standard methods. Among these is a method for releasing carbon from bone-specific proteins using ninhydrin (Nelson, 1991; Tisnérat-Laborde et al., 2003; Hodgins and Jull, 2004). This approach of selective collagen dissolution using ninhydrin is being applied to bones found in tar pits (see Jull et al., 2002; Hodgins and Jull, 2004). These environments are highly preservative, and thus rich repositories of animal and plant remains. However, the tar hydrocarbons that thoroughly contaminate the samples present large challenges for dating. 4.4. Surface cleaning and sample extraction by oxygen plasma We have begun a research program investigating the plasma oxidation methods used by Rowe (e.g. Rowe, 2001) to extract organic carbon from organic residues on geological carbonates. This method has to date been used for 14 C dating of rock art pictographs. We anticipate that the low temperature oxidation reactions accomplished by the plasma will have applications on other types of samples.
5. Can we observe climatic signals in the radiocarbon record? It has long been recognized that the fluctuations in the 14 C record were often coincident with major climatic events, particularly cold ones. A period of “apparent” radiocarbon age which is flat is observed at various major cold phenomena such as the Oldest Dryas (15.1–14.5 cal BP), Younger Dryas (12.9–11.6 cal BP), the 8.2 ka “cold event”, 2.5 ka and the Maunder Minimum (17–19th centuries). This can be observed in a plot of 14 C age (yr BP) vs calendar years for the glacial–interglacial transition, for example. Several of these events are observed in Fig. 4. Many authors have commented on these effects. It is well beyond the scope of this paper to present all the possible examples of these types of studies, so we will summarize a few areas of research which highlight the use of radiocarbon in the climatic record and related areas.
Application of accelerator mass spectrometry
13
5.1. Use of 14 C in modern corals and as a tracer in the ocean The in-growth of bomb 14 C into the oceans has been studied since the early GEOSEC cruises of the 1970s. 14 C variations in dissolved inorganic carbon in seawater were used to follow the different rates of carbon uptake from different parts of the world. These experiments were repeated in the 1980s and 1990s as part of the WOCE program. Both of these studies represent the distribution of radiocarbon at the time the samples were collected. Corals, which grow in the surface waters of the ocean, have since been the object of extensive study, because they preserve a record of the radiocarbon content of the surface ocean which allows temporal variations at a particular site to be studied. In addition to 14 C variations, modern corals also preserve a record of other geochemical proxies such as Sr–Ca and δ 18 O. This signal can also be correlated with the intensity of the El-Niño Southern Oscillation. Scientists at Arizona have continued their high frequency bomb pulse radiocarbon measurements of modern corals as part of their effort to generate time-slice maps of Pacific Ocean radiocarbon variations over the last 60 years. We have completed measurements on a drill core from Guadalcanal (Schmidt et al., 2004), Easter Island and the Marquesas, begun work on a record from Wallis Island, and have collected cores from Tokelau and Kiribati in collaboration with colleagues Thierry Correge of IRD (Noumea), and Julia Cole (Arizona). They have begun modelling efforts to combine these and other published and unpublished bomb pulse records (Fig. 7) with the WOCE and GEOSECS 14 C ocean surveys using a horizontal advective flux model based on multi-decadal wind stress records. Beck has also been involved with colleagues at ANU and Stanford University, to generate a record of bomb pulse 14 C variations from the Eastern Indian Ocean. This record was used in concert with another record from the Western Indian Ocean (Grumet et al., 2002) to generate an index of E–W 14 C for this period (Grumet et al., 2004). There have also been extensive studies of the 14 C record in the oceans, as part of the World Ocean Circulation Experiment (WOCE), Joint Global Ocean Flux Study (JGOFS) and other large international programs. 5.2. Forest fires There is a considerable record of forest-fire history from different regions of the world (see, e.g., Meyer et al., 1995, 2001; Turcq et al., 1998; Hallett and Walker, 2000; Pierce et al., 2004). A characteristic of many of these studies is evidence for marked periodicities, especially on century and millennial time scales of fire frequency. Typically, these studies involve either direct dating of charcoal found in alluvial and colluvial deposits (e.g. Meyer et al., 1995, 2001; Pierce et al., 2004) or studies of the charcoal record found in lake sediments (e.g. Long et al., 1998; Cumming et al., 2002). Meyer et al. (1995) demonstrated a spectacular record of forest fires from the Yellowstone National Park, which is now cited in all subsequent papers. In this record, Meyer et al. (1995) noted millennial-scale forcing and proposed that “Bond” cycles might be a forcing mechanism. Later studies showed similar periodicities, but a different phase relation at a location in southern Idaho (Pierce et al., 2004). Recently, Jull and Geertsema reported on radiocarbon dating of charcoal from paleosols and buried charcoal horizons in a unique sequence which potentially records the last 36,000
14
A.J.T. Jull et al.
Fig. 7. Location map of coral 14 C bomb pulse records to be used for genesis of 4-D Pacific Ocean radiocarbon variations. 14 C
years from a fan at Bear Flat, British Columbia (Fig. 8). This site included evidence for forest fire charcoal found over the last ca. 13,500 radiocarbon yr before present (yr BP) or 16,250 ± 700 calibrated yr BP. The latest evidence of fire is during the Medieval Warm Period. The charcoal ages show a periodicity in large fires on a millennial scale through the Holocene – an average of 4 fires per thousand years. Fire frequency is higher about 2,500–3,000 cal yr BP (10–11 fires/ka), ∼5,500 cal yr BP (∼5 fires/ka), ∼8,000 cal yr BP (∼8 fires/ka), and 9,000–10,000 cal yr (∼6 fires/ka). These authors concluded that fire frequency was driven by regional or global climate, as well as local phenomena. Over 50 discrete fire-related horizons have been observed. These charcoal ages show a periodicity in large fires on a millennial scale through the Holocene – an average of 4 fires per thousand years. Fire frequency is higher about 2,000–2,500 14 C yr BP (10–11 fires/ka), ∼5,000 14 C yr BP (∼5 fires/ka), ∼7,000 14 C yr BP (∼8 fires/ka), and 8,000 to 9,000 14 C yr (∼6 fires/ka). These intervals are also times of aboveaverage aggradation of the fan. Fire frequency appears to be related to climate. Recently, Alley noted that forest-fire frequency increases during North Atlantic cold events. In particular, times of higher fire frequency may follow North Atlantic ice-rafting events. Other studies in central BC have been undertaken with colleagues from the University of Northern British Columbia from sites in a wetter region of central British Columbia. A parallel study has been conducted in collaboration with colleagues at the University of New Mexico on the forest-fire history in Idaho, this record showing similar periodicities, but
Application of accelerator mass spectrometry
15
Fig. 8. Field photograph of some charcoal-rich horizons related to forest-fire debris flows at Bear Flat, British Columbia. The cyclicity of the forest fires is evident in the photograph.
with a different phase relation to the fires in British Columbia and Yellowstone Park (Pierce et al., 2004).
6.
10 Be
studies
We now routinely make measurements of the cosmogenic-radionuclide 10 Be with the new 3MV NEC AMS. 10 Be is produced in the atmosphere by spallation of oxygen and nitrogen by cosmic rays. The intensity of the cosmic-ray flux depends on galactic and solar sources, and modulation by the heliomagnetic and geomagnetic fields. After formation, 10 Be is quickly removed from the atmosphere by precipitation and deposited onto the surface of the Earth, where it is transported throughout the ocean and is eventually sequestered within marine sediments. A record of the cosmic-ray flux, modified by marine processes, may be interpreted from marine sediment cores and provides a valuable record of past geomagnetic and cosmicray phenomena. Our previous work documented the 10 Be record in the authigenic fraction of marine sediments from the Gulf of California, Leg 64, site 480 (McHargue et al., 1995), and the Blake Ridge, CH88-10P (McHargue et al., 2000). During the period of EAR01-15488, we completed work on another core from the Blake Outer Ridge, Leg 172, site 1061. The 10 Be data, normalized to the mass of the authigenic sediment fraction, is shown in Fig. 9. Normalization of 10 Be to the mass of the authigenic fraction provides a better proxy of paleomagnetic intensity in this core than normalization to the bulk sample mass or to 9 Be. In addition, normalizing 10 Be to the authigenic fraction provides good correlation with 10 Be from cores CH88-10P and DSDP 480 normalized in the same manner (not shown). This normalization makes the 10 Be increases at approximately 60,000 years (observed in polar ice), and at the Laschamp geomagnetic excursion and possible Mono Lake excursion more apparent. Future proposed
16
A.J.T. Jull et al.
Fig. 9. 10 Be results for the Blake Outer Ridge, DSDP Leg 172, site 1061 core. At the bottom of the figure 10 Be is normalized to the original sample mass. In the middle figure, it is normalized to 9 Be. In the top figure, 10 Be is normalized to the mass of the original sample removed by leaching, that is, the authigenic fraction of the sediment. Adapted from McHargue et al. (2000) and unpublished data.
work will revisit a core from the Gulf of California, in which the sediments are varved and the sedimentation rate is three to four times higher than that on the Blake Ridge. Of particular interest is a detailed analysis of an extreme short-term (50 years or less) 10 Be anomaly associated with the Mono Lake excursion in the Gulf of California (Leg 64, site 480), which is too short to be observable in Blake Ridge sediments. 7.
129 I
studies
Since the installation of the new NEC 3MV Pelletron accelerator we have measured more than 500 129 I samples. Sufficient data are now available to show a machine random error of
Application of accelerator mass spectrometry
17
0.4% for iodine samples. Repeated analysis of the low-level Woodward iodine standard has yielded results in the low 10−14 range for 129 I/127 I ratios. Chemical extraction techniques for a variety of environmental samples including seawater, corals, stalagmites and seaweed have been developed. With these techniques we can measure 129 I/127 I ratios in samples with as little as 5 µg total iodine content. We have recently produced an 84-year record showing anthropogenic iodine invasion into corals in the South Pacific, and a 20,000-year record of cosmogenically produced iodine in a stalagmite from the Bahamas islands (Biddulph, 2004; Biddulph et al., 2006, p. 612, this volume). We are collaborating with IAEA and other laboratories to produce iodine standard reference materials. Recent measurements of 129 I concentrations in fish from the North and Irish Seas are in good agreement with other laboratories. We are currently in the process of measuring other IAEA samples, including IAEA produced silver iodide and Mediterranean seawater. We have plans to continue radioiodine research in stalagmites and corals. There is some preliminary data suggesting that 129 I may track changes in the geomagnetic field and the cosmicray flux. Data from a stalagmite (Biddulph et al., 2006, this volume) seems to reproduce magnetic excursions that have been seen in other radioisotope records such as 10 Be and 36 Cl. The relatively long half-life of 129 I (15.7 million years) would enable us to analyze records that go back nearly 100 million years.
8. Megafaunal extinctions Radiocarbon dates were obtained by accelerator mass spectrometry on bones of extinct large mammals recovered from tar pits. Results on some samples of Glyptodon and Holmesina (extinct large mammals similar to armadillos) yielded results of >25 ka and >21 ka, respectively. We also studied the radiocarbon ages of 3 different samples of bones from the extinct Cuban ground sloth, Parocnus brownii, which yielded dates ranging from 4,960 ± 280 to 11,880 ± 420 yr BP. In order to remove the tar component the samples were cleaned by Soxhlet extraction in benzene. A report of our studies was presented at the 9th International AMS Conference (Jull et al., 2004b). 14 C dating of a tooth from a Madagascar Hippopotamus confirms that this member of the extinct “prehistoric” megafauna actually survived until well after European colonization. Gelatinized collagen subsamples independently dated by two AMS facilities yielded ages of 99±36 and 214±40 yr BP. Calibrated at 2σ , these dates give a range of 1639–1950 cal yr BP, placing a member of this extinct megafauna securely in colonial times. Dates on other extinct taxa show that many other prehistoric megafauna broadly overlapped the human presence on the island and may have survived until colonial times as well. A summary of all the available dates has been discussed by Burney et al. (2004). We continue to locate and try to date fossils from extinct megafauna, particularly from isolated or island refugia. In collaboration with Paul Martin (Arizona) and David Burney (Fordham University), we have located samples of extinct faunal remains from Madagascar, Cuba and other island refugia, as well as North American sites (Fig. 10). Megafaunal extinctions based on the effects of human expansion have been discussed for many other parts of the world, such as Australia, New Zealand, Madagascar, and the Americas.
18
A.J.T. Jull et al.
Fig. 10. Major Pleistocene extinction events on the continents and larger islands, excluding Afro-Asia. Patterns indicate timing of extinctions for each region (adapted from Martin and Klein, 1984).
Apart from some isolated evidence for earlier settlement with good associations, the predominant view is that early man arrived in the western hemisphere close to the end of the last Glacial (see Nuñez et al., 1994; Martin and Klein, 1984; Meltzer et al., 1997). The conventional model assumes that early humans arrived in the new world via a Bering land bridge. We note that the Bering Strait is only 30 m deep at its shallowest point and we can assume that the last sea-level rise would have closed off this route between Asia and the Americas. The Bering land bridge should have remained intact until ∼10,000 radiocarbon years BP (11,000 calendar years). We can understand this in terms of the sea-level rise history, which occurred in two stages of ∼60 m and ∼50 m, as shown by Bard et al. (1996) and Edwards et al. (1993). The rapid expansion of early man into central North America does not appear to have occurred until about 12,000 radiocarbon years BP, although the dating of the Monte Verde site in Chile (∼12,500 14 C yr BP, Meltzer et al., 1997) suggests a slightly older time. This period appears to be during the “Older Dryas”, which is observed in many records ∼12,200 14 C yr BP (e.g. Goslar et al., 2000). Dyke et al. (2001) have summarized the available radiocarbon records for the margins of the Laurentide ice sheet, which indicate that regions east of the Rocky Mountains were ice-free at that time. In addition, the results of Jull and Geertsema and others suggest that much of this region may have been ice-free even earlier. The extinction of many megafauna at the end of the late Pleistocene is well known (Martin and Klein, 1984). Hence, the exact time of these extinctions, and whether they are caused by climate alone, or by a combination of factors, such as the expansion of humans into previously unoccupied areas is a matter of great interest. Indeed, neither of these factors alone seems to explain all the observations. Taylor has summarized the earliest radiocarbon dates on human bones from the new world. In all cases, these dates are close to the dates for the last evidence for mammoths and other large megafauna (e.g. Martin and Klein, 1984). About one half of all species of large land
Application of accelerator mass spectrometry
19
mammals of North America disappeared at the close of the Pleistocene (Martin and Klein, 1984). At some locations, we can observe the interplay of climate, early man and megafauna. The Murray Springs site in Arizona, as discussed above, is only one of many which shows the interplay of climate, the expansion of humans and the disappearance of large mammals such as mammoths. Radiocarbon ages on algal mats, which overly the mammoth remains, also fall into the period of the Younger Dryas cold event at 10,300–10,600 yr BP. This evidence suggests that the Clovis expansion occurred during the European Allerød warm period, and that we can perhaps associate the Clovis “drought” of Haynes (1991) with the Intra-Allerød cold period (IACP) which occurred about 11,400 to 11,100 14 C yr BP and was followed by a short warmer epoch before the rapid onset of the Younger Dryas at 10,900 14 C yr BP. In addition, we can also compare these radiocarbon ages to the estimates of sea-level rise determined by Bard et al. (1996) in Barbados coral and Edwards et al. (1993) in the South Pacific. In either case, the algal mat deposits date to 10.2–10.6 ka, within the Younger Dryas epoch. This cold event post-dates Clovis, and all Clovis radiocarbon measurements (see Haynes, 1984, 1991, 1992) fall between these two sea-level rise events.
9. In situ terrestrial cosmogenic 14 C We have developed an extraction method for in situ 14 C from quartz, which overcomes numerous problems encountered using previous methods (Lifton et al., 2001). Degassed lithium metaborate (LiBO2 ) is used as a flux to dissolve the quartz sample at 1100–1200◦ C inside a mullite furnace tube. These temperatures allow the use of significantly cheaper tube furnaces than previous techniques. Furthermore, blank levels of in situ 14 C are low and quite stable ((1.3 ± 0.2) × 105 14 C atoms, 2σ ). Using this technique, quartz separates from eight samples of wave-cut quartzite benches from the well-dated Bonneville (17.4 ± 0.3 cal kyr) and Provo (16.8 ± 0.3 cal kyr) shorelines of Pleistocene Lake Bonneville, Utah, and from underlying deeply shielded locations were analyzed. Replicate analyses of five aliquots of a single sample demonstrate analytical precision better than 4% (2σ ). The precision we have attained allows measurement of in situ 14 C in a 5 g quartz sample after only ∼500 years of exposure at sea level and high latitude. Current work builds on these advances by addressing basic questions regarding in situ 14 C production and attenuation characteristics, and by applying in situ 14 C to fundamental problems in cosmogenic nuclide research. The Lake Bonneville shorelines will be the focus of the international CRONUS-Earth program and the first “intercomparison” site. We have completed field work in which we collected samples from surfaces at secular equilibrium along mid- and low-latitude altitude transects to assess the altitudinal and latitudinal dependence of integrated late Quaternary in situ 14 C production rates. We also have received samples from other investigators from Namibia, Antarctica, Australia and New Zealand to better constrain in situ 14 C production rate scaling globally. Results to date from these transects (ranging from sea level to nearly 4 km altitude) confirm the viability of using in situ 14 C in saturated surfaces to constrain in situ cosmogenic nuclide production rate scaling models (Lifton et al., 2002).
20
A.J.T. Jull et al.
10. Conclusion We have presented some applications of accelerator mass spectrometry which have applications to paleoclimate and environmental studies using the Arizona AMS laboratory. There are many varied uses of AMS and this paper only highlights a few chosen examples. We hope that this overview is useful to the non-AMS specialist and will allow the reader some insight into the breadth of applications of AMS.
Acknowledgements The authors are grateful for the technical support of the staff of the NSF Arizona AMS Laboratory and for support from NSF Grant EAR01-15488.
References Alley, R.B., Anandakrishnan, S., Jung, P., Clough, A. (2001). Stochastic resonance in the North Atlantic: Further insights. In: Seidov, D., Haupt, B.J., Maslin, M. (Eds.), The Oceans and Rapid Climate Change: Past, Present and Future. American Geophysical Union, Washington, DC, pp. 57–68. Bard, E., Hamelin, B., Arnold, M., Montaggioni, L., Cabioch, G., Faure, G., Rougerie, F. (1996). Deglacial sea-level record from Tahiti corals and the timing of global melt water discharge. Nature 382, 241–244. Baumgartner, S., Beer, J., Masarik, J., Wagner, G., Meynadier, L., Synal, H.A. (1998). Geomagnetic modulation of the 36 Cl flux in the GRIP ice core, Greenland. Science 279, 1330–1332. Beck, J.W., Richards, D.A., Edwards, R.L., Silverman, B.W., Smart, P.L., Donahue, D.J., Herrera Osterheld, S., Burr, G.S., Calsoyas, L., Jull, A.J.T., Biddulph, D. (2001). Extremely large variations of atmospheric 14 C concentration during the last Glacial period. Science 292, 2453–2458. Biddulph, D.A. (2004). Ph.D. thesis, University of Arizona. Biddulph, D.A., Beck, J.W., Burr, G.S., Donahue, D.J. (2006). Two 60-year records of 129 I from coral skeletons in the South Pacific Ocean. In: Povinec, P.P., Sanchez-Cabeza, J.A. (Eds.), Radionuclides in the Environment, International Conference on Isotopes in Environmental Studies: Aquatic Forum 2004. 25–29 October, Monaco. Elsevier, Amsterdam, pp. 612–618, this volume. Bird, M.I., Ayliffe, L.K., Fifield, L.K., Turney, C.S.M., Cresswell, R.G., Barrows, T.T., David, B. (1999). Radiocarbon dating of “old” charcoal using a wet oxidation, stepped-combustion procedure. Radiocarbon 41, 127–140. Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P., de Menocal, P., Priore, P., Cullen, H., Hajdas, I., Bonani, G. (1997). A pervasive millennial-scale cycle in North Atlantic Holocene and glacial climates. Science 278, 1257–1266. Bruhn, F., Duhr, A., Grooter, P.M., Minitrop, A., Nadeau, M. (2001). Chemical removal of conservation substances by Soxhlet-type extraction. Radiocarbon 43, 229–237. Burney, D.A., Burney, L.P., Godfrey, L.R., Jungers, W.L., Goodman, S.M., Wright, H.T., Jull, A.J.T. (2004). A chronology for late prehistoric Madagascar. Journal of Human Evolution 47, 25–63. Burr, G.S., Beck, J.W., Taylor, F.W., Récy, J., Edwards, R.L., Cabioch, G., Corrège, T., Donahue, D.J., O’Malley, J.M. (1998). High resolution radiocarbon calibration between 11.7 and 12.4 kyr BP derived from 230 Th ages of corals from Espiritu Santo Island, Vanuatu. Radiocarbon 40, 1085–1092. Cumming, B.F., Laird, K.R., Bennett, J.R., Smol, J.P., Salomon, A.K. (2002). Persistent millennial-scale shifts in moisture regimes in western Canada during the last six millennia. Proceedings of the National Academy of Sciences, USA 99, 16117–16121. Damon, P.E., Sonnett, C.P. (1991). In: Sonnett, C.P., Giampapa, M.S., Matthews, M.S. (Eds.), The Sun in Time. The University of Arizona Press, Tucson, AZ, pp. 361–388. Donahue, D.J., Jull, A.J.T., Linick, T.W., Toolin, L.J. (1990a). Radiocarbon measurements at the University of Arizona AMS Facility. Nuclear Instruments and Methods in Physics Research B 52, 224–228.
Application of accelerator mass spectrometry
21
Donahue, D.J., Linick, T.W., Jull, A.J.T. (1990b). Isotope-ratio and background corrections for accelerator mass spectrometry radiocarbon measurements. Radiocarbon 32, 135–142. Dyke, A.S., Andrews, J.T., Clark, P.U., England, J., Miller, G.H., Shaw, J., Veillette, J.J. (2001). Radiocarbon dates pertinent to defining the last glacial maximum for the Laurentide and Innuitian ice sheets, Geological Survey of Canada, Open File 4120, 54 pp. Edwards, R.L., Beck, J.W., Burr, G.S., Donahue, D.J., Chappell, J.M.A., Bloom, A.L., Druffel, E.R.M., Taylor, F.W. (1993). A large drop in atmospheric C-14/C-12 and reduced melting in the Younger Dryas, documented with Th-230 ages of corals. Science 260, 962–968. Fifield, L.K. (1999). Accelerator mass spectrometry and its applications. Reports of Progress in Physics 62, 1223– 1274. Goslar, T., Arnold, M., Tisnérnat-Laborde, N., Hatté, C., Paterne, M., Ralska-Jasiewiczowa, M. (2000). Radiocarbon calibration by means of varves versus 14 C ages of terrestrial macrofossils from Lake Go´sciaz and Lake Perespilno, Poland. Radiocarbon 42, 403–414. Grumet, N.S., Guilderson, T.P., Dunbar, R.B. (2002). Meridional transport in the Indian Ocean traced by coral radiocarbon. Journal of Marine Research 60, 725–742. Grumet, N.S., Abram, N.J., Beck, J.W., Dunbar, R.B., Gagan, M.K., Guilderson, T.P., Hantoro, W.S., Suwargadi, B.W. (2004). Coral radiocarbon records of Indian Ocean water mass mixing and wind-induced upwelling along the coast of Sumatra, Indonesia. Journal of Geophysical Research – Oceans 109, C05003. Hallett, D.J., Walker, R.C. (2000). Paleoecology and its application to fire and vegetation management in Kootenay National Park, British Columbia. Journal of Paleolimnology 24, 401–414. Hatté, C., Morvan, J., Noury, C., Paterne, M. (2001). Is classical acid-alkali-acid treatment responsible for contamination? An alternative proposition. Radiocarbon 43, 177–182. Haynes, C.V. (1984). Stratigraphy and Late Pleistocene extinctions in the United States. In: Martin, P.S., Klein, R.G. (Eds.), Quaternary Extinctions. University of Arizona Press, Tucson, AZ. Haynes, C.V. (1991). Geoarchaeological and paleohydrological evidence for a Clovis-age drought in North America and its bearing on extinction. Quaternary Research 35, 438–450. Haynes, C.V. (1992). Contributions of radiocarbon dating to the geochronology of the peopling of the New World. In: Taylor, R.E., Long, A., Kra, R.S. (Eds.), Radiocarbon After Four Decades. Springer-Verlag, New York. Hodgins, G.W.L., Jull, A.J.T. (2004). Radiocarbon dating of petroleum-impregnated bone from Tar Pits using the ninhydrin reaction. Abstract, 34th International Symposium on Archaeometry, 3–7 May, 2004, Zaragosa, Spain.. Jull, A.J.T., Burr, G.S. (2005). Accelerator Mass Spectrometry: Is the future bigger or smaller? Earth and Planetary Science Letters, submitted for publication. Jull, A.J.T., Donahue, D.J., Burr, G.S., Beck, J.W., McHargue, L.R., Hatheway, A.L., Lange, T.E., O’Malley, J.M., Biddulph, D. (2002). In: Aggarwal, S.K., Alamelu, D. (Eds.), Tenth ISMAS Workshop on Mass Spectrometry. Indian Society for Mass Spectrometry, Mumbai, India, pp. 25–34. Jull, A.J.T., Burr, G.S., Beck, J.W., Donahue, D.J., Biddulph, D., Hatheway, A.L., Lange, T.E., McHargue, L.R. (2003). Accelerator mass spectrometry at Arizona: Geochronology of the climate record and connections with the ocean. Journal of Environmental Radioactivity 69, 3–19. Jull, A.J.T., Burr, G.S., McHargue, L.R., Lange, T.E., Lifton, N.A., Beck, J.W., Donahue, D.J. (2004a). New frontiers in dating of geological, paleoclimatic and anthropological applications using accelerator mass spectrometric measurements of 14 C and 10 Be in diverse samples. Global and Planetary Change 41, 309–323. Jull, A.J.T., Itturalde-Vinent, M., O’Malley, J.M., McPhee, R.D.E., McDonald, H.G., Martin, P.S., Moody, J., Rincón, A. (2004b). Nuclear Instrument and Methods in Physics Research B 223, 668–671. Kutschera, W. (2005). Progress in isotope analysis at ultra-trace level by AMS. International Journal of Mass Spectrometry 242, 145–160. Lifton, N.A., Jull, A.J.T., Quade, J. (2001). A new extraction technique and production rate estimate for in situ cosmogenic 14 C in quartz. Geochimica et Cosmochimica Acta 65, 1953–1969. Lifton, N., Pigati, J., Jull, A.J.T., Quade, J. (2002). Altitudinal variation of in situ cosmogenic C-14 production rates: Preliminary results from the Southwestern US. Geochimica et Cosmochimica Acta 66 (15A), A457. Long, C.J., Whitlock, C., Bartlein, P.J., Millspaugh, S.H. (1998). A 9000-year fire history from the Oregon Coast Range, based on a high-resolution charcoal study. Canadian Journal of Forest Research 28, 774–787. Martin, P.S., Klein, R.G. (1984). Quaternary Extinctions: A Prehistoric Revolution. University of Arizona Press, Tucson, AZ.
22
A.J.T. Jull et al.
McGeehin, J., Burr, G.S., Jull, A.J.T., Reines, D., Gosse, J., Davis, P.T., Muhs, D., Southon, J. (2001). Steppedcombustion 14 C dating of sediment. Radiocarbon 43 (2A), 255–262. McGeehin, J., Burr, G.S., Hodgins, G., Bennett, S.J., Robbins, J.A., Morehead, N., Markewich, H. (2004). Steppedcombustion 14 C dating of bomb carbon in lake sediment. Radiocarbon 46, 893–900. McHargue, L.R., Damon, P.E., Donahue, D.J. (1995). Enhanced cosmic-ray production of 10 Be coincident with the Mono Lake and Laschamp geomagnetic excursions. Geophysical Research Letters 22, 659–662. McHargue, L.R., Donahue, D.J., Damon, P.E., Sonett, C.P., Biddulph, D., Burr, G.S. (2000). Geomagnetic modulation of the late Pleistocene cosmic-ray flux as determined by 10 Be from Blake Outer Ridge sediments. Nuclear Instruments and Methods in Physic Research B 172, 555–561. Meltzer, D.J., Grayson, D.K., Ardila, G., Barker, A.W., Dincauze, D.F., Haynes, D.V., Mena, F., Nuñez, L.A., Stanford, D.J. (1997). On the Pleistocene antiquity of Monte Verde, southern Chile. American Antiquity 62, 659–663. Meyer, G.A., Wells, S.G., Jull, A.J.T. (1995). Fire and alluvial chronology in Yellowstone National Park: Climatic and intrinsic controls on Holocene geomorphic processes. Geological Society of America Bulletin 107, 1211–1230. Meyer, G.A., Pierce, J.L., Wood, S.H., Jull, A.J.T. (2001). Fire, storms and erosional events in the Idaho batholith. Hydrological Processes 15, 3025–3038. Nelson, D.E. (1991). A new method for carbon isotopic analysis of protein. Science 251, 552–554. Nuñez, L.A., Varela, J., Casamiquela, R., Villagrán, C. (1994). Reconstrucción multidisciplinaria de la occupación prehistórica de Quereo, Centro de Chile. Latin American Antiquity 5, 99–118. Pierce, J.L., Meyer, G.A., Jull, A.J.T. (2004). Fire-induced erosion and millennialscale climate change in northern ponderosa pine forests. Nature 432, 87–90. Raisbeck, G.M., Yiou, F., Jouzel, J., Petit, J.R. (1990). 10 Be and 2 H in polar ice cores as a probe of the solar variability influence on climate. Philoophical Transactions of the Royal Society (London) A 330, 463–470. Richards, D.A., Beck, J.W. (2001). Dramatic shifts in radiocarbon dating the last glacial period. Antiquity 75, 482– 485. Rowe, M. (2001). Dating by AMS radiocarbon analysis. In: Whitley, D.S. (Ed.), Handbook of Rock Art Research. Alta Mira Press, pp. 139–166. Schmidt, A., Burr, G.S., Taylor, F.W., O’Malley, J., Beck, J.W. (2004). A semiannual radiocarbon record of a modern coral from the Solomon Islands. Nuclear Instruments and Methods in Physics Research B 223, 420–427. Slota, P.J., Jull, A.J.T., Linick, T.W., Toolin, L.J. (1987). Preparation of small samples for 14 C accelerator targets by catalytic reduction of CO. Radiocarbon 29 (3), 303–306. Stuiver, M., Polach, H. (1977). Reporting of 14 C data: Discussion. Radiocarbon 19, 355–363. Tisnérat-Laborde, N., Valadas, H., Kaltnecker, E., Arnold, M. (2003). AMS radiocarbon dating of bones at LSCE. Radiocarbon 45, 409–419. Tuniz, C., Bird, J.R., Fink, D., Herzog, G.F. (1998). Accelerator Mass Spectrometry: Ultrasensitive analysis for global science. CRC Press, Boca Raton, FL, 371 pp. Turcq, B., Siffedine, A., Martin, L., Absy, M.L., Soubies, F., Suguio, K., Volkmer-Ribeiro, C. (1998). Amazonia rainforest fires: A lacustrine record of 7000 years. Ambio 27, 139–142.
Further reading Alley, R.A., Clark, P.U., Keigwin, L.D., Webb, R.S. (1999). Making sense of millennial-scale climate change. In: Clark, P.U. et al. (Eds.), Mechanisms of Global Climate Change at Millennial Time Scales. Geophysics Monograph, vol. 112. American Geophysical Union, Washington, DC, pp. 385–394. Bird, M.I., Fifield, L.K., Santos, G.M., Beaumont, P.B., Zhou, Y., di Tada, M.I., Hausladen, P.A. (2003). Radiocarbon dating from 40 to 60 ka BP at Border Cave, South Africa. Quarternary Science Reviews 22, 943–947. Burr, G.S., Galang, C., Taylor, F.W., Gallup, C., Edwards, R.L., Cutler, K., Quirk, B. (2005). Radiocarbon results from a 13 ka BP coral from the Huon Peninsula, Papua New Guinea. Radiocarbon 46, 1211–1224. Cockburn, H.A.P., Summerfield, M.A. (2004). Geomorphological applications of cosmogenic isotope analysis. Progress in Physical Geography 28, 1–42. Druffel, E.R.M., Griffin, S., Hwang, J., Komada, T., Beupre, S.R., Druffel-Rodriguez, K.C., Santos, G.M., Southon, J. (2004). Variability of monthly radiocarbon during the 1760s in corals from the Galapagos Islands. Radiocarbon 46, 627–631.
Application of accelerator mass spectrometry
23
Gagan, M.K., Ayliffe, L.K., Beck, J.W., Cole, J.L., Druffel, E.R.M., Dunbar, R., Schrag, D.P. (2000). New views of tropical paleoenvironments from corals. Quaternary Science Reviews 19, 45–64. Gosse, J., Phillips, F.M. (2001). Terrestrial in-situ cosmogenic nuclides: Theory and application. Quarternary Science Reviews 20, 1475–1560. Grottoli, A.G., Gilli, S.T., Druffel, E.R.M., Dunbar, R.B. (2003). Decadal timescale shift in the 14 C record of a central equatorial Pacific coral. Radiocarbon 45, 91–99. Hallett, D.J., Lepofksy, D.S., Mathewes, R.W., Lertzman, K.P. (2003). 11,000 years of fire history and climate in the mountain hemlock rain forests of southwestern British Columbia based on sedimentary charcoal. Canadian Journal of Forest Research 33, 292–312. Hu, F.S., Kaufmann, D., Yoneji, S., Nelson, D., Shemesh, A., Huang, Y., Tian, J., Bond, G., Clegg, B., Brown, T. (2003). Cyclic variation and solar forcing of Holocene climate in the Alaskan subarctic. Science 301, 1890–1893. Hughen, K.A., Baillie, M.G.L., Bard, E., Beck, J.W., Bertrand, C.J., Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K., Damon, P.E., Edwards, R.L., Fairbanks, R., Friedrich, M., Guilderson, T.P., Kromer, B., McCormac, G., Manning, S., Bronk Ramsey, C., Reimer, P.J., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Tamalo, S., Taylor, F.W., van der Plicht, J., Wehenmeyer, C.E. (2005). MARINE04 Radiocarbon age calibration 0–26 ka cal BP. Radiocarbon 46, 1059–1086. Jull, A.J.T., Haynes, C.V. Jr, Donahue, D.J., Burr, G.S., Beck, J.W. (1999). Radiocarbon ages of early man in the New World and the influence of climate change. In: Evin, J. et al. (Eds.), Proc. 3rd International Conference “Archaeologie et 14 C”. Lyon, France, 6–10 April, 1998, pp. 239–343. Revu d’Archaeometrie, Suppl. 1999 et Soc. Préhist. Fr. Mémoire no. 26. Lertzman, K., Gavin, D., Hallett, D., Brubaker, L., Lepofsky, D., Mathewes, R. (2002). Long-term fire regime estimated from soil charcoal in coastal temperate rain forests. Conservation Ecology 6 (2), paper 5 [on-line journal]. Lifton, N.A., Pigati, J., Jull, A.J.T., Quade, J., Bierman, P., Kober, F. (2003). Testing cosmogenic nuclide production rate scaling models using in-situ 14 C from surfaces at secular equilibrium (abstract). Geochimica et Cosmochimica Acta 67, A253. Markgraf, V. (2001). Interhemispheric Climate Linkages. Academic Press, New York. Maslin, M., Seidov, D., Lowe, J. (2001). In: Seidov, D., Haupt, B.J., Maslin, M. (Eds.), The Oceans and Rapid Climate Change: Past, Present and Future. American Geophysical Union, Washington, DC, pp. 9–52. Masters, A.M. (1990). Changes in forest fire frequency in Kootenay National Park, Canadian Rockies. Canadian Journal of Botany 68, 1763–1767. McNichol, A.P., Jull, A.J.T., Burr, G.S. (2001). Converting AMS data to radiocarbon values: Considerations and conventions. Radiocarbon 43, 313–320. Meyer, G.A., Wells, S.G., Balling, R.C., Jull, A.J.T. (1992). Response of alluvial systems to fire and climate change in Yellowstone National Park. Nature 357, 147–150. Nuñez, L., Grosjean, M., Cartajena, I. (2001). In: Markgraf, V. (Ed.), Interhemispheric Climate Linkages. Academic Press, New York, pp. 105–117. Reimer, P.J., Baille, M.G., Bard, E., Bayless, A., Beck, J.W., Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K., Damon, P.E., Edwards, R.L., Fairbanks, R., Friedrich, M., Guilderson, T.P., Herring, C., Hughen, K.A., Kromer, B., McCormac, G., Manning, S., Ramsey, C.B., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Tamalo, S., Taylor, F.W., van der Plicht, J., Wehenmeyer, C.E. (2005). IntCal04 Terrestrial radiocarbon calibration 0–26 ka cal BP. Radiocarbon 46, 1029–1058.
24
Discriminating biogenic and anthropogenic chlorinated organic compounds using multi-isotope analyses of individual compounds Kazushi Aranamia,b,* , Steven J. Rowlandc , James W. Readmanb a Environmental Chemodynamics Section, Environmental Chemistry Division, National Institute for Environmental
Studies (NIES), Tsukuba, Ibaraki, 305-8506, Japan b Plymouth Marine Laboratory, Prospect Place, The Hoe, Plymouth, Devon PL1 3DK, United Kingdom c Petroleum and Environmental Geochemistry Group, School of Earth, Ocean and Environmental Sciences,
University of Plymouth, Drake Circus, Plymouth, Devon PL4 8AA, United Kingdom Abstract The number of known naturally occurring chlorinated organic compounds (COCs) is rapidly increasing and now totals approximately 2200. In order to investigate anthropogenic threats relating to COCs, the influence of natural COCs also needs to be understood. Compound-specific isotopic analyses (CSIA) of both stable (13 C) and radio(14 C) carbon on environmental samples has been used successfully to evaluate the origin and fate of some organic compounds (e.g. n-alkanes and polycyclic aromatic hydrocarbons (PAHs)). Recently CSIA of 13 C, 14 C and chlorine (37 Cl) isotopes have also been applied to COCs. In this paper, we review recent data for isotopic signatures and fractionations of COCs. The potential for discriminating the origins of COCs using CSIA is discussed. CSIA of 13 C is primarily applied to identify the origin of biosynthesized volatile COCs. CSIA of 37 Cl also has potential to discriminate biogenic COCs. In addition, stable isotopic fractionations of COCs associated with chloro-respiration offer potential for monitoring the environmental behavior of COCs. Keywords: Chlorinated organic compounds (COCs), Compound-specific isotopic analysis (CSIA), Stable carbon isotope (13 C), Stable chlorine isotope (37 Cl), Compound-specific radiocarbon analysis (CSRA)
1. Introduction Humans have synthesized many chlorinated organic compounds (COCs) which have been discharged into the environment. COCs made for industrial or agricultural practices were initially hailed as affording substantial improvements in the quality of life. However, their persistence, toxicity and bioaccumulative potential gave rise to concern (Carson, 1962; Colborn et al., 1996). COCs such as polychlorinated biphenyls (PCBs) or dichlorodiphenyl* Corresponding author. Address: National Institute for Environmental Studies, 16-2, Onogawa, Tsukuba, Ibaraki, 305-8506, Japan; phone and fax: (+81) 29 850 2902; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08002-2
© 2006 Elsevier Ltd. All rights reserved.
Discriminating chlorinated organic compounds
25
trichloroethane (DDT) have, more recently, been shown to have an endocrine disrupting influence on health (IPCS Assessment Report, 1995). Volatile COCs such as chlorofluorocarbons (CFCs) or hydrochlorofluorocarbons (HCFCs) have a catalytic influence depleting ozone in the stratosphere (IPCC Second Assessment Report, 1995). Thus, many people now consider COCs as an undesirable man-made legacy and many COCs have been discriminated from future use. Approximately 2200 natural COCs have been identified (Gribble, 2003), whilst thirty years ago, this number was only 150 (Siuda and DeBernardis, 1973). In many cases, these compounds have been investigated for pharmacological interest due to their biological activities, which include antifungal, antibacterial, antineoplastic, antiviral (e.g. anti-HIV), antiinflammatory, and other activities (Butler and Walker, 1993). They can be found in marine and terrestrial plants, marine animals, bacteria, fungi, some higher animals, and a few mammals, including humans. The oceans are the single largest source of biogenic COCs (Gribble, 1994, 2003; Winterton, 2000). The compounds have presumably evolved for chemical defense purposes (Paul et al., 1987; Pawlik, 1993) and are biosynthesized through chloroperoxidases (CPOs) (Butler and Walker, 1993; Urhahn and Ballschmiter, 1998; Ballschmiter, 2003). CPOs are enzymes secreted by bacteria or micro-algae associated with the host that catalyze chlorination as follows: Org-H + Cl− + H2 O2 + H+ → Org-Cl + 2H2 O. In contrast to chlorination, some bacterial de-chlorination reactions can occur under oxidative or reductive conditions through enzyme-catalyzed or metabolic processes (Fetzner, 1998). These have potential for use in bioremediation of materials contaminated with COCs. In the natural environment, biosynthesized COCs cycle through chlorination and dechlorination reactions (Fig. 1) which are highly stereo-selective (Fu et al., 1992) and result in isotopic fractionation (Reddy et al., 2002a). This suggests that stable isotopic signatures (37 Cl/35 Cl or 13 C/12 C) of individual COCs in environmental samples might allow discrimination between biogenic and anthropogenic origins. In addition, radiocarbon dating of individual COCs might also be used to distinguish between biogenic (contemporary 14 C) and
Fig. 1. Schematic for the biogeochemical cycle of COCs.
26
K. Aranami et al.
anthropogenic (14 C-free) compounds. Here, we review recent data on isotopic signatures and fractionations of COCs and discuss the potential for discriminating COC origins using compound-specific isotopic analyses (CSIA).
2. Biogenic chlorinated compounds Methyl chloride (CH3 Cl) and chloroform (CHCl3 ) are well known as major biogenic COCs. Terrestrial CH3 Cl sources include: biomass burning (Crutzen et al., 1979; Lobert et al., 1999), fungi (Harper, 1985) and higher plants (Yokouchi et al., 2002). Contributions from these are thought to exceed oceanic sources (Lovelock, 1975; Gschwend et al., 1985; Moore et al., 1996; Keene et al., 1999). For CHCl3 , terrestrial (soil) sources (Hoekstra et al., 1998; Haselmann et al., 2000; Laturnus et al., 2000) are thought to be comparable to oceanic sources (Nightingale et al., 1995; Laturnus et al., 2002; McCulloch, 2003). Trichloroacetate (TCA), which is used as a herbicide and is formed via the atmospheric breakdown of trichloroethene (TCE) and tetrachloroethene (PCE), is also produced naturally in soils even without CPOs in the presence of Cl− and H2 O2 (McCulloch, 2002; Schöler et al., 2003; Hoekstra, 2003; Fahimi et al., 2003; Ahlers et al., 2003). Persistent and bioaccumulative biogenic COCs, such as chlorinated anisoles (Haglund et al., 1997; Führer and Ballschmiter, 1998), chlorinated bipyrroles (Tittlemier et al., 1999, 2002; Vetter et al., 2000, 2003; Wu et al., 2002) and dioxins (Hashimoto et al., 1995; Silk et al., 1997; Hoekstra et al., 1999; Ferrario et al., 2000; Gaus et al., 2001; Green et al., 2001) have been reported in both terrestrial and marine environments (Fig. 2).
3. Stable carbon isotopes CSIA of stable carbon (13 C) in the complex mixtures found in environmental samples is usually measured using on-line gas chromatography/combustion/isotope ratio monitoring mass spectrometry (GC/C/IRMS or irm-GC/MS). Stable carbon isotope ratios are expressed in the delta notation δ 13 C (h) = (Rsample /Rstandard − 1) × 1000, where Rsample and Rstandard are the 13 C/12 C ratios of a sample and standard, respectively. The international standard for carbon is the Vienna Pee Dee Belemnite (VPDB) from the National Institute of Standards and Technology (NIST). A sample size of 10–100 ng C is typically required for continuous flow GC/C/IRMS and delivers a reproducibility of 0.1–1.0h (Eakin et al., 1992; Merritt et al., 1995; Rudolph et al., 1997). CSIA of oceanic sediments or atmospheric aerosols have been successfully applied to the investigation of the origin and fate of hydrocarbons (e.g. n-alkanes and polycyclic aromatic hydrocarbons (PAHs)). In early studies, the stable carbon isotopic signatures of individual n-alkanes seemed to reflect the origins of oceanic phytoplankton, terrestrial C3 and C4 plants, and petroleum (Freeman et al., 1990, 1994; Hayes et al., 1990; Rieley et al., 1991; Schoell et al., 1992, 1994a, 1994b; Hayes, 1993; Eglinton, 1994; Andrews et al., 1995;
Discriminating chlorinated organic compounds
27
Fig. 2. Examples of persistent biogenic COCs. 1 Haglund et al. (1997); 2 Führer and Ballschmiter (1998); 3 Vetter et al. (2000, 2003); 4 Wu et al. (2002); 5 Tittlemier et al. (1999, 2002); 6 Hashimoto et al. (1995); 7 Silk et al. (1997); 8 Ferrario et al. (2000); 9 Gaus et al. (2001); 10 Green et al. (2001); 11 Hoekstra et al. (1999).
Rogers and Savard, 1999). However, the larger isotopic fractionations associated with bacterial metabolism have more recently been shown to be of particular significance (Botz et al., 1996; Summons et al., 1998; Teece et al., 1999). Other investigations have demonstrated that 13 C measurements of individual PAHs provide a useful tool for discriminating between pyrolytic and petrogenic origins (O’Malley et al., 1994, 1996, 1997; Ballentine et al., 1996; Smirnov et al., 1998; Norman et al., 1999; McRae et al., 1999; Okuda et al., 2002a, 2002b, 2002c). Recent papers addressing PAHs have also shown that certain compounds, particularly perylene, in tropical sediments have depleted 13 C (McRae et al., 2000; Wilcke et al., 2002; Fahimi et al., 2003) and contemporary 14 C (Reddy et al., 2002b). This indicates the production of some natural PAHs, probably under anaerobic conditions.
28
K. Aranami et al.
CSIA of COCs has also been used to monitor isotopic fractionations associated with biodegradation of anthropogenic chlorinated chemicals, for example TCE and PCE (Hunkeler et al., 1999, 2003; Dayan et al., 1999; Lobert et al., 1999; Bloom et al., 2000; Slater et al., 2001, 2002; Bill et al., 2001; Song et al., 2002; Barth et al., 2002; Kirtland et al., 2003; Chu et al., 2004) and PCBs (Drenzek et al., 2001). CSIA of biogenic COCs has shown that atmospheric CH3 Cl has more depleted δ 13 C compared to non-methane hydrocarbons (NMHCs) or CFCs (Rudolph et al., 1997; Tsunogai et al., 1999; Thompson et al., 2002). Also, CH3 Cl emitted from terrestrial higher plants (Harper et al., 2001, 2003) and coastal salt marshes (Bill et al., 2002; Rhew et al., 2002) has extremely depleted e 13 C signatures. Identification of 13 C end-members of CH3 Cl by CSIA could prove useful for estimating contributions from the various sources to atmospheric CH3 Cl. Stable carbon isotopic ratios of volatile and persistent halogenated organic compounds are summarized in Fig. 3. Volatile compounds normally exist as gas phase and are reactive in the atmosphere. Persistent compounds are relatively stable in the environment and accumulate in the biosphere, but harmful to organisms and humans. The values of the volatile biogenic COCs such as CH3 Cl and CHCl3 tend toward more depleted δ 13 C, although the range of values is large. Unfortunately there is currently no data for persistent biogenic COCs. If, however, we look at the brominated persistent biogenic compound 2-(3 ,5 -dibromo-2 -methoxyphenoxy)3,5-dibromoanisole (Reddy et al., 2002c), this has a value (n = 1) relatively close to those of CH3 Cl. However, in order to apply CSIA to distinguish between anthropogenic and biogenic
Fig. 3. Stable carbon isotopic signatures of selected COCs (error bars represent the ranges of reported isotopic values). (a) 2-(3 ,5 -dibromo-2 -methoxyphenoxy)-3,5-dibromoanisole; (b) average value of background tropospheric CH3 Cl (Thompson et al., 2002); 1 Reddy et al. (2002c); 2 Drenzek et al. (2002); 3 van Warmerdam et al. (1995), Beneteau et al. (1999); 4 Holt et al. (1997); 5 Jendrzejewski et al. (1997); 6 Jendrzejewski et al. (2001); 7 Rudolph et al. (1997); 8 Tsunogai et al. (1999); 9 Harper et al. (2001, 2003).
Discriminating chlorinated organic compounds
29
persistent COCs, more 13 C data on persistent biogenic COCs are required. In addition, supplementary chemical information such as structural/molecular (enantiomer; isomer/congener) or isotopic (37 Cl or 14 C) signatures, which result from highly stereo-selective and isotopic fractionation or utilization of contemporary carbon of biosynthesized COCs, might be needed. 4. Stable chlorine isotopes Stable chlorine (37 Cl) isotope ratios are currently measured using dual-inlet stable isotope ratio mass spectrometry (SIRMS) (Long et al., 1993; Eggenkamp et al., 1995; Holt et al., 1997; Jendrzejewski et al., 1997; Rosenbaum et al., 2000) and thermal ionization mass spectrometry (TIMS) (Xiao and Zhang, 1992; Magenheim et al., 1994; Xiao et al., 1995, 2002; Rosenbaum et al., 2000; Numata et al., 2001; Holmstrand et al., 2004). Again, the information is expressed in the delta notation: δ 37 Cl (h) = (Rsample /Rstandard − 1) × 1000, where Rsample and Rstandard are the 37 Cl/35 Cl ratios of a sample and standard, respectively. There is no international standard for chlorine isotope ratios. The isotopic variations throughout the oceans are small and are typically less than analytical precision. This results from the relatively short circulation time of the ocean (1.5 × 103 yr) compared with the long oceanic residence time for chlorine (87 × 106 yr). Hence, Standard Mean Ocean Chloride (SMOC) has been proposed as the reference standard (Kaufmann et al., 1984; Holmstrand et al., 2004). A sample size of 3 µg Cl is required for TIMS and 300 µg Cl for SIRMS, both affording a reproducibility of <0.1–0.2h (Rosenbaum et al., 2000; Xiao et al., 2002; Holmstrand et al., 2004). Although there are only few data on stable chlorine isotopes in specific organic compounds, the potential for identifying different sources and degradation mechanisms (in combination with stable carbon isotopes) has been suggested (van Warmerdam et al., 1995; Beneteau et al., 1999; Jendrzejewski et al., 2001; Drenzek et al., 2002). Stable chlorine isotopic ratios of volatile and persistent COCs are summarized in Fig. 4. The values for CH3 Cl appear to be very variable (Tanaka and Rye, 1991; Holt et al., 1997; Jendrzejewski et al., 2001) but generally low (Tanaka and Rye, 1991). Moreover, the values for COCs biosynthesized by CPO-catalyzed reactions are estimated to be approximately from −14h to −10h (Reddy et al., 2002a). This suggests that biogenic COCs have more depleted isotopic end-members when compared to anthropogenic COCs. Other studies suggest that bacterial de-chlorination of anthropogenic chlorinated chemicals (such as TCE and PCE) is accompanied by some isotopic fractionation (Sturchio et al., 1998, 2003; Heraty et al., 1999; Reddy et al., 2000; Numata et al., 2002). It would appear that more depleted δ 37 Cl in the environment imply biosynthesized COCs or byproducts of chloro-respiration. 5. Radiocarbon Compound-specific radiocarbon analyses (CSRA) are often performed by measurement using accelerator mass spectrometry (AMS) after isolation of individual compounds from mixtures by methods such as preparative capillary gas chromatography (PCGC). Sample sizes of
30
K. Aranami et al.
Fig. 4. Stable chlorine isotopic signatures of selected COCs (error bars represent the ranges of reported isotopic values). 1 Reddy et al. (2000); 2 Drenzek et al. (2002); 3 Tanaka and Rye (1991); 4 van Warmerdam et al. (1995), Beneteau et al. (1999); 5 Holt et al. (1997); 6 Jendrzejewski et al. (1997); 7 Jendrzejewski et al. (2001).
25 µg C are required for AMS and the results obtained have an associated error of 10–20h (Eglinton et al., 1996). Radiocarbon (14 C) is produced from 14 N in the atmosphere. The 14 C is quickly oxidized to 14 CO2 , which is assimilated by plants during photosynthesis. Once the 14 C assimilation ceases, 14 C concentrations decrease through radioactive decay with a half-life of 5730 years. Hence, recently biosynthesized organic compounds contain “contemporary 14 C” and anthropogenic chemicals derived from petroleum are “14 C-free” (Reddy et al., 2002c). Whilst atmospheric 14 C production is affected by natural changes in sunspots and the Earth’s magnetic field (Bowman, 1990), there have also been anthropogenic perturbations such as fossil fuel combustion (the “Suess Effect” from the late 19th century) and nuclear bomb testing (the “Bomb Spike” during the 1950s and 1960s). Results are expressed as 14 C which is the per mill (h) deviation from the international standard for 14 C dating. For example, the “pre-bomb” 14 C concentration is defined as baseline of Standard Reference Material 4990B “Oxalic Acid” by 0.95 times (Stuiver and Polach, 1977). Although CSRA data in the environment have been used successfully to evaluate the origin and fate of n-alkanes (Lichtfouse and Eglinton, 1995; Eglinton et al., 1997; Pearson and Eglinton, 2000) and PAHs (Lichtfouse et al., 1997; Reddy et al., 2002b; Slater et al., 2002; Kanke et al., 2004; Mandalakis et al., 2004), few CSRA data on COCs have been published (Reddy et al., 2002c, 2004). Radiocarbon data for persistent COCs are summarized in Fig. 5. As expected, anthropogenic chemicals are “14 C-free” and biogenic COCs contain “contemporary 14 C”. An exception is 14 C values of toxaphene which are high (>200h) because the pesticide was derived from a non-petrochemical source (synthesized during the 1950s–1970s
Discriminating chlorinated organic compounds
31
Fig. 5. Radiocarbon dating of selected COCs (error bars represent the ranges of reported isotopic values). (a) 2-(3 ,5 -dibromo-2 -methoxyphenoxy)-3,5-dibromoanisole (b) 1,1 -dimethyl-3,3 ,4,4 -tetrabromo-5,5 dichloro-2,2 -bipyrrole. 1 Reddy et al. (2002c); 2 Reddy et al. (2004).
by the photo-chlorination of camphene, an isomerization product of α-pinene extracted from pine tree stumps) (Saleh, 1991). 14 C values of 1,1 -dimethyl-3,3 ,4,4 -tetrabromo-5,5 -dichloro-2,2 -bipyrrole (DBPBr4 Cl2 ) in marine animals are approximately −460h corresponding to conventional 14 C ages of ∼5000 years before present (BP) (Reddy et al., 2004). The authors suggest three possibilities to explain the depleted 14 C values: (1) approximately equal inputs from both natural and industrial sources, (2) naturally biosynthesized compounds derived from pre-aged carbon such as oceanic dissolved organic carbon (DOC) and (3) extremely persistent biosynthesized compounds with a relatively long residence time.
6. Conclusions Stable isotopic compositions of COCs in the environment depend on (1) isotopic components of the end-members of the source utilized, (2) isotopic fractionation associated with CPO-catalyzed chlorination during production of the COCs, (3) isotopic fractionation associated with metabolic de-chlorination, (4) isotopic residence times in the samples. Radiocarbon signatures, however, depend only upon the time passed (not entirely) since the atmospheric 14 C was assimilated by plants during photosynthesis. The isotopic data for COCs are summarized in Table 1. The larger is difference between anthropogenic and biogenic isotopic end members, the better is a tool for source identification. Therefore, CSIA of 13 C could be applied to identify the origin of biosynthesized volatile
32
K. Aranami et al.
Table 1 Summary of isotopic data of COCs Isotope Chemical property
Source
Isotopic end-member of source (h)
Enrichment factor with CPOchlorination (h)
Enrichment factor with chlororespiration (h)
Analytical method
Sample size
13 C
Anthropogenic Biogenic Anthropogenic Biogenic
−30 to −20 (ca. −35) −35 to −25 −75 to −40
? (>0) (>0) (20–40)
? ? 1–27 ?
GC/C/IRMS
10–100 ng
Anthropogenic Biogenic Anthropogenic Biogenic
−5 to 1 (−14 to −10) −3 to 4 −6 to 4
(∼3.5) ∼11 ? ?
(0–1.5) ? 0–13 ?
SIRMS
∼300 µg
TIMS
∼3 µg
Anthropogenic Biogenic Anthropogenic Biogenic
ca. −1000 −500 to 100 – –
– – – –
– – – –
Persistent Volatile
37 Cl
Persistent Volatile
14 C
Persistent Volatile
PCGC–AMS ∼25 µg
Notes: Values in parentheses are not directly measured. Bold values are thought to be used for discriminating biogenic COCs. Enrichment factors with chloro-respiration are usually expressed as negative values.
COCs, while CSIA of 37 Cl and CSRA would have a potential for discriminating persistent biogenic COCs. In particular, CSRA appears to be the most powerful tool for source identification, because an isotopic difference between dead and alive carbon is very large. However, the relatively large sample size required for AMS may limit the application of CSRA to discriminating biogenic and anthropogenic COCs. Moreover, stable isotopic fractionations of COCs associated with chloro-respiration could be used for monitoring the environmental behavior of COCs. Although a compound-specific multi-isotope approach has the potential to evaluate the origin and fate of COCs in the environment, isotopic data for individual COCs is very scarce to date. Advances in the compound-specific multi-isotope approach are needed to provide the essential temporal and process-related information necessary to generate predictive biogeochemical models to track anthropogenic disturbances.
Acknowledgements The authors are grateful to Dr. Kimiyoshi Kitamura of NIES (National Institute for Environmental Studies, Japan) and staff of National Marine Biological Library, Plymouth for their help with the literature search. This work was supported by Grant-in-Aid for JSPS (the Japan Society for the Promotion of Science) Fellows.
Discriminating chlorinated organic compounds
33
References Ahlers, J., Regelmann, H., Riedhammer, C. (2003). Environmental risk assessment of airborne trichloroacetic acid – a contribution to the discussion on the significance of anthropogenic and natural sources. Chemosphere 52, 531– 537. Andrews, J.E., Greenaway, A.M., Dennis, P.F. (1995). Combined carbon isotope and C/N ratios as indicators of source and fate of organic matter in a poorly flushed, tropical estuary: Hunts Bay, Kingston Harbor, Jamaica. Estuarine Coastal and Shelf Science 46, 743–756. Ballentine, D.C., Macko, S.A., Turekian, V.C., Gilhooly, W.P., Martincigh, B. (1996). Compound specific isotope analysis of fatty acids and polycyclic aromatic hydrocarbons in aerosols: Implications for biomass burning. Organic Geochemistry 25, 97–104. Ballschmiter, K. (2003). Pattern and sources of naturally produced organohalogens in the marine environment: Biogenic formation of organohalogens. Chemosphere 52, 313–324. Barth, J.A.C., Slater, G., Schuth, C., Bill, M., Downey, A., Larkin, M., Kalin, R.M. (2002). Carbon isotope fractionation during aerobic biodegradation of trichloroethene by Burkholderia cepacia G4: A tool to map degradation mechanisms. Applied and Environmental Microbiology 68, 1728–1734. Beneteau, K.M., Aravena, R., Frape, S.K. (1999). Isotopic characterization of chlorinated solvents – laboratory and field results. Organic Geochemistry 30, 739–753. Bill, M., Schuth, C., Barth, J.A.C., Kalin, R.M. (2001). Carbon isotope fractionation during abiotic reductive dehalogenation of trichloroethene (TCE). Chemosphere 44, 1281–1286. Bill, M., Rhew, R.C., Weiss, R.F., Goldstein, A.H. (2002). Carbon isotope ratios of methyl bromide and methyl chloride emitted from a coastal salt marsh. Geophysical Research Letters 29 (4), 1045. Bloom, Y., Aravena, R., Hunkeler, D., Edwards, E., Frape, S.K. (2000). Carbon isotope fractionation during microbial dechlorination of trichloroethene, cis-1,2-dichloroethene, and vinyl chloride: Implications for assessment of natural attenuation. Environmental Science & Technology 34, 2768–2772. Botz, R., Pokojski, H.D., Schmitt, M., Thomm, M. (1996). Carbon isotope fractionation during bacterial methanogenesis by CO2 reduction. Organic Geochemistry 25, 255–262. Bowman, S. (1990). Radiocarbon Dating. University of California Press, Berkeley. Butler, A., Walker, J.V. (1993). Marine haloperoxidases. Chemical Reviews 93, 1937–1944. Carson, R. (1962). Silent Spring. Houghton Mifflin, Boston. Chu, K.H., Mahendra, S., Song, D.L., Conrad, M.E., Alvarez-Cohen, L. (2004). Stable carbon isotope fractionation during aerobic biodegradation of chlorinated ethenes. Environmental Science & Technology 38, 3126–3130. Colborn, T., Dumanoski, D., Myers, J.P. (1996). Our Stolen Future. Dutton, New York. Crutzen, P.J., Heidt, L.E., Krasneck, J.P., Pollock, W.H., Seiler, W. (1979). Biomass burning as a source of atmospheric trace gases: CO, H2 , N2 O, NO, CH3 Cl and COS. Nature 282, 253–256. Dayan, H., Abrajano, T., Sturchio, N.C., Winsor, L. (1999). Carbon isotopic fractionation during reductive dehalogenation of chlorinated ethenes by metallic iron. Organic Geochemistry 30, 755–763. Drenzek, N.J., Eglinton, T.I., May, J.M., Wu, Q.Z., Sowers, K.R., Reddy, C.M. (2001). The absence and application of stable carbon isotopic fractionation during the reductive dechlorination of polychlorinated biphenyls. Environmental Science & Technology 35, 3310–3313. Drenzek, N.J., Tarr, C.H., Eglinton, T.I., Heraty, L.J., Sturchio, N.C., Shiner, V.J., Reddy, C.M. (2002). Stable chlorine and carbon isotopic compositions of selected semi-volatile organochlorine compounds. Organic Geochemistry 33, 437–444. Eakin, P.A., Fallick, A.E., Gerc, J. (1992). Some instrumental effects in the determination of stable carbon isotope ratios by gas-chromatography isotope ratio mass-spectrometry. Chemical Geology 101, 71–79. Eggenkamp, H.G.M., Kreulen, R., vanGroos, A.F.K. (1995). Chlorine stable isotope fractionation in evaporates. Geochimica et Cosmochimica Acta 59, 5169–5175. Eglinton, T.I. (1994). Carbon isotopic evidence for the origin of macromolecular aliphatic structures in kerogen. Organic Geochemistry 21, 721–735. Eglinton, T.I., Aluwihare, L.I., Bauer, J.E., Druffel, E.R.M., McNichol, A.P. (1996). Gas chromatographic isolation of individual compounds from complex matrices for radiocarbon dating. Analytical Chemistry 68, 904–912. Eglinton, T.I., Benitez-Nelson, B.C., Pearson, A., McNichol, A.P., Bauer, J.E., Druffel, E.R.M. (1997). Variability in radiocarbon ages of individual organic compounds from marine sediments. Science 277, 796–799.
34
K. Aranami et al.
Fahimi, I.J., Keppler, F., Scholer, H.F. (2003). Formation of chloroacetic acids from soil, humic acid and phenolic moieties. Chemosphere 52, 513–520. Ferrario, J.B., Byrne, C.J., Cleverly, D.H. (2000). 2,3,7,8-dibenzo-p-dioxins in mined clay products from the United States: Evidence for possible natural origin. Environmental Science & Technology 34, 4524–4532. Fetzner, S. (1998). Bacterial dehalogenation. Applied Microbiology and Biotechnology 50, 633–657. Freeman, K.H., Hayes, J.M., Trendel, J.M., Albrecht, P. (1990). Evidence from carbon isotope measurements for diverse origins of sedimentary. Nature 343, 254–256. Freeman, K.H., Wakeham, S.G., Hayes, J.M. (1994). Predictive isotope biogeochemistry – hydrocarbons from anoxic marine basins. Organic Geochemistry 21, 629–644. Fu, H., Kondo, H., Ichikawa, Y., Look, G.C., Wong, C.H. (1992). Chloroperoxidase-catalyzed asymmetric-synthesis enantioselective reactions of chiral hydroperoxides with sulfides and bromohydration of glycals. Journal of Organic Chemistry 57, 7265–7270. Führer, U., Ballschmiter, K. (1998). Bromochloromethoxybenzenes in the marine troposphere of the Atlantic Ocean: A group of organohalogens with mixed biogenic and anthropogenic origin. Environmental Science & Technology 32, 2208–2215. Gaus, C., Papke, O., Dennison, N., Haynes, D., Shaw, G.R., Connell, D.W., Muller, J.F. (2001). Evidence for the presence of a widespread PCDD source in coastal sediments and soils from Queensland, Australia. Chemosphere 43, 549–558. Green, N.J.L., Jones, J.L., Johnston, A.E., Jones, K.C. (2001). Further evidence for the existence of PCDD/Fs in the environment prior as 1900. Environmental Science & Technology 35, 1974–1981. Gribble, G.W. (1994). The natural production of chlorinated compounds. Environmental Science & Technology 28, A310–A319. Gribble, G.W. (2003). The diversity of naturally produced organohalogens. Chemosphere 52, 289–297. Gschwend, P.M., Macfarlane, J.K., Newman, K.A. (1985). Volatile halogenated organic-compounds released to seawater from temperate marine macroalgae. Science 227, 1033–1035. Haglund, P.S., Zook, D.R., Buser, H.R., Hu, J.W. (1997). Identification and quantification of polybrominated diphenyl ethers and methoxy-polybrominated diphenyl ethers in Baltic biota. Environmental Science & Technology 31, 3281–3287. Harper, D.B. (1985). Halomethane from halide ion – A highly effect fungal conversion of environmental significance. Nature 315, 55–57. Harper, D.B., Kalin, R.M., Hamilton, J.T.G., Lamb, C. (2001). Carbon isotope ratios for chloromethane of biological origin: Potential tool in determining biological emissions. Environmental Science & Technology 35, 3616–3619. Harper, D.B., Hamilton, J.T.G., Ducrocq, V., Kennedy, J.T., Downey, A., Kalin, R.M. (2003). The distinctive isotopic signature of plant-derived chloromethane: Possible application in constraining the atmospheric chloromethane budget. Chemosphere 52, 433–436. Haselmann, K.F., Ketola, R.A., Laturnus, F., Lauritsen, F.R., Gron, C. (2000). Occurrence and formation of chloroform at Danish forest sites. Atmospheric Environment 34, 187–193. Hashimoto, S., Wakimoto, T., Tatsukawa, R. (1995). Possible natural formation of polychlorinated dibenzo-p-dioxins as evidenced by sediment analysis from the Yellow Sea, the East-China Sea and the Pacific Ocean. Marine Pollution Bulletin 30, 341–346. Hayes, J.M. (1993). Factors controlling C-13 contents of sedimentary organic-compounds-principles and evidence. Marine Geology 113, 111–125. Hayes, J.M., Freeman, K.H., Popp, B.N., Hoham, C.H. (1990). Compound-specific isotopic analyses – A novel tool for reconstruction of ancient biogeochemical processes. Organic Geochemistry 16, 1115–1128. Heraty, L.J., Fuller, M.E., Huang, L., Abrajano, T., Sturchio, N.C. (1999). Isotopic fractionation of carbon and chlorine by microbial degradation of dichloromethane. Organic Geochemistry 30, 793–799. Hoekstra, E.J. (2003). Review of concentrations and chemistry of trichloroacetate in the environment. Chemosphere 52, 355–369. Hoekstra, E.J., De Leer, E.W.B., Brinkman, U.A.T. (1998). Natural formation of chloroform and brominated trihalomethanes in soil. Environmental Science & Technology 32, 3724–3729. Hoekstra, E.J., De Weerd, H., De Leer, E.W.B., Brinkman, U.A.T. (1999). Natural formation of chlorinated phenols, dibenzo-p-dioxins, and dibenzofurans in soil of a Douglas fir forest. Environmental Science & Technology 33, 2543–2549.
Discriminating chlorinated organic compounds
35
Holmstrand, H., Andersson, P., Gustafsson, O. (2004). Chlorine isotope analysis of submicromole organochlorine samples by sealed tube combustion and thermal ionization mass spectrometry. Analytical Chemistry 76, 2336– 2342. Holt, B.D., Sturchio, N.C., Abrajano, T.A., Heraty, L.J. (1997). Conversion of chlorinated volatile organic compounds to carbon dioxide and methyl chloride for isotopic analysis of carbon and chlorine. Analytical Chemistry 69, 2727–2733. Hunkeler, D., Aravena, R., Butler, B.J. (1999). Monitoring microbial dechlorination of tetrachloroethene (PCE) in groundwater using compound-specific stable carbon isotope ratios: Microcosm and field studies. Environmental Science & Technology 33, 2733–2738. Hunkeler, D., Aravena, R., Parker, B.L., Cherry, J.A., Diao, X. (2003). Monitoring oxidation of chlorinated ethenes by permanganate in groundwater using stable isotopes: Laboratory and field studies. Environmental Science & Technology 37, 798–804. IPCC (1995). Climate change. IPCC Second Assessment Report. IPCS (1995). DDT, aldrin, dieldrin, endrin, chlordane, heptachlor, hexachlorobenzene, mirex, toxaphene, polychlorinated biphenyls, dioxins and furans. IPCS Assessment Report. Jendrzejewski, N., Eggenkamp, H.G.M., Coleman, M.L. (1997). Sequential determination of chlorine and carbon isotopic composition in single microliter samples of chlorinated solvent. Analytical Chemistry 69, 4259–4266. Jendrzejewski, N., Eggenkamp, H.G.M., Coleman, M.L. (2001). Characterization of chlorinated hydrocarbons from chlorine and carbon isotopic compositions: Scope of application to environmental problems. Applied Geochemistry 16, 1021–1031. Kanke, H., Uchida, M., Okuda, T., Yoneda, M., Takada, H., Shibata, Y., Morita, M. (2004). Compound-specific radiocarbon analysis of polycyclic aromatic hydrocarbons (PAHs) in sediments from an urban reservoir. Nuclear Instruments & Methods in Physics Research, Section B – Beam Interactions with Materials and Atoms 223/224, 545–554. Kaufmann, R.S., Long, A., Bentley, H.W., Davis, S.N. (1984). Natural stable chlorine isotope variations. Nature 309, 338–340. Keene, W.C., Khalil, M.A.K., Erickson, D.J., McCulloch, A., Graedel, T.E., Lobert, J.M., Aucott, M.L., Gong, S.L., Harper, D.B., Kleiman, G., Midgley, P., Moore, R.M., Seuzaret, C., Sturges, W.T., Benkovitz, C.M., Koropalov, V., Barrie, L.A., Li, Y.F. (1999). Composite global emissions of reactive chlorine from anthropogenic and natural sources: Reactive chlorine emissions inventory. Journal of Geophysical Research – Atmospheres 104, 8429–8440. Kirtland, B.C., Aelion, C.M., Stone, P.A., Hunkeler, D. (2003). Isotopic and geochemical assessment of in situ biodegradation of chlorinated hydrocarbons. Environmental Science & Technology 37, 4205–4212. Laturnus, F., Lauritsen, F.R., Gron, C. (2000). Chloroform in a pristine aquifer system: Toward an evidence of biogenic origin. Water Resources Research 36, 2999–3009. Laturnus, F., Haselmann, K.F., Borch, T., Gron, C. (2002). Terrestrial natural sources of trichloromethane (chloroform, CHCl3 ) – An overview. Biogeochemistry 60, 121–139. Lichtfouse, E., Eglinton, T.I. (1995). C-13 and C-14 evidence of pollution of a soil by fossil fuel and reconstruction of the composition of the pollutant. Organic Geochemistry 23, 969–973. Lichtfouse, E., Budzinski, H., Garrigues, P., Eglinton, T.I. (1997). Ancient polycyclic aromatic hydrocarbons in modern soils: C-13, C-14 and biomarker evidence. Organic Geochemistry 26, 353–359. Lobert, J.M., Keene, W.C., Logan, J.A., Yevich, R. (1999). Global chlorine emissions from biomass burning: Reactive chlorine emissions inventory. Journal of Geophysical Research – Atmospheres 104, 8373–8389. Long, A., Eastoe, C.J., Kaufmann, R.S., Martin, J.G., Wirt, L., Finley, J.B. (1993). High-precision measurement of chlorine stable-isotope ratios. Geochimica et Cosmochimica Acta 57, 2907–2912. Lovelock, J.E. (1975). Natural halocarbons in the air and in the sea. Nature 256, 193–194. Magenheim, A.J., Spivack, A.J., Volpe, C., Ansom, B. (1994). Precise determination of stable chlorine isotopic-ratios in low-concentration natural samples. Geochimica et Cosmochimica Acta 58, 3117–3121. Mandalakis, M., Gustafsson, O., Reddy, C.M., Xu, L. (2004). Radiocarbon apportionment of fossil versus biofuel combustion sources of polycyclic aromatic hydrocarbons in the Stockholm metropolitan area. Environmental Science & Technology 38, 5344–5349. McCulloch, A. (2002). Trichloroacetic acid in the environment. Chemosphere 47, 667–686. McCulloch, A. (2003). Chloroform in the environment: Occurrence, sources, sinks and effects. Chemosphere 50, 1291–1308.
36
K. Aranami et al.
McRae, C., Sun, C.G., Snape, C.E., Fallick, A.E., Taylor, D. (1999). δC-13 values of coal-derived PAHs from different processes and their application to source apportionment. Organic Geochemistry 30, 881–889. McRae, C., Snape, C.E., Sun, C.G., Fabbri, D., Tartari, D., Trombini, C., Fallick, A.E. (2000). Use of compoundspecific stable isotope analysis to source anthropogenic natural gas-derived polycyclic aromatic hydrocarbons in a lagoon sediment. Environmental Science & Technology 34, 4684–4686. Merritt, D.A., Freeman, K.H., Ricci, M.P., Studley, S.A., Hayes, J.M. (1995). Performance and optimization of a combustion interface for isotope ratio monitoring gas-chromatography mass-spectrometry. Analytical Chemistry 67, 2461–2473. Moore, R.M., Groszko, W., Niven, S.J. (1996). Ocean–atmosphere exchange of methyl chloride: Results from NW Atlantic and Pacific Ocean studies. Journal of Geophysical Research – Oceans 101, 28529–28538. Nightingale, P.D., Malin, G., Liss, P.S. (1995). Production of chloroform and other low-molecular-weight halocarbons by some species of macroalgae. Limnology and Oceanography 40, 680–689. Norman, A.L., Hopper, J.F., Blanchard, P., Ernst, D., Brice, K., Alexandrou, N., Klouda, G. (1999). The stable carbon isotope composition of atmospheric PAHs. Atmospheric Environment 33, 2807–2814. Numata, M., Nakamura, N., Gamo, T. (2001). Precise measurement of chlorine stable isotopic ratios by thermal ionization mass spectrometry. Geochemical Journal 35, 89–100. Numata, M., Nakamura, N., Koshikawa, H., Terashima, Y. (2002). Chlorine isotope fractionation during reductive dechlorination of chlorinated ethenes by anaerobic bacteria. Environmental Science & Technology 36, 4389–4394. Okuda, T., Kumata, H., Zakaria, M.P., Naraoka, H., Ishiwatari, R., Takada, H. (2002a). Source identification of Malaysian atmospheric polycyclic aromatic hydrocarbons nearby forest fires using molecular and isotopic compositions. Atmospheric Environment 36, 611–618. Okuda, T., Kumata, H., Naraoka, H., Ishiwatari, R., Takada, H. (2002b). Vertical distributions and delta C-13 isotopic compositions of PAHs in Chidorigafuchi Moat sediment, Japan. Organic Geochemistry 33, 843–848. Okuda, T., Kumata, H., Naraoka, H., Takada, H. (2002c). Origin of atmospheric polycyclic aromatic hydrocarbons (PAHs) in Chinese cities solved by compound-specific stable carbon isotopic analyses. Organic Geochemistry 33, 1737–1745. O’Malley, V.P., Abrajano, T.A., Hellou, J. (1994). Determination of the C-13/C-12 ratios of individual PAH from environmental samples: Can PAH sources be apportioned. Organic Geochemistry 21, 809–822. O’Malley, V.P., Abrajano, T.A., Hellou, J. (1996). Stable carbon isotopic apportionment of individual polycyclic aromatic hydrocarbons in St John’s Harbour, Newfoundland. Environmental Science & Technology 30, 634–639. O’Malley, V.P., Burke, R.A., Schlotzhauer, W.S. (1997). Using GC-MS/Combustion/IRMS to determine the C-13/C-12 ratios of individual hydrocarbons produced from the combustion of biomass materials – application to biomass burning. Organic Geochemistry 27, 567–581. Paul, V.J., Hay, M.E., Duffy, J.E., Fenical, W., Gustafson, K. (1987). Chemical defense in the seaweed Ochtodes secundiramea (Montagne) Howe (Rhodophyta): Effects of its monoterpenoid components upon diverse coralreef herbivores. Journal of Experimental Marine Biology and Ecology 114, 249–260. Pawlik, J.R. (1993). Marine invertebrate chemical defenses. Chemical Reviews 93, 1911–1922. Pearson, A., Eglinton, T.I. (2000). The origin of n-alkanes in Santa Monica Basin surface sediment: A model based on compound-specific delta C-14 and delta C-13 data. Organic Geochemistry 31, 1103–1116. Reddy, C.M., Heraty, L.J., Holt, B.D., Sturchio, N.C., Eglinton, T.I., Drenzek, N.J., Xu, L., Lake, J.L., Maruya, K.A. (2000). Stable chlorine isotopic compositions of aroclors and aroclor-contaminated sediments. Environmental Science & Technology 34, 2866–2870. Reddy, C.M., Xu, L., Drenzek, N.D., Sturchio, N.C., Heraty, L.J., Kimblin, C., Bulter, A. (2002a). A chlorine isotope effect for enzyme-catalyzed chlorination. Journal of the American Chemistry Society 124, 14526–14527. Reddy, C.M., Pearson, A., Xu, L., McNichol, A.P., Benner, B.A., Wise, S.A., Klouda, G.A., Currie, L.A., Eglinton, T.I. (2002b). Radiocarbon as a tool to apportion the sources of polycyclic aromatic hydrocarbons and black carbon in environmental samples. Environmental Science & Technology 36, 1774–1782. Reddy, C.M., Xu, L., Eglinton, T.I., Boon, J.P., Faulkner, D.J. (2002c). Radiocarbon content of synthetic and natural semi-volatile halogenated organic compounds. Environmental Pollution 120, 163–168. Reddy, C.M., Xu, L., O’Neil, G.W., Nelson, R.K., Eglinton, T.I., Faulkner, D.J., Norstrom, R., Ross, P.S., Tittlemier, S.A. (2004). Radiocarbon evidence for a naturally produced, bioaccumulating halogenated organic compound. Environmental Science & Technology 38, 1992–1997. Rhew, R.C., Miller, B.R., Bill, M., Goldstein, A.H., Weiss, R.F. (2002). Environmental and biological controls on methyl halide emissions from southern California coastal salt marshes. Biogeochemistry 60, 141–161.
Discriminating chlorinated organic compounds
37
Rieley, G., Collier, R.J., Jones, D.M., Eglinton, G., Eakin, P.A., Fallick, A.E. (1991). Sources of sedimentary lipids deduced from stable carbon isotope analyses of individual compounds. Nature 352, 425–427. Rogers, K.M., Savard, M.M. (1999). Detection of petroleum contamination in river sediments from Quebec City region using GC-IRMS. Organic Geochemistry 30, 1559–1569. Rosenbaum, J.M., Cliff, R.A., Coleman, M.L. (2000). Chlorine stable isotopes: A comparison of dual inlet and thermal ionization mass spectrometric measurements. Analytical Chemistry 72, 2261–2264. Rudolph, J., Lowe, D.C., Martin, R.J., Clarkson, T.S. (1997). A novel method for compound specific determination of delta C-13 in volatile organic compounds at ppt levels in ambient air. Geophysical Research Letters 24, 659–662. Saleh, M.A. (1991). Toxaphene: Chemistry, biochemistry, toxicity and environmental fate. Reviews of Environmental Contamination and Toxicology 118, 1–85. Schoell, M., Mccaffrey, M.A., Fago, F.J., Moldowan, J.M. (1992). Carbon isotopic compositions of 28,30bisnorhopanes and other biological markers in a Monterey crude-oil. Geochimica et Cosmochimica Acta 56, 1391–1399. Schoell, M., Hwang, R.J., Carlson, R.M.K., Welton, J.E. (1994a). Carbon isotopic composition of individual biomarkers in gilsonites (UTAH). Organic Geochemistry 21, 673–683. Schoell, M., Simoneit, B.R.T., Wang, T.G. (1994b). Organic geochemistry and coal petrology of tertiart brown-coal in the Zhoujing mine, basin, south China. 4. Biomarker sources inferred from stable carbon-isotope compositions of individual compounds. Organic Geochemistry 21, 713–719. Schöler, H.F., Keppler, F., Fahimi, I.J., Niedan, V.W. (2003). Fluxes of trichloroacetic acid between atmosphere, biota, soil, and groundwater. Chemosphere 52, 339–354. Silk, P.J., Lonergan, G.C., Arsenault, T.L., Boyle, C.D. (1997). Evidence of natural organochlorine formation in peat bogs. Chemosphere 35, 2865–2880. Siuda, J.F., DeBernardis, J.F. (1973). Naturally occurring halogenated organic compounds. Lloydia 36, 107–143. Slater, G.F., Lollar, B.S., Sleep, B.E., Edwards, E.A. (2001). Variability in carbon isotopic fractionation during biodegradation of chlorinated ethenes: Implications for field applications. Environmental Science & Technology 35, 901–907. Slater, G.F., Lollar, B.S., King, R.A., O’Hannesin, S. (2002). Isotopic fractionation during reductive dechlorination of trichloroethene by zero-valent iron: Influence of surface treatment. Chemosphere 49, 587–596. Smirnov, A., Abrajano, T.A., Smirnov, A., Stark, A. (1998). Distribution and sources of polycyclic aromatic hydrocarbons in the sediments of Lake Erie. Part 1. Spatial distribution, transport, and deposition. Organic Geochemistry 29, 1813–1828. Song, D.L., Conrad, M.E., Sorenson, K.S., Alvarez-Cohen, L. (2002). Stable carbon isotope fractionation during enhanced in situ bioremediation of trichloroethene. Environmental Science & Technology 36, 2262–2268. Stuiver, M., Polach, H.A. (1977). Reporting of 14 C data. Radiocarbon 19, 355–363. Sturchio, N.C., Clausen, J.L., Heraty, L.J., Huang, L., Holt, B.D., Abrajano, T.A. (1998). Chlorine isotope investigation of natural attenuation of trichloroethene in an aerobic aquifer. Environmental Science & Technology 32, 3037–3042. Sturchio, N.C., Hatzinger, P.B., Arkins, M.D., Suh, C., Heraty, L.J. (2003). Chlorine isotope fractionation during microbial reduction of perchlorate. Environmental Science & Technology 37, 3859–3863. Summons, R.E., Franzmann, P.D., Nichols, P.D. (1998). Carbon isotopic fractionation associated with methylotrophic methanogenesis. Organic Geochemistry 28, 465–475. Tanaka, N., Rye, D.M. (1991). Chlorine in stratosphere. Nature 353, 707. Teece, M.A., Fogel, M.L., Dollhopf, M.E., Nealson, K.H. (1999). Isotopic fractionation associated with biosynthesis of fatty acids by a marine bacterium under oxic and anoxic conditions. Organic Geochemistry 30, 1571–1579. Thompson, A.E., Anderson, R.S., Rudolph, J., Huang, L. (2002). Stable carbon isotope signatures of background tropospheric chloromethane and CFC113. Biogeochemistry 60, 191–211. Tittlemier, S.A., Simon, M., Jarman, W.M., Elliott, J.E., Norstrom, R.J. (1999). Identification of a novel C10 H6 N2 Br4 Cl2 heterocyclic compound in seabird eggs. A bioaccumulating marine natural product? Environmental Science & Technology 33, 26–33. Tittlemier, S.A., Fisk, A.T., Hobson, K.A., Norstrom, R.J. (2002). Examination of the bioaccumulation of halogenated dimethyl bipyrroles in an Arctic marine food web using stable nitrogen isotope analysis. Environmental Pollution 116, 85–93. Tsunogai, U., Yoshida, N., Gamo, T. (1999). Carbon isotopic compositions of C-2-C-5 hydrocarbons and methyl chloride in urban, coastal, and maritime atmospheres over the western North Pacific. Journal of Geophysical Research – Atmospheres 104, 16033–16039.
38
K. Aranami et al.
Urhahn, T., Ballschmiter, K. (1998). Chemistry of the biosynthesis of halogenated methanes: C1-organohalogens as pre-industrial chemical stressors in the environment? Chemosphere 37, 1017–1032. Vetter, W., Alder, L., Kallenborn, R., Schlabach, M. (2000). Determination of Q1, an unknown organochlorine contaminant, in human milk, Antarctic air, and further environmental samples. Environmental Pollution 110, 401–409. Vetter, W., Wu, J., Althoff, G. (2003). Non-polar halogenated natural products bioaccumulated in marine samples. I. 2,3,3 ,4,4 ,5,5 -heptachloro-1 -methyl-1,2 -bipyrrole (Q1). Chemosphere 52, 415–422. van Warmerdam, E.M., Frape, S.K., Aravena, R., Drimmie, R.J., Flatt, H., Cherry, J.A. (1995). Stable chlorine and carbon isotope measurements of selected chlorinated organic solvents. Applied Geochemistry 10, 547–552. Wilcke, W., Krauss, M., Amelung, W. (2002). Carbon isotope signature of polycyclic aromatic hydrocarbons (PAHs): Evidence for different sources in tropical and temperate environments? Environmental Science & Technology 36, 3530–3535. Winterton, N. (2000). Chlorine: The only green element – towards a wider acceptance of its role in natural cycles. Green Chemistry 2, 173–225. Wu, J., Vetter, W., Gribble, G.W., Schneekloth, J.S., Blank, D.H., Gorls, H. (2002). Structure and synthesis of the natural heptachloro-1 -methyl-1,2 -bipyrrole (Q1). Angewandte Chemie – International Edition 41, 1740–1743. Xiao, Y.K., Zhang, C.G. (1992). High-precision isotopic measurements of chlorine by thermal ionization massspectrometry of the Cs2 Cl+ ion. International Journal of Mass Spectrometry and Ion Processes 116, 183–192. Xiao, Y.K., Zhou, Y.M., Liu, W.G. (1995). Precise measurement of chlorine isotope based on Cs2 Cl+ by thermal ionization mass-spectrometry. Analytical Letters 28, 1295–1304. Xiao, Y.K., Lu, H., Zhang, C.G., Wang, Q.Z., Wei, H.Z., Sun, A., Liu, W.G. (2002). Major factors affecting the isotopic measurement of chlorine based on the CS2 Cl+ ion by thermal ionization mass spectrometry. Analytical Chemistry 74, 2458–2464. Yokouchi, Y., Ikeda, M., Inuzuka, Y., Yukawa, T. (2002). Strong emission of methyl chloride from tropical plants. Nature 416, 163–165.
Further reading Fabbri, D., Vassura, I., Sun, C.G., Snape, C.E., McRae, C., Fallick, A.E. (2003). Source apportionment of polycyclic aromatic hydrocarbons in a coastal lagoon by molecular and isotopic characterization. Marine Chemistry 84, 123–135. Lollar, B.S., Slater, G.F., Ahad, J., Sleep, B., Spivack, J., Brennan, M., MacKenzie, P. (1999). Contrasting carbon isotope fractionation during biodegradation of trichloroethylene and toluene: Implications for intrinsic bioremediation. Organic Geochemistry 30, 813–820. Slater, J.F., Currie, L.A., Dibb, J.E., Benner, B.A. (2002). Distinguishing the relative contribution of fossil fuel and biomass combustion aerosols deposited at Summit, Greenland through isotopic and molecular characterization of insoluble carbon. Atmospheric Environment 36, 4463–4477.
39
Shift in stable water isotopes in precipitation in the Andean Amazon: Implications of deforestation or greenhouse impacts? A. Henderson-Sellersa , K. McGuffieb,* a Australian Science and Technology Organisation, Lucas Heights, Australia b University of Technology, Sydney, Australia
Abstract Changes in the O and H isotopes in precipitation have been linked to greenhouse warming, but no signal attributable to Amazonian deforestation has been reported. Recent data from the Andes exhibit a seasonally contrasting signal which is consistent with large-area removal of forest. Specifically, at Izobamba, in the far west of the basin, the seasonality in isotopic depletions has become enhanced between 1972 and 2000. The observed more negative isotopic ratios in the wet season are consistent with increases in runoff fraction and/or reductions in recycling through nonfractionating processes. The dry season result (statistically significant less negative isotopic ratios) is harder to explain and could be due to a decrease in fractionating recycling (i.e. partial evaporation from water bodies). Application of a simple isotopic catchment model suggests that these isotopic changes in precipitation may be the result of large-scale deforestation in the Amazon Basin. Isotopically-enabled numerical models are needed to establish regional validity. Keywords: Amazon, Stable water isotopes, Deforestation, Transpiration, Re-evaporation, Runoff, Greenhouse
1. Climate and land-use change in the Amazon Deforestation throughout the world’s humid tropics is acknowledged to be a serious issue but adequate estimates of rates, extent and possible recovery are hard to ascertain (Achard et al., 2002). The Intergovernmental Panel on Climate Change (IPCC) has pointed out that “for tropical countries, deforestation estimates are very uncertain and could be in error by as much as ±50%” (Watson et al., 2000) while Glantz et al. (1997) detail the problems of attempting to assess rates and processes of deforestation in a careful study focused on the Amazon. Despite the acknowledged uncertainties, all data point to continuing forest removal (Fig. 1(a)). Even recent and careful studies that point to lower rates still identify the Brazilian Amazon as * Corresponding author. Address: Department of Applied Physics, University of Technology, Sydney, P.O. Box 123, Broadway, NSW 2007, Australia; phone: (+61) 2 9514 2072; fax: (+61) 2 9514 2219; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08003-4
© 2006 Elsevier Ltd. All rights reserved.
40
A. Henderson-Sellers and K. McGuffie
(a)
(b)
Fig. 1. (a) Rates of deforestation in thousands of square kilometers per year and total area affected by deforestation as a percentage of original forested area. Rates for 1975–1983 are deduced from data presented in Fearnside (1987). Later data on area deforested annually (squares) from INPE (1998) and other sources. Rates for years without data have been estimated (thin line). (b) South America showing the Amazon Basin area within which forest is removed in GCM simulations and the location of Izobamba (filled circle) and other GNIP stations from which data are analyzed.
having among the highest deforestation rates in sampled ‘hot spots’: 4.4% and 2.4% in Acre and Rondonia, cf. 3.2–5.9% in Central Sumatra (Achard et al., 2002). Removing dense tropical moist forest and replacing it with agricultural pasture or scrub and grasses that regrow naturally causes significant shifts in hydrology and ecology. These have been measured for many years and models have been employed to try to predict the impacts of continued deforestation in the Amazon Basin (Fig. 1(b)). Henderson-Sellers and Gornitz (1984) conducted the first Amazonian deforestation simulation with a Global Climate Model (GCM) and since then many modelling groups have contributed additional estimates (see McGuffie et al., 1998). The joint impact of large-scale deforestation and greenhouse gas-induced change has only recently begun to be assessed. Zhang et al. (2001) and Costa and Foley (2000) both employed GCMs, the latter’s including a sophisticated land-surface scheme allowing stomatal closure feedback to be included. Table 1 lists the combined effects of greenhouse warming and deforestation predicted by these models. In all cases, temperatures increase and evaporation and precipitation both decrease. Adding greenhouse gases produces larger temperature increases but diminishes the impact on the hydrological cycle. Without stomatal closure in the CO2 enriched atmosphere, Zhang et al. (2001) find no change in the evaporation to precipitation ratio (E/P ) but including this feedback allows Costa and Foley (2000) to predict increased evaporative demand despite overall reductions in the hydrological cycling. At present, there are no data that can be used to assess the verity of predictions of the impact of future deforestation and greenhouse gas enrichment. The search for data with which to evaluate and, hopefully, improve GCM simulations is the basis of this paper.
Shift in stable water isotopes in precipitation in the Andean Amazon
41
Table 1 Results from two recent GCM simulations of the impacts of deforesting the Amazon (Defor.) and this combined (Both) with doubling the global CO2 content of the atmosphere (2 × CO2 ). The Costa and Foley (2000) model also incorporates stomatal response to CO2 changes, i.e. closure in enriched CO2 atmosphere Recent GCM simulations Precipitation (mm)
Evaporation (mm)
E/P (%)
Zhang et al. (2001) (no stomatal feedback) Deforestation +0.3 Both 2 × CO2 and defor. +2.9
−402 −317
−222 −179
55 56
Costa and Foley (2000) (with plant physiology) Deforestation +1.4 Both 2 × CO2 and defor. +3.5
−266 −153
−223 −146
83 95
Temperature (K)
2. Basin-scale change in the Amazon 2.1. Stable water isotopes The stable isotopes of hydrogen, oxygen and carbon are employed widely in Earth system science (e.g. Gat, 1996; Petit et al., 1999; Townsend et al., 2002) including in GCM simulations (e.g. Joussaume et al., 1984; Hoffman et al., 2000). Ratios of 18 O/16 O and 2 H/1 H have been used to quantify aspects of the hydro-climate of the Amazon Basin for over 30 years (Salati and Vose, 1984). Coupled with measurement of isotopes in water sources such as precipitation and vapor, stable isotope characteristics in rivers have been able to provide insight into basin-integrated hydro-climates (Gibson and Edwards, 2002). In the Amazon, interception of rainfall by the plant canopy is the source of re-evaporated isotopically ‘enriched’ water (e.g. Salati et al., 1979; Victoria et al., 1991). This process causes lower continental depletion in heavy isotopes here than in other major continents (Gat and Matsui, 1991). Improvement of isotope monitoring in large river basins (e.g. Gibson et al., 2002) may support a wide range of programmes exploring water budget management and modelling (Rozanski et al., 1993; Gat, 2000). In the Canadian Arctic, Gibson (2002) showed that stable water isotopes offer a means of monitoring the partitioning of evaporation and transpiration in basin-integrated discharge signals while, in semi-arid Australia, the downstream enrichment of heavier isotopes can be used to gauge progressive evaporative loss (Stone et al., 2003). The total latent fluxes in humid basins such as the Amazon can be much larger than the isotopically non-fractionating plant-mediated transpiration, resulting in evaporative enrichment of atmospheric moisture (Victoria et al., 1991; Henderson-Sellers et al., 2002). This paper extends the results of Henderson-Sellers et al. (2002) using stable water isotope data from Izobamba from 1972 up to 2001, to evaluate possible changes in the hydrological cycle over this time. 2.2. Application of stable isotopes to detecting Amazonian climatic variations The Amazon Basin recycles about half its rainfall with a water recycling time of about 5.5 days (Salati et al., 1979). Although field measurements of isotopes within the canopy (Leopoldo,
42
A. Henderson-Sellers and K. McGuffie
1981) are somewhat contradictory, the most likely source of observed differences is believed to be different moisture sources (e.g. Matsui et al., 1983; Vuille et al., 2003a) and the fate of water intercepted by the canopy. In a complete simulation of the Amazon’s forest hydrology, the land surface must correctly partition the moisture fluxes between water evaporation (fractionating), transpiration (non-fractionating), re-evaporated canopy intercepted rainfall (nonfractionating if complete) and runoff fraction (Martinelli et al., 1996). Taking advantage of the relatively straightforward nature of the atmospheric circulation over the Amazon Basin (e.g. Jones and Carvalho, 2002; Foley et al., 2002), Gat and Matsui (1991) employed a simple steady state model of the Amazon to demonstrate that some of the water recycling is from isotopically fractionating sources. Comparing their model with pre-1981 data from the International Atomic Energy Agency/World Meteorological Organization (IAEA/WMO) Global Network for Isotopes in Precipitation (GNIP, 1999), they suggested that 20–40% of the recycled moisture within the basin is derived from fractionating evaporation from sources such as lakes, the river or standing water. Victoria et al. (1991) combined isotopic observations from Belem and Manaus (1972–1986) and the model of Dall’Olio (1976) and Salati et al. (1979) to confirm that such fractionating evaporation contributes up to 40% of the total evaporation during the dry season (see also Gat, 2000). Henderson-Sellers et al. (2002) detected statistically significant temporal changes (1965– 1990) in water isotopic signatures in the central Amazon. Differences determined in deuterium excess were found to be consistent with recent GCM simulations only if there had been a relative increase in evaporation from non-fractionating water sources over the investigated period. They found no significant change in dry season isotopic characteristics despite earlier predictions of such land-use change impacts and concluded that the pre-1990 Amazonian stable isotope record is more consistent with the predicted effects of greenhouse warming (possibly combined with forest removal) than with the model-predicted effects of deforestation alone (Henderson-Sellers et al., 2002). Vuille et al. (2003a, 2003b) emphasize that δ 18 O is influenced by several factors but suggest that the dominant mode of ENSO (El Niño – Southern Oscillation) is likely to be important. However, Botta et al. (2002) used principal component analysis of Amazonian climate data to show that ENSO is not the major, or even the most important, mode of climate variability in the Amazon Basin. They found that the dominant mode of climate variability in the Amazon, which explains about 35% of the inter-annual variance of precipitation and 56% of the temperature variance, has a period of 24–28 years. By comparison, ENSO explains only 21% of the total variance in annual mean precipitation and temperature.
3. Precipitation isotopes in the Andean Amazon Stable isotopic data in rainfall continue to be collected by the IAEA/WMO in the Amazon Basin as part of a global monitoring program at around 550 stations worldwide (e.g. Rozanski et al., 1993; Gibson et al., 2002). Of the various stations that have operated within the area of the Amazon Basin, the most recent and extensive data sets are available from Izobamba (0.37◦ S, 78.55◦ W and 3058 m ASL) on the eastern flank of the Andes, situated in the upper part of the Amazon Basin. IAEA data now available for Izobamba bring the station record
Shift in stable water isotopes in precipitation in the Andean Amazon
43
Fig. 2. Monthly mean δ 18 O values (h) from Izobamba for all months (January–December) from 1972 to 2001 with least square linear trend lines shown in each monthly panel.
up to date and provide a record of 30 years (1972–2001), covering the periods of intense deforestation since the early 1970s (Fig. 1(a)) (McGuffie and Henderson-Sellers, 2004). Analysis of data for Izobamba shows upper Amazon Basin isotopic ratios shifting over 25 years (Fig. 2). The recent data show that the wet season (November–May) isotopic ratios have tended to become more negative while, in the dry season (June–October), ratios have tended to become more positive (Fig. 2). Although these changes are not all statistically significant (only May, June, July and November are statistically significant: shown by the black bars in Fig. 3), analysis of trend information for each month shows coherence in month-tomonth changes (Fig. 2). Seasonality in isotopic depletions in the upper part of the Amazon Basin, as represented by Izobamba, has increased over the last 25 years. Table 2 shows that at Izobamba the dry season depletions have decreased and the wet season depletions have increased. The dry season changes in both isotopic ratios are statistically significant. The wet season months where statistically significant changes are seen are those flanking the dry season (i.e. May and November) – see Fig. 3. These seasonal changes in 18 O and D signatures between 1972–1986 and 1987–2001 may indicate the impact of deforestation in the wet season: vegetation removal prompting less recycling and less re-insertion of heavy isotopes into the basin hydrological system. Figs. 2 and 3 suggest a shift away from non-fractionating sources (to fractionating sources) in the wet season and less effective recycling (i.e. a smaller fraction of precipitation is re-evaporated) of water (from fractionating sources) in the dry season. The hypothesis that these seasonally opposing changes could be the result of forest removal is examined in the next section using a simple model that allows for both fractionating evaporation (e.g. from lakes) and nonfractionating (transpiration).
44
A. Henderson-Sellers and K. McGuffie
Fig. 3. δ 18 O in precipitation differences (1987–2001 minus 1972–1986) at Izobamba together with monthly mean wind roses (from NCEP reanalysis, Kanamitsu et al., 2001) and mean total precipitation (mm). Statistical significance (95% level) in the δ 18 O differences is shown by black bars (significant difference) cf. gray (not significantly different). The wind roses show that the dominant air flow is from the forested basin (see Fig. 1(b)). Table 2 Wet (November–May) and dry (June–October) season average depletions of δ 18 O and δD (h) from precipitation at Izobamba between two periods (1972–1986 and 1987–2001) together with the full time extent standard deviation. Dry season differences are statistically significant (shown in bold) at the 95% level, with differences exceeding ±1 SD Season/Period
δ 18 O
Diff.
SD
δD
Diff.
SD
Wet (November–May) 1972–1986 1987–2001
−11.0 −12.0
−1.0
0.5 0.6
−77.8 −83.4
−5.6
4.8 4.8
Dry (June–October) 1972–1986 1987–2001
−10.6 −9.2
+1.4
0.5 0.4
−75.0 −61.3
+13.7
3.6 3.2
4. Modelling water isotopes at the basin scale No GCM stable water isotopic models have yet been applied to tropical deforestation simulations that include changes in lakes and open water surfaces (Henderson-Sellers et al., 2002; Vuille et al., 2003a, 2003b). However, Gat and Matsui (1991) developed a simple basin-scale representation of the relative impact of transpiration (and canopy re-evaporation) as compared to open water evaporation: the latter being fractionating while the former are not. The model gives the isotopic ratio in the precipitation, δp as δp = δp,o + γ log f,
(1)
where γ =
(l + (1 − h)(l + r))ε + l(1 − h)Ck , l + r(1 − h)
(2)
Shift in stable water isotopes in precipitation in the Andean Amazon
45
l is the fraction of the precipitation which is re-evaporated from lakes and other open water surfaces, r is the total runoff fraction of the basin, ε is the equilibrium isotopic fractionation factor, Ck is the kinetic fractionation factor and h is the relative humidity. The fraction transpired by the forest is therefore t = 1 − r − l. The quantity f represents the amount of ‘pristine’ oceanic moisture that remains after the advection over the forest canopy and δp,o the depletion of precipitation not influenced by the forest recycling of moisture. This model is applied to a location representative of the Andean Amazon which receives moisture that has passed over the Amazon Basin (e.g. Izobamba). It is assumed that only ∼20% of the moisture is unchanged from its oceanic composition and the relative humidity is ∼70% (i.e. f = 0.2 and h = 0.7). Using Equations (1) and (2), the relative impacts of greater lake evaporation, a larger transpiration fraction and greater runoff fraction in Fig. 4(a) can be compared with the shift in the Izobamba GNIP data in Fig. 4(b), suggesting explanations for the observed changes in isotopic depletions. The relationship of 18 O and D in precipitation has changed between the 1970s and the first few years of this century at Izobamba in three distinct ways: (i) values have become more depleted; (ii) the distribution slope has become less steep; and (iii) seasonal separation has become greater as the dry season precipitation becomes less depleted while the wet season precipitation has become more depleted. These alterations can be seen in Fig. 4(b) and interpreted by reference to Fig. 4(a). The lengthening of the data distribution in Fig. 4(b) especially towards more depleted values in the 1990s and 2000s is consistent with less non-fractionating recycling (i.e. less transpiration and/or canopy re-evaporation): a likely result of deforestation. The local D/18 O slope for monthly means shifts from 8.1 to 7.8. This change, although not statistically significant, is
Fig. 4. (a) Model results showing schematically the relative impacts of changes in transpiration (t), lake evaporation (l) and runoff (r) on the δ 18 O and δD relationship. Open block arrows show the effect of modifying each of these parameters. (b) Scatter plot of all available GNIP data for Izobamba. Temporal changes in the data distribution include a greater range in isotopic depletions, especially a tendency towards greater depletions recently (more points to the lower left in the 1990s and 2000s) and seasonal shifts. The latter are shown as pairs (solid filled circles) of four-monthly mean values for decadal wet and dry seasons (1970s and 2000s) displayed by offsetting the δD values by ±30h for the dry and wet season respectively. These seasonal movements are consistent with the model results in Fig. 4(a) of less non-fractionating evaporation (transpiration and/or canopy re-evaporation) in the wet season and more non-fractionating (transpiration) and/or less fractionating (open water and lake evaporation) in the dry season.
46
A. Henderson-Sellers and K. McGuffie
consistent with more evaporation from lakes or a greater runoff fraction (Fig. 4(a)), i.e. a result expected due to less vegetation impact on isotopic depletion. The seasonal ‘separation’ can be seen most readily by considering the (18 O, D) changes in, for example, the averages of the four wettest and four driest months in the decades of the 1970s and 2000s. In the dry season, these alter respectively from (−11.2, −79.8) to (−10.4, −74.1) and in the wet season from (−12.9, −92.5) to (−15.1, −110.8). These changes are illustrated in Fig. 4(b) by the pairs of points vertically displaced from the data distribution by ±30h of D for clarity. These movements underline the seasonal trends shown in Table 2: statistically significant in the dry season and clear (but not statistically significant) in the wet season. The results in Figs. 2 and 3 show: (i) during the dry season, isotopic ratios become less negative (statistically significant) at Izobamba (in air affected by the Amazon, e.g. Fig. 3); and (ii) opposite changes in the wet season. Comparing the catchment model with Izobamba data suggests that the recently observed increased data spread could be explained by less nonfractionating evaporation (transpiration and canopy re-evaporation) by the forest canopy over which the moisture has traveled and that a tendency for a decrease in the slope of the local meteoric line might imply either more lake evaporation or a greater runoff fraction (Fig. 4). These seasonal shifts suggest less non-fractionating evaporation (transpiration and/or canopy re-evaporation) in the wet season and more non-fractionating (transpiration) and/or less fractionating (from open water and lake evaporation) in the dry season.
5. Recent impacts of Amazonian deforestation Henderson-Sellers et al. (2002) proposed that their observed isotopic changes might be due to greenhouse intensification of the hydrological cycle masking any land-use change impact. Alternative explanations for their results include: isotope data to 1990 only were available; the statistically significant wet season changes reported might be related to ENSO events or other climatic variations that modify the regional circulation and hence affect the moisture climatology of the Amazon (cf. Vuille et al., 2003b; Botta et al., 2002); that no information on fluxes from simulated open water as a surface type in the Amazon GCM experiments was considered; and that the selected model sets failed to correctly simulate the relative components of transpiration and re-evaporated canopy interception at least in the Amazon dry season and perhaps throughout the year (Henderson-Sellers et al., 2004). Two recent assessments of the area of surface water suggest that its extent is significant. Richey et al. (2002) found that in May around 20% of the main Amazon River area is flooded. Specifically, they determined that, typically, the region is most flooded in May (350,000 km2 ), with an annual mean flooded area of 250,000 km2 (Richey et al., 2002). Foley et al. (2002) used the HYDRA land-surface scheme to simulate the extent of flooding in neutral (i.e. neither El Niño nor La Niña) years and found the areal range to be from 20,378 to 170,079 km2 . They also determined that, while both El Niños and La Niñas enlarge the minimum, La Niñas also increase the maximum (by 13,979 km2 ) while in El Niños the maximum area flooded is reduced. Overall, the estimates of open water range from about 350,000 km2 to about half this: an area that must be included in isotopic and other hydrological modelling. During the wet season, the isotopic ratios of 18 O and D in precipitation at Izobamba have become more negative over the period from the mid-1970s to 2000. These changes could
Shift in stable water isotopes in precipitation in the Andean Amazon
47
be caused by (i) an increase in the fraction of precipitation that appears as runoff fraction, (ii) a reduction in the fraction of recycling that occurs through non-fractionating processes, i.e. transpiration, and/or (iii) an increase in recycling that occurs through fractionating processes, i.e. evaporation from open bodies of water. Causes (i), (ii) and possibly (iii) could reasonably be associated with a loss of forest cover and thus be an isotopic signal of deforestation. During the dry season, isotopic ratios of 18 O and D in precipitation at Izobamba have changed in the opposite direction over the same period. These changes could be caused by a decrease in the fraction of water recycled through fractionating processes, i.e. evaporation from open bodies of water. These results are either complementary or contradictory to those of Henderson-Sellers et al. (2002) who showed a likely impact in the central Amazon due to an apparent intensification of hydrologic cycling, itself, perhaps, the result of greenhouse warming. To determine which of these apparent causes of observed isotopic precipitation disturbances is correct or whether both need to be invoked will require the use of isotopically-enabled Global Climate Models including appropriately complex and isotopically-enabled land-surface schemes. Such models could establish if the observed trends are more than station-specific and also examine alternative explanations including large-scale circulation changes.
References Achard, F., Eva, H.D., Stibig, H.-J., Mayaux, Ph., Gallego, J., Richards, T., Malingreau, J.-P. (2002). Determination of deforestation rates of the world’s humid tropical forests. Science 297, 999–1002. Botta, A., Ramankutty, N., Foley, J.A. (2002). Long-term variations of climate and carbon fluxes over the Amazon basin. Geophysical Research Letters 29, 1319, doi:10.1029/2001GL013607. Costa, M.H., Foley, J.A. (2000). Combined effects of deforestation and doubled atmospheric CO2 concentrations on the climate of Amazonia. Journal of Climate 13, 18–34. Dall’Olio, A. (1976). A composicao isotopica das precipitacoes de Brasil: modelos isotermicos e a influencia da evapotranspiracao na Bracia Amazonica. MSc. thesis. Piracicaba, Universidada de São Paulo, 180 pp. (In Portuguese.) Fearnside, P.M. (1987). Causes of deforestation in the Brazilian Amazon. In: Dickinson, R.E. (Ed.), The Geophysiology of Amazonia Vegetation and Climate Interactions. Wiley, New York, pp. 37–61. Foley, J.A., Botta, A., Coe, M.T., Costa, M.H. (2002). El Niño – Southern Oscillation and the climate, ecosystems and rivers of Amazonia. Global Biogeochemical Cycles 16 (4), 1132, doi:10.1029/2002GB001872. Gat, J.R. (1996). Oxygen and hydrogen isotopes in the hydrological cycle. Annual Review of Earth and Planetary Sciences 24, 225–262. Gat, J.R. (2000). Atmospheric water balance – The isotopic perspective. Hydrological Processes 14, 1357–1369. Gat, J.R., Matsui, E. (1991). Atmospheric water balance in the Amazon Basin: An isotopic evapotranspiration model. Journal of Geophysical Research 96 (D7), 13,179–13,188. Gibson, J.J. (2002). Short-term evaporation and water budget comparisons in shallow Arctic lakes using non-steady isotope mass balance. Journal of Hydrology 264, 242–261. Gibson, J.J., Edwards, T.W.D. (2002). Regional water balance trends and evaporation–transpiration partitioning from a stable isotope survey of lakes in northern Canada. Global Biogeochemical Cycles 16 (2), 1026, doi:10.1029/2001GB001839. Gibson, J.J., Aggrawal, P., Hogan, J., Kendall, C., Martinelli, L.A., Stichler, W., Rank, D., Goni, I., Choudury, M., Gat, J., Bhattacharya, S., Sugimoto, A., Fekete, B., Pietroniro, A., Maurer, T., Panarello, H., Stone, D., Seyler, P., Maurice-Bourgoin, L., Herzceg, A. (2002). Isotope studies in large river basins: A new global research focus. EOS 83 (52), 613–617. Glantz, M.H., Brook, A.T., Parisi, P. (1997). Rates and processes of Amazon deforestation. Available at http://www.isse.ucar.edu/rates/index.html.
48
A. Henderson-Sellers and K. McGuffie
GNIP, Global Network for Isotopes in Precipitation (1999). The IAEA/WMO GNIP Database, Release 3, October 1999. Available at http://isohis.iaea.org/search.asp. Henderson-Sellers, A., Gornitz, V. (1984). Possible climatic impacts of land cover transformations, with particular emphasis on tropical deforestation. Climatic Change 6, 231–258. Henderson-Sellers, A., McGuffie, K., Zhang, H. (2002). Stable isotopes as validation tools for global climate model predictions of the impact of Amazonian deforestation. Journal of Climate 15, 2664–2677. Henderson-Sellers, A., McGuffie, K., Noone, D., Irannejad, P. (2004). Using stable water isotopes to evaluate basinscale simulations of surface water budgets. Journal of Hydrometeorology 5 (4), 805–822. Hoffman, G., Jouzel, J., Masson, V. (2000). Stable isotopes in atmospheric general circulation models. Hydrological Processes 14, 1385–1406. INPE (1998). Average annual deforestation rate in the legal Amazon. Available at http://www.inpe.br/amz-04.htm. Jones, C., Carvalho, L.M.V. (2002). Active break phases in the South American monsoon system. Journal of Climate 15, 905–914. Joussaume, S., Sadourny, R., Jouzel, J. (1984). A general circulation model of water isotopes cycles in the atmosphere. Nature 311, 24–29. Kanamitsu, M., Kousky, V., van den Dool, H., Jenne, R., Fiorino, M. (2001). The NCEP–NCAR 50-year re-analysis: Monthly means CD-ROM and documentation. Bulletin of the American Meteorological Society 82, 247–268. Leopoldo, R.R. (1981). Aspetos hidrologicos at florista amazonica denga na regiao de Manaus. PhD thesis. University National Estado São Paulo, Botucato SP, Brasil, 180 pp. McGuffie, K., Henderson-Sellers, A. (2004). Stable water isotope characterization of human and natural impacts on land-atmosphere exchanges in the Amazon basin. Journal of Geophysical Research – Atmospheres 109, D17104, doi:10.1029/2003JD004388. McGuffie, K., Henderson-Sellers, A., Zhang, H. (1998). Modelling climatic impacts of future rainforest destruction. In: Maloney, B.K. (Ed.), Human Activities and the Tropical Rainforest. Kluwer, Dordrecht, The Netherlands, pp. 169–193. Martinelli, L.A., Victoria, R.L., Sternberg, L.S.L., Ribeirio, A., Moreira, M.Z. (1996). Using stable isotopes to determine sources of evaporated water to the atmosphere in the Amazon basin. Journal of Hydrology 183, 191–204. Matsui, E., Salati, E., Ribeiro, M.N.G., Reis, C.M., Tancredi, A.C.S.N.F., Gat, J.R. (1983). Precipitation in the Central Amazon Basin: The isotopic composition of rain and atmospheric moisture at Belem and Manaus. Acta Amazonica 13, 307–369. Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.-M., Basile, I., Bender, M., Chappellaz, J., Davisk, M., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V.Y., Lorius, C., Pépin, L., Ritz, C., Saltzmank, E., Stievenard, M. (1999). Climate and atmospheric history of the past 420000 years from the Vostok ice core, Antarctica. Nature 399, 429–436. Richey, J.E., Melack, J.M., Aufdenkampe, A.K., Ballester, V.M., Hess, L.L. (2002). Outgassing from Amazonian rivers and wetlands as a large tropical source of atmospheric CO2 . Nature 416, 617–620. Rozanski, K., Araguas-Araguas, L., Gonfiantini, R. (1993). Isotopic patterns in modern global precipitation. Climate Change in Continental Isotopic Records. Geophysical Monographs 78, 1–36. Salati, E., Vose, P.B. (1984). Amazon Basin: A system in equilibrium. Science 225, 129–137. Salati, E., Olio, A.D., Matsui, E., Gat, J.R. (1979). Recycling of water in the Amazon Basin: An isotopic study. Water Resources Research 15 (5), 1250–1258. Stone, D.W., Henderson-Sellers, A., Airey, P., McGuffie, K. (2003). Murray Darling basin isotope observations: An essential component of the Australian CEOP. EOS Transactions AGU 84 (46). AGU Fall Meeting Suppl., H22I-08, December 8–12, 2003. Townsend, A.R., Asner, G.P., White, J.W.C. (2002). Land use effects on atmospheric 13 C imply a sizable terrestrial CO2 sink in tropical latitudes. Geophysical Research Letters 29 (10), 1426, doi:10.1029/2001GL013454. Victoria, R.L., Martinelli, L.A., Mortatti, J., Richey, J. (1991). Mechanisms of water recycling in the Amazon Basin: Isotopic insights. Ambio 20 (8), 384–387. Vuille, M., Bradley, R.S., Werner, M., Healy, R., Keimig, F. (2003a). Modelling δ 18 O in precipitation over the tropical Americas: 1. Interannual variability and climatic controls. Journal of Geophysical Research 108 (D6), 4174, doi:10.1029/2001JD002038. Vuille, M., Bradley, R.S., Healy, R., Werner, M., Hardy, D.R., Thompson, L.G., Keimig, F. (2003b). Modelling δ 18 O in precipitation over the tropical Americas: 2. Simulation of the stable isotope signal in Andean ice cores. Journal of Geophysical Research 108 (D6), 4175, doi:10.1029/2001JD002039.
Shift in stable water isotopes in precipitation in the Andean Amazon
49
Watson, R.T., Noble, I.R., Bolin, B., Ravindranath, N.H., Verardo, D.J., Dokken, D.J. (2000). Land Use, Land Use Changes and Forestry: A Special Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, 388 pp. Zhang, H., Henderson-Sellers, A., McGuffie, K. (2001). The compounding effects of tropical deforestation and greenhouse warming on climate. Climatic Change 49, 309–338.
This page intentionally left blank
2. Oceanic radionuclide tracers
This page intentionally left blank
53
Southern Hemisphere Ocean Tracer Study (SHOTS): An overview and preliminary results M. Aoyamaa,* , M. Fukasawab , K. Hirosea , R.F.C. Mantourac , P.P. Povinecc , C.S. Kimd , K. Komurae a Geochemical Research Department, Meteorological Research Institute (MRI), Tsukuba 305-0052, Japan b Ocean Research Department, Japan Agency for Marine-Earth Science and Technology (JAMSTEC),
Yokosuka, Kanagawa 237-0061, Japan c Marine Environment Laboratory, International Atomic Energy Agency (IAEA-MEL), MC-98000, Monaco d Environmental Radioactivity Assessment Department, Korea Institute of Nuclear Safety (KINS),
Seoul, Republic of Korea e Low-Level Radioactivity Laboratory (LLRL), Institute of Nature and Environmental Technology,
Wake, Tatsunokuchi, Ishikawa 923-1224, Japan Abstract Approximately 900 seawater samples were collected in the subtropical gyres in the South Pacific, the South Atlantic and the South Indian Ocean in 2003–2004 during the BEAGLE2003 cruise. Preliminary results show that 137 Cs and 239,240 Pu concentrations in surface waters were comparable with those in the subtropical gyres in the North Pacific. The 137 Cs profile shows a smaller effect of subduction on water mass transport. The estimated 137 Cs inventory of 800 Bq m−2 is about a half of the inventory in the North Pacific at the same latitude. The 239,240 Pu concentration at sub-surface maximum is one order of magnitude lower than that observed in the North Pacific, which is a remarkable difference when compared with 137 Cs profiles. Measured 240 Pu/239 Pu atom ratios are 0.20–0.24, statistically not different from the global fallout ratio. The SHOTS collaboration will produce a comprehensive dataset on anthropogenic radionuclides in the Southern Ocean. The obtained results will contribute to better understanding of processes in the water column using radionuclides as tracers, and they will improve our knowledge of circulation processes in the ocean, important for better understanding of climate change. Keywords: Anthropogenic radionuclides, 137 Cs, Plutonium isotopes, SHOTS, Seawater, Water column, Pacific Ocean, Atlantic Ocean, Indian Ocean, Southern Ocean
1. Introduction The main source of anthropogenic radionuclides in the world ocean is global fallout from atmospheric nuclear tests carried out between 1945 and 1980. Several global studies have been carried out in the world ocean (e.g. GEOSECS (Bowen et al., 1980), WOCE (Schlosser et al., * Corresponding author. Address: MRI, 1-1 Nagamine, Tsukuba 305-0052, Japan; phone: (+81) 29 853 8719; fax: (+81) 29 853 8728; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08004-6
© 2006 Elsevier Ltd. All rights reserved.
54
M. Aoyama et al.
Fig. 1. BEAGLE2003 cruise track.
1999), WOMARS (Povinec and Togawa, 1998; Povinec, 2003)), however, anthropogenic radionuclides in the Southern Ocean have not been studied well and data on the distribution, inventory and behavior of anthropogenic radionuclides have been missing (Hamilton et al., 1996; Livingston and Povinec, 2000). Radionuclide data stored in the IAEA’s Global Marine Radioactivity Database (GLOMARD; IAEA, 2000; Povinec et al., 2005), as well as in the Meteorological Research Institute’s (MRI) HAM Pacific database (Aoyama and Hirose, 2004), include only a few hundred records for 137 Cs concentrations in seawater of the Southern Ocean, compared to several ten thousand records for the Northern Hemisphere Oceans. Radionuclide tracer studies in the Southern Hemisphere Oceans continue to be a major objective in worldwide marine radioactivity research. The ship opportunity to carry out sampling in the Southern Ocean was available during the Blue Earth Global Expedition 2003 (BEAGLE2003) conducted on the R/V Mirai during August 2003–March 2004 (Fig. 1) by the JAMSTEC. The BEAGLE2003 expedition was following the WOCE track in the Southern Ocean, revisiting WOCE stations with the aim of determining any changes in oceanographic parameters (Wijffels et al., 1998), distribution of 14 C and other tracers in the water column. In connection with the BEAGLE2003 expedition a project on the South Hemisphere Ocean Tracer Study (SHOTS) has been developed with the aim to study the distribution of radioactive and stable isotopes, and other non-radioactive tracers in the Southern Ocean, which plays a dominant role in the Earth climate. The project will focus on analysis of radionuclides in water samples for better understanding of circulation and water column processes. The main target radionuclides with the most frequent analyses will be tritium, radiocarbon, 90 Sr, 137 Cs and plutonium isotopes, however, other radionuclides will be analyzed as well, supported by oceanographic and non-radioactive tracer analyses such as nutrients. The co-analysis of nutrients and O2 with radionuclides offers a unique opportunity to determine the in situ phys-
Southern Hemisphere Ocean Tracer Study (SHOTS): An overview and preliminary results
55
ical and biogeochemical dynamics of anthropogenic radionuclides in the South Hemisphere oceans. The SHOTS international collaboration will produce a comprehensive dataset on anthropogenic radionuclides in the Southern Ocean. A complete hydrographic dataset from the BEAGLE2003 cruise will be produced by JAMSTEC. The obtained results will contribute to better understanding of processes in the water column using radionuclides as tracers, as well as for studying circulation processes in the ocean towards a better understanding of climate change. In this paper, we present and discuss preliminary results obtained from the South Pacific Ocean.
2. Sampling and analyses 2.1. Sampling onboard Sampling work in the subtropical gyres in the Southern Hemisphere Oceans was carried out at 93 stations which included both surface and water column stations (Fig. 1). The average distance for surface water stations was about 270 km. A CTD (SBE 9 plus) with a Rosette Multi-bottle Sampler (RMS) was used to measure oceanographic parameters at 491 stations throughout the cruise. The cruise details for hydrographic observations are available at the JAMSTEC web site (http://www.jamstec.go.jp/mirai/2003/data_2003.html). CTD/RMS water profile sampling for radionuclide measurements was done in 57 of 491 stations. The sampling depths at each station were 1, 100, 200, 400, 600, 800 and 1000 dbar, and approximately every 500 dbar below 1000 dbar (e.g. 1400, 2000, 2400, 3000, 3500, 4000, etc., down to the sea bottom). The location of stations and sampling depths for samples collected in the South Pacific Ocean are shown in Fig. 2. The sample volumes for surface water were around 85 L, while water column samples varied from 5 to 20 L depending on availability. The total number of samples is close to 900 (Table 1) with a total weight of about 22,000 kg. Seawater samples for radionuclide analyses were filtered onboard through a membrane filter with a pore size of 0.45 µm (Millipore HA), and acidified using nitric acid (Aoyama et al., 2000), while samples for tritium measurements were stored in 1 liter bottles without acidification. 2.2. Radionuclide analyses 239,240 Pu, 137 Cs and 90 Sr analyses of collected samples have been carried out in MRI (Aoyama
et al., 2000; Hirose et al., 2001), IAEA-MEL (La Rosa et al., 2001; Lee et al., 2001; Povinec et al., 2001), KINS (Kim et al., 2000, 2002) and LLRL (Komura, 1997; Hirose et al., 2005) following their respective radiochemical procedures. At LLRL, in order to achieve very low background in gamma spectrometry, high efficiency well type HPGe detectors, shielded with very low activity lead, were placed in the Ogoya underground facility, specially designed for low level counting. The detection limit for 137 Cs is 0.18 mBq for a counting time of 1 week. Since the volume of water column samples is relatively small (5–20) liters, the deep water samples are measured at Ogoya underground facility to obtain reliable results for 137 Cs.
56
M. Aoyama et al.
Fig. 2. SHOTS sampling locations and depths for the Pacific Ocean. Table 1 SHOTS samples
Surface Water column
Total
Volume (L)
Pacific Ocean
85 20 10–18 <10
51 133 19 218 421
Atlantic Ocean
Indian Ocean
Total
20 63 14 98
22 67 6 171
93 263 39 487
195
266
882
The samples from shallow and medium depths are measured at MRI. IAEA-MEL will also analyze 137 Cs in water samples using large volume HPGe well detectors in their underground facility. Plutonium in seawater samples analyzed at MRI was concentrated by co-precipitation with iron hydroxide after addition of FeCl3 , K2 S2 O3 and 242 Pu in acidified seawater. The plutonium isotopes (238 Pu and 239,240 Pu) were determined by α-spectrometry (OctêteTM (CANBERRA), SSD of 25 keV (FWHM)) after the chemical separation and purification using anion exchange resins (Dowex 1 × 2) (Hirose and Sugimura, 1985). The detection limit of 239,240 Pu in seawater samples, under the present analytical conditions (counting time: 106 s), is 0.5 mBq m−3 . Chemical yield of plutonium was in the range of 50–80%. Measurements of 240 Pu/239 Pu atom ratios were carried out by high-resolution inductively coupled plasmamass spectrometry (ICP-MS; PlasmaTrace 2, Micromass), described in detail elsewhere (Kim
Southern Hemisphere Ocean Tracer Study (SHOTS): An overview and preliminary results
57
et al., 2000, 2002). After the radiochemical separation (using anion exchange resins), a part the solution sample was injected into the high-resolution ICP-MS system by on-line purification and sequential injection technique including Sr-Spec and TEVA Spec resins. The sensitivity of high-resolution ICP-MS was 5.5 × 107 s−1 /ppb based on the measurement of 239 Pu. Background peak areas of 239 Pu and 240 Pu were 3.3 and 1.3 s−1 , respectively. To assure reliability of results, an intercomparison exercise has been carried out between participating laboratories using the IAEA-MEL reference material IAEA-381 (Irish Sea water). The intercomparison results for 137 Cs, 239 Pu, 240 Pu and 239,240 Pu were within the 95% confidence intervals.
3. Results and discussion As there are still many samples from the BEAGLE2003 expedition to be analyzed, in this paper we present preliminary results only for the South Pacific Ocean, and compare them with previous studies. We also compare and discuss our data with the data obtained for the North Pacific Ocean. 3.1.
137 Cs
and 239,240 Pu in surface water
137 Cs
concentrations in surface waters along the cruise track in 2003/2004 ranged from 1.2 to 1.5 Bq m−3 (Fig. 3). The 137 Cs concentrations in the South Pacific Ocean were the same order of magnitude as those in the North Pacific (Aoyama et al., 2001b; Hirose and Aoyama, 2003;
Fig. 3. 137 Cs and 239,240 Pu in surface waters in the subtropical gyres (20–32◦ S) in the South Pacific Ocean.
58
M. Aoyama et al.
Fig. 4. Temporal variation of 137 Cs activity in surface water in the region 37.5–25◦ S, 155◦ E–120◦ W.
Povinec et al., 2003). The spatial distribution of 137 Cs concentrations in surface waters appear to be homogeneous in the South Pacific and it has been decreasing since the early 1960s to the present as shown in Fig. 4. In the 1960s, relatively high 137 Cs levels were observed (3–10 Bq m−3 ) due to global fallout (Folsom, 1979; Saruhashi et al., 1975), which then decreased to 2.5 Bq m−3 in the 1990s (Bourlat et al., 1996). In 2003, the 137 Cs concentrations in surface waters decreased to 1.2–1.5 Bq m−3 . The obtained apparent half residence time of 137 Cs in surface water for this region during the period from 1965 through 2003 is about 20 years. This is in good agreement with 15–20 years in the subtropical gyres in both the North and the South Pacific, which were obtained in previous studies (Hirose and Aoyama, 2003; IAEA, 2005; Povinec et al., 2005). 239,240 Pu concentrations in surface waters in the South Pacific ranged from 1.2 ± 0.3 to 2.9 ± 0.6 m Bq m−3 . The surface 239,240 Pu concentration in the South Pacific, which showed a relatively larger variation compared with 137 Cs, was the same order of magnitude as that of the North Pacific. The 239,240 Pu concentrations observed in 2004 were almost the same as those in French Polynesia in 1993. 3.2.
137 Cs
and 239,240 Pu in the water column
The concentrations of 137 Cs and 239,240 Pu, and the 240 Pu/239 Pu atom ratios in filtered seawater samples at station P06C-175 (32.5◦ S, 177.7◦ E) are presented in Table 2. The 137 Cs profile in the water column of the western South Pacific Ocean in 2003 showed a weak subsurface maximum at 100 dbar depth, where the 137 Cs concentration was 1.52 ± 0.16 Bq m−3 , which then decreased gradually to 1000 dbar, where the 137 Cs concentration was 0.15 ± 0.08 Bq m−3 , one tenth of that observed at the surface. The 137 Cs concentration was at minimum at 2000 dbar, and then increased to 0.031 ± 0.017 Bq m−3 at 5000 dbar, as shown in Fig. 5. Although the
Southern Hemisphere Ocean Tracer Study (SHOTS): An overview and preliminary results
59
Table 2 Concentrations of 137 Cs and 239,240 Pu in the water column at station PO6C-175 in the South Pacific Ocean (32.5◦ S, 177.7◦ E) Sampling date 18 Aug. 2003 18 Aug. 2003 18 Aug. 2003 18 Aug. 2003 18 Aug. 2003 18 Aug. 2003 18 Aug. 2003 18 Aug. 2003 18 Aug. 2003 18 Aug. 2003
Depth (dbar) 0 100 200 400 600 800 1000 2000 3000–4000 5000
137 Cs (Bq m−3 )
239,240 Pu (mBq m−3 )
240 Pu/239 Pu atom ratio
1.26 ± 0.05MR
1.41 ± 0.24K 1.1 ± 0.3MR
0.22 ± 0.04K
4.69 ± 0.66K
0.22 ± 0.03K
4.81 ± 0.67K 1.96 ± 0.59K
0.24 ± 0.03K 0.20 ± 0.06K
1.52 ± 0.16MR 1.04 ± 0.13MR 0.53 ± 0.10MR 0.53 ± 0.16MR 0.29 ± 0.09MR 0.15 ± 0.08MR 0.011 ∓ 0.020LL 0.016 ± 0.008LL 0.031 ± 0.017LL
K KINS; LL LLRL, Kanazawa Univ.; MR MRI.
Fig. 5. 137 Cs profiles in the western South Pacific.
vertical 137 Cs profiles in the subtropical gyre in the North Pacific Ocean show a clear subsurface maximum (Hirose and Aoyama, 2003), the 137 Cs profiles at station P06C-175 in the South Pacific (Fig. 5) do not show such a clear subsurface maximum, probably due to water
60
M. Aoyama et al.
transport by subduction. We shall discuss this phenomenon in detail later when more 137 Cs profiles will be available. Maximum concentrations of 137 Cs in the subtropical gyre in 2002 in the North Pacific were 2–2.5 Bq m−3 , while in the South Pacific they were lower by a factor of two. We calculated the water column inventory of 137 Cs using the data obtained for station P06C-175. The 137 Cs inventory is 790 ± 80 Bq m−2 , about half of the corresponding subtropical North Pacific 137 Cs inventory. The lower 137 Cs inventory in the South Pacific reflects the geographical distribution of global fallout, although the 137 Cs water column inventory is largely affected by advection (Livingston et al., 2001; Aoyama et al., 2001a; Livingston and Povinec, 2002). The concentrations of 239,240 Pu in the water column gradually increased from the surface to the sub-surface maximum at 600–1000 m, where 239,240 Pu concentrations of 4.7 ± 0.7 mBq m−3 at 600 dbar depth, and 4.8 ± 0.7 mBq m−3 at 1000 dbar depth were observed, after which they decreased with depth (Table 3). Near the Rangiroa Atoll in French Polynesia, the 239,240 Pu maximum in 1993 was found at 500–600 m with concentrations around 9 mBq m−3 (Chiappini et al., 1999). At the GEOSECS station GX-263 (16.7◦ S, 167.1◦ W) in 1974, the 239,240 Pu maximum was found at 150–600 m, where the concentrations were 5–20 mBq m−3 (Livingston, 1985). The mid-depth maximum of the 239,240 Pu concentration has been observed in the South Pacific Ocean during the past three decades, similarly as in the North Pacific, although the 239,240 Pu maxima were one order of magnitude lower than those observed in the North Pacific Ocean (Table 3). The subsurface Pu maximum has been explained by the vertical transport of Pu associated with sinking particles, and the subsequent release of Pu from particulate matter to the water column (Bowen et al. 1971, 1980; Livingston and Anderson, 1983; Fowler et al., 1983). Pu is released in deep waters through the formation of small particles due to the breakdown of large organic-rich particles containing Pu, and the subsequent remineralization of Pu due to the microbial decomposition of organic particles. Although similar biogeochemical processes controlled the distribution of Pu, there are large differences in 239,240 Pu concentrations in deep waters observed between the two hemispheres. The reason of this discrepancy between Table 3 239,240 Pu concentrations in the water column in the Southern and the Northern Hemispheres Depth
0 200 450 600 800 1000 1100 2000
239,240 Pu (mBq m−3 )
P06C-175∗
St.3∗∗
1.41 ± 0.24
4.0 ± 0.3 10.4 ± 0.8 26 ± 2 43 ± 4 47 ± 4
4.69 ± 0.66 4.81 ± 0.67 1.96 ± 0.59
∗ P06C-175: 18 Aug. 2003, 32.5◦ S, 177.7◦ E (this work). ∗∗ St.3: 29 Oct. 1999; 30.6◦ N, 170.6◦ E (Povinec et al., 2003).
37 ± 2 19.4 ± 0.8
Southern Hemisphere Ocean Tracer Study (SHOTS): An overview and preliminary results
61
the profiles of 239,240 Pu in the North and the South Pacific is still unknown. One possible cause may be that inter-hemispheric transport processes for 239,240 Pu and 137 Cs in the early 1960s were quite different, e.g. smaller amounts of 239,240 Pu compared with 137 Cs were transported from the northern to the southern stratosphere after the large atmospheric weapons test in the early 1960s. This suggests that 239,240 Pu/137 Cs fallout activity ratios should be different between both hemispheres. We shall study this discrepancy in detail when more data on radioactive and non-radioactive tracers, and the oceanographic parameters will be available. 240 Pu/239 Pu atom ratios obtained from water column samples at station P06C-175 ranged from 0.20 to 0.24. The observed 240 Pu/239 Pu atom ratios in environmental samples are summarized in Table 4. The global fallout 240 Pu/239 Pu ratios (0.17–0.19), which showed no significant differences between the Southern and Northern Hemispheres, except hot spots, were determined from analysis of soil samples. On the other hand, the 240 Pu/239 Pu ratios in seawater of the western North Pacific and its marginal seas seem to be higher than that of global fallout. This has been explained by close-in fallout from tests conducted in the Pacific Proving Grounds, where higher 240 Pu/239 Pu ratios were observed (Komura et al., 1984; Buesseler, 1997; Lee et al., 2005). Recently Warneke et al. (2002) reported historical 240 Pu/239 Pu ratios using archived grass samples. The results revealed that higher 240 Pu/239 Pu ratios (Table 4) occurred in the post-moratorium period corresponding with highest plutonium deposition rates. If global fallout plutonium showed different behavior between terrestrial and oceanic conditions, there is a possibility that the global fallout 240 Pu/239 Pu ratios in seawater are different from a typical global fallout ratio of 0.18 previously reported. Chiappini et al. (1999) reported 240 Pu/239 Pu atom ratios in the water column near Mururoa in the South PaTable 4 Plutonium atom ratios of environmental samples in the Southern and Northern Hemispheres Region
Sample
240 Pu/239 Pu
Reference
Chiappini et al., 1999
Southern Hemisphere Coordinates Mururoa, Fangataufa Mururoa
Sediment
138◦ 50 W, 22◦ 00 S
∼0.05
Water
138◦ 55 W, 21◦ 45 S
Rangiora Antarctic
Water Glacier
147◦ 05 W, 14◦ 47 S 168◦ 40 W, 88◦ 22 S
Air-aerosol (1969–1970) Soil (1970–1971)
75◦ W, 40◦ S
∼0.10 (0–100 m) 0.15–0.16 (600–1000 m) ∼0.08 (1000–2300 m) 0.166 ± 0.01 (500–1000 m) 0.21–0.34 (pre-moratorium, 1952–1962) 0.09∼0.19 (post-moratorium, post-1962) 0.087–0.12
153◦ E, 27◦ 30 S 145◦ 08 E, 37◦ 52 S 171◦ 49 E, 42◦ 17 S 172◦ 35 E, 43◦ 31 S 177.7◦ E, 32.5◦ S
0.1768 ± 0.0027 0.1716 ± 0.0014 0.1845 ± 0.0015 0.1901 ± 0.0036 0.20–0.24
Brisbane Melbourne Greymouth South Canterbury South Pacific
Seawater
Koide et al.
Kelley et al.
This study
62
M. Aoyama et al.
Table 4 (Continued) Region
Sample
Rothamsted
Archived grass
Tsukuba, Japan
Rain water
240 Pu/239 Pu
Reference
Warneke et al., 2002
Northern Hemisphere 1952–1953 1954–1956 1957–1962 1963–1970 1971–1977 1999
Daejeon, Korea
Rain water
North-eastern Massachusetts Stevens Hill, Pennsylvania Bikini Enewetak Japan
Soil
0.06–0.135 0.241–0.306 0.112–0.162 0.181–0.221 0.169–232 0.19 (0.17–0.23) 0.19 (0.16–0.22) 0.18 (dry) (0.09–0.24) 0.19 (wet) (0.10–0.27) 0.173 ± 0.008
Soil
0.177 ± 0.002
Soil Soil Soil
0.297–0.320 0.062–0.117 0.168–0.179
Muramatsu et al.
North Pacific
Seawater
Buesseler, 1997
Japan Sea
Surface seawater
0.18–0.22 (shallow) 0.21–0.26 (deep) 0.21–0.33
2000
Yellow Sea Korea Strait North and Central Pacific Eastern Pacific NW Pacific Okhotst Sea Japan Sea Coastal Korea NW Pacific
Sediment
Sediment
Sediment 0–1 cm 8–9 cm
2000
0.18–0.33 0.21–0.30 0.34–0.37 0.18–0.29 0.15–0.20 0.15–0.21 0.16–0.23 0.15–0.24 0.133–0.388 0.239 ± 0.005 0.388 ± 0.005
Hirose et al., 2003
Hirose et al., 2004
Kim et al., 2003
Buesseler, 1997
Kim et al., 2003
Lee et al., 2005
cific in which lower 240 Pu/239 Pu ratios occurred in shallower and deep waters due to the French nuclear tests, although the 240 Pu/239 Pu ratios in the plutonium maximum layer coincided with the global fallout ratio. Therefore, it cannot be excluded that the global fallout plutonium in seawater showed slightly higher 240 Pu/239 Pu ratios rather than a typical global fallout ratio (0.18). These findings suggest that the observed 239,240 Pu maximum at station P06C-175 is originating from global fallout, mostly from nuclear weapons tests carried
Southern Hemisphere Ocean Tracer Study (SHOTS): An overview and preliminary results
63
out during the 1960s, and that there is not a contribution from the French nuclear testing at Mururoa and Fangataufa Atolls.
4. Conclusions Preliminary results obtained in the framework of the SHOTS project show that 137 Cs and 239,240 Pu concentrations in surface waters of the South Pacific Ocean were similar to those in the subtropical gyres in the North Pacific. The 137 Cs profile shows a smaller effect of subduction in water transport in this area compared with that of the North Pacific. The estimated 137 Cs inventory of 850 Bq m−2 in the South Pacific water column is about half of the 137 Cs inventory in the North Pacific at the same latitude. The 239,240 Pu concentration in the South Pacific water column showed a surface minimum, a mid-depth maximum and a gradual decrease with depth. However, the 239,240 Pu concentrations in waters deeper than the Pu maximum layer were one order of magnitude lower than that observed in the North Pacific, which is a remarkable difference observed between two hemispheres compared to 137 Cs. The measured 240 Pu/239 Pu atom ratios were 0.20–0.24, statistically not very different from the global fallout ratio.
Acknowledgements The authors thank the Captain and the crew of the research vessel Mirai for their assistance during the BEAGLE2003 expedition. They are indebted to Takeshi Kawano, Shuuichi Watanabe and Koji Yoshikawa, all at JAMSTEC, for their help during the BEAGLE2003 cruise. Several cruise participants from Marine Works Japan Ltd. assisted during water sampling. The authors are also indebted to Akira Takeuchi, Sang-Han Lee, Benjamino Oregioni and Janine Gastaud who worked hard during the cruise to collect and treat huge amounts of water samples. The authors also thank Dr. C. Jeandel and two unknown reviewers for constructive comments. The Agency is grateful for the support provided to its Marine Environment Laboratory by the Government of the Principality of Monaco.
References Aoyama, M., Hirose, K. (2004). Artificial radionuclides database in the Pacific Ocean: HAM database. The Scientific World Journal 4, 200–215. Aoyama, M., Hirose, K., Miyao, T., Igarashi, Y. (2000). Low lever 137 Cs measurements in deep seawater samples. Applied Radiation and Isotopes 53, 159–162. Aoyama, M., Hirose, K., Miyao, T., Igarashi, Y., Povinec, P.P. (2001a). Temporal variation of 137 Cs inventory in the western North Pacific. Journal of Radioanalytical and Nuclear Chemistry 248, 785–787. Aoyama, M., Hirose, K., Miyao, T., Igarashi, Y., Povinec, P.P. (2001b). 137 Cs activity in surface water in the western North Pacific. Journal of Radioanalytical and Nuclear Chemistry 248, 789–793. Bourlat, Y., Milliés-Lacroix, J.-C., Le Petit, G., Bourguignon, J. (1996). 90 Sr, 137 Cs and 239,240 Pu in world ocean water samples collected from 1992 to 1994. In: Guéguéniat, P., Germain, P., Métivier, H. (Eds.), Radionuclides in the Oceans: Inputs and Inventory. Les Editions de Physique, Les Ulis, pp. 75–93.
64
M. Aoyama et al.
Bowen, V.T., Wong, K.M., Noshkin, V.E. (1971). Plutonium-239 in and over the Atlantic Ocean. Journal of Marine Research 29, 1–10. Bowen, V.T., Noshkin, V.E., Livingston, H.D., Volchok, H.L. (1980). Fallout radionuclides in the Pacific Ocean: Vertical and horizontal distributions, largely from GEOSECS stations. Earth and Planetary Science Letters 49, 411–434. Buesseler, K.O. (1997). The isotopic signature of fallout plutonium in the North Pacific. Journal of Environmental Radioactivity 36, 69–83. Chiappini, R., Pointurier, F., Milliés-Lacroix, J.-C., Lepetit, G., Hemet, P. (1999). 240 Pu/239 Pu isotopic ratios and 239+240 Pu total measurements in surface and deep waters around Mururoa and Fangataufa atolls compared with Rangiroa atoll (French Polynesia). The Science of the Total Environment 237/238, 269–276. Folsom, T.R. (1979). Summary of Cs-137 concentrations measured at Scripps institution in North Pacific surface waters. EML-356. Fowler, S.W., Ballestra, S., La Rosa, J., Fukai, R. (1983). Vertical transport of particulate-associated plutonium and americium in the upper water column of the Northeast Pacific. Deep-Sea Research 30 (12A), 1221–1233. Hamilton, T.F., Milliés-Lacroix, J.-C., Hong, G.H. (1996). In: Guéguéniat, P., Germain, P., Métivier, H. (Eds.), Radionuclides in the Oceans: Inputs and Inventories. Les Editions de Physique, Les Ullis, pp. 29–58. Hirose, K., Aoyama, M. (2003). Analysis of 137 Cs and 239,240 Pu concentrations in surface waters of the Pacific Ocean. Deep-Sea Research II 50, 2675–2700. Hirose, K., Sugimura, Y. (1985). A new method of plutonium speciation in seawater. Journal of Radioanalytical and Nuclear Chemistry, Articles 92, 363–369. Hirose, K., Aoyama, M., Miyao, T., Igarashi, Y. (2001). Plutonium in seawaters of the western North Pacific. Journal of Radioanalytical and Nuclear Chemistry 248, 771–776. Hirose, K., Igarashi, Y., Aoyama, M., Kim, C.K., Kim, C.S., Chang, B.W. (2003). Recent trends of Plutonium fallout observed in Japan: Plutonium as a proxy for desertification. Journal of Environmental Monitoring 5, 302–307. Hirose, K., Kim, C.K., Kim, C.S., Chang, B.W., Igarashi, Y., Aoyama, M. (2004). Plutonium deposition observed in Daejeon, Korea: Wet and dry depositions of plutonium. The Science of the Total Environment 332, 243–252. Available at http://www.jamstec.go.jp/mirai/2003/data_2003.html. Hirose, K., Aoyama, M., Igarashi, Y., Komura, K. (2005). Extremely low background measurements of 137 Cs in seawater samples using an underground facility (Ogoya). Journal of Radioanalytical and Nuclear Chemistry 263, 349–353. IAEA (2000). Global marine radioactivity database (GLOMARD). IAEA-TECDOC-1146. IAEA, Vienna. IAEA (2005). Worldwide marine radioactivity studies (WOMARS). Radionuclide levels in oceans and seas. IAEA TECDOC-1429. IAEA, Vienna, 187 pp. Kim, C.S., Kim, C.K., Lee, J.I., Lee, K.J. (2000). Rapid determination of Pu isotopes and atom ratios in small amounts of environmental samples by an on-line sample pre-treatment system and isotope dilution high resolution inductively coupled plasma mass spectrometry. Journal of Analytical Atomic Spectroscopy 15, 247–255. Kim, C.S., Kim, C.K., Lee, K.J. (2002). Determination of Pu isotopes in seawater by an on-line sequential injection technique with sector field inductively coupled plasma mass spectrometry. Analytical Chemistry 74, 3824–3832. Kim, C.K., Kim, C.S., Chang, B.U., Chung, C.S., Hong, G.H., Hirose, K., Pettersson, H.B.L. (2003). 240 Pu/239 Pu atom ratios in sediments in the NW Pacific Ocean and its marginal seas. Journal of Radioanalytical and Nuclear Chemistry 258, 265–268. Komura, K. (1997). Challenge to detection limit of environmental radioactivity. Proceedings of the 1997 International Symposium on Environmental Radiation, pp. 56–75. Komura, K., Sakanoue, M., Yamamoto, M. (1984). Determination of 240 Pu/239 Pu ratio in environmental samples based on the measurement of LX/-ray activity ratio. Health Physics 46, 1213–1219. La Rosa, J.J., Burnett, W., Lee, S.H., Levy, I., Gastaud, J., Povinec, P.P. (2001). Separation of actinides, cesium and strontium from marine samples using extraction chromatography and sorbents. Journal of Radioanalytical and Nuclear Chemistry 248 (3), 765–770. Lee, S.H., Gastaud, J., La Rosa, J.J., Liong Wee Kwong, L., Povinec, P.P., Wyse, E., Fifield, L.K., Hausladen, P.A., Di Tada, L.M., Santos, G.M. (2001). Analysis of plutonium isotopes in marine samples by radiometric, ICP-MS and AMS techniques. Journal of Radioanalytical and Nuclear Chemistry 248 (3), 757–764. Lee, S.H., Povinec, P.P., Wyse, E., Pham, M.K., Hong, G.H., Chung, C.S., Kim, S.H., Lee, H.J. (2005). Distribution and inventories of 90 Sr, 137 Cs, 241 Am and Pu isotopes in sediments of the Northwest Pacific Ocean. Marine Geology 216, 249–263.
Southern Hemisphere Ocean Tracer Study (SHOTS): An overview and preliminary results
65
Livingston, H.D., Anderson, R.F. (1983). Large particle transport of plutonium and other fallout radionuclides to the deep ocean. Nature 303, 228–231. Livingston, H.D., Povinec, P.P. (2000). Anthropogenic marine radioactivity. Ocean and Coastal Management 43, 689–712. Livingston, H.D., Povinec, P.P. (2002). A millennium perspective on the contribution of global fallout radionuclides to ocean science. Health Physics 82, 656–668. Livingston, H.D. et al. (1985). Fallout nuclides in Atlantic and Pacific water columns: GEOSECS Data. Woods Hole Oceanog. Inst. Tech. Report WHOI-85-19. Livingston, H.D., Povinec, P.P., Ito, T., Togawa, O. (2001). In: Kudo, A. (Ed.), Plutonium in the Environment. Elsevier, Amsterdam, pp. 267–292. Povinec, P.P. (Ed.) (2003). Worldwide marine radioactivity studies (WOMARS). Deep-Sea Research II 50 (17–21) (2003) 2595–2846 (Special issue). Povinec, P.P., Togawa, O. (1998). Worldwide marine radioactivity studies – Assessing the picture. IAEA Bulletin 40, 11–17. Povinec, P.P., La Rosa, J., Lee, S.H., Mulsow, S., Osvath, I., Wyse, E. (2001). Recent developments in radiometric and mass spectrometry methods for marine radioactivity measurements. Journal of Radioanalytical and Nuclear Chemistry 248, 713–718. Povinec, P.P., Livingston, H.D., Shima, S., Aoyama, M., Gastaud, J., Goroncy, I., Hirose, K., Lang, H.-N., Ikeuchi, Y., Ito, T., La Rosa, J., Liong Wee Kwong, L., Lee, S.-H., Moriya, H., Mulsow, S., Oregioni, B., Pettersson, H., Togawa, O. (2003). IAEA’97 expedition to the NW Pacific Ocean – Results of oceanographic and radionuclide investigations of the water column. Deep-Sea Research II 50, 2607–2637. Povinec, P.P., Aarkrog, A., Buesseler, K.O., Delfanti, R., Hirose, K., Hong, G.H., Ito, T., Livingston, H.D., Nies, H., Noshkin, V.E., Shima, S., Togawa, O. (2005). 90 Sr, 137 Cs, and 239,240 Pu concentration surface water time series in the Pacific and Indian Oceans – WOMARS results. Journal of Environmental Radioactivity 81, 63–87. Saruhashi, K., Katsuragi, Y., Kanazawa, T., Sugimura, Y., Miyake, Y. (1975). 90 Sr and 137 Cs in the Pacific waters. Records of Oceanographic Works in Japan 13, 1–15. Schlosser, P., Bayer, R., Boenish, G., Cooper, L.W., Ekwurzel, B., Jenkins, W.J., Khatiwala, S., Pfirman, S., Smethie, W.M. (1999). Pathways and mean residence times of dissolved pollutants in the ocean derived from transient tracers and stable isotopes. The Science of the Total Environment 237/238, 15–30. Warneke, T., Croudace, I.W., Warwick, P.E., Taylor, R.N. (2002). A new ground-level fallout record of uranium and plutonium isotopes for northern temperate latitudes. Earth and Planetary Science Letters 203, 1047–1057. Wijffels, S.E., Hall, M.M., Joyce, T., Torres, D.J., Hacker, P., Firing, E. (1998). Multiple deep gyres of the western North Pacific: A WOCE section along 149◦ E. Journal of Geophysical Research 103, 12985–13009.
Further reading Aoyama, M., Hirose, K. (1995). The temporal and spatial variation of 137 Cs concentration in the western North Pacific and its marginal seas during the period from 1979 to 1988. Journal of Environmental Radioactivity 29, 57–74. Folsom, T.R., Mohanrao, G.J., Pillai, K.C., Sreekumaran, C. (1968). Distributions of Cs-137 in the Pacific. HASL-197, I-95–I-203. Haque, M., Nakanishi, T. (1999). Host phase of 239,240 Pu and 241 Am in the deep-sea sediment. Journal of Radioanalytical and Nuclear Chemistry 239, 565–569. Hirose, K., Sugimura, Y., Aoyama, M. (1992). Plutonium and 137 Cs in the Western North Pacific: Estimation of residence time of plutonium in surface waters. Applied Radiation and Isotopes 43, 349–352. Hirose, K., Amano, H., Baxter, M.S., Chaykovskaya, E., Chumichev, V.B., Hong, G.H., Isogai, K., Kim, C.K., Kim, S.H., Miyao, T., Morimoto, T., Nikiyin, A., Oda, K., Pettersson, H.B., Povinec, P., Seto, Y., Tkalin, A., Togawa, O., Veletova, N.K. (1999). Anthropogenic radionuclides in seawater in the East/Japan Sea. Results of the firststage Japanese–Korean–Russian expedition. Journal of Environmental Radioactivity 43, 1–13. IAEA (1998). The radiological situation at the atolls of Mururoa and Fangataufa. Main Report. Radiological Assessments Reports Series. International Atomic Energy Agency, Vienna.
66
M. Aoyama et al.
Kim, C.K., Kim, C.S., Chang, B.U., Choi, S.W., Chung, C.S., Hong, G.H., Hirose, K., Igarashi, Y. (2004). Plutonium isotopes in seas around the Korean Peninsula. The Science of the Total Environment 318 (1–3), 197–209. Suga, T., Hanawa, K., Toba, Y. (1989). Thermostat distribution in the North Pacific subtropical gyre: The central mode water and the subtropical mode water. Journal of Physical Oceanography 27, 140–152.
67
Plutonium isotopes in seawater of the North Pacific: Effects of close-in fallout K. Hirosea,* , M. Aoyamaa , C.S. Kimb , C.K. Kimb,* , P.P. Povinecc a Geochemical Research Department, Meteorological Research Institute, Tsukuba, Ibaraki 305-0052, Japan b Department of Radiochemical Analysis, Korea Institute of Nuclear Safety, Daejeon 305-336, Republic of Korea c Marine Environment Laboratory, International Atomic Energy Agency, Monaco
Abstract The behavior of 239,240 Pu in seawater was studied using recent 239,240 Pu data from the North Pacific Ocean. 239,240 Pu activity concentrations in surface waters of the North Pacific in 2001 and 2002 ranged from 1 to 10 mBq m−3 . 239,240 Pu water profiles at the mid-latitude of the eastern North Pacific and at the Equatorial western North Pacific showed typical patterns consisting of a surface minimum, a mid-depth maximum and a gradual decrease in concentration with increasing water depth, although some temporal changes in the distribution of 239,240 Pu were observed. 240 Pu/239 Pu atom ratios in seawater samples, ranging from 0.17 to 0.28, suggest that major sources of plutonium in the western North Pacific were both global fallout and close-in fallout from nuclear weapons tests carried out at the Marshall Islands. Plutonium concentrations in shallower water layers (<1000 m) of the North Pacific varied temporally due to physical and biogeochemical processes, whereas its behavior has been rather conservative in deep layers (>2000 m). Keywords: Plutonium isotopes, 240 Pu/239 Pu atom ratio, Vertical water profile, Inventory, Global fallout, Close-in fallout, North Pacific
1. Introduction Plutonium in seawater of the western North Pacific has originated mostly from global fallout from large atmospheric nuclear weapons tests, from which the major oceanic input occurred in the early 1960s (Harley, 1980; Perkins and Thomas, 1980; Hirose et al., 2001a). However, 239,240 Pu inventories in the water column of the North Pacific have been significantly greater than estimates based on global fallout deposition (Bowen et al., 1980). This additional source of plutonium in the North Pacific has been attributed to inputs from close-in fallout from the US nuclear explosions conducted at the Pacific Proving Grounds in the Marshall Islands in the 1950s (Bowen et al., 1980; Livingston et al., 2001). In order to better understand oceanic behavior of 239,240 Pu, it is important to elucidate the effects of close-in fallout on 239,240 Pu concentrations in the western North Pacific. * Corresponding author. Address: Meteorological Research Institute, Nagamine 1-1, Tsukuba, Ibaraki 305-0052, Japan; phone: (+81) 298 53 8719; fax: (+81) 298 53 8728; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08005-8
© 2006 Elsevier Ltd. All rights reserved.
68
K. Hirose et al.
Plutonium in the ocean is transported by physical and biogeochemical processes. The residence time of plutonium in surface waters of the open ocean ranges from 6 to 21 years, and is generally shorter than that of 137 Cs (Hirose et al., 2001b; Hirose and Aoyama, 2003a; Povinec et al., 2005). In contrast to 137 Cs, plutonium is a typical particle-reactive radionuclide; plutonium in particulate matter of surface waters occupies from 1% to 10% (Hirose et al., 2001b; Livingston et al., 1987), whereas particulate 137 Cs is less than 0.1% of the total (Aoyama and Hirose, 1995). Plutonium vertically moves with sinking biogenic particles (Livingston and Anderson, 1983; Fowler et al., 1983), and regenerates into soluble forms in deep waters as a result of microbial decomposition of particles. The biogeochemical processes are responsible for typical vertical concentration profiles of plutonium, showing a surface minimum, a mid-depth maximum and, thereafter, a decrease with increasing water depth (Hirose, 1997; Tsumune et al., 2003). Although plutonium moves vertically by biogeochemical processes, most of the plutonium introduced into the Pacific Ocean still exists in the water column. Plutonium inventories in sediment of the open ocean are generally less than 10% of the total inventories, except areas near Bikini and Enewetak Atolls (Livingston et al., 2001). Plutonium concentrations in the oceanic water are further affected by physical processes such as advection and upwelling (Hirose et al., 2002). For example, the observation that the plutonium maximum layer in the mid-latitude region of the North Pacific has been deepening with time (Livingston et al., 2001) is explained by a simple one-dimensional biogeochemical model (Hirose, 1997). The model, however, does not provide a mechanism for reducing the overall water-column inventory. Consequently, advection must play a significant role in determining the temporal trends in regional oceanic inventories of plutonium. North Pacific deep waters (below 2000 m) contain a significant amount of plutonium (Bowen et al., 1980; Povinec et al., 2003), whose input processes are still unknown. Plutonium in the environment consists of a number of different isotopes: 238 Pu, 239 Pu, 240 Pu, and 241 Pu. 239 Pu and 240 Pu are the most abundant plutonium isotopes in the marine environment, having long radioactive half-lives of 24,110 and 6,560 years, respectively. 238 Pu and 241 Pu, with shorter radioactive half-lives of 87.7 and 14.1 years, respectively, are also useful plutonium tracers, having, however, lower concentrations in the marine environment. The 240 Pu/239 Pu atom ratio, and 238 Pu/239,240 Pu and 241 Pu/239,240 Pu activity ratios in global fallout, depend on the specific nuclear weapons design and test yields, and therefore they may vary in the environment (Koide et al., 1979). The global fallout average 240 Pu/239 Pu atom ratio is 0.18, based on analyses of aerosols, soil, grass and ice core samples (HASL, 1973; Krey et al., 1976; Muramatsu et al., 1999; Warneke et al., 2002). However, different nuclear test series can be characterized by either higher or lower ratios. For example, fallout from Nagasaki, the Nevada and the Semipalatinsk test sites is characterized by generally lower 240 Pu/239 Pu ratios, 0.042, 0.035 (an average value) and 0.036, respectively (Komura et al., 1984; Hicks and Barr, 1984; Buesseler and Scholkovitz, 1987; Yamamoto et al., 1996), whereas elevated 240 Pu/239 Pu ratios (0.21–0.36) have been measured in soil samples from the Bikini atoll (Hisamatsu and Sakanoue, 1978; Muramatsu et al., 2001). Plutonium isotope signatures are therefore useful tools for identification of sources of plutonium, and for better understanding of its behavior in the marine environment. Buesseler (1997), by analyzing 240 Pu/239 Pu atom ratios in seawater samples, revealed that there is a significant contribution of close-in fallout from the Pacific Proving Ground nuclear testing to North Pacific deep waters.
Plutonium isotopes in seawater of the North Pacific
69
In this paper, we describe recent 239,240 Pu concentrations and plutonium isotope ratios in seawater collected in 2001 and 2002, discuss oceanic behavior of plutonium, and compare 239,240 Pu inventories in the water column, including data obtained from the IAEA’97 Pacific cruise (Povinec et al., 2003).
2. Sampling and analytical methods Surface water samples were collected in 2001 and 2002 aboard the R/Vs Mirai and Ryofumaru, which belong to the JAMSTEC (Japan Agency for Marine-Earth Science and Technology) and the Japan Meteorological Agency, respectively. Surface and deep waters were sampled using a submersible pump and 100-liter GoFlo sampler/CTD-rosette sampler, respectively. All water samples were filtered through a fine membrane filter (Millipore HA, 0.45 µm pore size) immediately after sampling. Sampling stations are shown in Fig. 1. Plutonium dissolved in seawater was co-precipitated with iron hydroxides from 20 to 100 L of seawater samples. Both dissolved and particulate plutonium were assayed using α-spectrometry and SF-ICP-MS, following radiochemical separation and purification using anion exchange resin, described in detail elsewhere (Hirose and Sugimura, 1985; Hirose et al., 2001b; Kim et al., 2000, 2002). The chemical yield was determined by the addition of a known amount of 242 Pu.
Fig. 1. Sampling locations in the central and western North Pacific Ocean. Closed circles: deep water (this work); double circles: vertical profiles (this work); triangles: IAEA’97 cruise (Povinec et al., 2003).
70
K. Hirose et al.
3. Results and discussion 3.1.
239,240 Pu
239,240 Pu
in surface waters
concentrations in surface waters of the North Pacific in 2002 ranged from 1 to 10 mBq m−3 (Table 1). In order to better understand behavior and fate of 239,240 Pu in the marine environment, we examined its temporal variations in surface waters, as the previous results were suggesting that temporal variations of surface 239,240 Pu depend on sea area (Hirose et al., 2001b; Hirose and Aoyama, 2003a). In order to elucidate temporal variations of plutonium in surface waters, the western North Pacific was divided into three areas: 20–35◦ N (KKR region: the Kuroshio and Kuroshio recirculation region), 10–20◦ N (NEC region: the North Equatorial Current region), and 0–10◦ N (ECC region: the Equatorial Counter Current and others). In the KKR region, the surface plutonium concentrations decreased exponentially during the period from 1967 to 2002, as shown in Fig. 2(a). Curve fitting was applied to determine the apparent residence time of 239,240 Pu in surface waters. In the mid-latitude region of the western North Pacific, the apparent residence time of surface 239,240 Pu was calculated to be 7.0 ± 0.5 years, which is in good agreement with estimates (6–8 years) (Hirose et al., 2001b). The apparent residence time in the KKR region is due to particle scavenging longer than the residence time determined by sediment traps in the eastern North Pacific (2.4 years; 0–80 m layer; Fowler et al., 1983). The short residence time of plutonium from the sediment traps may contain larger uncertainty due to large seasonal variability of particle export fluxes because the trap-deploying period was relatively short (2 weeks). Another process deducing the longer apparent residence time of plutonium in the KKR region could be in recycling of 239,240 Pu due to vertical mixing and upwelling of deep waters with higher 239,240 Pu concentrations, and its supply via horizontal advection. In the NEC region, surface 239,240 Pu gradually decreased during the sampling period from 1973 to 2002 (Fig. 2(b)). However, the apparent decrease rate of surface 239,240 Pu concentrations in the NEC region was slower than that in the KKR region. The apparent residence time of surface 239,240 Pu in the NCE region was calculated to be 7.9 ± 0.7 years. The longer apparent residence time of surface 239,240 Pu in the NEC region may be attributable to a lower particle scavenging in low productivity waters typical for the oligotrophic ocean, in addition to the recycling processes. In the ECC region, a temporal trend of surface 239,240 Pu is shown in Fig. 2(c). Although higher scavenging of 239,240 Pu than in the oligotrophic ocean is expected due to relatively high biological productivity, surface 239,240 Pu slowly decreased in the sampling period from 1973 to 2002 in comparison with the subtropical western North Pacific. Assuming that surface 239,240 Pu concentrations have been decreasing exponentially during the sampling period, we obtain an apparent residence time of 11 ± 4 years for surface 239,240 Pu in the ECC region. On the other hand, the surface 239,240 Pu levels in the ECC region seems to be almost constant after 1985 (Fig. 2(c)). Recent surface 239,240 Pu concentrations were higher than that expected from the exponential regression. These findings suggest that there is a supply of 239,240 Pu into ECC surface waters. Gradual increases of surface 137 Cs were observed in the western Equatorial Pacific (Aoyama and Hirose, 1995), which may be explained by horizontal advection of subsurface waters with higher 137 Cs concentrations, and Equatorial upwelling. Water
Plutonium isotopes in seawater of the North Pacific
71
Table 1 239,240 Pu concentrations and 240 Pu/239 Pu atom ratios in seawater of the North Pacific (the uncertainties represent 1 σ standard deviations) Date of sampling
Location
Depth (m)
239, 240 Pu
α-spectrometry (mBq m−3 )
240 Pu/239 Pu
ICP-MS (mBq m−3 )
atom ratio
Oct. 2002
0◦ 58 N 165◦ 10 E
surface 1448 1871 2800 2200 2256
9.4 ± 2.0 13.8 ± 2.0 14.6 ± 2.1 15.2 ± 2.7 9.2 ± 1.4 13.8 ± 1.9
0.20 ± 0.03 0.25 ± 0.03 0.21 ± 0.03 0.26 ± 0.02 0.24 ± 0.03 0.28 ± 0.04
Oct. 2002
17◦ 58 N 164◦ 54 E
surface 1800 2800 3800 5000
1.6 ± 0.4 5.9 ± 0.9 8.2 ± 2.1 5.2 ± 0.9 8.2 ± 0.2
0.23 ± 0.01 0.23 ± 0.01 0.21 ± 0.05 0.24 ± 0.03
Oct. 2002
27◦ 58 N 164◦ 57 E
surface 3500 5000
4.8 ± 1.0 12.2 ± 1.2 6.7 ± 0.9
Oct. 2002
7◦ 59 N 164◦ 54 E
surface 100 200 500 600 1000 1994 2800 3900 4800 5138
3.7 ± 0.6
11.0 ± 3.0∗ 8.4 ± 1.4∗ 17.9 ± 5.0∗ 6.1 ± 0.9∗ 5.5 ± 1.2
June 2001
41◦ 30 N 152◦ 00 E
1000
21.5 ± 2.3
June 2001
30◦ 38 N 152◦ 00 E
2000
21.2 ± 2.8
Aug. 2001
31◦ 01 N 150◦ 00 W
surface 100 300 500 700 1000 2000
9.2 ± 2.7 7.3 ± 1.5 16.0 ± 2.3 20.4 ± 2.6 26.9 ± 3.6 13.5 ± 2.1 18.5 ± 3.1
4.1 ± 0.1 15.1 ± 0.4 15.9 ± 1.2 25.1 ± 2.0 18.9 ± 1.3 16.0 ± 0.3∗ 11.0 ± 0.3∗ 10.7 ± 0.6∗ 6.5 ± 0.6∗
0.221 ± 0.004 0.164 ± 0.004 0.225 ± 0.016 0.218 ± 0.017 0.228 ± 0.015 0.232 ± 0.014 0.250 ± 0.007 0.247 ± 0.013 0.249 ± 0.022
0.23 ± 0.06 0.28 ± 0.01 0.25 ± 0.01 0.24 ± 0.01 0.23 ± 0.04
∗ The two portions from the same samples were used for plutonium analyses by α-spectrometry and ICP-MS.
72
K. Hirose et al.
Fig. 2. Temporal variation of surface 239,240 Pu concentrations in the western North Pacific. Open and closed circles are showed historical data and present data including the IAEA’97 Pacific cruise (Povinec et al., 2003), respectively. (a) Kuroshio and Kuroshio recirculation region (20–35◦ N); (b) the North Equatorial Current region (10–20◦ N); (c) the Equatorial Counter Current region (0–10◦ N).
Plutonium isotopes in seawater of the North Pacific
73
masses with higher radionuclide concentrations formed in the mid-latitude region of the central North Pacific are transported southwestward under the surface waters of the subtropical gyre on a time scale of several decades. In fact, higher 137 Cs concentration cores were observed in subsurface layers (100–500 m depth) of the subtropical gyre of the North Pacific (Hirose and Aoyama, 2003a). In the Equatorial Pacific, the subsurface waters with higher radionuclide concentrations are transported to the surface by Equatorial upwelling. The same mechanism has been proposed for the temporal change of 14 C in surface waters of the Equatorial Pacific (Quay et al., 1983). The recent trend of the surface 239,240 Pu concentrations in the western Equatorial Pacific may be maintained by a physical mechanism similar to that for 137 Cs and 14 C. It must be noted, however, that for surface 239,240 Pu in the Equatorial Pacific as well as in the subtropical North Pacific, there may be a contribution from close-in 239,240 Pu fallout from the Pacific Proving Grounds in the Marshall Islands. The spatial distribution of 239,240 Pu in seawater has been significantly affected by input processes such as global fallout, which showed a mid-latitude maximum in the Northern Hemisphere. However, the recently observed spatial distribution of surface plutonium in the North Pacific may be controlled by physical processes such as advection from close-in fallout, and upwelling of subsurface waters, including enriched 239,240 Pu, in addition to biogeochemical processes such as particle scavenging. As a result, there has not been during the 1990s a significant meridional distribution of surface plutonium in the western North Pacific (Hirose and Aoyama, 2003a, 2003b). 3.2. Vertical profiles of
239,240 Pu
The vertical profiles of 239,240 Pu concentrations in the water column of the eastern North Pacific (31◦ 01 N, 150◦ 00 W) (Fig. 3(a)) and the western Equatorial Pacific (7◦ 59 N, 164◦ 54 E) (Fig. 3(b)) showed a surface minimum and mid-depth maximum and, thereafter, a gradual decrease with increasing water depth, which is similar to previous studies (Bowen et al., 1980; Livingston et al., 2001; Nagaya and Nakamura, 1984). We compared the 239,240 Pu profile in the eastern North Pacific with the same GEOSECS site (31◦ 23 N, 150◦ 02 W) of 1973 (Livingston et al., 1985; Aoyama and Hirose, 2004). The depth of the 239,240 Pu maximum layer has shifted to deeper layers. The similar temporal change has been observed in the midlatitude region of the western North Pacific (Livingston et al., 2001). The peak 239,240 Pu concentration at around 400 m depth markedly decreased during the period of 28 years whereas the 239,240 Pu concentrations in deep waters (>1000 m depth) did not change. These findings suggest that the 239,240 Pu concentrations in mid depths (0–500 m depth) in the mid-latitude region of the North Pacific have been continuously decreasing during the past three decades. This change cannot be explained by a simple biogeochemical process including only vertical transport such as particle scavenging because there is no marked increase of the 239,240 Pu concentrations in deeper water layers (>1000 m). Therefore, the decrease of the 239,240 Pu concentrations in the mid-depth in the mid-latitude region of the North Pacific is lead by exchange with water masses having lower 239,240 Pu concentrations due to horizontal advection. Since there was no GEOSECS site adjacent to our station in the Equatorial North Pacific, we compared our data with a more eastern station (5◦ 53 N, 172◦ 01 W) in the Equatorial North Pacific. The vertical 239,240 Pu profile in the Equatorial North Pacific in the depth layers from the surface to 1000 m did not show significant change during the past three decades. A similar
74
K. Hirose et al.
(a)
(b)
Fig. 3. The vertical profiles of 239,240 Pu concentrations in the water column of the North Pacific. Open and closed circles show present observations and GEOSECS data (1973), respectively, (a) the eastern North Pacific; (b) the Equatorial North Pacific.
tendency has been observed in the subtropical western North Pacific (Livingston et al., 2001; Povinec et al., 2003). On the other hand, there is a significant difference in 239,240 Pu concentrations in deep waters between equatorial GEOSECS and present sites; deep 239,240 Pu concentrations in the present station near the Pacific Proving Ground nuclear test sites were higher than those in the eastern station (GEOSECS site). This finding suggests that 239,240 Pu from the close-in fallout did not spread in the central Equatorial Pacific, which is consistent with the previous observations that radionuclides from Bikini explosions were initially injected into sea areas northwards of Bikini (Miyake et al., 1955). 3.3. Deep 239,240 Pu The North Pacific deep waters contain significant amounts of 239,240 Pu (Bowen et al., 1980; Povinec et al., 2003). It is interesting to know the timescale of 239,240 Pu introduction into deep waters and processes which were responsible for its vertical transport into deep layers. Livingston et al. (2001) suggested that 239,240 Pu concentrations in deep waters (below 2000 m) have almost been maintained at constant level during the past two decades. In order to elucidate the present features in the distribution of deep 239,240 Pu, we have depicted latitudinal
Plutonium isotopes in seawater of the North Pacific
75
distributions of deep 239,240 Pu in two different layers (the upper layer: 2000–3500 m, and the lower layer: 4000–6000 m) along the 165◦ E and 180◦ E longitude, in which data (Aoyama and Hirose, 2004), collected from 1978 to 2002, and from 1973 to 1988, respectively, have been included. The results along the 165◦ E longitude are shown in Figs. 4(a),(b). The latitudinal distribution of the 239,240 Pu concentrations in the upper deep layer is showing two peaks; one peak occurs in the mid-latitude region (30–40◦ N), corresponding to the latitudinal pattern of global fallout. The second peak appears in the south subtropical region (10–15◦ N), corresponding to the close-in fallout from the Pacific Ground nuclear test sites. Lower 239,240 Pu concentrations in the upper deep layer occurred in the north subtropical (20–25◦ N) and equatorial regions (0–7/8◦ N). The 239,240 Pu concentrations in the lower deep layer were generally lower than that in the upper deep layer. The latitudinal distribution of the 239,240 Pu concentrations in the
Fig. 4. The latitudinal distribution of 239,240 Pu in deep waters. Open and closed circles show historical data and present data including the IAEA’97 Pacific cruise (Povinec et al., 2003), respectively. (a) The upper deep layer (2000–3500 m) along 165◦ E; (b) the lower deep layer (4000–6000 m) along 165◦ E; (c) the upper deep layer (2000–3500 m) along 180◦ ; (d) the lower deep layer (4000–6000 m) along 180◦ .
76
K. Hirose et al.
Fig. 4. (Continued.)
lower deep layer (Fig. 4(b)) is also showing peaks between 20–30◦ N and 10–15◦ N, however, the picture is not so for the upper deep layer. The latitudinal distributions of the 239,240 Pu concentrations in deep waters along the 180◦ E longitude (Figs. 4(c),(d)) show similar patterns as for the 165◦ E longitude. The 239,240 Pu concentrations in the upper deep layer along 180◦ E are slightly lower than those in the corresponding depth layer along 165◦ E. The 239,240 Pu concentrations in the lower deep layer showed, however, a peak in the mid-latitude region (30–40◦ N) with higher 239,240 Pu concentrations than in the upper deep layer. Buesseler (1997) observed an enhanced 240 Pu/239 Pu ratio, a typical signature of close-in fallout, in a sediment sample collected in the central North Pacific (38◦ 00 N, 179◦ 45 W). The close-in fallout plutonium may be preferentially transported into deeper layers including sediments because the chemical and physical properties of particles carrying close-in fallout plutonium are different from those of global fallout (Adams et al., 1960). Therefore, close-in fallout from Bikini or Enewetak Atolls may contribute to 239,240 Pu concentrations in deeper layer waters of the central North Pacific as well.
Plutonium isotopes in seawater of the North Pacific
77
These findings suggest that a geographical structure of deep 239,240 Pu concentrations exists in the western and central North Pacific. The subarctic Pacific is recognized as a highly productive area in the North Pacific. Therefore, it is expected that larger vertical 239,240 Pu transport and remineralization in deep waters occur as a result of higher particle export fluxes in the ocean interior. However, the geographical distributions of deep 239,240 Pu suggest that there is little contribution from biogeochemical cycling of 239,240 Pu to the North Pacific deep waters (>2000 m). Model studies conducted by Tsumune et al. (2003) suggest that there is less contribution of biogeochemical vertical transport to deep 239,240 Pu (>2000 m depth) although about four decades have passed since the surface injection of 239,240 Pu. These findings suggest that a fraction of fallout 239,240 Pu-bearing particles was rapidly transported to deep layers and dissolved in deep waters. Therefore, the geographical distributions of 239,240 Pu in deep waters reflect spatial distributions of global fallout and close-in fallout together. 3.4. Plutonium isotope ratios Recent developments in high resolution ICP-MS techniques allow us to determine plutonium isotopic ratios with high precision (Kim et al., 2000, 2002; Taylor et al., 2001). Measured 240 Pu/239 Pu atom ratios in water samples are summarized in Table 1. The 240 Pu/239 Pu ratios in the North Pacific waters ranged from 0.16 to 0.28, which are in the range of the 240 Pu/239 Pu ratio of global fallout (0.18) to that in close-in fallout (0.30). The 240 Pu/239 Pu ratios in shallower water layers tend to be lower than that in deep waters, although the depth trend of 240 Pu/239 Pu ratios is not statistically significant (the correlation coefficient R 2 = 0.1) because of larger analytical uncertainties obtained for the deep water samples. The 240 Pu/239 Pu atom ratios in deep waters (>2000 m) were around 0.25, which coincides with the results obtained by Buesseler (1997). The observed elevated 240 Pu/239 Pu ratios thus document a presence of close-in fallout plutonium in North Pacific waters. Significantly lower 240 Pu/239 Pu atom ratios were found in shallower water layers of the Equatorial North Pacific (7◦ 59 N, 164◦ 54 E), which may be due to intrusion of the subsurface waters carrying a global fallout 239,240 Pu signal as a result of southwestward motion. 3.5. Inventory of
239,240 Pu
in the water column
The 239,240 Pu inventory in the water column of the western North Pacific was calculated by interpolating the 239,240 Pu concentrations measured at each water depth. The results are summarized in Table 2. The 239,240 Pu inventory in the mid-latitude region of the western North Pacific decreased when comparing with GEOSECS data (Livingston et al., 2001; Povinec et al., 2003). The decrease in the 239,240 Pu inventory in the mid-latitude regions is attributable to a plutonium deficiency in the upper layer (0–1000 m depth). Similar trends in the North Pacific mid-latitude region have been observed also for 137 Cs (Aoyama and Hirose, 2003). On the other hand, there is no temporal decrease of the 239,240 Pu inventories in the subtropical and the Equatorial North Pacific water columns (Livingston et al., 2001), although the present 239,240 Pu inventory in the Equatorial North Pacific is larger than that of the corresponding GEOSECS station. The 239,240 Pu inventories in the western North Pacific shows a flat latitudinal distribution (100–130 Bq m−2 ), except for the Equatorial Pacific. The largest 239,240 Pu inventory (130 Bq m−2 ) was observed in the North Equatorial Current region (15◦ 30 N, 159◦ 31 W), the site being located northwards downstream of Bikini. It must
78
K. Hirose et al.
Table 2
239,240 Pu inventories in the western North Pacific (Bq m−2 )
Location 34◦ 60 N, 146◦ 00 E 30◦ 34 N, 170◦ 37 E 11◦ 26 N, 164◦ 52 E 11◦ 30 N, 161◦ 45 E 15◦ 30 N, 159◦ 31 E 7◦ 59 N, 164◦ 54 E
Depth (m)
239,240 Pu
5927 5472 4537 3690 5557 5138
101 122 115 110 129 67∗∗
239,240 Pu (close-in)
239,240 Pu (global)
239,240 Pu
GEOSECS 37∗ 51∗ 62∗ 64∗ 79∗ 26∗∗∗ (22–32)
64∗ 71∗ 53∗ 46∗ 50∗ 41∗∗∗ (45–35)
140 150 110
40
∗ The global and close-in 239,240 Pu inventories were calculated from 238 Pu inventories using Equation (1). The data of 239,240 Pu and 238 Pu concentrations in the water column obtained in the IAEA’97 cruise (Povinec et al., 2003) were used. ∗∗ This work. ∗∗∗ The inventories were calculated from 240 Pu/239 Pu atom ratios at each depth using Equation (2). Values in paren-
thesis were calculated from different 240 Pu/239 Pu atom ratios (0.30 and 0.36) of close-in fallout.
be noted that the present 239,240 Pu water column inventories in the western North Pacific are larger than that of total 239,240 Pu deposition in corresponding latitudes originating from global fallout (5 Bq m−2 (0–10◦ N) to 67 Bq m−2 (30–40◦ N)) (Hardy et al., 1973). Therefore for better understanding of plutonium behavior in the western North Pacific, it is important to know the contribution of close-in fallout plutonium from the Pacific Proving Ground testing. The plutonium isotope signature is a clue for estimation of individual contributions from global and close-in fallout 239,240 Pu, because the isotope ratios of global fallout plutonium (238 Pu/239,240 Pu and 240 Pu/239 Pu) differ from the close-in fallout ratios. The fallout plutonium originating from the Pacific Proving Ground testing (Bravo test, conducted on February 28, 1954) is characterized by lower 238 Pu/239,240 Pu activity ratio (0.001) and higher 240 Pu/239 Pu atom ratio (0.33), which were determined for archived samples directly contaminated by the Bravo test (Komura et al., 1984). The low 238 Pu/239,240 Pu activity ratios from the Bravo test were recorded in the Ross ice sheet, Antarctica, as well (Koide et al., 1979). The plutonium isotope ratios depend on individual nuclear explosions; for example, a higher 240 Pu/239 Pu atom ratio (0.36) was reported for the Ivy Mike explosion on November 1, 1952 (Diamond et al., 1960). If plutonium in the water column of the western North Pacific consists of only two components, i.e., global fallout and close-in fallout (the Bravo test as a representative), then 239,240 Pu inventories derived from individual fallout can be estimated from plutonium isotope data based on a simple mixing model. The 239,240 Pu water column inventory originating from global fallout based on 238 Pu/239,240 Pu activity ratios is estimated from the following equation: IWT,2 − IWT,1 RB IWG,1 = (1) , RG − RB where IWT,1 and IWG,1 are the total inventories of 239,240 Pu in the water column and from global fallout, respectively, and IWT,2 is the total inventory of 238 Pu in the water column.
Plutonium isotopes in seawater of the North Pacific
79
RG and RB denote the 238 Pu/239,240 Pu activity ratios in global fallout (0.03), and in close-in fallout (0.001), respectively. We calculated global fallout and close-in fallout 239,240 Pu inventories for the IAEA’97 cruise data (Povinec et al., 2003), which contains a complete data set including 238 Pu concentrations in the water column. The results are summarized in Table 2. On the other hand, the 239,240 Pu concentrations derived from the Bikini nuclear explosions can be estimated from the 240 Pu/239 Pu atom ratios by the following equation: −1 and AB = AO (1 + λR RB ) 1 + λR RB + (1 + λR RG )Q (2) −1 Q = (RO − RB )(RG − RO ) , where AB and AO are the Bikini-derived and observed 239,240 Pu concentrations, respectively, λR is the ratio of the radioactive decay constants of 239 Pu and 240 Pu, and RB , RG and RO are atom ratios in close-in fallout from Bikini, in global fallout and in analyzed water samples, respectively. The Bikini-derived 239,240 Pu inventory is calculated from the estimated Bikiniderived 239,240 Pu concentrations in each depth. The results are summarized in Table 2. It must be noted that the estimated inventories from global fallout and close-in fallout have significant uncertainties (less than 50%) because 238 Pu determinations have large uncertainties due to its low concentrations, and the variation range of the 240 Pu/239 Pu atom ratios between global fallout and close-in fallout is relatively narrow. The results suggest that a significant amount of 239,240 Pu derived from close-in fallout has spread over the subtropical region of the western North Pacific. The close-in fallout 239,240 Pu inventories peaked in the North Equatorial Current region (15◦ 30 N), and gradually decreased northward. A significant amount of the current 239,240 Pu inventory in the subtropical western North Pacific is still represented by the close-in fallout 239,240 Pu. The gap in the 239,240 Pu inventories derived from the close-in fallout, between the subtropical and Equatorial North Pacific, may be due to the fact that the close-in fallout occurred north of Bikini and Enewetak islands (Miyake et al., 1955). The 239,240 Pu inventory originating from global fallout is from 41 to 71 Bq m−2 , showing a latitudinal distribution with a maximum in the mid-latitude region, and a minimum in the Equatorial Pacific. Comparing with the latitudinal distribution of global fallout 239,240 Pu (Hardy et al., 1973) and GEOSECS data (Bowen et al., 1980), the 239,240 Pu inventories have gradually decreased in the mid-latitude region and increased in the Equatorial Pacific. Therefore, a latitudinal gradient of the 239,240 Pu water column inventory from global fallout in the western North Pacific is smaller than the 239,240 Pu deposition due to global fallout (Hardy et al., 1973). This finding suggests that a significant amount of global fallout 239,240 Pu, mostly deposited in the mid-latitude region, has been transported southward in the North Pacific during the last four decades, which is consistent with the 137 Cs inventory (Aoyama and Hirose, 2003).
4. Conclusions We examined the temporal and spatial distributions of the 239,240 Pu concentrations in seawater using present and historical databases (Aoyama and Hirose, 2004) with the aim to better understand the behavior of 239,240 Pu in the North Pacific Ocean. The observed 239,240 Pu concentrations in surface waters showed temporal variations; it decreased exponentially in the
80
K. Hirose et al.
mid-latitude and the North Equatorial Current regions, whereas in the Equatorial Pacific they have been stable since 1985. The vertical profiles of 239,240 Pu in the water column of the North Pacific are still showing typical patterns with a surface minimum, a mid-depth maximum and, thereafter, a gradual decrease with increasing water depth. The temporal variations in the 239,240 Pu vertical profiles depend on the sea areas; in the mid-latitude region, the 239,240 Pu maximum layer deepened with time and the corresponding 239,240 Pu concentrations decreased, whereas there does not seem to be any change of the 239,240 Pu vertical profiles in the subtropical and Equatorial North Pacific. The 239,240 Pu concentrations in deep waters (>2000 m depth) of the North Pacific have a spatial structure, which seems to reflect input patterns of global fallout and close-in fallout. This finding suggests that there is little temporal variability of deep 239,240 Pu over a time scale of several decades. Plutonium isotope signatures suggest that large amounts of 239,240 Pu were injected into the subtropical western North Pacific as close-in fallout from the Pacific Proving Ground nuclear testing in the early 1950s and spread due to advection and diffusion in the North Pacific during the past five decades. 239,240 Pu inventories from global and close-in fallout were estimated using plutonium isotope ratios. The close-in fallout 239,240 Pu still occupies a significant part of the 239,240 Pu water column inventory in the subtropical North Pacific. The inventory of the global fallout 239,240 Pu has showed a latitudinal distribution with high values in the mid-latitude regions, which gradually decrease to the south. Plutonium in seawater of the North Pacific showed different behavior between the shallower layer (<1000 m depth) and deep waters (>2000 m depth); in the shallower layer, 239,240 Pu shows larger temporal variability as a result of physical processes such as advection and mixing as well as biogeochemical processes, whereas deep 239,240 Pu seems to behave conservatively. Acknowledgements The authors thank the captain and the crew of the R/V Ryofu-maru for on-board water sampling, M. Fukasawa of JAMSTEC for providing water samples collected in the eastern North Pacific on the RV Mirai cruise, and I. Koshino for radiochemical analysis of plutonium. The authors also thank three anonymous reviewers for constructive comments and suggestions. This study has been supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan. The Agency is grateful for the support provided to its Marine Environment Laboratory by the Government of the Principality of Monaco. References Adams, C.E., Farlow, N.H., Schell, W.R. (1960). The compositions, structures and origins of radioactive fall-out particles. Geochimica et Cosmochimica Acta 18, 42–56. Aoyama, M., Hirose, K. (1995). The temporal and spatial variation of 137 Cs concentrations in the western North Pacific and marginal seas during the period from 1979 to 1988. Journal of Environmental Radioactivity 29, 57– 74.
Plutonium isotopes in seawater of the North Pacific
81
Aoyama, M., Hirose, K. (2003). Temporal variation of 137 Cs inventory in the Pacific Ocean. Journal of Environmental Radioactivity 69, 107–117. Aoyama, M., Hirose, K. (2004). Artificial radionuclides in the Pacific Ocean – HAM database. The Scientific World Journal 4, 200–215. Bowen, V.T., Noshkin, V.E., Livingston, H.D., Volchok, H.L. (1980). Fallout radionuclides in the Pacific Ocean: Vertical and horizontal distributions, largely from GEOSECS stations. Earth and Planetary Science Letters 49, 411–434. Buesseler, K.O. (1997). The isotopic signature of fallout plutonium in the North Pacific. Journal of Environmental Radioactivity 36, 69–83. Buesseler, K.O., Scholkovitz, E.R. (1987). The geochemistry of fallout plutonium in the North Atlantic: II. 240 Pu/239 Pu ratios and their significance. Geochimica et Cosmochimica Acta 51, 2623–2637. Diamond, H., Fields, P.R., Stevens, C.S., Studier, M.H., Fried, S.M., Inghram, M.G., Hess, D.C., Pyle, G.L., Mech, J.F., Manning, W.M. (1960). Heavy isotope abundances in Mike thermonuclear device. Physical Review 119, 2000–2004. Fowler, S.W., Ballestra, S., LaRosa, J., Fukai, R. (1983). Vertical transport of particulate-associated plutonium and americium in the upper water column of the Northeast Pacific. Deep-Sea Research 30, 1221–1233. Hardy, E.P., Krey, P.W., Volchok, H.L. (1973). Global inventory and distribution of fallout plutonium. Nature 241, 444–445. Harley, J.H. (1980). Plutonium in the environment – A review. Journal of Radiation Research 21, 83–104. HASL (1973). Global atmospheric plutonium-239 and plutonium isotopic ratios for 1959–1970. In: Fallout Program Quarterly Summary Report, HASL-237. US Department of Energy, pp. III-2–III-28. Hicks, H.G., Barr, D.W. (1984). Nevada test site fallout atom ratios: 240 Pu/239 Pu and 241 Pu/239 Pu. Lawrence Livermore National Laboratory, UCRL-53499/1, p. 4. Hirose, K. (1997). Complexation scavenging of plutonium in the ocean. In: Germain, P., Guary, J.C., Guéguéniat, P., Métivier, H. (Eds.), Radionuclides in the Oceans: Input and Inventories. Les Editions de Physique, Les Ulis, pp. 96–97. Hirose, K., Aoyama, M. (2003a). Analysis of 137 Cs and 239,240 Pu concentrations in surface waters of the Pacific Ocean. Deep-Sea Research II 50, 2675–2700. Hirose, K., Aoyama, M. (2003b). Present background levels of 137 Cs and 239,240 Pu concentrations in the Pacific. Journal of Environmental Radioactivity 69, 53–60. Hirose, K., Sugimura, Y. (1985). A new method of plutonium speciation in seawater. Journal of Radioanalytical and Nuclear Chemistry, Articles 92, 363–369. Hirose, K., Igarashi, Y., Aoyama, M., Miyao, T. (2001a). Long-term trends of plutonium fallout observed in Japan. In: Kudo, A. (Ed.), Plutonium in the Environment. Elsevier Science, Amsterdam, pp. 251–266. Hirose, K., Aoyama, M., Miyao, T., Igarashi, Y. (2001b). Plutonium in seawaters of the western North Pacific. Journal of Radioanalytical and Nuclear Chemistry, Articles 248, 771–776. Hirose, K., Miyao, T., Aoyama, M., Igarashi, Y. (2002). Plutonium isotopes in the Sea of Japan. Journal of Radioanalytical and Nuclear Chemistry, Articles 252, 293–299. Hisamatsu, S., Sakanoue, M. (1978). Determination of transuranium elements in a so-called “Bikini Ash” sample and in marine sediment samples collected near Bikini atoll. Health Physics 35, 301–307. Kim, C.S., Kim, C.K., Lee, J.I., Lee, K.J. (2000). Rapid determination of Pu isotopes and atom ratios in small amounts of environmental samples by an on-line sample pre-treatment system and isotope dilution high resolution inductively coupled plasma mass spectrometry. Journal of Analytical Atomic Spectrometry 15, 247–255. Kim, C.S., Kim, C.K., Lee, K.J. (2002). Determination of Pu isotopes in seawater by an on-line sequential injection technique with sector field inductively coupled plasma mass spectrometry. Analytical Chemistry 74, 3824–3832. Koide, M., Michel, R., Goldberg, E.D., Herron, M.M., Langway Jr., C.C. (1979). Depositional history of artificial radonuclides in the Ross ice sheet, Antarctica. Earth and Planetary Science Letters 44, 205–233. Komura, K., Sakanoue, M., Yamamoto, M. (1984). Determination of 240 Pu/239 Pu ratio in environmental samples based on the measurement of LX/X-ray activity ratio. Health Physics 46, 1213–1219. Krey, P.W., Hardy, E.P., Paxhucki, C., Rourke, F., Coluzza, J., Benson, W.K. (1976). Mass isotopic composition of global fall-out plutonium in soil. In: Transuranium Nuclides in the Environment, IAEA-SM-199/39. IAEA, Vienna, pp. 671–678. Livingston, H.D., Anderson, R.F. (1983). Large particle transport of plutonium and other fallout radionuclides to the deep ocean. Nature 303, 228–230.
82
K. Hirose et al.
Livingston, H.D., Bowen, V.T., Casso, S.A., Volchok, H.L., Noshkin, V.E., Wong, K.M., Beasley, T.M. (1985). Fallout Nuclides in Atlantic and Pacific Water Columns: GEOSECS Data. Woods Hole Oceanographic Institution, Woods Hole, MA, WHOI-85-19. Livingston, H.D., Mann, D.R., Casso, S.A., Schneider, D.L., Surprenant, L.D., Bowen, V.T. (1987). Particle and solution phase depth distributions of transuranics and 55 Fe in the North Pacific. Journal of Environmental Radioactivity 5, 1–24. Livingston, H.D., Povinec, P.P., Ito, T., Togawa, O. (2001). The behaviour of plutonium in the Pacific Ocean. In: Kudo, A. (Ed.), Plutonium in the Environment. Elsevier Science, Amsterdam, pp. 267–292. Miyake, Y., Sugiura, Y., Kameda, K. (1955). On the distribution of radioactivity in the sea around Bikini atoll in June, 1954. Papers in Meteorology and Geophysics 5, 235–262. Muramatsu, Y., Uchida, S., Tagami, K., Yoshida, S., Fujikawa, T. (1999). Determination of plutonium concentration and its isotopic ratio in environmental materials by ICP-MS after separation using ion-exchange and extraction chromatography. Journal of Analytical Atomic Spectrometry 14, 859–865. Muramatsu, Y., Hamilton, T., Uchida, S., Tagami, K., Yoshida, S., Robinson, W. (2001). Measurement of 240 Pu/239 Pu isotopic ratios in soils from the Marshall islands using ICP-MS. The Science of the Total Environment 278, 151–159. Nagaya, Y., Nakamura, K. (1984). Plutonium-239, plutonium-240, cesium-137, and strontium-90 in the central North Pacific. Journal of Oceanographical Society of Japan 40, 416–424. Perkins, R.W., Thomas, C.W. (1980). Worldwide fallout. In: Hanson, W.C. (Ed.), Transuranic Elements in the Environment. Tech. Inf. Center US Department of Energy, Washington, DC, pp. 53–82. Povinec, P.P., Livingston, H.D., Shima, S., Aoyama, M., Gastaud, J., Goroncy, I., Hirose, K., Hynh-Ngoc, L., Ikeuchi, Y., Ito, T., LaRosa, J., Kwong, L.L.W., Lee, S.-H., Moriya, H., Mulsow, S., Oregioni, B., Pettersson, H., Togawa, T. (2003). IAEA’97 expedition to the NW Pacific Ocean – Results of oceanographic and radionuclide investigations of the water column. Deep-Sea Research II 50, 2607–2637. Povinec, P.P., Aarkrog, A., Buesseler, K.O., Delfanti, R., Hirose, K., Hong, G.H., Ito, T., Livingston, H.D., Nies, H., Noshkin, V.E., Shima, S., Togawa, O. (2005). 90 Sr, 137 Cs and 239,240 Pu concentration surface water time series in the Pacific and Indian Oceans – WOMARS results. Journal of Environmental Radioactivity 81, 63–87. Quay, P.D., Stuiver, M., Broecker, W.S. (1983). Upwelling rates for the equatorial Pacific Ocean derived from the bomb 14 C distribution. Journal of Marine Research 41, 769–792. Taylor, R.N., Warneke, T., Milton, J.A., Croudace, L.W., Warwick, P.E., Nesbitt, R.W. (2001). Plutonium isotope ratio analysis at femtogram levels to nanogram levels by multicollector ICP-MS. Journal of Analytical Atomic Spectrometry 16, 279–284. Tsumune, D., Aoyama, M., Hirose, K. (2003). Numerical simulation of 137 Cs and 239,240 Pu concentrations by an ocean general circulation model. Journal of Environmental Radioactivity 69, 61–84. Warneke, T., Croudace, L.W., Warwick, P.E., Taylor, R.N. (2002). First ground-level fallout record of uranium and plutonium isotopes for northern template latitude. Earth and Planetary Science Letters 203, 1047–1057. Yamamoto, M., Tsumura, A., Katayama, Y., Tsukatani, T. (1996). Plutonium isotopic composition in soil from the former Semipalatinsk nuclear test site. Radiochimica Acta 72, 209–215.
83
Distribution of anthropogenic radionuclides in the water column off Rokkasho, Japan Shigeki Shima* , Shin-ichi Gasa, Ken-ichi Iseda, Tomoharu Nakayama, Hisao Kawamura Japan Marine Science Foundation, Mutsu, Aomori 035-0064, Japan Abstract Anthropogenic iodine-129, carbon-14, plutonium-239,240, strontium-90, caesium-137 and tritium were determined together with oceanographic parameters in the water column off Rokkasho where liquid wastes from a reprocessing plant will be discharged. This sea area is located at the mixing zone of the subtropical and the subarctic waters. The differences in radionuclide concentrations between the subtropical and the subarctic waters are presented and discussed. Except for 129 I, the observed concentrations of anthropogenic radionuclides in the water column offshore of Rokkasho can be explained by global fallout and processes in the water column. However, the atom ratio of iodine to caesium was ten times as high as that of global fallout, therefore contributions from reprocessing plants should be responsible for the observed elevated 129 I levels in the Northwest Pacific. Keywords: Anthropogenic radionuclides, 3 H, 14 C, 129 I, 239,240 Pu, Reprocessing plant, Seawater, Coastal water, Water column, Oceanic observation, Water masses, Rokkasho, Northwest Pacific
1. Introduction The first commercial facility for reprocessing nuclear spent fuel in Japan is to be opened in summer 2006 in Rokkasho, in the north-eastern part of Japan. The facility will routinely release gaseous and liquid wastes containing radionuclides into the surrounding environment. The liquid wastes will be discharged into coastal waters at a distance of 3 km from the coastal line. The Rokkasho area is situated offshore from Sanriku, in the boundary where the subarctic (Oyashio current – OC) and subtropical (Kuroshio current – KC) currents meet (Tomczak and Godfery, 2001). The Tsugaru Warm Current (TWC) flows into this region through the Tsugaru Strait from the Sea of Japan. These three water masses with different origins coexist in the surface layer of this domain. It has been known that there are differences in radionuclide concentrations between the subtropical and the subarctic gyre in the Pacific Ocean (Aoyama * Corresponding author. Address: Japan Marine Science Foundation, 4-24 Minato-machi, Mutsu, Aomori 035-0064, Japan; phone: (+81) 175 22 9111; fax: (+81) 175 22 9112; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08006-X
© 2006 Elsevier Ltd. All rights reserved.
84
S. Shima et al.
et al., 2001; Hirose and Aoyama, 2003). Therefore, it is important to understand the distribution of anthropogenic radionuclides in the region in advance of the operation of the nuclear reprocessing plant.
2. Oceanographic scope The area investigated in this paper is defined as the area from 39◦ N to 42.5◦ N and from 141◦ E to 143.5◦ E. It includes the Kuroshio/Oyashio mixed water region, where subtropical and subarctic waters meet and interact with each other. Therefore this region is also called the perturbed region (Kawai, 1972). The area is influenced by the TWC, the OC and the KC with several small or meso-scale eddies, including the Kuroshio warm-core ring. The Kuroshio begins west of the Philippines where the North Equatorial Current advects westward, continues northward east of Taiwan, flows along the eastern coast of the Japan Islands, and deviates from Honshu Island around 141◦ E, 35◦ N. The OC is formed by the Okhotsk Sea waters and the Kamchatka Current west of the Kamchatka Peninsula at about 55◦ N. It flows southward just south of Hokkaido Island and splits into two paths called the First and Second Oyashio Intrusion. A branch of the KC flows into the Japan Sea via the Korea Strait and flows out to the Pacific Ocean through the Tsugaru Strait (sill depth about 200 m) as the TWC. The volume transport from the strait varies seasonally, with volume transport of about 2.8 Sv in September and of only 1.2 Sv in March (Toba et al., 1982). This water advects partly southward along Honshu Island (the main island), while another part moves eastward against the advance of the OC. Since the path of the TWC varies seasonally and the nutrient-rich OC meets in this region, the area off Sanriku is known to be complex in oceanographic conditions and currents system, as well as in high fishing activities (Tomczak and Godfery, 2001; Conlon, 1982).
3. Materials and methods 3.1. Sampling and in situ measurements A CTD (SBE 20)/MBS with multi-bottle sampler SBE 32, SeaBirds Inc.) and LVS (largevolume sampler, 120 L Van Dom bottle) were used on October 2001 and June 2002 for water sampling at the stations shown in Fig. 1. In 2001 the stations were visited along the meridional section at 142.5◦ E, and in 2002 along the zonal section at 40.5◦ N, because of the behaviour of the TWC, i.e. within the TWC gyre in autumn, and across the coastal streamline of the TWC in early summer. For each station, surface and bottom water and three intermediate water samples (at 200, 300 and 500 m depths) were collected in order to obtain information on the distribution of anthropogenic radionuclides in this area. For 90 Sr, 137 Cs, and 239,240 Pu analyses, 120 L water samples were acidified with hydrochloric acid (pH = 1). One-litre samples for 14 C analyses were stored in special glass bottles with air-tight cups in order to avoid contamination from the surrounding air. Additionally, during each cruise at about hundred sites, including the water sampling stations, CTD observations were also conducted using the SBE 20 in order to get information on oceanographic conditions in this area.
Distribution of anthropogenic radionuclides in the water column off Rokkasho
85
Fig. 1. Location of sampling stations with bathymetry. The maximum depth of all stations is below 1,800 m.
3.2. Analytical methods Tritium in seawater was enriched using a solid polymer electrolyte (SPE) method, and its activity was determined in a low background liquid scintillation counter. Caesium-137 in seawater was pre-concentrated by adsorption on ammonium molybdophosphate (AMP), which was collected by filtration and counted by an HPGe detector with lead shielding. Carbon-14 (as dissolved inorganic carbon, DIC) from seawater was extracted as CO2 by acidification with phosphoric acid. After a purification of CO2 in a vacuum line, a graphite target was prepared, and the carbon isotope ratio was measured in the JAERI accelerator mass spectrometer (AMS) in Mutsu. δ 13 C measurements were carried out using an isotope mass spectrometer. The 14 C activity in seawater samples is expressed by 14 C (h), defined as 14 C = (Fm − 1) × 103 , where Fm (a fraction of modern carbon) is the measured AMS ratio of 14 C to 13 C, normalised to δ 13 C of −25h (Donahue et al., 1990). After recovery of iodine from seawater and its purification by the n-hexane extraction method, silver iodide targets for AMS were prepared. Iodine-129 was measured by AMS at the Isotrace Laboratory, University of Toronto. The 129 I concentrations are expressed in 106 atoms/L. Strontium in seawater was separated and purified by an ion exchange method and the activity of 90 Sr was determined by counting the β-rays emitted by its 90 Y daughter in radioactive equilibrium using a low background gas flow counter. Plutonium isotopes were separated by co-precipitation from seawater, purified by use of an anion exchange column, and electroplated onto a stainless steel disc. 238 Pu and 239,240 Pu activities were measured by α-spectrometry. Measured activity concentrations of radionuclides (except 14 C and 129 I) are expressed in mBq/L.
86
S. Shima et al.
4. Results and discussion 4.1. Oceanographic observations Vertical profiles of temperature and salinity above a water depth of 700 m varied seasonally and spatially (Fig. 2), however, below the 700 m water depth they were almost the same during sampling periods. Intrusion of water masses with lower temperatures and salinities appeared in the upper layer, as shown in Fig. 2. Salinity in the upper water layer of station 01W-1 was the lowest amongst all the stations. The highest salinity above the depth of 80 m was found at station 02W-1. The density (sigma-t) profiles also largely fluctuated in the upper layer, since they depend on in situ temperature and salinity. Moreover, the complexity of oceanographic conditions was inferred from a large difference (∼200 m) in water depths that have the same density in the upper layer. In addition to large vertical variations, the horizontal distributions of temperature and salinity were also complicated. For example, the distribution of temperature at 200 m depth is shown in Fig. 3. In June, the warmer TWC distributes along the coastal side of the Japanese main island and the colder Oyashio occupied the greater part of this area. The warmer water mass around latitude 41◦ N and longitude 143◦ E seems to derive from the warm core separated from the Kuroshio in the Kuroshio Extension region (Japan Coastal Guard, 2002). Although the TWC widely spread in this region in October, the expanse of the TWC is smaller than usual (Shima et al., 2000). The intrusion of the OC was recognised in the middle zone of this area. The distributions of temperature and salinity were also reflected in the
Fig. 2. Vertical profiles of temperature, salinity and density (sigma-t).
Distribution of anthropogenic radionuclides in the water column off Rokkasho
87
Fig. 3. Horizontal distribution of temperature at the depth of 200 m.
Table 1 Definition of water masses off Sanriku Water mass
Definition
TWC Kuroshio Oyashio Deep water Coastal Oyashio Surface water
33.7 S < 34.2, T 5 and σT 24 S > 34.2 and σT < 26.7 S > 33, T < 7 and σT < 26.7, except for T > 5 and S 33.7 σT > 26.7 > T , except for T > 5 and 33.7 S < 34.2 S < 33 and T < 2 Other than the above
water density. The difference in water depths having the same density reached hundreds of meters. Compared with the horizontal distributions of temperature, the isobathymetric line of the same density is deeper in the warm region than that in the colder one. It is well known that the TWC flows along the Japanese coastal line in winter to spring (called coastal mode) and forms a clockwise gyre in summer to autumn (called gyre mode) (Conlon, 1982; Hishida, 1987). The observed horizontal distributions of temperature appeared to reflect the characteristic flow pattern of the currents. The classification of water masses in the area was done following the objectives of research (Hanawa and Mitsudera, 1986; Yasuda et al., 1988; Inagake and Saitoh, 1998). In this paper, Hanawa’s criterion was adopted for classifying the water system distribution in the upper layer offshore of Sanriku (Table 1). According to this criterion, thirty large volume water samples were collected in these systems: TWC (6 samples), the Oyashio (4), the deep water (16), the
88
S. Shima et al.
surface water (3), and the Kuroshio (1). The classification of all sampling stations is shown in Figs. 4a, 4b, 4c. It can be seen that the TWC, the surface water, and the Kuroshio dominated over the Oyashio in this region. 4.2. Distribution of radionuclides 4.2.1. Carbon-14 As shown in Fig. 4a, 14 C values in surface waters were in excess to the modern carbon ratio, from 45 to 75h, except for station 01W-1 (−14h). The average value of 14 C in surface waters in the North Pacific Basin obtained from the WOCE project was 75h. The values in the western subarctic zone were lower than those in the subtropical one (Key et al., 2002). As the 14 C values in the subarctic gyre (45–50◦ N, 165◦ E in P13N line of the WOCE) are negative, the water with negative 14 C value observed in the subsurface of station 01W-1 must be of Oyashio origin. Vertical profiles of 14 C decreased with depth rapidly and monotonically, except for near the sea surface. In stations 01W-2, 01W-3 and 02W-3 where the mixing layer
Fig. 4a. Vertical profile of 14 C in the water column. The box plot on the right shows the water masses as classified in this paper.
Fig. 4b. Vertical profile of 129 I in the water column. The box plot on the right shows the water masses as classified in this paper.
Distribution of anthropogenic radionuclides in the water column off Rokkasho
89
Fig. 4c. Vertical profile of 239,240 Pu in the water column. The box plot on the right shows the water masses as classified in this paper.
Fig. 5. Plot of 14 C vs. density.
reached down to 200 m in depth, the 14 C values in surface waters were almost the same, or less than those at a depth of 200 m. Although the vertical profiles of 14 C in each cruise are divided into two groups, relationships between 14 C and σT (sigma-t) show no difference between the 14 C profiles. As shown in Fig. 5, the 14 C values are within uncertainties the same for densities of (24.5–26.5)σT , except for the value in surface water of station 01W-1. Most of the bomb-produced 14 C remains in this area in shallow waters with density < 26.5σT . Recently, instead of the SiO2 –14 C relationship (Broecker et al., 1995), a new method has been proposed to infer natural 14 C in the ocean (Rubin and Key, 2002). However, as alkalinity data are required, we could not use this method, as we do not have such data available for each cruise. Using Broecker’s relationship between 14 C and dissolved SiO2 , the bomb-produced 14 C has penetrated to an isopycnal of about 27.0.
90
S. Shima et al.
The average values of 14 C for each water mass are (40.7 ± 37.9)h for the TWC, (38.3 ± 36.2)h for the Oyashio and (105 ± 82.4)h for the deep water (see Table 1). Although there are large uncertainties in these values, the differences between water masses have been recognised. These differences in water masses offshore of Rokkasho may disappear due to disturbances in mixing at the outlet of the Tsugaru Strait, as well as due to high primary production. 4.2.2. Iodine-129 Concentrations of 129 I in surface waters are from (17 ± 4) × 106 to (23 ± 4) × 106 atoms/L, and it looks like there are not temporal and spatial variations similar to that observed for 14 C. Similar concentrations have been observed in surface waters of the Japan Sea ((12–31) × 106 atoms/kg) and in coastal water of Vladivostok harbour (58×106 atoms/kg; Cooper et al., 2001). Concentrations in the subtropical zone of the Northwest Pacific were similar (Povinec et al., 2000). The surface concentrations around Japan are almost two orders of magnitude lower than those in the Barents and Kara Seas in the Arctic, where European nuclear fuel reprocessing effluents dominate the 129 I signal (Smith et al., 1998). As mentioned later in discussion of activity ratios, the observed 129 I concentrations in the water column may be higher than that expected from global fallout. The vertical profiles of 129 I show rapid decrease with depth, as observed for 14 C, although the inversion of the 129 I concentration slightly appears in the upper layers (Fig. 4b). The 129 I concentrations of bottom water below 1,400 m depth were found to be under the detection limit. In the Pacific and the Sea of Japan, 129 I was found in deep waters, at several thousand meters. The observed 129 I profiles may be caused by shallower bathymetry, or differences in circulation of aged deep water (Schlosser et al., 2001). From the plot of 129 I concentrations as a function of sigma-t, the concentrations rapidly decrease below the isopycnal of about 26.7. The water depth of this isopycnal surface corresponds to 100–400 m. The plot of 129 I vs. 14 C shows that the 129 I concentration is decreasing with the water age (Fig. 6). The straight line in this figure is the 129 I concentration expected from global fallout. The difference between the observed and expected 129 I values suggests that another source of 129 I may influence its concentration offshore of Rokkasho. The average concentrations of 129 I in each water system are (15.8 ± 1.9) × 106 , (12.0 ± 2.6) × 106 and (6.8 ± 5.7) × 106 atoms/L for the TWC, Oyashio and deep waters, respectively. The 129 I concentration in the TWC is slightly higher than that in the Oyashio. If iodine in the ocean behaves similarly as radiocarbon, the 129 I concentration may be higher in the subtropical than in the subarctic zone. 4.2.3. Plutonium-239,240 The concentrations of 239,240 Pu in surface water were under the limit of detection (<0.005 mBq/L), except for station 02W-3 (0.0051 mBq/L). The concentrations in the intermediate layers and the bottom water ranged from 0.0058 to 0.025 mBq/L. These results are similar to those observed in 2001 in surface and bottom waters near to the coastal zone (Mishonoh and Inatomi, 2003). Since the concentrations become larger with increasing sampling depths and those at the bottom are smaller than those in the upper layers, the 239,240 Pu maxima seem to be between 500 m and 1,000 m (Fig. 4c). The observed vertical distributions of 239,240 Pu seem to be similar to those in the open Northwest Pacific (Povinec et al.,
Distribution of anthropogenic radionuclides in the water column off Rokkasho
91
Fig. 6. Relationship between 14 C and 129 I.
2003). From the plots of 239,240 Pu concentrations vs. densities, the 239,240 Pu concentrations rapidly increased in deeper waters at density of 26.7, and showed maxima at the isopycnal of ∼27σT . Higher concentrations of 239,240 Pu were also observed at water depths showing old radiocarbon ages. This is a result of scavenging and sinking of Pu into deeper water depths, as described in many references (for example, Hirose, 1997). The average concentrations of 239,240 Pu in each water system are 0.0064 ± 0.0017, 0.010 ± 0.0022 and 0.017 ± 0.004 mBq/L for the TWC, Oyashio and deep waters, respectively. The difference between the TWC and the Oyashio waters is very clear. Similar differences have been found in the North Pacific as well. Higher surface 239,240 Pu concentrations in the Subarctic Pacific can be explained by rapid recycling of scavenged 239,240 Pu due to large vertical mixing in winter (Hirose and Aoyama, 2003). 4.2.4. Distributions of short-lived radionuclides Concentrations of 3 H, 90 Sr and 137 Cs in surface waters ranged from 130 to 160, 1.4 to 1.8 and <0.6 to 2.8 mBq/L, respectively. The 3 H and 137 Cs profiles did not decrease with depth monotonously. Their vertical profiles reflected a complex structure of water masses in this area. The concentration vs. density plots show that waters from the surface to 26.7σT mix uniformly, and that the short-lived radionuclides are not present below 27σT . The average concentrations of 90 Sr for each water system are 1.51 ± 0.16, 1.06 ± 0.20 and 0.57 ± 0.42 for the TWC, Oyashio and deep waters, respectively. The averages of 137 Cs are 2.22 ± 0.49, 1.75 ± 0.25 and 0.48 ± 0.58 for the TWC, Oyashio and deep waters, respectively. In the late 1990s, 137 Cs concentrations in surface water of the North Pacific (around 50◦ N) were lower than those in the subtropical zone, and may have been affected by deeper convection in this sea area (Aoyama et al., 2001; Hirose and Aoyama, 2003). Lower 90 Sr and 137 Cs concentrations were observed in spring 2001 in several time-series stations
92
S. Shima et al.
located near the coastal zone (Mishonoh and Inatomi, 2003), which also could be a result of an intrusion of the Oyashio into the coastal area. 4.3. Radionuclide ratios Beasley et al. (1998) estimated that the 129 I/137 Cs atom ratio from global fallout in the North Atlantic Ocean for 1994 is 0.49. Edmonds et al. (1998) calculated the 129 I/137 Cs atom ratio from analysis of archived seawater samples collected in 1969 in the subequatorial Atlantic Ocean to be 4.1 ± 1.5 (decay corrected to 2003). The 129 I/137 Cs atom ratios observed in this work in surface waters of the Rokkasho area ranged from 4.4 to 7.1. These ratios are similar to that of Edmonds et al. (1998), about ten times higher than expected from global fallout. Such a disagreement between the observed values and the estimates for global fallout was also found in water samples of the Japan Sea (Cooper et al., 2001). The Rokkasho facility for uranium enrichment, and the disposal of low-level radioactive waste and the storage of spent nuclear fuel has been running from 1992 and 1995, respectively. A small demonstration plant for reprocessing fuel has been in intermittent operation since 1977 in Tokai, Japan. The effects of these facilities seem to be ignored, because the facilities in Rokkasho hardly release 129 I to the environment, and the reprocessing plant in Tokai is located about 700 km southward from the study area. Morgan et al. (1999) pointed out that higher than expected 129 I concentrations and resulting 129 I/137 Cs atom ratios may reflect direct atmospheric deposition of 129 I aerosols from nuclear facilities. However, we have no evidence of a significant local source, therefore further investigations in the North Pacific and marginal seas are necessary to explain the 129 I excess. The 239,240 Pu/90 Sr and 239,240 Pu/137 Cs activity ratios in surface water were 0.0030 and 0.0024, respectively, significantly lower than those expected from global fallout (239,240 Pu/ 90 Sr: 0.046, 239,240 Pu/137 Cs: 0.029, both decay corrected to 2003, UNSCEAR, 1993). As shown in Fig. 7, the activity ratios increased with depth from 0.003 to 0.034 for 239,240 Pu/90 Sr and from 0.0024 to 0.022 for 239,240 Pu/137 Cs, respectively. These results confirm again that plutonium is preferentially removed from surface waters. On the other hand, the activity ratios of 137 Cs/90 Sr ranged from 0.94 to 1.83 (average: 1.44), and are as expected from global fallout. 4.4. Radionuclide inventories Table 2 shows inventories of radionuclides for each station from the sea surface to the bot14 C was calculated as an average value. We adopted tom. The inventory of bomb-produced Broecker’s method using CO2 values from the oceanographic database of the Japanese Oceanographic Data Center. 129 I inventories included both the anthropogenic and natural ones, as we have no method to distinguish between them. The 129 I inventory in the Japan Sea (for a similar water depth) was three times as high as that in this area, since 129 I was found at deeper layers in the Japan Sea due to subduction of surface waters in winter. The calculated inventories are lower than those in the open Northwest Pacific Ocean. The calculated radionuclide inventory ratios are 1.1 ± 0.3 for 137 Cs/90 Sr, 0.023 ± 0.007 for 239,240 Pu/90 Sr and 0.027 ± 0.020 for 239,240 Pu/137 Cs, within uncertainties as expected from global fallout. The 137 Cs/90 Sr ratios in each water system were 1.46 for the TWC, 1.65 for
Distribution of anthropogenic radionuclides in the water column off Rokkasho
93
Fig. 7. Vertical profiles of activity ratios and the 129 I/137 Cs atom ratio vs. sigma-t.
Table 2 Inventories of radionuclides in the water column
(degree)
90 Sr Longi- 3 H tude (degree) (kBq/m2 ) (kBq/m2 )
129 I 137 Cs 239,240 Pu Bomb-produced 14 C∗ (1012 atoms/m2 ) (kBq/m2 ) (Bq/m2 ) (109 atoms/cm2 )
40.66 40.92 41.41 40.50 40.50 40.55
142.50 82 ± 16 142.33 75 ± 16 142.33 108 ± 20 143.05 40 ± 10 142.67 90 ± 28 142.25 79 ± 12
0.67 ± 0.14 0.89 ± 0.16 0.83 ± 0.19 0.84 ± 0.19 0.89 ± 0.18 0.61 ± 0.07
5.1 ± 1.2 3.9 ± 1.4 4.6 ± 1.2 4.3 ± 0.5 3.9 ± 0.6 3.9 ± 0.5
0.8 ± 0.2 1.2 ± 0.2 1.0 ± 0.2 1.1 ± 0.3 0.4 ± 0.1 0.9 ± 0.1
79 ± 23
0.79 ± 0.12
4.3 ± 0.5
0.9 ± 0.3 18 ± 7
Station Water No. depth (m)
Latitude
01W-1 01W-2 01W-3 02W-1 02W-2 02W-3
1,083 1,094 1,198 1,534 1,376 564
Average ± std. dev.
16 ± 3 16 ± 3 17 ± 4 24 ± 5 28 ± 5 7±1
>26.8σT
11
∗ Natural 14 C was estimated from the 14 C–SiO relation and CO . 2 2
the Oyashio, and 1.34 for the deep water, respectively, similar to that of global fallout. The 239,240 Pu/90 Sr and 239,240 Pu/137 Cs ratios in the two water systems were lower for upper water layers than that of the global fallout as expected, because of preferential removal of plutonium from surface waters.
94
S. Shima et al.
The 129 I/137 Cs ratio of inventories is approximately ten times higher than expected from global fallout, and may reflect possible effects from reprocessing plants.
5. Conclusions Relationships between oceanographic characteristics and distributions of anthropogenic and 239,240 Pu were examined in the water column offshore of Rokkasho. Although significant changes in the water quality such as temperature and salinity were observed in the upper water column, radionuclides were uniformly distributed downward in comparison with the same density. The activity ratios of studied anthropogenic radionuclides can be explained by global fallout and processes in the water column. However, the 129 I/137 Cs atom ratios were ten times as high as that of global fallout, therefore contributions from reprocessing plants should be responsible for the observed 129 I levels in the study area, and further studies will be needed in order to clarify the sources of 129 I. Compared with classification of water masses, the 14 C, 90 Sr and 137 Cs concentrations in the TWC were higher than those in the Oyashio. In contrast to those results, the 239,240 Pu concentrations were higher in the TWC. These tendencies can be caused by horizontal advection, vertical mixing, and biogeochemical processes in the Pacific surface waters. The observed distribution of radionuclides and oceanographic parameters indicated that part of the waters in the upper layer should have its origin from the subtropical gyre, however, contributions from the subarctic zone should also be taken into account, which makes the study area of great interest for future oceanographic investigations. 3 H, 14 C, 90 Sr, 129 I, 137 Cs
Acknowledgements A part of this research was conducted under a contract with the Government of Aomori Prefecture, Japan. We thank the captain and the crew of the R/V Daigo–Kaiyo–Maru for assistance during sampling, and Dr. K. Hasunuma for useful information on the oceanographic characteristics of the study area. We wish to thank the staffs of the institutes and the university for carrying out analysis of some of the radionuclides studied in the paper. We also thank two anonymous reviewers and Prof. P. Povinec for improving the manuscript.
References Aoyama, M., Hirose, K., Miyao, T., Igarashi, Y., Povinec, P.P. (2001). 137 Cs activity in surface water in the western North Pacific. Journal of Radioanalytical & Nuclear Chemistry 248 (3), 789–793. Beasley, T., Cooper, L.W., Grebmeier, J., Aagaard, K., Kelley, J.M., Kilius, L.R. (1998). 237 Np/129 I atom ratios in the Arctic Ocean: Has 237 Np from western European and Russian fuel re-processing facilities entered the Arctic Ocean? Journal of Environmental Radioactivity 39, 255–277. Broecker, W.S., Sutherland, S., Smethie, W., Peng, T.-H., Ostlund, G. (1995). Oceanic radiocarbon: Separation of the natural and bomb components. Global Biogeochemical Cycles 9, 263–288. Conlon, D.M. (1982). On the Outflow Modes of the Tsugaru Warm Current. La mer 20, 60–64. Cooper, L.W., Hong, G.H., Beasley, T.M., Grebmeier, J.M. (2001). Iodine-129 concentrations in marginal seas of the North Pacific and Pacific-influenced waters of the Arctic Ocean. Marine Pollution Bulletin 42, 1347–1356.
Distribution of anthropogenic radionuclides in the water column off Rokkasho
95
Donahue, D.J., Linick, T.W., Juli, A.J.T. (1990). Isotope-ratio and background corrections for accelerator mass spectrometry radiocarbon measurements. Radiocarbon 32, 135–142. Edmonds, H.N., Smith, J.N., Livingston, H.D., Kilius, L.R., Edmond, J.M. (1998). 129 I in archived seawater samples. Deep-Sea Research I 45, 1111–1125. Hanawa, K., Mitsudera, H. (1986). Variation of water system distribution in the Sanriku Coastal Area. Journal of Oceanography 42, 435–446. Hirose, K. (1997). Complexation scavenging of plutonium in the ocean. Radioprotection – Colloques 32, C2-225– C2-230. Hirose, K., Aoyama, M. (2003). Analysis of 137 Cs and 239,240 Pu concentrations in surface waters of the Pacific Ocean. Deep-Sea Research II 50, 2675–2700. Hishida, M. (1987). On southern movement and seasonal variation of Tsugaru warm current. Report of Hydrographic Researches 22, Japan Coastal Guard. (In Japanese.) Inagake, D., Saitoh, S. (1998). Description of the oceanographic condition off Sanriku, Northwestern Pacific, and its relation to spring bloom detected by the ocean color and temperature scanner (OCTS) images. Journal of Oceanography 54, 479–494. Japan Coastal Guard (2002). Quick Bulletin of Ocean Conditions, No. 26. (In Japanese.) Kawai, H. (1972). Hydrography of the Kuroshio Extension. In: Stommel, H., Yoshida, K. (Eds.), Kuroshio: Its Physical Aspects. University of Tokyo Press, pp. 235–352. Key, R.M., Quay, P.D., Schlosser, P., McNichol, A.P., von Reden, K.F., Schneider, R.J., Elder, K.L., Stuiver, M., Östlund, H.G. (2002). WOCE radiocarbon IV: Pacific Ocean results; P10, P13N, P14C, P18, P19 & S4P. Radiocarbon 44, 239–392. Mishonoh, J., Inatomi, N. (2003). The level of radioactivities in marine environment off the pacific coast of Aomori Prefecture. In: Inaba, J., Tsukada, H., Takeda, A. (Eds.), Radioecology and Environmental Dosimetry. Institute for Environmental Science, Aomori, Japan, pp. 372–376. Morgan, J.E., Oktay, S., Santschi, P., Schink, D. (1999). Atmospheric dispersal of 129 iodine from nuclear fuel reprocessing facilities. Environmental Science & Technology 33, 2536–2542. Povinec, P.P., Oregioni, B., Jull, A.J.T., Kieser, W.E., Zhao, X.-L. (2000). AMS measurements of 14 C and 129 I in seawater around radioactive waste dump sites. Nuclear Instruments & Methods in Physics Research B 172, 672– 678. Povinec, P.P., Livingston, H.D., Shima, S., Aoyama, M., Gastaud, J., Goroncy, I., Hirose, K., Huynh-Ngoc, L., Ikeuchi, Y., Ito, T., La Rosa, J., Kwong, L.L.W., Lee, S.-H., Moriya, H., Mulsow, S., Oregioni, B., Pettersson, H., Togawa, O. (2003). IAEA’97 expedition to the NW Pacific Ocean – Results of oceanographic and radionuclide investigations of the water column. Deep-Sea Research II 50, 2607–2638. Rubin, S., Key, R.M. (2002). Separating natural and bomb-produced radiocarbon in the ocean: The potential alkalinity method. Global Biogeochemical Cycles 16, 52-1–52-19. Schlosser, P., Bullister, J.L., Fine, R., Jenkins, W.J., Key, R., Lupton, J., Roether, W., Smethie, W.M. (2001). Transformation and age of water masses. In: Sieder, G., Church, J., Gould, J. (Eds.), Ocean Circulation & Climate: Observing and Modelling the Global Ocean. Academic Press, London, pp. 431–452. Shima, S., Nakayama, T., Iseda, K., Nishizawa, K., Gasa, S., Suto, K., Sakurai, S., Oguri, K., Kouzuma, K. (2000). Distribution and seasonal change of the Tsugaru warm current water off Rokkasho. In: Inaba, J. et al. (Eds.), Distribution and Speciation of Radionuclides in the Environment. Institute for Environmental Science, Aomori, Japan, pp. 289–296. Smith, J.N., Ellis, K.M., Kilius, L.R. (1998). 129 I and 137 Cs tracer measurements in the Arctic Ocean. Deep-Sea Research I 45, 959–984. Toba, Y., Tomizawa, K., Kurasawa, Y., Hanawa, K. (1982). Seasonal and year to year variability of the Tsushima– Tsugaru warm current system with its possible cause. La mer 20, 41–51. Tomczak, M., Godfery, J.S. (2001). Regional Oceanography: An Introduction. Pdf version I. UNSCEAR (1993). Sources and effects of ionizing radiation. United Nations, New York. Yasuda, I., Okuda, K., Hirai, M., Ogawa, Y., Kudo, H., Fukuhara, S., Muzuno, K. (1988). Short-term variations of the Tsugaru warm current in autumn. Bulletin of Tohoku Regional Fisheries Research Laboratory 50, 153–191. (In Japanese).
96
Artificial radionuclides in the Yellow Sea: Inputs and redistribution G.H. Honga,* , C.S. Chunga , S.-H. Leeb , S.H. Kima , M. Baskaranc , H.M. Leea , Y.I. Kima , D.B. Yang, C.K. Kimd a Korea Ocean Research and Development Institute, Ansan, Seoul, Republic of Korea b Marine Environment Laboratory, International Atomic Energy Agency, Monaco c Department of Geology, Wayne State University, Detroit, MI 48202, USA d Chemistry Unit, PCI, Agency’s Laboratories, International Atomic Energy Agency, Seibersdorf, Austria
Abstract The Yellow Sea is one of the marginal seas of the Northwest Pacific receiving large amounts of material from the continent via rivers and atmosphere. In order to understand the sources and present levels of key artificial radionuclides (90 Sr, 137 C and 239+240 Pu) in the Yellow Sea, the processes affecting their distribution in the water column and their burial in the sea floor, their concentrations were determined in seawater and bottom sediment samples collected from the Yellow Sea during 1994–2000, and from the East China Sea and the tropical Northwest Pacific during 1993 and 1994. The atmospheric and riverine inputs were also assessed at the mid-eastern coast of the Yellow Sea. The atmospheric deposition of radionuclides appears to be dominated by the long-range transport from the arid regions of the Asian continent with the highest values during the spring Asian dust storms and lowest in the summer wet monsoon period. The dry atmospheric deposition flux appeared to be particularly important for 239+240 Pu. Riverine fluxes of radionuclides dominated the total input due to the sheer size of the riverine water and sediment fluxes into the sea. The river input was seen in their distribution in the surface of the sea, particularly for 90 Sr in winter. In summer, the water column stratification segregates these radionuclides vertically, so they are depleted in the surface layer and enriched in the bottom layer. The half-removal rate for 90 Sr and 137 Cs was estimated to be 7 years. The levels of radionuclides in the Yellow Sea were higher than in the adjacent seas, and significant amounts of them have been exported from the Yellow Sea to the adjacent seas. Keywords: Radionuclides, 90 Sr, 137 Cs, 239+240 Pu, Atmosphere, River water, Seawater, Sediment, East China Sea, Yellow Sea, Northwest Pacific
1. Introduction As a part of the ongoing Radionuclides Study in the Seas (RADSEA) Program of the Korea Ocean Research and Development Institute, concentrations of artificial radionuclides (90 Sr, 137 Cs and 239+240 Pu) were determined in seawater and bottom sediments in the Yellow Sea * Corresponding author. Address: KORDI, Ansan P.O. Box 29, Seoul 425-600, Republic of Korea; phone: (+82) 31 400 6180; fax: (+82) 31 408 4493; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08007-1
© 2006 Elsevier Ltd. All rights reserved.
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
97
from 1994 to 2000. In addition, the atmospheric and riverine inputs of these radionuclides to the ocean were also assessed. The Yellow Sea, which has an area of 380 km2 and a mean depth of 44 m, is surrounded by the contiguous landmass of China and Korea, which is a part of the largest continental landmass on the Earth. Large and small rivers discharge into the Yellow Sea under prevailing topography and climate. The sea rests in a broad and tectonically stable area that was submerged by the latest post-glacial rise in the sea level. The general water circulation in the Yellow Sea can be characterized by the northward movement of the Yellow Sea Warm water that provides warm and saline water originating from the Kuroshio, and southward movement by the Yellow Sea Coastal Current on the western side. The residence time of the Yellow Sea Proper water is about 5–6 years and particles generally reside less than 2 months in the water column. Although the geochemistry of the Yellow Sea is mainly controlled by the ocean-derived water, salt, heat and chemical materials, processes that occur at the boundaries (air–sea, sediment–water and land–sea) significantly modify the current geochemical processes of the Yellow Sea, which in turn affects the East China Sea. In particular, influence of land and atmosphere is very significant due to the semi-enclosed nature of the sea and its location (situated in the down-wind side of the arid continent of the northern China). The Yellow Dust storm events are one of the salient geological features of the region (Hong et al., 1988). After a long cessation of active atmospheric nuclear weapons tests, levels of fallout radionuclides including 90 Sr, 137 Cs, and Pu isotopes in surface waters of the Pacific Ocean have decreased considerably over the past 5 decades and these radionuclides are now much more homogeneously distributed (Hamilton et al., 1996). In the marginal seas of the Northwest Pacific Ocean, the Yellow Sea serves as an ideal site for investigating the redistribution of artificial radionuclides that were once deposited on land and their subsequent transport to the sea by the interplay of transporting mediums of wind, river, and ocean currents due to the presence of large rivers and dust storms. Here, we attempt to report the current level of the key artificial radionuclides in the Yellow Sea as well as their redistribution in a continental margin setting. This work will also complement the monitoring of artificial radionuclides for the Yellow Sea prior to the 1980s (Zhu et al., 1991; Li et al., 1994). 2. Materials and methods 2.1. Soil core A single core was collected in 1994 using a PVC tube at a flat area within the Kwangrung National Arboretum (37◦ 44 N, 127◦ 00 E). This area is relatively flat and has been protected since 15th century as a part of the royal forest. The soil core was sliced at 1 cm intervals and dried at 105◦ C and ground with mortar to obtain particles less than 10-mesh size. 2.2. Atmospheric samples The atmospheric dust samples were collected by placing a dust collector on the roof of the Korea Ocean Research and Development Institute (KORDI) building in Ansan, which is lo-
98
G.H. Hong et al.
cated at the mid-eastern coast of the Yellow Sea (37◦ 17 N, 126◦ 50 E). Prior to rainout events, the dust collector was manually removed and replaced by a rain collector. The dust collector was deployed from May 1994 to February 1999. The surface area of the deployed collector was 1.0 m2 . Samples for dust analyses were collected by washing the sampler using deionized water. Samples for radionuclide analyses were collected by washing the sampler with 6 N HCl. About 98 rainout events and about 49 monthly dust samples were collected. The rainwater samples were filtered using 0.7 µm pore size GFF filters immediately after sampling and acidified (pH < 2) prior to storage. 2.3. Seawater Seawater samples were largely collected using an on-board pump from the sea surface and using a 10 liter sized Niskin water bottle mounted on a Rosette sampler which was also equipped with a SBE CTD sensor. About 100 surface water samples were collected, and 6 vertical profiles were obtained during the summer stratification period. Water samples were subject to onboard filtration to remove suspended particulate matter. 2.4. Bottom sediments Bottom sediments were collected using a grab sampler, box corer or a multiple-corer. About 52 grab samples for radionuclide analyses and 29 cores for 210 Pb analyses were collected. The length of cores was mostly less than 40 cm, and the cores were sliced at 0.5–1 cm intervals, frozen on board and later freeze-dried in the laboratory. 2.5. Analytical methods The filtered water samples were acidified, chemical carriers, and tracers of 242 Pu and 85 Sr were added followed by coprecipitation of manganese oxide for Pu isotopes and ammonium molybdophosphate (AMP) for 137 Cs and oxalate for 90 Sr (Hong et al., 1999a). The sediment and soil samples were mixed with oxalic acid and ashed to remove the organic matter. Subsequent to this, the residues were digested with concentrated. HF, HNO3 and HCl after the addition of 242 Pu and 85 Sr yield tracers. The supernatants were later coprecipitated by iron hydroxide for Pu and as oxalate for 90 Sr (Hong et al., 1999c). 137 Cs was directly counted in the HPGe gamma-spectrometer (Hong et al., 1999c). Down core 210 Pb massic activities were determined by alpha-spectrometry of its progeny (210 Po), which was released from dried, homogenized sediments by successive digestion with concentrated HF, HNO3 and HCl, and then spontaneously deposited on silver discs. 209 Po yield tracer was added to each sample prior to acid digestion. 210 Pb-derived sediment accumulation and 210 Pb flux to the bottom sediment were calculated based on a one-dimensional, two-layer, steady state constant 210 Pb flux/constant sedimentation model in which mixing occurs only in the surface mixed layer (Hong et al., 1997). Sediment organic carbon contents of bottom sediments were determined using a CHNS-O elemental analyzer (Carlo Erba model 1108) after treatment with HCl to remove carbonates (Hong and Burrell, 1988).
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
99
3. Results 3.1. Atmospheric input 3.1.1. Total 239+240 Pu deposition The massic activities of 239+240 Pu in residues collected from May 1994 to February 1999 in the mid-eastern coast of the Yellow Sea, Ansan (37◦ 17 N, 126◦ 50 E) are given in Table 1. The annual deposition of dust was found to be 65 ± 7 g m−2 yr−1 , with generally high values in the winter and low values in the summer season. This value is similar to that estimated from the dust deposition flux over the Yellow Sea (Zhang et al., 1993; Gao et al., 1997). Long-range transport of resuspended soil particles originating from the Chinese deserts and arid regions were largely attributed to the temporal variations of dust deposits, as well as the associated Table 1 Temporal variation of the depositional fluxes of dust and total 239+240 Pu, 238 Pu, and 137 Cs concentrations in residue at the mid-eastern coast of the Yellow Sea (Ansan, KORDI campus, 37◦ 17 N, 126◦ 50 E) for the period of 1994 to 1999. All quoted uncertainties are 1 sigma standard deviations Year
Month
Dust mass (g m−2 month−1 )
239+240 Pu
(mBq m−2 month−1 )
(Bq kg−1 )
5 6 7 8 9 10 11 12
7.98 3.03 1.40 1.30 2.47 3.49 3.00 5.20
1.24 ± 0.14 0.25 ± 0.04 0.13 ± 0.03 0.09 ± 0.04 0.22 ± 0.04 0.61 ± 0.13 0.18 ± 0.06 0.26 ± 0.10
0.155 ± 0.018 0.082 ± 0.012 0.094 ± 0.018 0.070 ± 0.030 0.090 ± 0.017 0.174 ± 0.036 0.060 ± 0.020 0.050 ± 0.020
1995
1 2 3 4 5 7 8 10 11–12
3.43 5.99 8.63 11.56 3.26 3.89 2.74 3.70 14.77
0.27 ± 0.07 1.50 ± 0.18 0.43 ± 0.09 1.62 ± 0.15 0.23 ± 0.07 0.19 ± 0.04 0.04 ± 0.01 0.28 ± 0.04 2.13 ± 0.15
0.080 ± 0.020 0.250 ± 0.030 0.050 ± 0.010 0.140 ± 0.013 0.070 ± 0.020 0.050 ± 0.010 0.014 ± 0.005 0.075 ± 0.011 0.144 ± 0.010
1996
1–2 3 4 5 6 7 8 9–10 11 12
11.39 6.70 7.33 6.18 3.58 2.62 2.28 4.20 12.30 6.68
1.37 ± 0.07 0.92 ± 0.09 2.20 ± 0.15 1.01 ± 0.09 0.34 ± 0.04 0.05 ± 0.03
0.120 ± 0.006 0.137 ± 0.013 0.300 ± 0.020 0.164 ± 0.015 0.094 ± 0.011 0.020 ± 0.010
0.32 ± 0.05 0.75 ± 0.07 0.70 ± 0.09
0.077 ± 0.013 0.061 ± 0.006 0.105 ± 0.013
1994
238 Pu
137 Cs
(Bq kg−1 )
(Bq kg−1 ) 4.0 ± 0.4
3.0 ± 0.5
0.006 ± 0.003 0.003 ± 0.001 0.002 ± 0.001
100
G.H. Hong et al.
Table 1 (Continued) Dust mass (g m−2 month−1 )
239+240 Pu
(mBq m−2 month−1 )
(Bq kg−1 )
1 2 3 4 5–6 7 8 9 10–11 12
4.97 5.86 6.40 6.00 11.50 2.39 3.94 6.25 7.70 2.32
0.20 ± 0.04 0.50 ± 0.06 1.06 ± 0.11 0.73 ± 0.08 1.00 ± 0.10 0.03 ± 0.01 0.22 ± 0.04 0.07 ± 0.01 0.57 ± 0.06 0.27 ± 0.04
0.040 ± 0.008 0.086 ± 0.010 0.165 ± 0.017 0.122 ± 0.014 0.087 ± 0.009 0.012 ± 0.005 0.056 ± 0.009 0.011 ± 0.002 0.074 ± 0.008 0.116 ± 0.016
0.009 ± 0.003
1998
1 2 5 6 7 8 10 11 12
8.54 9.87 2.27 11.38 2.50 3.05 2.81 3.89 2.54
0.97 ± 0.03 2.69 ± 0.10 0.70 ± 0.02 0.24 ± 0.01 0.05 ± 0.01 0.05 ± 0.01 0.17 ± 0.01 0.12 ± 0.01 0.35 ± 0.01
0.114 ± 0.004 0.273 ± 0.010 0.309 ± 0.011 0.021 ± 0.001 0.018 ± 0.001 0.016 ± 0.001 0.060 ± 0.002 0.031 ± 0.001 0.137 ± 0.005
0.041 ± 0.004 0.091 ± 0.032
1999
1 2
5.39 7.69
1.07 ± 0.04 0.98 ± 0.04
0.198 ± 0.007 0.128 ± 0.005
0.180 ± 0.006 0.025 ± 0.001
Year
1997
Month
238 Pu (Bq kg−1 )
137 Cs (Bq kg−1 )
0.005 ± 0.002 0.003 ± 0.001
0.005 ± 0.002 0.008 ± 0.004
0.035 ± 0.001
radionuclides, since the Yellow Dust Storm events are the most salient features observed in the late winter and early spring in the region (Zhang et al., 1993; Gao et al., 1997). It may be reasonably assumed therefore, that our sampling station is a regional representative in the Yellow Sea area. The range of annual dry deposition rate for 239+240 Pu varied between 6.3 and 8.1 mBq m−2 yr−1 . Monthly 239+240 Pu deposition in Ansan varied from 0.03 ± 0.03 to 2.69 ± 0.10 mBq m−2 and it showed a typical seasonal variation with high values in winter–spring and low values in summer–autumn (Fig. 1). Annual maxima in 239+240 Pu deposition appear to occur in the spring, although total dust flux often peaked in the winter– spring period (Fig. 2). The spring peak is more clearly seen in 239+240 Pu massic activity in the dust, where the 239+240 Pu massic activity in residue is the sum of dry and wet deposition of 239+240 Pu divided by the total weight of solid material obtained during the sampling period (Fig. 3). The average 239+240 Pu massic activity in dust in Ansan was 0.10 ± 0.07 Bq kg−1 . The Meteorological Research Institute (MRI, Tsukuba, Japan) has been maintaining an extensive monitoring effort for the atmospheric deposition of artificial radionuclides in Tsukuba (36◦ 03 N, 140◦ 08 E), located at similar latitude as Ansan, but approximately 1300 km east of Ansan, Korea (e.g., Hirose et al., 2003). Comparison with MRI data provides some insight on the atmospheric 239+240 Pu deposition in the coastal region of the Yellow Sea. The monthly total deposition of 239+240 Pu in Ansan varied with time similarly to that of Tsukuba, however, those values in Ansan were higher in the winter–spring period and a little lower in summer
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
101
Fig. 1. Monthly total 239+240 Pu deposition observed during the period from 1994 to 1999 in Ansan, mid-eastern coast of the Yellow Sea. Error bars were not included here since they are less than 10% for most samples. ∗ denotes the composite samples. Monthly total 239+240 Pu deposition data for the corresponding period observed in Tsukuba, Japan (Hirose et al., 2003) were also included for comparison.
Fig. 2. Monthly dust deposition observed during the period from 1994 to 1999 in Ansan, mid-eastern coast of the Yellow Sea. ∗ denotes the composite samples. Monthly dry residue deposition data for the corresponding period observed in Tsukuba, Japan (Igarashi et al., 2003) were also included for comparison.
than in Tsukuba, although there is a large year-to-year variation (Fig. 1). The average monthly total deposition of 239+240 Pu in Ansan was 0.55 ± 0.55 mBq m−2 , higher than in Tsukuba by a factor of 1.8 ± 1.3 for the corresponding period of observation in Ansan. Similarly, the dust deposition flux in Ansan was also greater than that in Tsukuba by as much as a factor of 1.6±0.4, based on the Tsukuba observation in 1990s (Igarashi et al., 2003) and Ansan (Fig. 2). It is interesting to note that 239+240 Pu massic activities in residue in Ansan were very similar to those in Tsukuba (0.10 ± 0.06 Bq kg−1 , Fig. 3) despite of the large distance between the Ansan and Tsukuba. Therefore, it may be regarded that the total deposition of 239+240 Pu is largely dependent on the dust deposition rate. The dust deposited in the region appears to be largely derived from arid regions of the China, and supplied via the long range transport by prevailing westerlies, particularly during the Yellow Dust Storm seasons in the spring.
102
G.H. Hong et al.
Fig. 3. Monthly 239+240 Pu concentration in residue observed during the period from 1994 to 1999 in Ansan, mid-eastern coast of the Yellow Sea. Error bars were not included here since they are less than 10% for most samples. ∗ denotes the composite samples. Monthly 239+240 Pu concentrations in residue data for the corresponding period observed in Tsukuba, Japan (calculated from Hirose et al., 2003; Igarashi et al., 2003) were also included for comparison.
Although we did not determine the total deposition fluxes of 90 Sr and 137 Cs, it would be instructive to estimate those in order to construct a broad-brush budget of the artificial radionuclides in the Yellow Sea. Based on the average activity ratios of 137 Cs/90 Sr of 2.1 and 239+240 Pu/137 Cs of 0.018 in the Taklimakan Desert, where the Chinese nuclear weapons test site Lop Nor is located, and dust deposit samples in Tsukuba for the former (Igarashi et al., 2001, 2003) and dust deposit samples for the latter (Hirose et al., 2003; Igarashi et al., 2003), and assuming that the dust falling at the eastern coast of the Yellow Sea were similar to those at Tsukuba, the total deposition of 137 Cs and 90 Sr appears to be 350–450 mBq m−2 yr−1 and 167–214 mBq m−2 yr−1 , respectively. These estimates appear to be a little higher than those observed in Tsukuba for the 1990s (140–350 mBq m−2 yr−1 for 137 Cs and 70–180 mBq m−2 yr−1 for 90 Sr) (Igarashi et al., 1996, 2003). Although higher 239+240 Pu/137 Cs activity ratios (0.038–0.058) were observed for two samples in Ansan (Table 1), they were not considered here due to the insufficient number of data available. The activity ratio of 137 Cs/90 Sr in the total deposition residues in Japan is significantly different from its values in either Japanese soil (5.3, Igarashi et al., 2003) or Korean Soil (11 in Table 5 and Kim et al., 1998). The activity ratios among these radionuclides in 1990s for the global fallout were 0.024 and 1.6 for 239+240 Pu/137 Cs and 137 Cs/90 Sr, respectively (Ikeuchi et al., 1999). The artificial radionuclide contamination of soil in Korean Peninsula will be presented later. 3.1.2. Wet deposition of 239+240 Pu, 90 Sr, and 137 Cs The wet deposition of the artificial radionuclides was determined on individual rainout event basis since 1994, and our data shown here did not cover the entire rainout events (Table 2). 239+240 Pu activity concentrations in individual precipitations varied as much as 3 orders
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
103
Table 2 Temporal variation of 239+240 Pu, 90 Sr and 137 Cs concentration in the individual rainout events at the mid-eastern coast of the Yellow Sea (Ansan, KORDI campus, 37◦ 17 N, 126◦ 50 E) for the period of 1994–2000. All quoted uncertainties are 1 sigma standard deviations 239+240 Pu (µBq l−1 )
Date
Precipitation (mm)
1994-5-3 1994-5-14 1994-5-15 1994-6-30 1994-7-1 1994-8-1 1994-8-10 1994-8-28 1994-9-2 1994-10-12 1995-1-15 1995-3-10 1995-3-22 1995-4-10 1995-4-22 1995-5-13 1995-5-20 1995-6-3 1995-6-18 1995-7-1 1995-7-3 1995-7-8–10 1995-7-11 1995-7-13 1995-7-18 1995-8-1 1995-8-8 1995-8-19 1995-8-20 1995-8-24 1995-8-26 1996-3-16 1996-3-21 1996-3-29 1996-4-16 1996-4-17 1996-4-29 1996-6-17 1996-6-24 1996-6-27 1996-6-28 1996-7-4 1996-7-15 1996-7-21 1996-8-26 1996-8-27
29 40 22.5 43.5 24.3 8.4 80.7 92.1 6.2 104.3
2.51 ± 0.58 1.06 ± 0.34 0.71 ± 0.21 1.28 ± 0.41 1.17 ± 0.21 0.63 ± 0.24 0.34 ± 0.15 1.45 ± 0.55 1.88 ± 0.37 0.51 ± 0.18
4.3 25.6 8.3 15 9.8 21.2 22.1 8.6 44.1 4.3 151.5 18.8 48.2 48.7 10.6 180.6 227.3 49.5 91.8 69.1 25.5 16.8 35.4 5.5 9.4 32.2 153.6 24.9 8.5 5.8 56.6 44.2 19.5 33.5 32.4
1.08 ± 0.25 1.16 ± 0.26 1.03 ± 0.6 5.07 ± 0.61 78.13 ± 5.19 2.88 ± 0.41 4.58 ± 0.69 1.86 ± 0.62 0.56 ± 0.25 33.39 ± 3.53 0.41 ± 0.16 2.12 ± 0.35 1.48 ± 0.29 0.68 ± 0.18 1.02 ± 0.42 0.25 ± 0.13 0.05 ± 0.05 0.29 ± 0.07 0.30 ± 0.08 0.5 ± 0.13 5.83 ± 0.43 4.13 ± 0.34 1.42 ± 0.18 131 ± 3.56 131 ± 3.56 3.16 ± 0.29 0.4 ± 0.11 1.2 ± 0.26 0.83 ± 0.17 0.83 ± 0.17 0.29 ± 0.13 0.10 ± 0.06 0.18 ± 0.08 0.56 ± 0.13 0.56 ± 0.13
238 Pu (µBq l−1 )
90 Sr
(mBq l−1 )
137 Cs (mBq l−1 )
0.23 ± 0.05 0.09 ± 0.03 0.21 ± 0.02
0.46 ± 0.06 0.38 ± 0.10
4.17 ± 1.08 1.42 ± 0.38
0.42 ± 0.11 4.58 ± 0.45
1.22 ± 0.13 0.41 ± 0.11 0.81 ± 0.11 0.58 ± 0.15
0.26 ± 0.03 0.16 ± 0.02 0.17 ± 0.03 1.52 ± 0.04 0.15 ± 0.03 0.14 ± 0.02 0.19 ± 0.03
0.10 ± 0.02 0.10 ± 0.04
104
G.H. Hong et al.
Table 2 (Continued) Date
Precipitation (mm)
1997-3-6 1997-3-14 1997-4-2 1997-5-7 1997-5-12 1997-5-29 1997-5-30 1997-6-25 1997-7-2 1997-7-7 1997-8-4 1998-3-20 1998-4-22 1998-5-2 1998-5-11 1998-5-16 1998-6-2 1998-6-25 1998-7-3 1998-7-8 1998-8-1 1998-8-11 1998-8-14 1998-9-29 1998-10-12 1999-1-29 1999-3-18 1999-3-22 1999-4-1 1999-4-6 1999-5-7 1999-5-19 1999-5-25 1999-6-10 1999-6-21 2000-6-27 2000-6-29 1999-7-20 1999-7-28 1999-8-2 2000-7-13 2000-7-24 2000-8-10 2000-8-26 2000-8-31 2000-10-2 2001-8-9 2001-8-16 2001-11-1
18.2 11.3 28.3 21 109.1 25.9 20.9 104.2 150.4 62.5 138.3 37.9 19 35.5 33.5 24.6 35.2 91.5 86.9 33.5 42 235.8 88.7 69.8 29.1 30.5 35.7 17.4 8.3 58.5 41.8 20.3 20.4 22 20.8 23.8 34.1 92 245.6 8.4 1.2 18.1 137.1 49.8 101.9 21.4 5 5.4
239+240 Pu (µBq l−1 )
238 Pu
(µBq l−1 )
0.77 ± 0.25 0.75 ± 0.34 1.16 ± 0.32 0.91 ± 0.23 0.91 ± 0.23 0.83 ± 0.18 0.89 ± 0.24 0.24 ± 0.10 0.46 ± 0.13 0.25 ± 0.15 5.0 ± 0.5 15.9 ± 1.0 2.15 ± 0.13 1.46 ± 0.09 1.82 ± 0.11 0.93 ± 0.06 0.31 ± 0.02 0.45 ± 0.03 0.69 ± 0.04 0.52 ± 0.03 0.47 ± 0.03 0.59 ± 0.04 0.40 ± 0.03 1.34 ± 0.08 2.36 ± 0.15 12.82 ± 0.81 12.66 ± 0.80 13.77 ± 0.87 4.21 ± 0.26 5.93 ± 0.37 5.57 ± 0.35 2.58 ± 0.16 1.61 ± 0.10 0.94 ± 0.06 0.53 ± 0.03
0.25 ± 0.01 0.08 ± 0.01
2.17 ± 0.42
90 Sr
(mBq l−1 )
0.37 ± 0.07 0.45 ± 0.09 0.74 ± 0.16 0.37 ± 0.04 1.85 ± 0.25 0.67 ± 0.14 0.88 ± 0.10 0.37 ± 0.09 0.23 ± 0.08 0.45 ± 0.09 0.16 ± 0.04 0.28 ± 0.04 0.25 ± 0.07 0.18 ± 0.06 0.08 ± 0.08 0.13 ± 0.07 0.06 ± 0.10 0.14 ± 0.03 0.08 ± 0.02 0.06 ± 0.02 0.05 ± 0.01 0.04 ± 0.02 0.4 ± 0.02 0.07 ± 0.01 1.09 ± 0.33 0.14 ± 0.02 1.46 ± 0.45 3.69 ± 0.27 0.53 ± 0.08 2.15 ± 0.68 0.67 ± 0.13 1.28 ± 0.19 0.71 ± 0.09 0.41 ± 0.06 0.21 ± 0.03 0.23 ± 0.03 0.34 ± 0.05 1.27 ± 0.13 0.33 ± 0.04 0.29 ± 0.05 0.21 ± 0.04 0.20 ± 0.03 0.11 ± 0.04 0.15 ± 0.03 0.09 ± 0.03 0.55 ± 0.09 0.08 ± 0.02 0.28 ± 0.03
137 Cs
(mBq l−1 )
1.04 ± 0.31 0.37 ± 0.22 0.35 ± 0.18 0.15 ± 0.11 0.43 ± 0.31 0.27 ± 0.09 0.10 ± 0.05 0.23 ± 0.13 0.09 ± 0.08 0.08 ± 0.05 0.04 ± 0.03 0.11 ± 0.06 0.09 ± 0.08
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
105
Fig. 4. Temporal variation of dissolved 239+240 Pu (µBq l−1 ) in individual rainout events in Ansan, Korea.
Fig. 5. Temporal variation of dissolved 90 Sr (mBq l−1 ) in individual rainout events in Ansan, Korea.
of magnitude, from less than 0.1–131 µBq l−1 with higher values in spring (Fig. 4). The average activity concentration of filtered rainwater was 6.6 ± 23.6 µBq l−1 for 69 rainout events in Ansan during the period for 1994–1998. 90 Sr concentration in the individual precipitation varied as much as 2 orders of magnitude from less than 0.1 to 3.7 mBq l−1 with higher values in spring (Fig. 5). The average activity concentration of filtered rainwater was 0.4 ± 0.4 mBq l−1 for 61 rain samples in Ansan during the period for 1994–1998.
106
G.H. Hong et al.
Fig. 6. Temporal variation of dissolved 137 Cs (mBq l−1 ) in individual rainout events in Ansan, Korea.
137 Cs
activity concentration in the individual precipitation varied as much as 2 orders of magnitude from less than 0.1–1.0 mBq l−1 (Fig. 6). The highest value was observed in April 1998. The activity ratios of 239+240 Pu/137 Cs and 137 Cs/90 Sr in precipitation varied from less than 0.001–0.064 with an average of 0.007 ± 0.005 and from 0.1 to 7.2 with an average of 2.6 ± 1.9, respectively. Highest values of these activity ratios were observed in the spring 1998. Much higher 239+240 Pu/137 Cs activity ratios (0.023) were found in Monaco rain in the late 1970s and again in the early 2000s (Thein et al., 1980; Lee et al., 2003b), which could be influenced by discharges from nuclear reprocessing facilities in Europe. The average rainfall during the period of 1994–2000 was 1260 ± 240 mm yr−1 and more than 2/3 of the annual precipitation was concentrated in the summer monsoon period of June–September in Incheon, near Ansan (Korea Meteorological Agency). The average rainout events larger than 0.1 mm day−1 were 95 ± 8 days yr−1 during the period of 1994–2000. The nature of the manual sampling adopted in this study significantly limited the coverage of rain sampling. However, there was a statistically significant correlation between the amount of precipitation and 239+240 Pu content, 137 Cs and 239+240 Pu, and 90 Sr and 239+240 Pu in the precipitation, as shown in Figs. 7–9, respectively. Therefore, these correlations and the amount of daily precipitation for the sampling period of 1994–1999 were utilized to estimate the annual wet deposition of artificial radionuclides in the eastern Yellow Sea coast (Ansan). Based on these calculations, wet depositional flux would be 2.6 ± 0.2 mBq m−2 yr−1 for 239+240 Pu, 371 ± 266 mBq m−2 yr−1 for 137 Cs, and 142 ± 146 mBq m−2 yr−1 for 90 Sr. It appears that the dry deposition was significant for 239+240 Pu, as it has also been observed by Hirose et al. (2004), and the wet deposition was important for 137 Cs and 90 Sr. A more close examination of the removal processes for 137 Cs and 90 Sr from the air in the Yellow Sea region is needed.
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
107
Fig. 7. Ln[Precipitation, mm] vs. ln[239+240 Pu, µBq l−1 ] of the individual rain events observed in Ansan, Korea.
Fig. 8. Ln[239+240 Pu, µBq l−1 ] vs. 90 Sr (mBq l−1 ) of the individual rain events observed in Ansan, Korea.
3.2. Soil Since we have only one soil core, the data should be regarded as preliminary. The massic activities of 90 Sr, 137 Cs and 239+240 Pu in the surface layer of the soil core were 5.2, 59.8 and 1.20 Bq kg−1 dry weight, respectively (Table 3). Activity ratios of 137 Cs/90 Sr,
108
G.H. Hong et al.
Fig. 9. Ln[239+240 Pu, µBq l−1 ] vs. 137 Cs (mBq l−1 ) of the individual rain events observed in Ansan, Korea.
Table 3 210 Pb, 137 Cs, 90 Sr, 239+240 Pu and 238 Pu massic activities in the soil from the mid-western part of the Korean Peninsula (Kwangrung, 37◦ 44 N, 127◦ 00 E). All quoted uncertainties are 1 sigma standard deviations
Depth (cm)
Soil mass (g cm−2 )
Total 210 Pb (Bq kg−1 )
137 Cs (Bq kg−1 )
90 Sr
(Bq kg−1 )
239+240 Pu (Bq kg−1 )
238 Pu (Bq kg−1 )
5.18 ± 0.61 (0–2 cm)
1.20 ± 0.04 (0–2 cm)
0.029 ± 0.003 (0–2 cm)
0.808 ± 0.03 1.095 ± 0.04 (3–5 cm)
0.031 ± 0.005 0.029 ± 0.005 (3–5 cm)
0.99 ± 0.04 (6–8 cm)
0.031 ± 0.003 (6–8 cm)
1.09 ± 0.05
0.036 ± 0.005
0.86 ± 0.02 (10–12 cm)
0.023 ± 0.002 (10–12 cm)
0.301 ± 0.03
0.006 ± 0.001
0.119 ± 0.01
0.002 ± 0.001
0.07 ± 0.01 (17–19 cm)
0.004 ± 0.001 (17–19 cm)
0–1
0.57
263 ± 8
59.8 ± 5.3
1–2 2–3 3–4
1.17 1.80 2.46
223 ± 38 157 ± 29 192 ± 36
44.7 ± 4.0 39.8 ± 3.7 41.5 ± 3.7
4–5 5–6 6–7
3.13 3.84 4.57
144 ± 24 133 ± 22
37.0 ± 3.3 37.7 ± 4.3
7–8 8–9 9–10 10–11
5.33 6.12 6.93 7.77
121 ± 24 138 ± 19 85 ± 17
37.7 ± 3.3 40.7 ± 3.7 24.0 ± 2.2
11–12 12–13 13–14 14–15 15–16 16–17 17–18
8.63 9.52 10.44 11.38 12.35 13.34 14.37
58 ± 12
4.7 ± 0.5
18–19
15.42
68 ± 14
1.3 ± 0.2
3.29 ± 0.58 (6–8 cm)
2.80 ± 0.57 (10–12 cm)
2.29 ± 0.47 (17–19 cm)
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
109
239+240 Pu/137 Cs, and 238 Pu/239+240 Pu in the top 1 cm section were 11.5, 0.020 and 0.024, re-
spectively. These values fall within the range of a much larger published database (Kim et al., 1998; Lee et al., 1997). The soil inventory of 90 Sr, 137 Cs and 239+240 Pu, and 238 Pu was found to be >5108, 3100, 100 and 3.2 Bq m−2 , respectively. Our values are in the high end of the earlier reported values (Hölgye and Filgas, 1995; Lee et al., 1996). The vertical distributions of 239+240 Pu and 137 Cs in the soil column are similar, the highest massic activities are at the surface, and remain nearly constant to the depth of 12 cm (Fig. 10). Then their massic activities decrease significantly, and at around 20 cm below the surface they are below our detection limits. The depth distribution pattern for 90 Sr shows a smoother decrease with depth, probably due to its losses at the surface. 210 Pb massic activity was also measured in the soil core to elucidate the retention behavior of these fallout radionuclides in the soil, since the excess 210 Pb is also supplied from the atmosphere into the soil. The downcore distribution of excess 210 Pb is very similar to that of 137 Cs and 239+240 Pu (Fig. 10). The 210 Pb concentration decreases logarithmically with depth, and it approaches to the supported level at approximately 12 cm below the surface. Therefore, the penetration rate of these radionuclides is estimated to be larger than 0.2 cm yr−1 . The apparent down core logarithmic decrease of the excess 210 Pb suggested that an apparently undis-
Fig. 10. Soil depth profiles of 90 Sr, 137 Cs, 239+240 Pu, and 210 Pb at the mid-western part of the Korean Peninsula (Kwangrung Arboretum, 37◦ 44 N, 127◦ E).
110
G.H. Hong et al.
turbed soil core sample was retrieved from the Kwangrung National Arboretum. Atmospheric 210 Pb flux was estimated by multiplying the decay constant of 210 Pb and the excess 210 Pb inventory (Nozaki et al., 1978). The estimated atmospheric 210 Pb flux calculated from the soil profile (Fig. 1) is 20 mBq cm−2 yr−1 , which is about 2/3 of the regional atmospheric 210 Pb flux (Nozaki et al., 1973). The denudation rate in this drainage basin, estimated using the total river discharge and suspended matter concentration, yields 0.047 g cm−2 yr−1 (Hong et al., 2002). If we assume the soil is eroded from the top of the soil, then, the annual 210 Pb removal flux would be 12 mBq cm−2 yr−1 , that could account for the remaining 1/3 of the regional atmospheric 210 Pb flux, although this estimate is based on only one soil core. However, this estimate will result unusually shorter drainage residence time for 210 Pb. For the sake of further discussion, we assume that the soil inventories of artificial radionuclides may be underestimated by as much as 1/3 of the total inventory in the region. 3.3. River input A few data were collected for the dissolved concentrations of 90 Sr, 137 Cs and 239+240 Pu of the major Korean rivers flowing into the Yellow Sea in 1999 and 2001. The range of values reported for the activity concentrations of 90 Sr, 137 Cs and 239+240 Pu were 2.8–4.3, 0.25–0.68 mBq l−1 , and 2.7–3.1 µBq l−1 , respectively (Table 4). Similar values for 90 Sr and 137 Cs were also observed in Ebro River in Spain (Pujol and Sanchez-Cabeza, 2000), but lower than in other measurements in North America (Cornet et al., 1995). Although the available data set is rather limited, a preliminary estimation on the riverine input of artificial radionuclides from the Korean Peninsula was carried out using the river discharge rate and the average dissolved activity concentrations and surface soil concentrations of artificial radionuclides (Table 5). The sum of the discharge from major rivers in South Korea (Han, Keum, Youngsan, Anseoung, Sapkyo, Mankyoung, and Dongjin) into the Yellow Sea is estimated to be 32.9 × 109 m3 yr−1 of water and 20.1 × 106 t yr−1 of suspended particulate matter. Furthermore, if we include all the major rivers from the China (Huanghe, Changjiang, and Aprock), then 217.7 × 109 m3 yr−1 of water and 504.9 × 106 t yr−1 of suspended particulate matter Table 4 Activity concentrations of artificial radionuclides in rivers discharging into the Yellow Sea. All quoted uncertainties are 1 sigma standard deviations River
Sampling date
River discharge∗ (m3 s−1 )
90 Sr
137 Cs (mBq l−1 )
Han
January 1999 April 1999 June 1999 September 1999 December 1999 June 2001 June 1999 January 2001
377.4 2,405.5 1,749.2 418.4 99.6 353.4 884.9 18.3
3.39 ± 0.10 2.31 ± 0.07 3.03 ± 0.11 3.04 ± 0.09 2.85 ± 0.07 4.33 ± 0.14 3.00 ± 0.07 3.20 ± 0.09
0.46 ± 0.16 0.68 ± 0.13
Keum Youngsan
(mBq l−1 )
239+240 Pu (µBq l−1 )
2.70 ± 0.16 3.07 ± 0.18
0.25 ± 0.13 0.32 ± 0.10
∗ Han River Flood Control Service, Ministry of Construction and Transportation, Republic of Korea.
2.94 ± 0.17 3.04 ± 0.18
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
111
Table 5 Riverine discharge of artificial radionuclides into the Yellow Sea from the south Korean Peninsula Riverine flux (S. Korean rivers)
90 Sr (109 Bq yr−1 )
137 Cs (109 Bq yr−1 )
239+240 Pu (109 Bq yr−1 )
Dissolved flux Particulate flux Total
103 104 208
60 1202 1262
0.10 24.1 24.2
(Hong et al., 2002) would additionally be flowing in to the Yellow Sea. The contribution from the Changjiang was taken to be 14% of the total discharge as estimated by Liu et al. (2003). The large excess of 239+240 Pu over the global fallout in the Yellow Sea was also attributed to the Changjiang River by Nagaya and Nakamura (1992). Most of the 239+240 Pu in river water is likely associated with particulate matter. For example, in the Rhone River, France, ∼80% of plutonium isotopes were found to be associated with the suspended material, and ∼20% was found to be exported as dissolved species (Eyrolle et al., 2004). 3.4. Seawater The dissolved activity concentrations of 90 Sr, 137 Cs and 239+240 Pu in the southeastern part of the Yellow Sea in the period of 1994–2000 are reported in Table 6. The average surface activity concentrations of 90 Sr, 137 Cs and 239+240 Pu in the Yellow Sea in 1999–2000 were found to be 2.1 ± 0.2, 2.7 ± 0.2 mBq kg−1 , and 4.2 ± 1.2 µBq kg−1 , respectively. The Yellow Sea is being fed from the East China Sea including a branch of Kuroshio Current originating from the western North Pacific subtropical gyre, and it is diluted with freshwater discharges from rivers in summer. Therefore, salinity in surface water of the Yellow Sea is highest in winter and lowest in autumn (Figs. 11–14). The temporal change in water characteristics was also reflected in the activity concentration of these artificial radionuclides. The average activity concentration of 90 Sr was higher in winter–spring (2.1 ± 0.4 and 2.4 ± 0.7 mBq kg−1 in February 1999 and April 2000, respectively) than summer–autumn (1.9 ± 0.3 and 2.0 ± 0.3 mBq kg−1 in August 1999 and September 2000, respectively). We observed a weak correlation of 90 Sr activity concentration and salinity (90 Sr (mBq kg−1 ) = 12.655–0.324 × Salinity, r = −0.523, P > 0.01), which may suggest that river waters are the main source of 90 Sr flowing into the Yellow Sea. The activity ratios of 137 Cs/90 Sr in the Yellow Sea varied from 1.1 to 1.5, showing lower values than that of the global fallout (1.6), which further suggest continuing terrestrial sources. However, there were no discernible relationship for 137 Cs and 239+240 Pu with respect to salinity in the Yellow Sea. The water column in the Yellow Sea is vertically well mixed during winter and stratified in summer. In general, 90 Sr activity concentrations were generally higher in surface water than in bottom water in winter (February 1999, Table 6) and 137 Cs and 239+240 Pu activity concentrations were lower in the surface water than in the bottom water (February 1999 and April 2000, Table 6), although the Yellow Sea is well mixed in winter and early spring period (Hong et al., 1999b). In summer, 239+240 Pu activity concentrations increased with depth more
112
Table 6 Activity concentrations of dissolved 90 Sr, 137 Cs, 239+240 Pu in the Yellow Sea (1994–2000). All quoted uncertainties are 1 sigma standard deviations Date
02-Jul-94 02-Jul-94 12-Dec-93 12-Dec-93 12-Dec-93 02-Jul-94 01-Jul-94 01-Jul-94 29-Jun-94
February 24– March 2, 1999
YS9406 S9406 CO9312 CO9312 CO9312 YS9406 YS9406 YS9406 YS9406
Stn.
Coordinates
Water depth (m)
Sampling depth (m)
90 Sr
Temp. (◦ C)
Sal.
0 0 0 0 0 0 0 0 0
22.28 20.47 13.11 14.33 17.71 21.28 19.88 22.70 20.84
32.70 32.42 32.87 33.11 34.27 32.57 32.32 31.87 31.11
3.29 ± 0.04 3.79 ± 0.06
(mBq kg−1 )
137 Cs (mBq kg−1 )
239+240 Pu (µBq kg−1 )
3.06 ± 0.25 2.93 ± 0.21 2.74 ± 0.20 2.74 ± 0.18 3.21 ± 0.21 3.78 ± 0.22 2.65 ± 0.22 2.72 ± 0.20 2.21 ± 0.16
2.3 ± 0.4 5.2 ± 0.6 8.1 ± 0.6 6.5 ± 0.6 12.7 ± 1.1 5.8 ± 0.6 5.3 ± 0.8 5.3 ± 0.6 4.1 ± 0.6
Latitude
Longitude
A3 A5 D2 D3 D5 B4 C1 C5 D3
33◦ 52.42 N 35◦ 07.45 N 34◦ 10.00 N 34◦ 00.00 N 33◦ 40.00 N 34◦ 30.53 N 33◦ 30.08 N 33◦ 59.94 N 33◦ 30.18 N
124◦ 00.00 E 125◦ 00.16 E 126◦ 00.00 E 126◦ 00.00 E 126◦ 00.00 E 124◦ 30.01 E 122◦ 30.18 E 124◦ 30.01 E 125◦ 59.90 E
R 3 4 5 6 8 11 12
37◦ 22 N 37◦ 19 N 37◦ 19 N 37◦ 17 N 37◦ 12 N 37◦ 09 N 37◦ 15 N 37◦ 18 N
126◦ 14 E 125◦ 55 E 125◦ 59 E 125◦ 58 E 125◦ 47 E 125◦ 57 E 126◦ 01 E 126◦ 05 E
20 34 9 10 46 35 43 33
0 0 0 0 0 0 0 0
19.6 16.9 16.96 17.1 17.7 17.2 17.2 17.5
31.55 31.68 31.74 31.75 31.92 31.77 31.75 31.85
2.61 ± 0.07 2.66 ± 0.08 2.47 ± 0.06 2.68 ± 0.15 2.76 ± 0.11 2.47 ± 0.06 2.47 ± 0.05
2.85 ± 0.11 2.65 ± 0.12 2.62 ± 0.18 2.81 ± 0.18 2.84 ± 0.10 2.71 ± 0.11 2.94 ± 0.16 2.57 ± 0.17
B1 B 1 B2
36◦ 54.9 N 36◦ 44.1 N 37◦ 07 N
126◦ 00 E 126◦ 00 E 125◦ 30 E
63 64 43
B 2 B3 B4 B5
36◦ 24.7 N 37◦ 00 N 36◦ 59.9 N 37◦ 0.12 N
125◦ 0.28 E 125◦ 00 E 124◦ 30.1 E 124◦ 0.35 E
35 54 80 80
C1 C 1
36◦ 00 N 35◦ 29.7 N
126◦ 27 E 124◦ 00 E
19 82
0 0 0 38 0 0 0 0 75 0 0
4.22 5.11 3.50 3.55 5.72 6.24 7.35 7.88 7.89 5.30 8.67
31.87 32.06 31.88 31.89 32.17 32.50 32.38 32.40 32.39 32.14 32.53
2.22 ± 0.07 2.46 ± 0.06 3.00 ± 0.08 1.79 ± 0.06 2.01 ± 0.08 2.39 ± 0.05 2.37 ± 0.03 2.60 ± 0.10 1.89 ± 0.04 2.08 ± 0.07 1.97 ± 0.12
2.29 ± 0.18 2.50 ± 0.23 2.36 ± 0.18 3.18 ± 0.29 2.36 ± 0.22 2.94 ± 0.24 2.75 ± 0.22 2.79 ± 0.18 3.12 ± 0.22 3.01 ± 0.21 2.39 ± 0.14
2.93 ± 0.04 3.37 ± 0.09 2.89 ± 0.05 2.74 ± 0.06
G.H. Hong et al.
July 4–6, 1995
Cruise
Table 6 (Continued) Date
Stn.
Coordinates Latitude
Longitude
Water depth (m)
C2 C3
36◦ 00 N 36◦ 00 N
126◦ 17.5 E 126◦ 00 E
27 43
C4 C5 C6
35◦ 59.9 N 36◦ 00 N 36◦ 00 N
125◦ 30.1 E 124◦ 59.7 E 124◦ 30.1 E
55 80 85
C7 D5 E2
36◦ N 35◦ 0.36 N 34◦ N
124◦ E 124◦ E 126◦ E
77 84 80
E3 E4 E5 E6 F1
34◦ N 33◦ 59.8 N 34◦ N 34◦ N 33◦ 35 N
125◦ 31.05 E 124◦ 59.6 E 124◦ 30 E 124◦ E 125◦ E
64 98 82 81 89
F2
33◦ 0.2 N
124◦ 15.1 E
65
A2 B1 B3
37◦ 25 N 36◦ 44.1 N 37◦ N
26◦ 10 E 126◦ 00 E 125◦ E
23 75 54
B5
37◦ N
124◦ E
80
90 Sr (mBq kg−1 )
137 Cs
239+240 Pu
(mBq kg−1 )
(µBq kg−1 )
32.19 32.37 32.37 32.45 32.52 32.48 32.51 32.51 32.55 32.69 34.05 32.64 32.67 32.81 33.04 33.92 34.00 33.67
2.19 ± 0.12 2.31 ± 0.08 1.94 ± 0.11 2.58 ± 0.18 1.50 ± 0.13 2.54 ± 0.13 2.04 ± 0.07 2.50 ± 0.07 2.40 ± 0.09 1.88 ± 0.05 1.56 ± 0.06 1.98 ± 0.06 2.01 ± 0.09 1.95 ± 0.06 1.47 ± 0.13 1.93 ± 0.09 1.88 ± 0.10 1.55 ± 0.08
2.98 ± 0.21 2.65 ± 0.19 2.62 ± 0.37 2.52 ± 0.19 2.65 ± 0.21 2.94 ± 0.23 2.47 ± 0.23 2.64 ± 0.22 3.01 ± 0.20 2.84 ± 0.18 2.50 ± 0.23 2.83 ± 0.26 2.41 ± 0.18 3.38 ± 0.22 2.11 ± 0.22 2.69 ± 0.29 2.67 ± 0.23 1.78 ± 0.24
27.13 31.16 31.33 31.34 31.82 32.14 31.93 32.45 32.40 32.40
1.98 ± 0.06 1.50 ± 0.07 2.10 ± 0.08 1.52 ± 0.06 1.72 ± 0.05 1.70 ± 0.05 2.24 ± 0.10 1.86 ± 0.06 2.22 ± 0.09 2.68 ± 0.10
2.25 ± 0.20 2.42 ± 0.21 2.88 ± 0.28 3.13 ± 0.30 2.08 ± 0.16 2.95 ± 0.27 2.14 ± 0.20 3.41 ± 0.39 3.34 ± 0.52 2.73 ± 0.16
Sampling depth (m)
Temp. (◦ C)
Sal.
0 0 39 0 0 0 80 0 0 0 75 0 0 0 0 0 84 0
5.74 6.41 6.33 6.34 7.12 7.59 7.57 8.30 8.45 8.71 12.65 8.32 9.36 10.14 10.51 12.35 12.62 11.98
0 0 0 10 30 49 0 20 40 75
22.84 20.39 24.90 22.39 16.95 11.70 26.14 17.39 9.25 9.25
2.61 ± 1.02 2.65 ± 0.79 4.18 ± 1.66 1.25 ± 0.89 5.52 ± 1.77 3.04 ± 0.63 5.76 ± 1.18 5.64 ± 0.80 3.46 ± 1.04
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
August 7–17, 1999
Cruise
113
114
Table 6 (Continued) Date
Stn.
137 Cs (mBq kg−1 )
239+240 Pu (µBq kg−1 )
30.23 31.57 31.64 31.51 32.48 32.07 32.16 32.66 33.00 31.66 31.66 32.54 32.54 31.72 32.06 32.56 32.55 30.53
2.11 ± 0.09 1.66 ± 0.05 1.62 ± 0.05 1.83 ± 0.05 1.80 ± 0.05 1.57 ± 0.08 2.03 ± 0.07 1.75 ± 0.07 2.13 ± 0.06 1.80 ± 0.06 1.77 ± 0.06 1.72 ± 0.07 1.94 ± 0.07 1.18 ± 0.03 1.58 ± 0.06 1.67 ± 0.08 2.12 ± 0.08 2.09 ± 0.09
2.69 ± 0.16 2.84 ± 0.21 2.49 ± 0.22 2.86 ± 0.25 2.98 ± 0.17 2.48 ± 0.20 2.89 ± 0.25 2.92 ± 0.22 2.72 ± 0.19 2.06 ± 0.15 2.70 ± 0.39 2.98 ± 0.29 3.40 ± 0.42 1.90 ± 0.16 2.40 ± 0.31 2.18 ± 0.22 2.96 ± 0.19 2.26 ± 0.20
2.03 ± 0.56 3.13 ± 1.41 2.06 ± 0.89 7.40 ± 2.40 3.61 ± 0.65 2.48 ± 0.44 3.84 ± 0.56 2.36 ± 0.47 6.04 ± 1.01 3.70 ± 1.16 3.59 ± 1.25 4.33 ± 0.66 11.29 ± 3.48 3.39 ± 1.37 2.50 ± 0.91 6.90 ± 1.20 7.17 ± 1.17 3.58 ± 0.78
31.62 32.05 32.36 31.30 31.92 32.46 32.47 32.13 32.13 32.13
2.04 ± 0.35 3.59 ± 0.39 2.02 ± 0.29 3.55 ± 0.50 2.34 ± 0.31 2.63 ± 0.17 2.01 ± 0.23 2.78 ± 0.14 2.65 ± 0.16 1.84 ± 0.23
2.12 ± 0.18 2.06 ± 0.12 2.33 ± 0.16 2.44 ± 0.22 2.25 ± 0.15 1.62 ± 0.09 2.70 ± 0.23 1.90 ± 0.20 2.27 ± 0.26 2.59 ± 0.15
2.75 ± 0.43 2.33 ± 0.52 4.02 ± 042 2.67 ± 0.43 3.41 ± 0.54 5.43 ± 0.63 3.62 ± 0.35 8.45 ± 0.96 7.99 ± 0.98
Sampling depth (m)
Temp. (◦ C)
C1 C5
36◦ N 36◦ N
126◦ 27 E 124◦ 30.1 E
19 80
C7
36◦ N
124◦ E
77
D4
35◦ N
124◦ 30 E
92
E4
34◦ N
124◦ E
98
F1
33◦ 35 N
125◦ E
89
0 0 20 40 75 0 20 40 70 0 20 50 87 0 20 75 93 0
24.54 25.83 24.52 6.55 8.49 26.17 24.52 10.23 10.10 25.06 25.03 8.78 8.74 24.84 22.55 8.98 8.93 25.38
B 1
36◦ 44.1 N 37◦ N 36◦ 59.9 N 36◦ N 36◦ N 36◦ N 36◦ N 35◦ N
126◦ 00 E 125◦ E 124◦ 30.1 E 126◦ 27 E 126◦ E 124◦ 30.1 E 124◦ E 125◦ 43 E
64 54 80 19 38 85 76 21
0 0 0 0 0 0 0 0 10 20
6.22 7.12 7.89 10.38 7.59 8.45 8.45 8.40 8.37 8.16
B3 B4 C1 C3 C6 C7 D1
Coordinates
(mBq kg−1 )
Longitude
Water depth (m)
Sal.
Latitude
90 Sr
G.H. Hong et al.
April 10–22, 2000
Cruise
Table 6 (Continued) Date
Cruise
Stn.
Coordinates Longitude
D2
35◦ N
125◦ 30 E
86
D3
35◦ N
125◦ E
87
D4
35◦ N
124◦ 30 E
92
D5
35◦ 0.36 N
124◦ E
84
124◦ E 124◦ 15.1 E 124◦ 15.1 E
81 65 0
E6 34◦ N F2 33◦ 0.2 N DEACHUNG-DO 33◦ 0.2 N September 14–27, 2000
B1 B3 B5
36◦ 54.9 N 37◦ N 37◦ 0.12 N
126◦ E 125◦ E 124◦ 0.35 E
63 54 80
C1 C5 C7
36◦ N 36◦ N 36◦ N
126◦ 27 E 124◦ 59.7 E 124◦ E
19 80 77
D1 D5 E3 E4
35◦ N 35◦ 0.36 N 34◦ N 33◦ 59.8 N
125◦ 43 E 124◦ E 125◦ 31.05 E 124◦ 59.6 E
74 84 64 98
E6
34◦ N
124◦ E
81
90 Sr
137 Cs
239+240 Pu
(mBq kg−1 )
(mBq kg−1 )
(µBq kg−1 )
32.15 32.24 32.34 32.60 − 32.60 32.59 32.57 32.57 32.75 32.43
2.81 ± 0.13 2.21 ± 0.22 1.95 ± 0.22 1.83 ± 0.05 1.73 ± 0.15 1.72 ± 0.04 2.54 ± 0.13 1.61 ± 0.12 1.59 ± 0.12 3.96 ± 0.35 2.52 ± 0.12 2.73 ± 0.09
1.96 ± 0.14 2.69 ± 0.18 2.34 ± 0.13 5.15 ± 0.25 2.08 ± 0.12 2.57 ± 0.15 2.28 ± 0.11 2.95 ± 0.18 2.23 ± 0.13 1.95 ± 0.12 1.95 ± 0.12 2.53 ± 0.23
5.76 ± 2.06 4.88 ± 0.70 6.20 ± 0.66 5.17 ± 0.64 4.09 ± 0.60 3.61 ± 0.68 6.02 ± 0.86 4.53 ± 1.15 8.13 ± 1.42 4.11 ± 0.68 2.61 ± 0.31
30.85 30.52 31.82 32.31 24.69 31.22 31.87 32.41 31.47 31.46 31.90 32.00 32.90 32.05
2.16 ± 0.08 2.10 ± 0.05 2.38 ± 0.07 2.09 ± 0.08 1.68 ± 0.09 1.77 ± 0.07 1.92 ± 0.09 2.13 ± 0.08 1.97 ± 0.09 2.53 ± 0.09 1.67 ± 0.09 1.90 ± 0.08 1.48 ± 0.10 2.03 ± 0.09
3.36 ± 0.16 3.03 ± 0.13 2.77 ± 0.12 3.50 ± 0.22 2.27 ± 0.14 3.08 ± 0.09 2.94 ± 0.07 4.03 ± 0.17 2.65 ± 0.15 2.43 ± 0.12 1.35 ± 0.08 2.65 ± 0.13 2.87 ± 0.09 2.70 ± 0.14
2.33 ± 0.34 2.78 ± 0.72 4.04 ± 0.59
Sampling Temp. depth (◦ C) (m)
Sal.
0 20 40 0 82 0 87 0 79 0 0 0
7.66 7.52 6.98 8.31 − 8.49 7.59 8.66 8.66 9.52 10.94
0 0 0 75 0 0 0 72 0 0 0 0 0 0
21.05 21.08 21.61 8.26 21.74 21.01 21.15 7.21 20.47 20.66 23.16 22.99 11.39 22.50
1.76 ± 0.33 3.46 ± 0.55 2.78 ± 0.43 5.23 ± 0.96 3.91 ± 0.65 2.55 ± 0.30 2.04 ± 0.32 2.21 ± 0.32
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
Latitude
Water depth (m)
2.06 ± 0.39 115
116
G.H. Hong et al.
Fig. 11. Distribution of temperature (T , ◦ C), salinity, 90 Sr (mBq kg−1 ) and 137 Cs (mBq kg−1 ) in the surface water of the Yellow Sea in February 1999.
than 137 Cs, while 90 Sr were stable, especially below the thermocline (Fig. 15), probably due to the particle scavenging in the surface and subsequent sinking and regeneration in the bottom layer of the sea. These nutrient-like features and similar 239+240 Pu activity concentrations at corresponding depths were found in the nearby East Sea (Sea of Japan) as well (Lee et al., 2003a). In general, spatial distribution patterns of radionuclides in the Yellow Sea were not pronounced for most of the seasons probably due to the presence of many rivers at various locations along the entire coast of the sea and active mixing of the water column due to its shallow depth (<100 m). However, the activity concentrations of 90 Sr, 137 Cs and 239+240 Pu tended
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
117
Fig. 12. Distribution of temperature (T , ◦ C), salinity, 90 Sr (mBq kg−1 ), 137 Cs (mBq kg−1 ) and 239+240 Pu (µBq kg−1 ) in the surface water of the Yellow Sea in August 1999.
118
G.H. Hong et al.
Fig. 13. Distribution of temperature (T , ◦ C), salinity, 90 Sr (mBq kg−1 ), 137 Cs (mBq kg−1 ) and 239+240 Pu (µBq kg−1 ) in the surface water of the Yellow Sea in April 2000.
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
119
Fig. 14. Distribution of temperature (T , ◦ C), salinity, 90 Sr (mBq kg−1 ), 137 Cs (mBq kg−1 ) and 239+240 Pu (µBq kg−1 ) in the surface water of the Yellow Sea in September 2000.
120
G.H. Hong et al.
Fig. 15. Vertical distribution of water temperature (◦ C), salinity, 90 Sr (mBq kg−1 ), 137 Cs (mBq kg−1 ) and 239+240 Pu (µBq kg−1 ) in the central part of the Yellow Sea (35◦ N, 124◦ 30 E) in August 1999.
to be higher in the lower salinity regions, which are the northwest and northeast corners of our sampling stations in summer (August 1999 and September 2000, Figs. 12 and 14). The 137 Cs and 239+240 Pu activity concentrations tended to be higher also in the trough traversing from northwest to the southeast direction in winter (February 1999 and April 2000, Figs. 11 and 3). The total estimated inventories of dissolved 90 Sr, 137 Cs and 239+240 Pu, based on the average activity concentrations in the water column and the water volume of the sea, were for the entire Yellow Sea in the years of 1999 and 2000, 35 ± 3, 45 ± 3 and 0.07 ± 0.02 GBq, respectively. 3.4.1. Temporal changes in radionuclide concentrations in the Yellow Sea Although water characteristics are changing geographically, seasonally and inter-annually due to variations of river discharges and inflow of the Pacific water, the temporal variations of 90 Sr and 137 Cs in the surface water of the Yellow Sea could be plotted against time using the historical data reported by Li et al. (1994), Zhu et al. (1991), Nagaya and Nakamura (1992), and Kim et al. (1997) (Fig. 16). The surface activity concentrations of these radionuclides decreased gradually from 1960 to ca.1994, and remain to be constant since 1994. The year marks of 1960 and 1994 are operationally defined here by virtue of sampling. However, activity concentrations of these radionuclides continue to decrease with time to the present in the Pacific Ocean (Povinec et al., 2005, and references therein). The apparent half-residence times of 90 Sr and 137 Cs in the Yellow Sea were 6.9 and 7.0 yr, respectively, based on the corresponding regression line of their exponential decrease curves for the period of 1960 to 1994. The apparent half-residence times of 90 Sr and 137 Cs in surface waters are influenced by oceanographic processes and radioactive decay (Hirose and Aoyama, 2003), These values are about a half of those found in the open Pacific Ocean, but similar to those in the North Sea (IAEA, 2005). 3.5. Bottom sediment The massic activities of 90 Sr, 137 Cs, 239+240 Pu measured in surface sediments in the Yellow Sea were <0.5–3.7, <0.2–11.2, <0.04–0.91 Bq kg−1 , respectively (Table 7). The average
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
121
Fig. 16. Temporal variation of 90 Sr and 137 Cs activity concentrations (mBq l−1 ) in the surface of the Yellow Sea for the last four decades. Correlation coefficient was calculated for the period of 1960–1994.
Table 7 90 Sr, 137 Cs and 239+240 Pu massic activities in surface bottom sediment of the Yellow Sea for the period of
1995–1999. All quoted uncertainties are 1 sigma standard deviations Location
90 Sr
137 Cs
239+240 Pu
238 Pu
Latitude
Longitude
(Bq kg−1 )
(Bq kg−1 )
(Bq kg−1 )
(Bq kg−1 )
Y9503-01 Y9503-03 Y9503-04 Y9503-05 Y9503-07 Y9503-08 Y9503-12 Y9503-13 Y9503-15 Y9503-32 Y9503-36
34◦ N 33◦ N 32◦ 30 N 32◦ N 32◦ 12 N 32◦ 54 N 36◦ 37 N 36◦ 13 N 36◦ 50 N 36◦ 10 N 35◦ 30 N
126◦ 30 E 125◦ 18 E 124◦ 42 E 124◦ 06 E 123◦ 12 E 122◦ 54 E 122◦ 30 E 123◦ E 124◦ 10 E 124◦ 30 E 125◦ E
0.47 ± 0.81 1.53 ± 0.62 2.37 ± 0.29 0.62 ± 0.33 0.00 ± 0.00 1.51 ± 0.52 11.18 ± 1.39 6.55 ± 0.79 1.60 ± 0.33 2.41 ± 0.65 2.59 ± 0.72
0.373 ± 0.038 0.199 ± 0.017 0.223 ± 0.016 0.224 ± 0.021 0.200 ± 0.019 0.186 ± 0.018 0.397 ± 0.036 0.232 ± 0.023 0.472 ± 0.029 0.045 ± 0.008 0.306 ± 0.022
0.009 ± 0.005 0.010 ± 0.003 0.007 ± 0.002
Y9605-0101 Y9605-0102 Y9605-0103 Y9605-0104 Y9605-0105 Y9605-0203 Y9605-0204 Y9605-0205 Y9605-0301
34◦ N 34◦ N 34◦ N 34◦ N 34◦ N 35◦ N 35◦ N 35◦ N 36◦ N
126◦ E 125◦ E 124◦ E 123◦ E 122◦ E 124◦ E 123◦ E 122◦ E 126◦ E
0.69 ± 0.40 1.33 ± 0.40 2.12 ± 0.51 2.99 ± 0.72 1.25 ± 0.47 2.09 ± 0.51 4.24 ± 1.03 2.21 ± 0.55 0.36 ± 0.27
0.360 ± 0.022 0.293 ± 0.033 0.411 ± 0.064 0.178 ± 0.015 0.054 ± 0.006 0.383 ± 0.024 0.556 ± 0.032 0.120 ± 0.014 0.101 ± 0.008
Y9608-42 Y9608-47 Y9608-49 Y9608-50 Y9608-51 Y9608-52 Y9608-53
32◦ N 36◦ N 36◦ N 37◦ N 37◦ N 37◦ N 37◦ N
123◦ E 124◦ E 122◦ 10 E 123◦ 11 E 124◦ E 125◦ E 126◦ E
0.16 ± 0.29 8.06 ± 1.02 3.55 ± 0.62 4.06 ± 0.71 2.08 ± 0.36 1.13 ± 0.28 0.91 ± 0.39
0.291 ± 0.018 0.720 ± 0.041 0.174 ± 0.021 0.303 ± 0.020 0.332 ± 0.023 0.200 ± 0.011 0.099 ± 0.007
Stn.
0.005 ± 0.001 0.020 ± 0.006 0.020 ± 0.004 0.010 ± 0.003 0.007 ± 0.002 0.015 ± 0.007 0.047 ± 0.020 0.010 ± 0.003 0.016 ± 0.001 0.015 ± 0.004
SOC (%)
0.20 0.32 0.29 0.01 0.36 0.93 0.92 0.18 0.60 0.13 0.60 0.76 0.16 0.48 0.91 0.84 0.01 0.01
0.021 ± 0.005 0.008 ± 0.003 0.006 ± 0.002 0.005 ± 0.001 0.004 ± 0.001
0.01 0.89 0.52 0.52 0.20 0.01 0.04
122
G.H. Hong et al.
Table 7 (Continued) Location
90 Sr
137 Cs
239+240 Pu
238 Pu
Latitude
Longitude
(Bq kg−1 )
(Bq kg−1 )
(Bq kg−1 )
(Bq kg−1 )
Y9704-01 Y9704-04 Y9704-10 Y9704-11 Y9704-15 Y9704-18 Y9704-19
33◦ 30 N 33◦ 10 N 35◦ 25 N 35◦ 25 N 36◦ 15 N 36◦ 50 N 36◦ 50 N
126◦ E 124◦ 30 E 123◦ 10 E 124◦ 10 E 125◦ 10 E 124◦ 10 E 125◦ 10 E
0.901 ± 0.055 0.231 ± 0.029 0.702 ± 0.045 0.562 ± 0.041 0.328 ± 0.021 0.336 ± 0.020 0.150 ± 0.010
0.010 ± 0.004
Y9902-B2 Y9902-B3 Y9902-B5 Y9902-C3 Y9902-C5 Y9902-O1 Y9902-D4 Y9902-F1 Y9902-F2 Y9902-E2
37◦ 7 N 37◦ N 37◦ 0.12 N 36◦ N 36◦ N 35◦ 27 N 35◦ N 33◦ 35 N 33◦ 0.2 N 34◦ N
125◦ 30 E 125◦ E 125◦ 30 E 126◦ E 124◦ 59.7 E 124◦ 38 E 124◦ 30 E 125◦ E 124◦ 15.1 E 126◦ E
YG9507-R YG9507-3 YG9507-4 YG9507-5 YG9507-6 YG9507-8 YG9507-11 YG9507-12
37◦ 22 N 37◦ 19 N 37◦ 19 N 37◦ 17 N 37◦ 12 N 37◦ 09 N 37◦ 15 N 37◦ 18 N
126◦ 14 E 125◦ 55 E 125◦ 59 E 125◦ 58 E 125◦ 47 E 125◦ 57 E 126◦ 01 E 126◦ 05 E
Stn.
1.80 ± 0.51 0.00 ± 0.00 5.42 ± 0.71 5.26 ± 0.77 1.66 ± 0.54 1.71 ± 0.58 0.15 ± 0.84 3.33 ± 0.58 1.35 ± 0.68 0.52 ± 0.33 0.88 ± 0.33 3.67 ± 0.91 0.94 ± 0.30 1.29 ± 0.22 1.14 ± 0.33
0.020 ± 0.006 0.038 ± 0.009 0.012 ± 0.003 0.016 ± 0.003 0.006 ± 0.002
SOC (%) 0.20 0.17 0.62 0.81 0.23 0.06 0.01
0.29 ± 0.29 2.18 ± 0.34 2.63 ± 0.33
3.37 ± 0.44 2.66 ± 0.33 1.89 ± 0.35 1.81 ± 0.29 2.84 ± 0.99 0.41 ± 0.23 3.93 ± 1.32 0.23 ± 0.56 1.86 ± 0.47 4.05 ± 1.53 3.32 ± 1.62
activities were 1.6 ± 1.2, 2.4 ± 2.1 and 0.29 ± 0.18 Bq kg−1 for 90 Sr, 137 Cs, 239+240 Pu, respectively. Sampling stations covered most of the South Yellow Sea, from the distal bottom deposits of the subaqueous Yellow River delta in a low sedimentation area, as well as from the periphery of the abandoned Yellow River delta of the high sedimentation area (Alexander et al., 1991). The 137 Cs massic activity of surface bottom sediment varied by more than a factor of five, being highest off the Shandong Peninsula and lowest in the mouth of the Yellow Sea (Fig. 17). The distribution pattern of 239+240 Pu activity of surface bottom sediments was similar to that of 137 Cs, but it varied only by a factor of two (Fig. 18). Bottom sediment 137 Cs distribution pattern is very similar to the sediment dispersal pattern in the sea. Surficial bottom sediment distribution in the Yellow Sea is characterized by the coarse, sandy deposits that flank the fine-grained (silt and clay) material in the central Yellow Sea, where high 137 Cs contents were found. The separate high 239+240 Pu contents found in the west of Cheju Island may be related to the fine-grained deposit flanking the southwest of the Korean Peninsula (Lee and Chough, 1989; Alexander et al., 1991). The spatial distribution of the sediment organic carbon content (%) was also very similar to that of 137 Cs, i.e. sediment organic carbon is more enriched in the fine-grained sediments than in the coarse sediments (Fig. 19). The association
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
Fig. 17. 137 Cs (Bq kg−1 ) massic activity in the surface sediment of the Yellow Sea.
Fig. 18. 239+240 Pu (Bq kg−1 ) massic activity in the surface sediments of the Yellow Sea.
123
124
G.H. Hong et al.
Fig. 19. Sediment organic carbon (%) in the surface sediment of the Yellow Sea.
of 137 Cs and 239+240 Pu with fine particles and sediment organic carbon content is well known (e.g. Baskaran et al., 1996). The sedimentation dynamics of 210 Pb was also investigated in order to provide bulk sediment accumulation rates, which will be used to derive sedimentary burial rates for artificial radionuclides, and elucidate the current depositional characteristics of the artificial radionuclides in the Yellow Sea (Table 8). Both 210 Pb and 90 Sr, 137 Cs and 239+240 Pu were largely supplied by the deposition from the atmosphere. They are strongly particle reactive, except 90 Sr, and thanks to the suitable half-life of 210 Pb (22.3 yr), it could serve the proxy for the artificial radionuclides for the past several decades. Since the atmospheric depositional flux is presumed to be relatively uniform over the entire Yellow Sea, any surplus or deficit against the mean values in the 210 Pb depositional flux would suggest that either a greater deposition, or a weaker deposition/erosion of 137 Cs and 239+240 Pu took place. The oceanic input of 210 Pb was estimated to be about one-fifth of the atmospheric input at the mouth of the Yellow Sea, based on the average residence time of water, the influx of Kuroshio Water, and concentration of 210 Pb in the Kuroshio (Hong et al., 1999b). Unfortunately, the artificial radionuclides were not determined for the same sediment samples subjected to 210 Pb analysis, however, close relationships between 210 Pb and artificial radionuclides are seen. A spatial distribution pattern of the total 210 Pb concentration (Fig. 20) in surficial sediments agrees reasonably with 137 Cs and 239+240 Pu (Figs. 17 and 18), considering the sampling stations were not the same. Since the average atmospheric 210 Pb deposition in the region is about 330 Bq m−2 yr−1 (Nozaki et al., 1973), a depositional 210 Pb flux over or below this value (Fig. 21), would be regarded as a deposition or an or erosion area in the central part of the sea, respectively. Good linear relationships between the excess 210 Pb and 137 Cs were re-
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
Fig. 20. 210 Pb massic activity (Bq kg−1 ) in the surface sediment of the Yellow Sea.
Fig. 21. 210 Pb flux (Bq m−2 yr−1 ) at the surface sediments of the Yellow Sea.
125
126 Table 8 210 Pb-derived sediment accumulation rates, surface mixed layer (SML) thickness, mixing coefficient, and 210 Pb depositional flux in the Yellow Sea. Sediment cores were taken in the period of 1991–1999 Station
Coordinates Latitude
Region Longitude
34◦ 00 N 125◦ 10.00 E Trough ◦ 35 51.30 N 124◦ 32.80 E Central part 35◦ 41 N 124◦ 37 E Central part
Y9404-08 Y9404-18
36◦ 54 N 33◦ 30 N
Y9503-3
3.9
Mixing SML coefficient residence time (cm2 yr−1 ) (yr)
100 82 87
48 28 59
0.02 0.03 0.04
0.04
123◦ 30 E 122◦ 30 N
Shandong coastal 67 Old Yellow River 23 delta
72 772
0.03 0.57
4.5
33◦ 00 N
125◦ 18 E
83
154
0.13
5.6
Y9503-08
32◦ 54 N
122◦ 54 E
29
192
0.11
6.5
Y9503-19 Y9503-32
35◦ 00 N 36◦ 10 N
125◦ 30 E 124◦ 30 E
Mouth of the Changjiang Old Yellow River delta Korean coastal Central part
47 72
128 76
0.08 0.06
8.5
Y9605-0203 35◦ 00 N Y9605-0205 35◦ 00 N
124◦ 00 E 122◦ 00 E
30 415
0.03 0.36
11.3
2.13
Y9605-0303 36◦ 00.3 N
124◦ 48 E
Central part 80 Old Yellow River 50 delta Central part 82
44
0.04
5.8
124◦ 30 E 123◦ 10 E 123◦ 30 E
Superficial 210 Pb flux (Bq m−2 yr−1 ) (Bq kg−1 ) 17 133
47 133
5
67 483
105 63
43
150
60
>9.5
44
250
50
>8.9
81
250 167
50 217
26
83 267
142 65
0.92
77
250
238
>3.9 1.35
73
210 Pb
Y9704-04 Y9704-09 Y9704-017
33◦ 10 N 34◦ 30 N 36◦ 15 N
Central part Central part Central part
82 80 75
214 46 39
0.18 0.06 0.06
6.8
3.9
36
133 83 117
48 147 252
Y9804-8 Y9804-10 Y9804-16
34◦ 35.43 N 124◦ 19.9 E Central part 35◦ 30.25 N 123◦ 29.84 E Central part 36◦ 05.19 N 124◦ 29.84 E Central part
85 71 81
44 130 40
0.04 0.23 0.04
8.7 8
10.8 17.3
156 37
283 450 117
130 217 253
G.H. Hong et al.
Y9104-13 Y9107-11 Y9111-5
Rate SML Water Accumulation depth depth (m) (mg cm−2 yr−1 ) (cm yr−1 ) (cm)
Station
Coordinates Latitude
Y9902-O1 Y9902-O2 Y9902-O3 Y9902-O4 Y9902-O5 Y9902-O6 Y9902-C1 Y9902-C6 Y9902-E2 Y9902-F1 Y9902-F2
Region Longitude
36◦ 00.1 N 124◦ 42.43 E Central part 36◦ 02.99 N 124◦ 48.80 E 0–7 cm section 7–13 cm section 35◦ 56.7 N 126◦ 36.4 E 0–5 cm section 0–6 cm section 36◦ 09.9 N 124◦ 30 E 9–17 cm section 35◦ 52.09 N 124◦ 52.1 E Central part 35◦ 40.83 N 123◦ 30 E 0–6 cm section 6–17 cm section 126◦ 27.00 E Korean coastal 36◦ 00 N 36◦ 00.02 N 124◦ 30.10 E 0–5 cm section 5–17 cm section Central part 33◦ 59.99 N 126◦ 00 E 33◦ 35 N 124◦ 15.06 E Central part 33◦ 00.2 N 124◦ 15.06 E Central part
Water Accumulation Rate SML depth depth (m) (mg cm−2 yr−1 ) (cm yr−1 ) (cm) 84 88 87 85 83 79 19 85 81 89 65
75 104 261 75 152 302 208 51 175 1591 20 227 61 162 119
0.05 0.07 0.18 0.07 0.13 0.27 0.12 0.07 0.24 1.04 0.02 0.17 0.05 0.14 0.12
7
Mixing SML coefficient residence time (cm2 yr−1 ) (yr) 0.05
118
210 Pb Superficial 210 Pb flux (Bq m−2 yr−1 ) (Bq kg−1 )
150 100
110 105
133 350
163 230
183 233
88 432
5 1
>54 >0.46
5 22
433 150
27 320
1 7 7
>0.62 0.35 0.06
16 38 54
100 117 100
88 112 78
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
Table 8 (Continued)
127
128
G.H. Hong et al.
Fig. 22. 210 Pb-derived sediment accumulation rate (mg cm−2 yr−1 ) in the Yellow Sea.
ported for their inventories and penetration depths in coastal sediments (Carpenter et al., 1987; Baskaran and Naidu, 1995). The 210 Pb-derived sediment accumulation rates revealed higher sediment accumulation rates along the coastal areas and lower sedimentation rates in the middle of the Yellow Sea (Fig. 22). They varied from 20 mg cm−2 yr−1 (0.02 cm yr−1 ) to 1590 mg cm−2 yr−1 (1.0 cm yr−1 ) with a geometric mean of 106 mg cm−2 yr−1 and the geometric standard deviation of 2.51 (Table 8). These values are similar to those obtained in previous studies (Alexander et al., 1991). Although the sediment accumulation rate varies a great deal in a relatively small area, a preliminary estimate of the sediment burial fluxes of particulate 90 Sr, 137 Cs and 239+240 Pu, based on the average 210 Pb-derived sediment accumulation rate, the average surface sediment deposition rates of 90 Sr, 137 Cs and 239+240 Pu for the entire Yellow Sea were found to be about 1.7 ± 1.3, 2.5 ± 2.2 and 0.3 ± 0.2 Bq m−2 yr−1 , respectively. However, high sedimentation regions of the Chinese coastal area were not included in this exercise, therefore the current estimate should be regarded as a minimum one. 3.6. Oceanic boundary conditions In order to compare levels of 90 Sr, 137 Cs and 239,240 Pu in the Yellow Sea (Table 6), oceanic boundary conditions were also determined in the Korea (Tsushima) Strait, East China Sea (Table 9) and the tropical Northwestern Pacific Ocean (Table 10) in 1993 and 1994. The average dissolved 90 Sr activity concentrations were 3.17 ± 0.39, 2.57 ± 0.48, 2.58 ± 0.32 mBq kg−1 in the Yellow, East China Sea and Tropical Pacific, respectively. The average dissolved 137 Cs activity concentrations were 2.89 ± 0.44, 2.83 ± 0.34, 2.56 ± 0.24 mBq kg−1 in the Yellow,
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
129
Table 9 Surface water distribution 90 Sr, 137 Cs, and 239+240 Pu activity concentrations in the Korea (Tsushima) Strait and East China Sea in December 1993 to September 1994. All quoted uncertainties are 1 sigma standard deviations Cruise
Stn.
Location
Date
Temp. Sal. (◦ C)
90 Sr
137 Cs
Latitude
Longitude
CO9312 A2 A4 A6 A8 A10
Korea Strait 34.4120◦ N 34.1862◦ N 33.9602◦ N 33.7343◦ N 33.5083◦ N
128.4590◦ E 128.6445◦ E 128.8297◦ E 129.0148◦ E 129.2000◦ E
06-Dec-93 06-Dec-93 06-Dec-93 06-Dec-93 07-Dec-93
14.61 17.67 18.70 20.31 19.05
33.42 33.87 34.11 34.58 34.30
3.34 ± 0.20 3.54 ± 0.30 3.20 ± 0.24 3.47 ± 0.15 2.26 ± 0.18
CO9312 CO9312 CO9312 CO9312 CO9312 CO9312 CO9312 CO9312 YS9406 YS9406 YS9406 CO9408 CO9408 CO9408 CO9408 CO9408
East China Sea∗∗ 30◦ 30.00 N 127◦ 0.00 E 30◦ 00.00 N 128◦ 20.00 E 33◦ 20.00 N 126◦ 00.00 E 33◦ 00.00 N 126◦ 00.00 E 31◦ 30.00 N 126◦ 00.00 E 31◦ 30.00 N 127◦ 20.00 E 31◦ 30.00 N 128◦ 20.00 E 31◦ 30.00 N 129◦ 20.00 E 31◦ 59.98 N 123◦ 00.08 E 32◦ 14.94 N 124◦ 00.07 E 32◦ 41.24 N 125◦ 45.08 E 32◦ 00.12 N 125◦ 59.83 E 29◦ 18.51 N 125◦ 15.14 E 28◦ 37.59 N 128◦ 00.04 E 29◦ 56.37 N 126◦ 44.95 E 29◦ 22.54 N 129◦ 00.11 E
10-Dec-93 10-Dec-93 12-Dec-93 12-Dec-93 08-Dec-93 08-Dec-93 08-Dec-93 08-Dec-93 30-Jun-94 30-Jun-94 30-Jun-94 05-Sept-94 03-Sept-94 02-Sept-94 01-Sept-94 01-Sept-94
20.37 22.27 18.35 18.90 19.57 21.40 21.24 20.75 24.48 19.48 21.51 28.43 28.82 28.53 28.42 28.98
34.37 34.78 34.52 34.44 34.13 34.64 34.60 34.51 30.11 31.14 31.82 31.61 34.04 34.63 33.95 34.56
3.23 ± 0.23 2.69 ± 0.22 1.84 ± 0.14 3.08 ± 0.23 3.02 ± 0.23 3.31 ± 0.19 2.88 ± 0.23 3.08 ± 0.25 2.73 ± 0.21 2.77 ± 0.22 2.79 ± 0.17 3.03 ± 0.22 2.56 ± 0.27 2.76 ± 0.16 2.93 ± 0.24 2.60 ± 0.21
H1 H6 D7 D9 B1 B4 B7 B10 E1 E3 E7 D10 L3 L11 J1 J9
239+240 Pu∗
(mBq kg−1 ) (mBq kg−1 ) (µBq kg−1 )
2.38 ± 0.04 2.56 ± 0.08 3.34 ± 0.06 2.82 ± 0.10 2.73 ± 0.07 2.84 ± 0.05 2.02 ± 0.03 1.87 ± 0.04
6.5 ± 0.6 3.9 ± 0.5 8.0 ± 0.6 7.0 ± 0.8 9.3 ± 0.8 5.2 ± 0.7 4.5 ± 0.5 8.1 ± 0.6 5.1 ± 0.6 4.9 ± 0.7 3.9 ± 0.5 7.0 ± 1.3 4.8 ± 1.7 6.3 ± 1.5 3.8 ± 0.9 5.0 ± 0.9
∗239+240 Pu data were reported earlier by Lee et al. (2003a). ∗∗ East China Sea was assigned the region south of Cheju Island (33◦ 30 N) for convenience.
East China Sea and tropical Northwestern Pacific, respectively. The highest activity concentration of dissolved 137 Cs was found in the Korea Strait (3.16 ± 0.52) where the chemical constituents are largely derived from the Yellow Sea, although the water mass in the Korea Strait originated from the Yellow Sea and the Kuroshio (Suk et al., 1996). The average 239+240 Pu activity concentrations were 6.14 ± 2.92, 5.83 ± 1.68 and 3.52 ± 1.22 mBq kg−1 in the Yellow Sea, East China Sea and tropical Northwestern Pacific, respectively. The radionuclide activity concentrations observed in the Yellow Sea were higher than in the Northwest Pacific (Fig. 23). In the early 1980s, the radionuclide concentrations were also higher in the Bohai Sea (northern part of the Yellow Sea) than in the East China Sea (Li et al., 1994). Therefore the Yellow Sea supply artificial radionuclides via water and particle transport to the North Pacific Ocean. A comprehensive study on the atom ratios of 240 Pu/239 Pu may reveal more information on the inter-basin redistribution of these radionuclides (Kim et al., 2004).
130
G.H. Hong et al.
Table 10 Surface water distribution 90 Sr, 137 Cs, and 239+240 Pu activity concentrations in the tropical Northwestern Pacific Ocean during April 1994. All quoted uncertainties are 1 sigma standard deviations Cruise
TG9404
Location
Water depth
Latitude
Longitude
25◦ N 20◦ N 15◦ N 10◦ N 5◦ N 2◦ N
132◦ E 133◦ 30 E 137◦ E 137◦ E 137◦ E 137◦ E
0 0 0 0 0 0 200 500 1000 2000
Temp. (◦ C)
Sal.
– – – 29.20 29.64 29.73 20.02 7.75 4.76 2.18
– – – 34.05 33.93 34.22 35.28 34.63 34.60 34.68
90 Sr
137 Cs
239+240 Pu
(mBq kg−1 )
(mBq kg−1 )
(µBq kg−1 )
2.71 ± 0.11 3.07 ± 0.08 2.66 ± 0.07 2.35 ± 0.07 2.54 ± 0.05 2.14 ± 0.10 2.05 ± 0.08
3.00 ± 0.13 2.58 ± 0.10 2.32 ± 0.10 2.60 ± 0.11 2.43 ± 0.10 2.43 ± 0.11
3.6 ± 0.7 3.3 ± 0.9 2.1 ± 0.8 5.1 ± 1.1 4.7 ± 0.8 2.3 ± 0.6 30.1 ± 2.7 4.3 ± 0.7 15.9 ± 2.2 8.9 ± 1.6
0.14 ± 0.02 0.15 ± 0.02
Fig. 23. Distribution of dissolved 90 Sr, 137 Cs, 239+240 Pu activity concentrations in the Northwest Pacific Ocean in the period of December 1993 to September 1994.
4. Conclusions 90 Sr, 137 Cs
and 239+240 Pu were studied in seawater and bottom sediments in the Yellow Sea during 1994–2000, along with atmospheric and riverine input of these radionuclides. From the present investigation, the following conclusions can be drawn: (i) The deposition of 239+240 Pu in the mid-eastern coast of the Yellow Sea (Ansan, Korea) was about 6.1–8.1 mBq m−2 yr−1 with highest values in the early spring. The monthly total deposition of 239+240 Pu in Ansan was very similar to that observed in Tsukuba, Japan, despite that Tsukuba is located further east more than 1000 km. Resuspension of surface soils from the arid regions of China and the subsequent deposition of fallout radionuclides from previously deposited debris on land has become an important source of radionuclides for the down wind sides including Yellow Sea
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
131
as well as Korean Peninsula. (ii) The wet deposition of these radionuclides during individual rainout events also showed the highest values in the spring, further supporting the tropospheric transport of these radionuclides in the region. (iii) The river water activity concentrations for 90 Sr, 137 Cs and 239+240 Pu were 2.8–4.3, 0.25–0.68 mBq l−1 and 2.7–3.1 µBq l−1 , respectively. (iii) The average activity concentrations of 90 Sr, 137 Cs and 239+240 Pu in surface water of the Yellow Sea were 2.17, 2.7 mBq kg−1 and 4.2 µBq kg−1 , respectively. They appeared to be supplied mainly from the land and their concentrations appeared to be higher towards the river mouth. During the summer stratification period, vertical segregation was pronounced for Cs and Pu isotopes. (iv) Despite the wide variation in the local water characteristics from which samples were taken, 90 Sr and 137 Cs activity concentrations exponentially decreased with time for the period of 1960–1994. Using the historical data the half-residence times of 90 Sr and 90 Cs in surface waters for this period were estimated to be 7 years. The current levels in the sea appeared to be relatively stable since the early part of the 1990s. (v) The average sedimentary 90 Sr, 137 Cs and 239+240 Pu massic activities in the seafloor were 1.6, 2.4 and 0.29 Bq kg−1 , respectively. The 210 Pb-derived sediment accumulation and mixing rates have been found useful for radionuclide sedimentation dynamics studies. (vi) The Yellow Sea appears to be currently supplying 90 Sr, 137 Cs, and 239+240 Pu to the East China Sea.
Acknowledgements The authors thank Professor Pavel P. Povinec (Marine Environment Laboratory, IAEA) for his continued encouragement and frequent advice which helped shape this paper. We also thank two anonymous reviewers for their constructive comments which improved greatly this paper. K.Y. Choi, S.K. Park, S.H. Chung, and D.S. Moon had provided laboratory assistance. This study was supported by Korea Ocean Research and Development Institute (PE89400 and BSPN 99383).
References Alexander, C.R., DeMaster, D.J., Nittrouer, C.A. (1991). Sediment accumulation in a modern epicontinental-shelf setting: The Yellow Sea. Marine Geology 98, 51–72. Baskaran, M., Naidu, A.S. (1995). 210 Pb-derived chronology and the fluxes of 210 Pb and 137 Cs isotopes into continental shelf sediments, East Chukchi Sea, Alaskan Arctic. Geochimica et Cosmochimica Acta 59, 4435–4448. Baskaran, M., Asbill, S., Santschi, P., Brooks, J., Champ, M., Adkinson, D., Colmer, M.R., Makeyev, V. (1996). Pu, 137 Cs and excess 210 Pb in Russian Arctic sediments. Earth and Planetary Science Letters 140, 243–257. Carpenter, R., Beasley, T.M., Zahnle, D., Somayajulu, B.L.K. (1987). Cycling of fallout (Pu, 241 Am, 137 Cs) and natural (U, Th, 210 Pb) radionuclides in Washington continental slope sediments. Geochimica et Cosmochimica Acta 51, 1897–1921. Cornet, R.J., Eve, T., Docherty, A.E., Cooper, E.L. (1995). Plutonium in freshwaters: Sources and behaviour in the Ottawa River Basin. Applied Radiation and Isotopes 46, 1239–1243. Eyrolle, F., Charmasson, S., Louvat, D. (2004). Plutonium isotopes in the lower reaches of the River Rhone over the period 1945–2000: Fluxes towards the Mediterranean Sea and sedimentary inventories. Journal of Environmental Radioactivity 74, 127–138. Gao, Y., Arimoto, R., Duce, R.A., Zhang, X.Y., Zhang, G.Y., An, Z.S., Chen, L.Q., Zhou, M.Y., Gu, D.Y. (1997). Temporal and spatial distributions of dust and its deposition to the China Sea. Tellus 49B, 172–189.
132
G.H. Hong et al.
Hamilton, T.F., Milliès-LaCroix, J.-C., Hong, G.H. (1996). 137 Cs (90 Sr) and Pu isotopes in the Pacific Ocean: Sources and trends. In: Guéguéniat, P., Germain, P., Métiver, H. (Eds.), Radionuclides in the Oceans: Inputs and Inventories. Les Editions de Physique, Les Ulis, pp. 29–58. Hong, G.H., Burrell, D.C. (1988). Sources, vertical distribution, and flux of particulate organic C, N, Biogenic Si, Mn and Fe in a southeastern Alaskan Fjord. In: Degens, E.T., Naidu, S. (Eds.), Transport of Carbon and Minerals in Major World Rivers, Lakes and Estuaries. Part 5. SCOPE/UNEP Sonderband Heft 68. Mitt. Geol-Paläont. Inst. Univ. Hamburg, Hamburg, pp. S185–S211. Hong, G.H., Zhang, J., Park, B.K. (1988). Health of the Yellow Sea. Earth Love Publication Association, Seoul, 342 pp. Hong, G.H., Kim, S.H., Chung, C.S., Kang, D.J., Shin, D.H., Lee, H.J., Han, S.J. (1997). 210 Pb-derived sediment accumulation rates in the southwestern East Sea (Sea of Japan). Geo-Marine Letters 17, 126–132. Hong, G.H., Kim, S.H., Lee, S.H., Chung, C.S., Tkalin, A.V., Chaykovskay, E.L., Hamilton, T.F. (1999a). Artificial radionuclides in the East Sea (Sea of Japan) Proper and Peter the Great Bay. Marine Pollution Bulletin 38, 933– 943. Hong, G.H., Park, S.K., Baskaran, M., Kim, S.H., Chung, C.S., Lee, S.H. (1999b). Lead-210 and polonium-210 in the winter well-mixed turbid waters in the mouth of the Yellow Sea. Continental Shelf Research 19, 1049–1064. Hong, G.H., Lee, S.H., Kim, S.H., Chung, C.S., Baskaran, M. (1999c). Sedimentary fluxes of 90 Sr, 137 Cs, 239,240 Pu and 210 Pb in the East Sea (Sea of Japan). The Science of the Total Environment 237/238, 225–240. Hong, G.H., Kremer, H.H., Pacyna, J., Chen, C.T.A., Behrendt, H., Salomons, W., Marshall Crossland, J.I. (Eds.) (2002). East Asia Basins: LOICZ Global Change Assessment and Synthesis of River Catchment-Coastal Sea Interaction and Human Dimensions. LOICZ Reports & Studies, vol. 26. LOICZ, Texel, p. 262. Hirose, K., Aoyama, M. (2003). Analysis of 137 Cs and 239,240 Pu concentrations in surface waters of the Pacific Ocean. Deep-Sea Research II 50, 2675–2700. Hirose, K., Igarashi, Y., Aoyama, M., Kim, C.K., Kim, C.S., Chang, B.W. (2003). Recent trends of plutonium fallout observed in Japan: Plutonium as a proxy for desertification. Journal of Environmental Monitoring 5, 302–307. Hirose, K., Kim, C.K., Kim, C.S., Chang, B.W., Igarashi, Y., Aoyama, M. (2004). Wet and dry deposition patterns of plutonium in Daejeon, Korea. The Science of the Total Environment 332, 243–252. Hölgye, Z., Filgas, R. (1995). Inventory of 238 Pu and 239,240 Pu in the soil of Chechoslovakia in 1990. Journal of Environmental Radioactivity 27, 181–189. IAEA (2005). Worldwide marine radioactivity studies (WOMARS). Radionuclide levels in oceans and seas. IAEATECDOC-1429. International Atomic Energy Agency, Vienna, 187 pp. Igarashi, Y., Otsuji-Hatori, M., Hirose, K. (1996). Recent deposition of 90 Sr and 137 Cs observed in Tsukuba. Journal of Environmental Radioactivity 31, 157–169. Igarashi, Y., Aoyama, M., Hirose, K., Miyao, T., Yabuki, S. (2001). Is it possible to use 90 Sr and 137 Cs as tracers for the Aeolian dust transport? Water, Air & Soil Pollution 130, 349–354. Igarashi, Y., Aoyama, M., Hirose, K., Miyao, T., Nemoto, K., Tomita, M., Fujikawa, T. (2003). Resuspension: Decadal monitoring time series of the anthropogenic radioactivity deposition in Japan. Journal of Radiation Research 44, 319–328. Ikeuchi, Y., Amano, H., Aoyama, M., Berezhnov, V.I., Chaykovskaya, E., Chumichev, V.B., Chung, C.S., Gastaud, J., Hirose, K., Hong, G.H., Kim, C.K., Kim, S.H., Miyao, T., Morimoto, T., Nikitin, A., Oda, K., Pettersson, H.B.L., Povinec, P.P., Tkalin, A., Togawa, O., Veletova, N.K. (1999). Anthropogenic radionuclides in seawater of the Far Eastern Seas. The Science of the Total Environment 273/238, 203–212. Kim, C.K., Kim, C.S., Yun, J.Y., Kim, K.H. (1997). Distribution of 3 H, 137 Cs and 239, 240 Pu in the surface seawater around Korea. Journal of Radioanalytical and Nuclear Chemistry 218, 33–40. Kim, C.S., Lee, M.H., Kim, C.K., Kim, K.H. (1998). 90 Sr, 137 Cs, 239+240 Pu and 238 Pu concentrations in surface soils of Korea. Journal of Environmental Radioactivity 40, 75–88. Kim, C.K., Kim, C.S., Chang, B.U., Choi, S.W., Chung, C.S., Hong, G.H., Hirose, K., Igarashi, Y. (2004). Plutonium isotopes in seas around the Korean Peninsula. The Science of the Total Environment 318, 197–209. Lee, H.J., Chough, S.K. (1989). Sediment distribution, dispersal and budget in the Yellow Sea. Marine Geology 87, 195–205. Lee, M.H., Lee, C.W., Hong, K.H., Choi, Y.H., Boo, B.H. (1996). Depth distribution of 239,240 Pu and 137 Cs in soils of South Korea. Journal of Radioanalytical and Nuclear Chemistry 204, 135–144. Lee, M.H., Lee, C.W., Boo, B.H. (1997). Distribution and characteristics of 239,240 Pu and 137 Cs in the soil of Korea. Journal of Environmental Radioactivity 37, 1–16.
Artificial radionuclides in the Yellow Sea: Inputs and redistribution
133
Lee, S.H., Gastaud, J., Povinec, P.P., Hong, G.H., Kim, S.H., Chung, C.S., Lee, K.W., Pettersson, H.B.L. (2003a). Distribution of plutonium and americium in the marginal seas of the northwest Pacific Ocean. Deep-Sea Research II 50, 2727–2750. Lee, S.H., La Rosa, J.J., Levy-Palomo, I., Oregioni, B., Pham, M.K., Povinec, P.P., Wyse, E. (2003b). Recent inputs and budgets of 90 Sr, 137 Cs, 239,240 Pu and 241 Am in the northwest Mediterranean Sea. Deep-Sea Research II 50, 2817–2834. Li, P., Yu, Y., Wu, Y. (1994). The development of marine radiochemistry in China. In: Zhou, D., Liang, Y. (Eds.), Oceanology of China Seas, vol. 1. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 189–200. Liu, S.M., Zhang, J., Chen, S.Z., Chen, H.T., Hong, G.H., Wei, H., Wu, Q.M. (2003). Inventory of nutrient compounds in the Yellow Sea. Continental Shelf Research 23, 1161–1174. Nagaya, Y., Nakamura, K. (1992). 239+240 Pu and 137 Cs in the East China and the Yellow Seas. Journal of Oceanography 48, 23–35. Nozaki, Y., Tsunogai, S., Nishimura, M. (1973). Lead-210 in the Japan Sea. Journal of the Oceanographical Society of Japan 29, 251–256. Nozaki, Y., DeMaster, D.J., Lewis, D.M., Turekian, K.K. (1978). Atmospheric Pb-210 fluxes determined fro soil profiles. Journal of Geophysical Research 83, 4047–4051. Povinec, P.P., Aarkrog, A., Buesseler, K.O., Delfanti, R., Hirose, K., Hong, G.H., Ito, T., Livingston, H.D., Nies, H., Noshkin, V.E., Shima, S., Togawa, O. (2005). 90 Sr, 137 Cs and 239,240 Pu concentration surface water time series in the Pacific and Indian Oceans – WOMARS results. Journal of Environmental Radioactivity 81, 63–87. Pujol, L., Sanchez-Cabeza, J.A. (2000). Natural and artificial radioactivity in surface waters of the Ebro River basin (Northeast Spain). Journal of Environmental Radioactivity 51, 181–210. Suk, M.S., Hong, G.H., Chung, C.S., Chang, K.I., Kang, D.J. (1996). Distribution and transport of suspended particulate matter, dissolved oxygen and major inorganic nutrients in the Cheju Strait. Journal of the Korean Society of Oceanography 31, 55–63. Thein, M., Ballestra, S., Yamato, Y., Fukai, R. (1980). Delivery of transuranic elements by rain to the Mediterranean Sea. Geochimica et Cosmochimica Acta 44, 1091–1097. Zhang, J., Liu, S.M., Lu, X., Huang, W.W. (1993). Characterizing Asian wind dust transport to the northwest Pacific Ocean: Direct measurement of the dust flux for two years. Tellus 45B, 335–345. Zhu, H., Li, S., Wu, F., Liu, Q., Yang, W. (1991). Radioactivity in the coastal waters of the Bohai and Yellow Seas of China. Journal of Environmental Radioactivity 14, 193–209.
This page intentionally left blank
3. Radionuclides in the European seas
This page intentionally left blank
137
Distribution of anthropogenic radionuclides in the water column of the south-western Mediterranean Sea S.-H. Leea,* , F.R. Mantouraa , P.P. Povineca , J.A. Sanchez-Cabezaa , J.-L. Pontisa , A. Mahjoubb , A. Noureddinec , M. Boulahdidd , L. Choubae , M. Samaalif , N. Reguiguif a International Atomic Energy Agency, Marine Environment Laboratory, 98000, Monaco b International Atomic Energy Agency, Department of Technical Cooperation, Vienna, Austria c Commissariat à l’Energie Atomique, Centre de Recherche Nucléaire d’Alger, 16000 Alger, Algeria d Institut des Sciences de la Mer et de l’Aménagement du Littoral, Dely Ibrahim, 16320 Alger, Algeria e Institut National des Sciences et Technologies Marines, TN-2060 La Goulette, Tunisia f Centre National des Sciences et Technologies Nucléaires, Sidi Thabet, TN-2020 Ariana, Tunisia
Abstract In the framework of the IAEA’s Technical Co-operation project RAF7/004, international research cruises were carried out in 2001 and 2004 in order to assess the distribution of anthropogenic radionuclides in the south-western (SW) Mediterranean Sea. Measured surface 239,240 Pu activity concentrations were slightly lower, while 241 Am concentrations were comparable to those observed in the north-western (NW) part of the Sea. The subsurface 239,240 Pu maximum was found in the Algerian Basin at 250 m, but in the Sardinia Channel it was much deeper (1000 m). 137 Cs activity concentration gradually decreased with depth, but higher values were observed at the bottom layers in the Sardinia and Sicily Channels. Higher 137 Cs activity concentrations observed in the Sicily Channel may be influenced by the Levantine Intermediate Water (LIW), carrying higher concentrations due to the Chernobyl accident. Higher concentrations of conservative anthropogenic radionuclides (90 Sr and 137 Cs) and lower concentrations of non-conservative radionuclides (239,240 Pu and 241 Am) were found in the bottom layer at the Sardinia Channel, possibly indicating an intrusion of surface waters. The accumulated inventories of 239,240 Pu, 241 Am, 137 Cs and 90 Sr are lower than those found in the NW Mediterranean, and also lower than the global fallout deposition in these latitudes. 238 Pu/239,240 Pu activity ratios indicate that plutonium in the SW Mediterranean is of global fallout origin. The 241 Am/239,240 Pu activity ratios are much lower than that of global fallout due to the enhanced scavenging of 241 Am from the water column. Keywords: Anthropogenic radionuclides, 90 Sr, 137 Cs, 241 Am, Pu isotopes, Water column, Mean residence time, Algerian Basin, Sardinia Channel, Sicily Channel, South Mediterranean
* Corresponding author. Address: IAEA-MEL, 4 Quai Antoine 1er, MC 98000, Monaco; phone: (+377) 97977229;
fax: (+377) 97977273; e-mail:
[email protected] RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08008-3
© 2006 Elsevier Ltd. All rights reserved.
138
S.-H. Lee et al.
1. Introduction The Mediterranean Sea is a semi-enclosed basin connected to the North Atlantic Ocean through the narrow and shallow Strait of Gibraltar, exchanging water, salt, heat and other substances. The basin is characterized by low precipitation and high evaporation. The surface water coming from Gibraltar and the intermediate water from the eastern Mediterranean cross the central Mediterranean region. The Algerian Current transports Modified Atlantic Water (MAW) across the Algerian Basin to the Sardinia Channel. East of the Sardinia Channel, a minor part of MAW penetrates into the Tyrrhenian Sea, and the major part enters the eastern Mediterranean through the Sicily Channel (Sammari et al., 1999). Below MAW, Levantine Intermediate Water (LIW) flows westward (Stansfield et al., 2003). Most of LIW and TDW (Tyrrhenian Deep Water) flow out from the Tyrrhenian Sea through the Sardinia Channel, along the Sardinian slope (Millot, 1999). As the straits play a key role in controlling the flows and the water exchange, the area between Sardinia, Sicily and Tunisia is thought to be an important region for the comprehension of exchange of water masses between the eastern and western Mediterranean basins (Astraldi et al., 1998). Marine radiotracer studies can effectively contribute to a better understanding of water dynamics and the behavior of contaminants in the Mediterranean Sea. Nevertheless, available data on isotope tracers for the Tunisia–Sardinia–Sicily area are scarce. In the framework of the IAEA’s Technical Co-operation project “Contamination Assessment of the South Mediterranean Sea (RAF/7/004)”, carried out in collaboration with several institutes from the North African countries, the IAEA-MEL organized four expeditions to sample seawater, sediment and biota in coastal zones of Morocco, Algeria and Tunisia. The present work aims to investigate the distribution of anthropogenic radionuclides such as plutonium isotopes, 241 Am, 137 Cs, and 90 Sr in the water column of the SW Mediterranean Sea, along the Algerian and Tunisian coasts.
2. Materials and methods Oceanographic cruises in the SW Mediterranean Sea were carried out from 8 August to 23 August 2001 along the Algerian coast and from 14 June to 29 June 2004 along the Tunisian coast using a research vessel “Mohamed Seddik Benyahia” owned by “Institut des Sciences de la Mer et de l’Aménagement du Littoral”, Algeria. The study area is shown in Fig. 1. A large volume water sampler (250 L) was used for collecting water column samples. All seawater samples were pumped through a membrane filter (0.45 µm pore, 293 mm diameter). Samples (∼150 L) were taken at several depths down to a maximum depth of 1500 m. They were acidified with HCl to pH 1–2 and tracers were added (242 Pu, 134 Cs, 243 Am and 85 Sr). The radiochemical and counting techniques used have already been described in detail by La Rosa et al. (2001) and Lee et al. (2005), therefore they will be only shortly described here. Plutonium and americium were coprecipitated with MnO2 . 8 M HNO3 was added to re-dissolve the precipitate followed by plutonium purification by anion exchange. For the separation of americium, the anion exchange column eluate fraction was evaporated. Ca-oxalate precipitations and TRU extraction chromatography were used for Am purification. An alcohol ion exchange step was then performed to separate rare earth elements from Am. The
Anthropogenic radionuclides in the water column of the south-western Mediterranean Sea
139
Fig. 1. Sampling stations in the SW Mediterranean Sea in 2001 (◦) and 2004 (•).
final column eluates, containing the purified Pu and Am isotopes, were evaporated to dryness, treated to remove organic traces, electrodeposited onto stainless steel discs and measured by alpha-ray spectrometry. After the MnO2 precipitation, an appropriate amount of ammonium molybdophosphate (AMP) was mixed with the re-acidified supernatant solution to adsorb Cs on the AMP. The AMP precipitate was dissolved with 10 mL of 10 M NaOH, transferred to a standardized container and diluted to an appropriate volume for gamma-ray spectrometry analysis. 90 Sr was analyzed by oxalate precipitation of the supernatants after sequential separation of 137 Cs. The oxalate fraction was ashed and strontium was precipitated as strontium nitrates in concentrated HNO3 . For further purification, the strontium nitrates were set aside for 14 days to enable full in-growth of 90 Y. Yttrium was precipitated as yttrium oxalate after the addition of stable yttrium, and analyzed in a gas proportional counter. The chemical recoveries of 85 Sr and 137 Cs were determined by gamma-ray spectrometry, and by gravimetric method for yttrium precipitates. IAEA-381 (Irish Sea water) reference material was used to assess data quality. The replicate analyses of investigated radionuclides were within the 95% confidence intervals.
3. Results and discussion 3.1. Algerian Basin 239,240 Pu activity concentrations in surface waters of the Algerian Basin ranged from 6.5 ±0.5
to 7.0 ± 0.4 µBq L−1 (Table 1 and Fig. 2). The 239,240 Pu activity concentration increased significantly with depth up to 250 m (21 ± 1 µBq L−1 ) and then remained constant up
140
Table 1 Activity concentrations of 239,240 Pu and 241 Am in the water column of the SW Mediterranean. The quoted uncertainties represent 1 sigma (σ ) standard deviations Sampling date
Location
Depth (m)
239,240 Pu (µBq L−1 )
238 Pu (µBq L−1 )
241 Am (µBq L−1 )
238 Pu/239,240 Pu
241 Am/239,240 Pu
St. 1 Jijel
13/Aug/01
36◦ 54 N, 05◦ 32 E
0 250 550 950 1200
6.5 ± 0.5 20.9 ± 0.9 22.2 ± 1.1 22.4 ± 1.0 21.7 ± 1.0
1.21 ± 0.23 0.80 ± 0.14 0.62 ± 0.15 0.90 ± 0.16 1.16 ± 0.19
1.29 ± 0.13 1.07 ± 0.12 1.51 ± 0.16 1.90 ± 0.18 2.02 ± 0.27
0.19 ± 0.04 0.038 ± 0.007 0.028 ± 0.007 0.040 ± 0.007 0.054 ± 0.009
0.20 ± 0.03 0.051 ± 0.006 0.068 ± 0.008 0.085 ± 0.009 0.093 ± 0.013
24.0 ± 0.4
Total inventory (Bq m−2 )
1.0 ± 0.1
1.9 ± 0.1
St. 2 Alger
18/Aug/01
36◦ 54 N, 03◦ 20 E
0 250
7.0 ± 0.4 7.7 ± 0.3
0.28 ± 0.07 0.40 ± 0.07
0.74 ± 0.10 0.70 ± 0.10
0.041 ± 0.010 0.052 ± 0.009
0.110 ± 0.020 0.090 ± 0.010
St. 1 Sicily Channel
20/June/04
37◦ 36 N, 11◦ 28 E
0
7.9 ± 1.2
–
–
–
–
St. 3 Sardinia Channel
24/June/04
38◦ 09 N, 09◦ 07 E
0 250 500 750 1000 1250 1483
8.8 ± 0.3 14.1 ± 0.6 18.7 ± 0.7 21.8 ± 0.7 20.5 ± 0.8 21.2 ± 0.9 6.9 ± 0.3
0.36 ± 0.06 1.22 ± 0.15 0.84 ± 0.11 0.75 ± 0.09 0.85 ± 0.13 0.68 ± 0.13 0.38 ± 0.07
0.57 ± 0.08 0.62 ± 0.08 0.71 ± 0.09 0.97 ± 0.11 0.92 ± 0.22 0.96 ± 0.12 0.74 ± 0.10
0.041 ± 0.007 0.087 ± 0.011 0.045 ± 0.006 0.035 ± 0.004 0.041 ± 0.007 0.032 ± 0.006 0.055 ± 0.010
0.065 ± 0.010 0.044 ± 0.006 0.038 ± 0.005 0.045 ± 0.005 0.045 ± 0.011 0.045 ± 0.006 0.107 ± 0.016
26.0 ± 0.3
1.2 ± 0.1
1.2 ± 0.1
Total inventory (Bq m−2 )
S.-H. Lee et al.
Station
Anthropogenic radionuclides in the water column of the south-western Mediterranean Sea
141
Fig. 2. Profiles of 239,240 Pu, 241 Am and 137 Cs in the water column of the Algerian Basin in 2001 and 2004.
to 1200 m. Plutonium and americium are characterized as particle reactive elements, therefore, their vertical profiles tend to be similar to that of nutrients. The observed low concentrations of plutonium (and americium, Fig. 2) in surface water are due to scavenging by particulate matter. The vertical transport of plutonium and americium associated with sinking particles, and the subsequent regeneration of these elements as a result of the decomposition of particulate matter, causes the subsurface 239,240 Pu and 241 Am maxima observed in the water column (Bowen et al., 1980; Fowler et al., 1983; Molero et al., 1995; Povinec et al., 2003; Lee et al., 2003a, 2003b). However 241 Am, due to its higher particle affinity, is more rapidly scavenged from the surface water than plutonium. The surface 239,240 Pu activity concentrations in the Algerian Basin are slightly lower than in the NW Mediterranean Sea (Lee et al., 2003a), in the Sardinia Channel and in the Gibraltar Strait (Benmansour et al., 2006). The 239,240 Pu activity concentration observed in 2001 (21 ± 1 µBq L−1 ) at 250 m depth at Jijel (St. 1) is higher than that at St. 2 (7.7 ± 0.3 µBq L−1 ), although both stations are located in the same basin. The local variations in plutonium concentrations might be due to large mesoscale turbulences that have been observed in the Algerian Basin (Bouzinac et al., 1998). The highest 137 Cs activity concentration (3.1 ± 0.2 mBq L−1 ) is observed at the surface layer (Fig. 2, Table 2), then it decreases with depth, and a higher activity is observed again at the bottom layer (2.9 ± 0.2 µBq L−1 ). This bottom water enrichment is thought to be the due to the influence of LIW coming from the eastern Mediterranean basin, where higher concentrations of 137 Cs were observed due to the Chernobyl fallout (Papucci et al., 1996). 3.2. The Sardinia Channel The Sardinia Channel is the passage between the Algerian Basin and the Tyrrhenian Basin, limited by Sardinia Island and the northern Tunisian coast. The 239,240 Pu, 241 Am, 137 Cs and 90 Sr profiles in the water column south of Sardinia Island are shown in Fig. 3. The 239,240 Pu activity concentration has a minimum at the surface (8.8 ± 0.3 µBq L−1 ), then it is gradually increasing until reaching a maximum at 800 m (22 ± 1 µBq L−1 ), much deeper than the
142
S.-H. Lee et al.
Table 2 Activity concentrations of 90 Sr and 137 Cs in the water column of the SW Mediterranean (2004). The quoted uncertainties represent 1 sigma standard deviations Station
Sampling date
Location
Depth (m)
St. 1 Sicily Channel
20/June/04
37◦ 36 N, 11◦ 28 E
0 150 350 650
38◦ 09 N, 09◦ 07 E
0 250 500 750 1000 1250 1483
37◦ 13 N, 08◦ 17 E
0 150 350 500
Total inventory (kBq m−2 ) St. 3 Sardinia Channel
24/June/04
Total inventory (kBq m−2 ) St. 4 Algerian Basin
26/June/04
Total inventory (kBq m−2 )
137 Cs (µBq L−1 )
90 Sr
137 Cs/90 Sr
1.9 ± 0.3 1.5 ± 0.2 2.1 ± 0.2 2.6 ± 0.2
1.48 ± 0.04 1.52 ± 0.04 1.60 ± 0.04 1.44 ± 0.04
1.31 ± 0.17 1.00 ± 0.12 1.31 ± 0.16 1.83 ± 0.10
(µBq L−1 )
1.3 ± 0.1
0.99 ± 0.01
1.98 ± 0.08 2.40 ± 0.17 1.73 ± 0.06 1.35 ± 0.06 1.55 ± 0.07 1.32 ± 0.07 2.18 ± 0.10
2.04 ± 0.05 2.15 ± 0.05 1.91 ± 0.05 1.83 ± 0.04 1.93 ± 0.05 1.47 ± 0.04 2.32 ± 0.06
2.58 ± 0.02
2.8 ± 0.2
3.1 ± 0.2 2.6 ± 0.2 2.5 ± 0.2 2.9 ± 0.2
– – –
1.3 ± 0.004
–
0.97 ± 0.05 1.12 ± 0.09 0.91 ± 0.04 0.74 ± 0.04 0.80 ± 0.04 0.90 ± 0.05 0.94 ± 0.05
maximum observed in the Algerian Basin. A remarkable decrease of 239,240 Pu and 241 Am activity concentrations was observed at the bottom layer. The activity concentrations of 90 Sr and 137 Cs in the water column at Sardinia Channel also varied with depth. A subsurface maximum concentration of 137 Cs was found at 250 m and then decreased with depth, although a higher concentration, comparable to the surface water concentration, was found in the bottom layer. The 90 Sr activity concentrations followed a similar but smoother pattern. Activity concentrations of 239,240 Pu, 241 Am, 90 Sr and 137 Cs observed near the bottom layer indicate an intrusion of surface waters, as documented by the temperature and salinity profiles (Fig. 3). During winter and early spring, cooling of superficial waters increases their density and induces vertical convection (Schmidt and Reyss, 1996; Bethoux et al., 2002). Also, strong mesoscale perturbations in the surface layer could affect the radiotracers distributions, as some eddies extend down to 2000 m (Bouzinac et al., 1998). Both these processes could transport high 90 Sr and 137 Cs and low 239,240 Pu and 241 Am concentrations with surface waters to the deep layer (Delfanti et al., 1994; Lee et al., 2003a). However, further studies are needed to better understand the observed radiotracers distributions in the Sardinia Channel. 3.3. The Sicily Channel The Sicily Channel provides a direct interface between the eastern and western Mediterranean. The two-layer water system in the Sicily Strait is maintained by the excess of evaporation and the specific circulation in the eastern Mediterranean (Astraldi et al., 1999). While the MAW
Anthropogenic radionuclides in the water column of the south-western Mediterranean Sea
143
Fig. 3. Profiles of 239,240 Pu, 241 Am and 90 Sr in the water column in the Sardinia Channel in 2004.
is found in the sea surface next to the Tunisian coast, the LIW flow fills the whole bottom section of the Strait. The observed surface 239,240 Pu activity concentration (7.9 ± 1.2 µBq L−1 ) is comparable with that found in the Sardinia Channel and the Algerian Basin surface waters (Table 1). The vertical profile of 137 Cs in the Sicily Strait (Fig. 4 and Table 2) shows a subsurface minimum (at 150 m, 1.5 ± 0.2 mBq L−1 ) and a bottom maximum (at 650 m, 2.6 ± 0.2 mBq L−1 ). However, the 90 Sr vertical profile shows a constant distribution with depth (Fig. 4). A few 90 Sr results were reported by Merino et al. (1997), who found 0.57–1.77 mBq L−1 in the Catalan Sea. A recent study (Lee et al., 2003a) showed 1.3 ± 0.1 mBq L−1 in surface waters of the NW Mediterranean, in agreement with 1.48–1.60 mBq L−1 found in the Sicily Strait. The 137 Cs maximum observed in deep water might be attributed to the presence of LIW, carrying higher levels of 137 Cs from the Chernobyl accident, but not for 90 Sr. Indeed, LIW carrying higher levels of 137 Cs moves westward, passes over the Sicily Strait and enters the western Mediterranean basin (Ozsony et al., 1993; Papucci et al., 1996; Papucci and Delfanti, 1999; Stansfield et al., 2003; Lee et al., 2003a). 3.4. Radionuclide inventories in the water column Radionuclide seawater inventories were calculated and compared with global fallout inputs estimated from data on the global distribution of fallout in this latitude band (UNSCEAR, 2001). The deposited 239,240 Pu inventory in the water column is 24.0 ± 0.4 Bq m−2 for the Algerian Basin and 26.0 ± 0.3 Bq m−2 for the Sardinia Channel, respectively. These inventories are much lower than that expected from global fallout because of the small water column
144
S.-H. Lee et al.
Fig. 4. Profiles of 137 Cs and 90 Sr in the water column in the Sicily Channel in 2004.
in the visited stations. The 238 Pu inventory is estimated to be 1.0 ± 0.1 Bq m−2 for the Algerian Basin and 1.2 ± 0.1 Bq m−2 for the Sardinia Channel, respectively. The 241 Am inventory in the water column is 1.9 ± 0.1 Bq m−2 for the Algerian Basin and 1.2 ± 0.1 Bq m−2 for the Sardinia Channel, respectively, which are lower than those observed in the NW Mediterranean (Lee et al., 2003a). The 137 Cs inventories in the water column in the Sicily Channel and in the Algerian Basin are 1.3 ± 0.1 and 1.3 ± 0.4 kBq m−2 , respectively. Higher 137 Cs inventory (2.58 ± 0.02 kBq m−2 ) has been observed in the Sardinia Channel. Nonetheless, those are much lower than the value (3.7 ± 0.1 kBq m−2 ) observed in the NW Mediterranean (Lee et al., 2003a) because of the smaller water column. The 90 Sr inventory in the water column at Sicily and Sardinia Channels is 0.99 ± 0.01 kBq m−2 and 2.8 ± 0.2 kBq m−2 , respectively. The 90 Sr inventory for Sardinia waters is even higher than for the NW Mediterranean (Lee et al., 2003a). The observed flat profiles of 90 Sr in the water column is possibly due to eddies. Similar flat profiles in the SW Mediterranean have also been observed by Noureddine et al. (2006) and Benmansour et al. (2006). 3.5. Radionuclide activity ratios 238 Pu/239,240 Pu
and 241 Am/239,240 Pu activity ratios in the water column are reported in Table 1. The activity ratios of 238 Pu/239,240 Pu are, within uncertainties, the same as the global fallout ratio (0.03) at these latitudes, indicating that global fallout is the main source of plutonium in the SW Mediterranean. The observed 137 Cs/90 Sr activity ratios in the water column ranged from 0.74 ± 0.04 to 1.8 ± 0.1 (Table 2), i.e. most of them are lower than the global fallout ratio of 1.6
Anthropogenic radionuclides in the water column of the south-western Mediterranean Sea
145
(UNSCEAR, 2001). This difference might be attributable to a higher scavenging rate of 137 Cs than 90 Sr in shallow coastal waters, due to its higher particle affinity, as the distribution coefficients (Kd s) for 137 Cs are higher by a factor of 500 and 10 than for 90 Sr for open ocean and coastal waters, respectively (IAEA, 2004). 4. Conclusions The distribution of anthropogenic radionuclides in the water column of the SW Mediterranean Sea gives useful indications for better understanding of oceanic processes and water transport in the area. Various oceanic processes such advection by water currents, convection by eddies and remineralization have affected the distribution of radionuclides in the water column. The main observations can be summarized as follows: (i) The surface 239,240 Pu activity concentrations in the SW Mediterranean Sea are slightly lower than the values reported for the NW region, probably because of the inflow of surface Atlantic waters; however, the surface 241 Am concentrations are comparable to those observed in NW Mediterranean waters. (ii) Higher 137 Cs activity concentrations were observed at the surface layer (MAW), and at bottom layers of the Sardinia and Sicily Channels. Higher activities observed in the deep waters of the Sardinia Channel might indicate the intrusion of surface waters. (iii) The inventories of 239,240 Pu, 241 Am and 137 Cs in the water column of the SW Mediterranean Sea are lower than those of global fallout deposition at this latitude. The inventories of 90 Sr in the water column are lower than that of global fallout deposition in the Sicily Channel, or comparable to that of global fallout in the Sardinia Channel. (iv) The 238 Pu/239,240 Pu activity ratios confirm that plutonium in the SW Mediterranean Sea is of global fallout origin. The 241 Am/239,240 Pu and 137 Cs/90 Sr activity ratios are lower than those of global fallout due to an enhanced scavenging of 241 Am and 137 Cs from the water column. Acknowledgements The authors would like to thank the Institut des Sciences de la Mer et de l’Aménagement du Littoral (Alger, Algeria), the Captain and the crew members of R/V Mohamed Seddik Benyahia, and the Algerian, Tunisian and IAEA-MEL colleagues participating in the cruises for assistance during seawater sampling. We also thank Ms. J. Gastaud and Ms. M.K. Pham for assistance during alpha- and gamma-spectrometry measurements. The IAEA is grateful for the support provided to its Marine Environment Laboratory by the Government of the Principality of Monaco. References Astraldi, M., Gasparini, G.P., Sparnocchia, S. (1998). Water masses and seasonal hydrographic conditions in the Sardinia–Sicily–Tunisia region. Rapports Commission Internationale pour l’exploration scientifique de la Mer Méditerranée 35, 122–123.
146
S.-H. Lee et al.
Astraldi, M., Balopoulos, S., Candela, J., Font, J., Gacic, M., Gasparini, G.P., Manca, B., Theocharis, A., Tintore, J. (1999). The role of straits and channels in understanding the characteristics of Mediterranean circulation. Progress in Oceanography 44, 65–108. Benmansour, M., Laissaoui, A., Benbrahim, S., Ibn Majah, M., Chafik, A., Povinec, P.P. (2006). Distribution of anthropogenic radionuclides in Moroccan coastal waters and sediments. In: Povinec, P.P., Sanchez-Cabeza, J.A. (Eds.), Radionuclides in the Environment, International Conference on Isotopes in Environmental Studies: Aquatic Forum 2004. 25–29 October, Monaco. Elsevier, Amsterdam, pp. 148–155, this volume. Bethoux, J.P., Durieu de Madron, X., Nyffeler, F., Tailliez, D. (2002). Deep water in the western Mediterranean: Peculiar 1999 and 2000 characteristics, shelf formation hypothesis, variability since 1970 and geochemical inferences. Journal of Marine System 33/44, 117–131. Bouzinac, C., Font, J., Millot, C., Vazquez, J. (1998). Circulation variability the channel of Sardinia observed from in situ and altimetric data. Rapports Commission Internationale pour l’exploration scientifique de la Mer Méditerranée 35, 128–129. Bowen, V.T., Noshkin, V.E., Livingston, H.D., Volchok, H.L. (1980). Fallout radionuclides in the Pacific Ocean: Vertical and horizontal distributions, largely from GEOSECS stations. Earth Planetary and Science Letters 49, 411–434. Delfanti, R., Papucci, C., Vives i Battle, J., Downes, A.B., Mitchell, P.I. (1994). Distribution of 137 Cs and transuranics elements in seawater of the western Mediterranean Sea (Algerian Basin, Balearic Sea). In: Cigna, A., Delfanti, R., Serro, R. (Eds.), The Radiological Exposure of the Population of the European Community to Radioactivity in the Mediterranean Sea. European Commission, Luxembourg, pp. 427–439. Fowler, S.W., Ballestra, S., La Rosa, J., Fukai, R. (1983). Vertical transport of particulate-associated plutonium and americium in the upper water column of the Northeast Pacific. Deep-Sea Research 12 (A), 1221–1233. IAEA (2004). Sediment distribution coefficients and concentration factors for biota in the marine environment. IAEA Technical Reports Series No. 422. La Rosa, J.J., Burnett, W., Lee, S.H., Levy, I., Gastaud, J., Povinec, P.P. (2001). Separation of actinides, cesium and strontium from marine samples using extraction chromatography and sorbents. Journal of Radioanalytical and Nuclear Chemistry 248, 765–770. Lee, S.-H., La Rosa, J.J., Levy-Palomo, I., Oregioni, B., Pham, M.K., Povinec, P.P., Wyse, E. (2003a). Recent inputs and budgets of 90 Sr, 137 Cs, 239,240 Pu and 241 Am in the northwest Mediterranean Sea. Deep-Sea Research II 50, 2817–2834. Lee, S.-H., Gastaud, J., Povinec, P.P., Hong, G.-H., Kim, S.-H., Chung, C.-S., Lee, K.W., Pettersson, H.B.L. (2003b). Distribution of plutonium and americium in the marginal seas of the Northwest Pacific Ocean. Deep-Sea Research II 50, 2727–2750. Lee, S.-H., La Rosa, J., Gastaud, J., Povinec, P.P. (2005). The development of sequential separation methods for analysis of actinides in sediment and biological materials using anion exchange resins and extraction chromatography. Journal of Radioanalytical and Nuclear Chemistry 263 (2), 419–425. Merino, J.A., Sanchez-Cabeza, J.A., Bruach, J.M., Masque, P., Pujol, L.I. (1997). Artificial radionuclides in high resolution water column profile from the Catalan Sea (the Northwestern Mediterranean). Radioprotection – Colloques 32, C2-85–C2-90. Millot, C. (1999). Circulation in the western Mediterranean Sea. Journal of Marine Systems 20, 423–442. Molero, J., Sanchez-Cabeza, J.A., Merino, J., Pujol, Ll., Mitchell, P.I., Vidal-Quadras, A. (1995). Vertical distribution of radiocaesium, plutonium and americium in the Catalan Sea (Northwest Mediterranean). Journal of Environmental Radioactivity 26, 205–216. Noureddine, A., Menacer, M., Boudjenoun, R., Benkrid, M., Boulahdid, M., Kadi-hanifi, M., Lee, S.-H., Povinec, P.P. (2006). 137 Cs in seawater and sediment along the Algerian coast. In: Povinec, P.P., Sanchez-Cabeza, J.A. (Eds.), Radionuclides in the Environment, International Conference on Isotopes in Environmental Studies: Aquatic Forum 2004. 25–29 October, Monaco. Elsevier, Amsterdam, pp. 156–164, this volume. Ozsony, E., Hecht, A., Unluata, U., Brenner, S., Sur, H.I., Bishop, J., Latif, M.A., Rozentruub, Z., Oguz, T. (1993). A synthesis of the Levantine Basin circulation and hydrography, 1985–1990. Deep-Sea Research II 40 (6), 1075– 1119. Papucci, C., Delfanti, R. (1999). 137 Cs distribution in the eastern Mediterranean Sea: Recent change and future trends. The Science of the Total Environment 237/238, 67–75. Papucci, C., Charmasson, S., Delfanti, R., Gasco, C., Mitchell, P., Sanchez-Cabeza, J.A. (1996). Time evolution and levels of man-made radioactivity in the Mediterranean Sea. In: Guéguéniat, P., Germain, P., Métivier, H. (Eds.), Radionuclides in the Oceans: Input and Inventories. Les éditions de physique, Les Ulis, pp. 177–197.
Anthropogenic radionuclides in the water column of the south-western Mediterranean Sea
147
Povinec, P.P., Livingston, H.D., Shima, S., Aoyama, M., Gastaud, J., Goroncy, I., Hirose, K., Huynh-Ngoc, L., Ikeuchi, Y., Ito, T., La Rosa, J., Liong Wee Kwong, L., Lee, S.-H., Moriya, H., Mulsow, S., Oregioni, B., Pettersson, H., Togawa, O. (2003). IAEA’97 expedition to the NW Pacific Ocean – Results of oceanographic and radionuclide investigations of the water column. Deep-Sea Research II 50, 2607–2637. Sammari, C., Millot, C., Taupier-Letage, I., Stefani, A., Brahim, M. (1999). Hydrological characteristics in the Tunisia–Sardinia–Sicily area during spring 1995. Deep-Sea Research I 46, 1671–1703. Schmidt, S., Reyss, J.-L. (1996). Radium as internal tracer of Mediterranean outflow water. Journal of Geophysical Research 101 (C2), 3589–3596. Stansfield, K., Gasparini, G.P., Smeed, D.A. (2003). High-resolution observations of the path of the overflow from the Sicily Strait. Deep-Sea Research I 50, 1129–1149. UNSCEAR (1993). Source and effects of ionizing radiation. United Nations, New York.
148
Distribution of anthropogenic radionuclides in Moroccan coastal waters and sediments M. Benmansoura,* , A. Laissaouia , S. Benbrahimb , M. Ibn Majaha , A. Chafikb , P.P. Povinecc a Centre National de l’Energie, des Sciences et des Techniques Nucléaires, Rabat, Morocco b Institut National de Recherche Halieutique, Casablanca, Morocco c International Atomic Energy Agency, Marine Environment Laboratory, Monaco
Abstract Concentrations of anthropogenic radionuclides (137 Cs, 239, 240 Pu and 241 Am) were measured in surface seawater, water column and sediment samples collected in Moroccan coastal waters. The average activity concentrations in surface water were 2.7 mBq L−1 , 8.7 µBq L−1 and 1.5 µBq L−1 for 137 Cs, 239, 240 Pu and 241 Am, respectively. The vertical distributions of 239,240 Pu in the water column of the southwest Mediterranean showed a sub-surface maximum at 500 m water depth, while 137 Cs concentrations were constant up to 900 m. On the contrary to the 239,240 Pu profile, which has been as expected, the 137 Cs profile has been affected by movement of water masses in the western Alboran Sea. The total inventories of 137 Cs and 239,240 Pu in the water column (0–900 m) were estimated to be 2170 and 23 Bq m−2 , respectively, and they were as expected for this water depth. The concentration profiles of 137 Cs and 239,240 Pu in the sediment core revealed the presence of sub-surface maximum located at the depth of 5 cm. The sedimentation rate of 0.12 cm yr−1 has been estimated using the 210 Pb profile, which indicated that the observed peaks of 137 Cs and 239,240 Pu originated in the sediment core following the maximum global fallout deposition in 1963. Keywords: Radionuclides, 137 Cs, 239,240 Pu, 241 Am, 210 Pb, Seawater, Sediment, Sedimentation rate, Strait of Gibraltar, Alboran Sea, Mediterranean Sea, Atlantic Ocean, Morocco
1. Introduction The main sources of anthropogenic radionuclides in the Mediterranean Sea are global fallout from atmospheric nuclear weapon tests carried out mainly in the fifties and the early sixties, and the Chernobyl accident which occurred in 1986 (e.g. Fukai et al., 1979; Delfanti et al., 1995; Molero et al., 1995; Papucci et al., 1996; León Vintró et al., 1999; Fowler et al., 2000; Lee et al., 2003). The global inputs of 137 Cs and 239,240 Pu to the Mediterranean Sea up to 1986 are estimated to be 12 PBq and 0.19 PBq, respectively (Holm et al., 1988; UNEP, 1992). The * Corresponding author. Address: CNESTEN, B.P. 1382, R.P. 10001, Rabat, Morocco; phone: (+212) 37819757; fax: (+212) 37803317; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08009-5
© 2006 Elsevier Ltd. All rights reserved.
Anthropogenic radionuclides in Moroccan coastal waters and sediments
149
Chernobyl accident significantly increased the total input of 137 Cs into the environment of the Mediterranean Sea. Papucci et al. (1996) estimated that about 2.5 PBq of 137 Cs was deposited after the accident mainly in the eastern and northern parts of the Mediterranean basin. The distribution of anthropogenic radionuclides in the southwest Mediterranean has not been well established. Therefore one of the objectives of the IAEA’s regional technical cooperation project RAF/7/004 “Contamination Assessment of the South Mediterranean Sea” has been to study present radionuclide levels in the Mediterranean Sea and to use radionuclides as tracers for better understanding of processes in the water column.
2. Materials and methods 2.1. Seawater and sediment samples Samples were collected in four stations alongside the Moroccan coast which is approximately 3500 km long, which 500 km are in the southwest Mediterranean Sea. Three stations were situated in the Mediterranean Sea (the Alboran Sea and Strait of Gibraltar) and one in the Atlantic Ocean. The cruise was carried out in December 1999 using the R/V Charif Al Idrissi of the Institut National de la Recherche Halieutique (INRH). The visited sampling stations are shown in Fig. 1 and their corresponding specifications are given in Table 1. Station 2 located in the Mediterranean Sea (35◦ 47 N, 04◦ 48 W) was extensively explored by collecting water profile samples at different depths until 900 m, and a bottom sediment core collected at the water depth of 800 m. Due to low concentrations of transuranic elements, large volumes of water (∼200 L) were sampled using 30 L Niskin bottles. The water samples were promptly filtered through membrane filters of 0.45 µm pore size to remove any suspended matter.
Fig. 1. Sampling stations in the Atlantic Ocean and the Mediterranean Sea.
150
M. Benmansour et al.
Table 1 137 Cs, 239,240 Pu and 241 Am activity concentrations in surface seawater and in the water column (reported uncertainties are ±1σ ; N.D. means not determined) Station
137 Cs (mBq L−1 )
239,240 Pu (µBq L−1 )
St. 1 (Mohammadia) 33◦ 46 N, 7◦ 28 W St. 2 (Mdiq) 35◦ 47 N, 4◦ 48 W 0m 250 m 500 m 900 m
2.95 ± 0.22
8.7 ± 0.9
2.32 ± 0.18 2.46 ± 0.18 2.50 ± 0.19 2.29 ± 0.16
St. 3 (Mdiq) 35◦ 47 N, 5◦ 15 W St. 4 (Tangier) 35◦ 52 N, 5◦ 51 W
241 Am
241 Am/239,240 Pu
239,240 Pu/137 Cs
N.D.
N.D.
0.0025 ± 0.0003
8.3 ± 0.9 23 ± 2 33 ± 5 27 ± 2
1.50 ± 0.20 N.D. 7.6 ± 1.2 N.D.
0.18 ± 0.03 N.D. 0.23 ± 0.05 N.D.
0.0035 ± 0.0004 0.0089 ± 0.0010 0.0114 ± 0.0020 0.0118 ± 0.0013
2.37 ± 0.14
8.8 ± 1.4
1.56 ± 0.50
0.17 ± 0.02
0.0037 ± 0.0006
3.11 ± 0.23
8.3 ± 0.9
N.D.
N.D.
0.0027 ± 0.0003
(µBq L−1 )
Appropriate tracers of 242 Pu and 243 Am, and a stable Cs carrier were added, after acidifying the samples with concentrated HCl to pH 1–2, to serve as indicators of chemical recoveries. Sequential separations of radionuclides were carried out on board by co-precipitating Pu and Am, along with the other actinides, with manganese dioxide MnO2 . 137 Cs was precipitated by physical adsorption onto AMP. The sediment core was collected using a 40 cm × 40 cm box corer (Ocean Instruments). The core was sectioned to series of horizontal slices of 0.5–1 cm thick. The sediment samples were freeze-dried, sieved and homogenized before a non-destructive gamma-ray spectrometry analysis was carried out. 2.2. Analytical methods 137 Cs, 210 Pb, 226 Ra activities in samples were measured using an HPGe detector (n-type) with
resolution of 2 keV and the relative efficiency of 50% at 1332 keV (in comparison with the 7.6 cm height × 7.6 cm in diameter NaI(Tl) detector). The stable Cs in water samples was measured by ICPMS to estimate the chemical recovery. For alpha-emitters such as 239,240 Pu and 241 Am suitable radiochemical methods were applied for their extraction from seawater and sediment samples (Lee et al., 2001; La Rosa et al., 2001). Pu was separated from Am by anion-exchange resin AG1 × 8. Am was co-precipitated with calcium oxalate and extracted into DDCP and finally separated from rare earths by anion exchange in mineral acids–methanol media. Both Pu and Am fractions were electrodeposited on stainless steel discs and the resulting alpha-sources were analyzed by alpha-ray spectrometry using silicon solid state detectors. The analysis of 210 Pb in sediment samples was also carried using its daughter product 210 Po. After the total digestion of a sediment sample and a spontaneous deposition of Po onto silver discs, the measurements were carried out by alpha-ray spectrometry.
Anthropogenic radionuclides in Moroccan coastal waters and sediments
151
3. Results and discussion 3.1. Seawater Table 1 lists 137 Cs, 239,240 Pu and 241 Am activity concentrations in surface waters for the four visited stations, as well as in water profile samples collected to a depth of 900 m at Station 2 located in the Alboran Sea. Some 241 Am activity concentrations are not given due to low recoveries of the radiochemical procedures. The 137 Cs activity concentrations in surface water of the Moroccan coast ranged from 2.32 to 3.11 mBq L−1 with an average concentration for the Mediterranean stations of 2.60 ± 0.21 mBq L−1 . These values are in the same range as those reported by other authors in previous studies regarding to the Mediterranean Sea area. Before the Chernobyl accident, which occurred in 1986, 137 Cs activity concentrations in different regions of the Mediterranean Sea ranged between 3.2–4.8 mBq L−1 and the average 137 Cs concentration was estimated to be 3.4 mBq L−1 (Fukai et al., 1980). The 137 Cs levels after 1986 increased mainly in the northern and eastern zones of the Sea. From recent Mediterranean studies (Molero et al., 1995; Merino et al., 1997; Sanchez-Cabeza and Molero, 2000; Lee et al., 2003), 137 Cs activity concentrations in surface seawater have been between 2.3 and 5.6 mBq L−1 . 137 Cs activity concentrations in the water profile of Station 2 (Fig. 2) do not show a noticeable surface maximum (Lee et al., 2003), as the concentration at 900 m water depth is comparable with the concentration at the surface. Such a homogeneous distribution of 137 Cs in the water column, which is behaving as a conservative tracer, may be due to movement of water masses in the area. Atlantic water, characterized by low salinity and density, enters the Mediterranean Sea via the Strait of Gibraltar as surface water (100–200 m water depth), while near the bottom, water masses with high salinity and density previously formed in the Mediterranean leave the Sea towards the Atlantic Ocean. The 239,240 Pu surface activity concentrations determined in this work range between 8.29 and 8.92 µBq L−1 and are comparable with the data reported by León Vintró et al. (1999) in the same zone ten years earlier, as well as with recent data published by Lee et al. (2003). The vertical distribution of 239,240 Pu in the water column of Station 2 shows a typical subsurface maximum at a depth close to 500 m (Fig. 2). This behavior particularly for plutonium was already observed in previous studies (Fukai et al., 1979; Mitchell et al., 1995; Merino et al., 1997; León Vintró et al., 1999; Fowler et al., 2000; Lee et al., 2003) and is usually
Fig. 2. Vertical distributions of 137 Cs and 239,240 Pu in the water column (Station 2).
152
M. Benmansour et al.
attributed to the removal of the plutonium from the surface in association with scavenging processes, its transport with suspended matter to a deeper water layers, and the subsequent particle dissolution/remineralization at depth, with the return of plutonium into the solution. Only a few results have been obtained for 241 Am in this work. However, on the basis of the results obtained for Stations 2 and 3 situated in the Alboran Sea, the mean 241 Am activity concentration in surface water has been 1.53 ± 0.18 µBq L−1 . This value is similar to that given by Lee et al. (2003) in a recent study in the northwest Mediterranean Sea (∼1.5 µBq L−1 ), and to that measured earlier by Molero et al. (1995) along the Spanish Mediterranean coast (∼1.0 µBq L−1 ). The 241 Am activity concentration at the depth of 500 m is 7.6 ± 1.2 µBq L−1 , much higher than at the surface. This is because of the fact that Am is even more particle reactive element than Pu, and its behavior in the water column is therefore similar to Pu. The total radionuclide inventories in the water column (Station 2, water depth 0–900 m) were estimated to be 2170 Bq m−2 for 137 Cs and 23 Bq m−2 for 239,240 Pu. They are both lower by a factor of two than the inventories reported in previous studies (Fowler et al., 2000; Lee et al., 2003), as the height of the water column at Station 2 has been smaller by about factor of two. The 239,240 Pu/137 Cs activity ratio increases with depth from 0.0035 to 0.012, as expected (Lee et al., 2003), suggesting a depletion of plutonium at the surface layer in comparison with cesium due to its preferential association with sinking particulate mater (Fowler et al., 2000; Lee et al., 2003). The 241 Am/239,240 Pu activity ratio is estimated to be 0.18 at the surface and 0.23 at the 500 m water depth, however, this change is within the reported uncertainties. 3.2. Sediment The vertical distributions of massic activities of 137 Cs and 239,240 Pu in the sediment core (Station 2) are plotted in Fig. 3. The both activity profiles revealed a presence of the subsurface maximum, located at the depth of ∼5 cm. It can be assumed that these maxima correspond to the maximum atmospheric deposition of anthropogenic radionuclides from global fallout which occurred in 1963 (Lee et al., 2003). The 137 Cs and 239,240 Pu activities ranged between 1.90–6.3 Bq kg−1 and 0.31–0.80 Bq kg−1 , respectively, and are within the range of those reported in the literature (Delfanti et al., 1995; Papucci et al., 1996; León Vintró et al., 1999). The total inventories were about 280 Bq m−2 for 137 Cs and 32 Bq m−2 for 239,240 Pu. It is worth noting that the activities and corresponding inventories of transuranics in sediment depend strongly on the sedimentation rate which is itself a function of local hydrodynamic conditions. However, as the suspended particle concentration in the Alboran Sea is very low, only a small fraction of transuranics is scavenged from the water column and deposited on the seafloor. In addition to activity profiles of the anthropogenic 137 Cs and 239,240 Pu, profiles of naturally occurring radionuclides such as 210 Pb and 226 Ra were determined in the sediment core, in order to estimate, through the excess 210 Pb, a sedimentation rate and to reconstruct the history of contaminants present in the sediment. 210 Pb found in the sediment has two components: (i) a supported 210 Pb, which comes from the 226 Ra decay that occurs in the sediment;
Anthropogenic radionuclides in Moroccan coastal waters and sediments
153
Fig. 3. Vertical distributions of 137 Cs and 239,240 Pu in the sediment core (Station 2).
Fig. 4. Excess 210 Pb profile in the sediment core (Station 2).
(ii) an unsupported or “excess” 210 Pb (210 Pbex ), which originates from 222 Rn decay in the atmosphere and the water column, and 210 Pb as its daughter product is deposited on the sediment. The second component is used in dating models of sediments. The 210 Pbex activity profile is shown in Fig. 4. The 210 Pbex activity in the top 3 cm of the sediment is almost uniform (∼600 Bq kg−1 ), which corresponds to the mixed sediment layer formed due to bioturbation processes in the sediment. Below the 3 cm layer, the activity profile is decreasing with depth following an exponential form, as a consequence of the radioactivity decay law. By applying the constant rate of supply of 210 Pbex model, a value of 0.12 cm yr−1 has been found for the average linear sedimentation rate. If the data is exploited for dating the sediment core, the peaks observed for anthropogenic 137 Cs and 239.240 Pu correspond to the period of 1962–1963, when the maximum global fallout deposition occurred.
154
M. Benmansour et al.
4. Conclusions The levels and behavior of anthropogenic radionuclides, with special emphasis on the 137 Cs, 239,240 Pu and 241 Am in Moroccan coastal waters have been presented in this paper. The most important findings can be summarized as follows: • The mean activity concentrations in surface water were around 2.7 mBq L−1 , 8.7 µBq L−1 and 1.5 µBq L−1 for 137 Cs, 239,240 Pu and 241 Am, respectively. These values are in an agreement with those reported in recent studies carried out in the western Mediterranean Sea, and are within the ranges of concentrations expected from global fallout. • The vertical distributions of 239,240 Pu in the water column of the southwest Mediterranean showed a sub-surface maximum at 500 m water depth, while 137 Cs concentrations were constant up to 900 m. On the contrary to the 239,240 Pu profile, which has been as expected, the 137 Cs profile has been affected by movement of water masses in the western Alboran Sea. • The total inventories of 137 Cs and 239,240 Pu in the water column (0–900 m) were estimated to be 2170 and 23 Bq m−2 , respectively, and they were as expected for this water depth. • The 239,240 Pu/137 Cs activity ratio was increasing with depth, suggesting a progressive separation of plutonium from cesium due to the Pu association with suspended matter. • The depth profiles of 137 Cs and 239,240 Pu in the sediment core showed a maximum peak which has been attributed to the maximum global fallout deposition which occurred in 1963. This result was confirmed by 210 Pb dating of the sediment core, estimating the sedimentation rate of 0.12 cm yr−1 .
Acknowledgements The authors thank colleagues from IAEA-MEL, CNESTEN and INRH for assistance during sampling and pre-treatment of collected samples. They also thank the Captain and the crew of the R/V Charif Al Idrissi for help during the sampling expedition. The support provided by IAEA for the RAF/7/004 project and for the sampling cruise is highly acknowledged. The IAEA is grateful for the support provided to its Marine Environment Laboratory by the Government of the Principality of Monaco.
References Delfanti, R., Desideri, D., Martinotti, W., Assunta Melti, M., Papucci, C., Queirraza, G., Testa, C., Triulzi, C. (1995). Plutonium concentration in sediment cores collected in the Mediterranean Sea. The Science of the Total Environment 173/174, 187–193. Fowler, S.W., Noshkin, V.E., La Rosa, J., Gastaud, J. (2000). Temporal variations in plutonium and americium inventories and their relation to vertical transport in the northwest Mediterranean Sea. Limnology and Oceanography 45, 446–458. Fukai, R., Holm, E., Ballestra, S. (1979). A note on vertical distribution of plutonium and americium in the Mediterranean Sea. Oceanologica Acta 2, 129–132. Fukai, R., Ballestra, S., Vas, D. (1980). Distribution of 137 Cs in the Mediterranean Sea. In: Management of Environment. Wiley Eastern Ltd, New Delhi, pp. 353–360.
Anthropogenic radionuclides in Moroccan coastal waters and sediments
155
Holm, E., Fukai, R., Whitehead, N.E. (1988). Radiocesium and transuranium elements in the Mediterranean Sea: Sources, inventories and environmental levels. In: International Conference on Environmental Radioactivity in the Mediterranean Area. SNE, Barcelona, pp. 601–617. La Rosa, J., Burnett, W., Lee, S.H., Levy, I., Gastaud, J., Povinec, P.P. (2001). Separation of actinides, caesium and strontium from marine samples using extraction chromatography and adsorbents. Journal of Radioanalytical and Nuclear Chemistry 248, 765–770. Lee, S.H., Gastaud, J., La Rosa, J., Liong Wee Kwong, L., Povinec, P.P., Wyse, E., Fifield, L.K., Hausladen, P.A., Di Tada, L.M., Santos, G.M. (2001). Analysis of plutonium isotopes in marine samples by radiometrics, ICPMS and AMS techniques. Journal of Radioanalytical and Nuclear Chemistry 248, 754–764. Lee, S.-H., La Rosa, J.J., Levy-Palomo, I., Oregioni, B., Pham, M.K., Povinec, P., Wyse, E. (2003). Recent inputs and budgets of 90 Sr, 137 Cs, 239,240 Pu and 241 Am in the northwest Mediterranean Sea. Deep-Sea Research II 50, 2817–2834. León Vintró, I., Mitchell, P.I., Condren, O.M., Downes, A.B., Papucci, C., Delfanti, R. (1999). Vertical and horizontal fluxes of plutonium and americium in the western Mediterranean and the strait of Gibraltar. The Science of the Total Environment 237/238, 77–91. Merino, J., Sanchez-Cabeza, J.A., Bruach, J.M., Masqué, P., Pujol, L.I. (1997). Artificial radionuclides in high resolution water column profile from the Catalan Sea (the Northwestern Mediterranean). Radioprotection Colloques 32, C2. Mitchell, P.I., Vives, J., Batlle, I., Downes, A.B., Condren, O.M., León Vintró, L., Sanchez-Cabeza, J.A. (1995). Recent observations on the physico-chemical speciation of plutonium in the Irish Sea and the Western Mediterranean. Applied Radiation and Isotopes 46, 1190–1995. Molero, J., Sanchez-Cabeza, J.A., Merino, J., Pujol, L.I., Mitchell, P.I., Vial-Quadras, A. (1995). Vertical distribution of radiocesium, plutonium and americium in the Catalan Sea (Northwest Mediterranean). Journal of Environmental Radioactivity 26, 205–216. Papucci, C., Charmasson, S., Delfanti, R., Gasco, C., Mitchell, P., Sanchez-Cabeza, J.A. (1996). Time evolution and levels of man-made radioactivity in the Mediterranean Sea. In: Guéguéniat, P., Germain, P., Métivier, H. (Eds.), Radionuclides in the Oceans: Input and Inventories. Les Editions de Physique, Les Ulis, pp. 177–197. Sanchez-Cabeza, J.A., Molero, J. (2000). Plutonium, americium and radiocesium in the marine environment close to the Vandellos I nuclear power plant before decommissioning. Journal of Environmental Radioactivity 51, 211– 228. UNEP, United Nations Environmental Programme (1992). Assessment of the state of pollution of the Mediterranean Sea by radioactive substances. MAP Technical Reports Series No. 62. Athens 66.
156
137 Cs
in seawater and sediment along the Algerian coast
A. Noureddinea,* , M. Menacera , R. Boudjenouna , M. Benkrida , M. Boulahdidb , M. Kadi-hanific , S.-H. Leed , P.P. Povinecd a Commissariat à l’Energie Atomique, Centre de Recherche Nucléaire d’Alger, Algiers, Algeria b Institut des Sciences de la Mer et de l’Aménagement du Littoral, Alger, Algeria c Universite des sciences et de la technologie Houari Boumediène, Alger, Algeria d International Atomic Energy Agency, Marine Environment Laboratory, Monaco
Abstract In order to define a baseline of natural and anthropogenic radionuclides, a sampling campaign along the Algerian coast was organised during the last quarter of 1999 to collect seawater and sediment samples. A total of 25 surface and water column samples were collected along the coast. In addition to a sediment core at Annaba station at the eastern coast, 41 surface sediment samples were collected each 25 km along the coast. 137 Cs activity concentrations in surface seawater ranged from 1.69 ± 0.11 to 3.3 ± 0.2 Bq/m3 with a mean value of 2.03 ± 0.13 Bq/m3 . A slight increase of 137 Cs activity (within a factor of two) from the western to the eastern coast of Algeria was observed. The 137 Cs water profiles had peaks at 80–100 m water depths. Massic activities of 137 Cs in surface sediment samples varied between 8.15 and 35.4 Bq/kg dry weight. The observed differences could be due to several reasons, namely the nature and composition of the sediment, and the water depth at the sampling location. Corresponding inventories of 137 Cs in the water column and sediment core were calculated, and a sedimentation rate using unsupported 210 Pb was estimated to be 0.2 cm/year. Keywords: Marine radioactivity, 137 Cs, 210 Pb, Sedimentation rate, Radionuclide inventory, Seawater, Sediment, Global fallout, Chernobyl accident, Mediterranean
1. Introduction The most common primordial radionuclides in the environment are 40 K and 238 U, 235 U and 232 Th, which are also parents of the three natural decay series, producing several radiogenic radionuclides with wide applications in environmental studies (e.g. 234 Th, 226 Ra, 222 Rn, 210 Pb, 228 Ra, etc.). In addition to that, a number of anthropogenic sources of radioactive contamination exist since the last century in the environment. These include global fallout from atmospheric * Corresponding author. Address: Centre de Recherche Nucléaire d’Alger (CRNA), 02 Bd. F. Fanon, BP 399 Algiers-Gare, 16000 Algiers, Algeria; phone: (+213) 21 43 44 44; fax: (+231) 21 43 42 80; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08010-1
© 2006 Elsevier Ltd. All rights reserved.
137 Cs in seawater and sediment along the Algerian coast
157
nuclear weapons test carried out during the 1950s and 1960s (with a distinct peak in 1963), radioactive substances released from nuclear reprocessing facilities and other nuclear plants, dumping of radioactive wastes, and nuclear accidents, especially that of Chernobyl which occurred in 1986 (e.g. Burton, 1975; Livingston and Povinec, 2000). For the Mediterranean region the most important sources of anthropogenic radionuclides have been global fallout, the Chernobyl accident (with a release of about 10 ± 2 PBq of 137 Cs; UNEP, 1992) and nuclear facilities on the Rhône River. Anthropogenic radionuclides are introduced into the marine environment following many pathways. In order to study the behaviour of radionuclides, their transport and cumulative deposition, it is important to determine their concentrations in seawater and sediment (e.g. Calmet and Grauby, 1988; Buffoni and Cappelletti, 1997; Delfanti et al., 2001; Lee et al., 2003). The aim of this work, which was carried out in the framework of the Algerian monitoring and contamination assessment programme, and the IAEA’s Technical Cooperation project ALG/2/006, has been to define a baseline of natural and anthropogenic radionuclides in seawater and marine sediment along the Algerian coast, to estimate the corresponding inventory of 137 Cs in the water column and sediment, and to evaluate sedimentation rates using the 210 Pb method.
2. Materials and methods 2.1. Sampling The sampling campaign was organised during September–October 1999 by the Laboratory of radiological impact studies of the Centre de Recherche Nucléaire d’Alger (CRNA), on board of the research vessel M.S. Benyahia (belonging to the Institut des Sciences de la Mer et de l’Aménagement du Littoral (ISMAL)) in order to collect surface seawater, water column samples, surface sediment and a sediment core along the Algerian coast. Sampling locations are shown in Fig. 1. A total of 25 surface and water profile seawater samples (40–70 L) were collected. Surface samples were taken at a depth of approximately 5 m below the water line by means of a pumping system. Water column samples were taken using a 50 L Niskin PVC water-bottle sampler. The collected unfiltered seawater was first acidified with HCl to pH ≈ 1–2, spiked with 134 Cs for determining the chemical yield, and coprecipitated with ammonium molybdophosphate (AMP) to preconcentrate the sample to a few litres, in order to bring it to the laboratory for analysis (e.g. Roos et al., 1994). After pre-treatment and drying, the AMP precipitates were analysed in the laboratory by direct gamma-ray spectrometry. A total of 41 surface sediments were collected each 25 km along the coast from Ghazaouet station (35◦ 10 90N, 02◦ 55 857E) situated on the West, to El-Kala (near Annaba station) on the East (36◦ 59 49N, 08◦ 26 93E) (not shown in Fig. 1). A Van Veen Grab sampler was used for the sediment sampling at water depths ranging from 13 to 170 m. In addition, a sediment core was collected using a box corer at Annaba station (on the eastern coast, 36◦ 56 504N, 07◦ 47 750E), at the water depth of around 50 m. A multiparameter probe was also used at the same station to determine temperature, salinity and pH profiles of the water column. On board,
158
A. Noureddine et al.
Fig. 1. Sampling stations for the water profiles and the sediment core collection.
surface layer sediments were put in plastic bags and labelled to record station co-ordinates, date and wet weight. The sediment core samples were cut with a metallic sheet each 2 cm until a depth of 22 cm. All sediment samples were then stored in a freezer to be later analysed in the laboratory by gamma-ray spectrometry. 2.2. Gamma-ray spectrometry A high purity germanium detector of 23% relative efficiency and resolution of 1.8 keV at 1332 keV (FWHM) has been used for analysis of gamma-ray emitters. The AMP seawater precipitates were placed in 100 cm3 plastic beakers, while the sediment samples after drying at 80◦ C, crushing into fine powder and homogenisation were put either into 100 cm3 plastic cylindrical containers or 500 cm3 Marinelli beakers. For the seawater samples two standards were prepared with the same geometry in which 134 Cs and 137 Cs spikes were injected (activity of 13 and 17 Bq, respectively) using two aliquots of wet AMP covering the different masses of AMP precipitates, in order to calculate the chemical yield using 134 Cs and to determine the detection efficiency using the 137 Cs spike. In the case of the sediments the detection efficiency curve was determined by measuring a standard sample of the same nature and geometry, in which a liquid radioactive solution of 152 Eu was injected. For the energies below 122 keV, a similar standard sample was spiked with a liquid radioactive source of 210 Pb. The seawater and sediment samples were counted until statistically reliable results were obtained (usually 48–72 h). A 4096 channels analyser with Genie-PC software (CANBERRA) was used to collect and analyse the spectra. The data quality was assured by analysing IAEA reference materials and by participation in intercomparison exercises: IAEA-381 – Irish Sea water for seawater samples (Povinec et al., 2002); IAEA-384 – Fangataufa lagoon sediment (Povinec and Pham, 2000); IAEA-385 – Irish Sea sediment, for sediment samples.
137 Cs in seawater and sediment along the Algerian coast
159
3. Results and discussion 3.1. Seawater activity concentrations in surface seawater samples ranged from 1.59 ± 0.11 Bq/m3 (observed at the westernmost station at Ghazaouet) to 3.31 ± 0.21 Bq/m3 (observed at the easternmost station at Annaba), with a mean value of 2.03 ± 0.13 Bq/m3 (Table 1). This would indicate a weak west–east gradient, however, the observed levels differ only by a factor of two. As Atlantic waters with lower 137 Cs levels are entering the Mediterranean Sea as surface waters, it should be possible to observe a west–east gradient along the Algerian coast. Generally, the activity concentrations of 137 Cs in surface seawater are comparable with those measured by other authors in different regions of the Mediterranean Sea (Delfanti et al., 2001; Lee et al., 2003; Benmansour et al., 2006; Lee et al., 2006). The activity concentrations of 137 Cs in water profiles (Fig. 2) appear to increase with a water depth at the Algiers station, having a maximum of 2.41 ± 0.15 Bq/m3 at 100 m, and probably 137 Cs
Table 1 137 Cs activity concentrations in seawater along the Algerian coast
Longitude
Latitude
Sampling depth (m)
137 Cs concentration (Bq/m3 )
Skikda
36◦ 56 107N
06◦ 55 857E
surface
1.69 ± 0.11
Annaba
36◦ 56 504N
07◦ 47 750E
surface 15 25 35
3.31 ± 0.21 1.92 ± 0.12 2.67 ± 0.15 2.10 ± 0.12
Azzefoun
36◦ 55 300N
04◦ 22 000E
surface
1.81 ± 0.11
Jijel
36◦ 45 500N
05◦ 53 500E
surface
1.89 ± 0.11
Algiers
35◦ 10 90N
02◦ 05 000E
surface 50 75 100 130 160
2.09 ± 0.12 2.28 ± 0.14 2.35 ± 0.14 2.41 ± 0.15 2.31 ± 0.15 2.26 ± 0.14
Bejaia Ténès
36◦ 44 32N
05◦ 07 000E
surface 30
1.92 ± 0.12 1.72 ± 0.10
Mostaganem
36◦ 00 20N
00◦ 02 20W
surface 15 35 55
1.92 ± 0.12 1.72 ± 0.10 1.80 ± 0.12 2.29 ± 0.14
Ghazaouet
35◦ 10 90N
02◦ 05 000W
surface 50 75 100 130 160
1.59 ± 0.10 1.63 ± 0.10 1.87 ± 0.13 1.87 ± 0.12 1.72 ± 0.11 2.02 ± 0.12
Station/location
Coordinates
160
A. Noureddine et al.
Fig. 2. 137 Cs water column profiles along the Algerian coast.
also at Ghazaouet station, having a maximum of 1.87 ± 0.13 Bq/m3 at 75–100 m. The other two stations (at Mostaganem and Annaba) were too shallow (55 and 35 m, respectively), so the subsurface 137 Cs maximum could not be observed. 137 Cs profiles in the water column of the south-western Mediterranean Sea obtained by other authors (Gheddou et al., 1999; Benmansour et al., 2006; Lee et al., 2006) seem to be similar to those obtained in this work. 3.2. Surface sediment Massic activities of 137 Cs in surface sediment ranged from 0.29 ± 0.04 Bq/kg dry weight (dw) to 13.1 ± 0.5 Bq/kg dw, with a mean value of 5.2 ± 0.2 Bq/kg dw. Surface sediments were sampled at different water depths (13–170 m) and different distances from the coast, having a variable grain size and organic carbon content (sand, mud). The variations in 137 Cs massic activities along the Algerian coast are shown in Fig. 3. The mean value of 5.2 ± 0.3 Bq/kg dw obtained for the Algerian coast is in a reasonable agreement with the mean value of 4.6 Bq/kg dw determined for the whole Mediterranean Sea (UNEP, 1992). Massic activities of 40 K in surface sediment samples, which depend on the nature of the sediments, ranged from 107 ± 4 Bq/kg dw to 630 ± 20 Bq/kg dw, with a mean value of 370 ± 20 Bq/kg dw. A weak correlation (R 2 = 0.51) is observed between 40 K and 137 Cs (Fig. 4), possibly due to their similar behaviour in the marine environment. The observed 137 Cs variations in surface sediments (Fig. 3) may be due to different water column depths and differences in physical (e.g. grain size) and chemical (e.g. organic carbon content) characteristics of the sediments.
137 Cs in seawater and sediment along the Algerian coast
161
Fig. 3. Massic activities of 137 Cs in surface sediments collected each 25 km along the Algerian coast (from W to E).
Fig. 4. Correlation between 40 K and 137 Cs massic activities in surface sediments.
3.3. Sediment core Massic activities of 137 Cs in the sediment core collected at Annaba station (36◦ 56 504N, 07◦ 47 750E) ranged from 0.16 ± 0.02 Bq/kg dw (at the bottom of the core at 22 cm) to 12.4 ± 0.7 Bq/kg dw in the second layer (2–4 cm), which forms a peak in the distribution of 137 Cs in the core (Fig. 5). The sedimentation rate of 0.2 cm/yr estimated using the 210 Pb method suggests that the 137 Cs peak presented in Fig. 5 may be due to the Chernobyl accident.
162
A. Noureddine et al.
Fig. 5. 137 Cs in the sediment core collected at Annaba station (36◦ 56 504N, 07◦ 47 750E) compared with the profile of the eastern station (37◦ 12 07N, 08◦ 17 87E) collected in 2004 (water depth 550 m).
3.4.
137 Cs
inventory
Based on the 137 Cs activity concentrations determined in 1999, the corresponding inventory of 137 Cs in the water column along the Algerian coast was estimated to be 86 ± 9 Bq/m2 , 370 ± 20 Bq/m2 , 100 ± 10 Bq/m2 and 280 ± 20 Bq/m2 for Annaba, Algiers, Mostaganem and Ghazaouet stations, respectively. These values are relatively low, as they strongly depend on the water column depth, but are comparable with those obtained by other authors in the south-western Mediterranean when normalised for similar water depths (Lee et al., 2003, 2006). Figure 5 compares the 137 Cs sediment profile in Annaba station with one collected in 2004 at the eastern Algerian coast (37◦ 12 07N, 08◦ 17 87E; water depth 550 m). The profiles are quite similar, higher 137 Cs massic activities observed in surface layers of the eastern core may be due to different grain size composition of the sediment. The 137 Cs inventory in sediment estimated for Annaba station is 1500 ± 30 Bq/m2 , by about a factor of two lower than the total deposition due to global fallout for the latitude belt of 40–60◦ N (UNSCEAR, 2003). Due to shallow water depth almost all 137 Cs has been deposited on sediment. The total inventory (i.e. water and sediment) for Annaba station is 1600 ± 30 Bq/m2 , at least by a factor of two lower than results reported elsewhere in the Mediterranean (e.g. Lee et al., 2003), however, for water depths greater than 1000 m.
137 Cs in seawater and sediment along the Algerian coast
163
4. Conclusions We can conclude that anthropogenic 137 Cs found in seawater and in sediment, along the Algerian coast was affected mostly by global fallout and the Chernobyl accident. The observed 137 Cs activity concentrations in surface water and in water profiles, and the total inventories of 137 Cs in the water column are in a reasonable agreement with those given by other authors. The observed radionuclide levels in surface and core sediments have been dependant on physical characteristics (e.g. grain size) and chemical composition (e.g. organic carbon content) of sediments.
Acknowledgements The authors wish to thank ISMAL Director M. Boulahdid, the scientific team and the crew who participated in the cruise on board of R/V M.S. Benyahia for support and assistance during sampling. Thanks are also due to the IAEA for support provided through the Technical Cooperation project ALG/2/006 (1996–1999). The IAEA is grateful for the support provided to its Marine Environment Laboratory by the Government of the Principality of Monaco.
References Benmansour, M., Laissaoui, A., Benbrahim, S., Ibn Majah, M., Chafik, A., Povinec, P.P. (2006). Distribution of anthropogenic radionuclides in Moroccan coastal waters and sediments. In: Povinec, P.P., Sanchez-Cabeza, J.A. (Eds.), Radionuclides in the Environment, International Conference on Isotopes in Environmental Studies: Aquatic Forum 2004. 25–29 October, Monaco. Elsevier, Amsterdam, pp. 148–155, this volume. Buffoni, G., Cappelletti, A. (1997). On the accumulation–dispersion processes of the 137 Cs tracer in the Italian seas. Journal of Environmental Radioactivity 37, 155–173. Burton, J.D. (1975). Radioactive nuclides in the marine environment. In: Riley, J.P., Skirrow, G.S. (Eds.), second ed. Chemical Oceanography, vol. 3. Academic Press, London. Calmet, D., Grauby, A., (1988). Distribution spatio-temporelle des radioéléments anthropogéniques du bassin Méditerranéen occidental 1980–1987. In: International Conference on Environmental Radioactivity in the Mediterranean Area. Barcelone, 1988, pp. 543–568. Delfanti, R. et al. (2001). The new distribution of the tracer 137 Cs in the eastern Mediterranean relationship to the deepwater transient. Rapport du 36ème Congrès de la CIESM. CIESM, Monte Carlo (Monaco), 36 pp. Gheddou, A., Noureddine, A., Menacer, M., Boudjenoun, R., Hammadi, A. (1999). Distribution of 137 Cs in surface and deep water in the central part of Algerian littoral. Marine Pollution, IAEA-TECDOC-1098. Lee, S.-H., Larosa, J.J., Palomo, I.L., Oregioni, B., Pham, M.K., Povinec, P.P., Wyse, E. (2003). Recent inputs and budgets of 90 Sr, 137 Cs, 239,240 Pu and 241 Am in the northwest Mediterranean Sea. Deep-Sea Research II 50, 2817–2834. Lee, S.-H., Mantoura, F.R., Povinec, P.P., Sanchez-Cabeza, J.A., Pontis, J.-L., Mahjoub, A., Noureddine, A., Boulahdid, L., Samaali, M., Reguigui, N. (2006). Distribution of anthropogenic radionuclides in the water column of the south-western Mediterranean Sea. In: Povinec, P.P., Sanchez-Cabeza, J.A. (Eds.), Radionuclides in the Environment, International Conference on Isotopes in Environmental Studies: Aquatic Forum 2004. 25–29 October, Monaco. Elsevier, Amsterdam, pp. 137–147, this volume. Livingston, H.D.L., Povinec, P.P. (2000). Anthropogenic marine radioactivity. Ocean and Coastal Management 43, 689–712. Povinec, P.P., Pham, M.K. (2000). Report on the Intercomparison Run IAEA-384 Radionuclides in Fangataufa Lagoon Sediment. IAEA/MEL/68. IAEA, Vienna, 46 pp.
164
A. Noureddine et al.
Povinec, P. et al. (2002). Certified reference material for radionuclides in seawater IAEA-381 (Irish sea water). Journal of Radioanalytical and Nuclear Chemistry 251, 369–374. Roos, P., Holm, E., Persson, R.B.R. (1994). Comparison of AMP precipitate method and impregnated Cu2 [Fe(CN)6 ] filters for the determination of radio-caesium concentrations in natural waters. Nuclear Instruments and Methods in Physics Research A 39, 282–286. UNEP (1992). Assessment of the state of pollution of the Mediterranean Sea by radioactive substances. MAP Technical Reports Series 62. Athens, 60 pp. UNSCEAR (2003). Sources and effects of ionising radiation. United Nations, New York, 123 pp.
Further reading Delfanti, R., Papucci, C., Alboni, M., Lorenzelli, R., Salvi, S. (1995). 137 Cs inventories in the water column and in sediments of the west Mediterranean Sea. Rapports Commission Internationale pour l’exploration scientifique de la Mer Méditerranée 34, 226–230. Merino, J., Sanchez-Cabeza, J.A., Bruach, J.M., Masqué, P., Pujol, L. (1997). Artificial radionuclides in a high resolution water column profile from the Catalan Sea (the Northwest Mediterranean). Radioprotection – Colloques 32 (C2), 85–90. Millot, C. (1987). Circulation in Western Mediterranean Sea. Oceanologica Acta 10, 2–8. Molero, J., Sanchez-Cabeza, J.A., Merino, J., Pujol, Ll., Mitchell, P.I., Vidal-Quadras, A. (1995). Vertical distribution of radiocaesium, plutonium and americium in the Catalan Sea (Northwest Mediterranean). Journal of Environmental Radioactivity 26, 205–216. Noureddine, A., Benkrid, M., Hammadi, A., Boudjenoun, R., Menacer, M., Khaber, A., Kecir, M.S. (2003). Radioactivity distribution in surface and core sediment of the central part of the Algerian coast: An estimation of the recent sedimentation rate. Mediterranean Marine Science 4 (2), 53–58. OECD (1971). The cycling of artificial radionuclides through marine food chains. Proc. Marine Radioecology, Hamburg.
165
Physical and chemical characteristics of 137Cs in the Baltic Sea Galina Lujanien˙ea,* , K˛estutis Jokšasb , Beata Šilobritien˙ea , Rasa Mork¯unien˙ec a Institute of Physics, Savanoriu ave 231, LT-02300 Vilnius, Lithuania ˛ b Institute of Geology and Geography, T. Ševˇcenkos 13, LT-2600 Vilnius, Lithuania c Vilnius Gedimino Technical University, Saul˙etekio al. 11, LT-2040 Vilnius, Lithuania
Abstract 137 Cs in seawater, suspended particles and bottom sediments was studied with the aim to better understand its behavior, redistribution and sink in the Lithuanian coastal area of the Baltic Sea. High massic activities of 137 Cs (over
1000 Bq/kg dry weight) associated with fine particles of bottom sediments can be attributed to an increase in the specific surface area and variations in mineralogical composition of sediments participating in sorption processes. A clear correlation was found both between the 137 Cs activity and the content of clay particles (r = 0.95, n = 16), as well as the amount of total organic carbon (r = 0.75, n = 16). 137 Cs activity of particulates depends strongly on the sampling depth of seawater. The study has indicated a complicated sorption–desorption behavior of 137 Cs in the Baltic Sea as a result of which it can be mobilized by suspended particles or released to seawater. The fate of 137 Cs can be considerably affected by mineralogical composition of suspended particles and bottom sediments, as well as by geochemical characteristics of seawater. Keywords: Cesium-137, Seawater, Speciation, Suspended particles, Bottom sediments, Sorption–desorption, Distribution coefficient (Kd ), Baltic Sea
1. Introduction The distribution of radionuclides between the particulate, colloidal and truly dissolved phases can strongly influence transport processes and bioavailability of contaminants in the aquatic systems. Among parameters that control the transport of radionuclides associated with particles, the size distribution of particles carrying radionuclides, their density and chemical composition are the most important. However, in the zone of interaction of fresh and saline waters under salinity and activity concentration gradients, the sorption–desorption process can lead to transformation of nuclide associations. In addition, the removal of particles of higher density and larger size occurring during settling can affect the properties of suspended matter * Corresponding author. Address: Institute of Physics, Savanoriu av. 231, LT-02300 Vilnius, Lithuania; phone: ˛ (+370) 5 264 48 56; fax: (+370) 5 260 23 17; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08011-3
© 2006 Elsevier Ltd. All rights reserved.
166
G. Lujanien˙e et al.
(e.g. mineralogical composition) and can result in different binding of radionuclides and their different removal from the water column. To assess the transfer and fate of radioactive contaminants in the marine environment the information on physical and chemical parameters of radionuclide carriers is required. Radioactive contamination of the Baltic Sea was caused by three main factors such as global fallout, discharges from the reprocessing plants and the fallout after the Chernobyl accident in April 1986. At present, the average activity concentration of 137 Cs in surface water of the Baltic Sea has been estimated to be about 60 mBq/L, while the worldwide average concentration due to global fallout is about 2 mBq/L (Livingston and Povinec, 2000). The Baltic Sea is a semi-enclosed, largest brackish water and unique shallow sea in which self-cleaning processes are slow and dissolved substances remain there for a long time. Measurements of total activity concentrations of 137 Cs as well as other artificial and natural radionuclides in the Baltic Sea water have been carried out in many riparian countries (Nielsen, 1997; Nielsen et al., 1999). However, there is a lack of information on the speciation of radionuclides in the Baltic Sea, their bioavailability, migration and on self-cleaning processes. There is no doubt that consistent pattern of radionuclide migration, processes of selfcleaning and their redistribution in the environment depend on their specific chemical forms or type of binding rather than on the total element or nuclide content. However, the direct determination of speciation or binding forms of radionuclides is difficult and often hardly possible due to very low concentrations of radionuclides typically found in the environment. Therefore, the determination of physico-chemical forms or fractions in practice, using sequential extraction methods, is a reasonable compromise to evaluate the associations of radionuclides in the environmental samples (e.g. with carbonate minerals, Fe/Mn oxides and organic substances). Although the results obtained by a sequential extraction procedure are operationally defined and cannot be used as input data for thermodynamic equilibrium models, they provide useful information about the behavior of radionuclides. Knowledge about association of radionuclides with geochemical phases is important from the point of view of radiation protection, as binding of radionuclides and the stability of geochemical phases provide data on bioavailability and migration ability of radionuclides. At present, the only available tool to study radionuclide associations at low-level concentrations is sequential extraction method. These methods were criticized because of incomplete selectivity of reagents used to dissolve one particular phase without the additional attack of other geochemical phases and readsorption of analytes after release. The results obtained by Khebonian and Bauer (1987) using synthetic models and their interpretation were intensively discussed since the appearance of their publication. Re-adsorption processes of some elements were observed for iron and organic rich sediments, and non-selective extraction was found during extraction of elements from anoxic sediments. Recently obvious advantages were achieved in analyses of speciation of heavy metals (Sahuquillo et al., 2003). The most widely used sequential procedure proposed by Tessier et al. (1979) was modified in order to avoid the mentioned problems by changing the extraction time, the extractant-sample ratio, the reagent concentration ant the extraction temperature. The aim of this study was to determine speciation of 137 Cs suspended particles and bottom sediments in order to better understand its behavior, redistribution and sink in the Lithuanian coastal area of the Baltic Sea.
Physical and chemical characteristics of 137 Cs in the Baltic Sea
167
2. Material and methods 2.1. Sampling Water and bottom sediment samples were collected during the expedition in the Lithuanian economical zone of the Baltic Sea, the Curonian Lagoon and near the seashore of the Baltic Sea and the Curonian Lagoon in 1999–2001. The sampling locations are shown in Fig. 1. The studied area is characterized by horizontal and vertical salinity gradient: (i) salinity ranges from 0.5h to 7h in surface waters and from 5h to 9h in near bottom waters; (ii) activity concentrations of 137 Cs vary from 1 mBq/L in fresh waters up to 100 mBq/L in seawater. Measurements performed in the Institute of Geology and Geography indicated a strong loading by particulate matter: in the Curonian Lagoon the concentration of suspended particles differed from 1 to 103 mg/L, in the Baltic near-mouth – from 2 to 29 mg/L and in the Baltic Sea open waters – from 1 to 41 mg/L (Galkus and Joksas, 1997). The seasonal variations are mostly related to plankton blooming periods. It should be noted that all studies were performed in the transitional–accumulation zone with a complex current regime. The intrusions of Baltic Sea water into the Curonian Lagoon caused by differences in the water level, different hydrometeorogical conditions and various anthropogenic activities can influence the accumulation of 137 Cs in the bottom sediments. The amount of sampled water was depended on the 137 Cs detection limit, and varied from 50 to 200 L. In addition, the suspended particle samples were collected in situ by filtering a large volume of water (∼400–1000 L) through two consecutive 5 µm and two consecutive 1 µm polypropylene Sediment Filter Cartridges (US Filter Plymouth Products). Bottom sediment samples were collected during different sampling campaigns in 1999– 2001. The bottom sediments in the Baltic Sea were collected using a Van Veen grab sampler available on the R/V “V˙ejas”. The Bottom Sampler acc. Ekman–Birge with an effective grasping area of 225 cm2 and weight of 3.5 kg was used for the bottom sediment sampling in the Curonian Lagoon. 2.2. Sample preparation The separation of the suspended matter was achieved using membrane filtration. Water (from 50 to 800 L) was filtered through Nuclepore (Dubna) membrane 0.2 µm and Filtrak 388 prefilters using the filtration equipment (Millipore) consisting of the Dispensing Pressure Vessel of 10 L and a Stainless Steel Filter Holder of 293 mm. It is very well known that the separation of suspended particles of 1–0.2 µm from large volume samples is a very time consuming and artefacts prone process, e.g. the size of separated particles can be changed significantly due to clogging (Salbu et al., 1985). To avoid the clogging effect the filtration process was carefully controlled and filters were changed on time. The activity of 137 Cs in some samples of suspended matter was close to detection limits for gamma-spectrometric detection, therefore cesium was separated radiochemically. The samples were digested using HF/HNO3 and HCl, and then cesium was precipitated as Cs3 Sb2 I9 . Chemical yield was determined gravimetrically.
168
G. Lujanien˙e et al.
Fig. 1. Sampling locations in the Baltic Sea and the Curonian Lagoon.
Physical and chemical characteristics of 137 Cs in the Baltic Sea
169
To determine aqueous speciation the filtered water passed through four subsequent cartridges filled with the impregnated Mtilon-T fiber. The Mtilon-T fiber developed in Moscow Textile Institute by modification of cellulose and polyacrylonitrile copolymer having thioamide functional groups is widely used for cesium pre-concentration (Vakulovsky et al., 1985). The Mtilon-T fibre impregnated with Ni Ferro cyanide (the efficiency – 99%) was used to separate ionic and non-ionic Cs species in seawater. The dependence of the preconcentration efficiency of 137 Cs from seawater on pH, temperature and the flow rate was studied. In order to provide a complete collection of ionic 137 Cs from seawater, four consecutive cartridges were used. The details of the separation method were described previously (Lujanien˙e et al., 1998). The ionic fraction of 137 Cs sorbed on the impregnated Mtilon-T was determined directly by gamma-spectrometry, while the fraction passing through cartridges so-called “non-ionic 137 Cs fraction” was separated from seawater radiochemically prior to measurements. For the determination of physical and chemical association of radionuclides, a separation of particles of different sizes from bottom sediment samples was performed using wet sieving and column settling techniques. The fractions >50 µm, 50–4 µm and <4 µm were separated and characterized as sand, silt and clay particles. In addition, a limited number of samples were separated into four fractions >50 µm, 50–4 µm, 4–1 µm and 1–0.2 µm. The separation of 1–0.2 µm fraction was achieved using membrane filtration. The modified Tessier sequential extraction method was used for the investigation of radionuclide chemical forms. The following extracting agents were used: F1 – 1 M NH4 Cl, pH = 7.0 (exchangeable); F2 – 1 M NH4 C2 H3 O2 , pH = 5 CH3 COOH (carbonate bound); F3 – 0.04 M NH2 OHHCl in 25% CH3 COOH (oxide bound); F4 – 30% H2 O2 at pH = 2 (HNO3 ), then 3.2 M NH4 C2 H3 O2 in 20% HNO3 (organically bound); F5 – 40% HF, HNO3 , or measured directly by gamma-spectrometry (residual). 2.3. Measurements Total carbon (TC) and TOC were determined using a LECO CS-125 analyzer. 133 Cs concentration in seawater was determined using ICP-MS. Grain size distribution was measured by the gravimetric pipette method. Clay minerals were identified by X-ray diffraction. XRD analyses were conducted using a D8 (Bruker AXS) X-ray diffractometer. Cs from suspended matter samples separated radiochemically was measured using the proportional Emberline FHT 770 T Multi-Low-Level-Counter. 137 Cs activities were measured with HPGe detector (resolution 1.9 keV at 1.33 MeV, the relative efficiency 42%). Measuring time varied according to sample activities. An efficiency calibration of the system was performed using calibration sources (prepared from a solution supplied by Amersham, UK) of different densities and geometry that were close to measured samples. Accuracy and precision of analysis was tested in intercomparison runs, organized by the Riso National Laboratory, Denmark (mineral, sea and lake water matrices). Precision of 137 Cs measurements by gamma-spectrometry was <10%. 137 Cs activities derived from betacounting had uncertainties from less than 10–20%, depending on the activity of the measured source.
170
G. Lujanien˙e et al.
3. Results and discussion 3.1. Aqueous speciation of 137 Cs Activity concentrations of 137 Cs in Baltic Sea water ranged from 14 to 100 mBq/L, and in Curonian Lagoon water they varied from 1 to 67 mBq/L. It should be noted that sampling was mainly performed in the zone of input of fresh water from the Curonian Lagoon, and a correlation between the 137 Cs activity concentration and salinity was observed (r = 0.76). Monitoring carried out at stations 4, 7, 46a and 65 in 1997–2002 by the Ministry of Environment also indicated a correlation of activity concentration of 137 Cs with salinity (r = 0.73) (Lujanien˙e et al., 2004). Data on 137 Cs speciation in Baltic Sea water obtained during expeditions in 1995–1997 show seasonal variations of 137 Cs forms: particulate, ionic and non-ionic, which probably consists of pseudocolloids and complexes of different origin (Table 1). Activity concentrations of 137 Cs as ionic species varied from 50 to 90 mBq/L and were found to be highest in autumn. The studies of aqueous speciation indicated that in Baltic Sea water soluble cesium constituted 70–99% of the total 137 Cs, and it was present in the ionic state (from 62% to 92%) most likely as Cs+ , both in the near-shore and in open waters. The amount of particulate cesium varied between 1% and 30% (Table 1). It was determined that up to 20% of 137 Cs was present in a non-ionic state, and it cannot participate in the exchange process since it is possibly associated with particles <0.2 µm in size. 3.2. Association of 137 Cs with suspended particles Suspended particles in near-shore and open seawaters of the Baltic Sea and the Curonian Lagoon were collected during three expeditions. The massic activity of 137 Cs in suspended particle samples collected in the Curonian Lagoon and separated using the 0.2 µm membrane ranged from 20 to 250 Bq/kg dry weight (dw). Insignificantly higher activities of 137 Cs from 20 to 370 Bq/kg dw were found in the near-shore zone of the Baltic Sea. The activities of 137 Cs in suspended particles from surface and bottom waters of the Baltic Sea differed from 80 to 840 Bq/kg dw and from 90 to 970 Bq/kg dw, respectively. Data on massic activity of particulate 137 Cs in surface water samples collected during different expeditions in the Curonian Lagoon and the Baltic Sea plotted against the concentration of suspended particles are presented in Fig. 2. The higher activity of 137 Cs associated with particles >0.2 µm corresponds to a small amount of particulate matter in the samples. It should be noted that concentrations Table 1 Seasonal variations of 137 Cs speciation (%) in seawater of the Baltic Sea near-shore Season
Particulate
Ionic
Non-ionic
Spring Summer Autumn Winter
8 ± 0.8 31 ± 2 7 ± 0.8 26 ± 2
74 ± 5 66 ± 4 87 ± 6 59 ± 5
18 ± 1.5 9 ± 0.9 14 ± 1.5 7±1
Physical and chemical characteristics of 137 Cs in the Baltic Sea
171
Fig. 2. Activity of particulate 137 Cs versus concentration of suspended particles in surface water samples.
Fig. 3. Activity of particulate (SP) and water soluble 137 Cs in Baltic Sea water (May 1999).
of particulate matter in samples collected in open waters of the Baltic Sea were lower and ranged from 1 to 11 mg/mL, and the highest activities of 137 Cs in suspended particles were found at the most remote station, about 100 km from the shore. Samples of particulate matter collected in open waters of the Baltic Sea had a wider range of 137 Cs activities. The massic activity of particulate 137 Cs in surface samples were lower as compared to the samples collected near bottom at all sampling stations during the sampling campaign in May 1999 (Fig. 3). A strong relation of massic activity of particulate 137 Cs to the sampling depth was found with the exception of the sample collected at the station 6b, which possibly could be explained by water circulation in this area (Fig. 1). This station is located at the crease of the Nemunas River and a different pattern of radionuclide behavior was expected to be characteristic for this area. The differences between the massic activities of particulate 137 Cs in surface and near-bottom samples increased with increasing sampling depth as well. The particular “excess” of 137 Cs activity on the near-bottom particles in comparison with the surface suspended matter is a function of the sampling depth, the activity concentration of 137 Cs in water and water salinity (Fig. 4). A direct correlation between the “excess” of activity of particulate 137 Cs and the sampling depth was found (r = 0.96, n = 6). We suppose that the
172
G. Lujanien˙e et al.
Fig. 4. Excess of 137 Cs activity on near-bottom suspended particles as compared to surface ones versus activity concentration in water and 137 Cs Kd for near-bottom samples versus sampling depth (May 1999).
increase in the activity of 137 Cs associated with suspended particles can be a result of sorption process during particle transport and settling in seawater, and this “excess” is a function of particle residence time in seawater. The accumulation of 137 Cs on suspended particles leads to an increasing of 137 Cs Kd values with the sampling depth (Fig. 4). In samples collected using the sediment filter cartridges with the nominal pore size of 1 µm the comparatively lower activities of particulate 137 Cs (from 20 to 60 Bq/kg dw) in surface water samples were observed. The differences in activities can be attributed to the different origin of suspended particles due to seasonal variations (the sampling was performed in October) or/and to the losses of small particles during filtration. However, it seems that suspended particles of the 0.2–1 µm size can considerably contribute to the total particulate activity of 137 Cs. An increase in activity of 137 Cs accompanied by an increase in concentration of suspended particles was observed in samples collected in October 2001 (Fig. 5). In addition, activity of particulate 137 Cs decreased with an increase of salinity and concentration of potassium (Fig. 5), indicating that a desorption of 137 Cs from particles >1 µm in size could take place. Data of different sampling campaign indicated the losses of large portion of fine particles up to 30% in the near-shore zone using the filters of 1 µm as compared with filters of the 0.2 µm pore size. In addition, this portion increased in open waters where fine particles are predominant. During transport from the Curonian Lagoon in the saline and fresh water mixing zone an intensive coagulation and deposition of particles take place (Fig. 6). However, not only a fractionation of suspended particles according to sizes (a decrease of suspended particles >1 µm and an increase in the amount of particles of 0.2–1 µm carrying 137 Cs depending on the distance from the Klaip˙eda Strait), but also changes in the chemical/mineralogical composition of suspended particles are possible. It seems that particles of various sizes and mineralogical composition can display different sorption ability. From particles of the >1 µm size 137 Cs can be released into seawater, on contrary, the fine fraction of suspended particles of 0.2–1 µm and possibly smaller, is responsible for sorption and removal of 137 Cs from the water column. Thus, different sorption–desorption behavior of cesium can be explained by different nature of particles. Possibly, the fine fraction of suspended particles collected in May 1999 consists mainly of clay minerals which can more effectively adsorb Cs even from brackish
Physical and chemical characteristics of 137 Cs in the Baltic Sea
173
Fig. 5. Activity of particulate 137 Cs versus concentration of suspended particles, salinity and concentration of potassium (October 2001).
Fig. 6. Variations of activity of particulate 137 Cs (Bq/kg), concentration of suspended particles (mg/mL) depending on the distance from the shore and the Klaip˙eda Strait.
seawater containing high concentrations of potassium. The clay minerals of Muscovite group can be responsible for an increase in the activity of particulate 137 Cs during particle transport and settling in seawater. The collected suspended matter exhibited rather different sorption ability due to different cut-off filters used during the sampling performed in October 2001. However, sorption–desorption behavior of 137 Cs depends not only on the origin of suspended particles, but also on the concentration of ions, which can compete with 137 Cs for sorption places. Thus, information about the concentration of stable cesium (133 Cs) at the site is also important for quantitative assessment of accumulation of radiocesium on suspended particles, because the concentration of 133 Cs is usually higher by 6–8 orders of magnitude in comparison with 137 Cs. During the sampling campaign in October 2001 the measured concentrations of 133 Cs ranged from 0.03 to 0.05 ppb, nevertheless, any relation of massic activity of particulate 137 Cs to the concentration of 133 Cs in seawater was not observed. Perhaps, high concentrations of potassium (which varied from 50,900 to 75,700 ppb) smothered the influence of 133 Cs. Another possible explanation could be in differences in speciation and a lack of equilibrium of 133 Cs and 137 Cs isotopes in seawater. If the equilibrium between 133 Cs and
174
G. Lujanien˙e et al.
137 Cs
isotopes has not been reached yet (after the Chernobyl accident), the quantitative assessment and applications of laboratory experiments for radiocesium transport are difficult because of absence of data on real Cs concentrations in seawater. The selective adsorption of Cs by clay minerals has been attributed to the large ionic radius, uncomplicated nature, and especially to its low hydration energy. Although cations with similar charge and ionic radii are expected to compete with cesium, the sequence of sorption ability of alkali elements Na+ < K+ < R+ < Cs+ is in good agreement with the sequence of effective ionic radii of alkali elements and the sequence of single ion hydration enthalpies of alkali elements (Richens, 1997). In the clay mineral muscovite, a negative fixed charge arises primarily from isomorphic substitution of Al3+ for Si4+ in the tetrahedral sheet comprising the siloxane site. In illite mineral whose composition is very close to that of muscovite, isomorphic substitution of Al3+ for Si4+ and partly of Fe2+ and Mg2+ for Al3+ enhances the stability of the Cs+ -siloxane surface complex (Jackson, 1962). Strongly sorbed Cs+ on fraered edge sites (FES), external basalt sites, or within the interlayer exists as an inner-sphere, dehydrated surface complex, which usually is much more stable than outer-sphere complexes (Kemner et al., 1997). Cesium sorbed to outer-sphere complexes can be easily desorbed and is distinguished for higher mobility in the environment, while inner-sphere sorption complexes can limit the Cs transport and bioavailability. In smectites, the isomorphous substitution in both tetrahedral and octahedral layers generates weak negative charges of sheets thus leading to the formation of a structural feature and resulting in wide ranges of cation exchange capacity, selectivity and swelling properties. Water and cations such as H+ , Na+ , Ca2+ , Mg2+ can easily penetrate into the smectite interlayer and participate in exchange processes and, for example, from a species as beidellite the sorbed cation can easily be ion-exchanged by the other cations. However, the existence of electro statically stable sorption site for a cation whose ionic size is small enough to enter the hole was estimated for montmorillonite (Onodera et al., 1998). The complicated behavior of cesium can be the result of peculiarities of its sorption to various sorption sites of different clay minerals and their mixtures. The specific surface area and mineralogical composition of particles of different sizes can differ considerably and can result in the variation of binding ability of suspended particles. 3.3. Association of 137 Cs with bottom sediments Massic activities of 137 Cs in bottom sediments of the Baltic Sea and the Curonian Lagoon ranged from 4 to 450 and from 0.4 to 596 Bq/kg dw, respectively. Studies performed on the particle size distribution and 137 Cs activities in bottom sediments indicated variations in 137 Cs activities associated with particles (62–1450 Bq/kg dw for particles 0.2–1 µm (n = 6), 62–836 Bq/kg dw for particles 1–4 µm (n = 20), 29–242 Bq/kg dw for particles 4–50 µm (n = 20) and 3–64 Bq/kg dw for particles >50 µm (n = 20). The data on the particle size distribution in six sediment samples are presented in Fig. 7. An increase in the massic activities of 137 Cs (over 1000 Bq/kg dw) observed in fine particles can be due to the increase in the specific surface area of sediments participating in sorption processes. Another possible explanation could be associated with variations in the mineralogical composition of different size particles. The presence of illite (hidromuskovite) mineral whose characteristic particles size distribution is from 0.002 to 2.9 µm, and smectite
Physical and chemical characteristics of 137 Cs in the Baltic Sea
175
Fig. 7. Massic activities of 137 Cs in bottom sediments and average physico-chemical characteristics of 137 Cs in bottom sediments of the Baltic Sea (F1 – exchangeable, F2 – carbonates, F3 – oxides, F4 – organic, F5 – residue).
with typical particles size distribution from 0.01 to 1.0 µm can be expected in fine fractions (Nelson, 2003). It seems therefore that the ability of bottom sediments to bind cesium varied with the particles size due to their different mineralogical composition. Moreover, not only an increase in 137 Cs activity of sediments with a decrease in particle sizes was observed, but also wide ranges of 137 Cs activities were found for smaller particles. Possibly heterogeneous origin of sediments is responsible for wide ranges of 137 Cs activities in fine fraction. It should be noted that a comparatively smaller amount of fine particles of <4 µm was found in samples collected in the Baltic Sea, where it ranged from 1% to 28% (on average – 6%), in comparison with Curonian Lagoon bottom sediment samples where it varied from 11% to 70% (with average value of 20%). The correlation of 137 Cs activity in bottom sediments with the amount of clay (r = 0.95, n = 16) and organic substances (r = 0.75, n = 16) was found (Lujanien˙e et al., 2004). Studies of 137 Cs geochemical fractionation performed on bottom sediments collected in the Curonian Lagoon indicated insignificant variations in partitioning of cesium between different geochemical fractions in the samples collected at different stations. In samples collected in the Baltic Sea, a wider range of fraction distribution of cesium was observed. In these samples higher percentage of 137 Cs associated with the exchangeable and the residual fraction was found. In general, the fraction distribution studies indicated a strong association of 137 Cs with sediment particles – about 70% of 137 Cs was found in residual fractions (Fig. 7), whereas the organic fraction comprised a small part of total activity of 137 Cs in the sediments. Thus, despite the strong correlation of 137 Cs activities with the total organic carbon (r = 0.75) and with the content of clay (r = 0.95), the geochemical fractionation indicated its strong association with the residual fraction, whereas the organic fraction comprised a small part of the total activity of 137 Cs in bottom sediments. This result is in good agreement with other investigations indicating that clay minerals, especially illite, can effectively immobilize cesium even in the presence of a high content of organic substances (Dumat and Staunton, 1999; Evans et al., 1983). Organic substances can inhibit sorption of Cs on clay minerals and thus increase its mobility in the environment, but perhaps they can also enhance a binding of cesium by minerals. A correlation between the total 137 Cs activities and organic substances usually found in the environmental samples confirms this. However, the mechanism of this phenomenon is not clear.
176
G. Lujanien˙e et al.
Average data on the association of 137 Cs with various geochemical fractions of particles of different sizes collected in the Baltic Sea are presented in Fig. 7. The geochemical partitioning of 137 Cs determined in the fine <4 µm size fraction can be attributed to the binding pattern of 137 Cs to clay minerals with specific sorption on the interlayer sites. The increase in percentage of exchangeable 137 Cs in the 4–50 µm size fraction of sediments can be explained by an increase in a specific surface area and in the number of sites available for exchange reactions on particles of this size range. The explanation of a comparatively high content of 137 Cs in the residual fraction of particles >50 µm and 4–50 µm, that is usually related to the fixation of Cs in the clay mineral crystal lattice, can be the coagulation or sticking of fine clay to the coarse sand and silt particles. The sticking particles present on the coarse can affect size-dependent distribution of radionuclide activities in soil and sediments observed in most cases (Ewais et al., 2000). It should be noted that analysis of bottom sediments using XRD indicated the presence of muscovite group clay minerals in the samples. Moreover, an interdependence of the muscovite mineral amount with 137 Cs activity of bottom sediment samples was found. The illite mineral was not identified due to technical reasons, and it is not clear whether it is present in the samples. Data on the size distribution of particles and speciation were used to calculate 137 Cs distribution coefficients (Kd ). The calculated Kd values have been increasing with decreasing particle size (Fig. 8). The Kd s were calculated for different geochemical fractions and the highest values were obtained for the residual fraction of particles smaller than 4 µm. These results and data obtained from two laboratory experiments indicated that residual fraction, which reflects the Cs sorption on clay minerals, is the most important in incorporating Cs to bottom sediments, while the role of other geochemical phases is relevant only during two months of the sorption experiment (Lujanien˙e et al., 2005). The laboratory conditions in many cases cannot correspond to the field ones and Kd s can differ considerably because of different environment conditions. In addition, the redistribution of radionuclides between different components of ecosystems is a time-dependent processes. In order to better understand the influence of site specific conditions we compared the 134 Cs
Fig. 8. Dependence of 137 Cs Kd on particle size and geochemical fraction of bottom sediments.
Physical and chemical characteristics of 137 Cs in the Baltic Sea
177
Kd s obtained from kinetic tracer experiments performed using the bottom sediments collected in the Curonian Lagoon and seawater labeled by 134 Cs (Lujanien˙e et al., 2005) with the 137 Cs Kd s calculated using data on the activity of 137 Cs in water and suspended particle samples collected in May 1999 (Fig. 3). The kinetic tracer experiments indicated that sediment particles can display intensive and selective sorption towards the Cs ions in seawater. It should be noted the results of these experiments were in good agreement with those published by Borretzen and Salbu (2002). It was supposed that the similar suspended particles are transported to the Baltic Sea from the Curonian Lagoon and can adsorb 137 Cs from seawater. On the other hand data obtained during the expedition in May 1999 indicated higher 137 Cs activities on particles >0.2 µm in near bottom samples in comparison with surface ones, and the differences (or “excess”) between the surface and near bottom activities increases with sampling depth. The 137 Cs Kd values changed in the similar way (Fig. 4) and the “excess” of Kd s could be regarded as a function of residences time of particles in the water column. The residence time of clay particles of density 2.7 g/cm3 and the average diameter (which varied from 5 to 0.5 µm depending on the distance from the shore) was evaluated using differences in the sampling depth and data on the sinking velocity of particles in a non-turbulent water column at 10◦ C (Ruchin, 1957). The 134 Cs Kd and 137 Cs Kd values obtained from two kinetic and one field experiments as function of time are presented in Fig. 9. It should be noted that the residence time of particles was estimated rather approximately due to very complex natural conditions and variability of different parameters. It is obvious, that size distribution of particles should be evaluated more precisely as well. Larger particles carrying 137 Cs are predominant in the near-shore zone and their residence time decreases, but at the remote sampling stations the spectrum of particle sizes changes towards the smaller ones (Fig. 6), and therefore they are sinking with lower velocities. In addition, turbulences in natural systems will reduce sinking velocities. Moreover, a halocline can serve as a particular trap for settling particles and increasing their residence time in the near bottom layer, where higher salinity and activity concentrations of 137 Cs can affect the sorption of 137 Cs to suspended particles. It should be noted that activity concentrations of 137 Cs in surface and near-bottom water sam-
Fig. 9. Cs Kd values as a function of time (data of two laboratory experiments (I exp., II exp.) and field experiment in the Baltic Sea (BS) in May 1999).
178
G. Lujanien˙e et al.
ples varied insignificant during the sampling campaign of May 1999. Usually the explanation of increase in activity concentrations of 137 Cs in near-bottom layer is linked to the anoxic conditions, and the diffusion of Cs ions from the bottom sediments. However, anoxic conditions are not characteristic for the study area, and an increase in concentrations of NH4 + ions (which are formed under anoxic conditions and are responsible for desorption of Cs from FES of clay minerals) in the near bottom layer was not observed (Jokšas, 2005). Thus, despite of complexity of natural conditions and variability of different parameters, this comparison indicated that the Cs Kd values obtained from various experiments differ insignificantly. The observed alterations can be mainly attributed to the differences in the size distribution of particles participating in the sorption process, to the variations in Cs concentrations in the kinetic experiments and turbulences in natural water systems. It seems that the 137 Cs K values obtained from measurements carried out in May 1999 reflect (with mentioned d uncertainties) sorption processes, however, 137 Cs Kd values calculated using the 137 Cs activity in bottom sediments can be attributed to the complicated long-term processes. Possibly, Kd data obtained from laboratory experiments can be used for short-term predictions (up to one year), but Kd values from field measurements are more suitable for appropriate long-term predictions.
4. Conclusions Our study has indicated complicated sorption–desorption behavior of 137 Cs in Baltic Sea waters as a result of which it can be mobilized by suspended particles or released to seawater. Thereby, the fate of 137 Cs can be considerably affected by mineralogical composition of suspended and bottom sediments, by characteristic geochemistry of seawater and the presence of other contaminants. Small 1–4 and 1–0.2 µm particles can play an important role in adsorption of 137 Cs from seawater. Particles distinguished for high specific adsorption ability can effectively remove Cs from water column; they can be easily re-suspended and transported over long distances.
Acknowledgements This work was performed under the auspices of IAEA under project LIT/7/002. In addition, we would like to express our gratitude to all project participants as well as to Prof. P.P. Povinec for encouragement, to Riso National Laboratory and Dr. Sven P. Nielsen for support.
References Borretzen, P., Salbu, B. (2002). Fixation of Cs to marine sediments estimated by a stochastic modelling approach. Journal of Environmental Radioactivity 61, 1–20. Dumat, C., Staunton, S. (1999). Reduced adsorption of caesium on clay minerals caused by various humic substances. Journal of Environmental Radioactivity 46, 187–195. Evans, D.W., Alberts, J.J., Clark III, R.A. (1983). Reversible ion-exchange fixation of cesium-137 leading to mobilization from reservoir sediments. Geochimica et Cosmochimica Acta 47, 1041–1049.
Physical and chemical characteristics of 137 Cs in the Baltic Sea
179
Ewais, T.A., Grant, A., Fattah, A.T.A. (2000). The role of surface coatings on sediments in sediment: Water partitioning of trace elements and radionuclides. Journal of Environmental Radioactivity 49, 55–64. Galkus, A., Joksas, K. (1997). Sedimentary Material in the Aquatic Transition Zone. Institute of Geography, Vilnius, 198 pp. (In Lithuanian.) Jackson, M.L. (1962). Interlaying of expansible layer silicates in soils by chemical weathering. In: Trans. 11th Nat. Conf. Clays Clay Min., pp. 29–46. Jokšas, K. (2005). Unpublished data. Livingston, H.D., Povinec, P.P. (2000). Anthropogenic marine radioactivity. Ocean & Coastal Management 43, 689– 712. Lujanien˙e, G., Mork¯unien˙e, R., Styra, B. (1998). Speciation of 137 Cs in the Baltic Sea water. Environmental Physics 1, 34–42. Lujanien˙e, G., Šilobritien˙e, B., Jokšas, K., Mork¯unien˙e, R. (2004). Behaviour of radiocesium in marine environment. Environmental Research, Engineering and Management 2 (28), 23–32. Lujanien˙e, G., Vilimait˙e-Šilobritien˙e, B., Jokšas, K. (2005). Accumulation of 137 Cs in bottom sediments of the Curonian Lagoon. Nukleonika 50 (1), 23–29. Nelson, S.A. (2003). Clay minerals. Geology, Mineralogy 211. Nielsen, S.P. (1997). Comparison between predicted and observed levels of 137 Cs and 90 Sr in the Baltic Sea. Radioprotection – Colloques 32, C2-387–C2-394. Nielsen, S.P., Bengtson, P., Bojanowsky, R., Hagel, P., Herrmann, J., Ilus, E., Jakobson, E., Motiejunas, S., Panteleev, Y., Skujina, A., Suplinska, M. (1999). The radiological exposure of man from radioactivity in the Baltic Sea. The Science of the Total Environment 237/238, 133–141. Kemner, K.M., Hunter, D.B., Bertsch, P.M., Kirkland, J.P., Elam, W.T. (1997). Determination of site-specific binding environments of surface sorbed cesium on clay minerals by Cs-EXAFS. Journal de Physique IV 7, 777–779. Khebonian, C., Bauer, C.F. (1987). Accuracy of selective extraction procedures for metal speciation in model aquatic sediments. Analytical Chemistry 59, 1417–1423. Onodera, Y., Iwasaki, T., Ebina, T., Hayashi, H., Torii, K., Chatterjee, A., Mimura, H. (1998). Effect of layer charge on fixation of cesium ions in smectites. Journal of Contaminant Hydrology 35, 131–140. Richens, D.T. (1997). The Chemistry of Aqua Ions, Syntheses, Structure and Reactivity. Wiley, Chichester. Appendices 2 and 3, 5. Ruchin, L.B. (1957). Methods of investigation of sedimentary material. Institute of Geology of Academy of Science USSR, vol. 1, 611 pp. (In Russian.) Sahuquillo, A., Rigol, A., Rauret, G. (2003). Overview of the use of leaching/extraction tests for risk assessment of trace metals in contaminated soils and sediments. Trends in Analytical Chemistry 22, 152–159. Salbu, B., Bjornstad, H.E., Lindstrom, N.S., Lydersen, E., Brevik, E.M., Rambaek, J.P., Paus, P.E. (1985). Size fractionation techniques in the determination of elements associated with particulate or colloidal material in natural fresh waters. Talanta 32, 907–913. Tessier, A., Campbell, P.G.C., Bisson, M. (1979). Sequential extraction procedure for the speciation of particulate trace metals. Analytical Chemistry 51, 844–851. Vakulovsky, S.M., Lishevskaja, M.O., Nikitin, A.I., Chumichiov, V.B., Shkurko, V.N. (1985). Method of preconcentration of radiocaesium from sea water using fiber sorbents. Trudy Gosudarstvennogo Okeanograficheskogo Instituta 174, 83–88.
Further reading von Gunten, H.R., Benes, P. (1995). Speciation of radionuclides in the environment. Radiochimica Acta 69, 1–29.
This page intentionally left blank
4. Radioecological studies
This page intentionally left blank
183
Comparison of the MARINA II dispersion model with CSERAM for estimating concentrations of radionuclides in UK waters Kamaljit Sihraa,* , Antony Bexona , John Aldridgeb a National Radiological Protection Board, Chilton, Didcot, Oxon, OX11 0RQ, UK b The Centre for Environment, Fisheries and Aquaculture Science, Lowestoft Laboratory, Pakefield Road,
Lowestoft, Suffolk, NR33 0HT, UK Abstract A strategy was agreed in 1998 by the OSPAR Commission to achieve, by 2020, near zero concentrations of anthropogenic pollutants and near to background levels for pollutants that also occur naturally. One of the uses of the MARINA II model was to test the feasibility of this strategy, given inherited activity concentrations. The model has been validated extensively within the OSPAR region using spatial measurements of nine radionuclides, including 99 Tc, 137 Cs and 239/240 Pu, over the period 1990–2000. Typically the model is found to agree with measurements to within a factor of three, with a marginal spatial bias towards underestimating activity concentrations. In this study, future estimates of activity concentration were tested in a comparison to CSERAM, a high-resolution, physically based model of the Irish Sea, for Sellafield discharges of 137 Cs and 239/240 Pu. Both models show good agreement to the year 2000, although differences of up to an order of magnitude can be observed close to the discharging source by 2020. Typically both models agree to within a factor of five elsewhere by 2020. This study illustrates that MARINA II is a computationally inexpensive but effective tool for calculations of activity concentration in radiation protection. Keywords: MARINA II, CSERAM, Marine modelling, Model comparison, OSPAR
1. Introduction A strategy was agreed in 1998 by the OSPAR Commission (OSPAR, 1998) to achieve, by 2020, close to zero concentrations of anthropogenic radioactive substances and close to background levels for pollutants that occur naturally. To determine whether or not this is achievable requires a comprehensive and robust marine model, together with knowledge of annual radioactive discharges to the marine environment. By modelling several different future discharge strategies to 2020, including a scenario where discharges stop in the year 2000, it is possible to estimate the contribution of historic discharges to predicted activity concen* Corresponding author. Address: National Radiological Protection Board, Chilton, Didcot, Oxon, OX11 0RQ, UK; phone: (+44) 1235 831600; fax: (+44) 1235 833891; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08012-5
© 2006 Elsevier Ltd. All rights reserved.
184
K. Sihra et al.
trations. Marine models, such as the MARINA II model, can be used to estimate activity concentrations for this purpose. MARINA II is a 72-compartment marine model of North European waters. It was developed from the MARINA model (CEC, 1990) in order to provide: better spatial resolution in the English Channel, Atlantic Ocean and Arctic Ocean; more realistic flows between compartments in the Irish Sea; a refined sediment model to improve remobilisation of radionuclides from the sediment layer to the water column. Activity concentrations are calculated annually; therefore seasonal effects are not resolved. However, the model is computationally efficient, a requirement necessary for radiation protection calculations where radiation doses are estimated over long periods. A brief description of the MARINA II model is given in Section 2; a more complete description of the model and its validation is given elsewhere (European Commission, 2002). Compartment models are typically tuned to observational data and therefore are most reliable over the time periods for which data are available. To test the model outside these conditions, MARINA II was compared to CSERAM, a physical, process-based, high-resolution model of the Irish Sea (Aldridge et al., 2003). This model attempts to describe the underlying physical processes directly, an approach that is clearly preferable, where data are available, to model dispersion over short timescales where physical effects, such as tides, need to be resolved. For longer term, radiation protection calculations however, the computational cost for routine multi-decadal estimates of activity concentrations are likely to be prohibitively expensive. The CSERAM model is significantly different in modelling philosophy and should therefore provide a useful insight into both the effectiveness and the limitations of the MARINA II model. The discharge scenario used in the comparison was of actual historical discharges of 239/240 Pu and 137 Cs from Sellafield from the 1950s until the end of 1999 (European Commission, 2002). Activity concentrations were calculated until 2020, as it is of interest to see the changes in concentration after the discharges cease. The discharges are presented in Section 3 and the results discussed in Section 4. 2. The dispersion models 2.1. The MARINA II model The MARINA II model is a 72-compartment model for estimating the dispersion of radionuclides in Northern European waters. It incorporates model developments carried out at the National Radiological Protection Board (NRPB), Centre d’études sur l’Evaluation de la Protection dans le domaine Nucléaire (CEPN) and the Risø National Laboratory and also recent improvements in representing marine sedimentation processes (European Commission, 1995, 2002; Lepicard and D’Ascenzo, 2000; Lepicard, 2001). The movement of water between various sea areas by processes of advection and diffusion is modelled by assuming instantaneous uniform mixing within each marine compartment with rates of annual transfer between adjacent compartments. The detail of the model compartments is greatest in northern European waters; however the model includes transfer to and recycling from the World oceans. This is important for very long-lived and mobile radionuclides, such as 14 C and 99 Tc, which adsorb weakly to sediment and potentially contribute to
Estimating concentrations of radionuclides in UK waters
185
Fig. 1. Irish Sea compartments in the MARINA II model.
collective doses for long timescales. However, for comparison with the CSERAM model, only the results for the compartments in the Irish Sea are considered; details of these compartments are given in Fig. 1. The absorption of radionuclides by sediments can lead to a significant removal from the water column, due to both the partitioning between the liquid phase and suspended sediments, and the subsequent removal of the activity from the water column to bed sediments. This partitioning is described in the model using a distribution coefficient (IAEA, 1985, 2004), defined as the ratio of the concentration of a radionuclide in dry sediment (in Bq t−1 ) to its concentration in filtered water (Bq m−3 ) at equilibrium. The movement of radionuclides within the seabed after being deposited from the overlying water and the return of radionuclides to the water phase is modelled using a multilayered bed structure. This allows relevant processes, such as molecular diffusion, porewater mixing, particle mixing and sediment turnover, to be taken into account. A distinction is made between deep and coastal waters in modelling the various processes due to factors such as the extent to which exchange with sediment occurs and the abundance of biota. A complete list of all of the model parameters is given in European Commission (2002). 2.2. The CSERAM model The approach adopted in CSERAM is to resolve and model, in a realistic way, the main physical, chemical, and biological processes responsible for the transport and re-distribution of radionuclides in both dissolved and particulate form. This compares with the more traditional ‘box model’ approach of the MARINA II model in which transport processes are parameterised via advective and diffusive fluxes between compartments representing the water column and the seabed.
186
K. Sihra et al.
Fig. 2. Compartments in the CSERAM model.
The more detailed approach relies on resolving the full tide, wind, and ‘diagnostic’ density driven flow fields derived from a numerical model (Aldridge, 1998; Aldridge et al., 2003). Where wind waves are responsible for significant sediment resuspension, a wave model (based on the JONSWOP spectrum; Hasselmann et al., 1973) is used to calculate the wave-induced stress. The third component of the modelling system (Aldridge et al., 2003) is a transport model capable of dealing with transport material in both dissolved and particulate (bound) phases. The hydrodynamic model is run on an extended grid which cover the Malin shelf and Celtic Sea regions, while the transport model is run in the Irish Sea only at a resolution of approximately 4 km (see Fig. 2). 2.3. Validation of the models Three radionuclides, 99 Tc, 137 Cs, and 239/240 Pu, were used in the validation of both models. These radionuclides were chosen either to represent a long time period of measurements (137 Cs), or to exhibit a range of behaviours in the marine environment, from a strong affinity to sediment (239/240 Pu) to high spatial dispersion (99 Tc).
Estimating concentrations of radionuclides in UK waters
187
Annual radionuclide discharge data from all significant sources to the Irish Sea were considered. Discharge data were collated, for discharges from the start of operation (indicated in parentheses) until 2000, for the sites located at Heysham (1982 for Heysham 1; 1987 for Heysham 2), Sellafield (1952), Wylfa (1971). For validation, the MARINA II model considered annual discharges from all of these sites, while the CSERAM model considered discharges using both monthly and annual discharge profiles. Observational data were provided by the Centre for Environment, Fisheries and Aquaculture Science, CEFAS, based on many measurement surveys dating back as far as the mid1970s (Bexon et al., 2003a). Furthermore, only a summary of the model validations in the Irish Sea, for the radionuclides used in this model comparison, is included here; for a complete description see European Commission (2002); Bexon et al. (2003b). For 137 Cs in seawater, the MARINA II model showed good agreement with observations pre-1974 and post-1986, with differences typically of better than a factor of two and with the model systematically calculating lower activity concentrations than the observations during the latter period. The variation with time estimated by the model was also in good agreement with the observations. Between 1974 and 1986, when activity concentrations in the Irish Sea were at their highest, the MARINA II model showed differences of a factor of two to three, with no systematic model bias in activity concentration. Similar results were seen in the comparison of the total Irish Sea inventory between the CSERAM model and observations. Largest differences of a factor of 1.5 were observed between 1974 and 1986, with the model calculating lower activity concentrations post-1986 by as much as a factor of two. It was also noted that the Irish Sea flushing rate needed to be increased post-1976 in order to accurately reproduce the peak activity concentrations (Jefferies et al., 1982). For 137 Cs activity concentrations in bed sediment, the CSERAM model generally predicted lower values than the measurements, while the MARINA II model showed a spread of values both greater and less than the measurements, dependent on geographical location. Outside the Irish Sea, the MARINA II model performed well for recent years lying within a factor of two of the measured data in most cases. Calculated activity concentrations of 239/240 Pu in the Irish Sea compared better with the observed measurements than the results for 137 Cs did. Both models predicted greater activity concentrations in seawater by a factor of 1.5 compared to measurements after 1980. However, there are large variations in the individual measurements that were not modelled either by MARINA II or the more physically based CSERAM model (although both models compare better to the average of the measurements). Both models showed strong remobilisation from the bed sediment when discharges, from Sellafield in particular, were reduced. Calculated activity concentrations in the bed sediment in the Irish Sea are in reasonable agreement with the observed data (with differences of up to a factor of three). A more extensive description of the model validation for the Irish Sea can be found in (Bexon et al., 2003b). However, it should be apparent that both models give a good if not perfect representation of the marine processes that disperse the radioactivity from the discharge points into the sea. It is interesting to find that the MARINA II spatially and temporally averaged activity concentrations agree with point observations from research vessels. It is possible that these measurements are not truly representative of the surrounding area. However, agreement also with the CSERAM model, with its higher temporal resolution, suggests that the MARINA II resolution and compartment sizes are appropriate to approximate the Irish Sea
188
K. Sihra et al.
environment. It should also be interesting to note that the physical processes modelled explicitly, and the input driving data, in CSERAM are sufficient to represent the movement of radionuclides in the Irish Sea accurately and with a reduced need for model parameterisations or empirically-based approximations. A comparison between MARINA II and CSERAM to the year 2020 should therefore provide a useful insight into the predictive ability of MARINA II in a regime outside of that which the model was tuned to.
3. The discharge scenario The discharge scenario used in the comparison was of actual historical discharges of 239/240 Pu and 137 Cs from Sellafield from the 1950s until the end of 1999 (see Fig. 3). Activity concentrations were calculated until 2020. Comparisons were produced for all Irish Sea compartments shown in Fig. 1, with the exception of the Irish Sea south compartment because it is outside the spatial extent of the CSERAM model. Model comparisons are discussed for activity concentrations in both filtered seawater and sediment.
4. Results and discussion 4.1. Caesium-137 The comparison of activity concentrations of 137 Cs in filtered seawater, between MARINA II and CSERAM, typically shows two different types of result. In the compartments close to the discharging source, there is close agreement (better than a factor of two) in calculated activity concentrations to the year 2000, while the site is discharging, and differences up to an order of magnitude by 2020, with the MARINA II model showing lower activity concentrations than CSERAM (see Fig. 4 for example). In the compartments further away from the discharging source (Irish Sea Northwest and Irish Sea West), the MARINA II model shows systematically
Fig. 3. Historical discharges from Sellafield.
Estimating concentrations of radionuclides in UK waters
189
Fig. 4. Comparison of the activity concentrations of 137 Cs in filtered water, in the Cumbrian Waters compartment, from historical discharges at Sellafield.
Fig. 5. Comparison of the activity concentrations of 137 Cs in bed sediment, in the Cumbrian Waters compartment, from historical discharges at Sellafield.
greater activity concentrations to the year 2000, by as much as a factor of three, but a much better comparison to 2020. For 137 Cs in sediment, there is a larger difference in massic activities between the two models than for filtered seawater by as much as a factor of 10 (see Fig. 5 for example). The CSERAM model estimates greater activities up to the early 1980s, when source discharges are high, but also a quicker reduction with time as discharges decrease, compared to MARINA II. This trend is representative of all compartments in the Irish Sea. Given the responses of the models to calculating activity concentrations in both filtered seawater and bed sediment, it is possible to conclude that the CSERAM model reaches equilibrium between liquid and sediment phases more quickly during discharges and shows more rapid remobilisation after the source is switched off. The differences in filtered seawater activity concentrations at 2020
190
K. Sihra et al.
for compartments close to the source maybe important for critical group doses. However, the differences are more than acceptable for collective dose calculations, to the UK population for instance. This is because collective doses are estimated by summing exposures from all compartments. 4.2. Plutonium-239/240 The comparison of 239/240 Pu activity concentrations in filtered water generally showed closer agreement between models than that for 137 Cs. Typically both models agreed to within a factor of two, as shown in Fig. 6, with the exception of the Cumbrian Waters, and Liverpool and Morecambe Bay, where differences up to a factor of 20 are observed in 2020. For the Irish Sea Northwest and West compartments, the MARINA II model showed greater activity concentrations than CSERAM, for all times to 2020, while for the other compartments, the MARINA II model showed lower activity concentrations. This spatial difference has been attributed to the CSERAM model having a time-dependent, sediment distribution coefficient until equilibrium is reached, while in MARINA II equilibrium conditions are assumed at all times. The comparison of massic activities of 239/240 Pu in sediment also showed closer agreement between models than that for 137 Cs. Both models were found to agree to within a factor of two to five for all Irish Sea compartments (see Fig. 7), with the MARINA II model calculating greater activities than CSERAM for all compartments except the Cumbrian Waters, and Liverpool and Morecambe Bay. It is possible that both the filtered water and the bed sediment differences at these locations are due to the differences in the treatment of site-specific details. The oscillations modelled in the western Irish Sea seabed activities (and water activity concentrations to some extent) by CSERAM are due to a seasonal density stratification effect, which mimics the induced summer circulation in the low tidal region to the west of the Isle of Man.
Fig. 6. Comparison of the activity concentrations of 239/240 Pu in filtered water, in the Irish Sea Northwest compartment, from historical discharges at Sellafield.
Estimating concentrations of radionuclides in UK waters
191
Fig. 7. Comparison of the activity concentrations of 239/240 Pu in bed sediment, in the Irish Sea Northwest compartment, from historical discharges at Sellafield. The oscillations observed in the CSERAM calculation are due to a seasonal density stratification effect.
5. Conclusion Both the MARINA II and CSERAM models show reasonably good agreement, especially considering that the models have been developed using very different modelling approaches. In particular, CSERAM is designed to provide activity concentrations in the Irish Sea on a high-resolution grid, compared to the seven compartments representing the Irish Sea in the MARINA II model. Spatial and temporal averaging of calculated activity concentrations is therefore likely to contribute to these observed differences. The MARINA II model also considers the transport of radionuclides throughout the whole of northern European waters and is optimised to give the best overall fit to all measurements in this region. Consequently, the effect of model differences close to the source at 2020 is likely to be reduced considerably for collective dose calculations in particular. Therefore, considering the computational cost of running the model, MARINA II provides an effective tool for radiation protection predictive modelling.
Acknowledgement The work was funded by the Food Standards Agency, as part of a wider study, under contract RO1049.
References Aldridge, J.N. (1998). CSERAM: A model for prediction of marine radionuclide transport in both particulate and dissolved phases. Radiation Protection Dosimetry 75, 99–103. Aldridge, J.N., Kershaw, P., Brown, J., Young, E.F., McCubbin, D., Leonard, K. (2003). Transport of plutonium (239/240 Pu) and caesium (137 Cs) in the Irish Sea: Comparison between observations and results from sediment and contaminant transport modelling. Continental Shelf Research 23, 869–899.
192
K. Sihra et al.
Bexon, A.P., Shaw, S., Sihra, K.S., Simmonds, J.R., Aldridge, J.N., Gurbutt, P.A., Smith, B.D. (2003a). Development of a Methodology for the Prediction of Doses from the Consumption of Marine Foodstuffs, for Past and Current Discharges. Volume 1 – Description of methodology and dataset CD-ROM. Food Standards Agency, London. Bexon, A.P., Shaw, S., Sihra, K.S., Simmonds, J.R., Aldridge, J.N., Gurbutt, P.A., Smith, B.D. (2003b). Development of a Methodology for the Prediction of Doses from the Consumption of Marine Foodstuffs, for Past and Current Discharges. Volume 2 – Technical background to the methodology. Food Standards Agency, London. CEC (1990). The radiological exposure of the population of the European Community from radioactivity in North European marine waters. Project “MARINA”, EUR 12483. EC, Luxembourg. European Commission (1995). Methodology for assessing the radiological consequences of routine releases of radionuclides to the environment. Radiation Protection 72. European Commission (2002). MARINA II. Update of the MARINA project on the radiological exposure of the European Community from radioactivity in North European marine waters. Radiation Protection 132. Hasselmann, K., Barnett, T.P., Bouws, E.C.H., Cartwright, D.E., Enke, K., Ewing, J.A., Gienapp, H., Hasselmann, D.E., Kruseman, P., Meerburg, A., Müller, P., Olbers, D.J., Richter, K., Sell, W., Walden, H. (1973). Measurements of wind-wave growth and swell decay during the Joint North Sea Wave Project (JONSWAP). Deutsche Hydrographische Zeitschrift 8 (Suppl. A), 95. IAEA (1985). Sediment kds and concentration factors for radionuclides in the marine environment. Technical Report Series No 247. IAEA, Vienna. IAEA (2004). Sediment distribution coefficients and concentration factors for biota in the marine environment. Technical Report Series No 422. IAEA, Vienna. Jefferies, D.F., Steele, A.K., Preston, A. (1982). Further studies on the distribution of 137 Cs in British coastal waters. I. Irish Sea. Deep Sea Research 29, 713–738. Lepicard, S. (2001). Review of marine models for impact assessments of radionuclide releases into the north European marine waters – contribution to MARINA II Working Group D. CEPN NTE/01/15. Lepicard, S., D’Ascenzo, L. (2000). Integration d’un decoupage pour l’Atlantique dans le logiciel POSEIDON (Version 3). Note CEPN NTE/00/09. OSPAR (1998). SINTRA Statement. Summary Record, OSPAR 98/14/1 Annex 45. OSPAR, London.
193
Assessment of the discharge of NORM to the North Sea from produced water by the Norwegian oil and gas industry T. Gäfvert* , I. Færevik, A.L. Rudjord The Norwegian Radiation Protection Authority (NRPA), Norway Abstract In order to obtain a reliable estimate of the discharge of NORM (Naturally Occurring Radioactive Materials) to the North Sea from the Norwegian oil and gas industry, produced water from 41 Norwegian offshore platforms has been sampled during a five-month period from the autumn of 2003 to early 2004 and analysed for 226 Ra, 228 Ra and 210 Pb. Together with data on the volume of produced water discharged from each production platform, the total activity of radium discharged to the North Sea in 2003 has been estimated. Activity concentration of 226 Ra and 228 Ra in the samples ranged from below the detection limit (0.3–1.3 Bq l−1 ) up to 16 and 21 Bq l−1 , respectively. For 210 Pb, all results except one were below the detection limit (0.2–1.5 Bq l−1 ). For some of the platforms, short-term variations in the activity concentration were observed. The discharge of 226 Ra and 228 Ra through produced water in 2003 from the Norwegian oil and gas industry into the North Sea was estimated to be 440 GBq and 380 GBq, respectively. Dividing the total discharged activity by the total volume of produced water discharged in 2003 gives an average activity concentration of 3.3 Bq l−1 for 226 Ra and 2.8 Bq l−1 for 228 Ra in produced water from the Norwegian continental shelf. Keywords: NORM, 226 Ra, 228 Ra, 210 Pb, Oil and gas industry, Produced water, Seawater, North Sea
1. Introduction In recent years the attention to discharges of natural radioactive elements from non-nuclear industries has increased. The MARINA II study, published by the European Commission in 2003, found that discharges from phosphoric acid production and the oil and gas industry, had contributed significantly to the total input of alpha-emitting nuclides to northern European marine waters. Since the discharge of phosphogyspsum has declined during the 1990s, the discharge of produced water from the oil and gas industry, containing enhanced levels of radium, has become the major contributor of alpha-emitters according to the MARINA II study (Betti et al., 2004; MARINA II, 2003). One problem in assessing the impact of discharges from * Corresponding author. Address: NRPA, Postboks 55, N-1332 Østerås, Norway; phone: (+47) 67 162576; fax: (+47) 67 147407; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08013-7
© 2006 Elsevier Ltd. All rights reserved.
194
T. Gäfvert et al.
non-nuclear industries is that there is generally less information available concerning the activity discharged compared with authorised discharges from the nuclear industry. A sampling programme was initiated in 2003 by the NRPA (Norwegian Radiation Protection Authority) in order to obtain a good estimate of the discharged activity of NORM, especially radium, through produced water from the Norwegian oil and gas industry. The objective was to obtain information on the activity concentrations of the alpha-emitter 226 Ra and the beta-emitters 228 Ra and 210 Pb in produced water from all 41 Norwegian platforms discharging into the North Sea, and to assess the total activity of 226 Ra and 228 Ra discharged in 2003. In order to see whether short-term variations in the activity concentration occur, samples were collected monthly from each platform during five consecutive months. The study is also considered important for work in the OSPAR radioactivity group (RSA), where relevant NORM industries will be identified, and discharge data collected during the next few years. 1.1. Produced water from the oil and gas industry Produced water constitutes the largest waste stream in terms of volume from oil and gas exploitation. In 2003, 135 × 106 m3 was discharged from the Norwegian offshore oil and gas industry into the North Sea (OLF, 2004), while 21 × 106 m3 was reinjected into the reservoirs. The produced water extracted together with the oil and gas consists of formation water naturally present within the wells, but can also be a mixture of formation water and seawater, if seawater has been injected in order to maintain pressure in the reservoir. Produced water contains a large amount of dissolved inorganic elements and organic compounds that have been leached from the surrounding geological formations (Neff, 2002). Constituents always present in varying amounts are dispersed oil, dissolved hydrocarbons, organic acids and phenols (Røe, 1998). Furthermore, produced water contains traces of production chemicals, such as scale inhibitors, that are injected into the reservoirs or the production line to facilitate production. The composition of produced water may also change over time. If seawater is injected, the composition of the produced water can change rapidly in the mixing zone if a breakthrough of seawater occurs, and in a long-term perspective the character of the water will change to that of seawater if large volumes of seawater are injected and the original formation water is extracted from the reservoir (Stephenson et al., 1994). When seawater is mixed with formation water, some elements and compounds present in the formation water will be diluted. Another effect is that elements such as barium may precipitate as sulphates, due to the low solubility product of BaSO4 . Due to the similar behaviour of barium and radium, coprecipitation of Ra with BaSO4 is also possible. An example of how Ba levels can change within a reservoir when seawater breaks through is presented in Fig. 1. This shows how Ba concentration in the produced water and seawater fraction (calculated from the concentration of sulphate ions) changed in one of the wells in the Statfjord Nord field during the period 1998–2004. Breakthrough occurred in the autumn of 2000, as shown by the increase in the fraction of seawater. Simultaneously, Ba levels declined. One year later, at the end of 2001, the seawater fraction fell sharply because production was changed to a zone in the reservoir where breakthrough had not yet occurred. The volume of produced water may also change over time. In the initial phase of production the water-to-oil ratio is generally low. If seawater or other types of water (sulphate free
Assessment of the discharge of NORM to the North Sea
195
Fig. 1. Barium concentration and seawater fraction in produced water in a North Sea well in the period 1998–2004 (unpublished data from Thingvoll, Statoil, 2004).
formation water can also be injected from nearby reservoirs) is injected the water-to-oil ratio will increase, and in the final phase of production the produced water volume can be several times larger than the volume of oil produced. The average water-to-oil ratio for the Norwegian oil and gas industry in 2003 has been reported to be 0.88 (OLF, 2004). 1.2. Natural radionuclides in produced water One component that has become the subject of much attention is naturally occurring radioactive elements, foremost radium, which may be present in the produced water. Four radium isotopes exist in nature (226 Ra from the uranium decay series, 228 Ra and 224 Ra from the thorium decay series, and 223 Ra from the actinium decay series) where 226 Ra and 228 Ra are the most long-lived, with physical half-lives of 1600 and 5.75 years, respectively. Radium enters the water from the surrounding rocks by direct alpha recoil into the water or chemical leaching (the latter may be facilitated by crystalline damage caused by recoil energy from previous alpha decays in the decay series). Low sulphate concentrations in formation water, due to the reducing environment, and high salinities are two factors that enable radium to remain in solution. If sulphate ions are present radium will coprecipitate with BaSO4 , and at a high salinity other positive ions will compete for adsorption sites on, for example, clay minerals in the reservoir (Kraemer and Reid, 1984; Bloch and Key, 1981). At high Cl− ion concentrations, the formation of soluble chloride complexes with radium can also explain enhanced radium activity concentration in formation water (Zukin et al., 1987). According to a review (Jonkers et al., 1997), 226 Ra levels in produced water samples have been reported to be in the range 0.002–1200 Bq l−1 .
196
T. Gäfvert et al.
At some locations, elevated activity concentrations of 210 Pb have been reported in formation water. Zukin et al. (1987) explained high levels (up to 97 Bq l−1 ), comparable to 226 Ra, in brines from a geothermal well in California, by the formation of soluble chloride complexes, high reservoir temperatures and high 222 Rn concentrations. Worden et al. (2000) reported high 210 Pb levels but low 226 Ra levels in a Triassic fluvial reservoir in the UK. The explanation of this was also the formation of soluble chloride complexes. Since the water was relatively oxidised and the concentration of sulphate was high, Ra could coprecipitate as barite (BaSO4 ). The low sulphide concentrations in the water prevented the formation of insoluble galena (PbS). 1.3. Radium in produced water from the Norwegian continental shelf In spite of the large volumes of produced water discharged to the North Sea into account, relatively few data have been published on the level of radium in produced water from the Norwegian oil and gas industry. An overview of published levels of radium in produced water sampled in the period 1995–2002 from the Norwegian continental shelf is presented in Fig 2. Strand et al. (1997) analysed produced water sampled at 11 Norwegian production platforms in 1996 with regard to 226 Ra and 228 Ra. Radium-226 ranged from below the detection limit up to 10.4 Bq l−1 , while 228 Ra was found in the range from below the detection limit up to 10.0 Bq l−1 . Of 27 samples in total, 23 samples showed activity concentrations of 226 Ra below 5 Bq l−1 . Røe (1998) published data on 226 Ra and 228 Ra levels in produced water sampled from 4 platforms (Brage, Oseberg F, Oseberg C and Troll) during the period October 1995 to August 1996, where 226 Ra and 228 Ra ranged from 6 to 9 Bq l−1 and <2 to 17 Bq l−1 , respectively. Strålberg et al. (2002) published 60 results (22 samples analysed by gamma spectrometry and 38 by the 222 Rn-emanation method) from the analysis of produced water during the period 1997 to 2001. The activity concentration of 226 Ra in 57 of the samples was in the range from below the detection limit up to 7.2 Bq l−1 , and for 228 Ra 21 of the 22 samples had concentrations in the range from below the detection limit up to 7.2 Bq l−1 . Two samples had
Fig. 2. Overview of published radium levels in produced water from the Norwegian continental shelf in the period 1995–2002 (Strand et al., 1997; Røe, 1998; Strålberg et al., 2002; Varskog, 2003).
Assessment of the discharge of NORM to the North Sea
197
significantly higher 226 Ra activity concentrations, 160 and 165 Bq l−1 . These activity concentrations do not seem to be representative of produced water from the Norwegian continental shelf and, due to the lack of information, are not easy to explain. The results of 41 produced water samples published by Varskog (2003) represent samples from each Norwegian production platform in the North Sea in 2002. Some samples from 2000 and 2001 were also included. Due to the analytical technique used (the 222 Rn-emanation method), results were only obtained for 226 Ra. The activity concentration of 226 Ra in these samples ranged from 0.04 to 15 Bq l−1 . The total activity of 226 Ra discharged from Norwegian production platforms in 2002 was estimated to be 306 GBq. The average activity concentration, 2.5 Bq l−1 was calculated by dividing the total discharged 226 Ra activity from the Norwegian continental shelf by the total volume of produced water discharged in 2002.
2. Materials and methods 2.1. Determination of 226 Ra, 228 Ra and 210 Pb activity All samples of produced water were collected at the platforms during the period September 2003–January 2004 (a few samples were also collected in February 2004). The samples were collected after separation of solids, oil and gas from the produced water, and each sample consisted of about two litres. In order to investigate the total discharge the samples were not filtered, neither were the samples acidified in connection with sampling. Bottles were carefully washed with dilute nitric acid later at the laboratory to remove radium and lead that could have been absorbed on the walls of the bottles. After the samples were collected they were sent to the Institute for Energy Technology (IFE) at Kjeller (Norway) where all analyses were carried out. The analytical method used has been developed and tested at IFE. In order to improve the radiochemical yield, the method was slightly modified during the project. The original and modified analytical methods are described below (Sidhu, IFE, personal communication, 2004). 2.1.1. Original method A 2-litre sample to which 133 Ba had been added as a yield determinant was first treated with permanganate under acidic conditions. Radium and 210 Pb were coprecipitated from the solution with MnO2 . By raising the pH, MnO2 was precipitated and then filtered from the solution. After the precipitate had been dried it was mixed with Al2 O3 and placed in a vial for gammaspectrometric analysis using lead-shielded HPGe detectors. The method was tested by adding known activities of 226 Ra and 210 Pb to produced water samples. Most of the samples showed a chemical yield above 80%. A number of samples, however, showed a chemical yield below 60–70%. The reason for this is unknown. These samples were also evaporated and the activities in the residue and the precipitate were analysed separately and then added. One possible reason for the lower yield is that complexing agents, such as scale dissolvers, could be present in the samples, preventing Ra from co-precipitating. Since many complexing agents are effective at elevated pH, the method was modified and MnO2 was precipitated at a low pH together with Pb(Ba)SO4 .
198
T. Gäfvert et al.
2.1.2. Modified method After a 2-litre sample of produced water had been acidified, Pb and Ba carriers were added together with 133 Ba, which was used as a radiochemical yield determinant. Potassium permanganate and sodium sulphate were then added and Ba, Pb and Ra were precipitated as sulphates. Part of the manganate is reduced to MnO2 in this initial step and the remaining manganate was reduced by addition of sodium sulphite. Both precipitates were then filtered from the solution, dried and homogenised together with Al2 O3 , in order to fill one of the vials used for gamma-spectrometric analysis at the laboratory. All samples analysed with this method had chemical yields above 90%. This method was also tested by adding known amounts of 226 Ra and 210 Pb to produced water samples. After been filtered from the solution and dried, the precipitates were stored for a sufficient period of time to ensure secular equilibrium between 228 Ra and 228 Ac. The samples were then analysed using lead-shielded HPGe detectors. Radium-228 was determined from the gamma radiation from 228 Ac. Radium-226 and 210 Pb were determined from their 186 and 46.5 keV gamma photons, respectively, corrected for self-absorption according to the method described by Cutshall et al. (1983). This method has previously been tested at the laboratory, and the results have been presented by Sidhu and Strålberg (2003). For most samples the detection limit was below 1 Bq l−1 for each nuclide, and the typical uncertainty (±2 SD) about 20%. Efficiency calibration of the HPGe detectors is routinely performed using a multi-nuclide solution traceable to a national standard (LMRI). The laboratory frequently participates in intercomparison tests arranged, for example, by the IAEA and NKS (Nordic Nuclear Safety Research) with satisfactory results (due to the less common intercomparison tests for radium, the results refer to gamma-emitting nuclides other than radium isotopes). 2.2. Discharge volumes Data on total discharge volumes of produced water during 2003 for all 41 production platforms were provided by the OLF (the Norwegian Oil Industry Association). Together with the results from the radioactivity analyses, these data were used to estimate the total activity of 226 Ra and 228 Ra discharged in 2003 from each production platform on the Norwegian continental shelf. The platform-specific data on discharge volumes during 2003 are shown in Fig. 3.
3. Results 3.1. Activity concentration of 226 Ra, 228 Ra and 210 Pb in produced water Overall, the response from the operators to the request to sample the produced water was good, and results were obtained for all 41 production platforms discharging produced water. In total 185 results were reported for 226 Ra, 228 Ra and 210 Pb. All results for 226 Ra and 228 Ra are presented in Fig. 4. The average activity concentration of 226 Ra in all analysed samples was 4.5 Bq l−1 , and ranged from below the detection limit up to 16 Bq l−1 . The majority of the results, 65%, were below 5 Bq l−1 . For 228 Ra, the average activity concentration in the samples was 3.9 Bq l−1 , ranging from below the detection limit up to 21 Bq l−1 . As
Assessment of the discharge of NORM to the North Sea
199
Fig. 3. Volume of produced water discharged from Norwegian oil and gas platforms in 2003 (data provided by the OLF).
Fig. 4. Overview of all results for 226 Ra and 228 Ra in produced water samples from the Norwegian oil and gas industry in the period September 2003–January 2004.
for 226 Ra, the majority of the results, 69%, were below 5 Bq l−1 . For 210 Pb, all samples except one showed activity concentrations below the detection limit (between 0.2 and 1.5 Bq l−1 ,
200
T. Gäfvert et al.
depending on parameters such as measurement time and detector efficiency). The one sample above the detection limit had an activity concentration of 0.4 ± 0.1 Bq l−1 . The average activity concentrations (and range) of 226 Ra and 228 Ra in produced water from each of the 41 production platforms are presented in Figs. 5 and 6, respectively. The results are arranged according to increasing average activity concentration of 226 Ra.
Fig. 5. Average activity concentration (and range) of 226 Ra in samples of produced water from 41 Norwegian production platforms collected in the period September 2003–January 2004.
Fig. 6. Average activity concentration (and range) of 228 Ra in samples of produced water from 41 Norwegian production platforms collected in the period September 2003–January 2004.
Assessment of the discharge of NORM to the North Sea
201
3.2. Estimated discharge of 226 Ra, 228 Ra and 210 Pb in 2003 from Norwegian platforms The discharge of 226 Ra and 228 Ra from each platform (see Figs. 7 and 8) was estimated by multiplying the volume discharged by each platform in 2003 (presented in Fig. 3) by the average activity concentration of 226 Ra and 228 Ra (Figs. 5 and 6). The total activity of 226 Ra discharged in 2003 was estimated to be 440 GBq (range: 310–590 GBq). For 228 Ra the total activity discharged was estimated to be 380 GBq (range: 270–490 GBq). In samples where the
Fig. 7. Average discharged 226 Ra activity (and range) from each Norwegian production platform in 2003.
Fig. 8. Average discharged 228 Ra activity (and range) from each Norwegian production platform in 2003.
202
T. Gäfvert et al.
activity concentration was below detection limit (about 14% of the total number of samples for 226 Ra), the detection limit was used as the activity concentration in the sample in order to obtain a conservative estimate. Assuming the samples with activity concentrations below the detection limit contain no radium, a total discharge of 430 GBq for 226 Ra and 370 GBq for 228 Ra would be derived. Since almost all analysed samples of 210 Pb were below the detection limit it is only possible to estimate an upper limit for the discharge of 210 Pb. Based on the results obtained in this study, the activity of 210 Pb discharged in 2003 was <92 GBq.
4. Discussion and conclusions The total activity of 226 Ra discharged in 2003 (440 GBq) agrees relatively well with the estimate for 2002 made by Varskog (2003) (306 GBq and an average 226 Ra activity concentration of 2.5 Bq l−1 ), also based on platform-specific data. The higher result in this study can partly be explained by the larger volume of produced water discharged in 2003 (135 × 106 m3 ), compared with 2002 (123 × 106 m3 ), and the fact that a different analytical method was used to analyse the samples collected in 2002. The samples collected in 2002 were analysed by the 222 Rn-emanation technique, which has a slightly lower detection limit than the method used to analyse the samples in this study. In samples with the lowest activity concentration in 2003, the activity concentration represents the detection limit and not the actual activity concentration in the produced water. Dividing the total activity discharged by the total volume of produced water discharged gives average activity concentrations of 3.3 Bq l−1 for 226 Ra, and 2.8 Bq l−1 for 228 Ra. This can also be compared with radium levels found in North Sea seawater. Activity concentrations of 226 Ra in the northern part of the North Sea in 2002 were found to be in the range 1–2.5 mBq l−1 (NRPA, 2004), showing that produced water contains about three orders of magnitude higher levels of radium than seawater. Due to the ageing of fields, it has been predicted that the discharged volume of produced water from the Norwegian oil and gas industry will increase in the years to come. The increasing trend is expected to continue to around 2010, before a fall in output due to the shutdown of old platforms and increased implementation of discharge reduction measures, such as, reinjection of produced water and plugging of water zones in the reservoirs. The predicted volume discharged in 2010 is about 200 × 106 m3 (Ministry of Petroleum and Energy, 2004). Assuming that the average activity concentration in the produced water will be the same as in 2003, a maximum discharge of about 660 GBq 226 Ra and 560 GBq 228 Ra in 2010 can be estimated. Another estimate of the activity of 226 Ra and 228 Ra discharged from the Norwegian oil and gas industry was made in the MARINA II study (MARINA II, 2003). In that study it was assumed that the average activity concentration of 226 Ra and 228 Ra, over the lifetime of an oil-producing reservoir was 10 Bq l−1 , each. The volume of produced water discharged was assumed to be 3 times the volume of the oil produced, which can be compared with the reported average water-to-oil ratio of 0.88 in 2003 for the Norwegian oil and gas industry (OLF, 2004). A small contribution was also assumed from gas production. With these assumptions it was estimated that 5.5 TBq of 226 Ra and 228 Ra, each, were discharged in 1997, compared with the 0.44 TBq and 0.38 TBq of 226 Ra and 228 Ra, respectively, found for 2003 in this
Assessment of the discharge of NORM to the North Sea
203
study. Since production of oil and gas from the Norwegian continental shelf has increased slightly since 1997 (OLF, 2004), the discharged activity should have been slightly higher than 5.5 TBq in 2003 with the assumptions used in the MARINA II study. According to the data presented in this study, it seems that both the activity concentration in the produced water and the volumes discharged were overestimated in the MARINA II study. The combination of relatively high activity concentration in the produced water and the large volumes of produced water discharged make the Troll B and C platforms the two main contributors of radium to the North Sea from the oil and gas industry on the Norwegian continental shelf in 2003. The total activities of 226 Ra and 228 Ra discharged from these platforms were estimated to be 193 GBq and 156 GBq, respectively, corresponding to 44% of the total discharged activity of 226 Ra and 41% of the total discharged activity of 228 Ra. The remaining platforms have an annual discharge of 226 Ra in the range <1 GBq to about 30 GBq. For 228 Ra the corresponding range is from <1 GBq to slightly below 25 GBq. By comparing Figs. 3 and 5 one can see that most platforms discharging large volumes of water also tend to show a lower activity concentration of radium in produced water, except for Troll B and C. This indicates that seawater injection has an impact on the radium levels in the produced water in a long-term perspective. Fields such as Statfjord and Gullfaks produce large amounts of oil, but have been in production since the mid 1980s or earlier. Over the years large volumes of seawater have been injected into the reservoirs and the lower radium levels could be due to dilution and coprecipitation with BaSO4 . It has, for example, been estimated from the sulphate ion concentration in the produced water, that 62% of the produced water at Statfjord A, B and C, in 2003, consisted of seawater previously injected into the reservoirs (Thingvoll, Statoil, personal communication, 2004). Due to a lack of data it was not possible to compare the radium levels in 2003 with levels during the initial phase of production from these fields. Troll B and C have been in production since 1995 and 1999, respectively, and are now among the largest oil-producing platforms on the Norwegian continental shelf, with a production of about 21 million m3 of oil in 2003 (Norsk Hydro, 2004). The reservoir naturally contains a large fraction of formation water compared to the oil. The oil is produced from horizontally drilled wells within thin oil-containing layers (thickness between 12 and 26 m). These special features result in the co-production of large volumes of produced water compared with other fields. Since no seawater is injected into the Troll field (Røe Utvik, Norsk Hydro, personal communication, 2004), the water quality is relatively stable and radium levels may not decrease over time due to dilution or coprecipitation with BaSO4 . The activity concentrations of 226 Ra and 228 Ra found in 1996 by Røe (1998) agree relatively well with the results from this study. The activity concentration of 226 Ra and 228 Ra in the produced water showed relatively large variations during the sampling period for many of the platforms (Figs. 5 and 6). For about 25% of the platforms the difference between the highest and the lowest observed activity concentration of 226 Ra was greater than 4 Bq l−1 . It was not possible to explain the observed variation in radium levels for each platform, but some parameters that may cause short-term variation are: 1. Many platforms produce oil and gas from several wells. If the production rates of the wells connected to the platform change, and the water chemistry is different in these wells, one can expect a variation in the activity concentration of radium in the produced water.
204
T. Gäfvert et al.
2. Breakthrough of seawater in the wells causes barium and probably radium levels in the produced water to change rapidly. 3. The total fraction of radium in the produced water was analysed in this study. If radium is present in suspended particulate material and colloids, such as BaSO4 , short-term variations in the particulate load in the produced water can be responsible for the variation of radium in the samples. Once or twice per week, the separators used to separate oil, gas and water are cleaned with high-pressure water to remove sand and scale particles. This, for example, may temporarily increase the load of radium-rich scale particles in the produced water. 4. Occasionally, scale removal operations take place at the platforms in order to remove scale in fixed equipment. This is usually done by dissolving the scale with acids or chelating agents. The liquid waste from these operations is then discharged together with the produced water. If a sample is collected during such an operation, elevated levels of radium can be expected. At some platforms scale dissolvers are also occasionally injected into wells when scale formation has slowed down the production in the well. If a sample is collected in connection with such an operation elevated levels of radium may also be observed. The use of scale inhibitors to prevent scale formation in wells, pipes and valves also leads to a larger fraction of radium in soluble form in the produced water. The purpose of collecting monthly samples for five consecutive months in this study was to detect possible variation and to use the information to determine a suitable sampling frequency for future surveys. Due to the relatively large number of platforms with a high variation it is desirable to continue sampling with a high frequency in order to be able to estimate the discharged activity. Collecting samples only once per year can, for certain platforms, lead to large over- or underestimates concerning the total amount of activity discharged annually. A sampling frequency of four samples per year from 2005 should ensure sufficient data for a reasonable estimate. This is also in line with the agreed OSPAR reporting procedure for discharges of radioactive substances from the oil and gas sector (OSPAR, 2004). Combining samples collected, for example, once or twice per week could be possible in order to get representative samples for a month. However, the extended storage time for the samples may be problematic for this sampling strategy. As described by Kraemer and Reid (1984), radium may coprecipitate rapidly with barium during storage, which later may constitute a problem when analysing the samples. A possible future strategy is to further reduce the number of samples collected from platforms showing low variation in activity concentrations of radium in the produced water and with a low annual discharged volume, while keeping the sampling frequency high for platforms with elevated levels of radium in the produced water, and where significant volumes of produced water are discharged.
Acknowledgements We wish to acknowledge the Norwegian Oil Industry Association (OLF) and the operators on the Norwegian continental shelf for assistance with knowledge and resources for this project.
Assessment of the discharge of NORM to the North Sea
205
References Betti, M., Aldave de las Heras, L., Janssens, A., Henrich, E., Hunter, G., Gerchikov, M., Dutton, M., van Weers, A.W., Nielsen, S., Simmonds, J., Bexon, A., Sazykina, T. (2004). Results of the European Commission Marina II study, Part II: Effects of discharges of naturally occurring radioactive material. Journal of Environmental Radioactivity 74, 255–277. Bloch, S., Key, R.M. (1981). Modes of formation of anomalously high radioactivity in oil-field brines. The American Association of Petroleum Geologists Bulletin 65, 154–159. Cutshall, N.H., Larsen, I.L., Olsen, C.R. (1983). Direct analysis of 210 Pb in sediment samples: Self-absorption corrections. Nuclear Instruments and Methods in Physics Research 206, 309–312. Jonkers, G., Hartog, F.A., Knaepen, A.A.I., Lancee, P.F.J. (1997). Characterization of NORM in the oil and gas production (E&P) industry. In: Radiological Problems with Natural Radioactivity in the Non-Nuclear Industry. Proc. Int. Symp., Amsterdam, 1997. KEMA, Arnhem. Kraemer, T.F., Reid, D.F. (1984). The occurrence and behaviour of radium in saline formation water of the U.S. Gulf Coast Region. Isotope Geoscience 2, 153–174. MARINA II (2003). Update of the MARINA Project on the radiological exposure of the European Community from radioactivity in North European marine waters. European Commission, Radiation Protection 132. Available at http://europa.eu.int/comm/energy/nuclear/radioprotection/publication_en.htm. Ministry of Petroleum and Energy (2004). Miljø 2004, Petroleumssektoren i Norge, Andresen, S. (Ed.), Olje- og energidepartementet, Oslo, ISSN 1502-0576 (in Norwegian). Neff, J.M. (2002). Bioaccumulation in Marine Organisms Effect of Contaminants from Oil Well Produced Water. Elsevier, ISBN 0-080-43716-8. Norsk Hydro (2004). Årsrapport til Statens forurensningstilsyn, Report (in Norwegian). Available at http://www.olf.no. NRPA (2004). Radioactivity in the Marine Environment 2002. Results from the Norwegian National Monitoring Programme (RAME). SrtålevernRapport 2004:10. Norwegian Radiation Protection Authority, Østerås. OLF, Norwegian Oil Industry Association (2003). Emissions and discharges from the Norwegian petroleum industry. Report. Available at http://www.olf.no. OSPAR (2004). OSPAR Convention for the protection of the marine environment of the North-East Atlantic. Meeting of the OSPAR Commission, Reykjavik, 28 June–1 July. Røe, T.I. (1998). Produced water discharges to the North Sea: A study of bioavailability of organic produced water compounds to marine organisms. PhD thesis. NTNU Trondheim, Norway. Sidhu, R., Strålberg, E. (2003). Self-absorption correction in the gamma analysis of low energy gamma emitters in LSA scale. Presented at the 3rd Dresden Symposium on Radiation Protection, Enhanced Naturally Occurring Radioactivity, ENOR III. Dresden, 3–7 March 2003. Stephenson, M.T., Ayers, R.C., Bickford, L.J., Caudle, D.D., Cline, J.T., Cranmer, G., Duff, A., Garland, E., Herenius, T.A., Jacobs, R., Inglesfied, C., Norris, G., Petersen, J.D., Read, A.D. (1994). North Sea produced water: Fate and effects in the marine environment. E&P Forum, Report No. 2.62/204. Strand, T., Lysebo, I., Kristensen, D., Birovljev, A. (1997). Deposits of naturally occurring radioactivity in production of oil and natural gas. Norwegian Radiation Protection Authority, Report 1997:1. Østerås, Norway. (In Norwegian.) Strålberg, E., Singh Sidhu, R., Varskog, P. (2002). Produsert vann og radioaktivitet–sammenfatning av eksisterande data. Report from Norsedecom, ND/E-05-02. (In Norwegian.) Varskog, P. (Ed.) (2003). Naturlige radionuklider i det marine miljø–en oversikt over eksisterende kunnskap med vekt påNordsjø-området, Norsedecom, ND/E-17-03. Norsedecom, Kjeller. Worden, R.H., Manning, D.A.C., Lythgoe, P.R. (2000). The origin and production geochemistry of radioactive lead (210 Pb) in NORM-contaminated formation waters. Journal of Geochemical Exploration 69/70, 695–699. Zukin, J.G., Hammond, D.E., Ku, T.-L., Elders, W.A. (1987). Uranium–thorium series radionuclides in brines and reservoir rocks from two deep geothermal boreholes in Salton Sea Geothermal Field, southeastern California. Geochimica et Cosmochimica Acta 51, 2719–2731.
206
Uranium mining and ore processing in Ukraine – radioecological effects on the Dnipro River water ecosystem and human health O. Voitsekhovitcha,* , Y. Sorokab , T. Lavrovaa a Ukrainian Hydrometeorological Institute, Nauki av., 37, 03028, Kiev, Ukraine b Centre for radioecological monitoring, Zhovti Vody, Ukraine
Abstract Results on the ecological impact of radionuclides from a uranium tailing dump in Dniprodzerzhinsk (Ukraine) and from leaching of contaminated water from an ore processing tailing pond close to the town of Zhovti Vody into the rivers are presented. It appears that most of the former uranium mining and milling operations have been conducted without sufficient and adequate care for the environmental consequences. An inventory of the waste related to the uranium industry was carried out during the recent decade, however, a complete picture of the spatial and temporal dispersion of radioactive and chemical contaminants into the surrounding environment is not clear. Close to uranium tailing and mines, hydrometallurgical and chemical plants in the studied areas also produce fertilizer and wastes containing radioactive substances and may have an affect on the aquatic environment, creating in some cases rather high concentrations of naturally occurring radionuclides in the river’s water and may constitute a risk to the population and environment. The highest levels of human exposure, which can be expected to be potentially received by inhabitants of settlements located on the banks of Zheltaya River are 0.10–0.15 mSv yr−1 . For people who (hypothetically) consume water from the Konoplyanka River the dose is expected to be 0.01–0.05 mSv yr−1 . Keywords: Radionuclides, Uranium isotopes, River water, Radiological risk, Ukraine
1. Introduction In the later years of the Soviet era, Ukraine produced about 1000 tonnes of uranium per year. The main area of uranium mining in Ukraine is situated in the vicinity of Zhovti Vody town and other sites in the Kirovograd and Dnipropetrovsk regions. The two main industrial sites for uranium milling were in operation in Zhovti Vody and Dniprodzerzhinsk towns (Fig. 1). Since Ukraine became independent, uranium production in Ukraine has significantly declined (Soroka, 2000). The Pridnieprovsky Chemical Plant (PCP) was the first uranium processing plant in Ukraine. The plant started up in 1948 using ores shipped from different regions of the for* Corresponding author. Address: Ukrainian Hydrometeorological Institute, Nauki av., 37, 03028, Kiev, Ukraine; fax: (+380) 442655363; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08014-9
© 2006 Elsevier Ltd. All rights reserved.
Uranium mining and ore processing in Ukraine
207
Fig. 1. Major potential sources of radionuclide releases into the environment at the Dnipro River basin.
mer Soviet Union and countries in Central Europe. The PCP is situated near the Dnipro River (Dniprovske Reservoir) in the city of Dniprodzerzhinsk (Fig. 1). The other major “hot spot” is around Zhovti Vody town, where several uranium mines, Hydrometallurgical Plant (which has processed uranium ores since 1959) and uranium ore tailing dumps are located (see area shaped in Fig. 1 around Zhovti Vody site). It is located at a former iron ore exploration site near the center of Ukraine’s main uranium province. The uranium ore mining region and areas where ore processing plants are located in the Dnipro River basin can be considered to be the second (after the Chernobyl area) significant source of radioactive pollution of aquatic systems in Ukraine. It is well known that uranium mining and milling has a number of potential impacts on the environment and human health. These effects include: • contamination of mine water with 238,234 U and other radionuclides (230,232 Th, 226 Ra, 210 Po, 210 Pb); • radon release from mines, waste rock dumps and mill tailings piles; • run-off of water from contaminated areas of the mine or mill and also from surface of tailing dumps in many cases covered by phosphogypsum, which also contains high level of radioactivity; • leaching and seepage of radionuclides from tailings and subsequent transport in water; • erosion of tailings storage systems leading to dispersal of tailings by wind and water; • contamination of underground and surface waters by toxic non-radioactive substances, such as heavy metals and chemicals used in the processing.
208
O. Voitsekhovitch et al.
The aim of the present study was to estimate the current effects of radioactive releases of technologically enhanced naturally occurring radioactive materials (TENORM) on the rivers crossing the area of mining and processing of uranium ore as a basis for development of an environmental monitoring program and further strategy for restoration of the industrial sites.
2. Materials and methods Analytical measurements of water samples collected in the Zheltaya River in 2002 were done using alpha-spectroscopic methods in the Laboratory of Environmental Protection and Ecological Technology of Ukrainian Research Institute of the Industrial Technologies (Zhovti Vody). Water samples collected in the Konoplyanka River and Dniprovske Reservoir during 2003 were measured using liquid-scintillation spectrometer (Quantulus) in the Marzeev Institute of Hygiene and Medical Ecology (Kiev). Bottom sediment coring was carried out in the Konoplyanka River settling pond and the reservoir; the analyses using HPGe (ORTEC) detectors were done in the Ukrainian Hydrometeorological Institute (UHMI). Samples of aquatic plants were taken during the field campaign of UHMI in 2003 and 2004. These samples were analyzed using photocalorimetry and HPGe gamma-ray spectrometry methods.
3. Results and discussion 3.1. Zhovti Vody site The mining and processing of uranium ores at Zhovti Vody has negatively affected the environment as well as the sanitary and hygienic state of the town since the start of operations in the 1950s (Korovin et al., 2000; TACIS, 1999). The mining of uranium ores ceased in 1990, but mining of iron ore at the Novaya mine is still in progress. The main sources of radioactive contamination at Zhovti Vody are the three tailings sites named “KBZh”, “Sch” and “R”, which have been characterized in previous studies (TACIS, 1999; Voitsekhovych et al., 2003). However, existing data are very limited and do not allow estimation of annual fluxes from these sites into the river (Soroka et al., 2002). The main water body in the vicinity of Zhovti Vody town is the Zheltaya River which is a tributary of the Ingulets River. At present its bed is silted, and only a small channel is cleared, which is transporting a significant amount of water during spring flood and rains. The treated mine waters from the Novaya mine are discharged into this river. The contaminated water from tailings also seeps into the Zheltaya River. In some cases, the seepage waters from tailings have high concentrations of uranium and other pollutants. These waters can exceed maximal permissible levels for chlorides, sulphates, and even of 210 Po in drinking water. There is therefore a need to assess their long-term impact on the environment. The recent studies (Soroka, 2000; Soroka et al., 2002; Voitsekhovych et al., 2003) have confirmed a contamination of the Zheltaya River with uranium. In the area of the Netesovka settlement, the concentration of total uranium in 2001 in water was 2400–5900 Bq m−3 . This exceeds the concentration at the source of the river by a factor of 80. The intervention level according to the National Radiation Safety Standards (NRBU-97/D-2000) is 1000 Bq m−3 . The high
Uranium mining and ore processing in Ukraine
209
Table 1 Activity concentrations of natural radionuclides in the Ingulets, Saksagan’ and Zheltaya Rivers Sampling place (May–June, 2002)
Activity (Bq m−3 ) 238 U
234 U
234 U/238 U 210 Pb
Zheltaya River in settlement Annovka Zheltaya River in settlement Annovka Drinking water in town Zhovti Vody Ingulets River (Iskra reservoir) Saksagan’ River (Kress reservoir)
1020 ± 100 1160 ± 115 1400 ± 90 1600 ± 130 36 ± 5 58 ± 6 31 ± 5 48 ± 6 55 ± 6 77 ± 7
<11 – <11 <11 <11
PCingest according to NRBU-97/D-2000
10000
500
10000
226 Ra
8±3 – 5±2 13 ± 3 7±4 1000
210 Po
20 ± 3 – 2.4±0.5 0.5±0.2 1.1±0.5
1.13 1.14 1.61 1.54 1.40
200
concentrations of uranium were also measured in water samples taken close to the settlement Annovka (Table 1), exceeding by ten times the activity in the Saksagan and Ingulets Rivers. Estimation of radiation doses arising from water consumption from the Zheltaya River was carried out using the most conservative scenario. Direct water use from the rivers for irrigation and use of water from wells near the shore was assumed. It was further assumed that the irrigation and drinking water had similar activity concentrations as the river water. It was found that for the population of the Annovka settlement the annual dose rates for the adult population due to 238 U, 234 U and 210 Po only could reach 0.10 mSv yr−1 . The maximum observed activity concentrations of uranium in water were below the permitted concentration for ingestion (PCingest ), as shown in Table 1. However, the chemical toxicity of uranium also needs to be considered. The averaged concentration of 238 U is estimated to be about 1.2 Bq L−1 in the water of the Zheltaya River (in 2002) at the most contaminated sites. This corresponds to a value if about 0.1 mg U per liter. This is much higher than the healthbased guideline concentration, which was established by USEPA as 0.03 mg of U per liter (Ground Water Quality Bureau, 2003) or recommended by WHO as 0.002 mg L−1 (Giddings, 1996). This is well below the limit based for radiological considerations. 3.2. Dniprodzerzhinsk site According to the National Reports on the State of the Environment in Ukraine (NRBU, 2000), the Dnipropetrovsk region is one of the most unfavorable in terms of contamination of air, adjacent catchment areas and water bodies. One of the reasons for this unfavorable situation is the large concentration of industrial enterprises, which include chemical, coke and metallurgical plants (Fig. 2). One of the metallurgical facilities, which have severely affected the environment in the region was the former Industrial Association “Pridnieprovsky Chemical Plant” (PCP). During the operation of the PCP, nine tailings dumps were created containing about 42 million tones of radioactive waste (RAW) with a total activity of 3.2 × 1015 Bq (Soroka, 2000). Some of the RAW is located within the territory of the industrial zone of Dniprodzerzhinsk (see positions 1–4 and 10 in Fig. 2). The tailing dam Suhachevskoe, I and II sections (position 8 and 9) and others (positions 5–7 on Fig. 2) are about 14–15 km to the southeast. The first (I) section of the Suhachevskoe tailings dam was operated from 1968 to 1983. It contains about
210
O. Voitsekhovitch et al.
Fig. 2. Scheme of the main local “hot spots” in the Dniprodzerzhinsk town and surrounding area. (a) Tailing “D” and Konoplyanka; (b) Suhachevskoe tailings, where 1, 2, 3, 10 – sites of RAW disposal in the territory of PCP and industrial area of town; 4 – tailing “D” (Dniprovske Reservoir); 5, 6, 7 – sites of uranium ores storage; 8 and 9 – tailings Suhachevskoe (I and II).
19 million tons of RAW that occupy a volume of 8.6 million m3 and a total area of about 90 ha. The area is partly covered with water with about 14 ha of dry beaches. Wind re-suspension of phosphogypsum and soil dust significantly contributes to pollution of the surrounding areas. The concentration of 226 Ra in the wastes varies from 560 to 13500 Bq kg−1 , 238 U from 170 to 880 Bq kg−1 , 230 Th from 350 to 83000 Bq kg−1 and 210 Pb from 500 to 14400 Bq kg−1 . The total activity in this section is about 710 TBq (Soroka et al., 2002; Voitsekhovych et al., 2003). The tailing pond Suhachevskoe (section II), covering the lower part of the gully, is formed by a dam crossing the gully (see position 9, Fig. 2). It was commissioned in 1983 and is still used for disposal of non-radioactive wastes (e.g. phosphogypsum). It contains 9.6 million tones of RAW occupying the volume about 5.5 million of m3 . The total area is 70 ha of which about 50 ha is covered with water. The total activity is about 270 TBq. The maximum radiation dose at the surface is low due to deposition of non-radioactive material on top of the uranium tailings. The content of the natural radionuclides in the sludge waters are: uranium from 400 to 2000 Bq m−3 ; 226 Ra from 50 to 240 Bq m−3 ; 230 Th from 50 to 100 Bq m−3 ; 210 Pb from 350 to 5000 Bq m−3 ; 210 Po from 40 to 90 Bq m−3 . This section of tailings “S” is fully engineered with clay and polyethylene barriers to minimize
Uranium mining and ore processing in Ukraine
211
Fig. 3. Radionuclides in silt deposits upstream of Konoplyanka River mouth and at the Dniprovske Reservoir at different distances from the river mouth.
seepage (Soroka et al., 2002; Voitsekhovych et al., 2003). The results of the 2003 UHMI radioecological expedition show (Fig. 3) elevated contents of natural radionuclides in the bottom sediment of Dniprovske Reservoir at a distance of 80 km down stream of Dniprodzerzhinsk town. Activity concentrations of 238 U in silt deposits of the Konoplyanka River at the points upstream of the release area were 20–30 Bq kg−1 , and for uranium in water in the range 50– 100 Bq m−3 . In the soils of Ukraine, the average content of Utotal is 30–40 Bq kg−1 and 226 Ra 14–25 Bq kg−1 . At the river course down stream of area of release, the content of 238 U in water increases to 250–400 Bq m−3 and to 120–150 Bq kg−1 in silt sediment. The highest activity of uranium in silt deposits of the Konoplyanka River were estimated to be up to 500 Bq kg−1 . The maximum levels of uranium in water of the Konoplyanka River were 10–20 times higher than those in the Dniprovske Reservoir (which were in the range 10–40 Bq m−3 ). Recent data on the accumulation of natural radionuclides in aquatic plants in the Konoplyanka River within the PCP Industrial area are presented in Fig. 4. Sampling was carried out for different parts of regionally typical macrophytes, reed and rush. These occupy the coastal littoral zones of the water bodies and have a dry mass annual growth rate of up to 2.0 kg m−2 . The total amount of aquatic plant biomass (dry weight) in the area covered mainly by reed was estimated to be up to (1−2) × 103 ton of dry biomass annually. The average content of uranium (total) in the macrophyte biomass was estimated twice higher of 238 U in Fig. 4, to be 30–40 Bq kg−1 dry mass (Fig. 4). These limited data allow us to estimate the annual uranium accumulation in the macrophytes and in particular in reeds (Phragmites australis), which is in the range of 0.4–0.8 GBq of Utotal , covering about 20–30% of the Konoplyanka River valley (see also Fig. 2a.). However, the total amount of reeds in this area drained by contaminated waters seeping from tailing site “D” could be significantly higher in order to create an expanded area of the plants. This could increase the expected annual rate of removal of TENORM radionuclides from the Konoplyanka River wetland valley by 2–4 times, and make the annual rate of uranium accumulation in the aquatic plants comparable with the annual amount of radionuclides discharged into reservoirs (5–10 GBq of Utotal ). Unfortunately, there is still little reliable data on the accumulation of natural radionuclides in the aquatic species of the rivers and reservoirs in this region. Radioecological monitoring
212
O. Voitsekhovitch et al.
Fig. 4. 238 U in the biomass (dry weight) of aquatic plants in the Konoplyanka River and the Dniprovske Reservoir.
of the accumulation of radionuclides in the macrophytes, phytoplankton, benthic organisms and fish of the rivers and reservoirs should be part of any future radioecological studies in this region. Preliminary results estimate potential doses via aquatic pathways (due to different water uses) for individuals living along the rivers and reservoirs affected by releases from the uranium tailings. The radiation exposure of the population of the Dnipro River basin is caused both by naturally occurring radionuclides (40 K, 238 U, 235 U and 232 Th decay chains, etc.) and man-made radionuclides, mainly fission products; especially 137 Cs and 90 Sr (see also Fig. 1). The pathways of human exposure include external exposure from deposited gamma-emitters, and internal exposure via ingestion of contaminated food and drinking water, as well as inhalation of air-borne radionuclides. The concentration of natural radionuclides in the Dnipro basin, and associated human exposure levels, are generally close to average worldwide levels. However, in the uranium mining and milling areas in the Dnipropetrovsk region of Ukraine, concentrations of uranium and its daughter radionuclides are significantly elevated in river water due to releases to the Zheltaya River and leakage from tailings into Dnipro tributaries and ultimately into the Dnipro River itself. If river water is used for drinking and/or irrigation, elevated levels of uranium compounds and its daughter radionuclides may enter the human body. Ingestion is the major pathway of human exposure due to past and present operations of the uranium industry. However, in the immediate vicinity of uranium tailings, a person could be subjected to external exposure from gamma-radiation and to internal exposure via inhalation of radon, its daughter products and possibly tailings dust. The highest levels of human exposure, which can be expected to be potentially received by inhabitants of settlements located on the banks of Zheltaya River are 0.10–0.15 mSv yr−1 . For people who (hypothetically) consume water from the Konoplyanka River the dose is expected to be 0.01–0.05 mSv yr−1 . Doses calculated from monitoring data are included as well as estimated contributions from 230 Th and radionuclides of the 235 U chain; in total they increase the dose by about 20% (Voitsekhovych et al., 2003). These estimates show that the calculated scenario dose rates exceed the Radiation Safety Standard of Ukraine (NRBU-97) value of 0.05 mSv for water use (based on a limit of 5% of 1 mSv for the water usage pathway (NRBU-
Uranium mining and ore processing in Ukraine
213
97/D-2000)). However, these streams are relatively small and known to be highly polluted with various contaminants; therefore, this water is not used for drinking, food preparation or other domestic needs. In any case the annual dose estimates are about 0.1 mSv yr−1 , as recommended by WHO. These assessments draw attention to this problem and indicate the need for more extensive monitoring and remedial work to reduce releases of radioactive substances into the rivers.
4. Conclusions The following conclusions and recommendations based on this and other studies can be made (Soroka, 2000; Voitsekhovych et al., 2003; Soroka et al., 2000): • A better monitoring program in the region has to be established. The pollution resulting from past and present operations in the Dniprodzerzhinsk industrial complex needs to be considered holistically in order to understand their respective contribution to pollution of the Dnipro basin and the effects of interactions between the major waste storage areas. • Radiation principles are not properly applied to justify the optimal set of actions for the management of radiation protection programs. • In any rehabilitation plan, particular attention should be given to tailings “D” and the Konoplyanka River, which is acting as a conduit for transfer of pollutants from the tailings impoundment into the Dnipro River. • Conservation of the tailing sites in Zhovti Vody and Dniprodzerzhinsk can be considered as one possible strategy for the future restoration options in the region. Although the actual radiological risks for human health and environmental impact are low, a potential long-term impact can be significant. Therefore further actions to control radionuclide fluxes from the uranium tailing sites have to be justified on a cost-benefit basis and implemented.
Acknowledgements The expertise of the existing data and original field studies have been carried out by authors during 2002–2004 in the framework of the technical cooperation project RER/9/72 “Preparation of the strategic action plan for the Dnipro River basin and development of SAP implementation mechanisms with regard to radioactive contamination of the Dnipro basin and its possible consequences”, under affiliation of UNDP and IAEA. The authors are grateful to Prof. D. Levins (Australia) and Prof. M. Balonov (IAEA) for their assistance and expertise devoted to this project.
References Giddings, M. (1996). Health criteria and other supporting information, Uranium. In: Guidelines for Drinking-Water Quality, Second edition. World Health Organization, Geneva, pp. 374–381.
214
O. Voitsekhovitch et al.
Ground Water Quality Bureau (2003). Recommendations for a uranium health-based ground water standard. Prepared for the New Mexico Environment Department, 71 pp. Korovin, V. et al. (2000). Problem of radioactive pollution as a result of Uranium ores processing. In: Proc. Conf. Scientific and Technical Aspects of International Cooperation in Chernobyl. Kyiv, pp. 469–476. NRBU (2000). Radiation safety standards of Ukraine Supplement: Radiation protection from the potential irradiation sources. NRBU-97/D-(2000). Ministry of Health, Kiev, 84 pp. Soroka, Y. (2000). Identification and characterization of radioactively contaminated sites in Ukraine and planning for environmental restoration activities. Site characterization techniques used in environmental restoration activities, TECDOC-1148, IAEA, Vienna, pp. 201–218. Soroka, Y., Korovin, V., Molchanov, A. (2000). Radioactive wastes of processing of uranium ores on Pridniprovsk Chemical Plant – their influence and problems of stabilization. In: Proc. Conf. Radioactive Waste Disposal. Berlin, September 4–6, 2000, Germany: Kontek Gesellschaft für technische Kommunication GmbH, pp. 383– 388. Soroka, Y., Korovin, V., Merkulov, V. et al. (2002). Assessment of effect of radioactive wastes after uranium ore processing in environment. In: Proc. IV Conf. “Ecology and Engineering. State, Consequences and Ways for Creation of Ecologically Clean Technologies”. Dniprodzerzhinsk, pp. 253–256. TACIS (1999). Assessment of urgent measures to be taken for remediation at uranium mining and milling tailings in the CIS. Regional Project, G 4.2/93-NUCREG 9308, 1999. Voitsekhovych, O., Soroka, Yu., Lavrova, T. (2003). Uranium mining and ore processing. In: M. Balonov and D. Levins (Eds.), Radioactive Contamination of the Dnipro River Basin. IAEA, Safety Series RER 9/072, submitted for publication.
215
90 Sr, 137 Cs, 238 Pu, 239,240 Pu
and 241Am distributions in an alpine wetland, Boréon (France) Maïa Schertz* , Hervé Michel* , Geneviève Barci-Funel, Vittorio Barci Laboratoire de Radiochimie, Sciences Analytiques et Environnement, Université de Nice – Sophia Antipolis, Nice, France Abstract The study zone is situated in the south-east of France. The sampling sites were chosen to enable future determination of the mass balance of radioactive contaminants in the zone. Analyses of 90 Sr, 137 Cs, 238 Pu, 239,240 Pu and 241 Am were made on several soil and lake sediment cores using sequential radiochemical separation methods. Vertical distributions and inventories of transuranics and fission products were calculated in order to better understand their behaviour and accumulation in soil and lake sediments. Activity ratios of analysed radionuclides were used to trace their origin in the environment. While 137 Cs observed in the cores has been of the Chernobyl accident origin, transuranics have been of global fallout origin. Keywords: Radionuclides, 90 Sr, 137 Cs, 238 Pu, 239,240 Pu, 241 Am, Radiochemical separation, Alpine wetland, Soil, Lake sediment
1. Introduction In order to better understand the accumulation of anthropogenic radionuclides (90 Sr, 137 Cs, 238 Pu, 239,240 Pu and 241 Am) and their behaviour in the environment, we have been studying their distribution in samples collected in an alpine wetland situated in the south-east of France. These radionuclides have been spread out over the Earth’s surface following atmospheric nuclear weapons tests carried out mainly during the early 1960s, as well as due to the Chernobyl nuclear power plant accident in 1986 (Ukraine). Soils and lake sediments were collected and analysed for 90 Sr, 137 Cs, 238 Pu, 239,240 Pu and 241 Am using sequential radiochemical separation methods (Michel et al., 2003). The core sampling sites were selected with the aim to determine the radionuclide mass balances, as well as for studying ecosystem phenomena, such as the erosion, migration and run-off. The obtained transuranics and fission products massic activity in soils and lake sediments have * Corresponding authors. Address: Laboratoire de Radiochimie, Université de Nice – Sophia Antipolis, 28 Avenue Valrose, 06108 Nice, France; phone: (+33) 4 92 07 63 61; fax: (+33) 4 92 07 63 64; e-mail:
[email protected];
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08015-0
© 2006 Elsevier Ltd. All rights reserved.
216
M. Schertz et al.
been used for the determination of their vertical distributions and inventories. Activity ratios of analysed radionuclides were used to trace their origin in the environment.
2. Site description The study site (1765 m altitude) is a part of the Boreon Valley in the protected national park of Mercantour-Argentiera, situated in the south-eastern France (44◦ 07 N, 7◦ 20 E), and covered to a large extent by spruce forests. Our study zone is an alpine wetland with its surroundings, constituted by a small lake (45 m long, 14 m wide, 630 m2 area and 0.5 m mean depth), a clearing and a steep catchment (2 km2 area and 2600 m maximum relief ) on the north and the east of the lake. The lake is supplied with water via two inlets: a stream (on the north) and a catchment run-off (on the south-east). This site was chosen for its suitable configuration, as the steep slope of the catchment area and run-off could provoke an accumulation of radionuclides in the clearing and in the lake sediments.
3. Sampling Figure 1 shows the study site with locations of soil and lake sediment samples, collected on June 21, 2002. The sediment cores were collected using plastic tubes of 10 cm in diameter. The soil cores were sampled using metal tubes of 16 cm in diameter. All the cores were extruded vertically and cut in centimetre layers. All collected samples were dried at 100◦ C
Fig. 1. Study site.
90 Sr, 137 Cs, 238 Pu, 239,240 Pu and 241 Am distributions in an alpine wetland, Boréon (France)
217
until constant weight was reached. They were then crushed to obtain a homogeneous powder, in which larger particles were discarded by sieving (2 mm mesh). Five lake sediment cores were collected at these sites: close to the stream water inlet (cores L1 and L2), in the middle of the lake (core M), near the run-off water inlet (core R2) and in the vicinity of the lake outflow (core R1). Five soil cores were collected at these sites: on the clearing (cores CL and CR), on the catchment slope (cores SL and SR) and on the catchment summit (core AF, not visible in Fig. 1). The summit site has not been eroded and not subjected to run-off from a higher altitude, therefore the core AF has been used for determination of the atmospheric fallouts, which have contaminated the site.
4. Methods Previously developed sequential radiochemical separation methods (Michel et al., 2003) were used to analyse fission products (90 Sr and 137 Cs) and transuranics (238 Pu, 239,240 Pu and 241 Am) in collected soil and sediment samples. Gamma-ray spectrometry (HPGe detector (ORTEC) with relative detection efficiency of 17%, and the energy resolution of 1.9 keV at 1.33 MeV) has been used to determine 137 Cs activities in the samples. Several co-precipitation steps and chromatographic separations on specific resins were applied on an aliquot of the sample in order to analyse the other elements (Fig. 2). 238 Pu,
Fig. 2. Different steps in the radiochemical separation method.
218
M. Schertz et al.
239,240 Pu
and 241 Am activities were determined by alpha-spectrometry, using boron doped silicon detectors (ORTEC) with 22% efficiency. 90 Sr was analysed using the liquid scintillation spectrometer (PACKARD Tri-CARB® 2000CA), having 68% detection efficiency for 90 Sr. The radiochemical methods have been validated using IAEA reference materials for soils (IAEA 375 and soil No. 6) and sediments (IAEA 135 and 385). The obtained activities were within the 95% confidence intervals.
5. Results and discussion Using the measured massic activities it has been possible to describe the activity profiles of soil and sediment cores, to calculate the radionuclide inventories which have been used to define different contamination zones at the site, and to calculate the activity ratios that have been used to determine the origin of radionuclides (Schertz et al., 2006). For all studied radionuclides and for all the cores, the activity profiles have similar characteristics. Representative soil and sediments profiles are shown for caesium in Fig. 3, for strontium in Fig. 4 and for transuranics in Fig. 5. For all the samples and all the analysed radionuclides, the measured activities have been above the detection limits. Higher 137 Cs activities have been measured in the soil than in the sediment samples. For 90 Sr both soil and sediment samples have shown similar contamination levels. Contrary to the 137 Cs, contamination from transuranics is more important in the sediment than in the soil. For all studied radionuclides and for all the cores, the activity profiles generally peak in the upper few centimetres. Nevertheless, the deepest core slices have still significant radionuclide activities. This shows that radionuclides have migrated deep into the soil and sediment (Goryachenkova et al., 1991). These repartitions are linked to the organic mat-
Fig. 3. Representative profiles of 137 Cs in soil and sediment.
90 Sr, 137 Cs, 238 Pu, 239,240 Pu and 241 Am distributions in an alpine wetland, Boréon (France)
219
Fig. 4. Representative profiles of 90 Sr in soil and sediment.
Fig. 5. Representative profiles of transuranics in soil and sediment.
ter percentage of the samples. Figure 6 shows the depth distribution of organic matter in sediment and soil cores, which shows similar variations as the radionuclide activity profiles (Lee et al., 1997). Organic matter level plays an important role in the retention and migration of radioelements. Even if this influence is distinct on caesium, strontium and transuranics, the general behaviour of these elements is the same (Bunzl et al., 1995; Szabo et al., 1997). Calculated radionuclide inventories for different sampling sites are presented in Table 1. For all the radionuclides studied, the two cores R1 and R2 contribute 60% to the total lake
220
M. Schertz et al.
Fig. 6. Representative profiles of organic matter in soil and sediment. Table 1 Radionuclide inventories in soil and sediment cores Core
90 Sr (102 Bq m−2 )
137 Cs (103 Bq m−2 )
239,240 Pu (101 Bq m−2 )
241 Am (101 Bq m−2 )
L1 L2 M R2 R1 SL CL SR CR AF
6±1 9±2 6±1 20 ± 2 6±2 6±1 5±1 4±1 5±2 5±1
30 ± 1 43 ± 1 24 ± 5 52 ± 1 52 ± 1 24 ± 1 15 ± 1 19 ± 4 37 ± 2 11 ± 1
23 ± 1 28 ± 3 20 ± 1 46 ± 1 38 ± 3 15 ± 2 11 ± 1 9±1 13 ± 1 12 ± 1
9±1 13 ± 2 8±1 15 ± 1 10 ± 2 6±2 5±1 4±1 4±1 5±1
contamination, whereas the three cores L1, L2 and M contribute only 40%. Therefore the lake can be divided in two different contamination zones (A and B as shown in Fig. 1), which represent the two lake contamination contributions: the catchment run-off on the south-east side and the stream contribution on the north-west side. Thus catchment run-off contribution is more important than stream contribution. The comparison between the core inventories for 137 Cs shows an important run-off in this site. Indeed the core AF (which represents the atmospheric fallout) is less contaminated than the other soil cores. Thus the catchment was contaminated directly by atmospheric fallout, but also indirectly by run-off on a higher altitude zone. For the other radionuclides, the core AF has about the same contamination than the other soil cores. Consequently in our site, the caesium has a greater mobility than strontium, americium and plutonium. Another point is the difference in contamination between the slope and the clearing. On the north-west side of the catchment area, the core SL is more contaminated than the core
90 Sr, 137 Cs, 238 Pu, 239,240 Pu and 241 Am distributions in an alpine wetland, Boréon (France)
221
Table 2 Comparison of measured activity ratios with theoretical ones Activity ratio
90 Sr/137 Cs
238 Pu/239,240 Pu
241 Am/239,240 Pu
Mean activity ratio Global fallout ratio Chernobyl theoretical ratio
0.028 ± 0.005 0.61 0.02
0.06 ± 0.02 0.03 0.44
0.40 ± 0.09 0.40
CL; contrary to the south-east side, where the core SR is less contaminated than the core CR. These data are linked to the location of the vegetation in the area, as shown in Fig. 1. The catchment area on the south-east side is covered by a thick vegetation, contrary to the north-west side, therefore organic matter levels are more important on the south-east side. For the catchment area on north-west side, where organic matter levels are low, radioelements have been retained on the catchment slope. Conversely, for the catchment area on the south-east side, where organic matter levels are high, radioelements have not retained on the catchment slope, and there has been an accumulation in the clearing. Therefore rich organic matter zones are more noticeable to run-off. Since the soil radionuclide inventories are lower than the sediment inventories, an important transfer of radionuclides occurs by run-off from the catchment to the lake water, and then to sediment due to fast sedimentation rates of some of the radionuclides. In relation to the total area of the site, 99.95% of the activity is spread out in the soil and only 0.05% is in the sediments. Consequently, the radioelement transfer and storage mechanism is very slow, especially in our site because the catchment area (2 km2 ) is much larger than the lake area (630 m2 ). The last study task was to determine radionuclide activity ratios. The Chernobyl accident and global fallout from atmospheric nuclear tests have specific activity ratios, therefore by comparison with our results, we can deduce the origin of the radionuclides. Table 2 compares the measured mean activity ratios with the expected ratios from global fallout and the Chernobyl accident. The observed mean activity ratio of 90 Sr/137 Cs (0.028 ± 0.005) is close to the Chernobyl accident theoretical ratio (0.02). Hence the fission products detected in our study site came from the Chernobyl accident (Irlweck and Khademi, 1996). The contamination due to global fallout is not visible because it is much lower in comparison with the Chernobyl fallout. The observed mean activity ratio of 238 Pu/239,240 Pu (0.06 ± 0.02) is the same as the global fallout ratio (0.03). The observed mean activity ratio of 241 Am/239,240 Pu (0.40 ± 0.09) is also the same as the global fallout ratio (0.40). Therefore the transuranics measured in our study site came from the atmospheric nuclear weapons tests (Momoshima et al., 1997), as we have not found any transuranics contaminants from the Chernobyl accident (Bunzl et al., 1994).
Acknowledgement The authors thank Philippe Abela for technical assistance provided from the first steps of this study.
222
M. Schertz et al.
References Bunzl, K., Förster, H., Kracke, W., Schimmack, W. (1994). Residence times of fallout 239+240 Pu, 238 Pu, 241 Am and 137 Cs in the upper horizons of an undisturbed grassland soil. Journal of Environmental Radioactivity 22, 11–29. Bunzl, K., Flessa, H., Kracke, W., Schimmack, W. (1995). Association of fallout 239+240 Pu and 241 Am with various soil components in successive layers of a grassland soil. Environmental Science and Technology 29 (10), 2513– 2518. Goryachenkova, T.A., Pavlotskaya, F.I., Myasoedov, B.F. (1991). Forms of occurrence of plutonium in soils. Journal of Radioanalytical and Nuclear Chemistry 147 (1), 153–157. Irlweck, K., Khademi, B. (1996). Radionuclide ratios of 90 Sr/137 Cs and 239(240) Pu/137 Cs in contaminated surface air after the Chernobyl accident in Austria. Journal of Radioanalytical Chemistry 203 (1), 79–85. Lee, M.H., Lee, C.W., Boo, B.H. (1997). Distribution and characteristics of 239,240 Pu and 137 Cs in the soil of Korea. Journal of Environmental Radioactivity 37 (1), 1–16. Michel, H., Schertz, M., Barci-Funel, G., Ardisson, G. (2003). Sequential radiochemical separations from Alpine Wetland soils (Boréon, France) with emphasis on 90 Sr measurement. Journal of Radioanalytical and Nuclear Chemistry 258 (2), 209–213. Momoshima, N., Kakiuchi, H., Maeda, Y., Hirai, E., Ono, T. (1997). Identification of contamination source of plutonium in environmental samples with isotopic ratios determined by inductively coupled plasma mass spectrometry and alpha-spectrometry. Journal of Radioanalytical and Nuclear Chemistry 221, 213–217. Schertz, M., Michel, H., Barci-Funel, G., Barci, V. (2006). Transuranic and fission product contamination in lake sediments from an alpine wetland, Boréon (France). Journal of Environmental Radioactivity 85 (2/3), 380–388. Szabo, G., Guczi, J., Nisbet, A. (1997). Investigation of the solid phase speciation of 90 Sr, 137 Cs, 239 Pu and 241 Am in soils determined by extraction and ultra-filtration methods. Journal of Radioanalytical and Nuclear Chemistry 226 (1–2), 255–259.
223
Concentrations and characteristics of uranium isotopes in drinking waters collected in Italy and the Balkan regions and their radiological impact on the public Guogang Jia* , Giancarlo Torri, Umberto Sansone, Piera Innocenzi, Silvia Rosamilia, Antonio Di Lullo, Stefania Gaudino Italian Environmental Protection Agency and Technical Services, 00144 Roma, Italy Abstract The concentrations and characteristics of uranium isotopes in drinking waters of Italy and the Balkan regions were studied. The obtained mean uranium activity concentrations in drinking waters in Italy are 20.6 ± 29.0 mBq l−1 for 238 U, 25.3 ± 35.6 mBq l−1 for 234 U and 1.00 ± 1.27 mBq l−1 for 235 U. The mean activity ratios are 1.33 ± 0.33 for 234 U/238 U and 0.047±0.038 for 235 U/238 U. The activity concentrations of uranium isotopes in waters of the Balkan regions are much lower than those in Italy. Depleted uranium has been detected in some water samples collected in the Balkan regions, but the measured concentrations do not constitute a health risk. The estimated average annual committed effective doses to the public due to water intake were 1.62 µSv yr−1 in Italy, 0.30 µSv yr−1 in Kosovo, 0.42 µSv yr−1 in Serbia and Montenegro, and 0.21 µSv yr−1 in Bosnia and Herzegovina. Keywords: Natural radionuclides, 238 U, 234 U, 235 U, Water, α-spectrometry, Radiation dose
1. Introduction Uranium is a naturally occurring, ubiquitous heavy metal found in the environment in various chemical forms in all soils, rocks, seas and oceans. It is also present in air, food and drinking water. Due to its widespread existence in nature, uranium isotopes (238 U, 234 U and 235 U) have become the most important sources of natural radioactivity and may contribute to some degree to external and internal doses to population. Uranium isotopes enter the human body mainly through ingestion, and by inhalation to a considerably smaller degree (UNSCEAR, 1993; Fisenne et al., 1987). Hence the internal radiation exposure for members of the public can be evaluated through the intake of the radionuclides from both food and water. Pietrzak-Flis et al. (2001) investigated the uranium * Corresponding author. Address: Italian Environmental Protection Agency and Technical Services, Via V. Brancati 48, 00144 Roma, Italy; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08016-2
© 2006 Elsevier Ltd. All rights reserved.
224
G. Jia et al.
intake fractions of the different pathways, and concluded that the uranium intake from water by man is the most important pathway which contributes the largest fraction (75.1%–76.9%) of total uranium intake. For the purpose of evaluation of the radiological impact of uranium isotopes in drinking water to the residents, water samples were collected in Italy and Balkan regions. The activity concentrations of 238 U, 234 U and 235 U in collected water samples were measured in the laboratory of the Italian Environmental Protection Agency and Technical Services (APAT). Based on the obtained activity concentrations of uranium isotopes in drinking water, annual consumption rate of the water, and the dose conversion factor per unit intake given by the International Commission on Radiological Protection (ICRP, 1996), the annual effective dose resulting from 1-year ingestion for the adult population was estimated. The obtained data can serve as basic information for evaluation of the radiological effects of uranium isotopes on human health of the regions.
2. Materials and methods 2.1. Apparatus and reagents The uranium sources were counted by α-spectrometry (Canberra, U.S.A.) with a counting efficiency of 31.2% and a background of less than or equal to 2×10−6 s−1 in the interested energy region. The apparatus for uranium electrodeposition (model PL320QMD; Thurlby Thandar Instruments Ltd, England) was composed of a Perspex cell of 25 mm internal diameter and a stainless-steel disk of 20 mm diameter. Chromatographic columns were 150 mm long and 9 mm internal diameter. U-232 standard solution, Microthene (microporous polyethylene, 60–140 mesh), tri-octylphosphine oxide (TOPO, 99%) and reference material (IAEA-381, seawater) were supplied by Amersham (G.B.), Ashland (Italy), Fluka (Switzerland) and IAEA (Monaco) respectively. FeCl3 was used to prepare the carrier solution for uranium separation in water sample and all other reagents were analytical grade. 2.2. Column preparation A solution (50 ml) of 0.3 M TOPO in cyclohexane was added to 50 g of Microthene; the mixture was stirred for several minutes until homogeneous and was then evaporated to eliminate cyclohexane at 50◦ C. The porous powder thus obtained contained about 10.4% TOPO. A portion (1.6 g) of the Microthene–TOPO powder, slurred with 3 ml concentrated HCl and some water, was transferred to a chromatographic column; after conditioning with 30 ml of 2 M HNO3 , the column was ready for use. 2.3. Sampling and sample preparation Drinking water is defined as water for everyday drinking, including commercially available bottled mineral waters, tap water, well water, lake water, river water, channel water and stream water, etc. The bottled mineral waters consumed in Italy were obtained from a supermarket
Uranium isotopes in drinking waters in Italy and the Balkan regions
225
in April 2004. Seventeen brands of mineral water samples were collected and most of them are very popular in Italy. About 10 litres of water for each brand were taken and part of it was used for uranium analyses. In order to verify a presence of depleted uranium (DU) in drinking water, and to evaluate its possible radiological impact, well water, tap water, lake water, river water, channel water and stream water samples from Balkan regions (Kosovo, Serbia and Montenegro, Bosnia and Herzegovina) were collected during UNEP missions from November 2000 to October 2002. The waters were sampled at the sites using polyethylene bottles. Immediately after sampling, without any filtering, the pHs of all water samples were adjusted to <2 by adding HNO3 to ensure that trace elements were kept in solution and to inhibit biological growth. More information about sampling sites is available in the literature (Jia et al., 2004, 2005; UNEP, 2001, 2002, 2003). 2.4. Method 2.4.1. Preconcentration of uranium in water Forty mg of Fe3+ (40 mg Fe3+ ml−1 ) as carrier, 0.03 Bq of 232 U as tracer and 20 ml of concentrated HNO3 are added to one litre of filtered water sample in a beaker. After boiling for 30 min, the beaker is removed to an electric-magnetic stirrer and the solution is adjusted to pH 9.5–10 with concentrated ammonia solution to co-precipitate uranium with iron (III) hydroxide. The solution is stirred for another 30 min and the precipitate is allowed to settle down for at least 4–6 h and preferable overnight. The supernatant is carefully syphoned off and the hydroxide slurry is centrifuged at 4000 rpm. The supernatant is discarded, the precipitate is dissolved with 3 ml of concentrated HNO3 , 17 ml of water and a few drops of 40% HF by heating. 2.4.2. Separation and determination of uranium The obtained solution passed through a preconditioned Microthene–TOPO column at a flow rate of 0.6–0.8 ml min−1 . After washing with 30 ml of 6 M HNO3 , 60 ml of 1 M HCl and 5 ml of water at the same flow rate, uranium is eluted with 30 ml of 0.025 M (NH4 )2 C2 O4 at a flow rate of 0.1 ml min−1 . The first 3 ml of eluant are discarded and the remains are directly collected in an electrodepositing cell. 0.62 ml of 8 M HNO3 is added to the cell and the solution is adjusted to pH 1.0–1.5 with 1:4 ammonia solutions. Uranium is electrodeposited on a stainless steel disk at a current density of 400 mA cm−2 for 4 h and counted by α-spectrometry. 2.4.3. Quality control of the method In order to assess the reliability of the uranium determination method in water samples, the IAEA-381 reference material (Irish Sea water) was analysed, and the obtained results are shown in Table 1. It is seen that the obtained 238 U, 234 U and 235 U activity concentrations are 38.1 ± 0.7 mBq kg−1 , 43.4 ± 0.9 mBq kg−1 and 2.00 ± 0.23 mBq kg−1 respectively, which are in a reasonable agreement with the IAEA information values (the 95% confidence interval is (38–48) mBq kg−1 for 238 U, (43–58) mBq kg−1 for 234 U and (1.4–2.4) mBq kg−1 for 235 U).
226
G. Jia et al.
Table 1 238 U, 234 U and 235 U massic activities (mBq kg−1 ) measured in the IAEA-381 reference material (Irish Sea water)∗ Sample No.
Sample weight (g)
U yield (%)
238 U
234 U
235 U
234 U/238 U
235 U/238 U
1 2 3 4
1000 1000 1000 1000
66.5 76.5 63.4 73.3
38.5 ± 1.4 37.9 ± 1.4 37.2 ± 1.4 38.9 ± 1.4
43.8 ± 1.6 42.2 ± 1.5 43.3 ± 1.6 44.3 ± 1.6
1.94 ± 0.20 2.20 ± 0.20 1.70 ± 0.19 2.16 ± 0.20
1.14 1.11 1.16 1.14
0.050 0.058 0.046 0.056
Mean ± 1σ
1000
69.9 ± 6.0
38.1 ± 0.7
43.4 ± 0.9
2.00 ± 0.23
1.14 ± 0.02
0.052 ± 0.005
∗ Information values (95% confidence interval) are 41 (38–48) mBq kg−1 for 238 U; 50 (43–58) mBq kg−1 for 234 U; 2.3 (1.4–2.4) mBq kg−1 for 235 U.
2.4.4. Lower limit of detection (LLD) The LLD is defined here as the 3σ of the blank count rate. Based on the counting efficiencies of the instruments (31.2%), the average radiochemical yield (77.3%), counting time (6 days) and the sample quantity (1 litre), the LLD of the method is 0.22 mBq l−1 for 238 U and 234 U, and 0.022 mBq l−1 for 235 U. 3. Results and discussion Uranium is present in natural water mainly as a result of leaching from mineral deposits. Uranium concentrations in waters vary from region to region due to different rocks composing the aquifer, the water composition and the distance from uraniferous areas. Uranium levels in drinking water are generally less than 1 µg l−1 , although concentrations as high as 700 µg l−1 have been measured in private supplies (WHO, 2004). It is reported that the typical 234 U/238 U activity ratios in natural water samples range from 0.8 to 10, while 235 U/238 U activity ratio is thought to have a quite uniform value of about 0.046 (Goldstein et al., 1977). The variation of the activity ratios between uranium isotopes is a very important characteristic, which can be used to distinguish the natural uranium (NU) from natural depleted uranium (NDU), reprocessed depleted uranium (RDU) and unprocessed depleted uranium (UDU). The differences in activity ratios can provide evidences for the uranium origin. The obtained activity concentrations of uranium isotopes in water samples collected in Italy and Balkan regions are given in Tables 2–5. The uncertainties given for individual analysis represent 1σ standard deviations, which are estimated from the uncertainties associated with the tracer (232 U) activity, the addition of the tracer to the sample, counting statistics of the samples, the blanks, etc. The DU fraction in possible DU contaminated water sample was calculated using the formula given by Jia et al. (2005). 3.1. Activity concentrations of uranium isotopes in Italian waters As shown in Table 2, the uranium activity concentrations in drinking waters in Italy are in the range of 0.21–103 (mean ±1σ deviation: 20.6 ± 29.0) mBq l−1 for 238 U, of 0.25–135 (25.3 ± 35.6) mBq l−1 for 234 U, and of LLD − 4.10 (1.00 ± 1.27) mBq l−1 for 235 U. The mean
Table 2 238 U, 234 U and 235 U activity concentrations (mBq l−1 ) in mineral water samples collected in April 2004 in Italy Name and origin
Water volume (l)
U yield (%)
238 U
234 U
235 U
234 U/238 U
235 U/238 U
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Blues Aura, Umbria Egeriam, Roma Guizza, Pescara Panna, Firenze Rocchetta-A, Perugia Lete, Caserta Vitasnella, Brescia San Gemini, Terni Rocchetta-B, Perugia Vera, Padova San Benedetto, Vinece Lieve, Perugia Ferrarelle, Caserta Uliveto, Pisa CSM tap water, Roma Maglana tap water, Roma Capannelle, Roma
1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 2.000 1.500 1.000 0.500
88.0 88.1 71.1 95.2 86.8 87.6 64.3 65.2 70.1 53.9 82.4 86.0 70.7 68.4 73.4 73.5 70.4
16.8 ± 0.9 59.8 ± 2.4 6.14 ± 0.29 7.21 ± 0.38 2.33 ± 0.15 20.5 ± 1.2 78.5 ± 4.2 14.2 ± 0.7 1.39 ± 0.20 12.0 ± 0.8 12.4 ± 0.7 2.91 ± 0.18 16.0 ± 0.8 4.13 ± 0.23 0.21 ± 0.06 7.22 ± 0.36 103 ± 3
17.1 ± 0.9 84.4 ± 3.3 6.43 ± 0.30 13.2 ± 0.6 3.37 ± 0.19 23.3 ± 1.3 71.4 ± 3.8 13.4 ± 0.7 2.09 ± 0.23 15.3 ± 1.0 21.6 ± 1.0 5.02 ± 0.24 20.7 ± 1.0 8.29 ± 0.37 0.25 ± 0.07 7.55 ± 0.37 135 ± 4
0.86 ± 0.17 3.31 ± 0.33 0.29 ± 0.05 0.35 ± 0.08 0.16 ± 0.04 0.98 ± 0.21 4.10 ± 0.54 0.42 ± 0.11 0.08 ± 0.05 0.99 ± 0.21 0.68 ± 0.14 0.20 ± 0.05 1.22 ± 0.18 0.27 ± 0.05
1.02 ± 0.08 1.41 ± 0.08 1.05 ± 0.07 1.83 ± 0.12 1.44 ± 0.12 1.14 ± 0.09 0.910 ± 0.068 0.947 ± 0.067 1.51 ± 0.27 1.28 ± 0.12 1.75 ± 0.13 1.72 ± 0.13 1.30 ± 0.09 2.01 ± 0.14 1.21 ± 0.49 1.05 ± 0.07 1.31 ± 0.05
0.051 ± 0.010 0.055 ± 0.006 0.047 ± 0.009 0.048 ± 0.011 0.070 ± 0.018 0.048 ± 0.010 0.052 ± 0.007 0.030 ± 0.008 0.054 ± 0.039 0.083 ± 0.018 0.055 ± 0.012 0.069 ± 0.016 0.077 ± 0.012 0.065 ± 0.014 ND∗ 0.071 ± 0.014 0.034 ± 0.004
0.500–2.000
77.3 ± 11.8
0.21–103
0.25–135
1.33 ± 0.33
0.047 ± 0.038
Mean or range
Uranium isotopes in drinking waters in Italy and the Balkan regions
Sample code
∗ ND: not detectable.
227
228
G. Jia et al.
activity ratios are 1.33 ± 0.33 for 234 U/238 U and 0.047 ± 0.038 for 235 U/238 U. It is seen that (1) the concentrations of uranium isotopes in analysed waters exhibit large variations; (2) the 235 U concentrations in the samples are low but measurable; (3) the radioactive disequilibria between 234 U and 238 U concentrations were observed in most of the samples; and (4) all the data in Table 2 indicate characteristics of natural uranium. Different countries and organisations have different regulations or guidelines for the drinking water quality evaluation. The WHO (1998) derived a guideline for drinking water uranium quality of 2 µg l−1 (24.9 mBq 238 U l−1 ). This value is considered to be protective for subclinical renal effects reported in epidemiological study. The uranium concentrations in most of the drinking waters in Italy are below the derived guideline value of WHO (1998), only three (about 18%) are over this limit. Currently the WHO (2004) has revised the guideline from 2 µg l−1 to 15 µg l−1 . It is reported that uranium concentrations are particularly high in granitic rocks and pegmatites, and in areas with sulfide mineralisation (UNEP, 2001). All the drinking waters in Table 2 are mineral waters. Geological condition, where there is high radiation background, high concentration of organic matter, iron hydroxide, carbonaceous material, clay minerals or sulphides, are the most important factors for concentrating uranium in drinking water. In fact, in some Italian ares, especially in the volcanic regions, minerals contain high levels of natural uranium and thorium. Due to the complex and/or redox reactions in water, uranium can be leached out and became soluble, e.g. uranyl carbonate (UO2 CO3 ) which is formed by action of CO2 under pressure on UO2 2+ (Partington, 1966). This could be the reason why the uranium concentrations in some Italian mineral waters are high. 3.2. Uranium isotopes in waters of Balkan regions During the field missions in Balkan regions (Kosovo, Serbia and Montenegro, Bosnia and Herzegovina) a great attention has been paid to possible water contamination with DU. The activity concentrations of uranium isotopes in water samples collected in Kosovo (Table 3) are in the range of 0.29–20.0 (3.80 ± 5.32) mBq l−1 for 238 U, 0.26–26.5 (4.77 ± 7.09) mBq l−1 for 234 U and 0.03–0.98 (0.20 ± 0.25) mBq l−1 for 235 U. The mean activity ratios range from 0.482 to 1.86 for 234 U/238 U and from 0.035 to 0.153 for 235 U/238 U. As shown in Tables 4 and 5, the activity concentrations of 238 U, 234 U and 235 U in waters of Serbia, Montenegro, Bosnia, and Herzegovina are similar to that of Kosovo. Looking at the data in Tables 3–5, it is observed that (1) the uranium isotope concentrations in the waters of Balkan regions are much lower than those found in Italy, (2) the uranium activities in most of the samples are dominated by the source of natural uranium, and (3) the uranium concentrations in water of Balkan regions are well below the guideline derived by WHO (1998, 2004) for drinking water. Moreover, three samples collected from Kosovo (WW1, WW41 and WW43) and two samples collected from Serbia and Montenegro (WW15 and WW16) show an anomalously 234 U/238 U activity ratios of 0.482–0.863, and the estimated DU fraction (m/m) in the samples are in the range of 30.4%–65.3%, indicating possible presence of anthropogenic DU in the samples. DU do appear in one sample collected from Bosnia and Herzegovina (WW3). However, low uranium concentrations associated with high relative uncertainties require further investigations.
Sample code
Water volume (l)
U yield (%)
238 U
234 U
235 U
234 U/238 U
235 U/238 U
FDU (%)
WW1 WW45 WW39 WW44 WW43 WW40 WW41 WW59 WW60
1.000 1.000 0.935 0.988 0.970 1.000 0.950 1.000 0.950
82.5 88.4 78.4 69.1 73.9 79.5 75.0 66.0 60.1
0.54 ± 0.10 0.64 ± 0.10 20.0 ± 1.0 0.51 ± 0.11 2.37 ± 0.20 3.88 ± 0.24 0.55 ± 0.10 0.29 ± 0.09 2.68 ± 0.19
0.26 ± 0.08 0.74 ± 0.11 26.5 ± 1.2 0.60 ± 0.12 1.98 ± 0.18 4.43 ± 0.26 0.31 ± 0.09 0.31 ± 0.09 5.00 ± 0.27
0.03 ± 0.03 0.07 ± 0.03 0.98 ± 0.18 0.04 ± 0.03 0.19 ± 0.06 0.20 ± 0.05 0.06 ± 0.03 0.04 ± 0.03 0.18 ± 0.05
0.482 ± 0.18 1.15 ± 0.25 1.33 ± 0.09 1.18 ± 0.34 0.836 ± 0.10 1.14 ± 0.10 0.565 ± 0.19 1.07 ± 0.46 1.86 ± 0.17
0.062 ± 0.050 0.105 ± 0.054 0.049 ± 0.009 0.083 ± 0.067 0.081 ± 0.025 0.051 ± 0.015 0.109 ± 0.061 0.153 ± 0.119 0.066 ± 0.020
65.3 ND ND ND 30.4 ND 57.1 ND ND
RW46
0.895
71.2
0.89 ± 0.13
1.05 ± 0.14
0.10 ± 0.04
1.18 ± 0.23
0.117 ± 0.053
ND
CW42
1.000
80.3
5.42 ± 0.31
5.49 ± 0.31
0.22 ± 0.06
1.01 ± 0.08
0.041 ± 0.012
12.9
LW47
0.970
71.7
6.77 ± 0.43
7.63 ± 0.46
0.24 ± 0.08
1.13 ± 0.10
0.035 ± 0.012
ND
SRW58
0.967
59.0
4.93 ± 0.30
7.76 ± 0.39
0.27 ± 0.07
1.58 ± 0.12
0.056 ± 0.015
ND
Mean or range
0.895–1.000
73.5 ± 8.6
0.29–20.0
0.26–26.5
0.03–0.98
0.482–1.86
0.035–0.153
ND–65.3
Uranium isotopes in drinking waters in Italy and the Balkan regions
Table 3 238 U, 234 U and 235 U activity concentrations (mBq l−1 ) in well water (WW), river water (RW), channel water (CW), lake water (LW) and stream water (SRW) samples collected in Kosovo in November 2000
229
230
Table 4 238 U, 234 U and 235 U activity concentrations (mBq l−1 ) in tap water (TW) and well water (WW) samples collected in Serbia and Montenegro in 2001 Sample code
Water volume (l)
U yield (%)
238 U
234 U
235 U
234 U/238 U
235 U/238 U
FDU (%)
0.500
86.2
4.54 ± 0.37
5.09 ± 0.39
0.34 ± 0.10
1.12 ± 0.13
0.076 ± 0.023
ND
0.500
85.6
1.17 ± 0.23
1.31 ± 0.24
0.09 ± 0.06
1.12 ± 0.30
0.077 ± 0.056
ND
TW8
0.500
91.3
0.83 ± 0.17
1.44 ± 0.10
0.051 ± 0.011
ND
TW13
0.990
61.8
5.20 ± 0.38
5.79 ± 0.40
0.28 ± 0.09
1.11 ± 0.11
0.054 ± 0.018
ND
WW14 WW15 WW16 WW17 WW23 WW27
1.000 1.000 1.000 1.000 1.000 1.000
60.9 63.0 49.5 54.5 61.3 40.1
0.58 ± 0.08 0.40 ± 0.10 0.52 ± 0.11 0.61 ± 0.14 21.9 ± 0.9 0.56 ± 0.12
0.74 ± 0.08 0.27 ± 0.09 0.43 ± 0.11 0.86 ± 0.15 28.1 ± 1.2 0.83 ± 0.14
0.04 ± 0.02 0.01 ± 0.03 0.02 ± 0.03 0.03 ± 0.04 0.88 ± 0.14 0.02 ± 0.02
1.26 ± 0.22 0.660 ± 0.277 0.816 ± 0.278 1.41 ± 0.39 1.28 ± 0.08 1.47 ± 0.40
0.074 ± 0.034 0.036 ± 0.065 0.036 ± 0.065 0.045 ± 0.064 0.040 ± 0.007 0.044 ± 0.032
ND 47.7 32.3 ND ND ND
Mean or range
0.500–1.000
65.4 ± 16.9
0.40–21.9
0.27–28.1
0.01–0.88
0.660–1.47
0.036–0.077
ND–47.7
16.2 ± 0.8
23.3 ± 1.0
G. Jia et al.
TW6 WW7
Sample code
Water volume (l)
U yield (%)
RW2
1.000
98.7
238 U
4.81 ± 0.26 12.2 ± 0.6
234 U
235 U
234 U/238 U
235 U/238 U
FDU (%)
5.34 ± 0.27
0.29 ± 0.06
1.11 ± 0.08
0.061 ± 0.013
ND
WW3
0.500
91.8
7.09 ± 0.45
0.41 ± 0.11
0.581 ± 0.05
0.034 ± 0.009
55.5
TW5
1.000
81.4
0.66 ± 0.12
0.95 ± 0.12
0.06 ± 0.03
1.44 ± 0.31
0.097 ± 0.053
ND
RW6
1.000
82.4
1.30 ± 0.14
2.44 ± 0.18
0.16 ± 0.05
1.87 ± 0.25
0.126 ± 0.040
ND
TW7 TW8
1.000 1.000
90.2 82.4
1.29 ± 0.13 0.60 ± 0.09
1.81 ± 0.15 1.02 ± 0.10
0.10 ± 0.04 0.06 ± 0.03
1.41 ± 0.19 1.70 ± 0.31
0.075 ± 0.030 0.099 ± 0.046
ND ND
RW9 RW10 RW11 RW12
0.922 1.000 1.000 1.000
79.3 77.4 67.2 60.4
0.33 ± 0.06 4.29 ± 0.25 1.30 ± 0.12 1.52 ± 0.21
0.41 ± 0.06 5.13 ± 0.28 2.95 ± 0.17 3.48 ± 0.30
0.01 ± 0.02 0.36 ± 0.07 0.07 ± 0.03 0.07 ± 0.06
1.24 ± 0.29 1.20 ± 0.09 2.27 ± 0.25 2.29 ± 0.38
0.036 ± 0.068 0.084 ± 0.017 0.055 ± 0.025 0.047 ± 0.037
ND ND ND ND ND
TW15
1.000
81.5
0.84 ± 0.10
1.90 ± 0.13
0.08 ± 0.03
2.26 ± 0.31
0.095 ± 0.036
RW17
0.875
69.6
2.11 ± 0.17
2.19 ± 0.17
0.17 ± 0.05
1.04 ± 0.11
0.080 ± 0.024
10.4
TW18
1.000
68.4
4.05 ± 0.29
4.67 ± 0.31
0.28 ± 0.08
1.15 ± 0.11
0.070 ± 0.020
ND
RW19
1.000
82.0
3.87 ± 0.20
5.23 ± 0.24
0.23 ± 0.05
1.35 ± 0.09
0.061 ± 0.013
ND
Mean or range
0.500–1.000
79.5 ± 10.4
0.33–12.2
0.41–7.09
0.01–0.41
0.581–2.29
0.034–0.126
ND–55.5
Uranium isotopes in drinking waters in Italy and the Balkan regions
Table 5 238 U, 234 U and 235 U activity concentrations (mBq l−1 ) in tap water (TW), well water (WW) and river water (RW) samples collected in Bosnia and Herzegovina in 12–24 October 2002
231
232
G. Jia et al.
3.3. Radiological impacts due to the uranium intake from water The annual consumption rate of water per person is assumed to be 730 l yr−1 (WHO, 1998). Using the observed activity concentrations in waters, the annual intake rate, and the dose conversion factors for uranium isotopes per unit intake given by the ICRP (1996), the annual committed effective doses for the adult population were estimated. As shown in Tables 6–9, Table 6 The annual effective doses for the adult population in Italy estimated from the annual intake rates of uranium which are calculated from the mean and maximum 238 U, 234 U and 235 U activity concentrations (mBq l−1 ) in waters Radionuclide
Intake rate (Bq yr−1 )
Concentration (mBq l−1 ) Mean
Maximum
Mean
235 U
20.6 ± 29.0 25.3 ± 35.6 1.00 ± 1.27
103 135 3.41
15.0 18.5 0.73
Total
46.9
241
34.2
238 U 234 U
Dose coefficient factor (µSv Bq−1 )
Maximum 75.2 98.6 2.49 176
The effective doses for the adult population (µSv yr−1 ) Mean
Maximum
0.045 0.049 0.047
0.68 0.91 0.03
3.38 4.83 0.12
–
1.62
8.33
Table 7 The annual effective doses for the adult population in Kosovo estimated from the annual intake rates of uranium which are calculated from the mean and maximum 238 U, 234 U and 235 U concentrations (mBq l−1 ) in waters Radionuclide
Intake rate (Bq yr−1 )
Concentration (mBq l−1 ) Mean
Maximum
Mean
Maximum
235 U
3.80 ± 5.32 4.77 ± 7.09 0.20 ± 0.25
20.0 26.5 0.98
2.77 3.48 0.15
14.6 19.3 0.72
Total
8.77
47.5
6.40
34.6
238 U 234 U
Dose coefficient factor (µSv Bq−1 )
The effective doses for the adult population (µSv yr−1 ) Mean
Maximum
0.045 0.049 0.047
0.12 0.17 0.01
0.66 0.95 0.03
–
0.30
1.64
Table 8 The annual effective doses for the adult population in Serbia and Montenegro estimated from the annual intake rates of uranium which are calculated from the mean and maximum 238 U, 234 U and 235 U concentrations (mBq l−1 ) in waters Radionuclide
Concentration (mBq l−1 ) Mean
238 U 234 U 235 U
Total
5.16 ± 7.32 6.67 ± 10.3 0.26 ± 0.34 12.1
Intake rate (Bq yr−1 ) Maximum
Mean
Maximum
21.9 28.1 0.88
3.77 4.87 0.19
16.0 20.5 0.64
50.9
8.83
37.1
Dose coefficient factor (µSv Bq−1 )
The effective doses for the adult population (µSv yr−1 ) Mean
Maximum
0.045 0.049 0.047
0.17 0.24 0.01
0.72 1.01 0.03
–
0.42
1.76
Uranium isotopes in drinking waters in Italy and the Balkan regions
233
Table 9 The annual effective doses for the adult population in Bosnia and Herzegovina estimated from the annual intake rates of uranium which are calculated from the mean and maximum 238 U, 234 U and 235 U concentrations (mBq l−1 ) in waters Radionuclide
Concentration (mBq l−1 )
Intake rate (Bq yr−1 )
Mean
Maximum
Mean
235 U
2.80 ± 3.11 3.19 ± 2.01 0.17 ± 0.13
12.2 7.09 0.41
2.04 2.33 0.124
Total
6.16
19.7
4.49
238 U 234 U
Dose coefficient factor (µSv Bq−1 )
Maximum 8.91 5.18 0.299 14.4
The effective doses for the adult population (µSv yr−1 ) Mean
Maximum
0.045 0.049 0.047
0.09 0.11 0.01
0.40 0.25 0.01
–
0.21
0.66
the average annual committed effective doses due to the water intake were 1.62 µSv yr−1 in Italy, 0.30 µSv yr−1 in Kosovo, 0.42 µSv yr−1 in Serbia and Montenegro, and 0.21 µSv yr−1 in Bosnia and Herzegovina. The corresponding values of maximum annual effective dose were 8.33 µSv yr−1 , 1.64 µSv yr−1 , 1.76 µSv yr−1 and 0.66 µSv yr−1 , respectively. The ICRP-recommended subsidiary dose limit for critical groups and principal dose limit for public are 5 mSv yr−1 and 1 mSv yr−1 respectively (ICRP, 1977). However, the planning dose limit for a given source for public is only 0.1 mSv yr−1 (ICRP, 1991). EU Council Directive 98/83/EC on the quality of water intended for human consumption requires the Member States to monitor the concentrations of radionuclides in public drinking water (EU, 1998). If the indicative dose exceeds 0.1 mSv yr−1 , the competent authorities should identify the cause and take justified precautions. From Tables 6–9 it is seen that the exposure levels of uranium from water ingestion in the population groups of Italy and the Balkan regions are well below 0.1 mSv yr−1 .
4. Conclusions The mean uranium activity concentrations in drinking waters in Italy are 20.6 ± 29.0 mBq l−1 for 238 U, 25.3 ± 35.6 mBq l−1 for 234 U and 1.00 ± 1.27 mBq l−1 for 235 U. The mean activity ratios are 1.33 ± 0.33 for 234 U/238 U and 0.047 ± 0.038 for 235 U/238 U. However, the activity concentrations of uranium isotopes in waters of the Balkan regions are much lower than that in Italy. DU do appear in some water samples collected in the Balkan regions, but the low DU concentrations do not constitute from a radiological point of view a health risk at the present time. The estimated average annual committed effective doses to the public due to the water intake were 1.62 µSv yr−1 in Italy, 0.30 µSv yr−1 in Kosovo, 0.42 µSv yr−1 in Serbia and Montenegro, and 0.21 µSv yr−1 in Bosnia and Herzegovina. The activity disequilibria between 234 U and 238 U were observed in most of the samples and the 234 U/238 U value in some sample can be as high as 1.83, therefore, it can lead to underestimate the actual exposure dose if the activity equilibrium of uranium is assumed. Radiation exposure from naturally occurring radionuclides (or medical procedures) is not subjected to the ICRP system of dose limitation, although they could be a major source of radiation exposure to man. The presented data provide basic information about the uranium
234
G. Jia et al.
radiation levels, and can also serve as reference values for a given area for doses from artificial radionuclides released to the environment as a result of human activity. However, for the overall external and internal dose evaluation of natural radioactivity in drinking water, all other important natural radionuclides, including the uranium series, such as 230 Th, 226 Ra, 210 Pb and 210 Po, the thorium series, such as 232 Th and 228 Th, and 40 K should also be determined.
References EU (1998). Council Directive 98/83/EC of 3 November 1998 on the quality of water intended for human consumption. Official Journal of the European Communities L 330/332 (December 1998). Fisenne, I.M., Perry, P.M., Decker, K.M., Keller, H.W. (1987). The daily intake of 234,235,238 U, 228,230,232 Th, and 226,228 Ra by New York City residents. Health Physics 53, 357–363. Goldstein, S.J., Rodriguez, M.J., Lujan, N. (1977). Measurement and application of uranium isotopes for human and environment monitoring. Health Physics 72, 10–18. ICRP (1977). Recommendations of the International Commission on Radiological Protection. ICRP Publication, vol. 26. Pergamon Press, Oxford. ICRP (1991). Recommendations of the International Commission on Radiological Protection. ICRP Publication, vol. 60. Pergamon Press, Oxford. ICRP (1996). Age-Dependent Doses to Members of the Public From Intake of Radionuclides: Part 5. Compilation of Ingestion and Inhalation Dose Coefficient. ICRP Publication, vol. 72. Pergamon Press, Oxford. Jia, G., Belli, M., Sansone, U., Rosamillia, S., Gaudino, S. (2004). Concentration, distribution and characteristics of depleted uranium (DU) in the Kosovo ecosystem: A comparison with the uranium behaviour in the environment uncontaminated by DU. Journal of Radioanalytical and Nuclear Chemistry 260 (3), 484–494. Jia, G., Belli, M., Sansone, U., Rosamillia, S., Gaudino, S. (2005). Concentration and characteristics of depleted uranium (DU) in water, air and biological samples collected in Serbia and Montenegro. Applied Radiation and Isotopes 63 (3), 381–399. Partington, J.R. (1966). General and Inorganic Chemistry for University Students, Fourth edition. Macmillan, London, St. Martin’s Press, New York, 760 pp. Pietrzak-Flis, Z., Rosiak, L., Suplinska, M.M., Chrzanowski, E., Dembinska, S. (2001). Daily intake of 238 U, 234 U, 232 Th, 230 Th, 228 Th and 226 Ra in the adult population of central Poland. The Science of the Total Environment 273, 163–169. United Nations Environment Programme, UNEP (2001). Depleted uranium in Kosovo: Post-conflict environmental assessment. (First published in Switzerland in by UNEP.) United Nations Environment Programme, UNEP (2002). Depleted uranium in Serbia and Montenegro: Post-conflict environmental assessment in the Federal Republic of Yugoslavia. (First published in Switzerland by UNEP.) United Nations Environment Programme, UNEP (2003). Depleted uranium in Bosnia and Herzegovina: Post-conflict environmental assessment. (First published in Switzerland in by UNEP.) United Nations Scientific Committee on the Effects of Atomic Radiation, UNSCEAR (1993). Sources and Effects of Ionising Radiation. Report to the General Assembly. United Nations, New York (with scientific annexes). WHO (1998). Guideline for Drinking-Water Quality, Second edition. Geneva (Switzerland). WHO (2004). Guideline for Drinking-Water Quality, Third edition. Geneva (Switzerland).
235
The radiological evaluation of uranium, radium and radon in metallic and thermo-metallic springs in Ikaria Island, the eastern Aegean Sea, Greece H. Floroua,* , K. Kehagiab , A. Savidoua , G. Trabidoua a NCSR “Demokritos”, INT-RP, Aghia Paraskevi 153 10, Athens, Greece b GAEC Aghia Paraskevi 153 10, Athens, Greece
Abstract The natural alpha and gamma-emitters 238 U, 226 Ra and 222 Rn were analysed in geothermal springs on Ikaria Island situated in the eastern Aegean Sea. The obtained results were evaluated in relation to the radiological regime of the wider area. The exposure dose rates in a network surrounding the island, and the committed effective doses to the human population due to potable spring water are also reported. Although, the measured concentrations of 238 U, 226 Ra and 222 Rn in spring water fall in the range of values reported in the literature, in general, they are higher if compared to other Greek regions. Elevated concentrations of 238 U, 226 Ra, 222 Rn were found occasionally, whereas evaporation of spring water during treatment procedures in spa installations have resulted in 238 U enrichment. Higher concentrations of 238 U and 226 Ra were observed in the hot springs. As the surrounding materials present elevated concentrations of the considered radionuclides, the metallic and thermo-metallic springs influence the ambient areas, acting as radionuclide carriers. In terms of the radiological risk, humans might suffer a limited influence from internal alpha-radiation due to systematic water consumption, whereas the concentration of uranium in one spring in terms of ecotoxicity was above the protective boundary level with regard to the predicted no-effect concentration (PNEC) values set for the protection of aquatic biota. Keywords: Radionuclides, Uranium, Radium, Radon, Spring water, Radiation dose, Greece
1. Introduction Some areas of Ikaria Island (37◦ 59 N, 22◦ 58 E – an area of 267 km2 in the eastern Aegean Sea, Greece) have shown in several environment studies elevated natural background radiation due to geological origin (Trabidou et al., 1996). The areas mainly spread in the neighbourhood of several metallic or/and thermo-metallic springs found on the Island. The measured gamma dose rates in the selected areas have a range 1.3–1.8 mGy yr−1 , and in control areas 0.5–0.8 mGy yr−1 , whereas the mean value for the total measured area is 1.2 mGy yr−1 * Corresponding author. Address: NCSR “Demokritos”, Aghia Paraskevi 153 10, Athens, Greece; phone: (+30) 210 6503809; fax: (+30) 210 6503050; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08017-4
© 2006 Elsevier Ltd. All rights reserved.
236
H. Florou et al.
Fig. 1. Sampling sites at Ikaria Island and gamma-radiometry dose rates (after Trabidou et al., 1996; Florou et al., 2004).
(Trabidou, 2004). The reported mean value for Greece (terrestrial and coastal sediment surface) is 0.7 mGy yr−1 (0.08 µGy h−1 ) (Kritidis and Florou, 1989). The reported radionuclide concentrations in abiotic materials (soil, ores, mud, etc.) are: up to 1050 Bq kg−1 for 238 U, up to 760 Bq kg−1 for 226 Ra, up to 260 Bq kg−1 for 228 Ra, up to 70 Bq kg−1 for 228 Th, up to 300 Bq kg−1 for 232 Th and up to 2700 Bq kg−1 for 40 K, with a wide range of levels (Trabidou et al., 1996; Florou et al., 2004). In the present study, the concentrations of 238 U, 226 Ra and 222 Rn were measured in water samples taken from four metallic springs (Fig. 1) using the alpha-spectrometry (after an appropriate radioanalytical treatment), the gamma-spectrometry and liquid scintillation counting. The findings were evaluated in terms of the radiological quality assessment of the examined areas. The published concentrations of 238 U and 226 Ra in soil (Trabidou, 2004) sampled from the investigated areas were also included in the evaluation. Radiological assessment for humans and ecotoxicity for non-humans was also carried out.
2. Experimental 2.1. Sampling The samples of spring water were collected during three expeditions, one per year from 2001 to 2003. As the study on the issue in Environmental Radioactivity Laboratory in Greece has started since 1996, the previous experience and the field parameters (i.e. temperature, pH, etc.) were used to ensure appropriate sampling. The sampling sites are shown in Fig. 1.
The radiological evaluation of 238 U, 226 Ra and 222 Rn in springs in Ikaria Island
237
2.2. Alpha-spectrometry One litre water samples were acidified with nitric acid to pH = 1 and then evaporated to dryness after adding 232 U as an isotopic tracer. Uranium was separated from 8 M HCl by anion exchange resin (Bio-Rad AG1-X4) column and eluted with 0.1 M HCl. The dried eluate was dissolved in 1 M ammonium sulfate, adjusted to pH = 2.5. The sources for alphaspectrometric measurements were prepared by electrodeposition of uranium onto stainless steel discs. The chemical yield of the method is 85–99% (Eaton et al., 1995). In samples with high salinity uranium was separated by cation exchange (Chelex-100) resin column and eluted with 2 M HNO3 , following by extraction with 30% tributyl phosphate and dodecan (Pashalidis and Tsertos, 2004). The chemical yield of the method is about 75%. The uranium sources were measured in alpha-spectrometers (Canberra). The uncertainties were calculated using the Curie method (95% confidence level). 2.3. Liquid scintillation counting Water samples, after aeration for 2 hours, were put into 40 ml glass vials. The vials were filled to the brim and sealed with Teflon-rubber discs that prevent leakage of radon from water. They were kept at least for one month to allow the equilibrium between 226 Ra and 222 Rn to be established. Then a 20 mL syringe was filled with 10 mL of the appropriate eluting cocktail (OPTI-FLUOR, blend of long chain alkylbenzenes with PPO and bis-MSB scintillators), and 10 mL of water from the bottom of the sampling vial (Zouridakis et al., 2002). The mixture was then transferred to a liquid scintillation vial. After 3 hours the equilibrium between radon and its daughters has been established and the counting was carried out in the LSC system. The measuring procedure has been tested by analysing 6 samples with known 226 Ra concentrations (15 Bq L−1 – 2 samples, 205 Bq L−1 – 4 samples). These samples were prepared by using a standard solution of 226 Ra (the standard reference material 4966, National Institute of Standards and Technology, USA). For the dilution municipal tap water was used drawn several months before. The samples were put into 40 mL glass vials filled to the brim and sealed with Teflon–rubber disc to prevent leakage of radon from water. The vials were kept at least for one month to allow the equilibrium between radium and radon to be established. The results of these measurements were in accordance with the reported concentrations of 226 Ra. Measurements were performed using the Packard Tri-Carb 2560 TRXL Liquid Scintillation Analyser, interfaced to a PC containing the Packard Applications Management System and spectrograph software packages. Samples were counted with a wide window setting of 25–900 keV as in the method described by Kappel et al. (1993). The rest of counting parameters and procedures, including equipment standardisation and normalisation, were set and carried out, as described in Pico-Rad 5.9 Program for radon analysis in water samples. For the determination of radon concentrations, all the samples have been measured by the LSC system for 6 hours. For all samples the counting was repeated at least five times √ to determine the mean value and the standard deviation. The lower limit of detection (2 B ) is 0.4 Bq L−1 . The background (B) was measured using 10 mL of cocktail and 10 mL of water drawn from municipal tap (prepared several months before the measurement, as described in Pico-Rad 5.9 Program). The background rate was 35 counts min−1 .
238
H. Florou et al.
2.4. Gamma-spectrometry Water samples were transferred to one litre Marinelli beakers, shaped in aluminium for radon measurements (Vitantonio de Palma, Milano, Italy). The pH was adjusted to 1 by adding nitric acid. For 222 Rn determinations, the beakers were sealed and stored 3 h prior to measurement to ensure the equilibrium between 222 Rn and its daughters. 222 Rn activities were derived from the analysis of the 295.2 and 352 keV lines of 214 Pb and the 609.4 keV line of 214 Bi, taking into account a correction factor for the decay between the sampling and measurement. It is estimated that the measurements have an uncertainty of 25% due to gas losses during sampling. For 226 Ra determinations, the samples were sealed in the measurement pots after the removal of 222 Rn by aeration of samples. The samples were kept sealed for at least 20 days to ensure the secular equilibrium between 226 Ra and its daughters. The samples were measured in a HPGe detector of 20% relative efficiency and 2 keV resolution. Quality assurance has been ensured by participation in intercomparison exercises organised by WHO (Sample No 60 SH 300, WHO – IRC, SCPRI – France, XI/93). 3. Results and discussion Figure 1 shows the published data obtained by the gamma-radiometry survey (Trabidou et al., 1996; Trabidou, 2004; Florou et al., 2004). It is seen that in the areas with springs outflows elevated levels of natural background radiation were observed, with maxima close to the upper limits of values reported in the literature (Gans, 1985; Bettencourt et al., 1988; Asikainen and Kahlos, 1979). The measured activity concentrations (mean values and standard deviations) of 238 U, 226 Ra and 222 Rn in spring water and soil at Ikaria Island, eastern Aegean Sea are presented in Table 1. Taking into account the wide range of observed radionuclide levels, the concentrations of 238 U and 226 Ra in spring water are characterised, in general, as reasonable, whereas evaporation of spring water during treatment procedures in spa installations have resulted in 238 U enrichment up to 500 times. The elevated concentrations of 226 Ra 5.0 ± 0.4 Bq L−1 (with a maximum value of 7.0 ± 0.5 Bq L−1 observed during summer) have been measured in the thermal spring Ik2. Comparatively, the analysed metallic spring water samples are generally characterised by low 238 U and high 226 Ra activities. The 238 U activity ranges from 1.3×10−3 to 1.09 Bq L−1 . In the hot spring water samples (Ik1a, Ik2 and Ik4) the 238 U activity is lower than in the cold spring water (Ik3). The determined 235 U/238 U activity ratio ranges from 0.045 to 0.052 (mean value 0.048 ± 0.020). These values are in accordance with the isotopic ratio value of the natural uranium (0.048) and ensure natural origin of uranium from lithospheric material. Additionally, the 234 U/238 U activity ratio was found to be in the range from 0.94 to 1.28. These values characterise the natural origin of uranium, as well (Table 1). The occurrence of 234 U/238 U radioactive equilibrium in spring waters is attributed to the fact that these two radionuclides display almost the same mobility during water–rock interaction. The 226 Ra concentrations range from 0.3 to 5.0 Bq L−1 , whereas the 226 Ra/238 U activity ratios are in the range of 1.09–3850. The highest activity ratios are observed in the hot spring water samples Ik1a, Ik2, Ik4 (Table 1). These elevated values for the hot spring waters are in a good accordance with the values reported by Hakam et al. (2001a, 2001b).
Table 1 Activity (mean values and standard deviations) of 238 U, 226 Ra and 222 Rn in spring water and soil at Ikaria Island, eastern Aegean Sea
Ik1a
Ik1b
Spring water
Activity ratio Activity ratio
238 U (mBq L−1 ) 234 U/238 U
5.34 ± 0.94
1092 ± 30.2
235 U/238 U
Soil∗∗
Spring water
Activity ratio Soil∗∗
238 U (Bq kg−1 ) 226 Ra (Bq L−1 ) 226 Ra/238 U
1.07 ± 0.25
0.045 ± 0.030
36 ± 13
1.3 ± 0.2
0.94 ± 0.04
0.049 ± 0.010
−
1.2 ± 0.3
−
1.3 ± 0.1
243
1.09
Spring water
Remarks
226 Ra (Bq kg−1 ) 222 Rn (Bq L−1 )
59 ± 17
2,467 ± 51
Spring outflow (field – 60◦ C)
−
967 ± 23
Fountain inside spa installation
−
142 ± 15
Outflow repository
120 ± 9
Spring outflow (field – 60◦ C)
Ik1c
25.75 ± 2.44
0.97 ± 0.13
0.045 ± 0.020
Ik2
1.30 ± 0.23
1.28 ± 0.25
0.052 ± 0.030 138 ± 16
5.0 ± 0.4
1.00 ± 0.08
0.048 ± 0.010
54 ± 12
0.4 ± 0.3
2.47
83 ± 9
64 ± 46
Spring outflow (field – 20◦ C), potable
1.15 ± 0.09
0.046 ± 0.020 141 ± 12
0.3 ± 0.3
9.47
118 ± 35
15 ± 6
Spring outflow (field – 30◦ C)
Ik3
Ik4
161.9 ± 9.1
30.09 ± 2.05
∗ 3 samples from each spring during three excursions. ∗∗ Trabidou et al. (1996); Trabidou (2004).
50.5 3,850
122 ± 38
The radiological evaluation of 238 U, 226 Ra and 222 Rn in springs in Ikaria Island
Metallic spring∗
239
240
H. Florou et al.
The radionuclide concentration ranges obtained from the present study and from the literature are shown in Table 2. As far as 222 Rn are concerned, the obtained results fall in the reported values. The highest value was observed in the hot spring Ik1a. Unlike in other metallic and thermo-metallic springs in Greece (Kritidis, 1991), the concentrations of 222 Rn are not in equilibrium with those of its parent 226 Ra (Table 1). It seems that 222 Rn is diluted directly from the lithospheric material around the spring sources, which emits it to the water body of the springs. Considering also the reported data of 238 U and 226 Ra in the ambient soil (Table 1), these are higher if compared to those in the literature (Table 2). Therefore, it seems that the geothermal springs are the main parameters influencing the ambient areas of the outflows, carrying on radionuclides from their repositories. Among the four springs, only Ik3 is occasionally used as potable water by the local population. From the radiological point of view, the concentrations of 238 U in Ik3 (Table 1) might result to overdose for humans with regards to the limit of 100 mBq L−1 for alpha-radiation (recommendations of the European Council under consideration). Considering 222 Rn, the levels in Ik3 are lower than the recommended limit of 100 Bq L−1 for radon internal exposure (European Commission, 2001). On the assumption of 0.5 L day−1 of human intake and the conversion factor of 2.8 × 10−7 Sv Bq−1 (IAEA, 1996), the committed effective dose due to 226 Ra is 0.02 ± 0.01 mSv yr−1 , which is much lower than the recommended limit of 0.1 mSv yr−1 (European Commission, 1998). In terms of ecotoxicity, Sheppard et al. (2005) has proposed the PNECs (predicted no-effect concentrations) of 0.005 mg L−1 as protective for most of aquatic invertebrates based on water concentration. From the measured concentrations of spring water in the field outflows, one can note that the average concentration of 238 U in the spring Ik3 (162 ± 9 mBq L−1 which corresponds to 13.05 µg L−1 ) is over twice the protective boundary level.
Table 2 Reported concentrations of 238 U, 226 Ra and 222 Rn in spring and groundwaters Sites
Water type
226 Ra (Bq L−1 )
238 U (mBq L−1 )
222 Rn (Bq L−1 )
Reference
Ikaria, Greece Migdonia, Greece Lesvos, Greece Mt. Etna, Italy Spanish Croatia
Springs Groundwater Thermal spas Groundwater Spa Thermal and mineral water Groundwater Hot springs Cold springs Groundwater
0.1–7
1.3–162 0.37–606
10–2600 8–160 10–304 1.8–52.7
Present study Zouridakis et al., 2002 Vogiannis et al., 2004 Alessandro and Vita, 2003 Duenas et al., 1998 Marovic et al., 1996
Groundwater
0.001–0.94
Groundwater
0.11–0.75
Swedish Morocco Rio de Janeiro, Brazil Fujian Province, Chinese India
0.002–1.37 0.07–4.4 up to 2.5 0.009–3.7 0.002–0.01 0.002–0.49
0.6–8.5 2.19–23.65 0.1–80
<3 0.7–3735
Salih et al., 2000 Hakam et al., 2001a Hakam et al., 2001b Almeida et al., 2004 Zhuo et al., 2001 Choubey et al., 2001
The radiological evaluation of 238 U, 226 Ra and 222 Rn in springs in Ikaria Island
241
4. Conclusions High concentrations of 222 Rn were found in spring water, which are not supported by its parent 226 Ra. The concentrations of 238 U and 226 Ra in spring water are characterised as reasonably high with higher values observed in the hot springs. As the surrounding materials present elevated concentrations of the considered radionuclides, the metallic and thermo-metallic springs influence the ambient areas acting as radionuclide carriers. In terms of the radiological risk, humans might suffer limited influence from internal alpharadiation due to systematic water consumption, whereas the concentration of uranium in one spring in terms of ecotoxicity was above the protective boundary level with regard to the predicted no-effect concentration (PNEC) values set for the protection of aquatic biota.
References Alessandro, W., Vita, F. (2003). Groundwater radon measurements in the Mt. Etna area. Journal of Environmental Radioactivity 65, 187–201. Almeida, R.M.R., Lauria, D.C., Ferreira, A.C., Sracek, O. (2004). Groundwater radon, radium and uranium concentrations in Regiao dos Lagos, Rio de Janeiro State, Brazil. Journal of Environmental Radioactivity 73, 323–334. Asikainen, M., Kahlos, H. (1979). Anomalously high concentrations of uranium, radium and radon in water from drilled wells in the Helsinki region. Geochimica et Cosmochimica Acta 43, 1681–1686. Bettencourt, R.O., Teixeira, M.M.G.R., Faisca, M.C., Vieira, I.A., Ferrador, G.C. (1988). Natural radioactivity in Portuguese mineral waters. Radiation Protection Dosimetry 24 (1/4), 139–142. Choubey, V.M., Bartarya, S.K., Saini, N.K., Ramola, R.C. (2001). Impact of geohydrology and neotectonic activity on radon concentration in groundwater of intermontane. Doon Valley, Outer Himalaya, India. Environmental Geology 40 (3), 257–266. Duenas, C., Fernandez, M.C., Enriquez, C., Carretero, J., Liger, E. (1998). Natural radioactivity levels in Andalusian spas. Water Research 32 (8), 2271–2278. Eaton, A.D., Clesceri, L.S., Greenberg, A.E. (1995). Standard Methods for the Examination of Water and Wastewater. American Public Health Association, Washington, DC 20005, 19th ed. ISBN 0-87553-223-3. European Commission (1998). On the quality of water intended for human consumption. Council Directive 98/83/EC/3.11.1998. European Council, Brussels. European Commission (2001). On the protection of the public against exposure to radon in drinking water supplies. Commission Recommendations of 20 December 2001/928/Euratom. European Council, Brussels. Florou, H., Tsytsugina, V., Polikarpov, G.G., Trabidou, G., Gorbenko, V.V., Chaloulou, Ch. (2004). Field observations of the effects of protracted low levels of ionizing radiation on natural aquatic population by using a cytogenetic tool. Journal of Environmental Radioactivity 75, 267–283. Gans, I. (1985). Natural radionuclides in mineral waters. The Science of the Total Environmental 45, 93–99. Hakam, O.K., Choukri, A., Moutia, Z., Chouak, A., Cherkaoui, R., Reyss, J.-L., Lferde, M. (2001a). Uranium and radium in groundwater and surface water samples in Morocco. Radiation Physics and Chemistry 61, 653–654. Hakam, O.K., Choukri, A., Reyss, J.L., Lferde, M. (2001b). Determination and comparison of uranium and radium isotopes activities and activity ratios in samples from some natural water sources in Morocco. Journal of Environmental Radioactivity 57, 175–189. IAEA (1996). International Basic Standards for protection against ionizing radiation and for the safety of radiation sources. SS No 115. IAEA, Vienna, 353 pp. Kappel, R.J.A., Keller, G., Kreienbrock, L., Nickels, R.M. (1993). An epidemiological study using passive radon measurements by liquid scintillation counting. In: Noakes, J.E., Schoenhofer, F., Polach, H.A. (Eds.), Liquid Scintillation Spectrometry, pp. 319–326. Kritidis, P. (1991). A radiological study of the Greek radon spa. In: Proc. Int. Symposium on Radon and Radon Reduction. Philadelphia (USA), April 2–5, 1991.
242
H. Florou et al.
Kritidis, P., Florou, H. (1989). Natural radioactivity in the environment and radioactive pollution. In: Proc. Nat. Conf. on Environmental Science and Technology, vol. B. Aegean University, Mytilini, September 1989, pp. 24–34. Marovic, G., Sencar, J., Franic, Z., Lokobauer, N. (1996). Radium-226 in thermal and mineral springs of Croatia and associated health risks. Journal of Environmental Radioactivity 33 (3), 309–317. Pashalidis, I., Tsertos, H. (2004). Radiometric determination of uranium in natural waters after enrichment and separation by cation-exchange and liquid–liquid extraction. Journal of Radioanalytical and Nuclear Chemistry 260 (3), 439–442. Salih, I., Pettersson, H., Lund, E. (2000). Determination of 222 Rn and 226 Ra in water using a large volume ionization chamber. Journal of Environmental Radioactivity 48, 235–245. Sheppard, S.C., Sheppard, M.I., Gallerand, M.O., Sanipelli, B. (2005). Derivation of ecotoxicity threshold for uranium. Journal of Environmental Radioactivity 79 (1), 55–83. Trabidou, G. (2004). Radiological study in the areas of radioactive springs (Ikaria island). PhD thesis. University of Athens and NCSR “Demokritos”, Athens, 350 pp. Trabidou, G., Florou, H., Angelopoulos, A., Sakeliou, L. (1996). Environmental study of the radioactivity of the spas in the island of Ikaria. Radiation Protection Dosimetry 63 (1), 63–67. Vogiannis, E., Nikolopoulos, D., Louizi, A., Halvadakis, C.P. (2004). Radon variations during treatment in thermal spas of Lesvos Island (Greece). Journal of Environmental Radioactivity 75, 159–170. Zhuo, W., Iida, T., Yang, X. (2001). Occurrence of 222 Rn, 226 Ra, 228 Ra and U in groundwater in Fuijian Province, China. Journal of Environmental Radioactivity 53, 111–120. Zouridakis, N., Ochsenkühn, K.M., Savidou, A. (2002). Determination of uranium and radon in potable water samples. Journal of Environmental Radioactivity 61, 225–232.
Further reading Kobal, I., Kristan, J., Ancik, M., Jerancic, S., Skofljanec, M. (1979). Radioactivity of thermal and mineral springs in Slovenia. Health Physics 37, 239–242. Kobal, I., Vaupotic, J., Mitic, D., Kristan, J., Ancik, M., Jerancic, S., Skofljanec, M. (1990). Natural radioactivity of fresh waters in Slovenia, Yugoslavia. Environment International 16, 141–154. Popit, A., Vaupotic, J., Kukar, N. (2004). Systematic radium survey in spring waters of Slovenia. Journal of Environmental Radioactivity 76, 337–347. UNSCEAR (1988). Sources, effects and risks of ionizing radiation. United Nation Scientific Committee on the effects of Atomic Radiation. 1988 Report to the General Assembly. United Nations, New York, 647 pp. (with Annexes). Vaupotic, J., Kobal, I. (2001). Radon exposure in Slovenia spas. Radiation Protection Dosimetry 97 (3), 265–270.
5. Isotope biomonitors
This page intentionally left blank
245
Bioaccumulation of radiocaesium in Arctic seals from Northeast Greenland JoLynn Carrolla,* , Kristina Rissanenb , Tore Haugc a Akvaplan-niva, Polar Environmental Centre, Tromsø, Norway b URadiation and Nuclear Safety Authority, Helsinki, Finland c Institute of Marine Research, Tromsø, Norway
Abstract Seals are high trophic level feeders that bioaccumulate many contaminants to a greater degree than most lower trophic level organisms. Their trophic status in the marine food web and wide-spread distribution make seals useful sentinels of arctic environmental change. In 1999 and 2000 seals were captured from the northeast coast of Greenland (75– 80◦ N) in order to document the levels and bioaccumulation potential of radiocaesium in the high latitude seal species: harp, ringed, and hooded seals. The results are compared with previous studies in order to assess geographic differences in bioaccumulation for Arctic seals. Concentrations of 137 Cs were determined in muscle, liver and kidney samples from a total of 25 juvenile and 3 adult seals. The mean concentration in muscle and liver samples for all animals was 0.36 ± 0.14 Bq/kg f.w. and 0.26 ± 0.08 Bq/kg f.w. The results are consistent with previous studies indicating low levels of radiocaesium in Arctic seals in response to a long term trend of decreasing levels of 137 Cs in the Greenland Sea region. Comparing levels in muscle tissue among the different seal species, the 137 Cs activity concentration of harp (0.36 ± 0.14 Bq/kg f.w.) and hooded seals (0.37 ± 0.14 Bq/kg f.w.) are similar within uncertainty limits while average concentrations in ringed seals are slightly lower (0.2 Bq/kg f.w.). Bioconcentration factors (BCFs) for the seals (30–110) correspond well with the variety of prey consumed by seals in this region indicating that diet selection is a dominant factor controlling bioaccumulation of radiocaesium in these Arctic seals. Keywords: Bioaccumulation, Radiocaesium, Seals, Greenland
1. Introduction Seals are among the most quantitatively important mammals in the upper part of the marine food chain, along with whales, walrus and polar bear (WWF, 2004). Due to the importance of seals in subsistence diets of Arctic indigenous peoples, they are a major focus of several on-going monitoring and protection efforts (AMAP, 2002). The results of this investigation provide critical information on concentrations of the radionuclide 137 Cs (radiocaesium) which is among the more pervasive elements found in the Arctic. * Corresponding author. Address: Akvaplan-niva AS, Polar Environmental Centre, Tromsø, phone: (+47) 7775 0314 +; fax: (+47) 7775 0301 +; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08018-6
© 2006 Elsevier Ltd. All rights reserved.
246
J. Carroll et al.
The main sources of radiocaesium to the Arctic are discharges from Europe’s nuclear reprocessing facilities. It is estimated that these facilities have contributed 10–15 PBq to the Arctic Ocean as compared to global fallout and Chernobyl which have contributed 4.1 PBq and 1–5 PBq, respectively (Aarkrog, 1994; Strand et al., 1997). Discharges from Sellefield (UK) and La Hague (France), the two largest reprocessing facilities, enter the Irish Sea and English Channel, respectively. After mixing with Chernobyl-laden water from the Baltic Sea in the North Sea, the plume of radiocaesium is transferred further northward by the Norwegian Coastal Current. Once in the Norwegian Sea, the caesium-laden current bifurcates with one branch leading into the Barents Sea (Barents Sea Overflow Current) and the other branch (West Spitzbergen Current) traversing around the west coast of Svalbard (Kershaw and Baxter, 1995; Kershaw et al., 1997; Heldal et al., 2003) (Fig. 1). Part of the northward flowing West Spitzbergen Current detaches and is re-circulated within the Greenland Sea. The transport time for radiocaesium from the North Sea into the Greenland Sea has been estimated to be on the order of 5 years (Dahlgaard, 1995). This caesium enriched water further mixes with lower concentration Arctic Ocean water entering the North Atlantic via the Fram Strait. A number of regional sources of radionuclides exist in the adjacent Norwegian, Barents, and Kara Seas. These secondary sources of radioactive contamination to northern European waters include fluvial inputs from nuclear facilities on the Ob and Yenisey Rivers, point source inputs from radioactive waste dump sites and nuclear test sites on the Russian coastline. Included as well are the nuclear operations related to the military-industrial complex in the Kola Peninsula. Several nuclear accidents have occurred in the region (AMAP, 1998; Matishov et al., 2002) and on-going activities in the region have been identified as possibili-
Fig. 1. Generalized pathway of radiocaesium transport from Sellefield (UK) and La Hague (France) into the Arctic Ocean. In Greenland waters, radiocaesium from Sellefield is returning from the Arctic Ocean. Seal bioconcentration factors (BCF) among different sub-regions of the European Arctic are also shown.
Bioaccumulation of radiocaesium in Arctic seals from Northeast Greenland
247
ties of accidents in the future (Strand et al., 1997). Of greatest concern are the Kola nuclear power plant reactors, nuclear submarines or icebreakers, waste disposal, nuclear weapons or military activities in Murmansk and Archangelsk regions, Novaya Zemlya installations, and military-industrial sites located within catchment areas of the rivers Ob and Yenisey. Levels of radioceasium around northeast Greenland began declining in the mid-1970s in response to a decline in discharges from Sellefield four years earlier (Matishov and Matishov, 2004). During the early to mid-1970s surface seawater concentrations in the Greenland Sea above 65◦ N were on the order of 9 Bq/m3 . In the early to mid-1990s, concentrations had decreased to approximately 7 Bq/m3 (Aarkrog et al., 2000). Among the wide variety of radionuclides present in the Arctic environment, radiocaesium is the only radionuclide that has been shown to biomagnify through marine food webs (Calmet et al., 1992; Kasamatsu and Ishikawa, 1997; Watson et al., 1999; Heldal et al., 2003; Zhao et al., 2001). And it has been hypothesized that features unique to Arctic ecosystems lead to increased vulnerability for organisms from contaminant releases (Fisher et al., 1999; Carroll and Carroll, 2003). These features include: the seasonal and spatial focus of primary productivity, strong benthic-pelagic coupling, a prevalence of large mammals as apex predators, and relatively large amplitude fluctuations in the lipid cycle of some species. Few studies have been conducted on Arctic marine mammals and in particular for seals (Aarkrog et al., 2000; Rissanen et al., 1997; Carroll et al., 2002). Better knowledge is needed of the bioaccumulation potential of apex predators for a variety of species, habitats, and environmental conditions. The present study provides an opportunity to examine the levels and bioaccumulation potential for radiocaesium in harp, ringed, and hooded seals from NE Greenland. The data are used in conjunction with previous work to compare differences among Arctic seal populations from several locations. This knowledge will lead to more effective use of seals and other marine mammals as indicators of environmental change in Arctic marine monitoring programmes.
2. Methods 2.1. Field sampling A research vessel was used to search for harp, hooded, and bearded seals along the drift ice edge areas of the east coast of Greenland in Fall 1999 and Summer 2000 (Table 1). Animals were weighed to the nearest 1/2 kg and sex and sexual maturity (juveniles and adults) were determined. Sub-samples of seal muscle, liver, and kidney were removed and frozen at −20◦ C until analyses were performed. Ages of seals were estimated by counting of growth layers on teeth extracted from individuals as described in Haug et al. (2004). Samples were analyzed at the Radiation and Nuclear Safety Authority in Finland. Muscle, liver and kidney samples from individual seals were cleaned, cut into small pieces, dried at 105◦ C and homogenized. Each sample was placed in a Marinelli beaker or in 100 or 35 ml plastic jars. Samples were counted on a high-purity, low background, gamma spectrometer for time periods ranging from 900–6800 minutes. The spectrometer was calibrated using matrixmatched standards in a similar geometry.
248
J. Carroll et al.
Table 1 Radiocaesium concentrations (Bq/kg f.w.) in seal muscle and liver tissue. Ringed, hooded, and harp seals were captured in October 1999 from Northeast Greenland Species
Weight (kg)
Age (yrs)
Sex
Muscle (Bq/kg f.w.)
Liver (Bq/kg f.w.)
Ringed Ringed Ringed Harp Harp Harp Harp Hooded Hooded Hooded Hooded Hooded Hooded Hooded Hooded Hooded Hooded Hooded Hooded Hooded Hooded Hooded
21 58 47 62 58 54 32 36 43 205 36 39 43 56 35 45 32 27 38 41 34 66
<1 10 >20 1 2 1 <1 <1 <1 10 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 1
M M M M M M M F F M M M M M M F F M M M F M
<0.2 0.2 ± 0.05 <0.2 0.39 ± 0.04 0.29 ± 0.06 0.30 ± 0.06 0.44 ± 0.05 0.25 ± 0.05 0.36 ± 0.09 0.44 ± 0.08 0.26 ± 0.07 0.45 ± 0.07 0.32 ± 0.05 0.72 ± 0.13 0.18 ± 0.04 <0.4 <0.2 <0.5 <0.3 0.37 ± 0.05 0.24 ± 0.04 0.40 ± 0.07
0.18 ± 0.03 0.32 ± 0.07 – 0.21 ± 0.05 <0.30 <0.13 0.36 ± 0.05 <0.17 0.32 ± 0.07 0.20 ± 0.04 0.16 ± 0.03 0.36 ± 0.06 0.24 ± 0.05 – <0.10 <0.20 – – – 0.20 ± 0.04 <0.24 –
Mean ± stdev Ringed
0.20 n=3
Mean ± stdev
0.36 ± 0.07
Harp Mean ± stdev
n=4 0.37 ± 0.15
Hooded
n = 15
Mean ± stdev All
0.36 ± 0.13 n = 22
0.25 ± 0.07 n=2
0.25 ± 0.08 n = 10 0.26 ± 0.08 n = 16
3. Results 3.1. Radiocaesium levels Samples of muscle, liver, and kidney taken from seals collected in 1999 and 2000 were analyzed as described in the previous section and are summarized in Table 1 and Table 2 respectively. The 137 Cs activity concentration of seal muscle for all animals (1999 and 2000 combined) was relatively low (mean = 0.37 ± 0.14 Bq/kg f.w.; n = 22). 137 Cs concentrations in liver were slightly lower than in muscle (mean = 0.26 ± 0.08 Bq/kg f.w.; n = 14).
Bioaccumulation of radiocaesium in Arctic seals from Northeast Greenland
249
Table 2 Radiocaesium concentrations (Bq/kg f.w.) in muscle, liver and kidney tissue from hooded seals collected from Northeast Greenland in July 2000 Species
Weight (kg)
Age (yrs)
Sex
Muscle (Bq/kg f.w.)
Liver (Bq/kg f.w.)
Kidney (Bq/kg f.w.)
Hooded Hooded Hooded Hooded Hooded Hooded
75 81 70 88 90 139
2 2 4 5 8 16
M M M F F M
0.44 ± 0.05 0.22 ± 0.04 0.23 ± 0.04 0.60 ± 0.06 0.34 ± 0.05 0.59 ± 0.06
0.34 ± 0.04 <0.1 <0.1 0.34 ± 0.05 0.17 ± 0.03 0.26 ± 0.04
0.20 ± 0.04 0.11 ± 0.04 0.16 ± 0.04 0.23 ± 0.04 0.35 ± 0.06 0.23 ± 0.04
0.40 ± 0.17
0.28 ± 0.08
0.35 ± 0.06
Mean ± stdev
Kidney samples were taken only during the 2000 expedition, and contained 137 Cs in concentrations similar to liver (mean = 0.21 ± 0.08 Bq/kg f.w.; n = 6). Comparing levels in muscle tissue among the different seal species, the 137 Cs activity concentration of harp and hooded seals are similar within uncertainty limits while ringed seals contain significantly lower radiocaesium concentrations despite similarities in animal sizes among the three species examined (Table 1). However there were no corresponding differences in 137 Cs activity concentration in liver samples among the three species examined during the investigation. 3.2. Radiocaesium bioconcentration factors Bioconcentration factors (BCFs) normalize radiocaesium tissue levels for spatial and temporal differences in environmental concentrations. The bioconcentration factor (BCF) is defined as BCF = CORG (Bq/kg fresh weight)/CSW (Bq/kg), CORG = concentration in the organism, CSW = concentration in seawater. The BCFs for NE Greenland seals were calculated taking the radiocaesium concentration measurements in individual muscle tissue samples and dividing these by a seawater concentration of 6.6 Bq/m3 . The resulting range of BCF values is from 30–110 (Table 3). This seawater concentration was estimated by extrapolating the long-term linear trend of decreasing seawater concentrations as measured in the waters surrounding NE Greenland and described by Aarkrog et al. (2000). Aarkrog et al. (2000) report that the mean seawater concentration during 1990–1997 in NE Greenland between 70–65◦ N, 75–70◦ N and 80–75◦ N were 7.1 ± 2.2, 7.1 ± 1.8 and 7.1 ± 1.3, respectively. There is little change to our determined BCF values as a result of the extrapolated decrease in seawater concentrations from the early to late 1990s. Certainly these minor differences do not alter the overall conclusions of this investigation.
250
J. Carroll et al.
Table 3 Bioconcentration factors (BCFs) in muscle tissue of marine mammals collected from different regions of the Northern European seas Location
Marine mammal
Time period
BCF value
Source
NE Greenland NE Svalbard White Sea/Kola Peninsula Greenland Barents/Norwegian Seas Svalbard Barents Sea
Seals Seals Seals Marine mammals Harbour porpoise Minke whales Minke whales
1999–2000 2000 1994–1995 1990–1997 1998–2000 1998 1998
30–110 34–130 60 ± 20 114 165 ± 5 70–150∗ 130–285∗
Present study Carroll et al., 2002 Rissanen et al., 1997 Aarkrog et al., 2000 Heldal et al., 2003 Born et al., 2002 Born et al., 2002
∗ Calculation made assuming a seawater concentration range of 2.0–4.4 Bq/kg for 137 Cs (Heldal et al., 2003).
4. Discussion 4.1. Radiocaesium levels in seals Seals living at the ice edge in the Greenland Sea are continuously exposed to radiocaesium. Conditions of chronic low level radioactive contamination, such as are present in the Greenland Sea today, lead to food chain transfer as the dominant pathway of exposure for marine organisms (Fowler, 1982; Rowan and Rasmussen, 1994; Zhao et al., 2001). Approximately 70% of the body burden of radiocaesium accumulates in muscle tissue with the additional 30% taken up by other parts of the organism (Anderson et al., 1990). Studies of dissected seal pups from Northwest Russia indicate that radiocaesium concentrations are highest in soft tissues (e.g. parathyroid gland, pancreas, cartilaginous tissues, ovaries) and lowest in bones and fat (Rissanen et al., 1999). Since radiocaesium levels in the seas surrounding the Arctic Ocean are currently very low, seals from NE Greenland would be expected to contain correspondingly low concentrations in their tissues and organs. Concentrations are highest in NE Greenland as compared to other sections of Greenland. This is a consequence of the influence of the recirculation currents originating from the West Spitzbergen Current which contain relatively high concentrations of radiocaesium derived mainly from the European reprocessing facilities (Aarkrog et al., 2000). The overall low mean radiocaesium concentration of 0.37 ± 0.14 Bq/kg in seal muscle, confirms the expectation that regional sources in addition to the main sources (reprocessing facilities, Chernobyl, global fallout) have thus far not had a significant impact on the marine ecosystem. Radiocaesium levels have not changed since 1990–1994 (0.40 ± 0.12 Bq/kg) and are in accordance with the long-term trend of slowly diminishing radiocaesium concentrations in seawater since 1970 (Aarkrog et al., 2000; Aarkrog, 1997a, 1997b). Concentrations in NE Greenland seals are slightly higher as compared to concentrations in seal pups taken from NE Svalbard during the same year (Spring 1999). In NE Svalbard the mean concentration for all animals was 0.23 ± 0.04 Bq/kg f.w. (n = 11) (Carroll et al., 2002). These differences mirror the higher radiocaesium concentrations detected in seawater from NE Greenland (∼6.6 Bq/m3 ) as compared to NE Svalbard (2.0–4.7 Bq/m3 ) and are
Bioaccumulation of radiocaesium in Arctic seals from Northeast Greenland
251
a consequence of the greater influence of Europe’s nuclear reprocessing facilities on the waters surrounding Greenland as compared to Svalbard. 4.2. Radiocaesium bioconcentration in seals The bioaccumulation of radiocaesium in seals represents the outcome of many factors involving a combination of physiological factors, such as absorption efficiency, and ecological factors including diet selection, feeding rates, and prey availability. Migration patterns will also determine the amount of radiocaesium accumulated from food and subsequently retained in the body due to variations in seawater concentrations along the main transport pathway to the Arctic from the main source areas (Fig. 1). Comparing different regions of the Arctic, seals have the lowest BCF values, followed by porpoises, while whales generally have the highest BCF values (Table 3). The BCFs reported for seals from NE Greenland and NE Svalbard further agree well with the value reported previously by Aarkrog et al. (2000) for marine mammals from all of Greenland’s coastal waters (BCF = 114) indicating that these various data sets provide a consistent picture of radiocaesium biouptake in Arctic seals. Among the wide variety of radionuclides present in the Arctic environment, radiocaesium can biomagnify through marine food chains (Anderson et al., 1990; Kasamatsu and Ishikawa, 1997; Fisk et al., 2001; Zhao et al., 2001). Biomagnification of radiocaesium implies that the trophic status of primary prey items is an important factor affecting the radiocaesium activity concentration of seals. As a result, a clear effect of diet selection on radiocaesium bioaccumulation would be expected. BCFs reported for prey items from the Barents and Norwegian Seas (Heldal et al., 2003) and in the waters surrounding Greenland (Aarkrog et al., 2000) for omnivorous fish, benthic organisms, and carnivorous fish are within an approximate range of 10–100. Thus the relatively low BCF values obtained for seals from the European Arctic indicate only a slight enhancement of radiocaesium in relation to prey species (Table 3). Given that life-times of seals (∼30 years) are significantly longer than for prey species (∼5–15 years) we would expect that over time, seals would bioaccumulate increasingly high levels of radiocaesium resulting in correspondingly higher BCF values. While this is apparently not the case for Arctic seals, studies of other marine mammals report a clear increase in bioaccumulation relative to potential lower trophic level prey items, for example, Harbour porpoises (BCF = 165 ± 5) and Minke whales (BCF = 130–285) from the Barents Sea (Table 3). Yet these findings are in contrast to those of Aarkrog et al. (2000) who report no significant differences in radiocaesium contamination between whales and seals taken from Greenland waters. The resulting low BCF values for seals from NE Greenland may be related to the relatively young age of seals in the present study in combination with diet selection factors. The BCF value would be expected to be relatively low if the diet selection of seals consists of organisms relatively low on the trophic pyramid. Analyses of stomach and intestinal contents from hooded and harp seals captured in the pack ice belt of the Greenland Sea in summer (July–August) of 2000 and winter (February–March) of 2001 reveal that the diet of both hooded and harp seals are comprised of relatively few prey taxa. In both harp and hooded seal species, pelagic amphipods of the genus Parathemisto, the squid Gonatus fabricii, polar cod (Boreogadus saida) and capelin (Mallotus villosus) constituted 63%–99% of the observed diet
252
J. Carroll et al.
biomass, irrespective of sampling period. Based on a simple 5 layer trophic pyramid (Heldal et al., 2003), the species determined by Haug et al. (2004) to dominate in the diets of Arctic seals (trophic level 5) are from trophic level 2 (amphipods), trophic level 3 (cephalopods, capelin) and trophic level 4 (polar cod). Differences in the relative dietary mix of these prey species and the clear importance of low trophic level prey in the diets of Arctic seals thus may explain why BCF values reported for Arctic seals are not substantially higher than their prey species. In comparison, BCFs for harbour porpoise from the Barents and Norwegian Seas are reported to be 165 ± 5 (Heldal et al., 2003) suggesting that harbour porpoises feed on a diet consisting of a larger proportion of higher trophic level species. Concerning the importance of diet selection processes, Haug et al. (2004) also report that the relative contribution of prey species to the diet varied both with seal species and sampling period/area, and there were large differences in dietary composition among individuals. In a related study, Christiansen et al. (2004) report that there are large differences among individual seals in their gastric properties (stomach temperature, acidity, pepsin concentration), properties related to the breakdown and assimilation of food, and hence may influence the transfer of radiocaesium from the digestive tract into the body. Differences were particularly pronounced in animals with body weights less than 100 kg which corresponds to the majority of individuals reported on in the present study. Their findings further indicate large differences in the feeding mode and diet composition of individuals: differences that would further be expected to influence biomagnification factors for radiocaesium in Arctic seals. Despite the complex array of sources of variation, and limited knowledge of the physiological and ecological processes controlling radiocaesium bioaccumulation, comparison of our more recent findings with earlier studies confirms that radiocaesium bioaccumulation factors in seals inhabiting different sub-regions of the European Arctic are generally similar. Furthermore, we suggest that the diets of Arctic seals, rich in species from lower trophic levels, lead to relatively low BCF values in these marine mammals. Clearly, to fully appreciate the importance of diet selection as a factor influencing biomagnification, diet composition and radiocaesium activity concentration must be examined simultaneously in a statistically relevant number of individuals. However, our findings would suggest that there is a strong effect of diet selection on radiocaesium bioaccumulation for seals and, to some degree, the levels found in NE Greenland seals reflect the levels found in prey items from their habitat areas, as confirmed in Haug et al. (2004) and Christiansen et al. (2004). Finally, the data provide further verification that the radiocaesium levels remain low throughout the European Arctic in accordance with expectations.
Acknowledgements This research was funded by Akvaplan-niva AS (Norway), the Institute of Marine Research (Norway) and the Radiation and Nuclear Safety Authority (Finland).
References Aarkrog, A. (1994). Radioactivity in polar regions – main sources. Journal of Environmental Radioactivity 25, 21–35.
Bioaccumulation of radiocaesium in Arctic seals from Northeast Greenland
253
Aarkrog, A. (1997a). Doses to the Arctic Population. In: Materials of 3th International Conference on Environmental Radioactivity in the Arctic. Tromsø, pp. 113–115. Aarkrog, A. (1997b). AMAP Greenland 1994–1996. Environmental Project No. 356, pp. 211–239. Aarkrog, A., Dahlgaard, H., Nielsen, S.P. (2000). Environmental radioactive contamination in Greenland: A 35 years retrospect. The Science of the Total Environment 245, 233–248. AMAP (1998). Arctic Pollution Issues. AMAP Assessment Report. Arctic Monitoring and Assessment Programme (AMAP). Oslo, Norway, 859+xii pp. AMAP (2002). Persistent Organic Pollutants in the Arctic. AMAP Assessment 2002. Anderson, S.S., Livens, F.R., Singleton, D.L. (1990). Radionuclides in grey seals. Marine Pollution Bulletin 21, 343–345. Born, E.W., Dahlgaard, H., Riget, F., Dietz, R., Øien, N., Haug, T. (2002). Regional variation of caesium-137 in minke whales Balaenoptera acutorostrata from West Greenland, the Northeast Atlantic and the North Sea. Polar Biology 25, 907–913. Calmet, D., Woodhead, D., Andre, J.M. (1992). 210 Pb, 137 Cs, and 40 K in three species of porpoises caught on the eastern tropical Pacific Ocean. Journal of Environmental Radioactivity 15, 153–169. Carroll, J., Wolkers, H., Andersen, M., Rissanen, K. (2002). Bioaccumulation of radiocaesium in Arctic seals. Marine Pollution Bulletin 44, 1366–1371. Carroll, M.L., Carroll, J. (2003). The Arctic Seas. In: Black, K.D., Shimmield, G.B. (Eds.), Biogeochemistry of Marine Systems. Blackwell, Oxford, pp. 127–147. Christiansen, J.S., Gildberg, A., Nilssen, K.T., Lindblom, C., Haug, T. (2004). The gastric properties of free-ranging harp (Pagophilus groenlandicus (Erxleben, 1777)) and hooded (Cystophora cristata (Erxleben, 1777)) seals. ICES Journal of Marine Science 61, 287–292. Dahlgaard, H. (1995). On Tc-99, Cs-137 and Sr-90 in the Kara Sea. In: Strand, P., Cooke, A. (Eds.), Environmental Radioactivity in the Arctic. Norwegian Radiation Protection Authority, Østerås, pp. 91–95. Fisher, N.S., Fowler, S.W., Boisson, F., Carroll, J., Rissanen, K., Salbu, B., Sazykina, T.G., Sjoeblom, K.-L. (1999). Bioconcentration factors and sediment partition coefficients in Arctic seas subject to contamination from dumped nuclear wastes. Environmental Science and Technology 33, 1979–1982. Fisk, A.T., Hobson, K.A., Norstrom, R.J. (2001). Influence of chemical and biological factors on trophic transfer of persistent organic pollutants in the Northwest polynya marine food web. Environmental Science and Technology 35, 732–738. Fowler, S.W. (1982). Biological transfer and transport processes. In: Kullenberg, G. (Ed.), Pollutant Transfer and Transport in the Sea, vol. II. CRC Press, Inc., Boca Raton, FL, pp. 1–65. Haug, T., Nilssen, K.T., Lindblom, L. (2004). Feeding habitats of harp and hooded seals in drift ice waters along the east coast of Greenland in summer and winter. Polar Research 23 (1), 35–42. Heldal, H.E., Føyn, L., Varskog, P. (2003). Bioaccumulation of 137 Cs in pelagic food webs in the Norwegian and Barents Seas. Journal of Environmental Radioactivity 65, 177–185. Kasamatsu, F., Ishikawa, Y. (1997). Natural variation of radionuclide 137 Cs concentrations in marine organisms with special reference to the effect of food habits and trophic level. Marine Ecology Progress Series 160, 109–120. Kershaw, P.J., Baxter, A. (1995). The transfer of reprocessing wastes from north-west Europe to the Arctic. Deep-Sea Research II 42, 1413–1448. Kershaw, P.J., Gurbutt, P., Woodhead, D., Leonard, K., Rees, J. (1997). Estimates of fluxes of 137 Cs in northern waters from recent measurements. The Science of the Total Environment 202, 211–223. Matishov, D.G., Matishov, G.G. (2004). Radioecology in Northern European Seas. Springer-Verlag, New York, 335 pp. Matishov, G.G., Matishov, D.G., Namjatov, A.E., Smith, J.N., Carroll, J., Dahle, S. (2002). Radioactivity near the sunken submarine ‘Kursk’ in the southern Barents Sea. Environmental Science and Technology 36 (9), 1919– 1922. Rissanen, K., Ikäheimonen, T.K., Matishov, D., Matishov, G.G. (1997). Radioactivity levels in fish, benthic fauna, seals and sea birds collected in the Northwest Arctic of Russia. Radioprotection – Colloques 32 (C2), 323–331. Rissanen, K., Pempkowiak, J., Ikäheimonen, T.K., Matishov, D.G., Matishov, G.G. (1999). 137 Cs, 239,240 Pu, 90 Sr and selected metal concentrations in organs of Greenland seal pups in the White Sea area. In: The 4th International Conference on Environmental Radioactivity in the Arctic. Edinburgh, Scotland, September 20–23. Rowan, D.J., Rasmussen, J.B. (1994). Bioaccumulation of radiocaesium by fish: The influence of physicochemical factors and trophic structure. Canadian Journal of Fisheries and Aquatic Sciences 51, 2388–2410.
254
J. Carroll et al.
Strand, P., Balonov, M., Aarkrog, A., Bewers, M.J., Howard, B., Salo, A., Tsaturov, Y.S. (1997). Arctic Pollution Issues: Radioactive Contamination. Norwegian Radiation Protection Authority, Østerås. Watson, W.S., Sumner, D.J., Baker, J.R., Kennedy, S., Reid, R., Robinson, I. (1999). Radionuclides in seals and porpoises in the coastal waters around the UK. The Science of the Total Environment 234, 1–13. WWF (2004). A Biodiversity Assessment of the Barents Sea Ecoregion. WWF Arktisprogrammet, WWF-Norge og WWF-Russland. Available at http://www.wwf.no. Zhao, X., Wang, W.-X., Yu, K.N., Lam, P.K.S. (2001). Biomagnification of radiocesium in a marine piscivorous fish. Marine Ecology – Progress Series 222, 227–237.
255
Anthropogenic radionuclides in biota samples from the Caspian Sea J. Gastauda,* , B. Oregionia , S.V. Pagavab , M.K. Phama , P.P. Povineca a International Atomic Energy Agency, Marine Environment Laboratory, Monaco b Tbilisi State University, Department of Nuclear Physics, Tbilisi, Georgia
Abstract Samples of fish flesh collected in 1999 in the south-western Caspian Sea in the Baku area, important for caviar production (sturgeon – russkyi osyotr, sevruga and beluga), as well as for consumption (roach and carp) were analysed for anthropogenic strontium, caesium, plutonium and americium, and natural polonium. The highest massic activities of 137 Cs were found in sevruga and beluga flesh (1.2–1.8 Bq/kg wet weight (ww)), while 90 Sr levels were between 5–12 mBq/kg ww, and plutonium and americium levels were close to limits of detection (∼0.2 mBq/kg ww). The observed plutonium and strontium levels are in the same range as in the Mediterranean Sea, whereas caesium has been accumulated in conditions of lower salinity in larger proportions. The 210 Po levels in fish were between 0.2–3 mBq/kg ww, in a fresh caviar (spawn) sample they were higher by a factor of 4 than in sturgeons, but comparable with levels observed in other species. The highest radionuclide levels, by one to two orders of magnitude, were measured in a macroalgae sample. The distribution of radionuclides seems to be more related to the species than to environmental conditions. The estimated concentration factors (CFs) for strontium and plutonium in fish and algae are in a reasonable agreement with IAEA recommended values. Caesium in the same species has been accumulated in larger quantities, so that the resulting CF is higher by a factor of two. The highest CFs were found for macroalgae, documenting that algae are suitable biomonitors of radioactive contamination. The measured activities of radionuclides in biota samples do not represent any radiological risk from their consumption. Keywords: Anthropogenic radionuclides, Natural radionuclides, 90 Sr, 137 Cs, 239,240 Pu, 241 Am, 210 Po, Marine biota, Fish, Sturgeon, Caviar, Algae, Caspian Sea
1. Introduction The Caspian Sea has been recently a subject of many scientific studies, mainly related to sea level changes and contamination by radionuclides and oil products (e.g. Klige and Myagkov, 1992; Kosarev and Yablonskaya, 1994; Vakulovsky and Chumichev, 1998). Two sampling expeditions were organised to the Caspian Sea in 1995 and 1996 by IAEA in collaboration with the Caspian Sea countries. The main aims were to investigate oceanographic conditions related with sea level changes, water dynamics, and the distribution of anthropogenic 90 Sr, * Corresponding author. Address: IAEA-MEL, 4 Quai Antoine 1er, MC-98000, Monaco; phone: (+377) 97977272; fax: (+377) 97977273; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08019-8
© 2006 Elsevier Ltd. All rights reserved.
256
J. Gastaud et al.
137 Cs
and 239,240 Pu in the water column (Froehlich and Povinec, 1998; Froehlich et al., 1999; Oregioni et al., 2003; Povinec et al., 2003). Considering the unique biodiversity of the Caspian Sea, there has also been interest to obtain information on radionuclide concentrations in biota samples, first of all in sturgeons and caviar, as their production is strongly linked with economical regional needs (http://www.caspianenvironment.org). The radioactive contamination of Caspian Sea biota has been investigated by analysing natural 210 Po and anthropogenic 137 Cs, 239,240 Pu and 241 Am in biota samples collected in April 1999 offshore Astrakhan, in the north Caspian Sea (Povinec et al., 2003). A new set of biota samples was collected in the Baku region in the south-western Caspian Sea – in June 1999 at Artom Island, and in November 1999 at Divichi and Neftechala districts; see Fig. 1. Collected biota samples were analysed for natural 210 Po, and anthropogenic 90 Sr, 137 Cs, 239,240 Pu and 241 Am. The obtained results are presented and discussed in this paper.
2. Description of samples and analytical methods 2.1. Biota samples Three different types of endemic sturgeons were sampled and flesh parts were analysed: • Acipenser gueldenstaedtii (common Russian name: russkyi osyotr), widely distributed over the whole area of the sea. It is a sand and silt bottom dweller, feeding with bottom organisms, mainly mollusks at the adult stage. The species lives at depths from 20 to 100 m. The spawning occurs at the end of May–beginning of June in the estuaries of rivers. • Acipenser stellatus (common Russian name: sevruga), spreading throughout the Caspian Sea, this species is more pelagic than the russkyi osyotr, as it lives in deep water zones. The food is mainly zoobenthos, higher crustaceans and fish (gobies), coming from silty to silty/sandy bottoms. The spawning occurs also in May–June in the estuaries of rivers. • Huso huso (common Russian name: beluga). It is a pelagic species covering the entire Caspian Sea. Feeding in shallow zones (1–30 m), but it may be found down to 130–180 m in winter. It is a predator in relation to other fish species like roaches, sprats and gobies. It also uses estuaries of rivers to spawn. A sample of fresh spawn from sturgeon collected at Artom Island was analysed as well. The other analysed fish samples were from the Cyprinidae family. Flesh samples from two species were analysed: • Rutilus frisii kutum (also called Caspian roach). It is also an endemic species which spreads in coastal areas at shell/sand bottoms (water depth 20–30 m). The individuals live at sea and spawn in rivers. The food is mainly mollusks, shrimps, amphipods and crabs at adult stage. • Cyprinus carpio (common name: carp). The species is defined as autochtonous. The fish prefer water bodies with stagnant and slowly flowing waters from 5 to 20 m deep. They burrow in silt for feeding with molluscs and aquatic plants at adult stage. The spawning occurs in rivers, e.g. in the Kura River.
Anthropogenic radionuclides in biota samples from the Caspian Sea
257
Fig. 1. Sampling sites in the Caspian Sea. Biota samples were collected at Artom Island (about 20 km east of the peninsula), and at Divichi and Neftechala coastal waters. Seawater sampling sites (1–13) are also shown (Povinec et al., 2003). Water depth is given in meters.
A sample of green macroalgae of the Cladophora species (Chlorophycea family) was also analysed. These algae are among the richest algae-macrophytes species found in the Azerbaijan sector of the sea. Their development in Baku area is abundant from 0 to 14 m of the water depth. They are also spread all over the world and have often been used for contamination surveys in coastal zones.
258
J. Gastaud et al.
2.2. Analytical procedures After collection, the samples were sliced, weighted, dried and the ratio dry/wet was calculated. Before any chemical treatment, the samples were freeze-dried, ground sieved, homogenised and canned to be counted by nondestructive gamma-ray spectrometry for the determination of 137 Cs. Radiochemical separation and purification procedures were carried out later for the determination of 210 Po, 239, 240 Pu and 241 Am, which were measured by alpha-ray spectrometry, whereas 90 Sr was measured by beta-ray counting using a gas proportional counter. The analytical procedures have already been described in previous publications (La Rosa et al., 2001; Lee et al., 2001), therefore they are not presented in this paper. 3. Results and discussion 3.1. Radionuclide concentrations Massic activities of natural 210 Po and anthropogenic 90 Sr, 137 Cs, 239,240 Pu and 241 Am in sturgeons, fish and algae samples are reported in Table 1. It is clear that the sturgeons tend to concentrate less of 210 Po than other coastal fish. The average massic activity for sturgeons is 0.6 ± 0.4 mBq/kg ww (wet weight), while for the roach and carp samples, the average activity is 2.5 ± 0.4 mBq/kg ww. The 210 Po activity in spawn (2.2 ± 0.4 mBq/kg ww) is higher by a factor of 4 than in sturgeons from the same area, but comparable with other fish species. The observed values are in a reasonable agreement with the compiled data obtained in the framework of the Marine Radioactivity Doses (MARDOS) Project (IAEA, 1995; Aarkrog et al., 1997). Table 1 Anthropogenic radionuclides and 210 Po in biota from the SW Caspian Sea (wet weight). The quoted uncertainties represent 1 sigma standard deviations Sampling area Artom Island
Sample type
Fish (sevruga) Fish (russkyi osyotr) Fish (beluga) Spawn (fresh) Fish (Cyprinus) Fish (Rutilus) Macroalgae (Cladophora) Divichi Fish (sevruga) Fish (russkyi osyotr) Neftechala Fish (Rutilus)
Sampling date
239,240 Pu
26-06-1999 26-06-1999
<0.3 <0.3
<0.15 <0.19
0.40 ± 0.13 0.61 ± 0.16
1.22 ± 0.07 1.14 ± 0.04
26-06-1999 26-06-1999 26-06-1999
<0.3 <0.4 <0.08
<0.06 <0.26 <0.08
0.16 ± 0.01 2.2 ± 0.4 2.30 ± 0.04
5.0 ± 2.2 1.76 ± 0.07 11.0 ± 4.5 0.15 ± 0.08 12.3 ± 2.4 0.22 ± 0.09
26-06-1999 26-06-1999
<0.12 51.0 ± 1.7
<0.14 14.2 ± 0.8
04-11-1999 04-11-1999
0.17 ± 0.07 <0.05 <0.13 0.47 ± 0.12
1.17 ± 0.15 0.53 ± 0.13
6.8 ± 1.7 0.95 ± 0.04 9.3 ± 2.0 0.50 ± 0.02
04-11-1999
<0.19
3.03 ± 0.22
0.48 ± 0.02
(mBq/kg)
241 Am (mBq/kg)
<0.20
210 Po
90 Sr
(mBq/kg)
(mBq/kg)
137 Cs (Bq/kg)
2.31 ± 0.24 9.8 ± 2.0 0.64 ± 0.03 69.5 ± 2.4 390 ± 20 4.35 ± 0.17
Anthropogenic radionuclides in biota samples from the Caspian Sea
259
The most abundant from the analysed anthropogenic radionuclides is 137 Cs. It is more concentrated in sturgeons (the average massic activity is 1.11 ± 0.46 Bq/kg ww) than in coastal fish (average activity is 0.45 ± 0.21 Bq/kg ww). The spawn sample has shown even a lower concentration, 0.15 ± 0.08 Bq/kg ww. The observed ranges of activities are in agreement with the compiled data from the MARDOS Project (IAEA, 1995; Aarkrog et al., 1997). The observed range of 90 Sr in sturgeons is 5–9 mBq/kg ww, slightly higher activities have been observed in coastal fish (10–12 mBq/kg ww) and in spawn (11 ± 4 mBq/kg ww). 239, 240 Pu levels in fish and spawn samples are below detection limits, except for one sevruga sturgeon from Divichi, with massic activity of 0.17 ± 0.07 mBq/kg ww. 241 Am results have also been below detection limits, except for a sample of russkyi osyotr from Divichi (0.47 ± 0.12 mBq/kg ww). It seems that the different fish species accumulate the radionuclides differently. These differences in fish, living in the open sea or in coastal areas cannot be explained by the differences in activities measured in seawater profile samples, as they do not change much with depth in the investigated shallow zones (Povinec et al., 2003). However, as no data for radionuclides in sediments or other kind of biota are available for the Caspian Sea, it is difficult to estimate the impact of the diet and environmental conditions on the accumulation of radionuclides by the fish. For all analysed radionuclides in the Cladophora macroalgae sample the measured activities are by one or two orders of magnitude higher than in the other biota samples. This is especially true in the case of 90 Sr, which shows the highest measured value, 390 ± 20 mBq/kg ww, but also for 210 Po (69 ± 2 mBq/kg ww) and 239,240 Pu (51 ± 2 mBq/kg ww). The results have confirmed previous observations that macroalgae are good biomonitors of radionuclides (IAEA, 1995). 3.2. Concentration factors Concentration factors (CFs) were calculated for plutonium, strontium and caesium (Table 2), as those radioelements were measured in the water column (unfiltered seawater samples) after the sampling cruises in 1995 and 1996 along the whole Caspian Sea basins (Oregioni et al., 2003; Povinec et al., 2003). As the sampling areas for biota are shallow coastal zones around Baku with water depths up to 100 m, the data on radionuclide concentrations in surface seawater at similar latitudes have been chosen for the calculations of CFs. The CFs were calculated as a ratio of the massic activity of a radionuclide in tissue (in Bq/kg ww) per activity in seawater (in Bq/kg). The calculated CFs for the fish species have been following corresponding radionuclide activities in fish samples. CFs for strontium are slightly higher in the coastal fish (the mean value is 1.58 ± 0.25) than in sturgeons (the mean value is 0.92 ± 0.24). Caesium is more concentrated in sturgeons. The sturgeons from the Artom Island collected in June 1999 have higher CFs (262 ± 64) than the two samples collected at Divichi in November 1999 (125 ± 54). The other coastal fish have CFs for 137 Cs 85 ± 40 in average. The CF for plutonium can only be calculated for the sevruga sturgeon from Divichi, situated north of the Baku area. The obtained value is 26 ± 11. The spawn has CFs both for strontium and caesium similar to coastal fish. As already mentioned, the differences in radionuclide concentrations in biota cannot be related only to their concentrations in surface seawater, which have varied negligibly in the
260
J. Gastaud et al.
Table 2 Concentration factors in biota from the SW Caspian Sea, compared with IAEA recommended values (IAEA, 2004) Sampling area
Sample type
Fish (sevruga) Fish (russkyi osyotr) Fish (beluga) Spawn (fresh) Fish (Cyprinus) Fish (Rutilus) Macroalgae Divichi Fish (sevruga) Fish (russkyi osyotr) Neftechala Fish (Rutilus)
239,240 Pu
This paper
90 Sr
IAEA (2004) This paper
137 Cs
IAEA (2004) This paper IAEA (2004)
Artom Island
7900 ± 1000 4×103 26 ± 11 1×102
0.7 ± 0.3 1.6 ± 0.6 1.7 ± 0.4 1.4 ± 0.3 56 ± 5 0.9 ± 0.2 1.2 ± 0.3
3×100 3×100 3×100 3×100 1×101 3×100 3×100
230 ± 20 222 ± 10
1×102 1×102
340 ± 20 29 ± 15 42 ± 17 122 ± 7 830 ± 50 163 ± 9 86 ± 5
1×102 1×102 1×102 1×102 5×101 1×102 1×102
92 ± 5
1×102
investigated areas. The species tend to accumulate radionuclides differently, that could also depend on their physical characteristics (e.g. size, age). This has also been confirmed for the macroalgae sample, for which the observed CFs have been the highest among the samples presented in this paper: 7900 ± 1000 for plutonium, 830 ± 40 for caesium and 56 ± 5 for strontium. As no data have been available for 210 Po in Caspian Sea water, we could not calculate a CF for this element.
4. Comparison of results with published data 4.1. Radionuclide concentrations As there are only very few available data concerning radionuclides in biota from the Caspian Sea, it is difficult to make an appropriate comparison. Timoschuk (1972) reported 90 Sr levels in several samples collected in May 1966 in the Caspian Sea. They were much higher (∼1 Bq/kg ww) than the results presented in this work, even if we take into account that these measurements were done on samples collected 30 years ago (more that one half-life of 90 Sr earlier). Data published for radionuclides in fish and algae from the Mediterranean Sea (Othman et al., 1994) and the Black Sea (Lazorenko et al., 2002, 2004) are comparable with the present data. The caesium is lower in the Mediterranean samples, but strontium levels are higher. This is understandable as analysed sardine fish samples from the Syrian coast had a small size and they were analysed as total flesh and bone samples which conditioned the strontium activities. Polonium results reported for the Black Sea samples of many species of pelagic, nearbottom and benthic (bottom) fish in 1999–2000 were from 0.4 to 5.3 Bq/kg ww (Lazorenko et al., 2002, 2004), comparable with the present results.
Anthropogenic radionuclides in biota samples from the Caspian Sea
261
Other recently published data from the Pacific Ocean (Hong et al., 2002) concern fish species from the open ocean: mackerel, flounder and pollock. In these cases the plutonium levels are comparable with those measured in the Caspian Sea fish samples, but the caesium activities are lower than the activities found in the sturgeon samples. Generally, the presented results are in agreement with data compiled for the MARDOS project (IAEA, 1995), except for higher concentrations of 137 Cs and 90 Sr found in Caspian Sea biota samples. 4.2. Concentration factors The calculated CFs for strontium in fish found in this work (0.7–1.7) are lower by a factor of two when compared with recommended value (3) for fish published recently by IAEA (2004). This is due to the fact the 90 Sr levels in seawater are in excess in the Caspian Sea, which also resulted in a low 137 Cs/90 Sr activity ratio in seawater with a mean value of 0.74 ± 0.15 at all water depths, compared to an expected global fallout ratio of 1.6 (Povinec et al., 2003). The calculated CFs for caesium in fish are from 30 to 230, i.e. when we take into account quoted uncertainties, they oscillate around the IAEA recommended value 100 (IAEA, 2004). The observed CF for plutonium is ∼30, compared with the IAEA recommended value of 100. Larger discrepancies in CFs have been found for the macroalgae sample, i.e. observed 60 vs. recommended 10 for strontium, 50 vs. 800 for caesium, and 8000 vs. 4000 for plutonium, however, we have to keep in mind that only one sample has been analysed. The CFs reported for a Cladophora macroalgae sample from the Syrian coast (Othman et al., 1994) are comparable for plutonium and strontium, but the CF for caesium in the Caspian Sea is higher by a factor of 10. Generally, the CFs in algae are much higher than in fish, that document advantages of algae as biomonitors of radionuclide contamination in the marine environment. 4.3. Activity ratios The activity ratio (AR)bio of a radionuclide x over a radionuclide y in biota can be expressed as (ARx/y )bio =
CFx (ARx/y )w , CFy
where (AR)w represents the activity ratio in seawater, i.e. the AR in biota is controlled by both, the AR in seawater and the corresponding CF. The 137 Cs/90 Sr activity ratios in all the analysed fish samples are in the interval 14–350, the weighted average 70 ± 15 is in agreement with the expected global fallout ratio of 50. This is a surprise as a lower ratio than the global fallout ratio has been expected due to elevated 90 Sr levels in seawater (the observed 137 Cs/90 Sr activity ratio was 0.74 ± 0.15 at all water depths, compared to an expected global fallout ratio of 1.6 (Povinec et al., 2003)). The higher 137 Cs/90 Sr ratio has been influenced by higher 137 Cs activities observed in fish samples and by a lower accumulation of 90 Sr in fish flesh. The 137 Cs/90 Sr activity ratio in the algae sample is 11 ± 1, higher than the expected global value of 7.5, documenting again that Cladophora macroalgae has accumulated more caesium than strontium.
262
J. Gastaud et al.
Measured 238 Pu activity in the algae sample (1.4 ± 0.5 mBq/kg ww) allowed to calculate the 238 Pu/239, 240 Pu activity ratio which is equal to 0.028 ± 0.09, the same as in the water column (Povinec et al., 2003), and represents the global fallout ratio (0.03). The 239, 240 Pu/90 Sr activity ratio for the same algae sample is 0.131 ± 0.008, what is much lower than the global fallout value for macroalgae (18), due to a preferential scavenging of plutonium from surface water. The same assumption is valid for the 239,240 Pu/137 Cs activity ratio of the same algae sample, 0.0117 ± 0.0006, that is again much lower than the global fallout ratio of 2.4. The sevruga sample from Divichi has the 239,240 Pu/90 Sr activity ratio 0.025 ± 0.012, and the 239, 240 Pu/137 Cs activity ratio 0.00018 ± 0.00007, both much lower again than the global fallout ratios of 1.5 and 0.03, respectively. These ratios were strongly influenced by low concentrations of plutonium found in surface water where fish live (Povinec et al., 2003), as well as by elevated levels of 137 Cs found in the Caspian Sea biota. The 241 Am/239, 240 Pu activity ratio calculated for the algae sample is 0.278 ± 0.018, which is lower compared to the global fallout ratio of 0.80, due to a higher accumulation rate of americium in algae, and its preferential scavenging from surface water.
5. Conclusions Samples of fish flesh collected in 1999 in the south-western Caspian Sea in the Baku area, important for caviar production (sturgeons such as russkyi osyotr, sevruga and beluga), as well as for a consumption (such as roach and carp) were analysed for anthropogenic strontium, caesium, plutonium and americium, and natural polonium. The main results may be summarised as follows: • The highest massic activities were measured for anthropogenic 137 Cs in sevruga and beluga (1.2–1.8 Bq/kg ww), while the 90 Sr levels were between 5–12 mBq/kg ww only. The observed plutonium and americium levels were close to limits of detection (∼0.2 mBq/kg ww). The observed plutonium and strontium levels are in the same range as in the Mediterranean an Black Seas, whereas caesium has been accumulated in conditions of lower salinity in larger proportions. The 210 Po levels were between 0.2–3 mBq/kg ww, in agreement with previous investigations. The measured 210 Po levels in a fresh caviar (spawn) sample were higher by about factor of four than in samples of sturgeon flesh, but comparable with other fish samples. The highest radionuclide levels, by one to two orders of magnitude, were measured in a macroalgae sample. • The distribution of radionuclides in biota seems to be related more to the species than to environmental conditions (a coastal or an open sea), but physical factors (like a size and an age of fish) should be considered as well in further studies. • The estimated CFs for strontium and plutonium in fish and algae samples are in a reasonable agreement with IAEA recommended values (IAEA, 2004). Caesium has been accumulated in larger quantities, so that the resulting CF is by a factor of two higher than the recommended value. • The highest CFs for all investigated radionuclides were found for macroalgae, confirming that algae samples are suitable biomonitors of radioactive contamination.
Anthropogenic radionuclides in biota samples from the Caspian Sea
263
• The observed 137 Cs/90 Sr activity ratios are in agreement with the global fallout ratio, while the 239,240 Pu/90 Sr and 239,240 Pu/137 Cs ratios are much lower due to preferential scavenging of plutonium from surface water. • The measured radionuclide concentrations in biota samples do not represent any radiological risk from their consumption. The lack of information on other components of the marine ecosystem avoids to make further assumptions on the importance of diet or environmental conditions on the distribution of radionuclides in marine biota. Presented data have demonstrated a need to undertake more targeted studies in biota, surveying the main fish species together with their physical parameters (size, age). The algae seem to be well spread in the Caspian Sea to track radioactive contamination along the shore, especially from natural radionuclides released by the growing oil industry. Radionuclide data for sediments, which are still missing, as well as new oceanographic investigations would bring more information, and thus provide a comprehensive picture on the unique environment of the Caspian Sea.
Acknowledgement The Agency is grateful for the support provided to its Marine Environment Laboratory by the Government of the Principality of Monaco.
References Aarkrog, A., Baxter, M.S., Bettencourt, A.O., Bojanowski, R., Bologa, A., Charmasson, S., Cunha, I., Delfanti, R., Duran, E., Holm, E., Jeffree, R., Livingston, H.D., Mahapanyawong, S., Nies, H., Osvath, I., Pingyu, L., Povinec, P.P., Sanchez, A., Smith, J.N., Swift, D.A. (1997). A comparison of doses from 137 Cs and 210 Po in marine food: A major international study. Journal of Environmental Radioactivity 34, 69–90. Froehlich, K., Povinec, P.P. (1998). The Caspian Sea water dynamics. In: EEZ Technology. ICG Publishing Ltd., London, pp. 37–41. Froehlich, K., Rozanski, K., Povinec, P., Oregioni, B., Gastaud, J. (1999). Isotope studies in the Caspian Sea. The Science of the Total Environment 237/238, 419–427. Hong, G.H., Kim, Y.I., Lee, S.H., Cooper, L.W., Choe, S.M., Tkalin, A.V., Lee, E.T., Kim, S.H., Cung, C.S., Hirose, K. (2002). 239+240 Pu and 137 Cs concentrations for zooplankton and nekton in the Northwest Pacific and Antarctic oceans (1993–1996). Marine Pollution Bulletin 44, 660–665. International Atomic Energy Agency, IAEA (1995). Sources of radioactivity in the marine environment and their relative contributions to overall dose assessment for marine radioactivity (MARDOS). IAEA-TECDOC-838, IAEA, Vienna. International Atomic Energy Agency, IAEA (2004). Sediment distribution coefficients and concentration factors for biota in the marine environment. IAEA Technical Reports Series No. 422. IAEA, Vienna. Klige, R.K., Myagkov, M.S. (1992). Changes in the water regime in the Caspian Sea. GeoJournal 27, 299–307. Kosarev, A.N., Yablonskaya, E.A. (1994). The Caspian Sea. Academic Press, The Hague. La Rosa, J.J., Burnett, W., Lee, S.H., Levy, I., Gastaud, J., Povinec, P.P. (2001). Separation of actinides, cesium and strontium from marine samples using extraction chromatography and sorbents. Journal of Radioanalytical and Nuclear Chemistry 248, 765–770. Lazorenko, G.E., Polikarpov, G.G., Boltachev, A.R. (2002). Natural radioelement polonium in main ecological groups of the Black Sea fishes. Biologiya Morya 28, 61–65 (in Russian). Lazorenko, G.E., Polikarpov, G.G., Osvath, I. (2004). Doses to the Black Sea fishes and mussels from naturally occurring radionuclide 210 Po. In: Book of Extended Synopses. IAEA-CN-109/78. IAEA, Vienna, pp. 242–244.
264
J. Gastaud et al.
Lee, S.H., Gastaud, J., La Rosa, J., Liong Wee Kwong, L., Povinec, P.P., Wyse, E., Fifield, L.K., Hausladen, P.A., Di Tada, L.M., Santos, G.M. (2001). Analysis of plutonium isotopes in marine samples by radiometric, ICP-MS and AMS techniques. Journal of Radioanalytical and Nuclear Chemistry 248, 757–764. Oregioni, B., Gastaud, J., Pham, M.K., Povinec, P.P. (2003). Anthropogenic radionuclides in the Caspian Sea. Water Resources 30, 86–91. Othman, I., Yassine, T., Bhat, I.S. (1994). The measurement of some radionuclides in the marine coastal environment of Syria. The Science of the Total Environment 153, 57–60. Povinec, P.P., Froehlich, K., Gastaud, J., Oregioni, B., Pagava, S.V., Pham, M.K., Rusetski, V. (2003). Distribution of 90 Sr, 137 Cs and 239,240 Pu in Caspian Sea water and biota. Deep-Sea Research II 50, 2835–2846. Timoschuk, V.I. (1972). Strontium in water of the Caspian and Azov Seas. In: Polikarpov, G.G. (Ed.), Marine Radioecology. AEC-tr-72-7299, TID-4500. US AEC, Washington, DC, pp. 185–195. Vakulovsky, S., Chumichev, V.B. (1998). Radioactive contamination of the Caspian Sea. Radiation Protection Dosimetry 75, 61–64.
Further reading International Atomic Energy Agency, IAEA (2005). Worldwide marine radioactivity studies (WOMARS). IAEATECDOC-1429. IAEA, Vienna.
265
210 Po
in fish, algae, mussel and beach sediment samples collected along the Turkish coast of the Black Sea
Nurdan Güngör* , Emin Güngör, B. Gül Göktepe, Güler Köksal Turkish Atomic Energy Authority, Çekmece Nuclear Research & Training Center, Istanbul, Turkey Abstract 210 Po in sea fish, mussel, algae and beach sediment samples collected at four selected stations (Sile, ¸ Trabzon, Sinop and Zonguldak) along the Turkish coast of the Black Sea are presented and discussed. The observed 210 Po massic activities in beach sediments and algae were 5.9–79 Bq kg−1 dry weight (dw) and 11.5–35 Bq kg−1 dw, respectively. The 210 Po levels in fish were between 0.8 and 2.4 Bq kg−1 wet weight (ww) in whiting (Merlangius sp.), 1.6–8.9 Bq kg−1 ww in horse mackerel (Trachurus sp.), and 5.8–6.3 Bq kg−1 in red mullet (Mullus sp.). Possible
causes of these variations are “vital effects”, e.g. individual fish size and interfamily differences. The average activity of 210 Po in mussels with shell length 5 cm was 25 ± 10 Bq kg−1 ww. Keywords: 210 Po, Marine organisms, Fish, Algae, Mussel, Beach sand, Black Sea, Turkish coast
1. Introduction The largest contribution to radiation doses received by humans from the consumption of seafood, as well as for marine fauna, comes from naturally occurring polonium (ICRP, 1990; Aarkrog et al., 1997). All polonium isotopes (33) are radioactive. 210 Po with the half-life of 138.5 days is the most long-lived one. 210 Po belongs to the natural 238 U decay series and it is considered as one of the most radiotoxic nuclides in the environment, and the main source of the internal radiation exposure (Aarkrog et al., 1997). 210 Po is one of the daughter products of 222 Rn and therefore its main source in the environment is the exhalation of radon gas from the ground into the atmosphere. All 222 Rn daughter products, including 210 Po, are transferred from the atmosphere by dry and wet deposition on land and seas. The other sources of 210 Po include burning of fossil fuels, uranium, phosphate and lead ore processing industries, oil and gas industry, agriculture, tetraethyl lead used for car engines, etc. (Belli et al., 2001). Environmental levels of 210 Po, especially in the aquatic ecosystems, can be therefore due to these technological activities higher as natural ones. * Corresponding author. Address: ÇNAEM, P.O. Box 1 Atatürk Airport, ˙Istanbul, Turkey; phone: (+90) 212 548 40 50; fax: (+90) 212 548 22 30; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08020-4
© 2006 Elsevier Ltd. All rights reserved.
266
N. Güngör et al.
210 Po
enters to the human body via either inhalation or ingestion. The major contribution to the radiation dose received by humans from 210 Po is via a consumption of marine products, especially fish and shellfish, which accumulate this radionuclide at high concentrations (Godoy, 1990; Carvalho and Fowler, 1994; Aarkrog et al., 1997). It has been shown that 210 Po activity in shellfish and fish depends on the place of their dwelling (Swift et al., 1995; Güngör et al., 2001). Therefore, it is important to determine 210 Po in marine biota in different locations and time of sampling, as its levels are controlled not only by environmental inputs but also by climate conditions and marine biogeochemical processes. A wide scope Regional Technical Co-operation Project RER/2/003 “Marine Environmental Assessment of the Black Sea Region” was implemented by the International Atomic Energy Agency (IAEA) in 1995–2003. This project was initiated in response to the needs of participating Member States – the six Black Sea coastal countries (Bulgaria, Georgia, Romania, Russian Federation, Ukraine and Turkey) with the aim to establish capabilities for reliable assessment of radionuclides in the Black Sea environment, and to apply tracer techniques to marine contamination studies. This particular study was carried out to investigate the distribution of 210 Po in fish, algae, mussel and beach sand samples collected along the Turkish coast of the Black Sea in the period June 1998–November 2001, as a contribution to the international coastal radioactivity monitoring programme carried out by the Black Sea countries. 2. Materials and methods 2.1. Sampling In the framework of the IAEA project a joint monitoring programme was initiated in 1996 in which marine samples were collected from the selected stations of each country and laboratory analyses were carried out twice a year by using a harmonized methodology. The four monitoring stations of along the Turkish coast of the Black Sea (Trabzon, Sinop, Zonguldak and Sile) ¸ are shown in Fig. 1. The sampling programme included these species: (i) fish: whiting Merlangius sp., horse mackerel Trachurus sp. and red mullet Mullus sp., (ii) mussel Mytilus gallopprovincialis, (iii) brown algae Cystoseria barbata, (iv) beach sediment and (v) surface water. A detailed analysis of anthropogenic as well as natural radionuclides of all the above samples were carried out twice a year (June 1998–November 2001). Sampling strategy, methodology for sampling and radionuclide analyses were carried out following the agreed procedures. The algae and beach sand samples were collected on the selected stations. Fish samples were either caught or purchased at markets of the Turkish Black Sea coasts. Samples of mussels of the same size, fish and algae were stored in plastic bags in an ice box, and transferred to the laboratory. 2.2. Analyses The samples were dried at 85◦ C to constant weight and homogenized. Analyses were made on dry tissues of the samples. Only the edible parts of the fish samples were analyzed. The beach
210 Po in fish, algae, mussel and beach sediments of the Turkish coast of the Black Sea
267
Fig. 1. Turkish monitoring stations in the Black Sea.
sand samples were dried in a porcelain tray at 105◦ C. After all samples were homogenized, sub-samples were taken for parallel analyses. The measurements of 210 Po were made using a standard method (Güngör et al., 2001). 209 Po (∼0.03 Bq) was used as a yield tracer. About 0.3–0.5 g of dry samples were digested using 65% HNO3 . After evaporation, the dry residue was washed with concentrated HCl. Polonium was plated onto a silver disc in 0.5 M HCl in presence of ascorbic acid. Alpha-spectrometry of 210 Po samples was carried out using silicon surface barrier detectors (ORTEC). Counting times were adjusted to obtain relative standard uncertainties of ∼5%. For checking the accuracy and precision of the analytical method, the IAEA-300 reference material was used.
3. Results and discussion We present in this paper results on 210 Po analysis of biota and beach sediment samples (Table 1). The range of 210 Po massic activities in Black Sea fish varies from 0.8 to 8.9 Bq kg−1 wet weight (ww). Although 210 Po levels decrease from benthic to pelagic and to demersal fish, there are no significant differences between fish species investigated in this work. The data of the Institute of Biology of the Southern Seas (IBSS, Ukraine) have shown that the 210 Po levels in 17 species of Black Sea fish (for the entire body of fish) collected along the Crimean coast in 1998–2000 ranged from 0.69 to 36 Bq kg−1 ww (Lazorenko and Polikarpov, 2001; Lazorenko et al., 2003, 2004). Our values are on the lower side, comparable with nonindustrially exposed environments. When we compare our results with the results from the global survey (IAEA, 1995; Aarkrog et al., 1997), the reported average activity of 210 Po in fish (2.4 Bq kg−1 ww) is close to our average value 3.5 ± 2.5 Bq kg−1 ww.
268
N. Güngör et al. Table 1 210 Po massic activities of fish and mussel samples in Bq kg−1 wet weight (uncertainties represent 1 sigma standard deviations) Location
Date
Sile ¸
June/1998 Dec./1998 July/1999 Nov./1999 June/2000 June/2001 Nov./2001 June/1998 June/1999 June/2000 June/2001 Nov./2001 June/1998 Dec./1998 July/1999 June/2000 Nov./2001 June/1998 Dec./1998 June/1999 Nov./1999 Dec./2000 June/2001 Nov./2001
Zonguldak
Sinop
Trabzon
Horse mackerel
Whiting
Red mullet
1.2 ± 0.2 2.5 ± 0.9
1.4 ± 0.4
3.3 ± 0.4 6.2 ± 0.6 1.6 ± 0.2
1.4 ± 0.2
5.8 ± 1.3
2.1 ± 0.3 8.9 ± 1.1
7.1 ± 0.7
8.1 ± 1.4 6.7 ± 0.6 3.4 ± 0.4 5.3 ± 0.5
1.8 ± 0.2 0.8 ± 0.1 2.4 ± 3.8 1.7 ± 0.4
6.3 ± 0.7
0.8 ± 0.1 2.0 ± 0.3 1.2 ± 0.5 1.6 ± 0.2 2.9 ± 0.3 1.9 ± 0.3
6.7 ± 1.3
Mussel 20.6 ± 1.7 9.3 ± 0.3 20.5 ± 1.8 29.4 ± 3.3 29.8 ± 2.1 11.7 ± 1.1 21.9 ± 1.6 32.1 ± 3.7 19.2 ± 1.8 17.1 ± 1.6 35 ± 16 29.8 ± 2.1 25.4 ± 1.3 51.6 ± 2.5 25.1 ± 2.2 26.5 ± 4.9 28.2 ± 1.9 11.9 ± 1.0 22.3 ± 2.6 17.8 ± 1.1 19.6 ± 2.9 20.4 ± 1.5 43.1 ± 3.7
The 210 Po levels in mussels and algae are higher than in all analyzed fish samples (Table 1), as they vary between 9 and 52 Bq kg−1 ww; the average value is 25 ± 10 Bq kg−1 . The MARDOS (IAEA, 1995) value is 15 Bq kg−1 ww in molluscs, therefore our results show elevated 210 Po levels in mussels. Mussels because of their high filtration rates, particle retention ability and ingestion efficiency, can accumulate 210 Po in higher quantities than the other species of the Black Sea hydrobionts (Göktepe et al., 1998, 2000; Turkish National Report, 2003). Swift et al. (1995) noticed that the accumulation of 210 Po by mussels Mytilus edulis depends on contamination levels in their environment. As one can see from Table 1, the 210 Po levels in mussels collected at four stations vary slightly. However, one of our stations (Sile), ¸ which is located between four large cities of the Turkish coast, is relatively less contaminated. The highest 210 Po massic activities in beach sediment samples were recorded during June 2000 at Zonguldak station. The most probable reason is the coal mining industry in Zonguldak. This area has the main coal reserves in Turkey and is the major domestic hard coal supplier. 210 Po levels in sediments from other stations varied between 5.4 ± 1.1 and 29.5 ± 1.0 Bq kg−1 dw, which are comparable with the results measured in the Aegean Sea (U˘gur-Tanbay and Yener, 2001).
210 Po in fish, algae, mussel and beach sediments of the Turkish coast of the Black Sea
269
4. Conclusions 210 Po
in sea fish, mussels, algae and beach sediments collected at four selected stations (Sile, ¸ Trabzon, Sinop and Zonguldak) along the Turkish coast of the Black Sea was determined. The observed 210 Po massic activities in beach sediments and algae were 5.9–79.2 Bq kg−1 dry weight (dw) and 11.5–34.8 Bq kg−1 dw, respectively; see Table 2. The levels in fish were between 0.8 and 2.4 Bq kg−1 wet weight (ww) in whiting (Merlangius sp.), 1.6– 8.9 Bq kg−1 ww in horse mackerel (Trachurus sp.), and 5.8–6.3 Bq kg−1 ww in red mullet (Mullus sp.). Possible causes of these variations are “vital effects”, e.g. individual fish size and interfamily differences. The average activity of 210 Po in mussels with shell length 5 cm was 25 ± 10 Bq kg−1 ww. The data accumulated as a result of this monitoring programme will be used for the further assessment of the Black Sea marine environment. The environmental pollution problems, ecological degradation and the rehabilitation efforts of the Black Sea are complex international cooperation issues, which have been gaining great interests in recent years. Further studies are
Table 2 210 Po massic activities of algae and beach sediment samples in Bq kg−1 dry weight (uncertainties represent 1 sigma standard deviations) Location
Date
Sile ¸
June/1998 Dec./1998 July/1999 Nov./1999 June/2000 June/2001 Nov./2001 June/1998 Dec./1998 June/1999 June/2000 June/2001 Nov./2001 June/1998 Dec./1998 July/1999 June/2000 June/2001 Nov./2001 June/1998 Dec./1998 June/1999 Nov./1999 Dec./2000 June/2001 Nov./2001
Zonguldak
Sinop
Trabzon
Algae
28.5 ± 3.5 19.2 ± 2.5 17.1 ± 1.7 8.7 ± 1.6 25.3 ± 2.4 30.3 ± 3.0 34.8 ± 2.4 24.7 ± 2.2 10.5 ± 1.3 23.9 ± 2.4 14.2 ± 1.2 17.3 ± 1.4 19.0 ± 2.7 25.0 ± 1.6 18.2 ± 2.3 24.8 ± 2.6 11.5 ± 1.6 18.9 ± 2.6 15.1 ± 2.3 16.9 ± 2.4 17.6 ± 1.8 16.2 ± 1.9 16.8 ± 1.9
Beach sediment 6.2 ± 1.0 15.0 ± 1.5 23 ± 5 5.9 ± 1.5 9.2 ± 0.9 6.7 ± 1.0 9.2 ± 1.5 10.4 ± 3.0 13.9 ± 3.3 8.1 ± 2.3 73 ± 6 6.6 ± 1.5 7.9 ± 1.0 6.5 ± 1.3 8.2 ± 1.5 8.4 ± 2.8 10.2 ± 1.6 5.4 ± 1.1 14.1 ± 1.0 19.2 ± 3.5 19.4 ± 2.8 20.9 ± 4.5 15.3 ± 1.3 17.7 ± 2.2 13.0 ± 1.6 29.5 ± 1.0
270
N. Güngör et al.
planned for assessment of doses due to 210 Po and other radionuclides contained in sea food (fish and mussel) consumed by Turkish people in the Turkish Black Sea coastal area.
Acknowledgements This work was carried out in the framework of the IAEA’s Regional Technical Co-operation Project RER/2/003 (2004) “Marine Environmental Assessment in the Black Sea Region”. We thank the project managers Ms. I. Osvath, Mr. M. Samiei and Mr. A. Chupov for their constant support, national scientific team members of the Black Sea countries for fruitful collaboration, and authorities in the Black Sea cities in Turkey for their interest in our work and for their help during sampling campaigns. We also thank two anonymous reviewers and Prof. P. Povinec for constructive comments and help with the manuscript.
References Aarkrog, A., Baxter, M.S., Bettencourt, A.O., Bojanowski, R., Bologa, A., Charmasson, S., Cunha, I., Delfanti, R., Duran, E., Holm, E., Jeffree, R., Livingston, H.D., Mahapanyawong, S., Nies, H., Osvath, I., Pingyu, L., Povinec, P.P., Sanchez, A., Smith, J.N., Swift, D. (1997). A comparison of doses from 137 Cs and 210 Po in marine food: A major international study. Journal of Environmental Radioactivity 34, 217–218. Belli, G.J.M., Blasi, M., Marchetti, A., Rosamilia, S., Sansone, U. (2001). Determination of 210 Pb and 210 Po in mineral and biological environmental samples. Journal of Radioanalytical and Nuclear Chemistry 247, 491–499. Carvalho, F.P., Fowler, S.W. (1994). A double-tracer technique to determine the relative importance of water and food as sources of 210 Po to marine prawns and fish. Marine Ecology – Progress Series 103, 251–264. Godoy, J.M.O. (1990). Determination of radionuclides in food and environmental samples. IAEA Interregional Training Course, Karlsruhe Nuclear Research Center, 23 April–25 May 1990. Göktepe, B.G., Samiei, M., Osvath, I. (1998). Marine environmental assessment in the Black Sea region. FISHECO98, First International Symposium on Fisheries and Ecology, Trabzon, 2–4 September 1998. Göktepe, B.G., Köksal, G., Osvath, I. (2000). Marine environmental assessment in the Black Sea region – A case for the Turkish coastal zone, I. Eurasia Conference on Nuclear Sciences and Its Applications, Izmir, 23–27 October 2000. Güngör, N., Topcuo˘glu, S., Kırba¸so˘glu, Ç. (2001). 210 Po and 210 Pb concentrations in biota from the Turkish coast of the Black Sea and Marmara Sea. Rapports Commission Internationale pour l’exploration scientifique de la Mer Méditerranée 36, 132. International Atomic Energy Agency (IAEA) (1995). Sources of radioactivity in the marine environment and their relative contributions to overall dose assessment for marine radioactivity (MARDOS). IAEA-TECDOC-838, IAEA, Vienna. International Commission on Radiation Protection (ICRP) (1990). Recommendations of the Int. Com. on Radiological Protection. ICRP Publication 60, Ann. ICRP, vol. 21 (1–3), Oxford. Lazorenko, G.E., Polikarpov, G.G. (2001). 210 Po in the Black Sea hydrobionts: The radioecolgy–ecotoxicology of continental and estuarine environments. In: Abstracts Book of Intern. Congress ECORAD 2001. Aix-en-Provence (France), 3–7 Sept. 2001, pp. 4–16. Lazorenko, G.E., Polikarpov, G.G., Osvath, I. (2003). Doses to the Black Sea fishes and mussels from naturally occurring radionuclide 210 Po. In: Intern. Conf. on Protection of the Environment from the Effects of Ionizing Radiation. 6–10 Oct. 2003, Stockholm, Sweden. Contributed Papers: IAEA-CN-109, pp. 242–244. Lazorenko, G.E., Polikarpov, G.G., Osvath, I. (2004). 210 Po in the North-West Black Sea ecosystem. Intern. Conf. on Isotopes in Environmental Studies – Aquatic Forum 2004 – 25–29 Oct. 2004, Monte Carlo, Monaco. Book of Extended Synopses: IAEA-CN-118, p. 405. Regional Technical Co-operation Project RER/2/003 (2004). Marine Environmental Assessment of the Black Sea. Working Material. IAEA, Vienna.
210 Po in fish, algae, mussel and beach sediments of the Turkish coast of the Black Sea
271
Swift, D.J., Smith, D.L., Allington, D.L., Winpenny, K. (1995). A laboratory and field study of 210 Po depuration by edible winkles (Littorina littorea L.) from the Cambrian Coast. Journal of Environmental Radioactivity 26, 119–133. Turkish National Report, IAEA TCP RER/2/003 (2003). Marine environmental assessment of the Black Sea region. Draft Final Report of Turkey. U˘gur-Tanbay, A., Yener, G. (2001). Accumulation rates and sediment deposition in the Gökova Bay in Aegean Sea Turkish Coast. Applied Radiation and Isotopes 55, 581–588.
272
210 Po
in mussels (Mytilus galloprovincialis) and sediments along the Turkish coast of the Aegean Sea
Aysun U˘gura,* , Güngör Yenera , Sayhan Topcuo˘glub , U˘gur Sunluc , Serpil Aközcana , Banu Özdena a Ege University, Institute of Nuclear Sciences, Bornova, Izmir, ˙ Turkey b Çekmece Nuclear Research Center, Istanbul, ˙ Turkey c Ege University, Faculty of Fisheries, Department of Hydrobiology, Bornova, Izmir, ˙ Turkey
Abstract 210 Po results for mussels (soft tissue) and sediment samples from 9 stations along the heavily industrialized Turkish
coastal zone of the Aegean Sea are presented and discussed. The mean massic activities in mussels varied between 190 ± 10 and 830 ± 40 Bq kg−1 dry weight (dw). The results show significant differences in concentrations for similar size mussels collected at different sites. 210 Po activities are considerably higher in spring mussels than in summer samples of the same size, especially in Çe¸sme, Bodrum and Didim. The highest activity (1200 ± 100 Bq kg−1 dw) was found in a Çe¸sme sample. Higher 210 Po activities were observed in smaller mussels than in larger individuals. The average 210 Po massic activities in sediments varied in the range of 20 ± 3 and 144 ± 7 Bq kg−1 dw. ˙Izmir Bay exhibited the highest polonium activities in sediments, likely due to specific sedimentation processes and other sediment characteristics. Keywords: 210 Po, Mussel, Sediment, Bioaccumulation, Aegean Sea
1. Introduction In coastal marine and estuarine environments, polonium mostly associates with sedimenting material, and mean residence times in coastal water masses between a few months and 2 years have been reported (Nozaki et al., 1991, 1998). For surface and coastal waters, the atmospheric input of radon daughters explains the main flux of 210 Pb and 210 Po, whereas in situ production
* Corresponding author. Address: Ege University, Institute of Nuclear Sciences, 35100, Bornova, ˙Izmir, Turkey;
phone: (+90) 232 3886466; fax: (+90) 232 3886466; e-mail:
[email protected] RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08021-6
© 2006 Elsevier Ltd. All rights reserved.
210 Po in mussels and sediments along the Turkish coast of the Aegean Sea
273
from 226 Ra in sediment is the main source in deeper oceanic waters (Bacon and Elzerman, 1980; Nozaki et al., 1991, 1998). 210 Po is a naturally occurring radionuclide (alpha-emitter) whose activity has been measured in algae, marine invertebrates and fish from different seas and oceans, since it is an important source of the natural radiation received internally by marine organisms, as well as by humans after seafood consumption (Fisher et al., 1983; Gouvea et al., 1992; Carvalho and Fowler, 1994; Carvalho, 1995; Hameed et al., 1997; Aarkrog et al., 1997). Stewart and Fisher (2003) indicated that 210 Po is ubiquitous in seawater, especially enriched in proteinaceous tissues of marine organisms, and may therefore be useful as a tracer of organic carbon flux in marine systems. In general, adsorption, absorption and ingestion are three distinct environmental processes for the entry of radioactive elements into marine organisms. A considerable accumulation of radioelements may occur through the food chain. This is particularly the case with filter feeders, which ingest detritus material with a high degree of radionuclide association, and so, the filter feeding mussels have been recognized internationally as first-order biological indicators of radioactive contamination. They have high bioconcentration factors for most contaminants found in the sea, they are abundant in coastal areas, and they have large geographical distribution, sessile nature, long life span and high filtration rates (Alam et al., 1999; Beiras et al., 2003). The extraction and processing of earth materials, or their industrial products or by-products, may increase the incorporation of radionuclides to the hydrosphere through surface and/or ground waters. In river basins supporting considerable volumes of human activity, many contaminants can be exported from urban or industrial areas via rainwater and runoff, and discharged into the aquatic ecosystem, e.g. Pujol and Sanchez-Cabeza (2000). Bottom sediments can accumulate radionuclides due to deposition of suspended sediment and by direct adsorption from overlying water. Marine sediments and mussels are commonly used as environmental matrices in chemical and radioactive monitoring programmes because they accumulate persistent contaminants at concentrations orders of magnitude above those in the water (e.g. Beiras et al., 2003). The Aegean Sea, in the Mediterranean Sea (640 km long, 320 km wide, 214,000 km2 ) is located off SE Europe between Greece and Turkey, Crete and Rhodes. Its maximum depth is 2500 m. The Dardanelles Strait connects the Aegean Sea with the Sea of Marmara and the Black Sea. The Turkish coastal zone of the Aegean Sea has been heavily industrialized in the last 25 years resulting in a considerable input of wastes to the coastal marine ecosystem. Consumption of fish and other edible marine organisms is relatively higher in the region than at other parts of Turkey. Therefore it is important to assess the additional radiation dose received by population consuming from seafood. The information on levels and on distribution of natural radionuclides is, however, sparse due to limited number of investigations conducted on this coast line (U˘gur et al., 2002; Tanbay et al., 1999). The aim of the present study has been to provide data on polonium levels along the Aegean Turkish coastal zone, and to assess its enhancement from the fossil fuel industry like coal-fired power plants (CPPs), oil and phosphate industry, all located along the coast and discharging wastes via rivers into the sea.
274
A. U˘gur et al.
2. Materials and methods 2.1. Sampling Mussels (Mytilus galloprovincialis) were collected at 9 stations indicated in Fig. 1. The sampling sites were chosen on the basis of possible sources for organic and trace metal contaminants, reflecting different anthropogenic impacts, and also to compare contamination levels at 9 stations placed along the coast. Sediments and mussels were sampled at the same sites from spring to summer. Immediately after collection, the shells were cleaned with a nylon brush and carefully rinsed with abundant distilled water in order to eliminate sediment and other impurities. The composite samples were weighed and oven dried to a constant weight at 80◦ C, mixed, ground, sieved through a 2 mm mesh, homogenized and analyzed. The mussels were
Fig. 1. Sampling locations in the Aegean Sea (Çanakkale, Ayvalık, Dikili, Foça, ˙Inciraltı, Çe¸sme, Didim, Bodrum, Marmaris).
210 Po in mussels and sediments along the Turkish coast of the Aegean Sea
275
of varying sizes from 1 to 9 cm long. Çe¸sme, Didim and Bodrum mussels were studied in three different groups according to their sizes and sampling time. Sediment samples from different locations were recovered using a Van-Veen grab (5 L) near the shore (10–40 m) of the stations. The collected sediments were oven-dried to constant weight at 80◦ C, sieved with 0.25 mm mesh size and homogenized before analysis. 2.2. Radiochemical procedures After addition of 208 Po tracer, each sample was completely dissolved with equal volumes of HCl and HNO3 . For sediment samples HF was also used in the dissolving process (HNO3 :HF:HCl = 3:3:35 mL). Polonium was spontaneously plated onto silver discs in 0.5 N HCl in the presence of ascorbic acid to reduce Fe+3 to Fe+2(3) (Flynn, 1968). Recoveries varied between 70% and 90% for mussels, and between 70% and 80% for sediment samples. Measurements of 210 Po were carried out using its 5.30 MeV alpha-rays, with 208 Po (5.11 MeV, t1/2 = 2.9 yr) as the internal tracer. Alpha-activities were measured using silicon surface-barrier detectors (Tennelec, 400 mm2 , 300 µm depletion depth), having efficiency of 29% for 226 Ra source. 210 Po activity was corrected for a recovery and the time of sampling. The separation of the two polonium peaks was usually good, but small corrections (below 10%) to adjust for tailing of the 210 Po peak into the 208 Po region were sometimes necessary. Counting period was adjusted to obtain a relative standard uncertainty of ∼5%. The detection limit was 3 mBq. The final activity calculations included all corrections (recovery, background, blanks, etc.).
3. Results and discussion 3.1. Mussel samples Mean massic activities of 210 Po in samples of all sizes, collected during a year at each station, are given in Table 1. Figure 2 shows the distribution of 210 Po activity in mussels (4–6 cm) collected in summer time versus the different sampling stations along the Aegean coast. Maximum levels (830 ± 50 Bq kg−1 dw) were found in mussels of Foça, where two years ago U˘gur et al. (2002) observed elevated levels of 210 Po (1200 ± 100 Bq kg−1 dw) in Mytilus galloprovincialis. Also in Dikili, another station near to Foça, 210 Po levels were quite high (720 ± 50 Bq kg−1 dw). A number of sources could be responsible for elevated 210 Po levels in the aquatic environment of this region. A fertilizer factory, 50 km from both sampling sites, has been operating for nearly 26 years. The other possible source, especially for Dikili, could be discharges via Bakırçay River. This is one of the largest fluvial systems discharging into the Aegean Sea (the river course is 110 km long). Along the Bakırçay course, large cities, agricultural areas and technological activities such as Soma CPP are present. The petrochemical industry on the coast of Alia˘ga could be another agent for high 210 Po levels observed in the region. Çe¸sme, Didim and Bodrum data are relatively high, but similar to other studies conducted in other parts of the Mediterranean Sea (McDonald et al., 1996). Bodrum and Didim are of particular concern because these regions have been contaminated by three major uraniferous
276
A. U˘gur et al. Table 1 Dry/wet weight and mean massic activity of 210 Po in mussels, collected during a year at the Aegean coast of Turkey. The uncertainties represent 1 sigma standard deviations Station
Dry/wet weight
210 Po (Bq kg−1 dw)
Çanakkale (Dardanelles) Ayvalık (Kydonia) Dikili (Aterneus) Foça (Phokaia) ˙Inciraltı (Smyrna) Çe¸sme (Cyssus) Didim (Didyma) Bodrum (Halicarnassos) Marmaris (Karya)
0.19 ± 0.02 0.17 ± 0.03 0.10 ± 0.01 0.14 ± 0.03 0.23 ± 0.04 0.16 ± 0.03 0.16 ± 0.05 0.22 ± 0.04 0.18 ± 0.03
220 ± 10 300 ± 20 720 ± 50 830 ± 50 260 ± 10 550 ± 20 510 ± 20 440 ± 20 190 ± 10
Fig. 2. Massic activities of 210 Po (Bq kg−1 dry weight) in mussels (4–6 cm) (Mytilus galloprovincialis) collected during summer (relative uncertainties are below 10%).
CPPs since 1982. Surprisingly, one of the highest 210 Po levels were found in mussels from Çe¸sme, where there is no evidence of a polonium source in this site, remote from industrial activities. A possible explanation could be the occurrence of hydrothermal sources in the region. Çe¸sme, Didim and Bodrum mussels are studied also in three different groups according to their sizes and sampling time. The results shown in Figs. 3 and 4 indicate that smaller mussels have higher 210 Po activities than larger ones in summer period in relation with the size effect influenced by the sampling time. On the other hand Fig. 4 shows that the levels for 4–6 cm size mussels are lower in the summer and higher in the spring. Such differences might be explained by the effect of age and the differences in the metabolic rates. Hence, the comparison between 210 Po concentrations in mussels along the coast has to be made for individuals of similar size, collected during the same season. The lowest mean 210 Po activities are observed in Marmaris, Çanakkale and ˙Inciraltı mussels. Marmaris is the station at far south of the Aegean Sea coast, whereas Çanakkale is located
210 Po in mussels and sediments along the Turkish coast of the Aegean Sea
277
Fig. 3. Massic activities of 210 Po (Bq kg−1 dry weight) in mussels (Mytilus galloprovincialis) of different size collected at three stations during summer (n = 5) (relative uncertainties are below 10%).
Fig. 4. Massic activities of 210 Po (Bq kg−1 dry weight) in mussels (4–6 cm) (Mytilus galloprovincialis) collected in spring and summer at three stations (n = 5) (relative uncertainties are below 10%).
at the northern end of the coast. In ˙Inciraltı, 210 Po is low in mussels, but high in sediments, as was also observed in a previous work (Saçan, 2004). The present results can be compared with 210 Po levels reported in a study conducted on Black Sea mussels (Topcuo˘glu, 2000; Topcuo˘glu et al., 2001). Although most of the data obtained so far belong to the fairly dry season, the 210 Po levels in Aegean Sea mussels are considerably higher than those in Black Sea mussels. This could be due to the fact that the Aegean region is highly industrialized and urbanized, and there are also intense agricultural activities with dense use of fertilizers. All these wastes, industrial and agricultural residues can be carried by rivers (Bakırçay, Gediz and Menderes) and streams into the sea. 3.2. Sediment samples The highest 210 Po massic activities in sediments along the coast were measured in ˙Izmir Bay (Table 2). This is possibly due to the fact that there are many streams flowing into the bay, passing through a number of industrial and agricultural areas. 210 Po level increases with additional material derived from coastal erosion, particulate matter flux from atmosphere and
278
A. U˘gur et al. Table 2 Dry/wet weight and mean activity of 210 Po in sediments from Aegean coast of Turkey (n = 3). The uncertainties represent 1 sigma standard deviations Station
Dry/wet weight
210 Po
Çanakkale (Dardanelles) Ayvalık (Kydonia) Dikili (Aterneus) Foça (Phokaia) ˙Izmir Bay (Smyrna) Çe¸sme (Cyssus) Didim (Didyma) Bodrum (Halicarnassos) Marmaris (Karya)
0.43 ± 0.16 0.67 ± 0.01 0.55 ± 0.16 0.77 ± 0.04 0.34 ± 0.18 0.80 ± 0.10 0.80 ± 0.07 0.86 ± 0.02 0.53 ± 0.13
35 ± 4 35 ± 3 43 ± 5 60 ± 20 144 ± 7 36 ± 4 30 ± 4 32 ± 4 20 ± 3
(Bq kg−1 dw)
biogenic activity within the ˙Izmir Bay. Particulate matter supplied by 11 streams is rapidly deposited. Limited water exchange of this area with the Aegean Sea play an important role in sediment deposition. In addition, the increase of nutrients has resulted in a higher degree of eutrophication in the inner part of Izmir Bay (Sunlu and Sunlu, 2001). Biological material rapidly sinks and deposit on the bottom sediment. The bottom of the bay is covered by soft mud. The carbon contents of Izmir Bay sediment range between 1.12% and 5.39% (TUB˙ITAK, 2002). Also, wastes from petrochemical and fertilizer industries at Alia˘ga region affect the contamination in ˙Izmir Bay, Dikili and Foça sediments. 210 Po levels in sediment of other stations varied between 20 ± 3 and 36 ± 4 Bq kg−1 , which are comparable with the supported 210 Po activities (in equilibrium with 226 Ra) previously reported for the same region (20.6 ± 0.9 to 35.6 ± 1.9 Bq kg−1 ) (U˘gur-Tanbay and Yener, 2001).
4. Conclusions Results showed that small mussels display higher 210 Po activities than larger individuals (summer sampling). A seasonal variability in 210 Po activities in mussels has been observed, i.e. activities measured in mussels of a given size class were higher in spring than in summer. However, it should be taken into account that the radionuclide concentrations in mussels do not only reflect their levels in the surrounding seawater. Indeed, a series of factors may also affect the accumulation of radionuclides in mussels, e.g. the sampling period (time of the year), the tide levels and the size of mussel (Baseda et al., 2002). Studies still continue to provide comparative data on 210 Po levels in different types of biota living in the same area, together with temporal changes, in order to assess a contribution of seafood to radiation doses to which the populations are exposed in the western Anatolia, as seafood, including mussels, constitute a substantial part of the local diet.
210 Po in mussels and sediments along the Turkish coast of the Aegean Sea
279
Acknowledgements This work has been supported by International Atomic Energy Agency (IAEA), Contract No. B5-TUR-31834, and the Ege University Sciences and Technology Research Center (EB˙ILTEM), Contract No. 2004/B˙IL/016.
References Aarkrog, A., Baxter, M.S., Bettencourt, A.O., Bojanowski, R., Bologa, A., Charmasson, S., Cunha, I., Delfanti, R., Duran, E., Holm, E., Jeffree, R., Livingston, H.D., Mahapanyawong, S., Nies, H., Osvath, I., Pingyu, Li., Povinec, P.P., Sanchez, A., Smith, J.N., Swift, D. (1997). A comparison of doses from 137 Cs and 210 Po in marine food; a major international study. Journal of Environmental Radioactivity 34, 69–90. Alam, M.N., Chowdhury, M.I., Kamal, M., Ghose, S., Matin, A.K.M.A., Ferdousi, G.S.M. (1999). Radionuclide concentrations in mussels collected from the southern coast of Bangladesh. Journal of Environmental Radioactivity 47, 201–212. Bacon, M.P., Elzerman, A.W. (1980). Enrichment of 210 Pb and 210 Po in the sea-surface microlayer. Nature 284 (5754), 332–334. Baseda, V., Fumega, J., Vaamonde, A. (2002). Temporal trends of Cd, Cu, Hg, Pb and Zn in mussel (Mytilus galloprovincialis) from the Spanish North-Atlantic coast 1991–1999. The Science of the Total Environment 288 (3), 239–253. Beiras, R., Bellas, J., Fernandez, N., Lorenzo, J.I., Cobelo-Garcia, A. (2003). Assessment of coastal marine pollution in Galicia (NW Iberian Peninsula); metal concentration in seawater, sediments and mussels (Mytilus galloprovincialis) versus embryo-larval bioassays using Paracentrotus lividus and Ciona intestinalis. Marine Environmental Research 56, 531–553. Carvalho, F.P. (1995). 210 Po and 210 Pb intake by the Portuguese population: The contribution of seafood in the dietary intake of 210 Po and 210 Pb. Health Physics 69 (4), 469–480. Carvalho, F.P., Fowler, S.W. (1994). A double-tracer technique to determine the relative importance of water and food as sources of polonium-210 to marine prawns and fish. Marine Ecology – Progress Series 103, 251–264. Fisher, N.S., Burns, K.A., Cherry, R.D., Heyraud, M. (1983). Accumulation and cellular distribution of 241 Am, 210 Po and 210 Pb in two marine algae. Marine Ecology – Progress Series 11, 233–237. Flynn, W.W. (1968). The determination of low levels of polonium-210 in environmental materials. Analytical Chimica Acta 43, 221–227. Gouvea, R.C., Santos, P.L., Dutra, I.R. (1992). Lead-210 and polonium-210 concentration in some species of marine molluscs. The Science of the Total Environment 112, 263–267. Hameed, P.S., Shaheed, K., Somasundaram, S.S.N. (1997). Bioaccumulation of 210 Pb in the Kaveri River ecosystem, India. Journal of Environmental Radioactivity 37 (1), 17–27. McDonald, P., Baxter, M.S., Scott, E.M. (1996). Technological enhancement of natural radionuclides in the marine environment. Journal of Environmental Radioactivity 32, 67–90. Nozaki, Y., Tsubota, H., Kasemsupaya, V., Yashima, M., Ikuta, N. (1991). Residence times of surface water and particle-reactive 210 Pb and 210 Po in the East China and Yellow Seas. Geochimica et Cosmochimica Acta 55 (5), 1265–1272. Nozaki, Y., Dobashi, F., Kato, Y., Yamamoto, Y. (1998). Distribution of Ra isotopes and the 210 Pb and 210 Po balance in surface seawaters of the mid Northern Hemisphere. Deep-Sea Research I 45, 1263–1284. Pujol, Ll., Sanchez-Cabeza, J.A. (2000). Natural and artificial radioactivity in surface waters of the Ebro River basin (Northeast Spain). Journal of Environmental Radioactivity 51, 181–210. Saçan, S. (2004). Periodic investigation of 210 Pb and 210 Po accumulation in surfacial sediments and black mussels (Mytilus galloprovincialis L. 1758) at ˙Izmir Bay. MSc in Nuclear Sciences. University of Ege, Turkey. Sunlu, F.S., Sunlu, U. (2001). Temporal variations of nutrients and chlorophylla in Urla Coast (Izmir Bay Aegean Sea – Turkiye). Rapports Commission Internationale pour l’exploration scientifique de la Mer Méditerranée 36, 420. Stewart, G.M., Fisher, N.S. (2003). Bioaccumulation of polonium-210 in marine copepods. Limnology and Oceanography 48, 2011–2019.
280
A. U˘gur et al.
Tanbay, A.U., Yener, G., Mulsow, S., Fowler, S.W., Duman, M. (1999). Natural and Man-Made Radionuclide Concentrations in Marine Sediments of Gökova Bay, Aegean Turkish Coast. International Atomic Energy Agency, Vienna, 350 pp. Topcuo˘glu, S. (2000). Black Sea ecology pollution research in Turkey of the marine environment. IAEA Bulletin 42 (4), 12–14. Topcuo˘glu, S., Güngör, N., Kırba¸so˘glu, Ç. (2001). To establish and compare for the coastal waters of Turkey, the 210 Po–210 Pb contents in anchovy fish and sea snails. Turkish Atomic Energy Authority, Research Contract No. 9712/R1/R0 (internal report). TUB˙ITAK (2002). The effects of ˙Izmir big channel waste-water treatment project to the lower trophic level of ˙Izmir Bay (Aegean Sea Turkey). TUB˙ITAK/YDBAG Project No. 102Y116. Project Coordinator: U. Sunlu, in press. U˘gur-Tanbay, A., Yener, G. (2001). Accumulation rates and sediment deposition in the Gökova Bay in Aegean Sea Turkish Coast. Applied Radiation and Isotopes 55, 581–588. U˘gur, A., Yener, G., Ba¸ssarı, A. (2002). Trace metals and 210 Po (210 Pb) concentrations in mussels (Mytilus galloprovincialis) consumed at western Anatolia. Applied Radiation and Isotopes 57, 565–571.
281
Stable nitrogen isotopes reveal weak dependence of trophic position of planktivorous fish on individual size: A consequence of omnivorism and mobility Antonio Bodea,* , Pablo Carreraa,# , Carmela Porteirob a Instituto Español de Oceanografía, Centro Oceanográfico de A Coruña, A Coruña, Spain b Instituto Español de Oceanografía, Centro Oceanográfico de Vigo, Vigo, Spain
Abstract Relationships between trophic position and body size in pelagic organisms were investigated using stable nitrogen isotopes (δ 15 N) in the northwestern shelf of the Iberian Peninsula. Samples included size-fractionated plankton, planktivorous fish, squid and dolphin. Values of δ 15 N increased with the logarithm of individual body mass when considering all organisms, but the slope of the regression line (0.195) was smaller than values reported for benthic communities. In addition, there was a large overlap in the range of δ 15 N values between planktivorous consumers, and consequently trophic position estimates were not related to individual size in most species. This suggests strong competition for planktonic prey, which may be minimized by migration. As an example, large (>20 cm) sardines (Sardina pilchardus) showed significant differences in δ 15 N between shelf zones, while small sardines did not. Such differences support a lower mobility of adults compared to juvenile sardines along the shelf. Keywords: Stable isotopes, 15 N, Plankton, Pelagic fish, Trophic position, Food web
1. Introduction Analyses based on the individual body size of organisms provide useful insights on the structure and the functioning of marine ecosystems because most metabolic processes are scaled to size, which also determines trophic relationships (Kerr and Dickie, 2001). Most primary producers in the sea are unicellular algae and the predators are generally larger than their preys, implying a continuous increase in body size from lower to upper trophic positions in the food web. Several studies demonstrated this increase in plankton (Fry and Quinones, 1994), benthos (France et al., 1998) and fish communities (Jennings et al., 2001, 2002), thus providing a means to characterize the trophic structure of communities and a diagnostic tool for changes in the ecosystems. * Corresponding author. Address: Instituto Español de Oceanografía, A Coruña, Spain; phone (+34) 985 205362; fax: (+34) 985 229077; e-mail:
[email protected] # Present address: Museo do Mar de Galicia, Vigo, Spain.
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08022-8
© 2006 Elsevier Ltd. All rights reserved.
282
A. Bode et al.
The estimation of the trophic position of individual marine organisms is complicated by the generally large variety of consumed prey in most species. For instance, primary consumers as zooplanktonic copepods are not exclusive herbivores and complement their trophic requirements with variable amounts of heterotrophic preys (e.g. Batten et al., 2001). Some fish display ontogenic changes in their diets as their feeding mechanisms develop. In this way, larvae of planktivorous fishes are generally zoophagous (Conway et al., 1994) while adults are able to filter significant amounts of phytoplankton (Varela et al., 1990). Therefore, the determination of the average trophic position of a given species is complicated by the variability of diets in time and space. The analysis of diets from gut contents generally require the examination of a large number of individuals of different sizes collected over the distribution range of the species and including the possible seasonal variability (e.g. Olaso et al., 1999). An alternative way is the determination of the natural abundance of stable isotopes of key elements as the nitrogen (Wada and Hattori, 1991; Vander Zanden and Rasmussen, 2001). Predators are more enriched in heavy isotopes than their preys, and its isotopic composition integrates the different diets over time (Tieszen et al., 1983). However, the comparison of isotopic signatures of species with very different generation times is complicated by the different turnover of the isotopes in each species (O’Reilly et al., 2002; Post, 2002). Ideally, the sampling must cover all possible states of the ecosystem and the whole range of life cycles, but this requirement is not feasible for a large marine ecosystem. Field studies overcome this difficulty by sampling at spatial scales large enough to cover most of the range of variability in the nutrient sources at the base of the food web (Sholto-Douglas et al., 1991; Jennings et al., 2001, 2002), which is the primary determinant of isotopic enrichment at upper trophic levels (Post, 2002). Such strategy is particularly suited to upwelling ecosystems where there are marked variations in the nutrient sources over relatively short spatial and temporal scales, as in the northern shelf of the Iberian Peninsula (Alvarez-Salgado et al., 2002). The objective of this study is to analyze the relationships between trophic position, estimated from the natural abundance of 15 N, and individual size of marine pelagic consumers, with emphasis on planktivorous fish species. The study was conducted in continental shelf waters of the north and northwest of the Iberian Peninsula. This region displays marked oceanographic gradients between the upwelling-influenced western shelf (Galicia) and the southern Bay of Biscay (Mar Cantábrico), the latter characterized by a seasonally stratified water-column caused by warm surface waters and the input of continental water (Fraga, 1981; Botas et al., 1990). Previous studies in this region showed significant relationships between the abundance of carbon and nitrogen isotopes and the size of some pelagic organisms (Bode et al., 2003), the seasonal variability of isotopic signatures in zooplankton (Bode and Alvarez-Ossorio, 2004) and the differential contribution of phyto- and zooplankton to the isotopic composition of sardines (Bode et al., 2004).
2. Materials and methods The natural abundance of 15 N was determined in samples from various components of the pelagic ecosystem in the study region. Plankton samples indicated the isotopic signature at
Dependence of trophic position of planktivorous fish on individual size
283
the base of the food web, while those of plankton consumers (fish and squids) and of piscivorous predators (dolphins) showed the propagation of the isotopic signal up the food web. Samples of plankton, pelagic fish and squids were collected during PELACUS cruises, between March–April, from 1998 to 2003 (Fig. 1). In addition, some samples of sardine (Sardina pilchardus) were obtained from commercial landings in June 2002 to test for seasonal differences in isotopic composition. Fish and squids were sampled by means of a pelagic trawl with vertical aperture of 24 m for 15–30 min of effective sampling. Each specimen was measured (±5 mm), weighed (±0.2 g) and dissected to obtain portions of white muscle, which were stored frozen for isotopic determinations. Plankton samples were collected during the night using a conical net of 20 µm mesh-size from 100 m depth to the surface at stations distributed over the shelf up to the shelf break (Bode et al., 2003, 2004). Samples were subsequently fractionated through sieves of 40, 80, 200, 500, 1000 and 2000 µm and each fraction was carefully washed with filtered seawater, transferred to glass-fibre filters and stored frozen. The plankton retained by the 2000 µm sieve (generally large salps) was not used in this study. In addition, water from the surface was prefiltered through a sieve of 20 µm mesh-size and subsequently filtered through Whatman glass-fibre GF/F filters to characterize seston between ca. 0.7 and 20 µm. These filters were stored frozen and processed as the filters with net plankton samples. Samples of the muscle of common dolphins (Delphinus delphis) were obtained from individuals stranded on the coast (Bode et al., 2003). Natural abundance of 15 N (δ 15 N) in all samples was measured in an isotope-ratio mass spectrometer (Finnigan Matt Delta Plus) coupled to an elemental analyzer (Carlo Erba CHNSO 1108) after oven drying plankton (50◦ C, 24 h) or freeze-drying consumer samples. The determinations were calibrated against atmospheric nitrogen. Precision (±1 se) of
Fig. 1. Map of sampling locations for plankton and planktivorous consumers. Different shelf zones (with ICES fishery code areas between brackets) are indicated.
284
A. Bode et al.
triplicate δ 15 N determinations was better than 0.03h (Bode et al., 2003, 2004; Bode and Alvarez-Ossorio, 2004). The relationships between isotopic composition and the individual size of organisms were studied by mean of linear regression of δ 15 N and log2 weight size-classes (Jennings et al., 2001, 2002). Size-fractionated plankton samples were assigned to individual weight classes determined from allometric equations (Rodríguez and Mullin, 1986; Bode et al., 1998, 2003) by assuming that the individual length was the geometric mean of the mesh-size limits for each fraction. The nominal lower and upper limits of the weight classes considered between plankton and dolphins were 1.86 ng and 131 kg, respectively. Estimates of trophic position using δ 15 N data (TPN ) were made following the procedure of Vander Zanden and Rasmussen (2001): TPN = 2 +
δ 15 Nc − δ 15 Nhz , 3.4
where δ 15 Nc and δ 15 Nhz are mean δ 15 N values of a given consumer and of herbivorous zooplankton, respectively, the latter used as the reference baseline value. The value of δ 15 Nhz was estimated from the measurements in the 200–500 µm fraction of plankton, mostly composed of copepods (Bode and Alvarez-Ossorio, 2004). The mean isotopic fractionation value between adjacent trophic levels was taken as 3.4h and the error of the estimation of TPN was computed from the measured errors in both δ 15 Nc and δ 15 Nhz and by assuming a food web of up to four trophic levels (Vander Zanden and Rasmussen, 2001). In addition, independent estimates of trophic position (TPF ) were obtained from the literature data on gut-contents for individual species as reported in FISHBASE (http://www.fishbase.org).
3. Results 3.1. Trophic position and individual size The samples collected over 6 years included seven plankton size-classes (Fig. 2), seven planktivorous fish species (Table 1), one squid genus (Alloteuthis) and one piscivorous predator (Delphinus delphis). These samples can be considered as representative of the productive phase of the pelagic ecosystem, as phytoplankton blooms (Bode et al., 2002) and the spawning activity of most pelagic fishes (Solá et al., 1992; ICES, 2003) peak during the spring in the study region. This is supported by the finding of maximum enrichment in stable C and N isotopes of zooplankton during the spring in the study area (Bode and Alvarez-Ossorio, 2004). In contrast, no seasonal variability was found in δ 15 N of sardines of similar size (19 cm) collected in the north Galicia area in March and in June 2002 (ANOVA, p > 0.05, n = 18), despite the expected differences in fat content (Oliver, 1951). Thus, in absence of other measurements, we assumed that the isotopic content measured in the muscle of pelagic consumers reflects their diets integrated at time and space scales relevant for the size of the ecosystem studied. When taking into account all the ecosystem components the mean values of δ 15 N were significantly related to the mean individual size (Fig. 2(a)). This result indicates that the trophic
Dependence of trophic position of planktivorous fish on individual size
285
(a)
(b) Fig. 2. Relationship between mean (±se) δ 15 N and individual size for the whole pelagic community (a) or for pelagic consumers (b). Individual size was expressed as log2 (w), being w the mean weight of doubling size classes. The ranges of values for plankton, planktivorous consumers and top predators (dolphins) are indicated by dashed rectangles in panel (a). The dashed line in panel (b) indicates the relationship found for a demersal fish community in the North Sea (Jennings et al., 2001) and the squares indicate size-classes composed entirely by squid or dolphin (see Section 3.1).
position of pelagic species increases with their individual size, as described for benthic ecosystems (e.g. Jennings et al., 2001, 2002). However, the slope of the regression line when planktivorous consumers are considered separately (Fig. 2(b)) is smaller than the slope for benthic consumers of similar size computed by Jennings et al. (2001).
286 Table 1 Mean and se values of individual length (cm), weight (g), natural abundance of 15 N (δ 15 N) and number of samples (n) analyzed for the components of the pelagic ecosystem considered in this study. Estimates of trophic position (TPN = 2 + [δ 15 Nc − δ 15 Nhz ]/3.4) from mean δ 15 N values of a given consumer (δ 15 Nc ) and herbivorous zooplankton (δ 15 Nhz ) are compared with trophic position (TPF ) estimates from dominant food items in the diet as reported in FISHBASE (http://www.fishbase.org). Only length and weight values for planktivorous consumers are given for clarity Length
n
Mean
se
Mean
se
Mean
se
– – – 18.11 25.55 29.40 17.14 15.08 4.58 13.80 24.42 –
– – – 0.15 0.97 1.08 0.17 0.54 0.42 0.32 1.34 –
– – –
– – –
51.64 148.19 217.47 34.42 15.20 2.80 54.65 156.26 –
1.21 12.83 25.96 1.87 1.54 0.67 3.51 25.42 –
5.18 5.62 6.79 10.75 11.34 10.58 10.51 10.73 10.07 10.74 10.35 13.10
0.11 0.15 0.11 0.04 0.06 0.15 0.12 0.10 0.19 0.06 0.09 0.36
TPN Mean
133 141 148 401 44 45 15 10 6 10 20 5
1.0 2.0 2.3 3.5 3.7 3.5 3.4 3.5 3.3 3.5 3.4 4.2
TPF
Dominant food
se
Mean
se
0.0 0.0 0.1 0.1 0.1 0.2 0.2 0.2 0.1 0.3
1.0 2.0 3.0 2.8 3.8 3.7 2.8 3.5 3.2 3.5 3.4 4.2
– – – 0.3 0.4 0.5 0.3 0.3 0.1 0.3 0.1 0.3
Phytoplankton Herbivorous zooplankton Zooplankton, phytoplankton, detritus Zooplankton, zoobenthos, fish, squids Zooplankton, zoobenthos, fish Zooplankton Zoobenthos, zooplankton Zooplankton, zoobenthos Zooplankton Zooplankton, zoobenthos, macrophytes Squids, pelagic and demersal fish
a Values of the smallest size-class considered. b δ 15 N value of the 200–500 µm size-class (mainly copepods). Most species are actually omnivores (e.g. Bode et al., 2003). c δ 15 N value of the 1000–2000 µm size-class (mostly carnivorous crustaceans, medusae, siphonophora, ctenophora and chaetognatha). d Not available in FISHBASE. Estimates of TP and diet are from other inshore squids (Loliginidae). F e Diet data from Santos et al. (2001).
A. Bode et al.
Phytoplanktona Herbivorous zooplanktonb Carnivorous zooplanktonc Sardina pilchardus Trachurus trachurus Scomber scombrus Engraulis encrasicholus Macroramphosus scolopax Alloteuthis spp.d Capros aper Boops boops Delphinus delphise
δ 15 N
Weight
Dependence of trophic position of planktivorous fish on individual size
287
3.2. Trophic position and species composition The mean values of δ 15 N for species of planktivorous consumers were remarkably similar despite large differences in size (Table 1). As a consequence, there is no significant relationship between δ 15 N and individual weight when planktivorous fish species spanning nine weight size-classes are considered (r 2 = 0.267, p > 0.05, n = 9). Furthermore, most species did not show a significant correlation between δ 15 N and individual weight (Fig. 3). The only excep-
(a)
(b) Fig. 3. Variability of mean (±se) δ 15 N (a, b) and individual size [log2 (w)] for the different species of planktivorous consumers. Species were separated in different panels for clarity. The regression lines in panel (a) indicate significant (p < 0.001) relationships for Sardina (continuous line) and Trachurus (dashed line).
288
A. Bode et al.
tions were Sardina pilchardus and Trachurus trachurus, the former showing decreasing δ 15 N values while the latter displayed a significant enrichment with increasing size. The similarity in δ 15 N translates in a narrow range of trophic positions occupied by the studied pelagic fishes, with a mean value of 3.5 trophic levels (Table 1). It must be noted that the average estimates of trophic position of pelagic consumers from δ 15 N (TPN ) are almost coincident with those derived from gut content studies (TPF , r = 0.705, p < 0.05, n = 9). 3.3. Mobility and plankton exploitation One way to avoid or alleviate food competition is the ability to exploit diverse habitats. As an example, we examined the differential mobility of S. pilchardus size-classes along the study area using δ 15 N as a tracer of migrations. This approach requires two main assumptions (e.g. Hanson et al., 1997). First, the feeding areas exploited by the target fish species will have a different isotopic signature at the base of the food web. Second, migrating fish will integrate the isotopic composition of its diet and therefore their values of natural isotopic abundance will be different from those of fish feeding only in the same area. Plankton from non-contiguous shelf zones showed significantly different mean δ 15 N values (ANOVA and Dunnett-C tests, p < 0.05), while those of central zones were equivalent (Fig. 4(a)). These results support the existence of a different isotopic signature at the base of the food webs in S Galicia and E Cantábrico areas, the former likely caused by new nitrogen inputs by the upwelling (Alvarez-Salgado et al., 2002). In contrast, δ 15 N values of small S. pilchardus individuals (<20 cm) in Galicia were similar to those of Mar Cantábrico (Fig. 4(b)), while significant differences appeared when large individuals were compared (ANOVA and Dunnett-C tests, p < 0.05). These differences were not due to differences in size between zones, as the mean length of fish >20 cm collected in Galicia and in the Mar Cantábrico was equivalent (ANOVA, p > 0.05).
4. Discussion The trophic position of marine fishes, estimated from δ 15 N, was reported to vary linearly with the logarithm of individual size in benthic assemblages (Jennings et al., 2001, 2002). However, our results indicate two main differences between benthic and pelagic species. First, the slope of the regression line between δ 15 N and individual size suggests that the average value of the predator/prey body mass ratio in pelagic ecosystems is higher than the value for benthic ecosystems. This ratio can be computed as 2(3.4 slope) , assuming an average trophic enrichment between adjacent trophic levels of 3.4h (Vander Zanden and Rasmussen, 2001), and amounts 1.8 × 105 in our study. In contrast, the predator/prey ratio for benthic communities lies near 102 (Jennings et al., 2001, 2002). The larger difference between the size of predators and preys in pelagic compared to benthic communities was related to the prevalence of filter feeding of plankton in the former, which allows for the concentration of highly diluted food of small size, as occurs with most planktonic preys (Cushing, 1978). In addition, there must be large differences in the predator/prey ratio between the different compartments considered. This is illustrated by the ratios between the nominal sizes of species of known feeding habits as the sardine (S. pilchardus) and the dolphin (D. delphis). The sardine (mean weight 52 g)
Dependence of trophic position of planktivorous fish on individual size
289
(a)
(b) Fig. 4. Mean (±se) δ 15 N of <200 µm plankton (a) and three size-classes of S. pilchardus (b) in different areas of the NW Iberian shelf. Letters indicate significant differences (ANOVA and Dunnett-C tests, p < 0.05).
consumes both phytoplankton and zooplanktonic copepods (e.g. Varela et al., 1990), which correspond to nominal weights in our size-classes of ca. 3 ng and 11 µg and to predator/prey ratios between 2×1010 and 5× 106 , respectively. The common dolphin (mean weight 104 kg) preys on squids (mean weight 2.8 g) and planktonic feeders, as sardines (Santos et al., 2001), with resulting predator/prey values between 4 × 104 and 2 × 102 . Furthermore, it must be taken into account that the predicted predator/prey body size ratios from the slope of the relationship between δ 15 N and individual weight is based on the assumption of a constant isotopic enrichment between trophic levels. Several studies, however, noted a relatively high variability in this enrichment across trophic levels and species (Vander Zanden and Rasmussen, 2001; Post, 2002; McCutchan et al., 2003).
290
A. Bode et al.
Second, the significance of the linear relationship between δ 15 N and individual size was reduced or even disappeared when only certain components, such as planktivorous consumers, were considered. Previous studies reported a moderate increase in δ 15 N with individual size in plankton (Fry and Quinones, 1994; Bode et al., 2003, 2004), which was attributed to the omnivory of most planktonic consumers. In our study, the relationship for pelagic consumers of Fig. 2(b) was significant because of the inclusion of squids and dolphins, both at the limits of the size range studied and with feeding habits that include benthic and pelagic preys (Mangold-Wirz, 1963; Santos et al., 2001). Such a lack of correlation with individual size of planktivorous fishes suggests a strong inter-specific competition for similar food resources, as indicated by the information on diets provided by FISHBASE (Table 1). Nevertheless, actual competition for food could be reduced because of dietary preferences unrelated to size, and because of the differences in mobility among species. Due to the relatively low variability in δ 15 N of plankton with size, even the consumption of preys of different size would produce similar δ 15 N values in the predators. The variability of δ 15 N with size of individual planktivorous fish species found in this study can be interpreted as the result of ontogenic changes in diet. In this way, S. pilchardus and T. trachurus could be considered as representative of two opposite feeding strategies. The decrease in the mean δ 15 N values of S. pilchardus as size increased was already described in this species and attributed to an increase in the filtering efficiency of phytoplankton with fish growth (Bode et al., 2003, 2004). The consumption of preys from a lower trophic position, less enriched in 15 N, would be responsible for a lower enrichment in larger and older individuals. In contrast, T. trachurus must be a consumer of animal preys all its life to display a consistent increase in δ 15 N with growth, as reported for demersal and benthic fishes and invertebrates (Lindsay et al., 1998; Jennings et al., 2002). Furthermore, the increase in the size range of preys consumed with growth of T. trachurus contributes to a high isotopic enrichment, as other fish and squids are reported preys of this species (Olaso et al., 1999). Between these extremes, other species showed no distinct pattern in δ 15 N with the growth in size because of a large plasticity in diets, which allows feeding on planktonic, nektonic and benthic preys (Table 1). The good agreement between trophic position estimations of pelagic fish from δ 15 N and independent estimates from gut-content analysis found in this study supports the use of an average trophic enrichment of 3.4h between trophic levels in the estimations of trophic position of consumers (Vander Zanden and Rasmussen, 2001). In turn, the validity of the average trophic enrichment value used further supports our estimations of mean predator/prey body mass ratios derived from the relationship between δ 15 N and body size (Fig. 2). The correspondence in both mean and se trophic position values is complete for some species, as D. delphis and Boops boops, while TPN estimates for the size-class containing carnivorous zooplankton are lower, and those for sardine higher, than those of TPF . As noted above, there are increasing evidences of a lower isotopic enrichment across plankton size-classes compared to the average value of 3.4h employed, thus confirming that the estimation of the trophic position of primary consumers cannot be made using a constant enrichment for the whole food web (Vander Zanden and Rasmussen, 2001). In the case of S. pilchardus, the average TPF value of 2.8 is the consequence of the frequent record of abundant phytoplankton cells in their gut (e.g. Varela et al., 1988). However, when estimated by weight or biovolume, the contribution of phytoplankton to the diet of S. pilchardus is largely reduced (e.g. Varela et al., 1990). Further-
Dependence of trophic position of planktivorous fish on individual size
291
more, isotopic studies showed that phytoplankton nitrogen is almost absent from the muscle of small sardines but increase in large individuals (Bode et al., 2004), thus producing TPN estimates equivalent to those of other planktivorous fishes (Table 1). As demonstrated for other fish species, it is likely that phytoplankton contributes to meet most of the carbon demands in the consumer, while protein-nitrogen is obtained mainly from zooplankton (Van der Lingen, 1998). These differences must be taken into account in future studies aiming to estimate trophic position in consumers of both phyto- and zooplanktonic preys. Our study also showed spatial differences in δ 15 N of large S. pilchardus similar to those of small plankton, supporting the existence of locally restricted populations. These results suggest that these fish have been feeding for relatively long periods in the same zones to acquire the isotopic signature of plankton. There is no information available on the turnover time of isotopes in the muscle of this species, but data for other fish species indicates that it may be in the order of several weeks (Tieszen et al., 1983). Therefore, our data support the hypothesis of a restricted mobility of large individuals of S. pilchardus between S Galicia and the other zones. In contrast, small and medium-sized individuals have similar average δ 15 N values in all zones, which can be explained either as the consequence of a larger mobility of small compared to large fish or as the preferential feeding on zooplankton by small fish in all zones. In both cases, the result would be the reduction in the competition between juveniles and adults. Our results agree with the expected migration of juvenile S. pilchardus individuals off the spawning areas in S Galicia and Mar Cantábrico areas, while large adults remain in these areas (Carrera and Porteiro, 2003). 5. Conclusions This study showed a linear relationship between the trophic position of pelagic consumers and the logarithm of individual biomass with a slope of 0.195, which is smaller that the values reported for benthic consumers. This difference suggests that marine pelagic food webs are less structured than benthic food webs, as indicated by predator/prey body-size ratios between 2 × 102 and 2 × 1010 . In addition, planktivorous fishes display variability in trophic position with size that depend on species composition. This variability suggests a strong competition for similar preys. Planktivorous fishes may reduce competition for food by exploiting a large variety of preys and habitats, as exemplified by migrant juvenile sardines having similar δ 15 N values in all shelf areas. In contrast, non-migrant adult sardines displayed larger δ 15 N values in upwelling compared to non-upwelling areas. Acknowledgements We are indebted to the captain and crew of R/V Thalassa for their collaboration during PELACUS cruises. The collaboration of all cruise participants, but especially of P. Iglesias, B. Castro, U. Autón and J. Valencia, made possible the collection and preparation of a large number of plankton and fish samples. S. Lens provided dolphin samples and M. Lema made the isotopic determinations at the SXAI (Universidade da Coruña, Spain). This research was supported in part by projects PELASSES (99/10), and SARDYN (QLRT-2001-00818) of the European Union, and is a contribution to the GLOBEC-Spain Programme.
292
A. Bode et al.
References Alvarez-Salgado, X.A., Beloso, S., Joint, I., Nogueira, E., Chou, L., Pérez, F.F., Groom, S., Cabanas, J.M., Rees, A.P., Elskens, M. (2002). New production of the NW Iberian shelf during the upwelling season over the period 1982–1999. Deep Sea Research 49, 1725–1739. Batten, S., Fileman, E.S., Halvorsen, E. (2001). The contribution of microzooplankton to the mesozooplankton diet in an upwelling filament off the north west coast of Spain. Progress in Oceanography 51, 385–398. Bode, A., Alvarez-Ossorio, M.T. (2004). Taxonomic versus trophic structure of mesozooplankton: A seasonal study of species succession and stable carbon and nitrogen isotopes in a coastal upwelling ecosystem. ICES Journal of Marine Science 61, 563–571. Bode, A., Alvarez-Ossorio, M.T., González, N. (1998). Estimations of mesozooplankton biomass in a coastal upwelling area off NW Spain. Journal of Plankton Research 20, 1005–1014. Bode, A., Varela, M., Casas, B., González , N. (2002). Intrusions of eastern North Atlantic central waters and phytoplankton in the north and northwestern Iberian shelf during spring. Journal of Marine Systems 36, 197–218. Bode, A., Carrera, P., Lens, S. (2003). The pelagic foodweb in the upwelling ecosystem of Galicia (NW Spain) during spring: Natural abundance of stable carbon and nitrogen isotopes. ICES Journal of Marine Science 60, 11–22. Bode, A., Alvarez-Ossorio, M.T., Carrera, P., Lorenzo, J. (2004). Reconstruction of trophic pathways between plankton and the North Iberian sardine (Sardina pilchardus) using stable isotopes. Scientia Marina 68, 165–178. Botas, J.A., Fernández, E., Bode, A., Anadón, R. (1990). A persistent upwelling off the Central Cantabrian coast (Bay of Biscay). Estuarine Coastal and Shelf Science 30, 185–199. Carrera, P., Porteiro, C. (2003). Stock dynamic of the Iberian sardine (Sardina pilchardus, W.) and its implication on the fishery off Galicia (NW Spain). Scientia Marina 67 (Suppl. 1), 245–258. Conway, D.V.P., Coombs, S.H., Fernández de Puelles, M.L., Tranter, P.R.G. (1994). Feeding of larval sardine, Sardina pilchardus (Walbaum), off the north coast of Spain. Boletin Instituto Espanol de Oceanografia 10, 165–175. Cushing, D.H. (1978). Upper trophic levels in upwelling areas. In: Boje, R., Tomczak, M. (Eds.), Upwelling Ecosystems. Springer-Verlag, Berlin, pp. 101–110. Fraga, F. (1981). Upwelling off the Galician coast, Northwest Spain. In: Richards, F.A. (Ed.), Upwelling Ecosystems. American Geophysical Union, Washington, DC, pp. 176–182. France, R., Chandler, M., Peters, R. (1998). Mapping trophic continua of benthic foodwebs: Body size-δ 15 N relationships. Marine Ecology – Progress Series 174, 301–306. Fry, B., Quinones, R.B. (1994). Biomass spectra and stable isotope indicators of trophic level in zooplankton of the northwest Atlantic. Marine Ecology – Progress Series 112, 201–204. Hanson, S., Hobbie, J.E., Elmgren, R., Larsson, U., Fry, B., Johansson, S. (1997). The stable nitrogen isotope ratio as a marker of food-web interactions and fish migration. Ecology 78, 2249–2257. ICES (2003). Report of the Working Group on the assessment of mackerel, horse mackerel, sardine and anchovy. ICES CM 2003/ACFM:08. ICES, Copenhagen. Jennings, S., Pinnegar, J.K., Polunin, N.V.C., Boon, T.W. (2001). Weak-cross species relationships between body size and trophic level belie powerful size-based trophic structuring in fish communities. Journal of Animal Ecology 70, 934–944. Jennings, S., Pinnegar, J.K., Polunin, N.V.C., Warr, K.J. (2002). Linking size-based and trophic analyses of benthic community structure. Marine Ecology – Progress Series 226, 77–85. Kerr, S.R., Dickie, L.M. (2001). The Biomass Spectrum. Columbia University Press, New York. Lindsay, D.J., Minagawa, M., Mitani, I., Kawaguchi, K. (1998). Trophic shift in the Japanese anchovy Engraulis japonicus in its early life history stages as detected by stable isotope ratios in Sagami Bay, Central Japan. Fisheries Science 64, 403–410. Mangold-Wirz, K. (1963). Biologie des Céphalopodes benthiques et nectoniques de la Mer Catalane. Vie Milieu 13 (suppl.), 1–285. McCutchan, J.H., Lewis, W.M.J., Kendall, C., McGrath, C.C. (2003). Variation in trophic shift for stable isotope ratios of carbon, nitrogen, and sulfur. Oikos 102, 378–390. Olaso, I., Cendrero, O., Abaunza, P. (1999). The diet of the horse mackerel, Trachurus trachurus (Linnaeus 1758), in the Cantabrian Sea (north of Spain). Journal of Applied Ichthyology 15, 193–198. Oliver, M. (1951). La sardina de la costa noroeste española en 1948 y 1949 (Estudio biométrico y biológico). Boletin Instituto Espanol de Oceanografia 42, 1–22.
Dependence of trophic position of planktivorous fish on individual size
293
O’Reilly, C.M., Hecky, R.E., Cohen, A.S., Plisnier, P.-D. (2002). Interpreting stable isotopes in food webs: Recognizing the role of time averaging at different trophic levels. Limnology and Oceanography 47, 306–309. Post, D.M. (2002). Using stable isotopes to estimate trophic position: Models, methods, and assumptions. Ecology 83, 703–718. Rodríguez, J., Mullin, M.M. (1986). Relation between biomass and body weight of plankton in a steady state oceanic ecosystem. Limnology and Oceanography 31, 361–370. Santos, M.B., Clarke, M.R., Pierce, G.J. (2001). Assessing the importance of cephalopods in the diets of marine mammals and other top predators: Problems and solutions. Fisheries Research 52, 121–139. Sholto-Douglas, A.D., Field, J.G., James, A.G., van der Merwe, N.J. (1991). 13 C/12 C and 15 N/14 N isotope ratios in the Southern Benguela Ecosystem: Indicators of food-web relationships among different size-classes of plankton and pelagic fish; differences between fish muscle and bone collagen tissues. Marine Ecology – Progress Series 78, 23–31. Solá, A., Franco, C., Lago de Lanzós, A., Motos, L. (1992). Temporal evolution of Sardina pilchardus (Walb.) spawning on the N-NW coast of the Iberian Peninsula. Boletin Instituto Espanol de Oceanografia 8, 97–114. Tieszen, L.L., Bouton, T.W., Tesdahl, K.G., Slade, N.A. (1983). Fractionation and turnover of stable carbon isotopes in animal tissues: Implications for δ 13 C analysis of diet. Oecologia 57, 32–37. Van der Lingen, C.D. (1998). Nitrogen excretion and absorption efficiencies of sardine Sardinops sagax fed phytoplankton and zooplankton diets. Marine Ecology – Progress Series 175, 67–76. Vander Zanden, M.J., Rasmussen, J.B. (2001). Variation in δ 15 N and δ 13 C trophic fractionation: Implications for aquatic food web studies. Limnology and Oceanography 46, 2061–2066. Varela, M., Larrañaga, A., Costas, E., Rodriguez, B. (1988). Contenido estomacal de la sardina (Sardina pilchardus Walbaum) durante la campaña Saracus 871 en las plataformas Cantábrica y de Galicia en febrero de 1971. Boletin Instituto Espanol de Oceanografia 5, 17–28. Varela, M., Alvarez-Ossorio, M.T., Valdés, L. (1990). Método para el estudio cuantitativo del contenido estomacal de la sardina. Resultados preliminares. Boletin Instituto Espanol de Oceanografia 6, 117–126. Wada, E., Hattori, A. (1991). Nitrogen in the Sea: Forms, Abundances, and Rate Processes. CRC Press, Boca Raton, FL.
This page intentionally left blank
6. Isotope hydrology
This page intentionally left blank
297
Radiocarbon loss from DIC in vadose water flow above the Judea Aquifer, Israel Israel Carmia,* , Mariana Stillera , Joel Kronfelda , Yoseph Yechielib , Miriam Bar-Matthewsb , Avner Ayalonb , Elisabetta Boarettoc a Tel-Aviv University, 69978 Tel-Aviv, Israel b Geological Survey of Israel, 30 Malkhei Israel Street, 95501 Jerusalem, Israel c Weizmann Institute of Science, 76100 Rehovot, Israel
Abstract The roof over the Soreq Cave in Israel was used to model the rate of transport in the vadose zone of the Mountain Aquifer of Israel. This aquifer is of great importance because it supplies 30% of the water consumption for the country. 14 C was measured in drip water from stalactites and in pool water in the cave, under varying thickness of cave roof, which represent different depths in the vadose zone. The flow rate of water through the roof has been determined with tritium analysis to be ∼1 m yr−1 . From this flow rate the corresponding atmospheric 14 C activity at the time of deposition of rain on top of the vadose zone could be determined. The ratio of the measured 14 C activity in drip and pool water to that of the atmosphere was then calculated. From this, the initial value of 14 C activity of water at the entrance to the aquifer was found to be 0.63 ± 0.03 relative to the atmosphere. A model for the chemical interaction between the DIC and the host rock was developed. The average rate of the reaction was found to be 3.5 ± 0.3% yr−1 . Keywords: Radiocarbon, Tritium, DIC, Soreq Cave, Groundwater, Stalactite
1. Introduction 1.1. The problem with radiocarbon dating of water A major problem in age dating of groundwater with inorganic 14 C is to determine the change in its activity in the dissolved inorganic carbon (DIC) of water percolating through the vadose zone down to the entrance to the aquifer. Many models have used δ 13 C information in the vadose zone to infer the initial value of 14 C activity at the entrance into an aquifer, and many of them are incorporated into the NETPATH program (Plummer et al., 1991). Recently, this approach has failed in two cases. In the first case Buckau et al. (2000) found in the Gorleben aquifer a contradiction between the tritium ages of fresh groundwater (<40 yr) and the 14 C * Corresponding author. Address: Tel-Aviv University, 69978 Tel-Aviv, Israel; fax: (+973) 36409282; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08023-X
© 2006 Elsevier Ltd. All rights reserved.
298
I. Carmi et al.
ages calculated with the prevailing methods (4000–8000 yr). In the second case Bouhlassa and Aiachi (2002) measured a 14 C value of 76 pMC (percent of modern carbon) in a fresh water aquifer in Morocco. Their simulation, using different models gave results for the initial value in the aquifer, varying between 22 and 85 pMC. An alternate approach is to study the processes in the vadose zone to determine from 14 C data in it the value of 14 C activity of water at the entrance to the aquifer. This is now possible by using Accelerator Mass Spectrometry (AMS) for the 14 C measurement, which requires only milligrams of carbon for the assay. In a previous paper we estimated the value at the entrance to the Judea Aquifer of Israel (Carmi et al., 2004). In this paper we present the results of a model for the rate of reaction between the DIC in the water and the host rock. We also recalculate the value of 14 C activity of water at the entrance to the aquifer, using a more robust data set. 1.2. Description of the research site The Judea Group aquifer of Israel is an important source of water in Israel, supplying about 30% of the annual water consumption of the country. The aquifer, built of carbonate rock of the Upper Cretaceous age, is several hundred meters thick. The anticlinal dip of the Judea Mountains is towards the Rift Valley to the east and towards the Shephela and the Mediterranean Sea to the west, where the Cretaceous rocks are overlain by younger formations. The karstic Soreq Cave is located in the vadose zone above the Judea Aquifer, within the steep western flank of the Judea Mountains crest at 40 km east of the Mediterranean Sea and at 400 m above sea level (Fig. 1(a)). The flow direction in this part of the aquifer is generally westward. The cave is developed within a 50–60 m thick dolomitic sequence of the Cenomenian Weradim Formation, which is part of the Upper Cretaceous Judea Group. The sequence is fractured and contains many karstic features. The overlying soil cover is irregular and usually occurs in patches of up to 40 cm thickness. The thickness of the roof above the cave varies between 10 m in the northwestern end of the cave and 40–50 m in the southeastern end. The cave ceiling is crossed with two main fracture systems in the west–east and northwest–southeast directions (Fig. 1(b)). Inside the cave there are various types of speleothems, straw and conical stalactites, stalagmites, curtains, cave corals, flowstones, etc. The cave is active and dripping of water through the tips of stalactites, through fractures and along the walls, is common and occur throughout the year, but are more intensive during winter months. The drippings accumulate in pools of varying sizes, from a few to about 500 liters.
2. Methods Water from stalactite drippings and from pools in the Soreq Cave was sampled in late winters 2002 and 2003. The sampling sites are shown in Fig. 1(b). Soil samples were collected in fall 2002 above the cave from depths of 0–15 cm and 16–30 cm. The samples were processed in the Radiocarbon laboratory of the Kimmel Center at the Weizmann Institute of Science and measured in the NSF AMS facility in Tucson, Arizona.
Radiocarbon loss from DIC in vadose water flow above the Judea Aquifer
(a)
299
(b)
Fig. 1. (a) The Soreq Cave in Israel. (b) The fracture system in the cave and the water-collection sources.
CO2 was extracted from the DIC of the water with phosphoric acid under vacuum. The soil (solid) samples were pre-treated by the acid–alkali–acid method (AAA). The alkali soluble fraction was also treated by acid to precipitate the humic fraction. The solid fractions (AAA and humic) were then oxidized to CO2 at 900◦ C. CO2 was converted to graphite in the lab and then sent to Tucson for 14 C measurements by AMS. δ 13 C was measured in the Department of Environmental Sciences and Energy Research at the Weizmann Institute of Science.
3. Results and discussion 14 C
activity and δ 13 C values for soil samples collected at two depths above the Soreq Cave are given in Table 1. Measurements of drip water and pool water samples, collected in 2002 and 2003, presented in Table 2 include DIC, 14 C activity and δ 13 C values.
300
I. Carmi et al. Table 1 14 C activity and δ 13 C values of soil above the Soreq Cave sampled in September 2002.
AAA is acid–alkali–acid pretreatment of sample Sample RTT
Depth below surface (cm)
Type of sample
4506 4507 4508 4509
0–15 0–15 16–30 16–30
After AAA Humic acid After AAA Humic acid
14 C activity
(pMC)
δ 13 C (h)
102.5 ± 0.7 101.9 ± 0.5 101.2 ± 0.5 97.8 ± 0.5
−23.8 −24.3 −23.5 −23.9
3.1. The difference between drip water and pool water Bar-Matthews et al. (1996) have described a large variation in the δ 13 C–DIC relation for drip waters, pool waters and stalactites-film water in the Soreq Cave. Figure 2(a) shows the δ 13 C–DIC relationship in samples from pool waters sampled in 2002 and 2003. The same parameters for stalactite drippings are presented in Fig. 2(b). In the drip water there is some correlation between the DIC and δ 13 C values (r 2 = 0.68). The largest DIC of 9.3 mmole C L−1 has δ 13 C = −13.5h and the lowest DIC of 3.3 mmole C L−1 has a δ 13 C value of −9.5h. In contrast, the pool waters are highly variable. There is no correlation between DIC and δ 13 C (r 2 = 0.03) and the DIC reaches only 6.4 mmole C L−1 . According to Bar-Matthews et al. (1996) the reason for the lower DIC and heavier δ 13 C in pools is that pool water loses light 12 CO2 to the cave atmosphere. In this work δ 13 C = −13.3 ± 0.4h was observed for more than 50% of the drip water, and δ 13 C = −9.5 ± 0.5h was observed for 75% of the pool waters. The exceptions are pool waters with very fast turnover times, for which the δ 13 C is the same as in the drip waters, and very slow drips, which exhale light CO2 to the atmosphere of the cave. The δ 13 C of the drip waters is similar to that of the Mountain Aquifer (Kroitoru, 1987). For our model, we normalized the measured 14 C to a uniform δ 13 C value of the drips, by modifying the 14 C activity by twice the difference in δ 13 C between the drip waters and pool waters (Table 2). 3.2. Determination of time of percolation (θ ), initial 14 C activity (Aθ ) and the Q values Kaufman et al. (2003) using tritium as a tracer in a linear flow model estimated the transit time of water in the vadose zone above the Soreq Cave at not more than 36 years. The average rate of water flow through the vadose zone was 1.07 ± 0.14 m yr−1 . A value of 1 m yr−1 is used in this work. This rate of flow places the date of deposition of rain above the cave, later sampled as drippings or pool waters, at the time of the post-thermonuclear fast decline of 14 C in the atmosphere, during the four last decades of the 20th century. Under thinner roof cover in the cave, water in drippings and pools take less time to arrive, than under a thicker roof cover. Yet, the decline in atmospheric 14 C activity, i.e. lower activities at shallower roof thickness is not observed in the 14 C data of drips and pools (Table 2). There is also no clear trend in the δ 13 C-depth relation (Table 2). We therefore adopt another approach for interpretation of the data.
Source
1-3 12-6 12-1 13-1 1-1 1-8 2-2 2-1 11-2 5-7 8-4 8-5 6-3
Roof thickness (m)
Type
12 15 15 20 20 20 30 30 30 40 40 40 50
D D P P D D D P D P P D P
2002 Sample (RTT)
2003 DIC (mmole L−1 )
δ 13 C (h)
Aθ (pMC)
Qθ
4326 4322 4323 4327
3.35 2.32 3.39 4.39
−9.5 −9.4 −9.6 −10.4
100.9 ± 0.6 98.2 ± 0.5 97.8 ± 0.7 82.6 ± 0.5
0.841 0.848 0.769 0.650
4321
5.56
−13.5
107.9 ± 0.6
0.643
4324
6.28
−9.5
92.0 ± 0.5
Sample (RTT)
DIC (mmole L−1 )
δ 13 C (h)
Aθ (pMC)
Qθ
4651 4619 4655 4620
4.74 6.55 3.35
−12.9 −13.6 −13.9 −10.2
101.1 ± 0.7 109.6 ± 0.6 106.7 ± 0.6 102.3 ± 0.6
0.873 0.915 0.892 0.884
4652 4653 4621 4654 4622 4623 4650
7.85 8.65 6.21 9.31 5.28 4.05 4.46
−12.8 −13.8 −10.7 −13.4 −13.6 −10.2 −10.6
103.0 ± 0.7 107.3 ± 0.7 105.5 ± 0.5 110.7 ± 0.6 112.2 ± 0.6 105.4 ± 0.6 100.3 ± 0.5
0.810 0.664 0.665 0.756 0.641 0.651 0.635
Radiocarbon loss from DIC in vadose water flow above the Judea Aquifer
Table 2 DIC, 14 C activity and δ 13 C values in drip (D) and pool (P) water collected in Soreq Cave in 2002 and 2003. The Qθ value is explained in the text. The standard deviation of Qθ is 1/100 that of the Aθ
301
302
I. Carmi et al.
Fig. 2. DIC and δ 13 C for pool water (a) and for drip water (b) sampled in 2002 and 2003.
Wigley (1975) had defined a parameter Q for an aquifer, which is the ratio of 14 C activity in fresh aquifer water (Aq ) to that of the atmosphere at the time of deposition on the ground (Aa0 ), Q = Aq /Aa0 . This definition regards the vadose zone as a black box and implies that the rate of chemical processes in the vadose zone (e.g. carbon isotope exchange between DIC, CO2 and CaCO3 ) do not change over time. The definition of Q is extended in this work to the whole of the vadose zone. The time of percolation in the vadose zone is denoted by θ and the 14 C activity at time θ is Aθ . Wherever available, we have used the θ derived from tritium measurements (Kaufman et al., 2003). When this was not possible we estimated θ from the average rate of percolation (1 m yr−1 ) and from the thickness of the roof of the cave above the source. The new definition is thus Qθ = Aθ /Aaθ . This parameter is given in Table 2 too. For every Aθ , Aaθ at θ years ago was selected from the detailed atmospheric data of Levin et al. (1985) and Levin and Kromer (1997). The late winter and spring data were used, as this is the period of intense rains over the cave and of the flourishing of the seasonal plants and pine trees above the ground. These plants provide CO2 through their roots and by their decay, to the percolating water. The atmospheric data of Levin et al. (1985) and Levin and Kromer (1997) is corrected for δ 13 C values and therefore the Aθ values are likewise compensated. The standard deviation (σ ) cited by Levin et al. (1985) and Levin and Kromer (1997) is small and practically does not influence the standard deviation of Aθ in cave waters. Qθ is a fraction and therefore its standard deviation is σQθ = σpMC /100. 3.3. Evaluation of the data The δ 13 C values of soil above the Soreq Cave (Table 1) show clearly that the source of organic carbon is C3 plants. 14 C activity was measured in two organic fractions of the soil: in soil treated by the AAA method and in the humic acid fraction. There is practically no difference
Radiocarbon loss from DIC in vadose water flow above the Judea Aquifer
303
between the 14 C content of the upper (0–15 cm) and lower (16–30 cm) layers of soil, which is reasonable considering the thinness of the soil and the depth to which the plants extend their roots. The average value is about 101 pMC. This value is lower than the atmospheric values of September 2002, which was 107.5 pMC (Levin and Kromer, 2004). Our results seem to be reasonable because the plants photosynthesize atmospheric CO2 but also pump up water that had already interacted with the carbonate of the soil. In Table 2 we show the DIC, 14 C activity and δ 13 C values in drip (D) and pool (P) water collected in Soreq Cave in 2002–2003. Apparently, no clear gradients of 14 C and δ 13 C in the roof of the cave are observed. We compared the averages of 3 pools on 2002 to the same pools on 2003: 102.3 ± 5.0 pMC and 107.1 ± 5.0 pMC, respectively. The difference between the two years is not very significant. However, the δ 13 C values are generally heavier for the 2002 water samples than those for the 2003 samples. The reason for this is that 2003 was a rainy year and fresh, less evaporated water was collected in the pools. The data for the dripping water source 1-8 (Table 2) is not used in the analysis because these samples are from an active fracture in the ceiling of the cave. The source is continually shifting and it has changed during these two years not only in DIC and isotopic composition, but also in its chemical composition. We also discount source 6-3 (Table 2) which is under the thickest roof of the cave. The reason is that this pool is in the deepest part of the cave and may have collected overflows from other pools and is therefore not a unique, well-defined source of water. 3.4. Qθ in the vadose zone and Qaqi value of the Judea Aquifer Using the parameter Qθ it was possible to observe its dependence on depth, which we could not observe with the Aθ parameter (Table 2). The parameter Qθ seems also to be insensitive to the change in Aθ from 2002 to 2003. Figure 3 shows the decrease of Qθ value with depth of the cave roof. At 40 m from the surface the data converges to Qθ = 0.63 ± 0.03, which is close to: (i) the Qaqi value reported previously for the Soreq Cave (Carmi et al., 2004)
Fig. 3. Change of Qθ value with roof thickness (depth) in the Soreq Cave. A value of Qθ = 1 at zero depth is assumed, two data points from 30 m below surface coincide and almost cannot be distinguished and two data points are not used in the figure and in the analysis (Section 3.3).
304
I. Carmi et al.
Qaqi = 0.60 ± 0.04, and (ii) the value reported for the Mountain Aquifer of Israel (Kroitoru, 1987) Qaqi = 0.62 ± 0.03. It is also not significantly different from the value reported for an aquifer in Jordan (Bajjali et al. (1997) cited by Clark and Fritz (1997), Qaqi = 0.59). We therefore conclude that Qθ at 40 m represents the initial 14 C activity of water entering the Mountain Aquifer, i.e. Q40 = Qaqi . At the depth of 30 m, two results (sources 2-2 and 2-1) are quite close to Qθ at 40 m and the third one (source 11-2) is higher, but the three results show that at 30 m the Qaqi value has not been attained. 3.5. A model for the reaction between DIC and the host rock We use a first-order reaction model to describe the interaction between the DIC and the host rock. Qθ decreases with time (i.e. with depth) with a rate constant ρ, and dQθ = −ρ(Qθ − Qeq )e−ρθ , dθ the rate of decrease of Qθ is proportional to the difference Qθ − Qeq . The boundary values are: At θ = 0, Qθ = Q0 = 1, and at θ = θeq , Qθ = Qeq . In the Soreq Cave we have found the initial value of the aquifer to be 0.63. The equilibrium value of Qθ , when the water reaches complete equilibrium with the host rock is assumed to be Qeq = 0.5. This value is not reached in the Mountain Aquifer but we use it in the model. The solution, given in (Fig. 4) in natural logarithm notation, is in a linear form. The slope, ρ, is the rate constant of the reaction (Guggenheim, 1926). We thus find that the rate constant for
Fig. 4. A Guggenheim plot of the data in the Soreq Cave (r 2 = 0.91). The rate constant in the vadose zone is ρ = 3.5% yr−1 (or 3.5% m−1 ).
Radiocarbon loss from DIC in vadose water flow above the Judea Aquifer
305
Fig. 5. Extrapolation of the model to 160 m shows that the decrease of Qθ with depth fits an exponential curve.
the decrease of Qθ in the vadose zone is ln(Qθ − Qeq ) = ln(Q0 − Qeq ) − ρθ and ρ = 3.5 ± 0.3% yr−1 . Because the flow rate is about 1 m yr−1 , it can also be said that Qθ decreases by 3.5 ± 0.3% m−1 . 3.6. The exponential decrease of Qθ with depth Our first-order reaction model assumes an exponential decrease of Qθ with depth and therefore we have to test whether our data are not in conflict with this assumption. In Fig. 5 we show an extrapolation of our exponential model to greater depths and it is seen that the available data do fit the model quite well. It is seen also that the equilibrium value Qeq = 0.5 is reached after a flow of about 160 m of carbonate rocks. This value has not been observed in the Mountain Aquifer. It thus seems that in the vadose zone of the Mountain Aquifer of Israel interaction of the DIC with the carbonate rock ends after ∼40–50 m.
4. Conclusions The water sources of the Soreq Cave proved to be a suitable model for the DIC–host carbonate rock interactions in the vadose zone above the Judea Aquifer of Israel. We were able to use 14 C activity data of water from stalactite drippings and pools to estimate: (i) the initial 14 C activity of water entering the aquifer, Qaqi = 0.63 ± 0.03; (ii) the reaction between the DIC of the water and the carbonate rock, which changes the 14 C activity of the DIC from the atmospheric 14 C value to the initial 14 C value at the entrance to the aquifer proceeds at a rate of ρ = 3.5 ± 0.3% yr−1 .
306
I. Carmi et al.
Acknowledgements We thank the Nature Reserve Authority of Israel for permission to work in the Soreq Cave. We also thank Vladimir Lyakonsky for help with the model, and Genia Mintz for help in the lab work.
References Bajjali, W., Clark, I.D., Fritz, P. (1997). The artesian groundwaters of northern Jordan: Insights into their recharge history and age. Journal of Hydrology 192, 355–382. Bar-Matthews, M., Ayalon, A., Matthews, A., Sass, E., Halicz, L. (1996). Carbon and oxygen isotope study of the active water–carbonate system in a karstic Mediterranean cave: Implications for paleoclimate research in semiarid regions. Geochomica et Cosmochimica Acta 60, 337–347. Bouhlassa, S., Aiachi, A. (2002). Groundwater dating with radiocarbon: application to an aquifer under semi arid conditions in the south of Morocco (Guelmime). Applied Radiation and Isotopes 56, 637–647. Buckau, G., Artinger, R., Geyer, S., Wolf, M., Fritz, P., Kim, J.L. (2000). 14 C dating of Gorleben groundwater. Applied Geochemistry 15, 583–597. Carmi, I., Kronfeld, J., Yechieli, Y., Boaretto, E., Bar-Matthews, M., Ayalon, A. (2004). A direct estimate of the initial concentration of 14 C in the Mountain Aquifer of Israel. Radiocarbon 46, 497–500. Proceedings of the 18th International Radiocarbon conference, N. Beavan and R.J. Sparks (Eds.). Clark, I.D., Fritz, P. (1997). Environmental Isotopes in Hydrology. Lewis Publishers, New York, 327 pp.1 Guggenheim, E.A. (1926). Philosophical Magazine 2, 538. Kaufman, A., Bar-Matthews, M., Ayalon, A., Carmi, I. (2003). The vadose flow above Soreq Cave, Israel: A tritium study of the cave waters. Journal of Hydrology 273, 155–163. Kroitoru, L. (1987). The characterization of flow systems in carbonate rocks defined by the groundwater parameters: Central Israel. PhD thesis. The Feinberg Graduate School of the Weizmann Institute of Science, Rehovot, Israel. Levin, I., Kromer, B. (1997). Twenty years of atmospheric 14 CO2 observations at Schauinsland station, Germany. Radiocarbon 39, 205–218. Levin, I., Kromer, B. (2004). The tropospheric 14 CO2 levels in mid-latitudes of the Northern Hemisphere. Radiocarbon 46, 1261–1272. Levin, I., Kromer, B., Schoch-Fischer, H., Bruns, M., Münnich, M., Berdau, D., Vogel, J.C., Münnich, K.O. (1985). 25 Years of tropospheric 14 C observations in Central Europe. Radiocarbon 27, 1–19. Plummer, L.N., Prestemon, E.C., Parkhurst, D.L. (1991). NETPATH – US Geological Survey Water Resources Investigations Report 91-4078. Reston, VA, USA. Wigley, T.M.I. (1975). 14 C dating of groundwater from closed and open systems. Water Resources Research 11, 324–328.
1 The authors cite the Bajjali paper as volume 187 of the Journal of Hydrology. The correct volume number is 192.
307
Stable water isotopes as tools for basin-scale water cycle: Diagnosis of the Murray–Darling A. Henderson-Sellersa , P. Aireya , K. McGuffieb,* , D.J.M. Stonea a Australian Science and Technology Organisation, Lucas Heights, Australia b University of Technology, Sydney, Australia
Abstract We examine the hypothesis that isotopic techniques are applicable to hydrological predictions in difficult-to-simulate semi-arid basins, using the Murray–Darling Basin as an example. Isotopic data from three aquifers in the Murray– Darling characterize precipitation intensity for evaluation of GCMs. Applying these to ‘good’ (water conserving) and ‘poor’ (non-water-conserving) climate model simulations of the Murray–Darling gives rise to large differences in rainfall amount (30–62%). Selecting only ‘good’ models shows a greater than 150 mm annual groundwater recharge loss in El Niño cf. La Niña climates. 2002–2003 El Niño drought data are used to refine isotopic calculation of water lost in evaporation from rivers and irrigation, giving a cumulative loss of 64% of river water during 2002 (cf. 80% using a previous method). This substantiates recent identification of this El Niño drought as evaporatively most extreme and we conclude that stable water isotopes, used synergistically with hydro-climate models, have great potential in future water resource predictions. Keywords: Water resources, Stable water isotopes, River models, Groundwater recharge, GCMs
1. Introduction: Isotopes in basin-scale hydrology 1.1. Stable water isotopes The stable water isotopes 1 H2 H16 O and 1 H1 H18 O exhibit systematic variations in the global water cycle (e.g. Dansgaard, 1964). Isotopic variability in precipitation is related mainly to the sources of moisture and their evolution, the cloud base temperature and precipitation amount. The isotopic composition of groundwater reflects that of the rainfall leading to effective recharge and hence varies with the geographical location of the recharge area and the climate prevailing at the time (e.g. Gat, 1996; Macumber, 2003). The stable isotope composition of surface water reflects the integrated effects of precipitation, modified by the influx of groundwater and evaporative enrichment of heavy isotopes * Corresponding author. Address: Department of Applied Physics, University of Technology, Sydney, P.O. Box 123, Broadway, NSW 2007, Australia; phone: (+61) 2 9514 2072; fax: (+61) 2 9514 2219; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08024-1
© 2006 Elsevier Ltd. All rights reserved.
308
A. Henderson-Sellers et al.
(e.g. Craig, 1961). Optimal use of irrigation water involves, among other issues, maximizing the proportion of water that is transpired. Water transpired by vegetation has approximately the same isotopic composition as root-zone soil water and groundwater (Zimmerman et al., 1967) so this component of total evaporative flux to the atmosphere does not affect the isotopic composition of residual water (Allison et al., 1983). Thus, stable isotope compositions are independent, sensitive monitors of both precipitation infiltration and evaporative loss. 1.2. Previous water isotope analysis in the Murray–Darling Basin The Murray River and its longest tributary, the Darling River, drain about 14% of the continent of Australia (Fig. 1(a)). Potential evaporation generally exceeds precipitation, which ranges from >1000 mm in the eastern highlands to <250 mm in the semi-arid interior. Consequently, the Murray discharges only about 5% of its basin’s annual precipitation (McMahon, 1982) and the Darling <1% (Simpson and Herczeg, 1991a, 1991b), placing it among the world’s lowest discharging major rivers. Surface runoff is controlled by storage reservoirs, weirs and canals used in the widespread irrigation in the basin. The very large, inter-annual fluctuations in hydrological components and forcings generated by the El Niño Southern Oscillation (ENSO) combine with the semi-arid climate to make simulating this basin’s hydrology in large-scale models very challenging. HendersonSellers et al. (2004) showed that only 7 of 20 GCMs evaluated produced basin-averaged
Fig. 1. (a) Map of the location of the Murray–Darling Basin and the groundwater aquifers in and close to the Murray–Darling Basin analyzed in this study. (b) Murray–Darling Basin in SE Australia showing the sampling stations.
Stable water isotopes as tools for water cycle
309
hydrologies that were valid. It has been postulated that global warming could produce a quasipermanent El Niño circulation with very significant consequences for the Murray–Darling (e.g. Timmerman et al., 1999). Indeed, warming has begun to affect the nature and impact of El Niños in Australia (e.g. Nicholls, 2004) although his proposal that evaporative fluxes are increasing has yet to be demonstrated. Stable isotopic characterization has played a role in determining the seasonal and geographical variations in evaporation and groundwater input across the Murray–Darling Basin since the 1980s. Herczeg et al. (1992) drew on data from Burtundy weir on the Darling River to show extreme enrichment of river water in the 1986–1987 El Niño (a 70h (δD = −40h to −30h) range) but little enrichment in the La Niñas of 1984–1985 and 1988–1990. Simpson and Herczeg (1991a, 1991b) established that the observed enrichment of stable isotopes in surface waters during the summer (October–March) offers the potential to deduce magnitudes of water evaporated. They observed a +39h change in δD between headwaters (δD = −43h) and the mouth (δD = −4h), attributable to evaporation since transpiration does not enrich the residual water. For effective mean humidities of 45, 55 and 65%, using the formulation of Gonfiantini (1986), the linear part of their curves (<50% evaporation) yields an enrichment of between 0.62h and 1h in δD for every 1% change in residual water. They concluded that isotopes effectively integrate evaporation losses from rivers and irrigation diversions and deduced that 40% (±15%) of the evaporative loss from the Murray River was from open water, a similar amount to plant transpiration during summer irrigation. Henderson-Sellers et al. (2004) evaluation of the simulations of three large basins over the AMIP II1 period, which included an isotope GCM (MUGCM: Noone and Simmonds, 2002), found the hydrology of the Murray–Darling to be less well simulated than either the Mississippi or the Amazon. The MUGCM predicted the seasonal range in δ 18 O and δD well and, although the annually-averaged values were ∼50% too depleted, the D excess was underestimated by only about 2–3h. 1.3. Objectives of this study This paper examines the potential of stable water isotopes in the context of the challenge of simulating large low-flow, semi-arid rivers such as the Murray–Darling at the resolution of global climate models (GCMs) (e.g. Henderson-Sellers et al., 2004). Two aspects of the Murray–Darling Basin’s hydrology are examined isotopically in the context of the significant fluctuations due to ENSO: precipitation intensity and river plus irrigation evaporation. The thrust of the first part (Section 3) is to demonstrate that the palaeo effects are due primarily to ‘selectivity’ or ‘rainfall distribution’ characteristics and not to temperature. Of course, we cannot exclude temperature as playing some role; however, we find that the relative magnitude of the role decreases with increasing aridity in warm climates and are now gathering parallel independent evidence from palaeoclimate studies in Lake Eyre and the Todd River (Magee et al., 2004). We recognize that in cool temperate climates, changes in the isotope ratios are forced principally by temperature so variations in the stable isotopic composition of dated groundwater can reconstruct palaeoclimatic change (e.g. Jouzel et al., 1998). Here, we invert this tool and use isotopic signatures as a means of determining precipitation intensity 1 General information on AMIP is available at http://www-pcmdi.llnl.gov/projects/amip/index.php.
310
A. Henderson-Sellers et al.
in the past and apply this to future simulations. In the second part (Section 4), we improve the model of Simpson and Herczeg (1991a) for evaporative enrichment of stable water isotopes.
2. Materials and methods 2.1. Palaeodata for isotopic analysis The analysis by Airey et al. (2003) has been extended here for Alice Springs and Melbourne, which characterize two end points of the Murray–Darling’s climate and hydrology, and with reference to daily data from Yass, central to the basin. Airey et al. (2003) found that matching the isotopic characteristics of groundwater to the rainfall isotopic characteristics implied the existence of a threshold rainfall intensity for groundwater recharge. We find the threshold above which groundwater recharge occurs is 65 mm month−1 for the Newer Basalt aquifer (Melbourne), 80 mm month−1 for the Mereenie Sandstone (Alice Springs) and between these for the Williams Creek aquifer (Yass) (Fig. 2a). (The weighting is by precipitation volume in Fig. 2a.) 2.2. Modern observations of isotopes Water samples were collected from 9 stations along the Darling River, monthly from July 2002 to January 2003 and weekly at Burtundy. Samples were also collected from Mildura on the Murray River on 28th October 2002 and 1st January 2003, and from Hay and Gundagai on the Murrumbidgee River on January 2nd 2003 (Fig. 1(b)). Their δD and δ 18 O values fall on a line
Fig. 2a. Variation of precipitation volume weighted mean δD (h) in rainfall exceeding threshold precipitation intensities as observed at Melbourne (monthly – diamonds), Yass (daily – triangles) and Alice Springs (monthly – circles) and the three dated Mereenie groundwaters plus the Williams creek and Newer Basalts groundwaters with (inset) for Alice Springs the GNIP-derived (IAEA, 1999) δD in rainfall plus the groundwater δD from the Mereenie aquifer.
Stable water isotopes as tools for water cycle
311
Fig. 3a. Stable isotopic data (diamonds) for the whole Darling River for 2002 together with relative humidity of 50% evaporation predictions (squares) and (inset) the data for the Darling River at Burtundy expanded showing the greatest determined evaporation which is logarithmic.
of slope ∼5 (Fig. 3a), with scatter only in water released from the Menindie lakes, which possesses unusual isotopic contents due to the long residence time there (inset in Fig. 3a).
3. Groundwater recharge potential from GCM predictions 3.1. Groundwater isotopes in semi-arid low latitudes Semi-arid low latitudes are characterized by low average rainfall, but high variability in monthly intensities. Recent work suggests that there is what we term a ‘threshold’ intensity below which there is no significant groundwater recharge (Airey et al., 2003). Our use of the word ‘threshold’ does not imply a ‘step’ function for rainfall recharge but rather the concept that below a certain monthly rainfall intensity, the probability of effective recharge rapidly decreases. High variability and large evaporative demand mean there is likely to be little antecedent water in the unsaturated soil zone prior to the infrequent heavy rainfall events which recharge groundwater. Nonetheless, mixing is important so that this methodology is best applied to deep (i.e. older) groundwater in arid zones. The residence time (age) of the groundwater must be such that the effect of mixing over the decadal timeframes have had time to disperse. This may take hundreds of years (hence the need for groundwater of >1000 years in age) and even then palaeoclimate fluctuations with return periods of hundreds to thousands of years or longer will not have dispersed. The reason that this methodology best applies to arid zones is that the interval between heavy rainfall events is such that the upper soil layers ‘dry out’: there are fewer ‘memory’ effects. There are many caveats that need to be applied to such basin-wide intensity–recharge relationships including separability of groundwater and precipitation isotopic characteristics and groundwater mixing. However, the negative correlation between precipitation amount and δD is well recognized as being largely dependent on precipitation processes and is thus
312
A. Henderson-Sellers et al.
likely to be conserved over time. Figure 2a inset shows the IAEA (1999) frequency distribution of Alice Springs rainfall δD extending over the range −120h to +60h together with the groundwater δD values from the Mereenie aquifer. Groundwater in the Mereenie can be classified into three distinct recharge periods: ‘recent’ (i.e. pre-nuclear but too young for credible 14 C dating; δD = −63.0 ± 0.6h); ‘recent + 1400 years’ (δD = −61.6 ± 0.8h); and ‘recent + 4100 years’ (δD = −59.6 ± 0.6h). The procedure indicates a threshold intensity of 84 mm month−1 for recent groundwater to 79 and 75 mm month−1 for the two older groundwaters. In temperate climates such as Melbourne, most of the recharge rainfall infiltrates with little isotopic discrimination, and the isotopic composition of the groundwater (Newer Basalts aquifer δD = −30.8h to δD = −32.8h) is little depleted with respect to the weighted mean average rainfall value (−28.2h). The corresponding values are from Williams Creek groundwater (−45h) and Yass rainfall (−28h). These data, from 15 months of daily rainfall, penetrate to very shallow groundwater with a threshold around 70 mm month−1 . 3.2. AMIP experiments and implications for groundwater resources Precipitation intensity offers a test of Global Climate Model (GCM) hydrological predictive skill (McGuffie et al., 1999). The Atmospheric Model Intercomparison Project phase II (AMIP II) offers an excellent opportunity for assessing a large range of current GCMs’ simulations of the contemporary climate (Gates et al., 1999). Partitioning the precipitation of the 7 ‘good’ and 13 ‘poor’ 2 models of Henderson-Sellers et al. (2004) shows large differences (Fig. 2b) for the AMIP II period (January 1, 1979 through March 1, 1996), which includes four
Fig. 2b. Frequency of monthly precipitation intensities from the 7 ‘good’ and 13 ‘poor’ AMIP (Atmospheric Model Intercomparison Project) II GCMs (Global Climate Models) for El Niños (warm and dry) for the Murray–Darling Basin and inset La Niñas (wet and cool) for ‘good’ models. 2 Henderson-Sellers et al. (2004) define ‘good’ models in terms of conservation of water and energy over the basin; ‘poor’ models do not conserve water.
Stable water isotopes as tools for water cycle
313
Table 1 Precipitation intensity thresholds applied to 7 ‘good’ (conserves water) and 13 ‘poor’ (fails to conserve water) AMIP II AGCMs for El Niño (dry and warm) and La Niña (wet and cool) periods Climate regime
>65 mm month−1
>80 mm month−1
>95 mm month−1
Dry and warm Good models (G) Poor models (P)
441 307
322 173
240 120
30%
46%
50%
598 346
506 218
409 154
42%
57%
62%
156 mm
184 mm
169 mm
% difference (G − P) Wet and cool Good models (G) Poor models (P) % difference (G − P) Good models Difference (wet − dry)
El Niños (1982–1983, 1986–1987, 1991–1992 and 1994–1995) and two La Niñas (1984–1985 and 1988–1989), prescribed in sea surface temperatures (Henderson-Sellers et al., 2003). Both model groups in Fig. 2b present a tri-modal distribution but the ‘good’ models have more high intensity rain. Above selected intensity thresholds, total precipitation stratified by ENSO-based climate regime differs by from 30% to 62% (Table 1). Even ‘good’ GCM simulations of the Murray–Darling predict differences in groundwater recharge (a vital resource) of 156–169 mm between wet and dry ENSO periods.
4. Stable water isotope data measure the extremity of drought In this section, drought-driven isotopic depletions in the Murray–Darling during the 2002– 2003 El Niño establish the degree of evaporation. 4.1. Interpretation of stable water isotope data in the 2002–2003 Darling drought The most depleted values (bottom left of Fig. 3a) lie in the range δD = −24h to δD = −33h, and are not from the Darling River, but from the Murray–Murrumbidgee (Fig. 3b). The data for the Darling River are quite distinct from those from the Murray– Murrumbidgee (Fig. 1(b)), ranging from Collarenabri (2018 km from the mouth) δD = −10h to Burtundy (112 km) δD = 50h, similar in magnitude to the variations observed by Herczeg et al. (1992) during 1984–1991, but with a much greater degree of evaporative enrichment. The most depleted/least evaporated samples (Fig. 3b) demonstrate that the Murray River lost 5.3% of water by evaporation at Mildura between October 28, 2002 and January 1, 2003. Extrapolating to the end of February gives about 10% of Murray River water lost by evaporation in the November to February irrigation season. Murrumbidgee River data obtained at Gundagai (10% loss) and Hay (19% loss) for January 2, 2003 indicate a cumulative loss of 9% along the reach from Gundagai to Hay between December 2002 and January 2003.
314
A. Henderson-Sellers et al.
Fig. 3b. Data for the Murray River (Mildura – filled circles) and 2 locations on the Murrumbidgee River (open circles) for January 2003 together with the global meteoric water line (GMWL) and the evaporative model for relative humidity of 50%.
5. Discussion: Stable water isotopes as integrators for disturbance diagnosis The year 2002 was noted for low flows over much of the Murray–Darling Basin: discharge varied from zero to only 10–20 megalitres per day for much of the year, while flowing steadily at 200–250 megalitres per day at Bourke 800 km upstream. The Darling ceased to flow upstream of Burtundy (from Bourke to Wilcannia) in the period November 2002 to April 2003. The Darling River isotopes reveal increased evaporation for successive downstream samples combined with an impressive 2002 drying trend: every station records a monotonic evaporative increase (Fig. 3a). We improve the model of Simpson and Herczeg (1991a) by using both δD and δ 18 O at different effective mean humidities. From Gonfiantini (1986), a relationship was derived from a starting isotopic composition of δD = −32h, δ 18 O = −5h (e.g. Simpson and Herczeg, 1991b). As the input rainfall has δD ≈ 30h (e.g. Henderson-Sellers et al., 2004), the water is already quite evaporated even at the most upstream stations of Collarenabri and Gunnedah (δD in the range −4h to −8h). This is confirmed by projecting onto the local meteoric water line, which gives the likely initial isotopic composition of the water body of δD = −32h. Using the relationship developed for both stable isotopes, we calculate the cumulative loss at about 64% as compared to the approximation of Simpson and Herczeg (1991a) of at least 80%. Determination of the extent of evaporative losses from rivers, and the irrigation systems they feed, can be achieved from flows measured in engineered constructions if they exist. For the Murray–Darling, we have demonstrated an alternative using isotopic characterization of river water: estimates agree well with hydrologic observations and improve on relationships deduced from 1980s’ data. At Burtundy (Fig. 3a), the cumulative loss from January 8, 2002 (40.4%) to January 7, 2003 (64%) was 22.5%, underlining the extremity of the 2002–2003 El Niño drought (Nicholls, 2004). If this is a harbinger of the future, groundwater will become
Stable water isotopes as tools for water cycle
315
even more important to this agricultural region both as a direct resource and by adding to river flow. Unfortunately, more prevalent El Niño conditions are also likely to reduce aquifer recharge, partly because there is less total precipitation and also because there are fewer very intense rainstorms (e.g. Fig. 2b). The total rainfall simulated as available for groundwater recharge decreases as a function of ENSO phase by between 156 mm yr−1 to 184 mm yr−1 depending on the intensity threshold selected (Table 1) using GCMs deemed ‘fit for purpose’. References Airey, P., Henderson-Sellers, A., Habermehl, M.A., Bradd, J., Chambers, S., Hughes, C. (2003). Sustainability of groundwater under climate change. IAEA Symposium on Isotope hydrology, May 2003. IAEA CN-104-P16. IAEA, Vienna. Allison, G.B., Barnes, C.J., Hughes, M.W., Leaney, F.W.J. (1983). Effect of climate and vegetation on oxygen-18 and deuterium profiles in soils. In: Isotope Hydrology 1983. IAEA, Vienna, p. 105. Craig, H. (1961). Isotopic variations in meteoric water. Science 133, 1702–1703. Dansgaard, W. (1964). Stable isotopes in precipitation. Tellus 16, 436–468. Gat, J.R. (1996). Oxygen and hydrogen isotopes in the hydrological cycle. Annual Review of Earth and Planetary Science 24, 225–262. Gates, W.L., Boyle, J.S., Covey, C., Dease, C.G., Doutriaux, C.M., Drach, R.S., Fiorino, M., Gleckler, P.J., Marlais, S., Phillips, T.J., Potter, G.L., Santer, B.D., Sperber, K.R., Taylor, K.E., Williams, D.N. (1999). An overview of the results of the Atmospheric Model Intercomparison Project (AMIP I). Bulletin of the American Meteorological Society 80, 29–55. Gonfiantini, R. (1986). Environmental isotopes in lake studies. In: Fritz, P., Fontes, J.Ch. (Eds.), Handbook of Environmental Isotope Geochemistry, vol. 2. Elsevier, Amsterdam, p. 113. Henderson-Sellers, A., Irannejad, P., McGuffie, K., Pitman, A.J. (2003). Predicting land-surface climates – better skill or moving targets? Geophysical Research Letters 30 (14), 1777–1780. Henderson-Sellers, A., McGuffie, K., Noone, D., Irannejad, P. (2004). Using stable water isotopes to evaluate basinscale simulations of surface water budgets. Journal of Hydrometeorology 5 (5), 805–822. Herczeg, A.L., Simpson, H.J., Dighton, J.C. (1992). Salinity, groundwater and evaporation in the Darling Basin. Murray Darling Basin Workshop, Renmark SA, 27–29 Oct. IAEA (1999). Global Network for Isotopes in Precipitation. The GNIP Database. Release 3, October 1999. Available at http://isohis.iaea.org/search.asp. Jouzel, J., Koster, R.D., Hoffman, G., Armengaud, A. (1998). Model evaluations of the water isotope-climate relationships used in reconstructing paleotemperatures. Vienna, 1998, pp. 485–502. Macumber, P.G. (2003). Lenses, plumes and wedges in the Sultanate of Oman: A challenge for groundwater management. In: Alsharhan, A.S., Wood, W.W. (Eds.), Water Resources Perspectives: Evaluation, Management and Policy. Development in Water Science, vol. 50. Elsevier, Amsterdam, The Netherlands, pp. 349–379. Magee, J.W., Miller, G.H., Spooner, N.A., Questiaux, D. (2004). Continuous 150 K monsoon record from Lake Eyre Australia: Insolation-forcing implications and unexpected Holocene failure. Geology 32 (10), 885–888, doi:10.1130/G20672.1. McGuffie, K., Henderson-Sellers, A., Holbrook, N., Kothavala, Z., Balachova, O., Hoekstra, J. (1999). Assessing simulations of daily variability with global climate models for present day and enhanced greenhouse climates. International Journal of Climatology 19 (1), 1–26. McMahon, T.A. (1982). World hydrology: Does Australia fit? In: Inst. Eng. Nat. Conf. Pub. No. 82/3. Melbourne, Australia, pp. 1–7. Nicholls, N. (2004). The changing nature of Australian droughts. Climatic Change 63, 323–336. Noone, D., Simmonds, I. (2002). Association between δ 18 O of water and climate parameters in a simulation of atmospheric circulation for 1979–95. Journal of Climate 15, 3150–3169. Simpson, H.J., Herczeg, A.L. (1991a). Stable isotopes as an indicator of evaporation in the River Murray, Australia. Water Resources Research 27 (8), 1925–1935. Simpson, H.J., Herczeg, A.L. (1991b). Salinity and evaporation in the River Murray basin, Australia. Journal of Hydrology 124, 1–27.
316
A. Henderson-Sellers et al.
Timmerman, A., Oberhuber, J., Bacher, A., Esch, M., Latif, M., Roeckner, E. (1999). Increased El Niño frequency in a climate model forced by future greenhouse warming. Nature 398, 694–696. Zimmerman, U., Ehhalt, D., Munnich, K.O. (1967). Soil water movement and evapotranspiration: Changes in the isotope composition of water. In: Isotopes in Hydrology. IAEA, Vienna, pp. 567–584.
317
Isotopic characteristics of the Sava River basin in Slovenia Nives Ogrinc* , Tjaša Kanduˇc, Janja Vaupotiˇc Department of Environmental Sciences, “Jožef Stefan” Institute, 1000 Ljubljana, Slovenia Abstract First results of isotopic research in the Sava River basin in Slovenia are presented. The seasonal variability of δ 18 O and δD of the river Sava reflects the isotopic composition of precipitation. Tritium concentrations up to 18.3 T.U. in the Sava River basin indicate that modern recharge occurs, except in the Krsko–Brezice area. Dissolved inorganic carbon (DIC) concentrations and its isotopic composition (δ 13 CDIC ) were used to trace watershed inputs and fluvial processes affecting the inorganic pool of the carbon budget. The total DIC flux of the river Sava was estimated to be 1.10 × 108 mol C/day in the spring of 2004, and lower, 4.4 × 107 mol C/day in the fall of 2004. An isotopic mass balance calculation at the Slovenian–Croatian border show that the relative contributions of the tributaries, exchange with the atmosphere, dissolution of carbonates, and degradation of organic matter to the total DIC flux were 70, −0.9, 20, 11 in the spring, and 62, −5, 26, 17 in fall of 2004, respectively. Keywords: Stable isotopes, 18 O, Deuterium, Tritium, 13 C, Surface water, Sava River
1. Introduction Rivers are critical components of the hydrological cycle, acting as drainage channels for surface water; the world’s rivers drain nearly 75% of the Earth’s land surface. Much of the world population lives along large rivers, relying on them for trade, transportation, industry, agriculture, and domestic water supplies. For sustainable management of the water supply, agriculture, flood-drought cycles and ecosystems, and human health, there is a basic need for improving the scientific understanding of water cycling processes in river basins. Isotopic analysis has become a valuable tool in global environmental change research (Kendall and McDonnell, 1998). Environmental isotopes of a few representative elements (such as oxygen, hydrogen, and carbon) in all source waters of a particular system, can help to make quantitative estimates on the proportions of the various components which contribute to the ultimate discharge (Yang et al., 1996; Pawellek et al., 2002). Isotopes of particular interest for hydrological studies include mainly the stable isotopes of water (18 O and D). They exhibit systematic variations in the water cycle as a result of isotopic fractionations * Corresponding author. Address: Department of Environmental Sciences, “Jožef Stefan” Institute, Jamova 39, 1000 Ljubljana; phone: (+386) 1 5885 387; fax: (+386) 1 5885 346; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08025-3
© 2006 Elsevier Ltd. All rights reserved.
318
N. Ogrinc et al.
that accompany phase changes and diffusion (Gat, 1996; Lachniet and Patterson, 2002). Coupled with measurements of isotopes in water sources and river discharge, such signatures can provide information on groundwater recharge/discharge processes, water balance, and snow mixing. On the other hand, tritium (3 H), a radioactive isotope characterized by a half-life of 12.3 years with measurable levels in the hydrosphere resulting from nuclear weapons tests, can be used as a transient tracer for following water masses (Acheampong and Hess, 2000; Maloszewski et al., 2002; Bolsunovsky and Bondareva, 2003). The isotopic composition of dissolved inorganic carbon (DIC) has also proven useful for study of river basin processes. It depends upon: (a) the extent to which atmospheric CO2 is in equilibrium with the water mass; (b) groundwater discharge; (c) leaching and degradation of soil CO2 and particulate organic carbon of natural terrestrial and domestic sewage origin, and (d) in situ photosynthesis (Mook and Tan, 1991). In addition, these isotopes can be used to quantitatively estimate storage, evaluate mixing between various reservoirs, and predict retention times between inputs and outflows of a riverine system. This contribution presents the first systematic chemical and isotopic composition (δ 18 O, δD, 3 H, δ 13 C) research of the river Sava and its tributaries in Slovenia in the study of the downstream evolution of river hydrogeochemistry. 2. Materials and methods 2.1. Investigation area The Sava catchment is the largest catchment in Slovenia covering almost half of the country (10,838 km2 ), and at the same time, one of the tributary streams of the river Danube, the second largest river in Europe. From the source of the Sava Dolinka to the national border with Croatia, its length is 219 km. The Sava Bohinjka and Sava Dolinka mainly drain Triassic, dolomites and limestones, while from their confluence at Radovljica the river accumulates alluvial gravel, sand and clay. From Ljubljana to Log it drains Permo-Carbonian shales, sandstones and conglomerates, and from Log to the border with Croatia it drains Triassic carbonates, Miocene marls, sandstones, claystones and limestones. Samples were taken for δ 18 O, δD and δ 13 CDIC over the Sava watershed and its tributaries (41 samples) during spring and fall in 2004, while for 3 H determination the samples were taken only in fall 2004. Samples of precipitation were also taken in both sampling seasons. Flow regimes of the river Sava are related to precipitation, topography and reveal upstream Alpine high-medium snow–rain regime and a downstream Alpine rain–snow regime (Hrvatin, 1998). The discharge conditions ranged from 1.76 to 144 during spring 2004 and from 0.7 to 144 m3 /s during fall 2004, respectively. We also sampled groundwaters (9 samples) at Krsko– Brezice from the Sava catchments to characterize the stable isotope (δ 18 O, δD, δ 13 CDIC ) exchange between river Sava and groundwaters. The sampling locations are shown in Fig. 1. 2.2. Analytical techniques Temperature, pH, conductivity, and concentration of dissolved oxygen were measured in the field after filtering through a 0.45 µm Micropore filter using standard methods and used (together with alkalinity) for calculation of pCO2 . Samples collected for determination of the
Isotopic characteristics of the Sava River basin 319
Fig. 1. Sampling locations in the Sava River basin in Slovenia. Groundwater samples near the Krsko–Brezice area are also marked.
320
N. Ogrinc et al.
isotopic composition of oxygen and hydrogen (deuterium and tritium) were not filtered but simply stored in containers at 4◦ C prior to analysis. Samples for determination of cations were acidified using “ultrapure” HNO3 . The samples for anion analysis needed no further treatment. Aliquots for dissolved inorganic carbon (DIC) were preserved with CuCl2 , while for carbon isotope analyses (δ 13 CDIC ) they were not preserved and both samples were capped in glass serum vials. Chemical analyses of water samples were performed by ICP–OES (e.g. Ca, Mg, Sr, Na, K, Si, Al – precision ±2%), ion chromatography (Cl, sulfate, nitrate – precision ±1%) and Gran’s titration for alkalinity. More details of analytical methods are provided in Szramek and Walter (2004). δ 13 CDIC was determined with a modified method of extraction (Capasso et al., 2003). In a glass septum tube (VACUTAINER Septum Tubes, Labco Limited, UK) 100% phosphoric acid was added (100–200 µl) and the closed septum tube was purged with pure He to remove any air contamination. 6 ml of water samples was introduced into the septum tube using a syringe and the acid–water reaction began immediately upon injection. The isotope ratios of extracted CO2 was then determined directly from the headspace by an Isotope Ratio Mass Spectrometer – IRMS (Europa Scientific 20-20) with an ANCA-TG preparation module for trace gas samples equipped with a Gilson autosampler. In order to determine the optimal extraction procedure for water samples, a standard solution was prepared with Na2 CO3 of a known δ 13 C value of −10.8 ± 0.23h. The oxygen-18 content was measured after equilibration with reference CO2 at 25◦ C for 24 h (Epstein and Mayeda, 1953), while reduction on Cr at 800◦ C was used to determine the deuterium content of water (Gehre et al., 1996). Both measurements were performed on a Varian MAT 250 mass spectrometer. Stable isotope results are reported using conventional delta (δ) notation in h relative to a suitable standard; δD and δ 18 O are reported relative to VSMOW, while δ 13 C values are reported relative to VPDB. The precision of measurements was ±0.2h for δ 18 O, ±3h for δD and ±0.2h for δ 13 C. For 3 H measurements the majorities of distilled water samples were electrolytically enriched and re-distilled again (Florkowski, 1975, 1981). 6 ml of scintillation cocktail was added to 6 ml of sample. Counting was performed on a TRI CARB 3170 TR/SL ultra low-level Liquid Scintillation Counter (LSC, Canberra Packard) with a precision of ±1.8 T.U. All results are expressed in Tritium Units (T.U.).
3. Results and discussion 3.1. Oxygen and hydrogen isotopes Stable isotope data of δD and δ 18 O are plotted in Fig. 2 together with the global meteoric water line (GMWL) (Craig, 1961). The average δ 18 O value determined in spring was −10.1 ± 0.4h, while in fall the average value was found to be −9.2 ± 0.3h. Therefore, the river Sava and its tributaries show a ∼1h depletion in δ 18 O in spring samples, relative to the fall. Seasonal variability was also reflected in the stable isotope values of precipitation (Fig. 2). The isotopically depleted rainfall samples collected during the major rainy season in the spring had a δ 18 O value of −11.8h and a δD value of −90h. The isotopic value of
Isotopic characteristics of the Sava River basin
321
Fig. 2. δD vs. δ 18 O for the Sava River basin. The data for precipitation in the sampling season, snow and annual precipitation are presented, together with the Global Meteoric Water Line (GMWL), after Craig (1961). Only data with both parameters (δ 18 O and δD) are presented.
precipitation in fall 2004 had δ 18 O and δD values of −7.8h and −48h, respectively. The lowest δ 18 O value of −12.4h was obtained near Savica, the source of the river Sava Bohinjka, in spring 2004. This data reflects the origin of water from snow melting occurring during the spring and the elevation effect. The isotopic composition observed in the Savica sample is lower than that observed in seasonal precipitation. In the fall of 2004 δ 18 O and δD values of the river samples approached the mean annual precipitation determined in the years 2000–2003 (Vreca et al., 2004). It seems that this presents the base flow of the river Sava which integrates the composition of precipitation over the drainage areas. All the groundwater samples taken in the Krsko–Brezice area (Fig. 1) in alluvial aquifers lie on or close to the GMWL (Fig. 2) and indicate that the groundwater has not been greatly fractionated by kinetic evaporation. The mean δ 18 O value of −9.3 ± 0.3h and δD value of −64 ± 3h are therefore reasonable representations of the groundwater in the study area. The similarity in the δ 18 O and δD values between river Sava and groundwater samples in this region indicates that during the rainy season a significant recharge between groundwater and river water occurs. The measured tritium content in the river Sava and its tributaries has an average value of 13.7 ± 2.4 T.U. This is close to the current rainfall value of 7.9 T.U., determined at the same time near Ljubljana, that indicates a residence time of the river water up to about 10 years, and a presence of young waters in the river watershed (Clark and Fritz, 1997). The maximum tritium content of 25.1 ± 2.0 T.U. was determined near the Krško Nuclear Power Plant. However, it was found that the international limits for tritium discharges from nuclear facilities have never been exceeded (Kozar Logar et al., 2005).
322
N. Ogrinc et al.
3.2. Riverine carbon cycle Groundwaters are Ca–Mg–HCO− 3 waters and are slightly oversaturated with respect to carbonate minerals and are near equilibrium for dolomite at pCO2 values up to twenty times atmospheric values. Waters draining end-member (calcite or dolomite) lithologies in groundwaters show Mg2+ /Ca2+ values from 0.25–1.0 (Szramek et al., in preparation). Geochemically similar to groundwater, surface waters have lower pCO2 values and are supersaturated for calcite. However, pCO2 values are still well above the atmospheric equilibrium pressure and represent a source of CO2 into the atmosphere. Bicarbonate (HCO− 3 ) is the dominant anion and forms over half of the total anions in the river. The watershed exhibit a stable concentration (Ca + Mg)/HCO− 3 ratio of 1:2, which is typical for rivers dominated by carbonate weathering (Barth et al., 2003). The Sava drainage-area-normalized HCO− 3 fluxes are about − 2 7–15 times higher than the world average (∼10 meq HCO3 /km ). Although nearly an order of magnitude increases in specific discharge in the Sava River basin occur during spring, the alkalinity concentrations do not vary significantly. This suggests that carbonate equilibria and mass transfer rates are capable of keeping pace with increasing water flux, and it appears that soil zone/groundwater source pCO2 controls the alkalinity of stream waters and thus, the weathering intensity of the carbonate watersheds. δ 13 C of DIC of the Sava River basin varied between −5.8 and −13.5h in spring 2004, while the fall 2004 values were somewhat higher, at −3.3 to −12.8h. Increased photosynthesis by algae during the summer could have resulted in preferential removal of 12 C, shifting the DIC towards heavier values, as demonstrated for various other river basins (Pawellek and Veizer, 1995; Yang et al., 1996; Aucour et al., 1999; Barth et al., 2003). An inverse relationship between the DIC concentration and δ 13 C DIC values in the river was found with three different sources: the source of the river, waters originating from carbonate Alpine catchments and downstream waters. The source of the river is characterized by high δ 13 CDIC and low DIC concentrations. Waters originating from carbonate Alpine catchments are also 13 C-enriched, while 13 C-depleted waters with δ 13 CDIC of −11h were found downstream. In all locations the groundwater samples were more depleted in 13 C when compared to river water, ranging between −10.9 and −15.5h. Although the details of the carbon flux trend are complicated by local flow patterns and characteristics, a stable isotope mass balance calculation of DIC sources was performed in the Alpine region and at the outflow of the river Sava near the Slovenian–Croatian border. It was assumed that the major inputs that contribute to DIC flux at the outflow of the river from Slovenia (DICout ) and to its isotopic composition originate from tributaries (DICtri ), degradation of organic matter (DICorg ), exchange with the atmosphere (DICex ) and from dissolution of carbonates (DICca ) (Yang et al., 1996): DICout = DICtri − DICex + DICorg + DICca , DICout δ 13 Cout = DICtri δ 13 Ctri − DICex δ 13 Cex + DICorg δ 13 Corg + DICca δ 13 Cca . DICout and DICtri were calculated from the concentrations of DIC and water discharge, while the corresponding δ 13 C values (δ 13 Cout , δ 13 Ctri ) were measured. DICex was calcu-
Isotopic characteristics of the Sava River basin
323
lated from pCO2 and a value of +1.4h and 1.2h was chosen for δ 13 Cex in spring and fall, respectively (Kanduˇc et al., in preparation). These values were calculated from the δ 13 C value of atmospheric CO2 and fractionation of carbon isotopes between atmospheric CO2 and aqueous bicarbonate (Deuser and Degens, 1967; Yang et al., 1996). The theoretical diffusion flux of CO2 between the river and the atmosphere, DICex , was calculated from the diffusion model of Broecker (1974) assuming two different limits of water turbulence: no stirring and moderate stirring of water, respectively (Mook, 1970; Yang et al., 1996; Aucour et al., 1999). This implies that the total loss of inorganic carbon from the river surface ranged from a low as 6.6 × 105 mol C/day to high as 1.23 × 106 mol C/day during spring 2004. The diffusive flux of CO2 during the fall was estimated to be higher, ranging between 1.5 × 106 mol C/day and 2.8 × 106 mol C/day. δ 13 Corg represents the average δ 13 C value of −26.7 ± 1.2h for particulate organic matter, while for δ 13 Cca the value of +1.4h was applied in our calculation. The measured δ 13 Cca of particulate inorganic carbon from the river basin ranged from 0.5 to 2.1h with an average of 1.4h, which is similar to the carbonates forming the Sava River watershed (Kanduˇc et al., in preparation). The only unknowns in the system were DICorg and DICca which were calculated by solving the proposed mass balance equations. The calculated contributions to the average DIC budget from DICtri :DICex :DICorg :DICca at the Slovenian–Croatian border were 70:−0.9:20:11 in the spring 2004 and 62:−5:26:17 in the fall of 2004, respectively. In the Alpine region the ratios were quite different and were calculated to be 16:−0.01:50:34 in spring and 20:−0.03:50:30 in fall of 2004, respectively. Evolution of the river Sava DIC concentration and δ 13 C appeared to be largely determined by mixing between the Sava and its main tributaries which was found also in other larger rivers (Yang et al., 1996; Aucour et al., 1999). The dissolution of carbonates represents ∼36% and is more pronounced in carbonate Alpine catchments. The proportion of DIC derived from organic sources is more pronounced in fall and overall it represents ∼23%.
4. Conclusions The seasonal variability of δ 18 O and δD values apparent in precipitation is likely to be reflected in the isotopic composition of surface waters. Therefore, more systematic research is needed and our results could not be used to extrapolate mean annual or seasonal conditions. The DIC fluxes and DIC δ 13 C mass-balance indicate that in the carbonate Alpine region carbonate dissolution played an important role in the inorganic carbon budget and was not changing during the season. However, the evolution of DIC concentration and the δ 13 C value at the Slovenian outflow of the river Sava is essentially determined by conservative mixing between the river and its main tributaries (Krka River). The fluvial respired CO2 from the oxidation of organic matter was estimated to be ∼23 % and more pronounced during the fall of 2004. It should be mentioned that these are the first estimates performed of the total carbon budget in the river Sava. A more detailed investigation has to be performed on riverine suspended matter and soil/sediment properties along the river. In particular, the linkages between soil profile properties (thickness, carbon content, carbon dioxide partial pressures) needs to be studied to better define how landscape type influences carbonate weathering fluxes.
324
N. Ogrinc et al.
Acknowledgements The authors thank Prof. A. Byrne for linguistic corrections, Stojan Žigon and Zdenka Trkov for δ 18 O and δD analyses. This research was conducted in the framework of the project L26458-792 funded by the Ministry of Higher Education, Science and Technology of the Republic of Slovenia and the Slovenian Research Agency (ARRS). Part of the project was financially supported by the IAEA under Contract No. 12642 of the Coordinated Research Programme (CRP) of the IAEA, entitled “Design criteria for a network to monitor isotopic composition of runoff in large rivers”, and the EU project SARIB, Contract No. INC-CT-2004-509160.
References Acheampong, S.Y., Hess, J.W. (2000). Origin of the shallow groundwater system in the southern Voltaian sedimentary basin of Ghana: An isotopic approach. Journal of Hydrology 233, 37–53. Aucour, A.-M., Sheppard, S.M.F., Guyomar, O., Wattelet, J. (1999). Use of 13 C to trace origin and cycling of inorganic carbon in the Rhône River system. Chemical Geology 159, 87–105. Barth, J.A.C., Cronin, A.A., Dunlop, J., Kalin, R.M. (2003). Influence of carbonates on the riverine carbon cycle in an anthropogenic catchment basin: Evidence from major elements and stable carbon isotopes in the Lagan River (N Ireland). Chemical Geology 200, 203–216. Bolsunovsky, A.Ya., Bondareva, L.G. (2003). Tritium in surface waters of the Yenisei River basin. Journal of Environmental Radioactivity 66, 285–294. Broecker, W.S. (1974). Chemical Oceanography. Hacout Brace Jovanovich, New York, 214 pp. Capasso, G., Favara, R., Grassa, F., Inguaggiato, S., Longo, M. (2003). Automated techniques for preparation and measuring stable carbon isotope of total dissolved inorganic carbon in water samples (δ 13 CDIC ). In: 7th International Conference on Gas Geochemistry. Freiberg, Germany, 22–26 September, 2003. Programme & Abstract Book, p. 38. Clark, I., Fritz, P. (1997). Environmental Isotopes in Hydrology. Lewis Publishers, New York, 328 pp. Craig, H. (1961). Isotopic variations in meteoric waters. Science 133, 1720. Deuser, W.G., Degens, E.T. (1967). Carbon isotope fractionation in the system CO2 (gas)–CO2 (aqueous)– HCO− 3 (aqueous). Nature 215, 1033–1035. Epstein, S., Mayeda, T. (1953). Variations of 18 O contents of water from natural sources. Geochimica et Cosmochimica Acta 4, 213–224. Florkowski, T. (1975). Low level tritium essay in water samples by electrolytic enrichment and liquid scintillation counting in IAEA Laboratory. IAEA-SM-252/63, 335 pp. Florkowski, T. (1981). Tritium electrolytic enrichment using metal cells, low level tritium measurement. In: Proc. Consultants Meeting, Vienna, 1979. IAEA TECDOC-246, 133. Gat, J.R. (1996). Oxygen and hydrogen isotopes in the hydrological cycle. Annual Review of Earth and Planetary Sciences 24, 225–262. Gehre, M., Hoefling, R., Kowski, P., Strauch, G. (1996). Sample preparation device for quantitative hydrogen isotope analysis using chromium metal. Analytical Chemistry 68, 4414–4417. Hrvatin, M. (1998). Discharge regimes in Slovenia. Geografski zbornik XXXVIII, 60–87. Kendall, C., McDonnell, J.J. (1998). Isotope Tracers in Catchment Hydrology. Elsevier Science, Amsterdam, 840 pp. Kozar Logar, J., Vaupotic, J., Kobal, I. (2005). Tritium measurements in Slovenia – chronology till 2004. Fusion Science and Technology 48 (1), 431–434. Lachniet, M., Patterson, W.P. (2002). Stable isotope values of Costa Rican surface waters. Journal of Hydrology 260, 135–150. Maloszewski, P., Stichler, W., Zuber, A., Rank, D. (2002). Identifying the flow systems in a karstic-fissured-porous aquifer, the Schneealpe, Austria, by modeling of environmental 18 O and 3 H isotopes. Journal Hydrology 256, 48–59. Mook, W.G. (1970). Stable carbon and oxygen isotopes of natural water in the Netherlands. In: Isotopic Hydrology. IAEA, Vienna, pp. 163–190.
Isotopic characteristics of the Sava River basin
325
Mook, W.G., Tan, F.C. (1991). Stable carbon isotopes in rivers and estuaries. In: Degens, E.T., Kempe, S., Richy, J.E. (Eds.), Biogeochemistry of Major World Rivers. Wiley, New York, pp. 245–264. Pawellek, F., Veizer, J. (1995). Carbon cycle in the upper Danube and its tributaries: δ 13 CDIC constrains. Israel Journal of Earth Sciences 43, 187–194. Pawellek, F., Frauenstein, F., Veizer, J. (2002). Hydrogeochemistry and isotope geochemistry of the upper Danube River. Geochimica et Cosmochimica Acta 66, 3839–3854. Szramek, K., Walter, L.M. (2004). Impact of carbonate precipitation on riverine inorganic carbon mass transport from a mid-continent, forested watershed. Aquatic Geochemistry 10, 99–137. Vreca, P., Kanduc, T., Zigon, S., Trkov, Z., Pivk, P., Lojen, S. (2004). Isotopic composition of precipitation in Slovenia. JSI Working Report, DP-9039. Yang, C., Telmer, K., Veizer, J. (1996). Chemical dynamics of the ‘St. Lawrance’ riverine system: δDH2 O , δ 18 OH2 O , δ 13 CDIC , δ 34 Ssulfate , and dissolved 87 Sr/86 Sr. Geochimica et Cosmochimica Acta 60, 851–866.
Further reading Ministrstvo za okolje in prostor Klimatografija Slovenije 1961–1990 padavine (1995). Plantprint. Ljubljana, 366 pp.
326
222 Rn
as a tracer for the estimation of infiltration of surface waters into aquifers
M. Schuberta,* , K. Knoellera , H.-C. Treutlera , H. Weissa , J. Dehnertb a UFZ – Centre of Environmental Research Leipzig–Halle, Leipzig, Germany b Sächsisches Landesamt für Umwelt und Geologie, Dresden, Germany
Abstract The estimation of the infiltration of lake and river waters into aquifers is of vital interest in many fields. The paper focuses on the use of radon as an indicator for the assessment of such infiltration processes. Radon is produced naturally in every aquifer matrix and thus as the naturally occurring compound is present in groundwater. On the contrary, its concentration in lake and river waters is negligible. It was shown that a qualitative monitoring of the infiltration of surface waters into aquifers is possible by measuring the radon concentrations in observation wells situated close to the respective surface water body. Radon concentrations in groundwater, that are significantly lower than the respective natural background indicate such infiltration processes. However, critical consideration of the boundary conditions such as homogeneity of the aquifer, natural radon background of surface water and influence of percolating rain is needed for sound data interpretation. Keywords: Radon, Tracer, Aquifer, Groundwater, River water, Infiltration, Migration, Elbe River, Germany
1. Introduction 1.1. The general suitability of radon as a groundwater tracer River or lake water infiltration into aquifers via the sediment/water interface is a significant pathway for material transport. On the one hand, the process plays a key role for the hydrological fate of surface water bodies and it may contribute substantially to a diffuse pollution of groundwater recourses. The evaluation of infiltration processes is hence of importance for general studies of alluvial ecosystems. Since the infiltration rate governs the capacity of drinking water wells that are mainly recharged by surface waters, it is as well essential for the monitoring of drinking water catchment areas. The extensions of groundwater protection zones and well capture zones adjacent to surface waters mainly depend on water infiltration rates. Generally speaking, the estimation of the infiltration of surface waters into aquifers is of vital interest in many fields. * Corresponding author. Address: UFZ – Centre of Environmental Research Leipzig–Halle GmbH, Permoserstr. 15, 04318 Leipzig, Germany; phone: (+49) 341 235 2002; fax: (+49) 341 235 2625; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08026-5
© 2006 Elsevier Ltd. All rights reserved.
222 Rn as a tracer for the estimation of infiltrationof surface waters into aquifers
327
Useful tools for assessing exchange processes between groundwater and surface water are tracers, i.e. identifiable substances, which can be used to follow the course of water migration physically. Generally the ideal tracer has the same chemical properties and shows the same behaviour as the molecule it replaces, i.e. as the respective water. Hence the ideal tracer should meet the following requirements: • • • • •
it should be susceptible to quantitative determination at low concentrations, it should behave chemically inert, it should not be absorbed by the aquifer matrix, it should be safe in terms of human health and sustainability of ecosystems, it should be easy to apply to the water course of concern.
If groundwater or surface waters are in focus of the investigation, such ideal tracers would be 2 H or 18 O since both isotopes are natural constituents of water. However, due to cost issues and to accessibility of suitable analytical equipment, the groundwater and surface-water tracers, which are used most commonly, are fluorescent tracer dyes or salts. Spores, microbial tracers, and stable or radioactive isotopes are used less frequently. The limitations of the suitability of dyes and ionic compounds as tracers arise from their possible anthropogenic background concentrations in water, from the dependence of their stability on chemical and physical conditions, and from the possibility of decomposition, dispersion, or retardation. Another general disadvantage of the mentioned commonly used tracers is the necessity of adding them to the investigated hydrological system, which might be environmentally critical or in contradiction to the respective legislations. An intrinsic, naturally occurring compound of the groundwater, which can easily be analysed on site, would make the perfect tool for tracing the groundwater migration processes. The radioactive noble gas isotope 222 Rn, hereafter referred to as radon, is such a natural constituent of all groundwater, which has the potential of being used as an aqueous tracer. If the use of radon as a tracer is compared to the use of other aqueous tracers, several advantages can be named. Besides its natural occurrence key pros are: • anthropogenically elevated radon concentration in the groundwater, as it is possible for many salt tracers, can virtually be excluded; • radon concentration in groundwater is not influenced by changing chemical conditions (Eh, pH) as it is observed for many other tracers; • radon as a groundwater tracer is not subject to photolytic decomposition as it is observed for some fluorescence tracer dyes; • due to its noble gas configuration radon shows a chemically inert behaviour, which hinders retardation due to physical/chemical interaction with the aquifer matrix, flocculation/precipitation, or biodegradation; • radon can be detected precisely and selectively because of its radioactive nature. 1.2. Radon as a naturally occurring compound of groundwater Radon is produced naturally in every aquifer matrix by radioactive decay of radium (226 Ra). Hence, the total amount of radon, which is generated in the mineral grains of the aquifer, is determined by the massic activity of 226 Ra in the mineral matrix (ARa ).
328
M. Schubert et al.
Radium activity of soils and aquifers ranges between 10 and 50 Bq/kg. However, only part of the radon produced in the mineral matrix enters the pore space where it is dissolved in groundwater. The share of the total amount of produced radon atoms that is able to emanate from the mineral matrix to reach the pore space is quantified by the emanation coefficient of the mineral grains (ε). Emanation coefficients of ordinary soils usually range from 0.2 to 0.4. Porosity (n) and dry density of the aquifer matrix (ρd ) are also influential, because higher porosities at a given dry density result in a lower radon concentration in the pore space, i.e. in the groundwater. A typical porosity value for sandy aquifers is 0.45; a typical density is 1700 kg/m3 . The radon concentration in the pore space can be quantified using Equation (1). Radon concentrations in groundwater that are in equilibrium with the 226 Ra concentration of the respective aquifer matrix (C ∞ ) vary between 7.5 and 75 Bq/l. εARa ρd (1) (Bq/l). n If radon free water enters a mineral matrix (e.g. sand) the radon concentration in the water approaches the equilibrium concentration C ∞ with time. Hence, the level up to which C ∞ is reached depends on the residence time of the water in the aquifer (Fig. 1). Since the half-life of radon is 3.8 days the equilibrium concentration is reached after about 20 days (five half-lives). The diffusion length of radon in water is only 2 cm. Therefore, radon migration in water is in fact only determined by groundwater transport processes, i.e. by advection and dispersion. C∞ =
1.3. Application of radon as a groundwater tracer Several possible applications for the use of radon as a groundwater tracer can be named. The use of radon concentration patterns in coastal zone waters as an indicator for submarine groundwater discharge was suggested by several authors (Burnett and Dulaiova, 2003; Corbett et al., 1997; Cable et al., 1996). Comparable studies that focused on evaluating discharge rates of groundwater into rivers have been performed by Ellins et al. (1990), Genereux and Hemond (1990) and Yoneda et al. (1991). Semprini et al. (2000) suggested
Fig. 1. Build-up of the radon concentration in radon-free water that enters a mineral matrix.
222 Rn as a tracer for the estimation of infiltrationof surface waters into aquifers
329
Fig. 2. Schematic illustration of the radon-in-water concentrations in a scenario with river water infiltrating an aquifer.
the use of radon as indicator for the assessment of residual aquifer contamination by nonaqueous phase liquids (NAPLs), an approach which is based on the strong partitioning of radon into NAPLs (Clever, 1979). Comparable results for NAPL contamination of the vadose zone of the soil have been published by Schubert et al. (2001). Furthermore, the monitoring of spatial and temporal variations of the radon concentration in groundwater and also in soil gas has been used to provide insight into the dynamics of seismic activities aiming to allow a prediction of earthquakes or volcanic eruptions (Notsu et al., 1983; De la Cruz Reyna et al., 1985). In the application discussed here, radon is used as an indicator for assessing the infiltration of lake water and river water into aquifers. The method is based on the fact that, unlike groundwater, open surface waters lack considerable contact with radium bearing minerals. Since the decay of matrix-hosted 226 Ra is the only source of significant radon production, lake and river waters show small radon concentrations. Radon activity concentrations in river waters, which may be caused by increased amounts of suspended matter occurring in periods of high and turbulent waters or during heavy rainfalls triggering an increased surface runoff, are hardly ever of significance for the application discussed here. If virtually radon free surface water enters an aquifer via the water/sediment interface, its radon concentration rises due to the decay of radium present in the aquifer mineral matrix until it reaches the radon background concentration typical for the aquifer matrix. As mentioned above, the surface water entering the aquifer reaches that background concentration after about 20 days. The time dependent concentration increase was used by Hoehn and von Gunten (1989) for estimating groundwater residence times. Generally, during the time slot between infiltration and establishing equilibrium concentration, infiltrating water can be identified by its comparably low radon concentration (Fig. 2).
2. Experimental 2.1. Determination of radon concentrations in groundwater samples A common way to determine radon concentration in water samples is liquid scintillation spectrometry (LSC). However, in this study, the LSC method was only used to verify the on-site
330
M. Schubert et al.
results of the field measurements, which will be described and discussed later on. A total of 22 LSC measurements were carried out to back-up the on-site results of which a total number of 88 was determined. The LSC method was applied as described in the following section. A commercially available toluene cocktail (Packard) was used as radon extractor/scintillator. 20 ml of the scintillation cocktail were added to a flask containing one litre of the water sample and shaken in an overhead shaker for 10 min to reach radon equilibrium distribution. After shaking, a 10 ml aliquot of the cocktail was transferred into a glass vial and 10 ml of fresh cocktail were added to the flask. The 10 min shaking procedure was repeated. Finally, another 10 ml of the cocktail were taken from the flask and added to the glass vial. For the determination of the radon activity of the cocktail in the glass vial a liquid scintillation spectrometer was used (TRI-CARB 2550 TR/AB; Packard). The measurement error of the LSC is <5% for counting rates between 40 and 8000 cpm. Each vial was measured five times during a period of about 5 days. Since immediate results were wanted on-site, another approach was used for the field measurements. The applied on-site method allows for an uncomplicated and quick determination of the radon concentration in water samples. The method is based on the partitioning of radon between air and water and is described in the following section. At room temperature and under partitioning equilibrium between air and water the radon concentration in air is about four times the concentration in the respective water (Clever, 1979). The applied method uses that significant affinity of radon to air to extract radon from a water sample into a gaseous phase, which can be analysed for its radon concentration much more easily. Since the partitioning coefficient, which governs the equilibrium distribution of radon between the air and water is examined well over a wide range of temperatures, the original radon concentration in the water sample can easily be obtained from the measured radon concentration in the gas phase. The equipment used for the on-site analysis of the radon concentration in water samples works as follows. In a closed circuit gas cycle a defined volume of air is bubbled through a defined volume of the water sample. The radon dissolved in the water partitions into the air stream corresponding to the partitioning coefficient at the given temperature. The equilibrium radon concentration in the gas stream can be determined continuously by any radon-in-air monitor that allows the measuring of gaseous radon concentrations in a flow through mode. The radon concentration in the water sample can finally be calculated from the measured radon concentration in the gas stream. The radon monitor used for the on-site survey described here was the AlphaGUARD monitor (Genitron Instruments, Frankfurt, Germany). The actual radon detector is an ionisation chamber, operated in a 3D-alpha spectroscopy mode. Besides the radon monitor and the gas pump, which maintains the closed circuit gas loop the equipment used consists of the following parts: stripping vessel, safety vessel, and plastic tubing (Fig. 3). The stripping vessel is a cylindrical glass bulb covered with a cap that allows the injection of the water sample without any contact of the sample to the outside air. The size of the stripping vessel is 620 ml for water sample volumes of 500 ml. The safety vessel, which is basically of the same design as the stripping vessel but with a volume of only 50 ml, is positioned in the gas stream behind the stripping vessel. Its main purpose is to hold back water droplets, ensuring that moisture is kept out of the in-air-radon monitor. All connections between stripping vessel, safety vessel, gas pump, and radon monitor are made of flexible
222 Rn as a tracer for the estimation of infiltrationof surface waters into aquifers
331
Fig. 3. Set-up for on-site radon measurement in water (schematic). Arrows indicate the direction of the closed circuit gas flow.
tubing, which has a particularly high resistance against radon diffusion and hence ensures a minimum radon loss due to diffusion during the experiment. Although the described method seems to be a quite straightforward technique, representative groundwater sampling for radon analysis is problematic. A major concern is the easy escape of radon from the water if the sample comes in contact with the atmosphere. The use of a submersible pump and the direct transfer of the sample into the stripping vessel via radon proof tubing without any air contact is the suggested way of sampling. Turbulence and air bubbles have to be avoided. Furthermore, it is essential to obtain a water samples representative for the aquifer. That can be ensured by monitoring pH, redox potential, temperature, and electric conductivity of the water during sampling. With the aim to obtain information concerning the radon extraction efficiency and data reproducibility a large number of experiments have been carried out applying the experimental set-up described above. However, the discussion of these experiments and of the respective results is beyond the scope of the presented paper. Hence, the respective data will be published elsewhere. However, summarising the results achieved it can be stated that the radon extraction efficiency at the given temperature is close to 100% if the stripping process is continued for about 10 minutes. That minimum stripping time is mandatory for reaching radon equilibrium concentration between the water sample and the closed air loop. If radon equilibrium concentration between water and air can be guaranteed due to a sufficient stripping time, the data reproducibility of the method does mainly depend on the radonin-gas monitor applied. The sensitivity of the detector used for the described experiments (AlphaGUARD) is 1 cpm at 20 Bq/m3 . Its background signal due to internal detector contamination is less than 1 Bq/m3 . The range of radon concentrations, which can be detected using the AlphaGUARD is from 2 Bq/m3 to 2000 kBq/m3 . The instrument calibration uncertainty is ±3%. 2.2. Exemplary on-site application The purpose of this study was to prove the possibility of estimating changing intensities of river water infiltration into an adjacent aquifer by the use of radon distribution patterns in the
332
M. Schubert et al.
Fig. 4. Geological situation at the study site (schematic) including monitoring wells.
Fig. 5. Elbe water level (mSL – meters above sea level) with sampling campaigns A to D.
groundwater. For a pilot study, the Elbe River valley was chosen (Fig. 4). In this particular part of the valley, the Elbe River is in direct hydraulic contact with a Pleistocene aquifer consisting of fluviatile and glaci-fluviatile sands and Holocene gravels. Close to the town of Torgau/Saxony ten observation wells were installed in the Elbe River valley perpendicularly to the river at one of its banks. The wells allowed for depth-depending sampling by a total of 22 built-in membrane pumps. The actual sampling depths are illustrated in Fig. 4. It was assumed that an estimation of the intensity of the infiltration of river water into the aquifer is possible by the monitoring of the radon concentrations in the observation wells. Changing radon concentrations in the wells were expected to indicate corresponding changes of the infiltration rate. The goal of the described pilot study was the monitoring of relative changes of the infiltration rate and the qualitative assessment of its dependence on the changing river water level. The observation period was characterised by considerable variations in the river water level as displayed in Fig. 5.
222 Rn as a tracer for the estimation of infiltrationof surface waters into aquifers
333
3. Discussion Figure 6 shows selected results of the monitoring program. Four sampling campaigns (A to D) are discussed. The data shown were achieved by using the AlphaGUARD radon monitor and the on-site equipment described above. There is a highly reasonable agreement between the field results and the LSC measurements, which were carried out to verify the field data. Maximum deviations for single samples of 2.5 Bq/l were determined. Hence the field method applied could be considered as suitable for the purpose of the study. The first sampling campaign (marked with A in Figs. 5 and 6) was carried out during a period of low river water levels. The radon distribution in the aquifer, as it was determined from groundwater samples taken in the monitoring wells at several depths, is illustrated as A in Fig. 6. The normal radon background concentration of the groundwater is about 16 Bq/l. The much lower radon concentrations in the groundwater, which were determined only immediately adjacent to the riverbed, indicate only moderate infiltration rates of river water into the sandy aquifer. The second sampling campaign (B in Figs. 5 and 6) was carried out during a short but intense period of high river water levels. In that case, the radon distribution in the aquifer indicates a much higher infiltration rate, caused by the abruptly rising river water level of the river Elbe. A third sampling campaign (C in Figs. 5 and 6) was characterised by a constantly falling river water level and thus by declining infiltration rates. This gave rise to a beginning recovery of the radon concentrations in the aquifer close to the riverbed. However, since campaign C was carried out only 5 days after campaign B, radon concentrations in the just recently in-
Fig. 6. Radon distribution patterns (Bq/l) determined by the four sampling campaigns A, B, C and D. The black crosses (+) indicate the groundwater sampling locations (mSL – meters above sea level).
334
M. Schubert et al.
filtrated water were still depleted with regard to the background. A front of infiltrated river water moving through the aquifer away from the river still displays a significantly low radon level. The final sampling campaign (D in Figs. 5 and 6) is characterised by a return to normal river water levels. The radon distribution in the aquifer shows a pattern similar to the one in the first sampling campaign; the infiltration rate returns back to normal values. 4. Conclusions The results show that a qualitative monitoring of the intensity of the infiltration of surface waters into aquifers is possible by measuring the radon concentrations in observation wells situated close to the respective surface water. Changes of the infiltration rate result in changing radon concentrations in the observation wells. Hence, radon can be used as a naturally occurring tracer to assess and estimate infiltration from surface waters into aquifers. In the case of aquifers with a homogeneous radium distribution and an uniform grain size the ingrowths of the radon concentration of the infiltrating water can also be used to estimate the residence time and hence the migration velocity of the water in the aquifer as it was shown by Hoehn and von Gunten (1989). However, such conclusions could not be drawn in the study presented here, due to the inhomogeneity of the aquifer. In any case critical consideration of the boundary conditions, such as homogeneity of the aquifer, natural radon background of the surface water and influence of percolating rainwater is mandatory for sound data interpretation. References Burnett, W.C., Dulaiova, H. (2003). Estimating the dynamics of groundwater input into the coastal zone via continuous radon-222 measurements. Journal of Environmental Radioactivity 69, 21–35. Cable, J.E., Burnett, W.C., Chanton, J.P., Weatherly, G. (1996). Modeling groundwater flow into the ocean based on 222 Rn. Earth and Planetary Science Letters 144, 591–604. Clever, H.L. (Ed.) (1979). Krypton, Xenon and Radon – Gas Solubilities. Solubility Data Series, vol. 2. Pergamon Press, Oxford. Corbett, D.R., Burnett, W.C., Cable, P.H. (1997). Tracing of groundwater input into Par Pond, Savannah River site by Rn-222. Journal of Hydrology 203, 209–227. De la Cruz Reyna, S., Mena, M., Segovia, N., Chalot, J.F., Seidel, J.L., Monnin, M. (1985). Radon emanometry in soil gases and activity in ashes from El Chichon volcano. Pure and Applied Geophysics 123, 407–421. Ellins, K.K., Roman-Mas, A., Lee, R. (1990). Using 222 Rn to examine groundwater/surface discharge interaction in the Rio Grande De Manati, Puerto Rico. Journal of Hydrology 115, 319–341. Genereux, D.P., Hemond, H.F. (1990). Naturally occurring radon-222 as a tracer for streamflow generation: Steady state methodology and field example. Water Resources Research 26, 3065–3075. Hoehn, E., von Gunten, H.R. (1989). Radon in groundwater. Water Resources Research 25, 1795–1803. Notsu, K., Abiko, T., Wakita, H. (1983). Radon concentration changes in groundwater related to volcanic activity of Usu volcano. Bulletin of the Volcanological Society of Japan 28, 305–308. Schubert, M., Freyer, K., Treutler, H.-C., Weiß, H. (2001). Using soil gas radon as an indicator for ground contamination by non-aqueous phase-liquids. Journal of Soils and Sediments 1, 217–222. Semprini, L., Hopkins, O.S., Tasker, B.R. (2000). Laboratory, field and modeling studies of radon-222 as a natural tracer for monitoring NAPL contamination. Transport in Porous Media 38, 223–240. Yoneda, M., Inoue, Y., Takine, N. (1991). Location of groundwater seepage points into a river by measurement of 222 Rn concentration in water using activated charcoal passive collectors. Journal of Hydrology 124, 307–316.
335
Monitoring of geochemical and geophysical parameters in the Gran Sasso aquifer Wolfango Plastinoa,b,* a Department of Physics, University of Roma Tre, Rome, Italy b National Institute of Nuclear Physics, Section of Roma Tre, Rome, Italy
Abstract Since May 1996, groundwater monitoring (for 222 Rn concentration, pH, electrical conductivity, pressure of dissolved gases and temperature) has been carried out at the Gran Sasso National Laboratory of the National Institute of Nuclear Physics using multiparametric equipment. This monitoring scheme has been designed with the aim to better define the geophysical properties of the Gran Sasso aquifer and its radon source(s). The time series analyses show strong anomalies in pH and radon concentrations, highly correlated with the known 1997–1998 Umbria–Marche seismic sequence, which occurred in the central Apennines, Italy. Keywords: Radon, Radon daughters, Geophysics, Groundwater, Aquifer, Earthquake
1. Introduction Chemical–physical groundwater monitoring has been performed in seismic areas to estimate possible correlations between its spatial–temporal variations and deformation phenomena of tectonic interest (Igarashi et al., 1995). Nevertheless, non-tectonic factors related to chemical– physical groundwater variations have an important role as well (Plastino and Bella, 2001; Shapiro et al., 1985). It is important therefore to define correctly geological, hydrogeological and meteorological characteristics of the site, if variations induced by tectonic phenomena are assessed (Roeloffs, 1999). Another important aim is to test continuously the stability and reliability of the monitoring system during the measurement period (Plastino and Bella, 2001), and/or using new detectors, independent on environmental noise parameters such as temperature, acid concentrations, humidity and air pressure (Plastino et al., 2002). * Address: Department of Physics, University of Roma Tre, Via della Vasca Navale, 84, I-00144 Roma, Italy; phone:
(+39) 06 55177277; fax: (+39) 06 5579303; e-mail:
[email protected] RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08027-7
© 2006 Elsevier Ltd. All rights reserved.
336
W. Plastino
2. Equipment and method The automatic multiparametric equipment is made of stainless steel, fitted with pneumatic valves. It consists of a system for the geochemical analysis of groundwater (e.g., temperature, electrical conductivity and pH), for the extraction of dissolved gases (pressure of dissolved gases) and the detection of alpha-particles arising from 222 Rn and its short-lived daughters 218 Po and 214 Po. The groundwater to be analyzed flows through a tube inserted into the rock down to a depth of about three meters, and a sampling period of twelve hours has been selected (Plastino and Bella, 2001). The measurement site is located at the Gran Sasso National Laboratory near the main overthrust fault; particularly, in the well-drained cretacic formations. In fact, the underground laboratory is characterized by an overthrust fault which separates water masses belonging to two distinct creeks, and then interesting to better define the radon groundwater transport processes through the fault itself.
3. Results The temporal variations of the geochemical parameters observed during the period from May 1996 to June 1999 are shown in Fig. 1. Data for groundwater temperature from July 1997 to September 1997 are not available due to an equipment problem during that period. The observed data indicate: (i) the temporal variation of groundwater temperature has a standard deviation of 0.2◦ C; (ii) the temporal variation of electrical conductivity has a standard deviation of 0.2 mS/m apart from large fluctuations in October–November 1996 and in August–September 1998; (iii) during the period from September 1996 to March 1997, the pH increased by approximately 0.4; from April 1997 to August 1997 a pH buffering effect occurred, and the pH then decreased, reaching approximately its initial value by November 1997; (iii) a large fluctuation in 222 Rn concentration in groundwater occurred in November 1996. Furthermore, the mean value of the 222 Rn concentration decreased from January 1997 to March 1997, and then decreased again from April 1997 to September 1997, and then suddenly increased on September 15th, remaining high until November 1997. The temporal variations of the setting parameters from May 1996 to June 1999 are shown in Fig. 2. The environmental temperature shows a periodic component with T = 180 days and some spikes due to ventilation activity inside the underground laboratory. The instrument background, and its response to the calibration 241 Am source counted before each measurement, were both stable over the period of the measurements. The gas pressure in the counting cell appears to rise steadily, with two abrupt rises: the first in July 1997 (∼1.5 kPa) and the second in September 1997 (∼2 kPa). The pressure had decreased to approximately its initial value by September 1998.
Monitoring of geochemical and geophysical parameters in the Gran Sasso aquifer
337
Fig. 1. The temporal variations of the groundwater temperature, electrical conductivity, pH and 222 Rn monitored during the period from May 1996 to June 1999.
338
W. Plastino
Fig. 2. The temporal variations of the environmental temperature, background counting, 241 Am counting and pressure of gases monitored during the period from May 1996 to June 1999.
Monitoring of geochemical and geophysical parameters in the Gran Sasso aquifer
339
4. Discussion This work has focused on investigating possible correlations between observed variations in groundwater chemical–physical parameters and strain processes of tectonic interest. The occurrence probability of hydrogeochemical spike-like anomalies has been estimated with a time series approach and has been computed avoiding the influence of nontectonic factors, i.e. meteo-climatic and stratum parameters. An autoregressive linear model to describe the residual time series have been developed (Plastino and Bella, 2001; Box and Jenkins, 1976), and the presumed anomalies on the obtained white noise series were identified when the modulus was greater than three times the estimated standard deviation. In order to correlate the groundwater chemical–physical spike-like anomalies with seismic activity, we estimated the ε deformation parameter (Dobrovolsky et al., 1979) during the period from May 1996 to June 1999 in an area of 150 km from measurement site by the ING catalogue. We first selected seismic events and clusters assuming a lower limit for ε of 0.1 × 10−8 . The residual time series of the geochemical parameters and ε deformation parameter during the period from May 1996 to June 1999 are shown in Fig. 3. Fifteen spike-like anomalies in the series of electrical conductivity, twenty-six in the pH series and ten in 222 Rn series have been detected. Also, a comparison between earthquake time series (ε parameter) and alert period suggested by the occurrence of electrical conductivity, pH, and 222 Rn spikelike anomalies is reported in Fig. 3. The electrical conductivity spike-like anomalies are probably related to the two seismic clusters which occurred in October–November 1996 and August 1998, located northwest and west of the measurement site, respectively. Also the 222 Rn spike-like anomalies recorded in November 1996 appear to coincide with the former seismic sequence. These spike-like anomalies are co-seismic, and probably are due to change in groundwater dynamics, i.e. variations in the velocity of liquid-phase fluids, related to strain processes associated with earthquakes (Andrews, 1977). Also, the hydrological properties of the Gran Sasso–Sirente catchment basin, particularly the fluid flows in the Gran Sasso National Laboratory direction, can justify this phenomenology. The different polarities observed on these anomalies may be related to the different dips of the seismogenetic structures detected. Furthermore, two different trends on pH and 222 Rn, as a different chemical and physical changes of the Gran Sasso aquifer, have been emphasized. In fact, from September 1996 to March 1997, the pH increased and 222 Rn decreased. Then, from April 1997 to September 1997, coinciding with the foreshock activation stage including the Massa Martana earthquake, a pH buffering effect occurred without any long time variations on it and the 222 Rn concentration decreased again. This phenomenology due to the occurrence of transient compression phase, producing a change of the carbon dioxide content in groundwater from the beginning of January 1997 to September 1997. The carbon dioxide variations may have resulted in the pH increase, the 222 Rn decrease and the absence of electrical conductivity variations recorded before the Umbria–Marche seismic sequence, by changing the chemical properties of the CaCO3 –CO2 –H2 O system, and the dynamic transport properties of radon in groundwater (Toutain and Baubron, 1999). Finally, the pH and 222 Rn spike-like anomalies recorded during September 1997 are probably related to the Umbria–Marche main shock and emphasize pre-co-post seismic characteristics.
340
W. Plastino
Fig. 3. The residual time series of the electrical conductivity, pH, 222 Rn, and the ε deformation parameter (Dobrovolsky et al., 1979) during the period from May 1996 to June 1999. The estimated 3σ thresholds used to identify the anomalies, are also shown.
Monitoring of geochemical and geophysical parameters in the Gran Sasso aquifer
341
5. Conclusions The groundwater monitoring at Gran Sasso National Laboratory has shown a good reliability and stability of the automatic multiparametric equipment during all the monitoring period. The geochemical and geophysical parameters have pointed out a shallow aquifer behavior with high dynamics, typical for karst units. The data collected during three years were a useful tool to characterize the background in the groundwater monitoring parameters, and then to define the anomalies occurred on them and the possible link with deformation phenomena of tectonic interest. The time series analyses show strong anomalies in pH and radon concentrations, highly correlated with the known 1997–1998 Umbria–Marche seismic sequence, which occurred in the central Apennines, Italy. Furthermore, to better constrain these hypothesis we need to realize a groundwater monitoring network, that will better define the spatial-temporal variations of the geochemical and geophysical parameters and possible correlations with the deformation phenomena of tectonic interest.
Acknowledgement The author wish to thank Prof. Eugenio Coccia, Director of Gran Sasso National Laboratory, for his kind collaboration.
References Andrews, J.N. (1977). Radiogenic and inert gases in ground waters. In: Paquet, H. and Tardy, Y. (Eds.), Proc. 2nd Int. Symp. Water Rock Interaction. Strasbourg, France, pp. 334–342. Box, G.E.P., Jenkins, G.M. (1976). Time Series Analysis Forecasting and Control. Holden Day, Oakland, CA. Dobrovolsky, I.P., Zubkov, S.I., Miachkin, V.I. (1979). Estimation of the size of earthquake preparation zones. Pure and Applied Geophysics 117, 1025–1044. Igarashi, G., Saeki, S., Takahata, N., Sumikawa, K., Tasaka, S., Sasaki, Y., Takahashi, M., Sano, Y. (1995). Groundwater radon anomaly before the Kobe earthquake in Japan. Science 269, 60–61. Plastino, W., Bella, F. (2001). Radon groundwater monitoring at underground laboratories of Gran Sasso (Italy). Geophysical Research Letters 28, 2675–2678. Plastino, W., De Felice, P., De Notaristefani, F. (2002). Radon gamma-ray spectrometry with YAP:Ce scintillator. Nuclear Instruments and Methods in Physics Research A 486, 146–149. Roeloffs, E. (1999). Radon and rock deformation. Nature 399, 104–105. Shapiro, M.H., Rice, A., Mendenhall, M.H., Melvin, D., Tombrello, T.A. (1985). Recognition of environmentally caused variations in radon time series. Pure and Applied Geophysics 122, 309–326. Toutain, J.P., Baubron, J.C. (1999). Gas geochemistry and seismotectonics: A review. Tectonophysics 304, 1–27.
This page intentionally left blank
7. Groundwater–seawater interactions
This page intentionally left blank
345
Coastal water exchange rate studies at the southeastern Brazilian margin using Ra isotopes as tracers Joselene de Oliveiraa,* , Mathew Charetteb , Mathew Allenb , Elisabete de Santis Bragac , Valdenir Veronese Furtadoc a Divisão de Radiometria Ambiental, Centro de Metrologia das Radiações,
Instituto de Pesquisas Energéticas e Nucleares, São Paulo, Brazil b Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution,
Woods Hole, MA 02543, USA c Departamento de Oceanografia Química e Geológica, Instituto Oceanográfico da Universidade de São Paulo,
São Paulo, Brazil Abstract The São Paulo Bight is the arc-shaped part of the southeastern Brazilian margin extending from 23◦ S to 28◦ S. To assess the cross-shelf Ra distributions in São Paulo Bight, four shore-perpendicular profiles were collected up to 100 km offshore from 23 to 26 February 2003 (summer). All samples studied here were taken in the selected area between latitudes 23◦ 15 S and 25◦ 50 S and longitudes 44◦ W and 46◦ W, in order to estimate coastal mixing rates and groundwater discharge fluxes. The activity concentrations of 223 Ra in surface seawater varied from 0.002 to 0.4 mBq L−1 , 224 Ra in excess from 0.02 to 2.5 mBq L−1 , 226 Ra from 1.2 to 1.8 mBq L−1 and 228 Ra from 0.4 to 4.4 mBq L−1 . The 223 Ra/224 Ra and 228 Ra/226 Ra activity ratios observed in seawater samples ranged from 0.03 to 2.6 and from 0.3 to 2.4, respectively. These results seem to indicate that Ra isotopes from 232 Th series prevail in a major number of samples, when compared with Ra isotopes from 238 U and 235 U series. Considering the results obtained in the summer 2003, shore-perpendicular profiles of 223 Ra and 224 Ra in surface waters along the coast were modeled to yield eddy diffusion coefficients. Keywords: Ra isotopes, Marine environment, Seawater, Coastal water exchange rates, São Paulo Bight, Brazil
1. Introduction The fate of contaminants and natural compounds in estuaries and coastal waters is determined by a set of biological, geochemical and physical interactions. Although scientists have a basic understanding of the major sources, sinks, and transformations for many substances, to assess * Corresponding author. Address: Instituto de Pesquisas Energéticas e Nucleares, Av. Lineu Prestes, 2242, Cidade Universitária, São Paulo, SP, CEP05508-000, Brasil; phone: (+55) 11 3816 9293; fax: (+55) 11 3816 9118; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08028-9
© 2006 Elsevier Ltd. All rights reserved.
346
J. de Oliveira et al.
offshore fluxes of dissolved materials we need to know coastal water residence times. Until today to quantify residence times in coastal areas remains a difficult task, since only a few methodologies are available to assess water exchange in this dynamic region, where currents, waves, tides, river flow and groundwater discharge usually play together a complex role. The knowledge of the four natural Ra isotopes in aqueous systems can be used to constrain important environmental processes occurring at the land–sea margin. The high variability of the 226 Ra/228 Ra activity ratio in waters, as has been reported by several authors (Moore and Arnold, 1996; Rama and Moore, 1996; Charette et al., 2001), clearly demonstrates the suitability of Ra isotope relationships to determine fluxes and mixing rates of continental waters into ocean and estuaries, and exchanges between ground-water and surface water. Two main geochemical characteristics control the production and input of Ra isotopes in coastal areas: the existence of particle-reactive Th isotopes in sediments as direct radiogenic parents, and the vastly different environmental behavior of Ra in fresh water and salt water media. The short-lived Ra isotopes of the 232 Th series, which are highly depleted in the ocean basins due to their rapid decay and the strong depletion in parent Th isotopes, have been used to track advection from the coasts. Because of the 5.7 year half-life of 228 Ra, this isotope has been used effectively to estimate oceanic horizontal eddy diffusivities and coastal water residence times over timescales of less than 30 years. By using both 228 Ra and 224 Ra (t1/2 = 3.7 days), timescales of less than 10 years can be investigated. The cycling of Ra in the oceans can be considered as the most interesting phase of radium geochemistry. The fact that in the oceans 226 Ra and 228 Ra exist in excess of their respective parents, 230 Th and 232 Th (more than tenfold in excess for 226 Ra), led some authors to hypothesize that the excess radium was being supplied by diffusion from deep sea sediments. This hypothesis was later confirmed by deficiencies of the radium isotopes observed in deep sea sediments. The potential of these isotopes for studying oceanic circulation was recognized because they are the only ones among other natural tracers which are injected directly into bottom water from underlying sediments. The first measurements of radium in the ocean water were made due not so much to its radiological significance, but to the promising role of oceanic radium as a tool for understanding marine geochemical processes. It was Koczy (1958) who first proposed using Ra as a natural tracer in the ocean for the calculation of vertical diffusion coefficients in the water column. This was extended later by Koczy and Szabo (Koczy and Szabo, 1962) for estimating the renewal time of water masses in the Pacific and Indian Oceans. Numerous investigations have been conducted since then to obtain a fuller understanding of 226 Ra distributions in the major oceans of the world, with considerable attention devoted to the variability of 226 Ra concentrations with regard to depth and latitude, correlation with barium, silica concentrations, salinity, etc. (Broecker et al., 1967). 226 Ra concentrations in surface seawater (0–500 m) appear to be in a narrow range, and nearly uniform in the Pacific (0.7–3.7 mBq L−1 ) and Atlantic Oceans (0.7–3.0 mBq L−1 ). The Indian Ocean has levels of 226 Ra which cover a narrower range (1.1–2.2 mBq L−1 ), however. At the increasing depths, a trend of increasing concentrations is observed uniformly in all of the oceans, which, according to most investigators, results from the injection of 226 Ra from thorium (230 Th) bearing sediments in the ocean floor (IAEA, 1990; Moore, 1969). Reports of an interesting correlation of silica and barium concentrations with those of Ra in collateral seawater samples have been made by some researchers, calling attention to the
Coastal water exchange rate studies at the southeastern Brazilian margin
347
parallel geochemical behavior among these elements. It appears that radium behavior in the marine environment closely parallels that of barium, both of which are carried by marine diatoms, which are highly siliceous organisms (Moore, 1999). 226 Ra concentrations in coastal waters are slightly higher relative to open ocean water. Radium is released by coastal sediments, which form an important source of 226 Ra migration to the ocean. This was corroborated by measurements in South Carolina (USA) performed by Elsinger and Moore (1980) who concluded that a desorption mechanism can quantitatively explain the increase of 226 Ra in brackish water. Moore (1981) in his investigations at Chesapeake Bay, demonstrated the clear influence of salinity on the desorption of 226 Ra from particulates: at salinities lower than 0.5h, about 12% of 226 Ra in the water is in the soluble phase, while above 5h salinity, over 80% is in the dissolved phase. The spread of 226 Ra in coastal waters is more or less uniform (0.7–6.6 mBq L−1 ), except for estuarine and shelf water in west central Florida. There the 226 Ra levels can vary from 1.8 to 54 mBq L−1 for surface water and from 1.8 to 22 mBq L−1 for deep water (IAEA, 1990). These high concentrations are attributed to the geology of that area, known for its rich phosphate deposits, seepage of groundwater and to the existence of active geothermal springs. It is generally observed that nearshore water, particularly from restricted bays and sounds, has higher 228 Ra levels than does open ocean surface water. However, these concentrations tend to become lower as one moves away from the coast, thus showing a close correlation between concentration and proximity to land (coastal activities ranging from 0.7 to 11.5 mBq L−1 ). Some nearshore Indian coastal waters displays significant 228 Ra levels (13.7–38 mBq L−1 ) due to significant occurrence of monazite sands (IAEA, 1990). Moore (1969) has discussed the use of 228 Ra in the oceans as a natural tracer for studying marine processes occurring within a 3–30 year timescale. He has also described the applicability of 228 Ra/226 Ra activity ratios as tracers for studying lateral and vertical movements within the ocean. Nearshore water, such as surface water close to continents and coastal water in contact with terrigenous sediments with limited circulation with open ocean, appear to have very high 228 Ra/226 Ra activity ratios. For example, some of the highest 228 Ra/226 Ra ratios (7.1) have been measured in the brackish water of Mississippi Sound (IAEA, 1990). Similar enhanced ratios have also been observed in Chesapeake Bay (1.8); Long Island Sound (1.4); Davis Bay, Mississippi (2.5); Narragansett Bay, Rhode Island (2.0); Wellington Harbour, New Zealand (1.7); Penang Harbour, Malaysia (2.0) and few other places (Moore, 1969). When nearshore water mixes with oceanic water, the 228 Ra/226 Ra ratio goes down. However, within surface water, large variations in this activity ratio have been observed by Moore (1999). Atlantic surface water generally has higher ratios than Pacific surface water. For example, the Atlantic has activity ratios of 0.09–2.41 and the Pacific 0.01–0.2. According to Moore, the 228 Ra/226 Ra ratio is determined by a balance between the supply of 228 Ra to the water body and the lateral and vertical mixing rates. The activity ratios in the Atlantic and Pacific Oceans indicate that the mixing rate of the surface water is considerably longer than ten year mean life of 228 Ra. The 228 Ra/226 Ra activity ratios in the surface water of the Mediterranean, Caribbean and Black Seas are all less than unity and are also in a close range. Inversely, the surface water of the Indonesian seas show typically enhanced ratios, from 0.26 to 3.8, which is due to considerable diffusion of 228 Ra from the continental shelves. Other regions such as New York Bight (0.39–1.99), Chesapeake Bay (1.17–4.08), Kalpakkam, India (2.4–4.9) and Bombay, India
348
J. de Oliveira et al.
(1.6–1.9) further support the fact that estuarine and coastal water is rich in 228 Ra owing to diffusion from nearshore and estuarine sediments, thereby leading to enhanced 228 Ra/226 Ra ratios, as observed. Moore (1981), following his studies in Chesapeake Bay on 226 Ra and 228 Ra flux rates and comparing his data with those from other bay and estuarine regions, has proposed that the flux rate of 228 Ra from the sediments should be greater than that of 226 Ra, owing to the faster growth of 228 Ra from its parent 232 Th, and the effects of bioturbation.
2. Materials and methods The São Paulo Bight is the arc-shaped part of the southeastern Brazilian margin extending from 23◦ S to 28◦ S. To assess the cross-shelf Ra distributions in São Paulo Bight, four shoreperpendicular profiles were collected up to 100 km offshore from 23 to 26 February 2003 (summer), on board the R/V Prof. Besnard. All samples studied here were taken in the selected area between latitudes 23◦ 15 S and 25◦ 50 S and longitudes 44◦ W and 46◦ W, in order to estimate coastal mixing rates and ground water discharge fluxes (Fig. 1). This region is considered a tropical coastal area and the main geologic/geomorphologic feature is the presence of granites and migmatites of a mountain chain locally called Serra do Mar (altitudes up to
Fig. 1. Location of the seawater sampling stations, collected on the São Paulo State continental margin, Southeastern Brazil. Ilha Grande Bay horizontal profile is represented by IG, Ubatuba by UB, São Sebastião by SS and Santos by SB.
Coastal water exchange rate studies at the southeastern Brazilian margin
349
1,000 meters), which reaches the shore in almost all of the study area, and limits the extension of the drainage systems and of Quaternary coastal plains. Wave action is the most effective hydrodynamic phenomenon responsible for the bottom sedimentary processes in the coastal area as well as in the adjacent inner continental shelf. The terrestrial input of sediments is strongly dependent on the rainfall regime, leading to a higher contribution of sediments during summer season. During the summer, the advance of the South Atlantic Central Water (SACW) over the coast leads to the displacement of the Coastal Water (CW), rich in continental suspended materials, and to the transportation of these sediments to the outer portions of the continental shelf. During winter, the retreat of the SACW and the decreasing of the rainy levels restrict the input of sediments from the continental areas. The mean annual rainfall is roughly 1,803 mm, the maximum rainfall rates being observed in February. Sea level varies from 0.5 to 1.5 m, the highest values occurring in months August/September due to greater volume of warm waters of Brazil Current (Mesquita, 1997). For the purposes of pre-concentration of Ra isotopes from large volume of seawater samples described in this paper, Acrylic fiber (Cia. Sudamericana do Brasil, 3.0 denier) was treated with a hot solution of saturated KMnO4 for approximately 10 minutes. The KMnO4 oxidizes specific sites on the acrylic molecule and deposits MnO2 at these sites. The prepared fiber was washed with purified water free of radium and was kept in plastic bags for use. This produces Mn fiber having sub-micrometric sized particles of MnO2 chemically bonded to the fiber. The MnO2 constitutes 8–10% by mass of the Mn fiber. This procedure was conducted in a 5 L beaker scale. Large volume seawater samples (196 L) were pumped from 5 m below the surface into plastic drums on the R/V Prof. Besnard. The sample volume was recorded and the seawater was percolated through a column of manganese coated acrylic fiber to quantitatively remove radium from seawater (Moore, 1996). Temperature and salinity profiles were obtained at each station using a CTD, from Seabird Scientific Inc. Samples for dissolved oxygen, salinity and nutrients were also collected in each station. Water samples for nutrients were frozen until the time of analysis. The analytical procedures adopted for these determinations were vanadium reduction followed by chemiluminescence detection of NOx for nitrate–nitrite, phenate method for ammonia and ascorbic acid method for phosphate. These measurements were performed at the Nutrients Laboratory (LABNUT), from the Oceanographic Institute of University of São Paulo. In an onshore laboratory each Mn fiber sample was partially dried with a stream of air and placed in an air circulation system described by Moore and Arnold (1996). Helium was circulated over the Mn fiber to sweep 219 Rn and 220 Rn generated by 224 Ra and 223 Ra decay in a 1.1 L scintillation cell. The alpha particles from the decay of radon and its daughters were recorded by a photomultiplier tube (PMT) attached to the scintillation cell described previously. Signals from the PMT were routed to a delayed coincident counter system adapted for Ra measurements (Moore and Arnold, 1996). The delayed coincidence system utilizes the difference in decay constants of the short-lived Po daughters of 219 Rn and 220 Rn to identify alpha particles derived from 223 Ra and 224 Ra captured on the Mn fiber. The expected error of the short-lived Ra isotope measurement is 10%. After completing the 224 Ra and 223 Ra measurements, the Mn fiber samples were aged for 5 weeks to allow excess 224 Ra to equilibrate with 228 Th, also adsorbed to the Mn fiber. The
350
J. de Oliveira et al.
samples were measured again to determine 228 Th activity and this value was used to correct the 224 Ra activity to its unsupported activity. Following these analyses, the Mn fiber was leached in a beaker with 200 mL HCl under controlled heating, to quantitatively remove the longer-lived Ra isotopes. For the radiochemical separation of 226 Ra and 228 Ra, carriers of stable barium (20 mg) and lead (20 mg) were added to the water sample in the presence of 5 mL of 1 M citric acid and 5 mL of 40% hydroxylamine hydrochloride solutions. The radium was co-precipitated as BaPb(Ra)SO4 by adding 50 mL of 3 M H2 SO4 . The precipitate was dissolved with alkaline EDTA. When the pH is adjusted to 4.5 with glacial acetic acid, Ba(Ra)SO4 is re-precipitated, while interfering elements remain in the solution. The Ba(Ra)SO4 precipitate was transferred to a 2 mL polypropylene tube and sealed to avoid the escape of 222 Rn. 226 Ra and 228 Ra were measured by gamma spectrometry of a Ba(Ra)SO4 precipitate in a WeGe well germanium detector, after 21 days from the precipitation. The detector is a 200 cm3 coaxial intrinsic germanium crystal, from Canberra, model GCW4023, with relative efficiency of 40%, well diameter of 16 mm and depth of 40 mm (near 4π counting geometry), and typical resolutions of 1.4 keV and of 2.3 keV at 122 keV and 1332 keV, respectively. The standard configuration includes a vertical slimline cryostat with 30 liter Dewar, HV power supply and a preamplifier with 3 meter bias, model 2002C. The 226 Ra activities were determined (after >20 days sample storage) by taking the mean activity of three separate photopeaks of its daughter nuclides: 214 Pb at 295.2 keV and 351.9 keV, and 214 Bi at 609.3 keV. The 228 Ra content of the samples was determined from the 911 keV and 968 keV gamma-ray peaks of 228 Ac. Both measurements were performed at the Marine Chemistry and Geochemistry Laboratory of Woods Hole Oceanographic Institution. 2.1. Radium isotope disequilibrium to delineate coastal dynamics and groundwater input In the natural radioactive series, there are four radium isotopes: 226 Ra (t1/2 = 1620 years); 223 Ra (t 224 Ra (t 1/2 = 5.75 years); 1/2 = 11.3 days); 1/2 = 3.66 days). Each isotope is 230 produced from the decay of a thorium parent: Th (t1/2 = 7.54 × 104 years); 232 Th (t1/2 = 1.40 × 1010 years); 227 Th (t1/2 = 18.7 days); 228 Th (t1/2 = 1.91 years), respectively. Because thorium remains tightly bound to particles while radium daughters are mobilized into the marine environment, sediments provide a continuous source of Ra isotopes to seawater, at rates set by their respective decay constants. Measurements of the Th isotope activities in the sediments and the distribution coefficient of Ra between the sediments and water provide a means of quantifying the potential input of each isotope to the ocean. Two short-lived radium isotopes 223 Ra and 224 Ra can be used as tracers to measure the rate of exchange of coastal waters (Moore, 1998). Shore-perpendicular profiles of 223 Ra and 224 Ra in surface waters along the coast may be modeled to yield eddy diffusion coefficients. Coupling the exchange rate with offshore concentration gradients, the offshore fluxes of dissolved materials are estimated. For systems in steady-state, the offshore fluxes must be balanced by new inputs from rivers, groundwater, sewers or other sources. Also, it was observed that barium and 226 Ra contents can be powerful indicators of groundwater input in marine systems, since they have high relative concentrations in the fluids and low reactivity in the coastal ocean. An estimate of the 226 Ra offshore flux is made applying the eddy diffusion coefficients to the 226 Ra offshore gradient. Complementary data of 226 Ra in subsurface fluids provides a 228 Ra (t
Coastal water exchange rate studies at the southeastern Brazilian margin
351
mean of calculate the fluid flux necessary to support the 226 Ra concentrations found in the marine environment. The Ra distribution may be expressed by a simple one-dimensional horizontal diffusion model, in which the distribution is in balance between eddy diffusion and radioactive decay (Moore, 1998) ∂ 2A dA = Kh 2 − λA. dt ∂x At the steady-state, this expression can be written as the following: λ , Ax = A0 exp −x Kh
(1)
(2)
where Ax is the activity at distance x from coast, A0 is the activity at distance zero from coast, λ is the decay constant and Kh is the horizontal eddy diffusion coefficient. Based on the activity ratio of the short-lived Ra isotopes, ages or coastal residence times can be estimated using the following equation: 224 −λ224 t 224 Ra Ra e = 223 , (3) 223 Ra Ra i e−λ223 t obs where [224 Ra/223 Ra]i is the initial activity ratio measured in nearshore water (at distance zero from coast) and [224 Ra/223 Ra]obs is the activity ratio measured at the distance x from coast, respectively; λ224 Ra is the decay constant for 224 Ra (0.189 d−1 ) and λ223 Ra is the decay constant for 223 Ra (0.06 d−1 ).
3. Results and discussion The location of all stations studied as well as the results of the cruise carried out in February 2003 are shown in Tables 1–4. During the period of this investigation, the activity concentrations of 223 Ra in surface seawater varied from 0.002 to 0.41 mBq L−1 , 224 Ra in excess from 0.02 to 2.53 mBq L−1 , 226 Ra from 1.22 to 1.81 mBq L−1 and 228 Ra from 3.9 Table 1 Location of Ra isotopes transect collected at Ilha Grande Bay (23/February/03) – R/V Prof. Besnard cruise Sample/ Distance Latitude Station offshore (km) IG-1 IG-2 IG-3 IG-4 IG-5 IG-6 IG-7 IG-8
5 10 25 40 55 70 85 100
S23◦ 05.073 S23◦ 05.076 S23◦ 14.483 S23◦ 19.358 S23◦ 24.283 S23◦ 33.763 S23◦ 44.315 S23◦ 44.310
Longitude
Time Volume T (◦ C) (L)
Water column Sampling Salinity depth (m) depth (m)
W44◦ 27.648 W44◦ 27.645 W44◦ 24.823 W44◦ 22.937 W44◦ 21.686 W44◦ 18.550 W44◦ 15.288 W44◦ 15.284
08:51 11:05 12:54 14:08 15:31 17:24 19:15 21:50
20 30 34 40 49 73 90 120
196 196 196 196 196 196 196 196
28.7 28.7 28.0 28.0 26.0 28.0 28.0 27.0
5 5 5 2 5 2 3 5
34.5557 34.5703 34.6386 34.7264 35.0016 – 36.1107 36.4020
352
J. de Oliveira et al.
Table 2 Location of Ra isotopes transect collected at Ubatuba Bay (24/February/03) – R/V Prof. Besnard cruise Sample/ Distance Latitude Station offshore (km) UB-1 UB-2 UB-3 UB-4 UB-5 UB-6 UB-7 UB-8
100 85 70 55 40 25 10 5
S24◦ 12.976 S23◦ 04.470 S24◦ 04.467 S23◦ 47.764 S23◦ 44.100 S23◦ 40.398 S23◦ 36.738 S23◦ 33.105
Longitude
Time Volume T (◦ C) (L)
Water column Sampling Salinity depth (m) depth (m)
W44◦ 30.513 W44◦ 38.664 W44◦ 38.667 W44◦ 54.600 W44◦ 58.186 W45◦ 01.810 W45◦ 05.658 W45◦ 08.811
03:12 06:13 08:39 10:23 11:43 13:07 14:27 15:45
130 130 77 56 47 36 29 12.5
196 196 196 196 196 196 196 196
23.2 24.0 28.0 28.0 28.0 28.5 28.0 29.0
5 5 5 5 5 5 5 5
36.5313 36.2379 34.6841 34.3567 34.6022 34.5027 34.5735 34.8372
Table 3 Location of Ra isotopes transect collected surrounding São Sebastião (25/February/03) – R/V Prof. Besnard cruise Sample/ Distance Latitude Station offshore (km) SS-1 SS-2 SS-3 SS-4 SS-5 SS-6 SS-7 SS-8 SS-9
5 10 25 40 55 70 85 100 125
S24◦ 12.976 S23◦ 04.470 S24◦ 04.467 S23◦ 47.764 S23◦ 44.100 S23◦ 40.398 S23◦ 36.738 S23◦ 33.105 S23◦ 33.105
Longitude
Time Volume T (◦ C) (L)
Water column Sampling Salinity depth (m) depth (m)
W44◦ 30.513 W44◦ 38.664 W44◦ 38.667 W44◦ 54.600 W44◦ 58.186 W45◦ 01.810 W45◦ 05.658 W45◦ 08.811 W45◦ 08.811
03:12 06:13 08:39 10:23 11:43 13:07 14:27 15:45 15:45
22 35.7 48 49 58 71.3 79 90 116
196 196 196 196 196 196 196 196 196
29.0 28.5 28.0 28.0 28.0 28.0 27.5 27.5 27.5
5 5 5 5 5 5 5 5 5
33.8869 34.0712 33.7698 33.9977 33.8463 34.0948 34.5466 35.5124 36.5842
Table 4 Location of Ra isotopes transect collected at Santos Bay (26/February/03) – R/V Prof. Besnard cruise Sample/ Distance Latitude Station offshore (km) SB-1 SB-2 SB-3 SB-4 SB-5 SB-6 SB-7 SB-8 SB-9 SB-10 SB-11 SB-12
115 100 85 70 55 40 25 10 8 6 4 2
S25◦ 20.008 S25◦ 10.613 S25◦ 00.694 S24◦ 51.636 S24◦ 42.398 S24◦ 32.841 S24◦ 23.589 S24◦ 18.785 S24◦ 13.922 S24◦ 09.302 S24◦ 04.912 S24◦ 00.013
Longitude
Time Volume T (◦ C) (L)
Water column Sampling Salinity depth (m) depth (m)
W45◦ 50.389 W45◦ 53.742 W45◦ 57.083 W46◦ 00.639 W46◦ 04.216 W46◦ 07.578 W46◦ 11.183 W46◦ 12.927 W46◦ 14.457 W46◦ 16.471 W46◦ 18.212 W46◦ 19.852
01:17 04:04 06:21 08:16 10:01 11:42 13:22 14:37 15:55 16:50 17:40 18:41
118 90 84 71 63 57 44 36.6 34 30 21 11
196 196 196 196 196 196 196 196 196 196 196 196
27.5 27.5 28.0 28.5 28.5 29.0 28.5 28.0 29.0 30.0 30.0 30.0
5 5 5 5 5 5 3 5 5 5 2 2
35.7475 35.7096 35.4429 35.1437 34.6412 34.6473 33.5494 33.8592 34.0345 32.8669 32.3302 31.7591
Coastal water exchange rate studies at the southeastern Brazilian margin
353
to 4.42 mBq L−1 . The 223 Ra/224 Ra activity ratios observed in seawater samples ranged from 0.03 to 2.6, whereas 228 Ra/226 Ra activity ratios varied in the interval from 0.3 to 2.4. These results seems to indicate that Ra isotopes from 232 Th series prevail in a major number of samples, when compared with Ra isotopes from 238 U and 235 U series. The 226 Ra activity concentrations found in surface seawater samples studied at the southeastern Brazilian margin are of the same order of magnitude than those reported by other author in the southeastern coast of United States (typical values from 1.33 to 2.83 mBq L−1 ) (Moore, 1999). The 226 Ra distribution in surface seawater samples from all transects showed a narrow range along the coast. Since the half-life of 226 Ra (1620 yr) is comparable to the mean ocean circulation time established from 750 to 1000 yr (Broecker and Peng, 1982), the 226 Ra should be very well mixed in seawater (uniform concentrations along the coast). Deviation of this behavior is expected to occur only close to the margins or the bottom (in this case one positive 226 Ra signal could be used to identify a groundwater input). In the case of 228 Ra, as it has the half-life of 5.7 yr, the activities are higher close to the margins, decreasing with distance from shore. In coastal areas the short-lived Ra isotopes, 223 Ra and 224 Ra, are flushed from the sediments and are regenerated from their thorium parents in sediments on a time scale of days. This provides a continuous source of 223 Ra and 224 Ra activity to the overlying seawater that is not accompanied by large additional 226 Ra and 228 Ra, which are regenerated more slowly. Groundwater discharge directly into sea water may also provide significant additions of Ra activity to the estuaries and ocean shelf bottom water (Moore, 1996; Rama and Moore, 1996). In Figs. 2–5 we show the distributions of 223 Ra, 224 Ra, 226 Ra and 228 Ra along all the four horizontal profiles established in Ilha Grande, Ubatuba, São Sebastião and Santos Bay. These horizontal profiles were sampled from 5 to 100 km offshore. Concentrations of dissolved oxygen, nitrate, silicate phosphate, ammonium and salinity determined in each profile are also presented in those figures.
(a)
(b)
Fig. 2. (a) Activity concentration of the natural Ra isotopes determined in samples, (b) concentrations of dissolved oxygen (DO), nitrate, silicate, ammonium, phosphate and salinity of samples collected at Ilha Grande Bay transect.
354
J. de Oliveira et al.
(a)
(b)
Fig. 3. (a) Activity concentration of the natural Ra isotopes determined in samples, (b) concentrations of dissolved oxygen (DO), nitrate, silicate, ammonium, phosphate and salinity of samples collected at Ubatuba transect.
(a)
(b)
Fig. 4. (a) Activity concentration of the natural Ra isotopes determined in samples, (b) concentrations of dissolved oxygen (DO), nitrate, silicate, ammonium, phosphate and salinity of samples collected at São Sebastião transect.
(a)
(b)
Fig. 5. (a) Activity concentration of the natural Ra isotopes determined in samples, (b) concentrations of dissolved oxygen (DO), nitrate, silicate, ammonium, phosphate and salinity of samples collected at Santos Bay transect.
Coastal water exchange rate studies at the southeastern Brazilian margin
355
In Figs. 6 and 7 we show the Ra ages of the samples collected at each horizontal profile. These ages were estimated using Equation (3), described in Section 2.1, which provides a technique for measuring coastal residence times, over time scales of just few days. This expression takes into account the variation of the short-lived Ra isotopes activity ratio (223 Ra/224 Ra) along the distance offshore (km). One assumption required to validate this model is that the 223 Ra/224 Ra activity ratio must be constant in each of the source terms (i.e. input nearshore/ocean/seabed) at least over the period of time of this study. The short halflives of 223 Ra and 224 Ra likely satisfy this assumption, but in situ time series experiments
(a)
(b)
Fig. 6. Ra ages of the samples collected at (a) Ilha Grande and (b) Ubatuba transects, calculated using the short-lived isotopes activity ratio versus distance offshore (km).
(a)
(b)
Fig. 7. Ra ages of the samples collected at (a) São Sebastião and (b) Santos Bay transects, calculated using the short-lived isotopes activity ratio versus distance offshore (km).
356
J. de Oliveira et al.
are needed to confirm this. 223 Ra and 224 Ra isotopic data used to derive the age dates were already presented in Tables 5–8. According these results, ages for Ilha Grande transect varied from 4 days (at a distance zero from the coast) to about 14 days (100 km offshore). As for São Sebastião transect, ages estimated varied from 9 days to 20 days and for Santos Bay from 17 days to 35 days. Apparent ages calculated by this method reflect the time elapsed since water became enriched in Ra and was isolated from the source. These calculations assumed that 100% of the initial concentration of the Ra isotope present in sediments in the near shore region was transferred to the seawater (end-member fraction = 1). These ages imply that exchange times of the coastal waters across the regions studied are rapid, of the order of 23 days in average, during the period of investigation. In the case of Ubatuba transect, the ages plot evidenced that the water mass studied was not moving horizontally by advection offshore, since older waters were found closest to the shore, ages decreasing with distance offshore. This can indicate for example, that there is lateral movement of the water mass (it is very common the passage of ocean currents parallel to the coastline in this area during the summer season, causing displacement of the younger coastal water mass). Table 5 Concentration of the natural Ra isotopes and nutrients determined in the samples from Ilha Grande Bay transect 224 Ra 226 Ra 228 Ra Sample/ 223 Ra DO Nitrate Silicate Ammonium Phosphate Station (mBq/100 L) (mBq/100 L) (mBq/100 L) (mBq/100 L) (mL/L) (µmol/L) (µmol/L) (µmol/L) (µmol/L)
IG-1 IG-2 IG-3 IG-4 IG-5 IG-6 IG-7 IG-8
0.85 2.9 2.1 2.4 0.52 1.5 1.0 0.69
24.8 15.7 29.8 23.4 8.4 5.1 8.5 3.5
129 127 148 154 144 140 137 128
234 196 197 231 137 103 62.0 38.7
4.69 4.15 4.63 4.74 4.97 5.08 4.85 4.78
0.1 0 0.01 0.13 0.24 0.1 0.05 5.29
1.46 1.04 1.67 1.43 14.30 2.57 4.64 4.19
1.43 1.58 1.41 2.58 1.44 1.31 0.74 0.93
0.35 0.30 0.26 0.33 0.28 0.26 0.23 0.13
Table 6 Concentration of the natural Ra isotopes and nutrients determined in the samples from Ubatuba Bay transect 224 Ra 226 Ra 228 Ra Sample/ 223 Ra DO Nitrate Silicate Ammonium Phosphate Station (mBq/100 L) (mBq/100 L) (mBq/100 L) (mBq/100 L) (mL/L) (µmol/L) (µmol/L) (µmol/L) (µmol/L)
UB-1 UB-2 UB-3 UB-4 UB-5 UB-6 UB-7 UB-8
0.25 0.20 1.2 4.9 27.2 30.4 0.48 3.3
6.3 5.8 6.1 17.2 22.0 – 22.8 54.5
136 143 135 139 148 138 150 151
45.3 42.2 176 226 197 177 233 235
3.82 4.87 4.76 4.70 4.71 4.68 4.80
0.05 0.04 0.18 0.14 0.23 0.02 0.01
0.56 0.69 2.98 0.83 1.75 1.08 4.05
1.09 1.07 1.28 0.88 1.18 0.77 1.30
0.08 0.09 0.10 0.15 0.57 0.13 0.10
Coastal water exchange rate studies at the southeastern Brazilian margin
357
Table 7 Concentration of the natural Ra isotopes and nutrients determined in the samples from São Sebastião transect 224 Ra 226 Ra 228 Ra Sample/ 223 Ra DO Nitrate Silicate Ammonium Phosphate Station (mBq/100 L) (mBq/100 L) (mBq/100 L) (mBq/100 L) (mL/L) (µmol/L) (µmol/L) (µmol/L) (µmol/L)
SS-1 SS-2 SS-3 SS-4 SS-5 SS-6 SS-7 SS-8 SS-9
5.1 41.2 9.1 13.5 2.2 2.4 1.7 2.2 –
28.9 27.9 51.4 28.2 19.1 6.5 6.8 3.4 –
144 146 149 140 148 131 123 – –
287 265 273 292 298 153 65.4 – –
2.97 4.62 4.64 4.66 4.67 4.63 4.69 3.86 3.22
0.01 0.18 0.01 0.00 0.03 0.02 0.01 0.00 0.09
3.22 4.14 9.71 4.99 5.81 1.59 3.76 0.01 0.33
0.94 1.84 1.84 1.47 18.22 0.04 1.24 2.63 1.09
0.18 0.28 0.23 0.28 0.33 0.12 0.06 0.11 0.10
Table 8 Concentration of the natural Ra isotopes and nutrients determined in the samples from Santos Bay transect 224 Ra 226 Ra 228 Ra Sample/ 223 Ra DO Nitrate Silicate Ammonium Phosphate Station (mBq/100 L) (mBq/100 L) (mBq/100 L) (mBq/100 L) (mL/L) (µmol/L) (µmol/L) (µmol/L) (µmol/L)
SB-1 SB-2 SB-3 SB-4 SB-5 SB-6 SB-7 SB-8 SB-9 SB-10 SB-11 SB-12
0.56 5.3 11.6 – 2.1 10.1 27.5 – – 27.5 27.3 16.3
3.6 6.1 – 3.2 5.1 3.9 – 9.8 4.7 43 26.4 253
123 139 149 133 124 143 145 167 137 162 168 181
75.9 74.3 111 120 134 142 177 254 204 314 333 442
3.40 4.40 4.10 4.57 4.67 4.67 – 4.84 4.90 4.59 – –
0.10 0.02 0.07 0.09 0.03 0.06 – 0.02 0.05 0.00 – –
0.00 6.30 8.28 5.80 1.96 3.58 – 1.38 1.06 1.71 – –
1.90 1.59 3.28 26.05 23.63 12.44 – 20.00 6.48 1.41 – –
0.18 0.14 0.17 0.10 0.18 0.16 – 0.13 0.23 0.15 – –
In Figs. 8 and 9 we show the distribution of the natural logarithm of the short-lived Ra isotopes activity ratio versus distance offshore in Ilha Grande, Ubatuba, São Sebastião and Santos Bay transects. Analysis of the short-lived Ra activity ratios trends resulted in mixing coefficients (Kh ) of about 5.71 km2 day−1 for Ilha Grande, 9.46 km2 day−1 for Ubatuba, 6.0 km2 day−1 for São Sebastião and 8.6 km2 day−1 for Santos Bay, respectively.
4. Conclusions The application of the four naturally occurring Ra isotopes and a one-dimensional advection– diffusion model was shown as a tool to assist in the interpretation of coastal ocean circulation at the southeastern Brazilian margin. Since they do not require steady-state conditions with respect to mixing, this isotopic technique can supply useful data which coupled with salinity
358
J. de Oliveira et al.
(a)
(b)
Fig. 8. Distribution of the natural logarithm of the short-lived Ra isotopes activity ratio versus distance offshore in (a) Ilha Grande Bay and (b) Ubatuba transects.
(a)
(b)
Fig. 9. Distribution of the natural logarithm of the short-lived Ra isotopes activity ratio versus distance offshore in (a) São Sebastião and (b) Santos Bay transects.
or any other tracer distributions, provide powerful constraints to follow the circulation patterns and calculate fluxes of several dissolved materials to the ocean. Obviously, additional work during different conditions shall be carried out to estimate average exchange times and seasonal variations. As it was indicated and quantified in a previous research work, carried out in Ubatuba using 222 Rn as a tracer, that there is a significant inflow of subsurface fluids at rates in excess of several cm per day in the same embayments studied here (Oliveira et al., 2003), we intend to use the Ra data set and the residence times obtained to perform a mass balance (integrating river and groundwater end-member concentrations) to quantify the groundwater input for the same area.
Coastal water exchange rate studies at the southeastern Brazilian margin
359
Acknowledgements This research work was supported by International Atomic Energy Agency, Research Contract No. 12151, and by Fundação de Amparo à Pesquisa no Estado de São Paulo, Process No. 2002/08154-9. We also would like to thank the crew of R/V Prof. Besnard, who assisted during the sample collection and processing.
References Broecker, W.S., Li, Y.H., Cromwell, J. (1967). Radium-226 and radon-222 concentration in the Atlantic and Pacific Oceans. Science 158, 1307–1310. Broecker, W.S., Peng, T.H. (1982). Tracers in the Sea. Lamonh-Doherty Geological Observatory, Palisades, NY, 690 pp. Charette, M.A., Buesseler, K.O., Andrews, J.E. (2001). Utility of radium isotopes for evaluating the input and transport of groundwater derived nitrogen to a Cape Cod estuary. Limnology and Oceanography 46, 465–470. Elsinger, R.J., Moore, W.S. (1980). Radium-226 behaviour in the Pee Dee River–Winyah Bay estuary. Earth and Planetary Science Letters 48, 239–249. International Atomic Energy Agency, IAEA (1990). The environmental behaviour of radium. Technical Reports Series No. 310, vol. 1, Part 4, pp. 419–450. Koczy, F.F. (1958). Natural radium as a tracer in the ocean. In: Peaceful Uses of Atomic Energy, vol. 1. Proc. 2nd Int. Conf., Geneva, 1958. United Nations, Geneva. Koczy, F.F., Szabo, B.J. (1962). Renewal time of bottom water in the Pacific and Indian Oceans. Journal of the Oceanographical Society of Japan 20th Anniversary volume, 590–599. Mesquita, A.R. (1997). Marés, circulação e nível do mar na Costa Sudeste do Brasil. Relatório Fundespa, São Paulo. Moore, W.S. (1969). Oceanic concentrations of radium-228. Earth and Planetary Science Letters 6 (2), 437–446. Moore, W.S. (1981). Radium isotopes in the Chesapeake Bay. Estuarine Coastal and Shelf Science 12, 713–723. Moore, W.S. (1996). Large groundwater inputs to coastal waters revealed by 226 Ra enrichments. Nature 380, 612– 614. Moore, W.S. (1998). Application of 226 Ra, 228 Ra, 223 Ra, and 224 Ra in coastal waters to assessing coastal mixing rates and groundwater discharge to oceans. Earth and Planetary Science 107 (4), 343–349. Moore, W.S. (1999). The subterranean estuary: A reaction zone of groundwater and seawater. Marine Chemistry 65, 111–125. Moore, W.S., Arnold, R. (1996). Measurement of Ra-223 and Ra-224 in coastal waters using a delayed coincidence counter. Journal of Geophysical Research 101 (C1), 1321–1329. Oliveira, J., Burnett, W.C., Mazzilli, B.P., Braga, E.S., Farias, L.A., Christoff, J., Furtado, V.V. (2003). Reconnaissance of submarine groundwater discharge at Ubatuba coast, Brazil, using 222 Rn as a natural tracer. Journal of Environmental Radioactivity 69, 37–52. Rama, A., Moore, W.S. (1996). Using the radium quartet for evaluating groundwater input and water exchange in salt marshes. Geochimica et Cosmochimica Acta 60, 4245–4252.
360
Submarine groundwater discharge in the southeastern Mediterranean (Israel) Y. Weinsteina,* , G. Lessa , U. Kafrib , B. Herutc a Bar-Ilan University, Ramat-Gan, 52900 Israel b Geological Survey of Israel, Malkei Israel 30 St., Jerusalem, Israel c Israel Oceanographic and Limnological Research, P.O. Box 8030, Haifa, 31080 Israel
Abstract Low salinities and relatively high 222 Rn activities were found in several sites along the northern coast of Israel. Seawater salinities were as low as 37.4 ppt (compared with 39.6–39.8 ppt in eastern Mediterranean water) and 222 Rn activities were up to 3.5 dpm/l, implying the existence of Submarine Groundwater Discharge (SGD). 226 Ra activities were usually low (<0.3 dpm/l), suggesting that the discharge is mainly of fresh water. Very high radon activities (up to 1,800 dpm/l) with strong tidal variability were found in small tidal flat springs in northwestern Carmel. High activities of 222 Rn were not found at any site south of Mt. Carmel, and low salinities were documented just at two sites. The lack of SGD in the south is probably due to the efficient confinement of the Cretaceous carbonate aquifer and to the reduced hydraulic heads in the overlying Pleistocene granular aquifer. On the other hand, the SGD in the north is from the Cretaceous aquifer, either by direct discharge to the sea or through the overlying Pleistocene aquifer. The latter becomes possible due to the absence of confining layers or via transversal faults. Keywords: Submarine groundwater discharge, 222 Rn, 226 Ra, Aquifer, Salinity, Seawater, Mediterranean
1. Introduction Submarine Groundwater Discharge (SGD) was recently shown to be a major process in coastal systems. It may locally account for a significant fraction of the fresh water inflow to the ocean (e.g. Moore, 1996; Burnett et al., 2001; Taniguchi et al., 2002), and estimates of global fluxes may be as high as 10% of river flow to the ocean (Church, 1996). Moreover, SGD was recognized as a major pathway for nutrients from land to coastal water (e.g. Johannes, 1980; Valiela et al., 1990; Corbett et al., 1999; Kelly and Moran, 2002). In a lot of places, SGD is associated with seawater intrusion, and it is accepted, though not studied enough, that these two processes are complementary (e.g. Taniguchi et al., 2002). It is also understood now that the SGD itself is not just of fresh water but rather it may include significant amount of recycled seawater (Taniguchi et al., 2002). * Corresponding author. Address: Bar-Ilan University, Ramat-Gan, 52900 Israel; phone: (+972) 3 531 8806; fax: (+972) 3 534 4430; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08029-0
© 2006 Elsevier Ltd. All rights reserved.
Submarine groundwater discharge in the southeastern Mediterranean
361
SGD is usually described in relation to Karstic aquifers like the coast of Florida or the Croatian Adriatic coast or in soft sediment phreatic coastal aquifers like off the German coast at the North Sea and the Baltic Sea (e.g. Bussmann and Suess, 1998). In all these cases, discharge is controlled by the hydraulic gradient between land and ocean, and the flow is regulated by the conductivity of the aquifers. Faults could play a major role in the submarine discharge from confined aquifers. However, there is no literature about this aspect, even though it is well known that faults may behave as pathways for flow in other submarine environments (e.g. the accretionary zone, Carson and Westbrook, 1995). The southeastern Mediterranean coast is a semi-arid region. It is densely populated, and consequently coastal aquifers frequently suffer from over-exploitation and contamination. Several authors have highlighted the role of seawater intrusion in the aquifers along the Israeli coast (e.g. Melloul and Gillad, 1992; Kafri and Arad, 1979). The extent and patterns of SGD in this area was never studied. In the current paper we present new data of geochemical tracers in shallow water from the Israeli coastline with implications to the role of faults in SGD.
2. Hydrogeology In the Mediterranean coast of Israel (Fig. 1), two aquifers are discharging to the sea, the Pleistocene granular (Nilotic sands) coastal aquifer, and the Late Cretaceous (Cenomanian– Turonian) carbonate aquifer (Fig. 2). The coastal area can be hydrogeologically subdivided into two main regions, the southern coastal plain and the northern region. In the southern coastal plain, possible discharge to the sea could be just from the Pleistocene aquifer. The Cretaceous aquifer is confined under a thick sequence of clayey layers, with apparently no direct contact to the coastal water. On the other hand, in the northern region, the Cretaceous aquifer may play an important role in the discharge to sea. The region includes two areas (Mt. Carmel and Rosh Haniqra, Fig. 3) where the Cretaceous aquifer is directly exposed to sea. In other areas in this region, the Cretaceous aquifer is covered by Senonian and younger sequence. However, the Cretaceous aquifer is either not confined (e.g. the Mt. Carmel coastal plain) or that discharge is allowed via transversal faults, which are very common in this region (Fig. 3). Faults in this area were already suggested by some authors as pathways for subsurface seawater intrusion inland (Kafri and Arad, 1979).
Fig. 1. Location map of the study area.
362
Y. Weinstein et al.
Fig. 2. A hydrogeologic section of the coastal area showing the Pleistocene granular coastal aquifer and the Cretaceous aquifer.
Activities of radon in aquifer water from the northern area were studied by Kafri et al. (1997) and Kafri (2001). In Galilee (northeast of Mt. Carmel, Fig. 3), activities of fresh water from the phreatic Cretaceous aquifer were found to be between 200 and 650 dpm/l. The same aquifer, when confined under the phosphates-bearing Senonian sequence, exhibits somewhat higher radon values (Kafri, pers. comm.).
3. Methodology 3.1. Study area Our study focused on the Mediterranean coast of Israel (Figs. 4–6). We measured activities − − + of 222 Rn and 226 Ra, salinity and nutrients (ortho-PO3− 4 , NO3 + NO2 , NH4 and Si(OH)4 ) in shallow water sites, from Rosh Haniqra in the north to Ziqim in the south, a coastline 170 km long. The shoreline is usually smooth, opened to the sea, with the exceptions of Dor, Akhziv South and Shavei Zion, which are partly enclosed inlets and the Shiqmona site, which is on a tidal flat. Most sites were located close to shore, at 1–2 meter depth. Additionally, we studied surface water at several deeper sites along the shore (depths of 10–40 meters), at Haifa Bay, and along a cross-shore transect 90 km long, from Haifa to the northwest (Figs. 5(a) and 6(a)). At the northwestern side of Mt. Carmel, where the Cretaceous aquifer is directly discharging to the sea, we identified two seepage sites. The northern site, Shiqmona (Fig. 4(a)), is located on Cretaceous limestone tidal flats. Flow in the direction of the open sea was identified at several fractures cutting the limestone in the general direction of E–W. The width of the area with up-to-date identified leaking fractures is about 150 meters. We conducted measure-
Submarine groundwater discharge in the southeastern Mediterranean
363
Fig. 3. Map of faults at the northern coast (Eytam, 1988).
ments at four discharge points and focused on the southernmost spot (Shiqmona 3), including measurements during low and high tide and at various distances from the identified discharge point up to 1 km offshore. The southern site, Dado Beach (Fig. 4(a)), includes two spots of clear emanation of fresh water, in this case through a thin veneer of Pleistocene sands. 3.2. Sampling and analytical methods Water was sampled for 222 Rn and 226 Ra using 4.5 liter evacuated glass bottles. In few of the sites, water was filtered through a 100 µm filter. Bottles were sealed immediately after sampling to prevent gas loss.
364
Y. Weinstein et al.
(a)
(b)
Fig. 4. A map showing salinity (ppt, minimum values) in coastal water. (a) Full length of Israeli coastline, (b) northwestern Carmel and Haifa Bay. See Table 1 for more details.
Sampling at the shallow sites was performed directly in the water, while at deeper stations it was done from a rubber boat or from the RV Shiqmona using a peristaltic pump. Water samples for nutrients were collected by 15 ml plastic scintillation vials that were pre-washed with 10% hydrochloric acid and sample-rinsed three times. Samples were kept frozen until analysis. Radon gas was extracted and counted using a modified emanation technique described by Mathieu et al. (1988). Radon was stripped from the water using commercial stripping boards (Storm King Associates) and was transferred for counting in alpha decay counters (Applied techniques, model AC/DC-DRC-MK10-2). Water was then left in sealed bottles for another three weeks and then re-measured for 222 Rn after it reached equilibrium with 226 Ra. In some cases, radon was re-measured after a shorter period and corrected accordingly to steady-state concentrations. Counting times were adjusted to sample activity. Radon results were decaycorrected for the radium measured in the bottle and to the time of sampling to obtain in situ activities. Salinity was measured in situ using YSI6600 probe and PC6600 software. Nutrients (ortho-phosphate, nitrate + nitrite, silicic acid and ammonium) were determined using a segmented flow Skalar SANplus System Instrument as detailed in Kress
Submarine groundwater discharge in the southeastern Mediterranean
(a)
365
(b)
Fig. 5. A map showing 222 Rn activities (dpm/l, maximum values) in coastal water. (a) and (b) as in Fig. 4.
and Herut (2001) and IOC–SCOR–UNESCO (1994). The precisions for ortho-phosphate, nitrate + nitrite, ammonium and silicic acid were 0.003, 0.02, 0.02 and 0.06 µM, respectively, and the detection limits were 0.008, 0.075, 0.072 and 0.03 µM, respectively.
4. Results Activities of 222 Rn and of 226 Ra and salinity are presented in Table 1 and in maps (Figs. 4–6). The area of Haifa Bay is shown in higher resolution (Figs. 4(b), 5(b), 6(b)). 4.1. Salinity About half of the coastal sites show salinities lower than that of typical eastern Mediterranean water (39.6–39.8 ppt; Fig. 4). Salinities get as low as 37.4 ppt, and they often show large variability at the same site. Most low salinity sites are at the northern coast, and just two such sites were located south of the Carmel area. Both at Dado Beach and Shiqmona, at the foot of Mt. Carmel, where seepage was identified, salinities are significantly lower. At Shiqmona 3 (Fig. 7), salinities at the discharge point vary between 2–3 and 20 ppt during low and high tide respectively (Table 1).
366
Y. Weinstein et al.
(a)
(b)
Fig. 6. A map showing 226 Ra activities (dpm/100 kg, maximum values) in coastal water. (a) and (b) as in Fig. 4.
Table 1 Activities of radon and radium, and salinity in water samples from this study (n.d. for not determined) Station Date
Location
North
East
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12
10 km offshore NW of Haifa 13 km offshore NW of Haifa 18 km offshore NW of Haifa 28 km offshore NW of Haifa 50 km offshore NW of Haifa 90 km offshore NW of Haifa Haifa Bay Yagur F. (13 m depth) Yagur F. (31 m depth) Offshore Atlit Offshore Taninim Offshore Tel Aviv
32◦ 53.970 32◦ 54.660 32◦ 55.550 32◦ 56.910 32◦ 59.870 33◦ 09.070 32◦ 51.229 32◦ 51.070 32◦ 51.960 32◦ 43.580 32◦ 31.480 32◦ 06.600
34◦ 55.020 34◦ 52.680 34◦ 51.080 34◦ 44.8990 34◦ 31.950 34◦ 09.580 35◦ 02.089 34◦ 57.910 34◦ 57.025 34◦ 55.116 34◦ 52.510 34◦ 42.630
29.2.04 29.2.04 29.2.04 29.2.04 29.2.04 29.2.04 15.3.04 15.3.04 15.3.04 15.3.04 15.3.04 15.3.04
222 Rn 226 Ra (dpm/l) (dpm/100 kg)
0.44 0.52 0.62 0.28 0.24 0.38 0.76 0.44 0.38 0.47 0.43 0.12
3.8 6.5 3.0 7.8 4.9 4.0 12.4 6.3 11.8 4.8 8.9 9.0
Salinity (ppt) n.d. n.d. n.d. n.d. n.d. n.d. 39.16 39.24 39.53 39.54 39.64 39.56
Submarine groundwater discharge in the southeastern Mediterranean
367
Table 1 (Continued) 222 Rn 226 Ra Salinity (dpm/l) (dpm/100 kg) (ppt)
Station Date
Location
North
East
S1-S3 S4 S5 S7 S8 S9 S11 S12 S14 S16 S17 S18 S19 S20 S21 S23 S24 S25 S26 S27 S28 S29 S30 S31 S32 S34 S36
20.6.04 20.6.04 20.6.04 20.6.04 5.7.04 5.7.04 5.7.04 5.7.04 5.7.04 19.7.04 19.7.04 19.7.04 19.7.04 19.7.04 19.7.04 19.7.04 5.9.04 5.9.04 5.9.04 8.11.04 8.11.04 8.11.04 8.11.04 8.11.04 8.11.04 8.11.04 15.11.04
32◦ 36.750 32◦ 48.343 32◦ 49.900 32◦ 49.579 32◦ 30.760 32◦ 23.750 32◦ 23.140 32◦ 19.660 32◦ 07.300 33◦ 05.553 33◦ 05.253 33◦ 04.476 33◦ 03.340 33◦ 02.582 32◦ 58.957 32◦ 07.300 31◦ 55.795 31◦ 44.682 31◦ 36.770 32◦ 49.494 32◦ 51.454 32◦ 53.614 32◦ 55.229 32◦ 48.763 32◦ 46.445 32◦ 49.021 32◦ 49.021
34◦ 54.980 34◦ 57.333 34◦ 58.100 34◦ 57.400 34◦ 53.766 34◦ 51.920 34◦ 51.780 34◦ 50.840 34◦ 46.930 35◦ 06.314 35◦ 06.342 35◦ 06.322 35◦ 06.190 35◦ 05.291 35◦ 04.847 34◦ 46.930 34◦ 41.907 34◦ 36.005 34◦ 30.201 35◦ 02.608 35◦ 03.934 35◦ 04.717 35◦ 04.558 35◦ 01.007 34◦ 57.192 34◦ 57.261 34◦ 57.261
0.40 0.70 0.17 0.75 1.85 0.16 0.21 0.08 0.11 0.69 1.18 0.82 3.48 2.56 416.07 1076.14 1800.75
S35
15.11.04
32◦ 49.021
34◦ 57.261
446.05
S37 S38 S39 S40 S41 S42 S43 S44 S45 S46 S47 S48 S49 S50
15.11.04 15.11.04 15.11.04 15.11.04 15.11.04 15.11.04 15.11.04 29.11.04 29.11.04 29.11.04 29.11.04 29.11.04 29.11.04 29.11.04
Dor Beach Hof Ha’Carmel Val Tal IOLR Caesaria Alexander River mouth Beit Yanai Nathanya – Migdalor Beach Tel Aviv – Tel Baruch Rosh Haniqra Rosh Haniqra south Betzet south Akhziv andarta Akhziv port (south) Shavei Zion Rishon Le Zion Palmahim Nitsanim Zikim Kiryat Haim Kiryat Yam Frutarom Akko Shemen beach Dado spring Shiqmona 3 (mid tide) Shiq 3 (low tide) – 0.7 m from spring Shiq 3 (low) – 1.5 m from spring Shiq 3 (low) – open sea Shiqmona offshore 140 m Shiqmona offshore 400 m Shiqmona offshore 1100 m Shik 3 (high tide) – 0.7 m Shiq 3 (high) – open sea Qishon River – Julius bridge Shik 3 (mid tide) – 0.7 m Haifa Bay 1 Haifa Bay 2 Haifa Bay 3 Haifa Bay 4 Haifa Bay 5 Haifa Bay 6 (inlet of Qishon River)
32◦ 49.021 32◦ 49.047 32◦ 49.105 32◦ 49.261 32◦ 49.021 32◦ 49.021 32◦ 48.012 32◦ 49.021 32◦ 50.384 32◦ 50.149 32◦ 49.965 32◦ 49.544 32◦ 48.906 32◦ 48.978
34◦ 57.242 34◦ 57.172 34◦ 57.001 34◦ 56.557 34◦ 57.261 34◦ 57.242 35◦ 02.01 34◦ 57.261 34◦ 58.075 34◦ 59.630 35◦ 00.189 35◦ 01.145 35◦ 00.952 35◦ 01.375
1.59 0.14 0.13 0.19 0.42 0.19 0.33 0.25 0.27
7.1−21.8 6.9 11.6 9.1 5.3 10.8 0.0 0.0 8.7 9.6 8.2 0.0 8.6 7.8 8.3 8.7 9.0 11.1 7.8 9.5 13.0 10.4 12.4 26.5 106.5 313.6 402.7
38.88–39.45 n.d. 38.36, 39.8 39.45 39.17–39.43 n.d. 38.22−39.41 39.80 38.30–39.26 37.38–38.48 39.64 37.87 39.27–39.7 38.92–39.27 38.29–39.20 39.04–39.84 39.85 39.84 40.03 38.75 39.03 38.80 38.40–38.61 38.26 2.90 6.09 4.02
250.2
21.56
6.66 10.2 0.48 7.3 0.66 4.5 7.78 0.0 40.50 151.6 5.30 5.3 11.17 118.2 1076.83 1539.9 0.79 16.7 0.74 8.6 0.62 16.7 1.85 12.2 1.97 11.8 8.16 0.0
39.00 39.00 39.63 39.24 25.9–31.3 39.90 35.00 4.50 n.d. n.d. n.d. n.d. n.d. n.d.
368
Y. Weinstein et al.
4.2. Radon activities In general, radon activities in the coastal water are low, usually less than 50 dpm/100 kg (0.5 dpm/l; Fig. 5(a)). However, some of the low salinity sites show relatively high activities of radon (Fig. 5). This includes several sites in the Rosh Haniqra area, and the site of Dor next to Mt. Carmel with activities of 70–180 dpm/100 kg (0.7–1.8 dpm/l), and all sites of the Haifa Bay (0.7–3.5 dpm/l). Higher activities were documented in water samples taken from the Qishon River (1 km upstream and inlet, 11.2 and 8 dpm/l, respectively, Fig. 5(b)), which discharges to the Haifa Bay. At Dado Beach and Shiqmona (Carmel area) discharge points, activities get as high as 415 and 1796 dpm/l, respectively. At Shiqmona 3, activities measured at November 15, 2004, varied between 40 and 1796 dpm/l (a factor of 45) during high and low tide, respectively (Table 1). On the other hand, activities measured 30 meters from the discharge point, at the surface water of the open sea (off the tidal plate), vary very little (6.6 and 5.3 dpm/l). A section from the Shiqmona 3 discharge point to 1,100 meters offshore is shown in Fig. 7. The section
Fig. 7. A picture of the site of Shiqmona 3 with measured low tide salinities (white background) and radon activities (green background), including offshore radon activities.
Submarine groundwater discharge in the southeastern Mediterranean
369
shows a clear increase in surface water Rn activity toward 1,100 meters offshore (about 20 meter depth). Offshore surface water at 20–40 meters depth show relatively low Rn (<50 dpm/100 kg; Fig. 5(a)) except for the above-mentioned offshore Shiqmona and for Haifa Bay (7.7 and 0.5–2.0 dpm/l, respectively). The cross-shore transect off Haifa shows a slight increase in radon activities of surface water toward 15 km offshore (61 dpm/100 kg, Table 1) and a decline further away. 4.3. Radium activities 226 Ra is very low in the coastal water, both close to shore and at 15–40 meters depth. It is usually lower than 12 dpm/100 kg (Fig. 6(a)) except for Dor with 22 and the Haifa Bay with up to 27 dpm/100 kg (Fig. 6(b)). In several samples no radium was detected at all (Table 1). At the sites of the northwestern Carmel, radium activities are significantly higher, with 1.1 and 4.1 dpm/l at Dado Beach and Shiqmona, respectively. At Shiqmona 3, activities from November 15, 2004 vary between 4.1 and 1.6 dpm/l during high and low tide, respectively. Activities in samples from the cross-shore transect are low, varying between 3 and 8 dpm/100 kg (Fig. 6(a)).
4.4. Nutrients − − + The dissolved inorganic nutrient concentrations (ortho-PO3− 4 , NO3 + NO2 , NH4 and Si(OH)4 ) were measured in 34 sites along the shoreline. Several sites (excluding the Qishon River estuary) may be affected by known land-base point sources of nutrients (mostly coastal streams; Herut and Galil, 2000), which influence the spatial distribution of dissolved nutrient concentrations at shallow coastal water (5–30 m water depths). Nonetheless, the dissolved nutrient concentrations at the Dado and Shikmona discharge points were approximately 100-fold − higher in NO− 3 + NO2 and less so in Si(OH)4 than in most other sampling sites (Table 2) or than expected in shallow coastal water (Herut et al., 2003) or at the continental shelf (Herut et al., 2000). These preliminary data show that at certain discharge sites containing high radon
Table 2 Dissolved inorganic nutrient concentrations (µM) and radon activity (dpm/l) in water from sites sampled during this study. Coastal water (5–30 m water depth) nutrient data are from the 2000–2003 National Monitoring Reports on the Quality of the coastal Mediterranean marine environment in Israel (e.g. Herut et al., 2003). Parentheses represent extreme values Area
222 Rn
PO3− 4
− NO− 3 + NO2
NH+ 4
Si(OH)4
Shoreline stations∗ Shikmona springs Dado spring Shallow coastal water
0.2–1.8 1800 416 <0.5
0.03–0.29 1.1 0.1 bdl–1
0.1–22 336 490 bdl–3 (10)
0.1–2.6 2.5 2 bdl–3 (10)
0.8–5.2 200 70 0.6–4 (10)
Note: bdl = below detection limit. ∗ Excluding stations at Haifa Bay.
370
Y. Weinstein et al.
activities the nitrate concentrations were much higher than the maximum values found in coastal water (Table 2). It remains to further examine the relationship between radon activity and nutrient concentrations at stations along the granular and Cretaceous aquifers.
5. Discussion As shown in Table 1 and Fig. 5 radon activities along the Israeli coast are mostly lower than 3.5 dpm/l. 226 Ra activities are significantly lower, mostly in the order of less than 12 dpm/100 kg (Fig. 6), in accordance with the very low Ra activities (0–2 dpm/l) found in fresh water from boreholes along the coast (Gutman, unpublished results). This implies relatively high excess radon in almost all sites, which must arrive from the underlying sediment. We do not yet have any data regarding the activities of radium and radon in the sediments. Therefore, we cannot decipher whether and what part of the excess radon is derived by advection. However, sites with relatively high radon activities (>0.5 dpm/l) are always associated with low salinities (Table 1, Fig. 4). Moreover, low salinities are not restricted just to high radon sites. They were observed also at some relatively low radon sites (Table 1), which are far from any surface runoff (e.g. Val Tal and Beit Yanai, Table 1 and Figs. 4 and 5), implying that groundwater discharges also at these sites (except for part of the Haifa Bay, see below). The very low activities of 226 Ra even at most high radon sites (Table 1, Fig. 6) suggests that recycling of seawater could not play a major role in this discharge, since radium is more mobile in saline water (e.g. Krishnaswami et al., 1991; Sturchio et al., 2001; Porcelli and Swarzenski, 2003). The high radon sites (>0.5 dpm/l, Fig. 5(a)) are all located at the northern coast, from the Mt. Carmel coastal area to Rosh Haniqra. The southern coast is characterized by low radon (and radium) activities and by salinities usually close to typical east Mediterranean (39.6–39.8 ppt, Fig. 4(a)). Few reported submarine seeps (Galili and Inbar, 1987; Galili, personal comm.), are also restricted to the northern coast. The difference between the northern and southern coastal areas is probably the result of their different hydrogeological regime and it is mainly about whether or not the Cretaceous aquifer is discharging to the coastal water. At the south, it is just the Pleistocene aquifer that discharges to the sea, while the Cretaceous aquifer is confined by a relatively thick clayey confining layers (Fig. 2) and cannot discharge to the coastal water. Over-exploitation of the Pleistocene aquifer during the last several decades caused significant reduction of water levels (sometimes to below sea level), therefore discharge to sea was diminished, as reflected in the low radon and the mostly normal salinities. On the other hand, at the northern coast, where the confining layers do not always exist, and even when exist (e.g. Haifa Bay) they are dissected by faults (Kafri and Ecker, 1964, Fig. 3), the Cretaceous aquifer discharges (‘leaks’) to the sea. The preferred flow along faults and the higher hydraulic gradients allow significant discharge of fresh water, as reflected in radon activities and salinity. This hypothesis is supported by the proposed evidence of seawater intrusion along faults in the northern coast (Kafri and Arad, 1979). In the northwestern Carmel, the Cretaceous aquifer is directly exposed to sea. It is in this area, where we found the fresh water seepage sites of Shiqmona and Dado Beach. The activities of radon in this water (measured during low tide, with salinities of 4 and 2.9 ppt, at Shiqmona 3 and Dado Beach, respectively), is very high (1,800 and 415 dpm/l, respectively),
Submarine groundwater discharge in the southeastern Mediterranean
371
and that of Shiqmona is significantly higher than all previously measured activities in borehole water from the Cretaceous aquifer in northern Israel (<650 dpm/l, Kafri et al., 1997; Kafri, 2001) and actually in any fresh water from this area. Similarly, as shown above, nutrient data at both sites is very high, at this stage with unidentified source. The source for this high-activity high-nutrient water should be further studied in the future. At Shiqmona 3 discharge point, radon activities vary by a factor of 45 between low and high tide (1,800 and 40 dpm/l, respectively), while Ra activities vary just by a factor of 2.7 (4 and 1.5 dpm/l). Salinity also doesn’t vary to such an extent as the radon does (4.0 and 25.9 ppt, respectively). We explain this by mixing of the discharging high-Rn fresh water with tidal flat water, the latter being itself a mixture of seawater with relatively low-Rn fresh water sources discharging at other spots on the tidal flat. It could also be that the large range of Rn activity reflects a combined effect of variability in discharge (or in the dilution by seawater) together with a varying flux of radon into the fresh water. During low tide there is both higher proportion of fresh water in the discharge point as well as increased activities of radon in the fresh water component. This is supported by the tidal patterns commonly reported in soil radon (e.g. Steinitz et al., 2003; Steinitz, personal comm.). This possibility should be further studied. Activities of Rn are relatively high at Haifa Bay, reaching 3.5 dpm/l at Akko. However, we suspect that the southwestern part of the bay is affected by discharge from the Qishon River, where we found high activities of Rn and Ra (8 and 11 dpm/l Rn at the outlet and 1 km upstream, respectively, and 118 dpm/100 kg Ra at the upstream site). The high activities in river water could either be due to groundwater seepage from the river bed or – alternatively – the result of contamination from the nearby fertilizing industry. This will be studied in the near future.
Acknowledgements The research was supported by BSF Grant 2002381. Y. Gertner from the IOLR was of great help in the field and in the lab and L. Izraelov helped with the nutrient analyses. Radon analyses could not be performed without the technical assistance of Y. Sherer from the Hebrew University. The authors also wish to thank Y. Yechieli and B. Lazar for the useful discussions.
References Burnett, W.C., Kim, G., Lane-Smith, D. (2001). A continuous radon monitor for assessment of radon in coastal ocean waters. Journal of Radioanalytical and Nuclear Chemistry 249, 167–172. Bussmann, I., Suess, E. (1998). Groundwater seepage in Eckerfärde Bay (western Baltic Sea); effect on methane and salinity distribution of the water column. In: Richardson, M.D., Davis, A. (Eds.), In: Modeling Gassy Sediment Structure and Behavior, vol. 18. Pergamon Press, Oxford, pp. 1795–1806. Carson, B., Westbrook, G.K. (1995). Modern fluid flow in the Cascadia accretionary wedge: A synthesis. In: Carson, B., Westbrook, G.K., Musgrave, R.J., Suess, E. (Eds.), Proceedings of the Oceanic Drilling Program, Scientific Results, vol. 146, pp. 413–424. Church, T.M. (1996). An underground route for the water cycle. Nature 380, 579–580.
372
Y. Weinstein et al.
Corbett, D.R., Chanton, J., Burnett, W., Dillon, K., Rutkowski, C., Fourqurean, J. (1999). Patterns of groundwater discharge into Florida Bay. Limnology and Oceanography 44, 1045–1055. Eytam, Y. (1988). The shallow structure and the geological processes of the inner shelf off northern Israel in the late Pleistocene. PhD thesis. Tel-Aviv Univ., 100 pp. (In Hebrew, English abstr.) Galili, E., Inbar, M. (1987). Underwater clay exposures along the Israeli coast, submerged archaeological remains and sea-level changes in the northern Carmel coast. Horizons in Geography 22, 3–34 (in Hebrew). Herut, B., Galil, B. (2000). Environmental evaluation of the marine system along the coast of Israel (SE Mediterranean). In: Sheppard, C. (Ed.), Seas at the Millennium: An Environmental Evaluation. Elsevier, London, pp. 253– 265. Herut, B., Almogi-Labin, A., Jannink, N., Gertman, I. (2000). The seasonal dynamics of nutrient and chlorophyll a concentrations on the SE Mediterranean shelf-slope. Oceanologica Acta 23, 771–782. Herut, B., Shefer, E., Cohen, Y. (2003). Quality of the coastal Mediterranean marine environment in Israel during 2002. IOLR Rep. H39/03. (In Hebrew, Executive Summary and figures in English.) IOC–SCOR–UNESCO (1994). Protocols for the Joint Global Ocean Flux Study (JGOFS) core measurements. Manual and Guides 29. Johannes, R.E. (1980). The ecological significance of the submarine discharge of groundwater. Marine Ecology – Progress Series 3, 365–373. Kafri, U. (2001). Radon in groundwater as a tracer to assess flow velocities: Two test cases from Israel. Environmental Geology 40, 392–398. Kafri, U., Arad, A. (1979). Current subsurface intrusion of Mediterranean seawater. A possible source of groundwater salinity in the Rift Valley system. Israel Journal of Hydrology 44, 267–287. Kafri, U., Ecker, A. (1964). Neogene and Quaternary subsurface geology and hydrogeology of the Zevulun plain. Geological Survey of Israel Bulletin 37, 13. Kafri, U., Lang, B., Vulkan, U. (1997). Radon in groundwaters of northern Israel. Report GSI/16/97, Geological Survey of Israel. Report ZD/188/97, Soreq Nuclear Research Center. Kelly, R.P., Moran, S.B. (2002). Seasonal changes in groundwater input to a well-mixed estuary estimated using radium. Limnology and Oceanography 47 (6), 1796–1807. Kress, N., Herut, B. (2001). Spatial and seasonal evolution of dissolved oxygen and nutrients in the Southern Levantine Basin (Eastern Mediterranean Sea), Chemical characterization of the water masses and inferences on the N : P ratios. Deep Sea Research I 48, 2347–2372. Krishnaswami, S., Bhushan, R., Baskaran, M. (1991). Radium isotopes and 222 Rn in shallow brines, Kharaghoda (India). Chemical Geology 87, 125–136. Mathieu, G., Biscayne, P., Lupton, R., Hammond, D. (1988). System for measurements of 222 Rn at low levels in natural waters. Health Physics 55, 989–992. Melloul, A., Gillad, D. (1992). Importance of the seawater interface monitoring network for the coastal aquifer. Report Hydro/1992/4, Hydrological Service. Moore, W.S. (1996). Large groundwater inputs to coastal waters revealed by 226 Ra enrichments. Nature 380, 612– 614. Porcelli, D., Swarzenski, P.W. (2003). The behavior of U- and Th-series nuclides in groundwater. Reviews in Mineralogy and Geochemistry 52, 317–361. Steinitz, G., Begin, Z.B., Gazit-Yaari, N. (2003). Statistically significant relation between radon flux and weak earthquakes in the Dead Sea rift valley. Geology 31 (6), 505–508. Sturchio, N.C., Banner, J.L., Binz, C.M., Heraty, L.B., Musgrove, M. (2001). Radium chemistry of groundwaters in Plaeozoic carbonate aquifers, mid-continent, USA. Applied Geochemistry 16, 109–122. Taniguchi, M., Burnett, W.C., Cable, J.E., Turner, J.V. (2002). Investigation of submarine groundwater discharge. Hydrological Processes 16, 2115–2129. Valiela, I., Costa, J., Foreman, K., Teal, J.M., Howes, B., Aubrey, D. (1990). Transport of water-borne nutrients from watersheds and their effects on coastal water. Biogeochemistry 10, 177–198.
373
Submarine groundwater discharge investigations in Sicilian and Brazilian coastal waters using an underwater gamma-ray spectrometer Pavel P. Povineca,* , Isabelle Levy-Palomoa , Jean-Francois Comanduccia , Joselene de Oliveirab , Benjamino Oregionia , Agata M.G. Priviterac a International Atomic Energy Agency, Marine Environment Laboratory, MC-98000, Monaco b Instituto de Pesquisas Energeticas e Nucleares, São Paulo, Brazil c University of Palermo, National Research Group for the Defence Against Hydrogeological Disasters,
Palermo, Italy Abstract Submarine groundwater discharge (SGD) in coastal zones was monitored using in situ underwater gamma-ray spectrometry of radon decay products (214 Bi). Several sites were visited during the IAEA’2002 expedition to southeastern Sicily, where SGD variations were observed in the Donnalucata boat basin. The continuous monitoring carried out for 3 days at the site closest to the coast revealed an anticorrelation dependence of 222 Rn concentration with tide and salinity. The 222 Rn activity concentrations in seawater varied from 2.3 kBq m−3 (during high tides) to 4.8 kBq m−3 (during low tides). In situ gamma-ray spectrometric measurements were also carried out during the IAEA–UNESCO’2003 expedition to Ubatuba (Brazil). The results obtained during 5 days of continuous monitoring in Flamengo Bay confirmed an anticorrelation between the 222 Rn activity concentration in seawater (which varied between 1.5 and 5.2 kBq m−3 ) and tide/salinity, however, the relationship seems to be more complicated than was observed off Donnalucata. It was confirmed that at both Donnalucata and Ubatuba sites the variations in 222 Rn concentrations were caused by sea level changes, as tide effects induce variations of hydraulic gradients, which increase 222 Rn concentrations during decreasing sea level, and conversly, during high tides the 222 Rn concentrations are decreasing. Keywords: Radionuclides, 222 Rn, 214 Bi, Gamma-ray spectrometry, Groundwater, Seawater, Submarine groundwater discharge, Coastal zone, Mediterranean Sea, Sicily, Southwest Atlantic, Brazil
1. Introduction In situ gamma-ray spectrometry has been recognised as a powerful tool for continuous analysis of gamma-ray emitters in seawater (Wedekind, 1973; Povinec et al., 1996; Wedekind et al., 1999, 2000; Povinec et al., 2001a; Nies et al., 2003). Following successful operations of * Corresponding author. Address: IAEA-MEL, 4 Quai Antoine 1er, MC-98000, Monaco; phone: (+377) 97977272; fax: (+377) 97977273; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08030-7
© 2006 Elsevier Ltd. All rights reserved.
374
P.P. Povinec et al.
IAEA-MEL’s in situ, fully automatised gamma-ray spectrometers for continuous monitoring of anthropogenic radionuclides in seawater (Povinec et al., 2001b; Osvath et al., 2005), it has been possible to apply this technique also for monitoring natural radionuclides such as 40 K, radon and its daughter products, which are commonly used as tracers for studying marine and coastal processes (Povinec et al., 2001a; Levy-Palomo et al., 2004). One of the frequently studied coastal processes is Submarine Groundwater Discharge (SGD), because of its potential importance for management of fresh water resources in coastal areas (Taniguchi et al., 2002; Kontar et al., 2002). Several isotope techniques for SGD quantification have been developed using stable (2 H, 18 O, 87/86 Sr, etc.) and radioactive (3 H, 14 C, Ra isotopes, 222 Rn, etc.) isotopes (e.g. Moore, 2000; Burnett et al., 2001, 2002; Moore, 2005; Schiavo et al., 2005; Weinstein et al., 2005). New technologies developed in recent years have been used to carry out temporal and spatial monitoring of SGD via analysis of radon or its daughter products emitting alpha-rays (Burnett et al., 2001; Burnett and Dulaiova, 2003; de Oliveira et al., 2003; Burnett and Dulaiova, 2005), or by analysing radon decay products emitting gamma-rays (Povinec et al., 2001a; Levy-Palomo et al., 2004). A Coordinated Research Project (CRP) on “Nuclear and Isotopic Techniques for the Characterisation of SGD in Coastal Zones” has been jointly organised by the IAEA’s Marine Environment Laboratory (Monaco) and the Isotope Hydrology Section (Vienna), with the aim to develop new isotope techniques for studying SGD. The CRP has been carried out in cooperation with UNESCO’s Intergovernmental Oceanographic Commission (IOC) and the International Hydrological Programme (IHP), and collaboration with several laboratories in Brazil, India, Italy, Japan, Russia, Slovenia, Turkey and USA was established (Povinec et al., 2005a). In the framework of the CRP two expeditions were carried out (in June 2001 and March 2002) to the Ionian Sea (offshore Sicily), and one in November 2003 to Ubatuba (São Paulo region, Brazil). The choice of such geologically different regions was based on the strategy, developed in the framework of the IAEA–UNESCO SGD cooperation, to visit and study SGD sites with different geological and hydrological conditions, which could primarily affect SGD in the region. In the present paper we present results of monitoring of SGD via in situ analysis of radon decay products in seawater using an underwater gamma-ray spectrometry. The field work was carried out offshore Donnalucata in Sicily and Ubatuba in Brazil.
2. Study areas 2.1. Donnalucata boat basin (Sicily) The studied area belongs to the Hyblean Plateau that forms the south-eastern (SE) part of Sicily (Aureli and Privitera, 2005). It is primarily of carbonate in origin (Triassic–Jurassic and partly Cretaceous). The eastern part was influenced by volcanic activity, while the western part was formed from carbonate sediments. The groundwater circulation appears mostly along cracks, fractures and karst hollows, producing numerous springs on beaches, as well as submarine springs in the sea (Aureli and Privitera, 2005). In situ underwater gamma-ray spectrometry measurements were carried out from 16 to 25 March 2002 in the south-eastern Sicily during the IAEA’2002 expedition. The continuous
Submarine groundwater discharge investigations in Sicilian and Brazilian coastal waters
375
SGD monitoring was done mainly in the Donnalucata boat basin, where several sites, situated close to manual or automatic seepage meters posts (Taniguchi, 2005) were visited. 2.2. Flamengo Bay (Ubatuba region, Brazil) In contrast to the Sicily sites, which represent a typical karstic region with low massic activities of 238 U (∼10 Bq kg−1 ) and 232 Th (∼5 Bq kg−1 ), the Brazilian site is a tropical coastal area characterised by granite rocks where the concentrations of 238 U and 232 Th in collected rock samples were higher by a factor of 5 and 10, respectively. The main geologic/geomorphologic feature is the presence of granites and migmatites of the mountain chain Serra do Mar (altitudes up to 1,000 meters), which reaches the shore in almost all of the study area and limits the extension of the drainage systems and of Quaternary coastal plains. The mean annual rainfall is around 1,800 mm, the maximum rainfall rates being observed in February. Sea level varies from 0.5 to 1.5 m, the highest values occurring in months August/September due to greater volume of warm waters of the Brazil Current (Mesquita, 1997). Continuous underwater gamma-ray spectrometry measurements were carried out in several bays along the Ubatuba coast. We discuss here results obtained for Flamengo Bay only, where measurements were done between 22–26 November, 2003, at the Oceanographic Base of the University of São Paulo, about 20 meters from the coast.
3. Methods The underwater gamma-ray spectrometer consists of a 5 cm in diameter and 15 cm long NaI(Tl) scintillation crystal housed together with photomultiplier, high voltage power supply and signal processing electronics in a stainless steel tube one meter long and 10 cm in diameter (Fig. 1). Additional sensors for monitoring of temperature, water depth and wave impacts are located in front of the NaI(Tl) detector. The detector unit was connected via a 70 m long, double armoured steel coaxial cable to a PC with processing electronics and multichannel
Fig. 1. Underwater gamma-ray spectrometer with electronics.
376
P.P. Povinec et al.
analyser. The PC and an auxiliary low voltage power supply are located on a ship or in a car when operating close to the coast. The energy calibration of the gamma-ray spectrometer was carried out using 137 Cs and 60 Co radioactive sources. The energy resolution for 662 keV (137 Cs) gamma-rays was 6.5%. The efficiency calibration was done using a 137 Cs source dispersed in a polyethylene tank 1 m in diameter and 1 m high, filled with fresh water. The tank was also filled with seawater with natural concentration of 40 K (1461 keV), 226 Ra, 222 Rn and its daughter products. 222 Rn being pure alpha-emitter (half-life 3.83 days), decays to 218 Po (half-life 3.05 months), which then decays to 214 Pb (half-life 26.8 months) and then to 214 Bi (19.7 months), which has been used as the most suitable gamma-ray emitter for analysis of 222 Rn in environmental samples, as it emits high intensity gamma-rays of different energies. The corresponding 214 Bi peaks used in spectra evaluations were either 609, 1120 or 1765 keV peaks, depending on background conditions during real measurements, assuring the best factor of merit (the ratio εγ /(b)−1/2 , where ε is the photopeak efficiency, γ is the photon emission intensity, and b is the corresponding background). Background measurements were carried out with the detector immersed in the tank filled with fresh water. The detection limit for 222 Rn measurements in seawater is ∼0.05 kBq m−3 and the reported uncertainties are ∼20%. The data acquisition system evaluates gamma-ray spectra every minute. Later the obtained spectra are usually integrated to one hour intervals (depending on the type of measurement, e.g. continuous long-term monitoring at one site or spatial mapping), and the activity concentration of selected radionuclides in fresh water or seawater is calculated. The system is fully automatised and can be operated without any technical surveillance. Continuous salinity measurements were carried out using a small conductivity/temperature/ depth DST-CTD sensor (Star-Oddi, Iceland). The precision of salinity measurements was ±0.01. Water level data measured by Burnett and Dulaiova (2005) during the Sicily mission, and data recorded by the Oceanographic Institute of the University of São Paulo for the Brazilian mission have been used in the present study.
4. Results It has been a great challenge to investigate SGD using underwater gamma-ray spectrometry in such different geological and hydrological environments, as presented by the Sicilian and Brazilian coasts. First we shall present results obtained for the Sicilian and Brazilian coasts separately, and then we shall compare and discuss results obtained for both sites. 4.1. Donnalucata boat basin (Sicily) Several sites were visited in the Donnalucata boat basin where seepage measurements were also carried out using automated and manual seepage chambers. Seepage rates up to ∼30 cm day−1 were observed (Taniguchi, 2005). Highest 222 Rn activity concentrations (up to 3.7 kBq m−3 ) were recorded at the sites close to the coast, where salinities were the lowest (35.8); on the contrary, in open sea conditions, where salinity was the highest (38.7), the 222 Rn activity concentrations were the lowest (∼0.1 kBq m−3 ). A recirculation
Submarine groundwater discharge investigations in Sicilian and Brazilian coastal waters
377
Fig. 2. 222 Rn activity concentration in seawater, salinity and tide with time from 20 to 22 March, 2002 in the Donnalucata boat basin, Italy.
of groundwater + seawater mixtures, having higher salinities but lower 222 Rn activity concentrations (due to its radioactive decay), may be responsible for the observed lower 222 Rn concentrations at sites where high seepage rates were observed. Time series of 222 Rn in seawater, salinity and tide were recorded at a site close to the coast. Results presented in Fig. 2 document that after the maximum tide the 222 Rn activity concentration of seawater was minimum (down to 2.3 kBq m−3 ), and after the minimum tide the 222 Rn activity concentration was at maximum (up to 4.8 kBq m−3 ) with a delay of about one hour. A shift of approximately two hours was observed between the maximum tide and the salinity maximum. It is interesting to see that even small changes in the water level, observed e.g. during March 21 and 22, made corresponding changes in 222 Rn activity concentrations. 4.2. Flamengo Bay (Ubatuba, Brazil) Time series of 222 Rn activity concentration in seawater, salinity and tide recorded from November 22 to 26, 2003, in Flamengo Bay are shown in Fig. 3. The 222 Rn activity concentration in seawater varied between 1.5 and 5.2 kBq m−3 , while the tide varied between 4.4 and 5.6 meters. Generally, the salinity record follows the tide record, however, on November 26 a delay in the salinity record was observed. The usual inverse relationship between the 222 Rn activity concentration in seawater and tide/salinity was not observed during November 22, despite of large variations in water level. The observed changes in salinity during this time were, however, also much smaller than during 25th and 26th November. The usual inverse relationship between the 222 Rn activity concentration in seawater and tide/salinity is, however, well seen
378
P.P. Povinec et al.
Fig. 3. 222 Rn activity concentration in seawater, salinity and tide with time during measurements in Flamengo Bay, Brazil.
from 23rd to 25th November, when a few hours shift between the tide minimum/maximum and the maximum/minimum in the 222 Rn activity concentration was observed. 5. Discussion The results obtained for Sicilian and Brazilian waters confirm the inverse relationship between tide and 222 Rn activity concentration. During decreasing sea level 222 Rn concentration increases, and opposite, during high tides 222 Rn concentration is decreasing. While 222 Rn activity concentration in Sicilian waters followed the tide with a delay of only one hour, in Brazilian waters a delay of several hours was observed. This may be caused by different geological/hydrological conditions at both sites. The Sicilian coast is characterised by carbonate rocks with cracks, which facilitate groundwater transport to the sea, while for the Brazilian coast, granite rocks have lower transport capabilities (IAEA, 1998). In spite of different geological/hydrological settings, the 222 Rn activity concentrations in seawater at Sicilian and Brazilian sites were very similar, between 2.3 and 4.8 kBq m−3 , and 1.5 and 5.2 kBq m−3 , respectively. Due to a factor of 5 higher 238 U concentration in granite than carbonate rocks, higher 222 Rn activity concentrations along the Brazilian coast would be expected. However, in Flamengo Bay, similarly as it was observed in the Donnalucata boat basin, the SGD may be represented by a mixture of recirculated groundwater and seawater, having a lower 222 Rn concentration. In the Donnalucata region we have a clear end-member represented by the groundwater beach spring, having an average 222 Rn activity concentration of ∼16 kBq m−3 (varying between 12 and 18 kBq m−3 , depending on tide; Povinec et al., 2005b). In the case of Flamengo Bay we need to wait for availability of tritium and stable isotope data, which could help to quantify the composition of SGD at this site. The variations in 222 Rn activity concentrations in the Donnalucata boat basin, measured at several sites using alpha-ray spectrometry, were also reported by Burnett and Dulaiova (2005).
Submarine groundwater discharge investigations in Sicilian and Brazilian coastal waters
379
They found spatial changes in 222 Rn activity concentrations from 0.05 to 2.5 kBq m−3 (the maximum concentration was measured on March 21st at a site about 10 meters far from our site). They showed that the observed variations in 222 Rn activity concentrations can be related to SGD fluxes, and thus can be used for characterisation of SGD. Such large changes in SGD, observed in a relatively small area, document again why the isotopic characterisation of SGD is important for estimation of real groundwater fluxes to the sea.
6. Conclusions In situ underwater gamma-ray spectrometry measurements carried out during the expeditions in coastal waters of Donnalucata (Italy) and Ubatuba (Brazil) showed the ability of the method to monitor SGD and to study its temporal and spatial variations. This new method of SGD investigations represents a robust technique that can be applied for long-term, continuous monitoring of radon in seawater and/or groundwater. Time series measurements of 222 Rn activity concentration offshore Donnalucata (activity concentrations between 2.3 and 4.8 kBq m−3 ) and Ubatuba (1.5–5.2 kBq m−3 ) generally confirmed an anticorrelation between the 222 Rn activity concentration and tide/salinity. However, the relationship at the Ubatuba site seems to be more complicated, probably due to different geological/hydrological conditions. It has been confirmed that variations in 222 Rn activity concentrations are caused by sea level variations as tide effects induce variations of hydraulic gradients, which increase 222 Rn concentrations during decreasing sea level, and conversly, during high tides 222 Rn activity concentrations are decreasing.
Acknowledgements The authors would like to thank colleagues who participated in the expeditions to Sicily and to Brazil for fruitful collaboration. The authors also thank Prof. W.C. Burnett and the Oceanographic Institute of the University of São Paulo for providing water level data. The University of Palermo (Italy), Instituto de Pesquisas Energeticas e Nucleares at São Paulo (Brazil), and the University of São Paulo are acknowledged for logistical support during the field-work. The Captain and the crew of the R/V Albacora are acknowledged for assistance during the oceanographic investigations offshore Ubatuba. The Agency is grateful for the support provided to its Marine Environment Laboratory by the Government of the Principality of Monaco.
References Aureli, A., Privitera, A.M.G. (2005). Structural, geological and hydrological characteristics of the Hyblean Plateau, Continental Shelf Research, submitted. Burnett, W.C., Dulaiova, H. (2003). Estimating the dynamics of groundwater input into the coastal zone via continuous radon-222 measurements. Journal of Environmental Radioactivity 69, 21–35. Burnett, W.C., Dulaiova, H. (2005). Radon as a tracer of submarine groundwater discharge into a boat basin in Donnalucata, Sicily. Continental Shelf Research, submitted.
380
P.P. Povinec et al.
Burnett, W.C., Kim, G., Lane-Smith, D. (2001). A continuous radon monitor for assessment of radon in coastal ocean waters. Journal of Radioanalytical and Nuclear Chemistry 249, 167–172. Burnett, W.C., Chanton, J., Christoff, J., Kontar, E., Krupa, S., Lambert, M., Moore, W., O’Rourke, D., Paulsen, R., Smith, C., Smith, L., Taniguchi, M. (2002). Assessing methodologies for measuring groundwater discharge to the ocean. EOS 83, 117–123. de Oliveira, J., Burnett, W.C., Mazzilli, B.P., Braga, E.S., Farias, L.A., Christoff, J., Furtado, V.V. (2003). Reconnaissance of submarine groundwater discharge at Ubatuba coast, Brazil, using 222 Rn as a natural tracer. Journal of Environmental Radioactivity 69, 37–52. IAEA (1998). The radiological situation at the atolls of Mururoa and Fangataufa. The Main Report, IAEA, Vienna, 282 pp. Kontar, E.A., Burnett, W.C., Povinec, P.P. (2002). Submarine groundwater discharge and its influence on hydrological trends in the Mediterranean Sea. In: Tracing Long-Term Hydrological Change in the Mediterranean Sea. CIESM Series, vol. 16. CIESM, Monaco, pp. 109–113. Levy-Palomo, I., Comanducci, J.-F., Povinec, P.P. (2004). Investigation of submarine groundwater discharge in Sicilian and Brazilian coastal waters using underwater gamma-spectrometer. In: Book of Extended Synopses, IAEACn-118. IAEA, Vienna, pp. 228–229. Mesquita, A.R. (1997). Marés, circulação e nível do mar na Costa Sudeste do Brasil. Relatório Fundespa, São Paulo. Moore, W.S. (2000). Ages of continental shelf waters determined from 223 Ra and 224 Ra. Journal of Geophysical Research 105, 22117–22122. Moore, W.S. (2005). Radium isotopes as tracers of submarine groundwater discharge in Sicily. Continental Shelf Research, submitted. Nies, H., Herrmann, J., Schilling, G. (2003). Aktuelle Belastung von Nord- und Ostsee durch Radionuklide und Informationen aus dem automatischen Radioaktivitätsmessnetz des BSH. In: Proc. 12 Fachgespräch “Überwachung der Umweltradioaktivität”. BMU, Bonn, pp. 445–454. Osvath, I., Povinec, P.P., Livingston, H.D., Ryan, T.P., Mulsow, S., Commanducci, J.-F. (2005). Monitoring of radioactivity in NW Irish Sea water using a stationary underwater gamma-ray spectrometer with satellite data transmission. Journal of Radioanalytical and Nuclear Chemistry 263, 437–440. Povinec, P.P., Osvath, I., Baxter, M.S. (1996). Underwater gamma-spectrometry with HPGe and NaI(Tl) detectors. Applied Radiation and Isotopes 47, 1127–1133. Povinec, P.P., Osvath, I., Mulsow, S., Commanducci, J.-F. (2001a). La surveillance in situ de la radioactivite marine au large de Monaco. Rapport du 36e Congres de la CIESM, vol. 36. CIESM, Monaco, 155 pp. Povinec, P.P., La Rosa, J., Lee, S.-H., Mulsow, S., Osvath, I., Wyse, E. (2001b). Recent developments in radiometric and mass spectrometry methods for marine radioactivity measurements. Journal of Radioanalytical and Nuclear Chemistry 248, 713–718. Povinec, P., Aggarwal, P., Aureli, A., Burnett, W.C., Kontar, E.A., Kulkarni, K.M., Moore, W.S., Rajar, R., Taniguchi, M., Comanducci, J.-F., Cusimano, G., Dulaiova, H., Gatto, L., Hauser, S., Levy-Palomo, I., Ozorovich, Y.R., Privitera, A.M.G., Schiavo, M.A. (2005a). Characterisation of submarine ground water discharge offshore of south-eastern Sicily – SGD Collaboration. Journal of Environmental Radioactivity, submitted. Povinec, P.P., Comanducci, J.-F., Levy-Palomo, I., Oregioni, B. (2005b). Monitoring of submarine groundwater discharge along the Donnalucata coast in the south-eastern Sicily using underwater gamma-ray spectrometry of radon decay products. Continetal Shelf Research, submitted. Schiavo, M.A., Grassa, F., Hauser, S. (2005). The role of isotopes in understanding the groundwater circulation in coastal areas of the south-eastern Sicily (Italy). Applied Radiation and Isotopes, submitted. Taniguchi, M., Burnett, W.C., Cable, J.E., Turner, J.V. (2002). Investigation of submarine groundwater discharge. Hydrological Processes 16, 2115–2129. Taniguchi, M. (2005). Submarine groundwater discharge measured by seepage meters in Sicilian coastal waters. Continental Shelf Research, submitted. Wedekind, Ch. (1973). Gamma-ray spectrometer probe for the measurements of radioactive pollution in the sea. Health Physics 25, 51–55. Wedekind, Ch., Schilling, G., Grüttmüller, M., Becker, K. (1999). Gamma-radiation monitoring network at sea. Applied Radiation and Isotopes 50, 733–738. Wedekind, Ch., Schilling, G., Grüttmüller, M., Becker, K. (2000). Marine environmental radioactivity monitoring by “in-situ” radiation detection. Kerntechnik 65, 190–194.
Submarine groundwater discharge investigations in Sicilian and Brazilian coastal waters
381
Weinstein, Y., Less, G., Kafri, U., Herut, B. (2005). Submarine groundwater discharge in the southeastern Mediterranean. In: Povinec, P., Sanchez-Cabeza, J.A. (Eds.), Radionuclides in the Environment, International Conference on Isotopes in Environmental Studies: Aquatic Forum 2004. 25–29 October, Monaco. Elsevier, Amsterdam, pp. 360–372, this volume.
382
Isotope hydrochemical investigation of saline intrusion in the coastal aquifer of Karachi, Pakistan A. Mashiatullaha,* , R.M. Qureshia , M.A. Tasneema , T. Javeda , C.B. Gayeb , E. Ahmadc , N. Ahmadd a Pakistan Institute of Nuclear Science and Technology, Islamabad, Pakistan b Isotope Hydrology Section, International Atomic Energy Agency, Vienna, Austria c World Wide Fund for Nature (WWF), Karachi, Pakistan d Postgraduate Centre for Earth Sciences, University of the Punjab, Lahore, Pakistan
Abstract Environmental stable isotope parameters δ 18 O and δ 2 H in water molecules, δ 13 C in Total Dissolved Inorganic Carbon (TDIC) and δ 34 S in SO4 , have been used in conjunction with physico-chemical tools to study the extent and origin of saline intrusion in the coastal aquifer system of Karachi. Physico-chemical data show that the shallow groundwater is moderately saline. Shallow wells in close proximity of Karachi coast have much higher values of electrical conductivity, salinity, contents of aqueous chloride and sulfate as compared to all other locations relatively far away from the coast. The mean stable isotope contents of 18 O and 2 H indicate that the shallow aquifer system is recharged by a mixture of fresh water of mainly Indus River origin and the polluted waters of the Layari and Malir Rivers and their tributaries, both under natural infiltration conditions and artificially induced infiltration conditions. Much depleted values of δ 13 C (less than −6h V-PDB) indicate the impact of pollution from the Layari and Malir Rivers into the shallow groundwater environment. Relatively deep groundwater is mostly saline and has high electrical conductivity and salinity as compared to shallow groundwater. Physico-chemical data of deep groundwater show that the deep wells have relatively higher values of electrical conductivity and salinity as compared to the shallow wells. The hydrochemical and stable isotope results indicate that the confined aquifer hosts a mixture of rainwater from the hinterlands and surrounding regions around coastal Karachi, as well as sea trapped water/seawater through intrusion under natural infiltration conditions or under induced recharge conditions. The present investigations prove seawater intrusion and existence of trapped seawater salinity and build-up of salt-water up-coning in the shallow and deep confined aquifer in coastal Karachi. Keywords: Stable isotopes, δ 2 H, δ 13 C, δ 18 O, δ 34 S, Surface water, Groundwater, Deep well, Shallow well, Karachi Sea, Pakistan
1. Introduction Karachi metropolis is located on the northern boundary of the Arabian Sea and hosts a coastline extending up to ∼80 km. It is by far the most populous (∼12 million inhabitants) and * Corresponding author. Address: Pakistan Institute of Nuclear Science and Technology, P.O. Nilore, Islamabad, Pakistan; fax: (+92) 51 9290275; e-mail:
[email protected],
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08031-9
© 2006 Elsevier Ltd. All rights reserved.
Saline intrusion in the coastal aquifer of Karachi, Pakistan
383
the largest industrial base (∼1000 large industrial units) of Pakistan. The drainage system in Karachi mainly comprises the Layari River and the Malir River systems. In general, there are five sources of recharge to groundwater reserves in and around coastal Karachi. These include: (i) rainfall, (ii) Indus River water supply system, (iii) Hub River/Hub Lake water supply system, (iv) Layari/Malir Rivers and their contributory channels that drain domestic, industrial and agricultural wastewater, and (v) seawater. Contribution to groundwater recharge by local precipitation seems very small due to very poor frequency of rainfall events in coastal Karachi. The long term (15 years annual record) mean monthly average precipitation for Karachi is between 0–15 mm during the months of January to June, 23–91 mm during the months of July to September and 0–7 mm during the months of October to December (IAEA, 1992). As the rainfall intensity in the study area is very low (around 75–100 mm per year), the city of Karachi suffers from a deficit of precipitation, and it appears that the contribution to shallow groundwater storage from rainfall in the Karachi metropolis is very little except for the very shallow thin sandy lenses. Nevertheless, rainfall in the hinterlands and other areas surrounding Karachi may significantly contribute to the confined groundwater flow system. It appears that the remaining four sources can play a significant role in recharge to shallow and deep groundwater system in coastal Karachi. The pollution inventories in the water courses are quite alarming as a number of shallow to deep pumping wells (called hydrants) and mechanized hand-pumps are installed along these rivers causing a threat of artificially induced recharge to the local groundwater system. In addition, the long term pumping of energized tube-wells installed in the immediate vicinity of seashore can also cause intrusion of seawater into the coastal groundwater system. Some small-scale sporadic groundwater quality surveys involving classical hydrochemical have been made in the past to estimate groundwater pollution status in coastal Karachi. There is a vast literature on application of classical hydrochemical techniques and isotope hydrogeochemical techniques for evaluation of saline intrusion in coastal aquifers (Freeze and Cherry, 1997; Goswani, 1969; Klein and Ratzlaff, 1989; Yurtsever and Payne, 1978; Kulkarni et al., 1979). In the present investigation, the stable isotope contents of 18 O, 2 H (in water molecules), 13 C in the Total Dissolved Inorganic Carbon (TDIC) and the related hydrochemical data (major cation analysis) of all surface water samples and the coastal groundwater samples collected are statistically evaluated to postulate the origin of groundwater and associated salinity in the shallow and deep aquifer system in coastal Karachi. This paper documents results of a first ever study on the evaluation of groundwater recharge characteristics and origin of salinity in coastal Karachi using environmental isotope techniques.
2. Methods Standard field sample collection/preservation methods were used for subsequent chemical, and stable isotopic analysis in the laboratory. 2.1. Field sampling Surface water samples were collected from various locations along polluted streams/rivers namely: Layari River, Malir River and local sea (shallow seawater off Karachi coast). Shallow
384
A. Mashiatullah et al.
Fig. 1. Map of Karachi showing sampling points.
groundwater and deep groundwater were collected from different localities of Karachi city. Number of samples collected are as follows: Indus River (1), Layari River (3), Malir River (2), Karachi Sea (5), shallow groundwater (12) and deep groundwater (16). Figure 1 shows the location of sampling points. Geo-location of each sampling point was determined with a standard GPS (M/S Garmin). Shallow groundwater samples (12) were collected from hand-pumps and dug wells. All water samples were collected in leak-tight/lined cap plastic bottles or glass bottles. 2.2. Field in-situ analyses In the field, the water samples were immediately analyzed for specific physiochemical parameters such as pH, temperature, dissolved oxygen, turbidity, electrical conductivity and
Saline intrusion in the coastal aquifer of Karachi, Pakistan
385
salinity. Turbidity was measured with a portable turbidity meter (Model 6035, JENWAY). Electrical conductivity and temperature were measured with a portable conductivity meter (Model HI 8633, M/S HANNA Instruments). Dissolved oxygen was measured with a portable DO Meter (Model 9070, JENWAY). Salinity was measured with a portable Salinometer (refractometer). 2.3. Laboratory analyses In the laboratory, the collected water samples were analyzed for stable isotope content of 18 O and 2 H in the water molecule; 13 CTDIC in TDIC, 34 S from sulfate of water and major ions −1 such as Cl− , SO−2 4 , HCO3 (Clark and Fritz, 1997; Qureshi et al., 2001). Chloride contents were determined by ion selective electrodes with an Orion Microprocessor Ion Analyzer/901. Carbonates and bicarbonates were measured by titration. Sulfate concentrations were determined by turbidimetric/spectrophotometric method (Hitachi 220-A Double Beam Spectrophotometer). The environmental stable isotope analyses were performed using a modified Varian Mat GD-150 Mass Spectrometer. All stable isotope data are expressed in conventional δ (h) notation and referred to the standards namely: SMOW (Standard Mean Ocean Water) for 18 O and 2 H analyses, PDB (Pee-Dee Belemnite) for 13 C analysis of TDIC, and CDT (Canyon Diablo Troilite) for 34 S. The overall analytical uncertainties are ±0.1h for 18 O and 13 C 2 TDIC and ±1h for δ H measurements.
3. Results and discussion 3.1. Surface water sources 3.1.1. Local precipitation No significant rainfall events occurred in coastal Karachi during the sampling period. However, stable isotope data on precipitation for the period from 1961 to 1975 are available from the IAEA Precipitation Network for Karachi Station (IAEA Precipitation Network Code: 41780000, Lat. 24◦ 90 N, Long. 67◦ 13 E, Altitude: 23 meters above mean sea level). The following stable isotope indices of precipitation in Karachi as quoted by IAEA were used for interpretation purposes: Long Term Weighted Mean δ 18 O (water): −3.93 ± 1.94h V-SMOW. Long Term Weighted Mean δ 2 H (water): −23.5 ± 18.1h V-SMOW. Long Term Monthly Correlation between δ 18 O and δ 2 H:
δ 2 H = 7.56δ 18 O ± 0.34 + [3.41 ± 1.50]. 3.1.2. Indus River The Indus River (IR) water sample was collected from the River course near Thatta city whereby the Indus River water is partly diverted to Karachi for irrigation and drinking water purposes. Typical chemical indices and stable isotope indices of oxygen and hydrogen in the Indus River water molecules as used in evaluation of coastal groundwater recharge characteristics are given in Table 1.
386
Table 1 Physico-chemical and isotopic ranges of Indus River, Layari River, Malir River and Karachi Sea in shallow and deep groundwater samples Indus River
Layari River
Malir River
Karachi Sea
Shallow groundwater
Deep groundwater
PH E.C. (mS/cm) Salinity Turbidity (NTU) HCO−1 3 (ppm) Cl−1 (ppm) SO−2 4 (ppm) −1 SO−2 4 /Cl −1 Cl/HCO3 δ 18 O (h V-SMOW) δ 2 H (h V-SMOW) δ 34 SSO4 (h CDT) δ 13 C (h PDB)
7.56 0.48 1 36 108 14 13 0.69 0.22 −5.9 −48.12 6.28 −1.66
7.46–8.40 1.5–9.02 1–5.0 54–76 502–660 431–1300 33–195 0.06–0.11 1.48–3.39 −5.4 to −2.7 −44.4 to −31 7.2–8.9 −7.2 to −0.2
7.7–7.9 3.2–3.4 3.5–5 97–98 490 2020 271 0.10 6.52 −0.95 to −0.66 −27.5 to −5.91 11.6–14.2 −4.5 to −2.4
7.7–8.5 49.3–53.7 31.0–39.0 52.6–195.5 145–196 21580–25230 2080–2320 0.064–0.06 190–300 0.27–1.1 4.9–8.6 17.6–19.5 −3.9 to 0.8
6.75–8.3 0.9–12.5 1–5.1 13.9–95 246–520 58–580 26–220 0.10–0.79 0.33–6.35 −6.74 to −4.41 −53.89 to −33.06 2.69–9.7 −11.23 to −1.72
6.94–7.94 1.9–32.5 1.3–7.4 2.7–59.6 102–760 500–12800 61–2220 0.05–0.36 1.8–215 −6.5 to −3.04 −72.59 to −26.7 −17.97 to −2.7 −12.04 to −2.47
A. Mashiatullah et al.
Parameters (units)
Saline intrusion in the coastal aquifer of Karachi, Pakistan
387
3.1.3. Local polluted rivers The Layari River and the Malir River represent the local rivers in the study area as these facilitate the main drainage of domestic and industrial waste generated by the entire Karachi metropolis to the Karachi Sea. Thus, the two rivers get polluted downstream. The Layari River (LR) was monitored at three locations along its flow from North Karachi (upstream region) to Sher-Shah Bridge (downstream region) near Keamari/Karachi Harbour. The range of variation (along the flow direction) in typical chemical and stable isotope contents of oxygen and hydrogen in the water molecules, δ 34 S of aqueous sulfate of water and carbon isotope in TDIC of LR water as used in evaluation of coastal groundwater recharge characteristics are shown in Table 1. There is a good correspondence between electrical conductivity (EC) and salinity along the river. Generally, the EC and salinity values tend to decrease downstream. Maximum values (9.02 mS/cm) of EC were observed at the origin of the Layari stream near Yousuf Goth area. In this zone, the Layari stream receives minor spring water, domestic wastewater from small isolated dwellings and wastewater from industries (pharmaceutical industry, electronic industry, etc.) which host tube-wells with quite high salinity values. Turbidity levels in the river water fluctuate depending upon the concentration of inputs from industrial and domestic sector. High concentrations of Cl− (1300 ppm) coupled with mildly alkaline pH values are found in the upstream regions of the river. Significantly, high values of Cl− in the upstream region indicate that the source of water in the river is the saline water discharged from deep tube-wells installed in the nearby industrial complexes. The δ 13 CTDIC , δ 18 O (water) and δ 34 S values are also quite enriched in this zone of LR as compared to local shallow groundwater, and are in fact relatively closer to sea values. The Malir River (MR) was monitored at two locations along its flow from Karachi East to Qayyum Abad Bridge where the MR joins Ghizri Creek on the north-west side of Karachi coast. The range of variation (along the flow direction) in typical chemical and stable isotope contents of oxygen and hydrogen in the water molecules and of carbon in the TDIC and 34 S of aqueous sulfate of water in MR water, used in evaluation of coastal groundwater recharge characteristics, are shown in Table 1. Like LR, there is good correspondence between EC and salinity along the flow in MR. The reducing conditions of the river become adverse as it receives more and more industrial effluents and sewage from Korangi Industrial Trading Estate (KITE) zone and Qayyum Abad/Ghizri Creek area along its course towards the sea. 3.1.4. The Karachi Sea The coast of Karachi is about 40 km long. Shallow seawater samples were collected during the high tide period from 5 locations in the inter-tidal zone along Karachi coast. The range of variation in chemical and stable isotope contents of oxygen and hydrogen in the water molecules, and of carbon in the TDIC in Karachi seawater, used in evaluation of coastal groundwater recharge characteristics are shown in Table 1. The mildly alkaline pH values of ∼8 for open seawater off Karachi coast generally conform to those for normal ocean waters. EC values for Karachi seawater range between 49.3 and 53.7 mS/cm, while the salinity values are ∼38. The EC values higher than 50 mS/cm correspond to relatively non-polluted open seawaters. The seawater temperature off Karachi coast is fairly constant about 25.5◦ C. Slightly higher temperature is observed near Ghizri coast, which is attributed to input of relatively warmer wastewaters of industrial and domestic origin. Turbidity values of open seawater are higher than the on-shore surface water sources. This is attributed to much higher contents of partic-
388
A. Mashiatullah et al.
ulate matter in seawater as compared to on-shore surface water sources. The lowest turbidity values are observed along Clifton coast, as this coast is relatively free of pollution along southeast side of the Karachi coast. Cl− contents of seawater off Karachi coast range from 21,500 to 25,230 ppm, while the SO−2 4 concentrations are in the range of 2080–2210 ppm. The stable 13 carbon isotope contents (δ CTDIC ) of TDIC vary in the range of −3.9 to +0.8h in different zones off Karachi coast. This is indicative of different levels and sources of dissolved inorganic carbon in seawater due to input of domestic and industrial wastewater into the sea from key industrial trading estates via polluted drains. The highest δ 13 CTDIC value of +0.8h and δ 34 S values of 19.1h corresponds to relatively non-polluted seawater along the northwest coast of Karachi (Buleji Coast). The lowest δ 13 CTDIC value of −3.9h corresponds to highly polluted seawater in Korangi Creek that receives industrial and domestic waste drains from KITE. 3.2. Groundwater in the coastal aquifer Shallow groundwater samples (n = 12) were obtained from hand pumps (n = 4), dug wells (n = 1) and shallow mini-bores with centrifugal pumps (n = 7), installed at depths less than 45 m in the coastal aquifer of Karachi. Relatively deep groundwater was obtained from pumping wells (n = 16) installed at depths between 50 and 100 m in the coastal aquifer of Karachi. These cased wells also tap various proportions of shallow groundwater in addition to deep groundwater. Table 1 presents the ranges of physico-chemical and environmental stable isotope data of shallow groundwater samples and shallow mixed deep groundwater samples collected from the municipal jurisdiction of Karachi. 3.2.1. Shallow groundwater Physico-chemical data show that shallow groundwater is moderately saline. The mean chemi18 2 cal concentrations of Cl− , HCO− 3 and SO4 and the mean isotope content of O, H in shallow 34 13 groundwater, S of aqueous sulfate of water and C in TDIC are shown in Table 2. The mean stable isotope contents of 18 O and 2 H indicate that the shallow aquifer system is recharged by a mixture of fresh waters of mainly IR and polluted waters of LR and MR and their tributaries, both under natural and artificially induced infiltration conditions. Much depleted values of δ 13 CS.G. (less than −6h) indicate the impact of pollution from LR and Table 2 Results of chemical and isotopic analysis of shallow groundwater samples Groundwater
Concentration
Mean Cl− (shallow groundwater) Mean HCO− 3 (shallow groundwater) Mean SO− 4 (shallow groundwater) Mean δ 18 O (shallow groundwater)
280 ± 380 ppm (n = 12) 360 ± 110 ppm (n = 12) 74 ± 55 ppm (n = 12) −5.98 ± 0.66h V-SMOW (n = 12) −47.62 ± 5.41h V-SMOW (n = 12) −7.52 ± 2.9h PDB (n = 12) 6.21 ± 2.1h CDT (n = 12)
Mean δ 2 H (shallow groundwater) Mean δ 13 C (TDIC-shallow groundwater) Mean δ 34 S (sulfate-shallow groundwater)
Saline intrusion in the coastal aquifer of Karachi, Pakistan
389
MR into the shallow groundwater environment. It is noteworthy that polluted river waters are mixtures of public supply water of local aquifer origin, Hub River (draining spring water and flood water) origin and IR origin. Majority of shallow groundwater samples have Cl/HCO−1 3 ratios higher than 0.66 coupled with a trend towards increase in the Cl− content. Generally, most of coastal shallow groundwater samples have Cl− /HCO−1 3 ratios between 0.6 and 6.36 −1 − (n = 6) (Cl /HCO3 ratios of <0.66 are attributed to groundwater of freshwater origin (rain and river recharge). Higher Cl− /HCO−1 3 ratios in shallow coastal groundwater samples 18 coupled with a trend towards δ O values of local seawater indicate diffusion/intrusion of seawater. These samples indicate that the zone of diffusion of seawater into groundwater has extended to a considerable lateral extent into the coastal aquifer. 3.2.2. Deep groundwater Physico-chemical data of deep groundwater show that the deep wells have relatively higher values of EC and salinity as compared to the shallow wells. Further, the deep groundwater is quite saline. The mean chemical concentrations of Cl− , HCO−1 3 and SO4 and the mean isotope content of 18 O, 2 H and 34 S in shallow mixed deep groundwater, and 13 C in TDIC in shallow mixed deep groundwater are shown in Table 3. The hydrochemical and stable isotope results indicate that the confined aquifer hosts a mixture of rainwater from hinterlands and surrounding regions around coastal Karachi, as well as sea trapped water/seawater through intrusion under natural infiltration conditions, or under induced recharge conditions. 3.3. Groundwater recharge characteristics and seawater intrusion 3.3.1. Origin of coastal groundwater recharge The analysis of natural variations of the heavy isotope contents of oxygen (18 O) and hydrogen (2 H) represents one of the classical applications of isotope hydrology in studying the origin and dynamics of groundwater. Due to their different contents observed in groundwater and seawater, 18 O and 2 H have been used to assess or confirm seawater intrusion as the main mechanism of salinization. Figure 2 shows the δ 18 O versus δ 2 H plot of groundwater in coastal Karachi vis-à-vis isotopic indices of local precipitation (rain), Indus River water and local Table 3 Results of chemical and isotopic analysis of deep groundwater samples Groundwater
Concentration
Mean Cl− (shallow mixed deep groundwater) Mean HCO− 3 (shallow mixed deep groundwater) Mean SO4 (shallow mixed deep groundwater) Mean δ 18 O (shallow mixed deep groundwater) Mean δ 2 H (shallow mixed deep groundwater) Mean δ 13 CTDIC (shallow mixed deep groundwater) Mean δ 34 SSO4 (shallow mixed deep groundwater)
3900 ± 400 ppm (n = 16) 340 ± 160 ppm (n = 16) 520 ± 530 ppm (n = 15) −5.0 ± 1.0h V-SMOW (n = 16) −43.0 ± 11.0h V-SMOW (n = 16) −8.47 ± 3.0h PDB (n = 16) 6.85 ± 4.9h CDT (n = 16)
390
A. Mashiatullah et al.
Fig. 2. δ 18 O versus δ 2 H plot for groundwater, coastal Karachi (GMWL = Global Meteoric Water Line).
seawater. The long term mean stable isotopic indices of 18 O and 2 H in precipitation as well as the Indus River fall just below the Global Meteoric Water Line (GMWL), while the surface water samples (Hub Lake water and the river water, local seawater) and the groundwater samples fall well below the GMWL. There is a large scatter in the stable isotope plot of oxygen vs. hydrogen for shallow groundwater and deep groundwater. The δ 18 O vs. δ 2 H values of shallow groundwater and deep groundwater mainly range between the mean δ 18 O vs. δ 2 H values of the Indus River (δ 18 O = −5.9h and δ 2 H = −48.12h) and mean δ 18 O vs. δ 2 H values of precipitation (δ 18 O = −3.9 ± 1.94h and δ 2 H = −23.5 ± 18.1h), as well as shallow seawater along Karachi coast (δ 18 O = +0.76h and δ 2 H = 6.85h). The shallow groundwater samples cluster around the mean δ 18 O value of −5.98 ± 0.65h. The deep groundwater samples cluster around the mean δ 18 O value of −5.0 ± 1.05h. Isotopically, some of these deep groundwater samples partly overlap the shallow groundwater. Nevertheless, the large shift in δ 18 O values of shallow and less deep groundwater towards right of Local Meteoric Water Line (LMWL) is attributed to mixing of various proportions of recharge of local precipitation and the mixed waters from polluted rivers and seawater. The possibilities of major contribution to groundwater recharge of shallow/phreatic aquifer directly by the local rainfall seems very small due to extremely poor frequency of rainfall events, the rainfall intensities in Karachi and high evaporation rates. The long term (15 yr annual record) mean monthly average precipitation for Karachi is between 0–15 mm during the months of January to June, 23–91 mm during the months of July to September and 0–7 mm during the months of October to December (IAEA, 1992). Due to the current drought conditions in the area, the direct recharge from precipitation is negligible. This leaves seawater as the 2nd major source of recharge to coastal groundwater system. Under the present surface water supply practices, drought conditions and significant withdrawal of groundwater, the stable isotope index of 18 O and 2 H in shallow and less deep groundwater may shift towards the isotopic index of rainwater/seawater in association with higher groundwater salinities.
Saline intrusion in the coastal aquifer of Karachi, Pakistan
391
3.3.2. Origin of coastal groundwater salinity Gonfiantini and Araguás (1988) compiled a number of case studies on seawater intrusion carried out under different hydrogeological settings. The examples described in their paper illustrate both simple and complex approaches used to characterize coastal aquifers and to shed light on the nature of salinization. Yurtsever and Payne (1978) presented a typical case of interactions between three components or water types, in which most of the observed increase in salinity was due to the upward leakage of deep groundwater, and only in few cases a direct seawater intrusion was active. In other cases, these tools were used to show a complex geochemical behavior of dissolved salts in groundwater systems, due to water rock-interaction (adsorption, cation-exchange), long residence time of groundwater in the aquifer or the presence of deep brines (connate water). From the Ghijben–Herzberg relationship, the freshwater–saltwater equilibrium requires that the water table or the piezometric surface lies above the sea level and slopes downward towards the ocean. Freeze and Cherry (1997) described a simple and useful scheme for recognizing groundwater salinity, based on the contents of Total Dissolved Solids (TDS). Seawater intrusion into coastal aquifers is usually assumed when an increase in (TDS) or EC of the extracted groundwater is observed. Seawater encroachment is the most common mechanism operating in well fields located at short distance from the coastline, mainly due to intensive pumping in selected sites. The TDS of the seawater is approximately 35000 ppm. Surface water or groundwater containing more than 2000–3000 ppm of TDS is generally too salty to drink. Cl−1 usually plays a minor role in groundwater, but it is a dominant ion of seawater. In contrast, HCO−1 3 is usually the most abundant negative ion in groundwater. There is such a large difference between the proportions of Cl− and HCO−1 3 in groundwater and in sea− water, that the ratio between these two ions (i.e. the ratio: Cl /HCO−1 3 ) is a useful index of the presence of seawater in groundwater. Hence, an increase of chloride content in groundwater is the most reliable indicator of the first stage of salt-water intrusion in groundwater. Goswani (1969) presented field studies at Digha (India) and, making use of Cl−1 content of water samples, he delineated the groundwater body on isochlor of 500 ppm (TDS = 100 ppm) and Cl− /HCO−1 3 ratio of 0.66. The boundary zone of the 300–500 isochlors demarcates the diffusion zone of seawater (Freeze and Cherry, 1997). To verify possible seawater intrusion in shallow groundwater and mixed deep groundwater and/or existence of trapped seawater in deep groundwater, the concentrations of Cl− , as −2 18 well as Cl− /HCO−1 3 ratios, are plotted against δ O values, and the concentrations of SO4 −2 − are plotted against SO4 /Cl ratios for shallow groundwater samples in Figs. 3, 4 and 5, respectively. It may be realized that the extrapolated or forecast trends for shallow groundwater samples do not fall on the data points for local seawater (or other tropical seawaters). However, the extrapolated or forecast trends for shallow groundwater samples, excluding the near shore coastal groundwater sample, project towards higher salinities. Lastly, the extrapolated − or forecast trends for shallow mixed deep groundwater samples (with high SO−2 4 and Cl 18 contents, Cl− /HCO−1 3 ratios greater than 0.66 and relatively enriched δ O values) project towards very high salinities, and tend to pass through the data points for local seawater (or other tropical seawater). The present investigations, therefore, prove seawater intrusion/existence of trapped seawater salinity and build-up of salt-water up-coning in the shallow and deep confined aquifer in coastal Karachi.
392
A. Mashiatullah et al.
Fig. 3. Chloride versus 18 O trend line for shallow and deep aquifers, coastal Karachi.
18 Fig. 4. Trend line of Cl− /HCO−1 3 ratios versus O shallow and deep groundwater in coastal Karachi.
−2 − Fig. 5. Trend line of SO−2 4 versus SO4 /Cl ratios for shallow and deep groundwater in coastal Karachi.
Saline intrusion in the coastal aquifer of Karachi, Pakistan
393
4. Conclusion The studies carried out with conjunctive use of stable isotope techniques and conventional non-nuclear chemical and biological techniques have successfully facilitated recognition of seawater intrusion in the shallow groundwater and deep groundwater systems in coastal Karachi. The present investigations prove seawater intrusion/existence of trapped seawater salinity and build-up of salt-water up-coning in the shallow and deep confined aquifers in coastal Karachi.
Acknowledgements These studies have been performed with financial assistance provided by the International Atomic Energy Agency and Pakistan Institute of Nuclear Science and Technology/Pakistan Atomic Energy Commission under IAEA-Research Contract PAK-11322. Special thanks are due to Dr. E. Ahmed for in-kind provision of base camp laboratory facilities and partial transport for fieldwork, and to Dr. M. Ahmed and Mr. M.R. Sheikh for facilitating chemical analysis.
References Clark, I.D., Fritz, P. (1997). Environmental Isotopes in Hydrology. Lewis Publishers, New York, 328 pp. (Chapters 1 and 10). Freeze, R.A., Cherry, J.A. (1997). Groundwater. Prentice Hall, Englewood Cliffs, NJ, 604 pp. Gonfiantini, R., Araguás, L. (1988). Los isótopos ambientales en el estudio de la intrusión marina. In: LopezCamacho Camacho, B. (Ed.), Tecnología de la Intrusión Marina en acuíferos costeros. IGTE, Almuñecar, Spain, pp. 135–190. Goswani, A.B. (1969). Study of salt-water encroachment in the coastal aquifer at Digha, Midnapore district west Bengal. Bulletin of International Assocociation of Hydrological Sciences, India 13, 77. International Atomic Energy Agency (1992). Statistical Treatment of Data on Environmental Isotopes in Precipitation. STI/DOC/10/331. IAEA Technical Reports Series No. 331. IAEA, Vienna, Austria, ISBN92-0-100892-9, 781 pp. Klein, H., Ratzlaff, K.W. (1989). Changes in saltwater intrusion in the Biscayne aquifer, Hialeah–Miami springs area, Dade County, Florida. U.S. Geological Survey Water-Resources Investigations Report No. 87-4249. Kulkarni, K.M., Navada, S.V., Nair, A.R., Rao, S.M., Shivanna, K., Sinha, U.K., Sharma, S. (1979). Drinking water salinity problem in Coastal Orissa–India – Identification of past transgression of sea by isotope investigation. IAEA-SM-349/18. In: Proceedings of the International Symposium on Isotope Techniques in the Study of Past and Current Environmental Changes in the Hydrosphere and Atmosphere. Vienna, Austria. Qureshi, R.M., Mashiatullah, A., Rizvi, S.H.N., Khan, S.H., Javed, T., Tasneem, M.A. (2001). Marine pollution studies in Pakistan by nuclear techniques. The Nucleus 38, 41–52. Yurtsever, Y., Payne, B.R. (1978). Application of environmental isotopes to groundwater investigations in Qatar. IAEA-SM-228/24. In: Proceedings of the International Symposium on Isotope Hydrology. Neuherberg, Germany.
This page intentionally left blank
8. Coastal radionuclide studies
This page intentionally left blank
397
Temporal variations and behaviour of 90Sr and 137 Cs in precipitation, river water and seawater in Japan Yoshihiro Ikeuchi* Japan Chemical Analysis Center, Chiba 263-0002, Japan Abstract The 90 Sr and 137 Cs concentrations in precipitation (wet and dry deposition), in soil, river water, coastal surface seawater and seabed sediment were determined during 1963–1999 at 10–47 sites in Japan. The 90 Sr and 137 Cs concentrations in precipitation decreased during this time by more than 3 orders of magnitude, showing 4 peaks after the large-scale Chinese atmospheric nuclear weapons tests, and one sharp peak of 137 Cs due to the Chernobyl accident in 1986. The temporal variations in soil (0–20 cm depth), river water, seawater and seabed sediment had decreased gradually. Specific features were an increase of 137 Cs/90 Sr ratio in soil with time, higher 90 Sr than 137 Cs concentrations observed in river water, a lower 137 Cs/90 Sr ratio in seawater than in precipitation, and higher 137 Cs than 90 Sr concentrations in seabed sediment. The 90 Sr and 137 Cs cumulative depositions by precipitation were calculated from 1963 to 1999, having maximum values of 2,600 MBq/km2 and 3,700 MBq/km2 in 1965, respectively. 27% of 137 Cs and 51% of 90 Sr were removed from soil (0–20 cm depth) during 1980–1999. However, when comparing cumulative depositions by precipitation and concentrations in coastal surface seawater, the removal rate of 137 Cs was during 1980–1999 about 1.2 times higher than for 90 Sr, documenting that more 137 Cs was removed to seabed sediment, and more 90 Sr came to coastal waters via rivers. Keywords: Environmental radioactivity, Radionuclides, 90 Sr, 137 Cs, Temporal variation, Precipitation, Soil, Seawater, Sediment, Japan coast
1. Introduction Large amounts of 90 Sr and 137 Cs have been deposited globally due to fallout from atmospheric nuclear weapons tests conducted by USA, USSR, UK, France and China during 1945–1980, with maximum annual deposition in 1963. A smaller amount of 137 Cs was deposited over Japan after the Chernobyl accident which occurred in 1986. Therefore, large amounts of 90 Sr and 137 Cs were introduced into soil, river water and seawater via precipitation (dry and wet deposition). It has been therefore important to investigate the behaviour and fate of these * Address: Japan Chemical Analysis Center, 295-3, San-no-cho, Inage, Chiba 263-0002, Japan; phone: (+81) 43 424 8661; fax: (+81) 43 423 5326; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08032-0
© 2006 Elsevier Ltd. All rights reserved.
398
Y. Ikeuchi
radionuclides in order to assess the degree of contamination of the terrestrial and marine environments. Environmental samples were collected annually by the Japanese prefectural institutes from 1963–1999. Subsequent analyses of 90 Sr and 137 Cs were performed by governmental institutes, mainly by the Japan Chemical Analysis Center (JCAC). The resulting data has since been archived at the Ministry of Education, Culture, Sports, Science and Technology (MEXT); formerly the Science and Technology Agency (STA) of Japan. In this paper, temporal variations of 90 Sr and 137 Cs concentrations in precipitation, soil, river water, coastal surface seawater and seabed sediment during 1963–1999 are described. The behaviour of 90 Sr and 137 Cs in these environmental systems are discussed as well, specifically their transport between different compartments, e.g. between precipitation and soil, precipitation and seawater, soil and river water, river water and seawater, and finally between seawater and seabed sediment.
2. Sampling and analytical procedures Precipitation samples were collected during 1963–1999 once per month at 22–47 sites, soil samples were collected once per year at 30–47 sites, river water samples were collected twice per year at 29–47 sites, seawater and seabed sediment samples were collected once per year at 10–13 sites (9 sites on the Pacific side, 4 sites on the sea of Japan side). The sampling sites of seawater and seabed sediment were the same. 90 Sr and 137 Cs, collected from 0.5 m2 surface areas of monthly precipitation, 100 g of soil, 100 l of river water, 40 l of seawater and 100 g of seabed sediment were determined mainly in the Japan Chemical Analysis Center. 90 Sr was separated as carbonate, oxalate and nitrate. Ba and Ra impurities were removed as chromates. After scavenging of 90 Y, SrCO3 was recovered and weighed to obtain the chemical yield. After 2 weeks, 90 Y separated from the 90 Sr was analysed using a low background gas flow counter. 137 Cs in the supernatant solution at carbonate precipitation phase was separated as Cs3 PO4 · 12MoO3 , and 87 Rb impurity was removed by the cation exchange method. Cs2 PtCl6 was precipitated and weighed to obtain chemical yield. 137 Cs was analysed using a low background gas flow counter. The detection limits for both 90 Sr and 137 Cs (3σ criterion in counting statistics) were ∼0.2 mBq/l for river water, ∼0.5 mBq/l for seawater and ∼0.4 Bq/kg for seabed sediment.
3. Results and discussion The results are presented as the average values (arithmetic means) in order to investigate the behaviour and fate of 137 Cs and 90 Sr over the territory of Japan. The observed 90 Sr and 137 Cs concentrations in the environment were different due to sampling at different latitudes, or sampling either on the side of the Japan Sea, or on the Pacific side. The concentrations of 90 Sr and 137 Cs in precipitation decreased by more than 3 orders of magnitude, showing 4 peaks after the large-scale Chinese atmospheric nuclear weapons tests, and one sharp peak of 137 Cs
90 Sr and 137 Cs in precipitation, river water and seawater in Japan
399
due to the Chernobyl accident in 1986 (Fig. 1). On the other hand, the concentrations in soil (0–20 cm depth), river water, seawater and seabed sediment had decreased gradually, showing 137 Cs levels of 2.0–4.5, 0.03–0.57, 0.98–2.0 and 9.2–33 times higher than 90 Sr levels, respectively (Figs. 2–5). The interesting features that can be observed from the presented data sets (Figs. 2–5) include: • • • •
the increase of 137 Cs/90 Sr ratio in soil with time (from 2.0 to 4.5); higher 90 Sr than 137 Cs concentrations in river water; lower 137 Cs/90 Sr ratios in seawater than in precipitation; higher 137 Cs than 90 Sr concentrations in seabed sediment.
Fig. 1. Average 90 Sr and 137 Cs depositions by precipitation at 22–47 sampling sites.
Fig. 2. Average 90 Sr and 137 Cs depositions in soil (0–20 cm depth) at 30–47 sampling sites.
400
Y. Ikeuchi
Fig. 3. Average 90 Sr and 137 Cs concentrations in river water at 29–47 sampling sites.
Fig. 4. Average 90 Sr and 137 Cs concentrations in seawater at 10–13 sampling sites.
Fig. 5. Average 90 Sr and 137 Cs massive activities in seabed sediment at 10–13 sampling sites.
90 Sr and 137 Cs in precipitation, river water and seawater in Japan
401
The 90 Sr and 137 Cs cumulative depositions by precipitation in 1962 (Figs. 6 and 7) were calculated from the data for the Northern Hemisphere (UNSCEAR, 1982), which were updated using more recent data published by Monetti (1996) comparing the analysed data in 1963 (Fig. 1). The 90 Sr and 137 Cs cumulative depositions by precipitation during 1963–1999 (Figs. 6 and 7) were calculated from analysed data (Fig. 1) considering decay corrections for 90 Sr and 137 Cs. It was found that maximum 90 Sr and 137 Cs cumulative depositions by precipitation were in 1965: 2,600 MBq/km2 for 90 Sr and 3,700 MBq/km2 for 137 Cs. List et al. (1965) reported for the maximum 90 Sr cumulative deposition in 1963 and early 1964 on soil at 30–45◦ N values between 1,000 and 1,600 MBq/km2 . About 2,200 MBq/km2 of 90 Sr was deposited on Japanese soil in 1965–1967 according to results published by Meyer et al. (1968), in a reasonable agreement with our value.
Fig. 6. 137 Cs cumulative deposition for precipitation and soil.
Fig. 7. 90 Sr cumulative deposition for precipitation and soil.
402
Y. Ikeuchi
The 137 Cs/90 Sr activity ratio for the maximum cumulative deposition is 1.4 ± 0.1, a value close to 1.6, derived in the UNSCEAR (1982) Report. The cumulative deposition of 137 Cs by precipitation in 1972 and 1975 was almost the same as the observed deposition in soil (Fig. 6). On the contrary, for 90 Sr the observed cumulative deposition in soil was only 0.6 times of the cumulative deposition by precipitation (Fig. 7). Comparing the 137 Cs/90 Sr ratios in soil (0–20 cm depth) with 137 Cs/90 Sr ratios in cumulative deposition, we see that the ratios increased from 1.2 in 1962 to 2.4 in 1999 (Fig. 8). This indicates that 90 Sr has been removed from soil (0–20 cm depth) much faster than 137 Cs. Comparing cumulative depositions by precipitation and in soil after the Chinese atmospheric nuclear weapons tests, we find that in average 27% of 137 Cs and 51% of 90 Sr has been removed from soil during 1980–1999 (Figs. 9 and 10). This documents that the removal rate of 90 Sr from soil was about twice than that for 137 Cs.
Fig. 8. Average 137 Cs/90 Sr ratios in precipitation and soil.
Fig. 9. 137 Cs correlation – soil vs. precipitation.
90 Sr and 137 Cs in precipitation, river water and seawater in Japan
403
Fig. 10. 90 Sr correlation – soil vs. precipitation.
Fig. 11. 137 Cs correlation – seawater vs. precipitation.
A comparison of cumulative 90 Sr and 137 Cs deposition by precipitation and corresponding concentrations in seawater is more difficult, because we are dealing with different units (actually, water column and sediment inventories should be used in such comparison). Knowing the ratio of 90 Sr (and 137 Cs) in precipitation and seawater, and comparing 1/100 of cumulative deposition by precipitation and concentration of seawater (for convenience), we find that during 1980–1999, after the Chinese atmospheric nuclear weapons tests, around 55% of 137 Cs and 47% of 90 Sr have been removed from surface seawater (Figs. 11 and 12). The removal rate of 137 Cs was thus 1.2 times higher than that of 90 Sr. This ratio indicates that more 137 Cs than 90 Sr has been removed from seawater to seabed sediment (as expected, because Cs is more particle reactive and its Kd – distribution coefficient is higher than for Sr). This ratio, however, could also be influenced by the fact that more 90 Sr than 137 Cs has been transported by rivers to coastal waters.
404
Y. Ikeuchi
Fig. 12. 90 Sr correlation – seawater vs. precipitation.
4. Conclusions Several conclusions can be made from the present study: • Maximum cumulative deposition by precipitation in Japan in 1965 was 3,700 MBq/km2 and 2,600 MBq/km2 for 137 Cs and 90 Sr, respectively. • Comparing cumulative depositions for precipitation and soil during 1980–1999, 27% of 137 Cs and 51% of 90 Sr were removed from soil after the Chinese atmospheric nuclear weapons tests. The removal rate of 90 Sr was found to be about twice than that of 137 Cs. • Comparing cumulative depositions for precipitation and coastal surface seawater during 1980–1999, the removal rate of 137 Cs was found to be about 1.2 times than that of 90 Sr, probably due to the fact that more 137 Cs was removed to seabed sediment.
Acknowledgements The author wishes to thank the staff of prefectural institutes who collected environmental samples, as well as the colleagues of the JCAC who assisted in the analysis of samples. He would also like to express his gratitude to Prof. H. Hirao (Japan Chemical Analysis Center), and Prof. P.P. Povinec (International Atomic Energy Agency–Marine Environment Laboratory) for support and suggestions. The described research work was funded by the Ministry of Education, Culture, Sports, Science and Technology (MEXT, former the Science and Technology Agency (STA)) of Japan.
References List, R.J., Machta, L., Alexander, L.T., Allen, J.S., Meyer, M.W., Valasis, V.T., Hardy Jr., E.P. (1965). Strontium-90 on the Earth’s surface. In: Klement Jr., A.W. (Ed.), Radioactive Fallout from Nuclear Weapons Tests. AEC Symposium Series, No. 5 CONF-765. AEC, Washington, DC, pp. 359–368.
90 Sr and 137 Cs in precipitation, river water and seawater in Japan
405
Meyer, M.W., Allen, J.S., Alexander, L.T., Hardy, E. (1968). Strontium-90 on the Earth’s Surface. Summary and Interpretation of a World-Wide Soil Sampling Program: 1961–1967 Results. USAEC Report TID-24341. Health and Safety Laboratory. Monetti, A. (1996). Worldwide Deposition of Strontium-90 through 1990. USDOE Report EML-579. EML, New York, 56 pp. UNSCEAR (1982). Ionizing radiation: Sources and biological effects. United Nations Scientific Committee on the Effects of Atomic Radiation. Report to the General Assembly, United Nations, New York.
Further reading IAEA (1986). Summary report on the post-accident review meeting on the Chernobyl accident. Safety Series No. 75INSAG-1. IAEA, Vienna, pp. 33–34. Norris, R.S., Burrows, A.S., Fieldhouse, R.W. (1994). British, French and Chinese Nuclear Weapons. Nuclear Weapons Databook, vol. V. Natural Resources Defense Council (NRDC), Washington, DC.
406
Isotope fractionations and radiocarbon ages of beach rock samples collected from the Nansei Islands, southwest of Japan Kunio Omoto* Department of Geography, College of Humanities and Sciences, Nihon University, Tokyo, Japan Abstract Four hundred and twelve beach rock samples were collected from 157 sites on 29 islands consisting of the Nansei Islands, in the southwestern part of Japan. They were radiocarbon dated and their isotope fractionations (δ 13 C) were analyzed. The values of isotope fractionations of different materials making up the beach rocks ranged between +9.4h and −6.0h with an average of 1.99h. The isotope fractionations outside the range of 0 ± 2h suggest that these beach rocks were subjected to the influence of meteoric water when they cemented. Radiocarbon ages suggest that beach rocks in the Nansei Islands began to form about 6,900 yr BP and some of them are at present still under development at the islands. Keywords: Isotope fractionation, δ 13 C, Radiocarbon dating, Beach rock, Sea level change, Holocene, Nansei Islands, Japan
1. Introduction Beach rocks are commonly formed on tropical and subtropical sandy beaches where they represent small-scale “cuesta” topography with thin slabs (strata) that dip seaward at less than 15 degrees. They are believed to have been formed within the range of the inter-tidal zone, where beach sediments, including fragments of shell, coral, foraminifera and other marine biocarbonate have been cemented by calcium carbonate and magnesian calcite (Fig. 1). Therefore, they not only provide good sample material for radiocarbon dating but also provide a good indicator of past sea level changes. Scoffin and Stoddart (1983) summarized that the origin of beach rock cements fell into three main categories (or combinations of these): (i) seawater, (ii) meteoric water and (iii) biological activity. Moore (1973) reported on the stable isotope composition of beach rocks collected from Grand Cayman which were enriched in 13 C when compared to Gulf Stream marine bicarbonate, however, it was well within the range of reported values from marine cements and * Address: Department of Geography, Nihon University, 24-40, 3 Chome, Sakurajosui, Setagaya-Ku, Tokyo 1568550, Japan; phone: (+81) 35317 9723; fax: (+81) 35317 9429; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08033-2
© 2006 Elsevier Ltd. All rights reserved.
Isotope fractionations and radiocarbon ages of beach rock samples of Nansei Islands
407
Fig. 1. Beach rock observed at Okinawa Prefecture, southwest Japan.
ooliths. The isotope fractionation of beach rocks was considered one of a key to solving the origin of calcium carbonate (Omoto, 2001). If the calcium carbonate originates in seawater, it must indicate nearly the same isotope fractionation of marine organisms, otherwise it was brought by meteoric water, or by both interactions. The aim of this paper is to identify the range of isotope fractionation for beach rock samples and to consider the origin of calcium carbonate. Another purpose is to reconstruct the geo-history of the Nansei Islands, especially in the formative periods of beach rocks, viz. late Holocene sea-level change and tectonic movements, based on the radiocarbon ages and evidences obtained from the field surveys.
2. Description of samples and analyses 2.1. Samples The Nansei Islands are located between Kyusyu Island and Taiwan Island, and consist of 3 major islands groups, namely Amami Islands, Okinawa Islands and Sakishima Islands (Fig. 2). On the sandy beaches along the Nansei Islands, beach rocks are frequently observed within the range of the inter-tidal zone. A number of beach rock radiocarbon ages collected from the Nansei Islands have been reported by numerous researchers (Omoto, 2004b), however, a great deal of the dates reported previously (except Kimura et al., 2003) could not be included because they have not been corrected yet for isotope fractionations and the marine reservoir effect. As mentioned by Stuiver and Polach (1977) it is essential to correct marine organism radiocarbon ages in years BP. The newly determined radiocarbon and isotope data were summarized in the previous report (Omoto, 2005). Three different types of sample materials viz. fossil shells, fossil corals and calcarenite (or calcirudite) samples were collected from each sampling site in order to compare their ages.
408
K. Omoto
Fig. 2. Index map of the surveyed islands.
Sample materials for radiocarbon dating were usually collected from the most landward slab (bed), which was believed to have been formed in the initial stage of beach rocks formation. However, it has been difficult to identify them below the present mean sea level, because they have usually been covered with green algae. A total of 412 beach rock samples were collected from 157 sites on 29 islands of the Nansei Islands (Table 1). Two samples collected from Yonaguni Island (Kimura et al., 2003) were added to this paper. The beach rock samples consisted of 146 calcarenite samples (35.3%), 116 fossil corals (28.0%) and 152 fossil shells (36.7%) (Table 1). 2.2. Analyses Analyses of isotope fractionations and radiocarbon dating of samples were carried out at the Radiocarbon Dating Laboratory of Nihon University by the author following the radiocarbon dating manual (Omoto, 1993). Isotope fractionations (δ 13 C) were measured mostly by IsoPrime and partly by Optima or MAT-252 mass spectrometers, using the same CO2 sample as for β-counting (radiocarbon dating). The results ranged between 9.40h and −6.0h, the average value for all samples from the 29 Nansei Islands was 1.99h (Table 2). The radiocarbon ages were corrected for isotope fractionations (δ 13 C), and then 400 yr (Stuiver and Braziunas, 1993) were subtracted as a correction for the mean global ocean reservoir effect. The calculations for the age corrections based on the isotope fractionations of each sample indicated the values between 573 yr and 311 yr, with the average value of 446 yr. It is noteworthy that the average value of isotope fractionations of 187 samples (∼45.7%) over 412 beach rock samples was outside the range of 0 ± 2h (Fig. 3). These samples consist of 67 calcarenite samples (46.2%), 85 fossil corals (73.3%) and 35 fossil shells (23.2%).
Isotope fractionations and radiocarbon ages of beach rock samples of Nansei Islands
409
Table 1 Beach rocks collected from the Nansei Islands. Compiled by Omoto (2005) Name of island Amami Islands Amami-Ohshima Island Kakeroma Island Yoro Island Tokuno Island Okinoerabu Island Yoron Island Subtotal Okinawa Islands Iheiya Island Izena Island Okinawa Island Ie Island Aguni Island Kume Island Oha Island Oh Island Hatenohama (beach) Zamami Island Tokashiki Island Aka Island Geruma Island Subtotal Sakishima Islands Miyako Island∗ Tarama Island Ishigaki Island∗ Hatoma Island Yonaguni Island∗∗ Iriomote Island Kohama Island Taketomi Island Kuro Island Hateruma Island Subtotal Total
Calcarenite
Coral
Shell
Total
Sites no.
12 2 0 6 11 13
7 0 1 4 10 4
7 0 0 6 5 4
26 2 1 16 26 21
11 2 1 4 11 12
44
26
22
92
41
4 3 10 13 3 5 0 7 7 2 0 0 0
7 1 23 2 1 9 0 2 1 4 1 1 1
8 3 27 11 2 14 3 3 1 12 1 1 0
19 7 60 26 6 41 3 12 9 18 2 2 1
6 4 28 6 4 2 3 3 5 6 2 2 1
54
47
77
178
72
16 5 7 0 3 5 1 3 5 3
15 1 18 0 2 3 0 2 0 2
16 2 16 2 2 4 1 1 1 8
47 8 41 2 7 12 2 6 6 13
12 5 10 1 3 5 1 2 3 3
48
43
53
144
45
146
116
152
414
157
∗ Calcirudite samples are included. ∗∗ Two samples reported by Kimura et al. (2003) are included.
The maximum and minimum values of isotope fractionation for calcarenite samples were 9.4h and −3.3h respectively, while the average value was 1.71h (Table 2). The maximum and minimum values of isotope fractionation for coral samples were 7.04h and −6.0h respectively, the average value was 1.54h. In addition, the maximum, minimum and average
410
Table 2 Statistical data of isotope fractionations of beach rocks collected from the Nansei Islands. Compiled by Omoto (2005) Island/mat.
Amami Islands Amami-Oshima Island Kakeroma Island Yoro Island Tokuno Island Okinoerabu Island Yoron Island Okinawa Islands Iheiya Island Izena Island Okinawa Island Ie Island Aguni Island Kume Island Oha Island Oh Island Hatenohama Zamami Island Tokashiki Island Aka Island Geruma Island Subtotal
Coral
no.
max
min
avg
12 2 0 6 11 13
2.05 2.2 – 2.62 3.18 3.82
−2.1 1.94 – −3.3 0.15 1.57
0.382 2.07 – 1.208 2.338 2.624
44
3.8
−3
4 3 10 13 3 5 0 7 7 2 0 0 0
2.18 1.42 5.63 2.89 1.59 3.43 – 2.89 4.06 3.15 – – –
54
5.6
no.
Shell max
min
avg
7 0 1 4 10 4
1.79 – −1.7 1.68 4.28 3.09
−1.1 – −1.7 −3.3 −0.9 1.34
1.723
26
4.3
−2.3 0.53 0.6 0.15 1.18 1.2 – 0.17 1.21 2.28 – – –
1.005 1.023 2.168 1.902 1.317 1.97 – 1.714 2.897 2.715 – – –
6 1 23 2 1 7 0 2 1 1 1 1 1
3.86 −3.5 2.48 3.51 −0.2 2.31 – 1.33 1.1 2.74 −0.6 0 −1.2
−2
1.945
47
3.9
−3 1.96 −3.5 −6 0.63 −0.2 −2.1 – 0.62 1.1 2.74 −0.6 0 −1.2 −6
Total
no.
max
min
avg
0.481 – −1.65 1.303 2.113 2.283
7 0 0 6 5 4
3.35 – – 4.67 4.77 4.27
−0 – – 2.74 3.33 2.77
1.223 – – 3.672 4.004 3.42
1.431
22
4.8
−0
2.887 −3.49 0.04 2.07 −0.16 0.284 – 0.975 1.1 2.74 −0.59 0 −1.23
9 3 27 11 2 7 3 3 1 10 1 1 0
0.5 1.15 4.11 6.03 3.42 2.54 3.06 3.37 1.12 5.07 2.89 3.27 –
0 −5.2 −4.7 0.95 2.72 0.52 2.21 1.12 1.12 1.18 2.89 3.27 –
0.525
77
6
−5
no.
max
min
avg
26 2 1 16 26 21
3.35 2.2 −1.7 4.67 4.77 4.27
−2.1 1.94 −1.7 −3.3 −0.9 1.34
0.635 2.07 −1.65 2.156 2.572 2.71
2.92
92
4.8
−3.3
0.275 −1.117 1.444 2.973 3.07 1.983 2.763 2.053 1.12 3.323 2.89 3.27 –
19 7 60 26 6 19 3 12 9 13 2 2 1
3.86 1.42 5.63 6.03 3.42 3.43 3.06 3.37 4.06 5.07 2.89 3.27 −1.2
−2.3 −5.2 −6 0.15 −0.2 −2.1 2.21 0.17 1.1 1.18 −0.6 0 −1.2
1.89
178
6
−6
1.927 2.317 0.54 1.027 2.368 1.655 1.354 2.763 1.676 2.5 3.185 1.15 1.635 −1.23 1.548
K. Omoto
Subtotal
Calcarenite∗
Island/mat.
Sakishima Islands Miyako Island Tarama Island Ishigaki Island Hatoma Island Yonaguni Island∗∗ Iriomote Island Kohama Island Taketomi Island Kuro Island Hateruma Island Subtotal Total
Calcarenite∗
Coral no.
Shell
no.
max
min
avg
16 5 7 0 3 5 1 3 5 3
2.44 2.38 9.4 – 2.35 8.4 1.82 4.11 1.95 3.3
−2.5 −0.2 3.45 – 0.67 −0.4 1.82 2.95 0.98 2.48
0.855 1.612 6.737 – 1.313 3.31 1.82 3.337 1.544 2.83
48
9.4
−3
2.446
43
7
−3
146
9.4
−3
1.71
117
7
−6
15 1 18 0 2 3 0 2 0 2
∗ Include calcirudite. ∗∗ Two samples reported by Kimura et al. (2003) are included.
max
min
avg
2.66 −0.3 7.04 – 3.57 2.44 – 4.09 – 1.81
−2 −0.3 −3.3 – 0.1 0.51 – 1.74 – 1.4
0.445 −0.29 1.519 – 1.835 1.683 – 2.915 – 1.605
Total
no.
max
16 2 16 2 2 4 1 1 1 8
6.98 3.2 9.4 4.93 6.7 8.37 2.53 5.16 2.91 4.45
1.197
53
9.4
1.54
152
9.4
min
avg
no.
max
avg
−2.5 −0.3 −3.3 3.41 0.1 −0.4 1.82 1.74 0.98 1.4
1.709 1.679 3.25 4.17 2.373 3.618 2.175 3.5 1.772 3.039
9.4
−3.3
2.575
9.4
−6
1.989
1.36 2.46 0.11 3.41 2.3 3.5 2.53 5.16 2.91 2.09
3.749 2.83 3.671 4.17 4.5 5.455 2.53 5.16 2.91 3.476
47 8 41 2 7 12 2 6 6 13
6.98 3.2 9.4 4.93 6.7 8.4 2.53 5.16 2.91 4.45
0.1
3.81
144
3.04
414
−5
min
Isotope fractionations and radiocarbon ages of beach rock samples of Nansei Islands
Table 2 (Continued)
411
412
K. Omoto
Fig. 3. Isotope fractionations vs. radiocarbon ages of beach rock samples collected from the Nansei Island. Compiled by Omoto (2005).
values of isotope fractionation for fossil shells were 9.4h, −5.2h and 3.04h, respectively (Table 2). Figure 3 illustrates the relationship between radiocarbon ages and isotope fractionations of beach rock samples collected from the Nansei Islands.
3. Discussions 3.1. Isotope fractionation The average value of isotope fractionation (δ 13 C) for marine organisms and carbonates is within a range of 0 ± 2h (Geyh and Schleicher, 1990). Although the average isotope fractionation of beach rock samples collected from the Nansei Islands was 1.99h, the average isotope fractionation for samples from 15 islands is outside the range of 0 ± 2h (the upper limit for marine organisms and biocarbonates). In addition, the maximum value of the isotope fractionation for a beach rock sample taken from Ishigaki Island was +9.4h (Table 2 and Fig. 4). The values presented in this figure are more than 4 times larger than those reported by Geyh and Schleicher (1990). Figure 5 shows a histogram of δ 13 C values of beach rock samples collected from the Nansei Islands. The three different materials comprising the beach rocks clearly indicated a separated center of gravity (median value). The median of shell samples (Tridacna squamosa sp.) shifts to a larger δ 13 C values (heavy) while the coral samples shift to smaller δ 13 C values (light). The calcarenite samples showed intermediate values, as they are composed of fine grain size beach sediments including mixed fragments of coral, shell, foraminifera and other marine carbonates.
Isotope fractionations and radiocarbon ages of beach rock samples of Nansei Islands
413
Fig. 4. Histogram of isotope fractionations of beach rock samples collected from the Nansei Islands. Compiled by Omoto (2005). Numbers: 1–91, Amami Islands: 92–269, Okinawa Islands: 92–269, Sakishima Islands: 270–414.
Fig. 5. Histogram of isotope fractionations for three sample materials comprising beach rocks.
414
K. Omoto
3.2. Origin of calcium carbonate The cement materials that formed beach rock could have originated in several ways. Beach rocks were formed in the tropical or subtropical beaches by evaporation of seawater in the inter-tidal zone at times of low tide in which evaporation would cause a rise in the ionic concentration and a consequent precipitation of calcium carbonate cement (Dana, 1851). Later after Dana’s proposal, Field (1919) suggested that a solution of calcium carbonate from inland sands and groundwater followed by precipitation as the solution seeped out through the beach at low tide would have cemented the beach rocks. As reported by Stoddart and Cann (1965), beach rock had aragonite as the primary cement and a secondary filling of voids by calcite. It was considered therefore that aragonite was probably derived from seawater. High-magnesian calcite and aragonite cements, and the magnesian calcite contained in the beach rock were believed to appear both as a primary precipitation and then replacement of aragonite had occurred (Taylor and Illing, 1971). These materials are believed to be still forming today from sea-derived waters. Calcium carbonate could originate not only from seawater but also from meteoric water when beach sediments were cemented (Omoto, 2000). The hypothesis was based on two pieces of evidences, (i) location of beach rocks, and (ii) limestone caves. In the Nansei Islands several beach rocks are observed at levels apparently above the present high-tide levels. They are frequently located in small river mouths and near sand dunes. Otherwise, they are also situated on tectonically uplifted positions. Limestone caves have been developed frequently in the surveyed islands, and continue to develop below the present sea level. Meteoric water that soaked through beach sediments and underground water flows must have influenced the ecosystem of marine organisms when they reached seawater. The δ 13 C values obtained from Okinoerabu Island demonstrate the results of these interactions. In these cases a figure for mean global reservoir correction is no more valid and an isotope correction is not possible. We noted that the carbon cycles between the atmosphere–land sea systems may be of importance for understanding the origin of calcium carbonate, and the significance of the δ 13 C values greater than 0 ± 2h. However, further information and research on the food cycle of the marine organisms is needed to demonstrate this hypothesis. We expects to obtain accurate and reliable data by separating aragonite, calcite and Magnesian calcite under a microscope, and to analyze δ 13 C and to determine their radiocarbon ages by accelerator mass-spectrometry. 3.3. Formative ages of beach rocks The formative ages of beach rocks have been estimated using the radiocarbon ages of fossil coral, shell and foraminifera samples embedded in beach rocks. However, their radiocarbon ages should be interpreted with caution, because these may be transported deposits, older than the beach formation (Pirazzoli, 1996). For example, Russel (1959) reported that an empty cartridge case used in the World War II was found in the beach rock. Yonetani (1963) found a silver-comb in beach rock at Amami Oshima Island, Kagoshima Prefecture, and he estimated that it was produced probably within 200 years. Takenaga (1965) found a fragment of glass of Shoyu bottle at Yoron Island, Kagoshima Prefecture, and he estimated that it was produced within 100 years. He also proposed that the beach rock might be formed within a year or one season.
Isotope fractionations and radiocarbon ages of beach rock samples of Nansei Islands
415
This evidence indicate that beach rocks have been formed within a short time. If it took a long time to form beach rocks, then the materials embedded in beach rocks should indicate different plural radiocarbon ages. But as they have never indicated such different plural ages, the elapsed time should be short, negligible when we consider the formation age of beach rock. One calcarenite sample collected from Bise Point, on the west coast of Okinawa Island, gave a radiocarbon age of 6890 ± 90 yr BP (Omoto et al., 2003). This is the oldest age among the islands surveyed (Fig. 3 and Fig. 5), and it indicates also that the beach rock of the Okinawa Islands began to form at ∼6,900 yr BP. The ages of beach rock formations are different for the islands surveyed, as they were formed continuously between 4,800 yr BP and the present (Fig. 5). 3.4. Late Holocene sea level The influence of tectonic movements in relation to sea level change was recently addressed by Omoto (2004a). Figure 6 illustrates the relationship between whole radiocarbon dates and elevations of beach rock samples collected from the Nansei Islands. It may be assumed that almost all of the beach rocks were formed at a sea level higher than that at present. However, a great deal of beach rock samples was situated within a range corresponding to the present inter-tidal zone (±1 m). We must also consider whether sample materials were collected from a higher or the highest slabs (beds). As previously mentioned, we were unable to collect fossil coral and shell samples located below the present mean sea level, because they were usually covered with green algae.
Fig. 6. Elevations vs. radiocarbon ages of beach rock samples collected from the Nansei Islands. Compiled by Omoto (2005).
416
K. Omoto
A total of 18 beach rocks marked at elevations higher than 2 m were collected from the floor of a limestone cave at Nyatie Gama, on the southern coast of Ie Island (Omoto, 1998). Four other coral samples plotted on lower than −1 m elevation were collected from off Hentona Fishing Port (Omoto, 2004c) where submarine beach rocks were reported (Sunamura, 1983). The former elevations have no relation to the past sea level because they were formed under influence of ground water filtered through the Ryukyu limestone bed. Therefore these beach rock samples should be excluded from the discussion on the past sea-level changes. On the west coast of Okinawa Island, the southern coast of Ie Island and the southeast coast of Kume Island, beach rocks were located at obviously higher elevations above the present high tide level. These facts indicate that the islands have been uplifted by local tectonic movements (Omoto, 1998, 2004a, 2004c), otherwise these beach rocks would have formed in a different environment. Beach deposits were influenced by meteoric water whose calcium carbonate cemented the soft beach deposits and then formed the beach rock (Omoto, 2000, 2003). In this case the elevations of the beach rocks never indicate the past sea level. Based on these pieces of evidence, the past sea level existed at a level similar to the present one since at least 5,000 yr BP, except for several uplifted coasts.
4. Summary A total of 414 beach rock samples (including 2 data of Kimura et al., 2003) were collected from 158 sites on 29 islands of the Nansei Islands. They were radiocarbon dated and their isotope fractionations were measured. The results of this study may be summarized as follows: (i) Isotope fractionation (δ 13 C) of all beach rock samples ranged between +9.40h and −6.0h, with an average of 1.99h. The average values of isotope fractionations of 15 islands are outside a range of 0 ± 2h. These results suggest that beach rocks during their formation were under the influence of an obviously different origin of calcium carbonate occupied by meteoric water. (ii) In the Nansei Islands, beach rocks were formed within the range of the inter-tidal zone between approximately 6,900 yr BP and present. (iii) The elevations and radiocarbon ages of the beach rock samples suggest that many islands have been emerged by neo-tectonic movements. However, the late Holocene sea level has remained similar to the present one since at least the past 5,000 yr BP.
Acknowledgements The author is grateful to Professor Toshio Kawana of Ryukyu University who helped with the bibliographical research on beach rocks, to Dr. Pavel P. Povinec, Scientific Secretary of IAEA, anonymous reviewers and to Prof. William D. Patterson, Department of English Literature, College of Humanities and Sciences, Nihon University for their useful suggestions and improvements of the manuscript. Field surveys, radiocarbon dating and isotope analyses were financially supported by the Nihon University and the Ministry of Education, Culture, Sports, Sciences and Technology.
Isotope fractionations and radiocarbon ages of beach rock samples of Nansei Islands
417
References Dana, J.D. (1851). On coral reefs and islands. American Journal of Science (Ser. II) 11, 357–372. Field, R.M. (1919). Investigations regarding the calcium carbonate oozes at Tortugas and the beach rock at Loggerhead Key. Carnegie Institute of Washington Year Book 18, 197–198. Geyh, M.A., Schleicher, H. (1990). Absolute Age Determination. Springer-Verlag, New York, Berlin, Heidelberg, 503 pp. Kimura, M., Asato, S., Nakamura, T., Sugiyama, M., Matsuura, M. (2003). 14 C age measurement of carbonate samples recovered from Tuguru-hama ruins and beach rocks in Yonaguni Island, Japan. In: Summaries of Researches Using AMS at Nagoya University, vol. XIV. Cent. Chronol. Res. Nagoya Univ., Nagoya, Japan, pp. 170–190. Moore, C.H. (1973). Intertidal carbonate cementation Grand Cayman, West Indies. Journal of Sedimentary Petrology 43 (3), 591–602. Omoto, K. (1993). The radiocarbon dating manual for Nihon University. Dept. Geogr., Coll. Human and Sci. Nihon Univ., 1993, pp. 1–102 (in Japanese). Omoto, K. (1998). Radiocarbon dates of beach rock samples collected from Ie Island, Okinawa Prefecture, Japan – Especially on the beach rocks collected from limestone cave of Nyatie Gama. Proc. Ann. Meet. Jap. Coral Reef Soc. 46 (in Japanese). Omoto, K. (2000). A preliminary report of radiocarbon ages on the beach rock samples collected from Iriomote Island, southwest of Japan. Annals of Geography, The Chiri Shiso 42 (1), 17–30 (in Japanese). Omoto, K. (2001). Radiocarbon age corrections for beach rock samples based on isotope fractionations (a preliminary report). Quartery Journal of Geography 53 (3), 192–193 (in Japanese). Omoto, K. (2003). Radiocarbon ages and δ 13 C values of beach rock samples collected from Yonaguni Island and Hateruma Island southwest of Japan. Proc. Inst. Natur. Sci. Nihon Univ. 38, 1–17 (in Japanese). Omoto, K., Chigono, S., Kanno, K. (2003). Radiocarbon ages and δ 13 C values of beach rock samples collected from west coast of Okinawa Island. Proc. Gener. Meet. Assoc. Jap. Geogr. 64, 125 (in Japanese). Omoto, K. (2004a). Radiocarbon ages and δ 13 C values of beach rock samples collected from Kume Island, Oh Island, and Hateno-hama (beach), west of Okinawa Island. Proc. Inst. Natur. Sci. Nihon Univ. 39, 15–31 (in Japanese). Omoto, K. (2004b). Radiocarbon ages and isotope fractionations of beach rock samples collected from the Nansei Islands, southwest of Japan. Proc. 18th Int. Radiocarbon Conf., Wellington, 2003. Radiocarbon 46 (2), 539–550. Omoto, K. (2004c). Radiocarbon ages and isotope fractionations of beach rock samples collected from Okinawa Islands, southwestern part of Japan. Proc. 10th Int. Coral Reef Symp., Okinawa, 2004. J.C.R.S., p. 33. Omoto, K. (2005). The data sets of radiocarbon ages and isotope fractionations of beach rock samples collected from the Nansei Island, southwestern part of Japan. Proc. Inst. Natur. Sci. Nihon Univ. 40, 1–27 (in Japanese). Pirazzoli, P.A. (1996). Sea-Level Changes the Last 20000 Years. Willey, Chichester, New York, Brisbane, Toronto, Singapore, 211 pp. Russel, R.J. (1959). Origin of beach rock. Zeitschrift fur Geomorphologie, N.F. 3, 227–236. Scoffin, T.P., Stoddart, D.R. (1983). Beach rock and intertidal cements. In: Goudie, A.S., Pye, K. (Eds.), Chemical Sediments and Geomorphology: Precipitates and Residua in the Near-Surface Environment. Academic Press, London, New York, Paris, San Diego, San Francisco, São Paulo, Sydney, Tokyo, Toronto, pp. 401–425. Stoddart, D.R., Cann, J.R. (1965). Nature and origin of beach rock. Journal of Sedimentary Petrology 35, 243–247. Sunamura, T. (1983). Change of coastal landform. In: Mizuyama (Ed.), Fundamental Studies on Antitative Predictions with Relation to Change of Landforms. Rep. Sci. Res. Grant in Fund (1981 and 1982), pp. 29–39 (in Japanese). Stuiver, M., Polach, H.A. (1977). Discussion reporting of 14 C data. Radiocarbon 19 (3), 355–363. Stuiver, M., Braziunas, T.F. (1993). Modeling atmospheric 14 C influences and 14 C ages of marine samples to 10,000 BC. Radiocarbon 35 (1), 137–189. Takenaga, K. (1965). Beach rock and lagoon on Yoron Island, Ryukyu Archipelago. Geogr. Review Jap. 38, 739–755 (in Japanese). Taylor, J.C.M., Illing, L.V. (1971). Development of recent cemented layers within intertidal sand flats, Qatar, Persian Gulf. In: Bricker, O.P. (Ed.), Carbonate Cements. In: Johns Hopkins Univ. Studies in Geology, vol. 19. Johns Hopkins Univ. Press, pp. 27–31. Yonetani, S. (1963). Preliminary notes on beach rock at the south-west Island of Japan. Kagoshima Univ. Bunka Hokoku Shigakuhen 9 (12), 1–28.
418
K. Omoto
Further reading Omoto, K. (1999). Radiocarbon ages of beach rock and fossil coral samples collected from Aguni Island, southwestern part of Japan. Chronological view on the late Holocene sea-level change of Aguni Island. Annals of Geography, The Chiri Shiso 40 (2), 15–28 (in Japanese). Omoto, K. (2004d). Isotope correction ages for Holocene fossil coral, shell and calcarenite samples collected from the Nansei Islands, southwest of Japan. Proc. Gen. Meet. Assoc. Jap. Geogr. 66, 92.
419
A model of recent sedimentation in the Cananeia–Iguape estuary, Brazil R.T. Saitoa , R.C.L. Figueirab,c , M.G. Tesslerc , I.I.L. Cunhaa,* a Radiochemistry Division, IPEN-CNEN, São Paulo, Brazil b Cruzeiro do Sul University (UNICSUL), São Paulo, Brazil c Instituto Oceanográfico, Universidade de São Paulo, São Paulo, Brazil
Abstract Coastal systems, in particular estuaries, are the first depositional environment to receive sediments transported by rivers to the coastal ocean. As a consequence of the geochemical processes involved, a part of the sediment may be deposited in the estuary and the remainder flows into the ocean, being transported and deposited under the influence of tides and maritime currents. This study was carried out in the Cananeia–Iguape estuary on the southern coast of São Paulo state, Brazil. The vertical distribution of 210 Pb and 137 Cs in sediments has been used as a tool for estimating the sedimentation rates in the coastal environment. Concentrations of 210 Pb and 137 Cs, as well as heavy metals (lead, zinc, copper) were determined in four sediment cores collected in the Cananeia–Iguape estuary. The estimated sedimentation rates were from 5.3 mm yr−1 to 12.7 mm yr−1 . The highest sedimentation rate obtained for Valo Grande corresponds with an accelerated expansion of the sand and clay banks of the Mar Pequeno channel, a growth of the mangrove areas, and a decrease in the depth of the main channel of navigation, a fact that has been affecting the navigation in the area very seriously. The data obtained for the metals showed a sedimentary dynamics in agreement with that obtained in the studies of sedimentation rates, thus contributing to the understanding of the hydrodynamic mechanisms of the system. Keywords: Radionuclides, 210 Pb dating, 137 Cs, Heavy metals, Coastal sediment, Estuary, Sedimentation model, Brazil
1. Introduction Intensive human activities in the estuarine regions in recent decades, have significantly increased sediment inflows at these sites and, consequently, of the accumulation of sediment in the estuarine channels and in the coastal ocean. The use of 210 Pb to date sediments up to 100 years old is a very important tool for establishing a geochronology of the coastal environment (Ravichandran et al., 1995). Other radionuclides, such as 137 Cs, are often used to determine sedimentation rates in addition to the data provided by 210 Pb measurements (Somayajulu et al., 1999). * Corresponding author. Address: Rua Manoel Alonso Esteves 66 – Jardim Bonfiglioli, São Paulo Capital, CEP 05589-020, Brazil; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08034-4
© 2006 Elsevier Ltd. All rights reserved.
420
R.T. Saito et al.
This paper presents and discuss radionuclide levels, as well as contents of heavy metals in sediment cores collected from the Cananeia–Iguape estuary. Sedimentation rates have been evaluated at these locations from the unsupported 210 Pb and 137 Cs profiles. The main objective of this paper is to elucidate the sedimentation process in the Cananeia–Iguape estuary and to develop a suitable sedimentation model. 1.1. Cananeia–Iguape estuary The estuary consists of several channels, the most important of them being the Mar Pequeno, Iguape, Cubatão and Cananeia; several islands, such as Comprida, Iguape, Cananeia and Cardoso; there are also rivers, such as the Ribeira of Iguape. Near the discharge of the Ribeira of Iguape River, there is a group of islands close to the continent, separated by a series of narrow channels, which are interconnected and flow into the Atlantic Ocean through three outlets (Ararapira, Cananeia and Icapara). Comprida Island, a barrier island approximately 70 km long, separates the Cananeia–Iguape (25◦ S, 48◦ W) system from the ocean. The mouth of the Ribeira of Iguape River is located in the northeast of this system. The largest drainage system of the south-eastern Brazilian seashore, draining all the crystalline coastal mountainous complex behind the coastal plain, is connected to the Cananeia–Iguape estuary only by the Valo Grande channel. This artificial channel connects the Ribeira of Iguape River directly to the Mar Pequeno channel. About 60% of the Ribeira of Iguape River discharge flows at present through the internal channels of the Cananeia–Iguape estuary, causing an increasing silting up of the channels by the deposition of the muddy sediments in suspension carried by the drainage of the Ribeira of Iguape. Thus, the continental material is transferred to the maritime system on the southern seashore of São Paulo State, not only at the mouth of the Ribeira of Iguape River but also at the river mouths of the Cananeia–Iguape estuary. Sediments continue to be deposited in the channels and internal seas, mainly in the Mar Pequeno channel, and along the islands in the channels. The silting up of the area is taking place so fast that whereas, some decades ago large ships were able to dock in Iguape city, crossing the Cananeia bar, this is now impossible, due to the increasing obstruction at two regions towards the Icapara bar and at Cananeia. In consequence, the present depth of the lagoons and channels is of less than 15 m. The Valo Grande channel was completed in 1852, then about 4 meters wide and 7 meters deep. Now, it is 250 m wide and 15 m deep. 1.2. Dynamics of the channels The Cananeia–Iguape system is formed predominantly of sandy sediments disposed along the surfaces of bottom of all of the channels. This general pattern is only modified in the concave banks of the channel’s bends and in the Mar Pequeno channel, where equal mixtures of thick and fine sediments are found beside sandy sediments due to an alternation of energy flows, resulting from the changes in the direction and intensity of currents, that make the deposition of fine sediments, loaded in suspension, mainly from the Ribeira of Iguape River and also from the mangroves areas, possible. Beginning in this area, towards Iguape city, the mixed sediments are gradually replaced by silt and clay sediments, until Valo Grande, where very fine sediments (loamy silts), originating from the Ribeira of Iguape River are deposited.
A model of recent sedimentation in the Cananeia–Iguape estuary, Brazil
421
2. Experimental 2.1. Sediment sampling The core samples were collected by the Oceanographic Institute of the University of São Paulo, Brazil, at four estuary stations, that show different sediment inflows to the coastal system (Fig. 1): T1 (Ponta do Arrozal, in the south of Cananeia Island, between Cananeia Island and Cardoso Island); T2 (Ponta do Frade, along the Comprida Island); T3 (Valo Grande, near the mouth of Ribeira of Iguape, at Mar Pequeno channel); and T4 (Carapara River, NW of Cananeia Island). The sediment was collected with a cylindrical PVC container of 50 cm height and 7 cm in diameter. The cores were sliced into 2 cm thick layers, dried and homogenized, and transferred to plastic containers appropriate for gamma-ray spectrometry measurements. The sand and mud contents (silt and clay) were analyzed, as well as organic matter contents and humidity were determined in each core. 2.2. Methodology for 210 Pb, 226 Ra and 137 Cs analysis A low background HPGe gamma-ray spectrometer (ORTEC), with a resolution of 1.9 keV at 1332.40 keV photopeak of Co-60, together with MAESTRO II software was used for analysis of gamma-ray emitters. 210 Pb was assayed by means of its 47 keV photopeak. The method consisted of detector calibration, determination of detector counting efficiency, cumulative
Fig. 1. Cananeia–Iguape estuary and the sampling sites (T1, T2, T3 and T4).
422
R.T. Saito et al.
countings of both background and samples in regular intervals, photopeak smoothing and linear regression (Figueira et al., 1997; Saito et al., 2001). 226 Ra was assayed by means of 609 keV photopeak (214 Bi), and 661 keV peak was used for analysis of 137 Cs. For 226 Ra measurements, the packed samples were stored for 20 days at least in order to reach radioactive equilibrium. Unsupported 210 Pb was calculated by subtracting the 226 Ra (supported 210 Pb) activity from the 210 Pb total activity profile (Ravichandran et al., 1995). The analytical quality control was achieved by regular participation in international intercomparison exercises organized by the International Atomic Energy Agency (IAEA) as well as by analyzing reference materials.
3. Results and discussion 3.1.
210 Pb, 226 Ra
and 137 Cs in sediment cores
Results of radionuclide analyses of sediment cores collected in the Cananeia–Iguape estuary are given in Fig. 2. 210 Pb levels in the estuary varied from 6.1 to 167 Bq kg−1 . The lowest values were observed in the T2 core (from 6.1 to 53 Bq kg−1 ), and the highest in the T1 (from 9.8 to 167 Bq kg−1 ) and T3 (from 31 to 120 Bq kg−1 ) cores. The 210 Pb levels were much higher in the upper layers of the core and decreased with the depth. These results are consistent with observations published in the literature. The 137 Cs levels obtained in the Cananeia–Iguape estuary ranged from 0.28 to 6.1 Bq kg−1 . The presence of 137 Cs in the Brazilian sediments is due to global fallout. The content of organic matter and the grain size composition are showed in Tables 1 and 2. The mud (silt–clay) content varied from 2.0 to 83.3%. The lowest values were found in the T2 core and the highest in T3. The sediments of the T1 core presented a prevalence of mud up to a depth of 6 cm, later it was noticed a gradual increase of fine sand. The T2 core presented a prevalence of fine sand along the whole sediment column. The T3 core was the only one that presented a prevalence of mud in the whole sediment column. The T4 core presented a prevalence of fine sand, with a gradual increase of this component along the sediment column. The grain size composition of the cores reflects the dynamics of the circulation in the channels. The T3 core receives the largest load of fine sediments from the discharge of the Ribeira of Iguape River. The other three points are located opposite the Cananeia–Iguape system and are not so affected by the discharges from this river, but they are receiving loads from small local rivers. The organic matter content presented a great variation along all of the cores analyzed, with levels varying from 0.4 to 10.6%. The lowest and the highest values were observed in the T4 and T1 cores, respectively. The higher levels of organic matter reflect the intense local biological activity, furthermore, the cores were collected close to the mangrove. The statistical treatment of data showed a correlation between the 210 Pb levels and the mud (silt–clay) content. In relation to the organic matter levels, no correlation was observed as the biological activity might have affected these values. Samples collected close to the mangrove might have increased organic matter content. In general, the grain size composition followed the expected pattern, the sediments with higher levels of mud (silt–clay) had higher radionuclide concentrations.
A model of recent sedimentation in the Cananeia–Iguape estuary, Brazil
423
Fig. 2. Massic activities of radionuclides in sediment cores.
3.2. Sedimentation rates Figure 3 presents the levels of unsupported 210 Pb and 137 Cs as a function of the core depth. The sedimentation rates derived by the 210 Pb method 5.3 (T1), 9.8 (T2), 12.7 (T3) and 6.2 (T4) mm yr−1 . The 137 Cs profiles showed distinct peaks corresponding to the maximum global fallout observed in 1963–1964. The sedimentation rates were therefore estimated using
424
R.T. Saito et al. Table 1 Contents of organic matter (OM) and grain size composition (T1 and T2 cores) Depth (cm)
T1 OM (%)
Sand (%)
Mud (%)
T2 OM (%)
Sand (%)
Mud (%)
0–2 2–4 4–6 6–8 8–10 10–12 12–14 14–16 16–18 18–20 20–22 22–24 24–26 26–28 28–30 30–32 32–34 34–36 36–38 38–40
6.8 6.4 10.6 6.4 9.2 3.1 5.2 5.4 5.2 2.7 2.7 3.1 1.5 2.2 6.1 9.3 7.9 4.7 2.4 1.4
20.2 17.8 34.3 48.6 51.1 50.0 61.2 66.7 69.0 69.3 73.7 67.6 51.3 43.7 49.5 42.8 43.0 65.2 59.6 83.6
59.1 61.7 50.2 37.9 36.3 37.1 27.8 23.8 22.0 22.0 18.7 23.9 38.8 43.6 38.4 43.0 44.6 26.8 30.0 11.3
1.0 2.9 3.2 3.5 1.4 1.2 5.0 4.2 2.9 4.3 2.8 4.0 3.4 2.4 3.7 5.1 2.7
91.1 90.7 87.0 72.2 72.4 85.3 89.9 87.9 87.5 92.8 90.7 91.6 91.3 92.3 93.4 93.0 93.3
4.7 3.8 5.3 5.3 3.0 4.3 4.4 5.2 6.6 2.8 3.6 3.9 3.5 2.9 2.9 2.1 2.2
the 137 Cs method as well (Huh and Su, 1999), and they were in good agreement (within the uncertainties, which varied up to 10%) with the 210 Pb estimations. The core sampled at Valo Grande (T3) presented the highest sedimentation rate (12.7 mm yr−1 ). This is the site directly influenced by the drainage of the Ribeira of Iguape River, which is the main source of sediments discharged into the coastal system (Tessler et al., 1987). The Carapara River core (T4) and that of the Ponta do Arrozal (T1) are representative of a deposition environment, with a prevalence of recent marine conditions and of continental sediments originating from the crystalline coastal mountain complex and from the quaternary sandy formations of the coastal plain. These continental sources have contributed with smaller amounts of sediments than those deposited in Valo Grande. The Ponta do Frade (T2) is located at an intermediate site of the system, and presented an intermediate value of sedimentation rate (9.8 mm yr−1 ) compared to the sampling stations T1 and T3. These results indicate that this area of Ponta do Frade receives muddy sediments resulting from the drainage of the Ribeira of Iguape River, transported by the ebb tide towards to Cananeia Island. The values obtained reflect both the dynamics of the circulation of the estuarine channels and the contribution of the main sources of the sediments found in the Cananeia–Iguape system. The sedimentation rate (12.7 mm yr−1 ) obtained for Valo Grande (T3) is higher than for any of the other sites. It confirms three local events: an accelerated expansion in recent decades of the sand and clay banks of the Mar Pequeno channel; a growth of the mangrove areas; and
A model of recent sedimentation in the Cananeia–Iguape estuary, Brazil
425
Table 2 Contents of organic matter (OM) and grain size composition (T3 and T4 cores) T3
T4
Depth (cm)
OM (%)
Sand (%)
Mud (%)
Depth (cm)
OM (%)
Sand (%)
Mud (%)
0–4 4–7 7–10 10–13 13–16 16–19 19–22 22–25 25–28 28–31 31–34 34–37 37–40 40–43 43–46 46–49 49–52 52–55 55–58 58–61 61–64 64–67 67–70
2.6 4.1 9.8 3.9 4.1 6.3 6.8 6.2 4.2 5.6 5.3 5.5 7.0 5.3 6.3 6.7 8.0 8.0 7.3 5.0 6.4 5.8 6.1
15.2 14.7 14.5 17.5 13.8 10.4 11.2 10.5 15.8 16.4 13.2 7.7 10.2 9.9 10.7 6.6 4.4 9.8 14.6 16.5 10.5 18.6 13.5
77.2 78.1 75.8 74.1 76.6 79.3 77.9 80.3 75.3 74.4 76.5 80.5 80.0 78.9 77.7 82.0 83.3 78.5 74.9 72.3 78.7 69.3 75.8
0–2 2–4 4–6 6–8 8–10 10–12 12–14 14–16 16–18 18–20 20–22 22–24 24–26 26–28 28–30 30–32 32–34 34–36 36–38 38–40 40–42 48–50 58–60
1.1 3.8 2.8 1.7 0.7 0.7 0.6 1.0 1.6 0.4 1.6 0.6 0.6 0.8 0.4 3.9 3.1 2.6 2.0 2.1 3.1 3.0 1.8
59.2 47.8 60.6 56.7 72.8 80.9 77.7 76.3 72.2 74.5 76.1 79.5 82.6 84.2 83.4 85.8 91.6 91.8 93.4 89.7 88.5 90.4 91.2
28.6 36.9 29.2 30.2 16.2 11.3 14.2 16.4 17.9 17.1 16.3 13.9 10.9 9.8 10.3 7.9 3.3 3.1 2.0 4.1 3.6 3.8
a decrease in the depth of the main channel of navigation, a fact that has been affecting the navigation in the area very seriously. 3.3. Geochemical characterization The chemical analyses of lead, copper and zinc were carried out by the Actilab Company (Canada), by using the SW 846 EPA 3050B EPA method. The solutions resulting from the acid lixiviation were analyzed by atomic absorption spectrometry. The concentrations obtained for copper, zinc and lead are showed in the Tables 3 and 4. Lead represents one of the most abundant elements and has been intensively mined in the region near the mouth of the Ribeira of Iguape River. The highest levels of these metals were found between the confluence of the Valo Grande and the Mar Pequeno channel, as far as the neighborhood of the Pedra do Tombo, probably, incorporated to the fine sediments loaded in suspension by the Ribeira of Iguape River. Pedra do Tombo (located in the Mar Pequeno channel, in the half of Comprida Island) corresponds to the point of encounter of tides of opposite directions, it avoids the propagation of those fine sediments transported by the Ribeira of Iguape River for all of the lagoon channels. Due to this, the area of Comprida Island shows low lead levels, around
426
R.T. Saito et al.
Fig. 3. Massic activities of 210 Pb (unsupported) and 137 Cs in sediment cores. Table 3 Concentrations of lead, copper and zinc (µg g−1 ) in cores T1 and T2 (total uncertainties below 10%) Depth (cm)
Pb
T1 Cu
Zn
T2 Pb
Cu
Zn
0–2 2–4 4–6 6–8 8–10 10–12 12–14 14–16 16–18 18–20 20–22 22–24 24–26 26–28 28–30 30–32 32–34
26 25 19 17 16 18 16 13 12 11 12 12 13 10 16 23 19
15 15 11 10 11 10 9 8 8 7 8 9 9 17 10 12 10
55 57 41 40 40 39 32 29 28 25 26 28 33 43 41 47 44
8 12 11 12 12 8 11 9 9 5 6 4 4 4 3 4 5
8 7 8 7 7 7 7 7 7 6 7 6 7 6 6 6 6
14 18 18 19 18 18 19 18 18 15 16 11 14 13 12 14 13
A model of recent sedimentation in the Cananeia–Iguape estuary, Brazil
427
Table 4 Concentrations of lead, copper and zinc (µg g−1 ) (T3 and T4 cores) Depth (cm)
T3 Pb
Cu
Zn
T4 Pb
Cu
Zn
0–2 2–4 4–6 6–8 8–10 10–12 12–14 14–16 16–18 18–20 20–22 22–24 24–26 26–28 28–30 30–32 32–34 34–36 36–38 38–40 40–42 42–44
101 113 114 84 106 132 120 105 101 110 127 131 141 151 144 153 150 163 159 150 166 131
37 39 39 32 36 41 35 35 32 34 37 36 39 41 39 41 40 43 41 38 44 36
120 126 127 113 127 133 123 117 108 115 123 121 129 126 124 121 121 124 120 118 133 121
14
9
31
12
7
29
8
6
22
35
7
24
13 9
7 8
28 25
8
7
18
7
9
22
5
8
16
2 4 4
8 8 7
12 13 15
3
5
10
4 µg g−1 . No high concentration of the heavy metals, such as those presented at Valo Grande station, for the other regions of the channels, was found. The lead concentrations found in sediments of the Valo Grande (T3) varied from 84 to 166 µg g−1 , above the regional background for the Valo Grande region, that is of 16 µg g−1 (CPRN, 1978), considerably above the concentration of 40 µg g−1 , threshold concentration given by Prater and Anderson (1977) for highly polluted sediments. It was verified that there is a decrease in the lead concentrations in the upper layers, which probably correlates with the reduction in the metal mining activities in the upper area of the Ribeira valley (Moraes, 1997). The concentrations of zinc and copper in the sediment collected in Valo Grande (T3) followed the same pattern as that presented by lead. The concentrations of zinc in Valo Grande varied from 108 to 133 µg g−1 , much higher than the regional background for this region (46 µg g−1 ). Copper concentrations varied from 32 to 44 µg g−1 , and were also above the background value for the Valo Grande region (18 µg g−1 ). 3.4. Sedimentation model The small Mandira basin (located southeast) drains into the most inland part of the coastal system, in the Cubatão Sea, a channel between Cananeia Island and the continent. The slope
428
R.T. Saito et al.
in the plain and the proximity of the crystalline mountainous complex result in a low transport capacity of these water-courses, which contribute predominantly with suspended sediments (mud and organic matter) to the southern most points of the system. On the other hand, the large drainage basin of the Ribeira of Iguape River, which flows into the channels of this system exclusively through the Valo Grande channel, in the proximity of the Iguape city, is a significant source of finer sediments (silts and clays), loaded in suspension, associated with large amounts of organic matter. The flood and ebb tide currents propagated in opposite directions, flow to the Mar Pequeno channel, near the Iguape city, and consequently extensive areas of deposition of mud (silts and clays), rich in organic matter, are formed. This area of the most intense sedimentation has also been studied by Souza (1995), who concluded that, besides prolonging this depositional area towards the south, there is a transport of sediments, close to the bottom of the channel, in the direction of the mouth of Cananeia (SE). Thus, the sediments in suspension discharged by the Valo Grande, that are not deposited in the area of Pedra do Tombo by the flood tide, are carried towards the southeastern mouth, through the channels that outline the Cananeia Island (the Cubatão and Cananeia Seas). The concave curves on the inner sides of the bends (SE region) favor areas of low circulation, resulting in the deposition of mud. Nowadays, the expansion of the mangrove is associated predominantly with these areas in which the sandy substratum is being enriched by the addition of continental mud, and also with the sand banks in the middle of the channels, which result from the erosion of the banks of the convex faces. These sand banks in the middle of the channels are an obstacle to the flow of the waters, leading to a fall in the energy of the transport of the currents and, consequently, to the deposition of the suspended mud. Tessler et al. (1987) has demonstrated that the coastal system does not present a capacity to export sandy sediments from the bottom of the channels into the ocean, except for the plumes of suspended sediments. As a result of the characteristics of the internal dynamics of the system, most of the sediments loaded by the Ribeira of Iguape River are discharged through the Icarapara mouth (NE). In relation to the suspended sediment loaded by Ribeira of Iguape River towards the Cananeia mouth, a greater part of it is deposited at the bends of Cananeia and Cubatão channels, before it reaches the outlet to the Ocean. The Cananeia–Iguape system presents a general tendency to an accentuated silting up process as a result of the pronounced contribution of sediments of continental origin. However, the northeastern part of the system is more strongly submitted to the influence of the continental contribution than does southeastern part of the study area.
4. Conclusions Several observations made in this study may be summarized as follows: • The sedimentation rates obtained by the 210 Pb method in the sediment columns along the Cananeia–Iguape estuary were of 5.3 mm yr−1 (T1), 9.8 mm yr−1 (T2), 12.7 mm yr−1 (T3) and 6.2 (T4) mm yr−1 . The highest sedimentation rate obtained for Valo Grande (T3) corresponds with an accelerated expansion of the sand and clay banks of the Mar Pequeno channel, a growth of the mangrove areas, and a decrease in the depth of the main channel of navigation, a fact that has been affecting the navigation in the area very seriously.
A model of recent sedimentation in the Cananeia–Iguape estuary, Brazil
429
• Data obtained for the metals (lead, copper and zinc) in the sediments showed a sedimentary dynamics in agreement with that obtained in the studies of sedimentation rates, thus contributing to the understanding of the hydrodynamic mechanisms of the system. The highest levels of metals measured in the sediments were a consequence of the mining activities carried out in the upper of the Ribeira valley, that results in an increment in the drainage of the Ribeira of Iguape River of enriched particles of metals originating from the lixiviation of wastes or the washing of ores. • There are three main sources of sediments in the channels of the Cananeia–Iguape system: the drainage systems that flow directly into the channels (Mandira and Ribeira of Iguape systems), the internal biological production of the system, and the sandy sediments resulting from the erosion of the Cananeia coastal stringers, i.e. of the banks of the concave margins of the channels. • The fine sediments, transported in suspension from the basin of Ribeira of Iguape River, are deposited predominantly in the channel which separates Comprida Island from the continent, and in the Mar Pequeno channel, close to the Valo Grande channel. During the ebb tide processes, the material not deposited in this channel is mainly transferred to the SE of the area, through the Cananeia Sea channel. Part of the material is deposited in this SE area, in the coincident areas with mangroves, now, undergoing expansion, as observed at the Ponta do Arrozal. • The other internal channels, that are not directly affected by the drainage of the Ribeira of Iguape River, nor by the ebb tide dynamics of the Mar Pequeno channel present lower sedimentation rates than do the more external channels of the system (Cananeia Sea and Mar Pequeno).
Acknowledgements The authors would like to thank FAPESP, CNPq, CNEN and the Oceanographic Institute of the University of São Paulo for the financial support.
References CPRN, Coordenadoria de Pesquisas de Recursos Naturais (1978). Relatório técnico: Projeto Geoquímico no Vale do Ribeira. São Paulo, 326 pp. Figueira, R.C.L., Silva, L.R.N., Cunha, I.I.L. (1997). Instrumental analysis of low levels of 137 Cs in marine samples by gamma-spectrometry. International Conference on the Radiological Accident of Goiania – 10 Years Later, Goiania, Brazil. Huh, C., Su, C. (1999). Sedimentation dynamics at the East China Sea elucidated from 210 Pb, 137 Cs and 249+241 Pu. Marine Geology 160, 183–196. Prater, B.L., Anderson, A. (1977). 96-hour bioassays of Otter Creek. Journal of the Water Pollution Control Federation 49, 2099–2106. Moraes, R.P., (1997). Transporte de Chumbo e Metais associados no rio Ribeira de Iguape, São Paulo, Brasil. Dissertação de Mestrado. Instituto de Geociências, USP, São Paulo. Ravichandran, M., Baskaran, M.P., Santschi, P.H., Bianchi, T.S. (1995). Geochronology of sediments in the Sabine– Neches estuary, Texas, USA. Chem. Geol. 125, 291–306.
430
R.T. Saito et al.
Saito, R.S., Figueira, R.C.L., Tessler, M.G., Cunha, I.I.L. (2001). Geochronology of sediments in the Cananeia– Iguape estuary in the southern continental shelf of São Paulo State, Brazil. Journal of Radioanalytical and Nuclear Chemistry 250, 109–115. Somayajulu, B.L.K., Bhushan, R., Sarkar, A., Burr, G.S., Jull, A.J.T. (1999). Sediment deposition rates on the continental margins of the eastern Arabian Sea using 210 Pb, 137 Cs and 14 C. The Science of the Total Environment 237/238, 423–429. Souza, L.A.P. (1995). A planície costeira Cananéia–Iguape, litoral sul do Estado de São Paulo: um exemplo de utilização de métodos geofísicos no estudo de áreas costeiras. Dissertação (Mestrado). Instituto Oceanográfico, Universidade de São Paulo. Tessler, M.G., Suguio, K., Robilotta, P.R. (1987). Teores de alguns elementos traços metálicos em sedimentos pelíticos da superfície de fundo no Sistema Cananéia–Iguape (SP). In: Simposio sobre ecosystemas da costasud e sudeste Brasileira, vol. 2, pp. 255–263.
9. Modelling of environmental processes
This page intentionally left blank
433
Simulation of the advection–diffusion–scavenging processes for 137Cs and 239,240Pu in the Japan Sea Masanao Nakano* Japan Nuclear Cycle Development Institute, Tokai-mura, Japan Abstract An advection–diffusion–scavenging model incorporated into the Oceanic General Circulation Model (OGCM) has been developed and validated by calculating 137 Cs and 239,240 Pu water profiles in the Japan Sea and comparing them with experimental data. A reasonable agreement between the calculated and observed 137 Cs and 239,240 Pu concentrations in the water column have been obtained. The optimised horizontal and vertical diffusion coefficients of 3 × 107 cm2 /s and 0.3 cm2 /s, respectively, have been applied in the present calculations. The simulations have contributed to better understanding of the origin and behaviour of 137 Cs and 239,240 Pu in the Japan Sea, which has been different when compared with the Northwest Pacific Ocean. Keywords: OGCM, Advection, Diffusion, Scavenging, Radionuclides, 137 Cs, 239,240 Pu, Water column, Japan Sea
1. Introduction In order to assess long-term (over 10 years) effects of discharges of radioactive contaminants from nuclear facilities to the world ocean, a Long-term Assessment ModEl for Radioactivity in the oceans (LAMER) has been developed. The LAMER model has been validated by simulating the advection–diffusion–scavenging processes of anthropogenic radionuclides (137 Cs, 239,240 Pu) in the world ocean from the past atmospheric nuclear weapons tests. Although the LAMER modelled only essential and indispensable diffusion processes of 137 Cs and 239,240 Pu, it successfully reproduced the distribution of these radionuclides, after several decades of their dispersion in the water column of the world ocean (Nakano and Povinec, 2003a, 2003b). It has been challenging to apply the LAMER technique of assessing the distribution of anthropogenic radionuclides in the water column to the Japan Sea because of its economical and environmental importance for the Southeast Asia, and the fact that over 30 nuclear facilities have been installed along the Japan Sea coast. The advection–diffusion–scavenging processes * Address: Japan Nuclear Cycle Development Institute, 4-33 Muramatsu, Tokai-mura, Naka-gun, Ibaraki-ken, 319-1194, Japan; phone: +81-29 282 9377; fax: +81 29 282 3838; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08035-6
© 2006 Elsevier Ltd. All rights reserved.
434
M. Nakano
Fig. 1. The Japan Sea, the East China Sea and the Pacific Ocean.
of 137 Cs and 239,240 Pu in the Japan Sea were simulated and validated using the observed data in a similar way as it was done for the Pacific Ocean. The Japan Sea is situated in the northwest of the Pacific Ocean (Fig. 1). The area of the Japan Sea is 1.3 × 106 km2 , and the average and maximum depths are 1,350 m and 3,700 m, respectively. All the straits of the Japan Sea are narrow and shallow, the maximum depth of the Tsushima and Tsugaru Straits is 140 m, and that of Soya and Mamiya Straits is only 60 m and 10 m, respectively. The inflow to the Japan Sea is limited only by the Tsushima Warm Current (Wadachi, 1987). The Yamato Rise and the North Yamato Rise with depths below 500 m divide the Japan Sea into the Japan Basin in the north (3,000–3,500 m depth), the Tsushima Basin in the southwest (1,500–2,500 m depth), and the Yamato Basin in the south-east (2,500–3,000 m depth) (Wadachi, 1987).
2. Model description 2.1. Oceanic General Circulation Model (OGCM) A modified version of the robust diagnostic OGCM described by Fujio et al. (1992) has been applied to the Japan Sea in the present study. The Japan Sea is also called a “mini ocean” as its
Simulation of the advection–diffusion–scavenging processes for 137 Cs and 239,240 Pu in the Japan Sea
435
basic properties are similar to the ocean. The model covers the Japan Sea with real topography and divides it horizontally into 0.5◦ ×0.5◦ grids and vertically into 17 water levels. The model consists of the equations of motion, continuity, state, advection and diffusion. Based on the annual average hydrographical data of Levitus (1982) and the wind stress data of Hellerman and Rosenstein (1983), the annually averaged velocity fields have been determined diagnostically. The details of the equations have already been described by Nakano and Povinec (2003a), and therefore they will not be repeated here. The inflow (+) and outflow (−) conditions of the three straits have been fixed as +2.2 × 106 m3 /s for Tsushima Strait (Isobe, 1994), −1.4 × 106 m3 /s for Tsugaru Strait (Shikama, 1994), and −0.8 × 106 m3 /s for Soya Strait (Kanari et al., 1984), the same during a year. Only the annual mean current has been used for tracking because the period of tracking is at least 10 years, and the effects of seasonal currents could therefore be neglected. 2.2. Tracking and scavenging models The procedures and parameters used for advection, diffusion and scavenging processes are similar to those described by Nakano and Povinec (2003a, 2003b). The tracking, the random walk, and the one-dimensional, two-phase scavenging model were used for the advection, the diffusion and the scavenging process, respectively. The mixed layer has also been considered in the model. The details and equations have already been described by Nakano and Povinec (2003a, 2003b). Because of the scale difference of the grid system, the time steps (dt) of the tracking and the scavenging were fixed as 2 days and 0.1 day, respectively. 2.3. Concept of the model validation Since 1945, around 948 PBq of 137 Cs and 10.87 PBq of 239,240 Pu have been input on the Earth as global fallout from 543 atmospheric nuclear weapons tests (UNSCEAR, 2000). The fallout from most of nuclear tests has been distributed globally, however, in the Pacific Ocean local fallout from the tests carried out at Bikini and Enewetak Atolls has to be considered as well. The total fallout to the Japan Sea can be divided into the following three categories: • The ocean component: The fallout onto the Pacific, Atlantic and Indian Oceans, and its dispersion into the Japan Sea via the Tsushima Strait. • The East China Sea component: The fallout onto the Yangtze catchment area and its transport to the East China Sea, and then via the Tsushima Strait into the Japan Sea. • The Japan Sea component: Direct fallout onto the Japan Sea. 2.3.1. Ocean component The ocean component of the fallout has been estimated using the global oceanic model (Nakano and Povinec, 2003a, 2003b). Using the UNSCEAR (2000) data, the contributions from global and local fallout were estimated and after the introduction of advection–diffusion– scavenging processes concentrations of 137 Cs and 239,240 Pu at the entrance of Tsushima Strait were calculated (Fig. 2).
436
M. Nakano
Fig. 2. Inflow of 137 Cs and 239,240 Pu into the Japan Sea via the Tsushima Strait.
2.3.2. The East China Sea component As mentioned by Nagaya and Nakamura (1992) and Lee et al. (2003), 239,240 Pu supply to the East China Sea via Yangtze River should be taken into account. It is expected that due to atmospheric nuclear weapons tests conducted at Lop Nor from 1964 to 1980, and subsequent river runoff large amounts of anthropogenic radionuclides were transported to the East China Sea. Concentrations of 239,240 Pu in surface seawater collected in 1987 offshore the estuary of the Yangtze River (123◦ 30 E, 30◦ 00 N) were 61.3–84.3 mBq/m3 (Nagaya and Nakamura, 1992). These values were twenty times higher than those in the Pacific Ocean. On the other hand, the concentration of 137 Cs in surface seawater collected at the same stations was 1.95–2.86 Bq/m3 (Nagaya and Nakamura, 1992), similar to those measured in the Pacific. The large percentage of the East China Sea (the average depth is 188 m) has continental shelf. Some radionuclides (e.g. Pu isotopes) in the East China Sea tend to be easily scavenged and deposited in the sediment. They can also be resuspended by the strong Tsushima Warm Current and transported. The continental shelf behaves therefore as a reservoir of these radionuclides. Using the observed values of 239,240 Pu in surface seawater collected around the Tsushima Strait between 1977 and 1995, its inflow from the East China Sea can be estimated as 239,240 Pu inflow in mBq/m3 = 2.28 × 1046 × exp(−0.0528 × dominical year). (1) The calculated inflow of presented in Fig. 2.
137 Cs
and
239,240 Pu
into the Japan Sea via the Tsushima Strait is
2.3.3. The Japan Sea component It has been assumed that the decay-corrected ratio of 137 Cs/90 Sr and 239,240 Pu/90 Sr in global fallout has been constant since 1945, so it may be possible to estimate the annual and latitudinal fallout for 137 Cs and 239,240 Pu on the basis of 90 Sr data reported by UNSCEAR (2000).
Simulation of the advection–diffusion–scavenging processes for 137 Cs and 239,240 Pu in the Japan Sea
437
Fig. 3. Annual deposition of 137 Cs and 239,240 Pu from global fallout onto the Japan Sea (calculated from UNSCEAR, 2000).
The estimated annual oceanic deposition of 137 Cs and 239,240 Pu from global fallout onto the Japan Sea is shown in Fig. 3. This estimation does not include the possible contributions from atmospheric nuclear weapons tests conducted in China. Using the estimated global fallout deposition and inflow data, the simulations for the advection–diffusion–scavenging processes of 137 Cs and 239,240 Pu in the water column were carried out, and the obtained results were compared validated with the experimental values published in the literature.
3. Results and discussion 3.1. Three-dimensional velocity fields The annually averaged velocity field in the surface layer (10 m) is shown in Fig. 4(a). In general, it does not have mesoscale eddies and is smooth. It differs from the calculated field presented by Kim and Yoon (1999) because their model does not include seasonal changes and a fine grid system. The velocity pattern diagram placed in NAO (1994) agrees well with Fig. 4(a). The purpose of this study has not been to reproduce the temperature, salinity and velocity fields, but to simulate the dispersion of radionuclides in the ocean. In this concept, only the most important processes were modelled. Unless the mesoscale eddies are reproduced, their effects could be included within the diffusion coefficients. A simulation of dispersion of radionuclides in the ocean on a time scale of a few decades dose not need to consider seasonal changes in the velocity field. However, in this study, the diffusion coefficients also include the differences in seasonal variations. Additionally, when comparing the seasonal field with that of Kim and Yoon (1999), Fig. 4(a) reproduces well the main Tsushima current, though the detail flows are different from their seasonal velocity fields.
438
M. Nakano
Fig. 4. Annually averaged velocity fields at (a) 10 m, (b) 600 m, and (c) 2,000 m.
The velocity fields in the middle layer (600 m) and the deep layer (2,000 m) are shown in Figs. 4(b) and 4(c), respectively. As they were not measured, the calculated results cannot be compared with experimental observations. 3.2. Determination of diffusion coefficients The calculated 137 Cs and 239,240 Pu water profiles were compared with the observed data in the water column, extracted from the Global Marine Radioactivity Database (GLOMARD) (IAEA, 2000) at 91 stations for 137 Cs and at 75 stations for 239,240 Pu, which were collected between the 1960s and 2000. The diffusion coefficients ranging from 1×105 to 2×108 cm2 /s for the horizontal diffusion and from 0.1 to 1.0 cm2 /s for the vertical diffusion, were used in sensitivity study, and the obtained results were compared with observed vertical profiles. The results were treated as in agreement if the ratios of the calculated and the observed values were within factor 2. Results giving unsuitable descriptions of 137 Cs and 239,240 Pu water profiles were counted for each combination of coefficients.
Simulation of the advection–diffusion–scavenging processes for 137 Cs and 239,240 Pu in the Japan Sea
439
The best agreement between the experimental and calculated water profiles was obtained with the horizontal and vertical diffusion coefficients of 3 × 107 cm2 /s and 0.3 cm2 /s, respectively. 3.3. Comparison of calculated and measured 137 Cs and 239,240 Pu data in seawater The water profile stations in the Japan Sea (listed in Table 1 and shown in Fig. 5) were selected on the basis of possible spatial and chronological developments in the water column. Table 1 Sampling stations in the Japan Sea Name
Sampling year
Longitude
Latitude
Reference
Cs1 Cs2 Cs3 Cs4 Cs5 Cs6
1964 1977 1984 1993 1996 1997
137.02E 135.19E 135.48E 136.67E 131.50E 136.34E
41.18N 37.73N 38.30N 40.59N 36.59N 41.00N
Aoyama and Hirose, 2004 Nagaya and Nakamura, 1981 Nagaya and Nakamura, 1987 MSA, 1995 MSA, 1998 MSA, 1999
Pu1 Pu2 Pu3 Pu4 Pu5 Pu6
1984 1993 1993 1994 1997 1999
135.48E 132.80E 136.67E 134.57E 137.43E 135.47E
38.30N 38.68N 40.59N 40.00N 41.45N 37.92N
Nagaya and Nakamura, 1987 Yamada et al., 1996 MSA, 1995 MSA, 1996 MSA, 1999 Ito et al., 2003
Fig. 5. Sampling stations in the Japan Sea.
440
M. Nakano
Fig. 6. Comparison of calculated (lines) and observed (points) 137 Cs water profiles.
As an example, the calculated and measured 137 Cs and 239,240 Pu vertical profiles for 6 stations are shown in Figs. 6 and 7, respectively. Generally, the model is reproducing well the observations, although in some cases it predicts higher 137 Cs and 239,240 Pu concentrations. As shown in Fig. 8(a), each measured 137 Cs concentration (374 results) over 1 Bq/m3 was compared with the calculated one, and about 80% and 97% of calculated results were within factor 2 and 5 with measured ones, respectively. Similarly for 239,240 Pu, each measured concentration (352 results) over 10 mBq/m3 was compared with the calculated one (Fig. 8(b)), and about 74% and 95% of calculated results were within factor 2 and 5 with measured ones, respectively. The ability of the model to predict temporal changes in radionuclide concentrations was tested using the data set gathered by the Japan Chemical Analysis Centre (JCAC, 2004). The data collected from the Shimane area (shown in Fig. 5) contains results on concentrations of 137 Cs in seawater samples collected mainly around nuclear power stations in Japan. Both the concentration levels and the time trend agree well with the calculations shown in Fig. 9.
Simulation of the advection–diffusion–scavenging processes for 137 Cs and 239,240 Pu in the Japan Sea
441
Fig. 7. Comparison of calculated (lines) and observed (points) 239,240 Pu water profiles.
3.4. The origin of 137 Cs and 239,240 Pu in the Japan Sea The contributions of 137 Cs and 239,240 Pu from the ocean, the Japan Sea and the East China Sea components were determined for the water depth of 10 and 2,000 m at 137◦ E, 41◦ N. For 137 Cs (Fig. 10) the Japan Sea component was dominant during the 1940s and mid-1950s, while the ocean component was dominant at the remaining period. After 1970, over 90% of 137 Cs remaining in the sea was from the ocean component. This result implies that 137 Cs deposited onto the Japan Sea was partially transported as surface water out of the sea via Tsugaru and/or Soya Straits, but also down to the 2,000 m water depth, where the contribution from the Japan Sea component was even a little larger than at the surface. For 239,240 Pu (Fig. 11) the Japan Sea component was dominant for surface and deep (2,000 m) waters before 1960, and the ocean component was dominant afterwards. The East China Sea component was not negligible after 1970, as it contributed about 30% of 239,240 Pu in surface seawater.
442
M. Nakano
Fig. 8. Comparison of calculated and measured concentrations of (a) 137 Cs and (b) 239,240 Pu in seawater samples collected in the Japan Sea from the 1960s to 2000.
Fig. 9. Comparison of calculated and measured concentrations of 137 Cs in surface seawater collected from the Shimane area.
3.5. Comparison of water profiles in the Japan Sea with that in the Pacific Ocean Figure 12 compares the calculated 137 Cs water profiles at 137◦ E, 41◦ N (the Japan Sea) with profiles calculated for 160◦ E, 41◦ N (the Pacific Ocean). It has been expected that the deposition of radionuclides from global fallout has been the same for the latitude belts. However, the calculated concentration of 137 Cs in surface seawater in 1960 in the Japan Sea was 32 Bq/m3 , which was four times larger than the concentration (8 Bq/m3 ) in surface seawater at the same latitude belt in the Pacific Ocean. The high 137 Cs concentrations in the Japan Sea could orig-
Simulation of the advection–diffusion–scavenging processes for 137 Cs and 239,240 Pu in the Japan Sea
443
Fig. 10. Contributions for 137 Cs from the ocean component and the Japan Sea component at the depths of (a) 10 m and (b) 2,000 m.
inate from close-in fallout at the Bikini and Enewetak nuclear weapons test sites, which was transported into the Japan Sea via surface Kuroshio and Tsushima Currents. When comparing 137 Cs water profiles in the Pacific Ocean with that in the Japan Sea (Fig. 12) we observe deeper penetration depths of 137 Cs in the Japan Sea. For example in 1975, concentrations over 1 Bq/m3 of 137 Cs were estimated at the depth of 2,000 m, while in the Pacific most of 137 Cs was situated above 600 m, even in the year 2000. This is caused by different vertical velocities, 10−4 –10−3 cm/s and 10−5 –10−4 cm/s for the Japan Sea and the Pacific Ocean, respectively. Therefore 137 Cs from the surface of the Japan Sea is faster transported downward by the vertical advection. On the contrary to 137 Cs, the calculated maximum concentration of 239,240 Pu for 1960 (Fig. 13) in the water profile in Japan Sea (137◦ E, 41◦ N) was 73 mBq/m3 , which was less than half of the maximum concentration (190 mBq/m3 ) in the water profile in the Pacific
444
M. Nakano
Fig. 11. Contributions for 239,240 Pu from the East China Sea component, the ocean component and the Japan Sea component at the depths of (a) 10 m and (b) 2,000 m.
Ocean (160◦ E, 41◦ N). The 239,240 Pu that originated from global fallout and Bikini and Enewetak test site was transported downward by the scavenging processes, and horizontally by the Kuroshio Current. The water with maximum 239,240 Pu concentration could not cross the shallow Tsushima Strait, which depth is only 140 m. Therefore the calculated maximum 239,240 Pu concentration in the Japan Sea is only half of that in the Pacific Ocean. But with passing time, this difference has been decreasing. No difference was recognised in 1990, and in 2000 the difference has reversed. The maximum 239,240 Pu concentration for 2000 in the Japan Sea and the Pacific Ocean was calculated as 38 mBq/m3 and 31 mBq/m3 , respectively. This result implies that the 239,240 Pu flown from Tsushima Strait advected and scavenged downward in the Japan Sea, and it could not cross over the Tsugaru Strait. As a consequence, the maximum 239,240 Pu concentration and its inventory in the Japan Sea became greater than that in
Simulation of the advection–diffusion–scavenging processes for 137 Cs and 239,240 Pu in the Japan Sea
(a)
445
(b)
Fig. 12. Calculated water profiles of 137 Cs in seawater at 137◦ E, 41◦ N ((a): the Japan Sea) and at 160◦ E, 41◦ N ((b): the Pacific Ocean).
(a)
(b)
Fig. 13. Calculated water profiles of 239,240 Pu in seawater at 137◦ E, 41◦ N ((a): the Japan Sea) and at 160◦ E, 41◦ N ((b): the Pacific Ocean).
446
M. Nakano Table 2 Calculated depth (m) with maximum 239,240 Pu concentration in the water column Year
The Japan Sea (137◦ E, 41◦ N)
The Pacific Ocean∗ (160◦ E, 41◦ N)
1960 1970 1980 1990 2000
400 500 500–1,500 500–3,000 500–3,000
100 200 500 500–1,000 500–2,000
∗ The values for the Pacific Ocean were estimated using the worldwide model (Nakano and Povinec, 2003b).
the Pacific Ocean. The comparison of water depths with maximum 239,240 Pu concentrations in the Japan Sea and the Pacific Ocean is presented in Table 2. 3.6.
137 Cs
and 239,240 Pu inventories in the Japan Sea
The balance of 137 Cs in the Japan Sea has been estimated by simulation of advection– diffusion–scavenging processes from 1945 to 2005. As shown in Fig. 14(a), the largest contribution to the total 137 Cs inventory in the Japan Sea was from the Pacific Ocean via Tsushima Strait, 44.7 PBq (92%). The contribution from global fallout deposited onto the Japan Sea is only 3.9 PBq (8%). The amount of outflow via Tsugaru and Soya Strait was 38.3 PBq (79%). 7.7 PBq (16%) of 137 Cs has decayed, and only 2.6 PBq (5%) has remained in Japan Sea waters in 2005. For 239,240 Pu (Fig. 14(b)), 153 TBq (77%) flew from the Pacific Ocean via Tsushima Strait, and 45 TBq (23%) was deposited onto the Japan Sea as global fallout. Only 66 TBq (33%) flew out from the Japan Sea via Tsugaru and Soya Straits, 88 TBq (44%) settled to the seabed, and 39 TBq (22%) has remained in Japan Sea waters in 2005. From the total input only 5% of 137 Cs, but 64% of 239,240 Pu have remained in the Japan Sea in 2005. Predictions for the future show that 239,240 Pu will mainly accumulate in sediment in the Japan Sea, while 137 Cs will flow out from the Japan Sea to the Northwest Pacific Ocean.
4. Conclusions An advection–diffusion–scavenging model incorporated into the OGCM has been developed for the Japan Sea and validated by calculating 137 Cs and 239,240 Pu water profiles in the Japan Sea and comparing them with experimental data. The calculated water profile concentrations of 137 Cs and 239,240 Pu have been in a reasonable agreement with experimental data. This is to our knowledge for the first time that OGCM has been developed for the prediction of 137 Cs and 239,240 Pu concentrations in the water column in the Japan Sea. The simulations performed in this study have contributed to better understanding of origin and behaviour of 137 Cs and 239,240 Pu in the Japan Sea, which have been much different
Simulation of the advection–diffusion–scavenging processes for 137 Cs and 239,240 Pu in the Japan Sea
(a)
447
(b)
Fig. 14. Estimated balance of (a) 137 Cs and (b) 239,240 Pu in the Japan Sea from 1945 to 2005.
when comparing with the Pacific Ocean. The simulations also helped to integrate fragmentary observational data into a complex assessment of present and future 137 Cs and 239,240 Pu inventories in the Japan Sea. Further work is in progress to combine the Japan Sea model with the LAMER model, and to assess radiation doses to humans and biota from nuclear facilities constructed at the Japan Sea coast. Acknowledgements The author would like to express his appreciation for support he received from Dr. D. Tsumune and Prof. P.P. Povinec, and for their careful reading of the manuscript and useful comments. He also thanks all scientists who made available their 137 Cs and 239,240 Pu data for the Japan Sea. References Aoyama, M., Hirose, K. (2004). Artificial radionuclides database in the Pacific Ocean: HAM database. The Scientific World Journal 4, 200–215. Fujio, S., Kadowaki, T., Imasato, N. (1992). World Ocean circulation diagnostically derived from hydrographic and wind stress fields. 1. The velocity field. Journal of Geophysical Research 96, 11163–11176. Hellerman, S., Rosenstein, M. (1983). Normal monthly wind stress over the World Ocean with error estimates. Journal of Physical Oceanography 13, 1093–1104. IAEA (2000). Global marine radioactivity database (GLOMARD). IAEA-TECDOC-1146, IAEA, Vienna. Isobe, A. (1994). Kaiyo Monthly 26 (12), 802–809 (in Japanese). Ito, T., Aramaki, T., Kitamura, T., Otosaka, S., Suzuki, T., Togawa, O., Kobayashi, T., Senjyu, T., Chaykovskaya, E.L., Karasev, E.V., Lishavskaya, T.S., Novichkov, V.P., Tkalin, A.V., Shcherbinin, A.F., Volkov, Y.N. (2003). Anthropogenic radionuclides in the Japan Sea: Their distributions and transport processes. Journal of Environmental Radioactivity 68, 249–267. Japan Chemical Analysis Center (JCAC) (2004). Environmental radioactivity database. Available at http://search.kankyo-hoshano.go.jp/servlet/search.top (in Japanese). Kanari, S., Koga, M., Aota, M. (1984). Velocity profiles and vorticity structure of Soya Warm Currents obtained with a free-fall electro-magnetic velocity profiler. Geophysical Bulletin of Hokkaido University, Japan 44, 67–76 (in Japanese).
448
M. Nakano
Kim, C.H., Yoon, J.H. (1999). A numerical modeling of the upper and the intermediate layer circulation in the East Sea. Journal of Oceanography 55, 327–345. Lee, S.H., Gastaud, J., Povinec, P.P., Hong, G.H., Kim, S.H., Chung, C.S., Lee, K.W., Petterson, H.B.L. (2003). Distribution of plutonium and americium in the marginal seas of the northwest Pacific Ocean. Deep-Sea Research II 50, 2727–2750. Levitus, S. (1982). Climatological Atlas of the World Ocean. NOAA Professional Paper, vol. 13. U.S. Govt. Printing Office, Washington, DC, 173 pp. MSA (Hydrographic Department, Maritime Safety Agency, Ministry of Transport) (1995). Annual reports of radioactivity survey for 1993. MSA (Hydrographic Department, Maritime Safety Agency, Ministry of Transport) (1996). Annual reports of radioactivity survey for 1994. MSA (Hydrographic Department, Maritime Safety Agency, Ministry of Transport) (1998). Annual reports of radioactivity survey for 1996. MSA (Hydrographic Department, Maritime Safety Agency, Ministry of Transport) (1999). Annual reports of radioactivity survey for 1997. Nagaya, Y., Nakamura, K. (1981). Artificial radionuclides in the western Northwest Pacific (I): 90 Sr and 137 Cs in the deep waters. Journal of the Oceanographic Society of Japan 37, 135–144. Nagaya, Y., Nakamura, K. (1987). Artificial radionuclides in the western Northwest Pacific (II): 137 Cs and 239,240 Pu inventories in water and sediment columns observed from 1980 to 1986. Journal of the Oceanographic Society of Japan 43, 345–355. Nagaya, Y., Nakamura, K. (1992). 239,240 Pu and 137 Cs in the East China and the Yellow Seas. Journal of Oceanography 48, 23–35. Nakano, M., Povinec, P.P. (2003a). Oceanic general circulation model for the assessment of the distribution of 137 Cs in the world ocean. Deep-Sea Research II 50, 2803–2816. Nakano, M., Povinec, P.P. (2003b). Modelling the distribution of plutonium in the Pacific Ocean. Journal of Environmental Radioactivity 69, 85–106. National Astronomical Observatory (NAO) (1994). Chronological Scientific Tables (Rika nenpyo), Tokyo. (In Japanese.) Shikama, N. (1994). Kaiyo Monthly 26 (12), 815–818 (in Japanese). United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) (2000). Exposures to the public from man-made sources of radiation. In: Sources and Effects of Ionizing Radiation. United Nations, New York. Wadachi, K. (1987). Encyclopaedia of Oceanography. Tokyodo Syuppan, Tokyo. (In Japanese.) Yamada, M., Aono, T., Hirano, S. (1996). 239+240 Pu and 137 Cs distributions in seawater from the Yamato Basin and the Tsushima Basin in the Japan Sea. Journal of Radioanalytical and Nuclear Chemistry, Articles 210, 129–136.
449
A biokinetic model for the uptake and depuration of radioiodine by the edible periwinkle Littorina littorea J. Vives i Batllea,* , R.C. Wilsona , P. McDonalda , T.G. Parkerb a Westlakes Research Institute, Westlakes Science and Technology Park, Moor Row, Cumbria, UK b British Nuclear Group plc, Sellafield, Cumbria, CA20 1PG, UK
Abstract A dynamic model for the prediction of radioiodine concentrations in the edible periwinkle Littorina littorea following uptake from the marine environment has been developed and successfully tested. This model incorporates calibration information generated from both laboratory experiments and knowledge of the local ecosystem. Central to the design is a biokinetic sub-model for winkles containing compartments for the shell, gastro-intestinal tract, soft tissue and remainder. The model is found to reproduce uptake, depuration and basic internal distribution as experimentally observed for winkles exposed to radioiodine in both seawater or labelled red seaweed (Chondrus crispus). For winkles in the field the model approximates satisfactorily the available monitoring data for 129 I in winkles from the west Cumbrian coast (UK), spanning several years. Use of the model in assessment mode predicts that, under the current discharge regime for 129 I, consumption of local winkles and chronic exposure to the winkles themselves pose no radiological significance. Keywords: Periwinkle, Littorina littorea, Iodine, Biokinetic model, Uptake, Depuration
1. Introduction The objective of the present study was to develop a biokinetic model to predict the uptake, depuration and distribution of radioiodine within edible periwinkles (Littorina littorea) from the UK Cumbrian coast. Although not physiologically based, the model was intended to provide an integrated representation of the available experimental data (e.g. biological half-lives) as a tool for assessments, more refined than the usual concentration factor approach, which is generally valid only in environments under equilibrium or near-equilibrium conditions. Winkles are an organism of interest because their consumption contributes significantly to the dose received by critical group consumers of seafood in the vicinity of the British Nuclear Group nuclear reprocessing plant at Sellafield in Cumbria, UK (BNFL, 1974–2002). Many radionuclides are discharged under authorisation from the Sellafield site, several of which * Corresponding author. Address: Westlakes Research Institute, The Princess Royal Building, Moor Row, Cumbria, CA24 3LN, UK; fax: (+44) 0 1946 514091; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08036-8
© 2006 Elsevier Ltd. All rights reserved.
450
J. Vives i Batlle et al.
(99 Tc, 137 Cs, Pu-α) have been investigated with relation to the winkle (McDonald et al., 1992, 1993; Swift, 1989; Swift et al., 1995) in order to provide data that can be used in modelling studies. However, one of the radionuclides discharged, 129 I (t1/2 = 1.57 × 107 years), has not been investigated to the same extent and, particularly, no published data are available on the response of radioiodine concentrations in shellfish to aquatic releases from Sellafield. For modelling purposes, the use of concentration factors (CF) and biological half-lives (T b1/2 ) provides a simple method for estimating the uptake, concentration and depuration of a radionuclide in biological materials from ambient seawater and food. The concentration factor (CF) is defined as the ratio of activity concentration in biota (Bq kg−1 , wet weight) to ambient seawater activity concentration (Bq m−3 ). The biological half-life, Tb1/2 , is defined as the time required for a biological system to eliminate, by natural processes, half the amount of a substance (such as a radioactive material) that has been absorbed into that system. The biological half-life is used as a index for understanding the speed at which substances metabolise. Transfer rate constants are the real transfer parameters for the model, and these can be calculated from biological half-lives using the simple relationship Tb1/2 = ln 2/transfer rate. A multi-stage depuration process (e.g. initial fast clearance combined with organ retention and further depuration) leads to a multiphasic depuration curve, supported by different biological half-lives and complex partitioning between the various body parts. This representation is further complicated by the fact that uptake and depuration operate at the same time. It is this more dynamic situation that the present model was designed to address. It was determined early in this work that information available on the uptake of iodine by winkles was insufficient, on its own, to support the development of a full assessment model. A number of studies reported the concentration factor (CF), not for winkles, but for other mollusc species. These species include mussels (Sombrito et al., 1982; Whitehead et al., 1988; Shunhua et al., 1997) and clams (Mayr et al., 1988; Cuvin-Aralar and Umaly, 1988, 1991). From these studies it appears that iodine CFs are generally higher than the IAEA (2004) reference value for molluscs of 10−2 m3 kg−1 but close to the generic value of 10−1 m3 kg−1 reported by Preston and Jefferies (1969). It also appears that, although not edible, the shell is an important sink for iodine. Uptake from seawater appears to be the predominant route of intake to the detriment of uptake from food. Information was also found on depuration kinetics (for mussels) suggesting that turnover is fast, with biological half-lives of 2–3 days and elimination following a simple exponential fall (Sombrito et al., 1982; Shunhua et al., 1997; Cuvin-Aralar and Umaly, 1988, 1991). The direct applicability of the above data to winkles is limited: concentration factors are routinely used in dose assessment calculations, so it is important that the values used should match the organism and nuclide of interest where possible. Inter-species variability also dictates that biological half-lives should relate to specific organisms. A knowledge gap was therefore identified, whilst the information available suggested a wide range of effects that might be expected in winkles. Laboratory studies were initiated to investigate how winkles assimilate radioiodine, by exposing them to labelled seawater and seaweed (Chondrus crispus). Such experiments involved the use of 131 I (simple to acquire and to detect) as surrogate tracer for the 129 I released into the Irish Sea by Sellafield. Experimental results are given in Vives i Batlle et al. (2005) and Wilson et al. (2005). Taken together, our results indicated that the way winkles incorporate iodine (particularly depuration) is more complex than other molluscs.
Uptake and depuration of radioiodine by the edible periwinkle Littorina littorea
451
The present study capitalises on the above experimental work and seeks to prove the hypothesis that a relatively simple biokinetic model is capable of reproducing the recent laboratory experiments on uptake and depuration of radioiodine by winkles including biological half-lives, concentration factors and basic body distribution of the activity. It is also hypothesised that the same model can provide a reasonable approximation to the available data for 129 I in winkles from the Cumbrian coastal area without requiring a separate calibration.
2. Model description The winkle biokinetic model developed in this study is consistent with equivalent models previously developed for 99 Tc in winkles and lobsters (Vives i Batlle et al., 2002; Olsen and Vives i Batlle, 2003). It was constructed and equations solved numerically using the modelling software ModelMaker® , version 3 (Walker, 1997; Hess et al., 1999). For the execution of the model the four available solution methods (Euler, Mid-point, Runge–Kutta and Bulirsh–Stoer) were tested. Runge–Kutta integration with accuracy of 0.001 and a minimum value of 10−10 was chosen, as this is a good all-purpose solver, which proved to be slightly more stable for this particular model than the other methods for a wide range of parameter variations. The model as a whole can be described as a first-order differential equation system containing 12 compartments, representing integrators in the model. The model parameters encapsulating the calibration information are detailed in Table 1. A conceptual representation of the Table 1 Parameters for the biokinetic model following optimisation Parameter
Value
Parameter
Seawater sub-model Coastal box volume Flushing T1/2 Recirculation T1/2 Feedback T1/2 Sediment T1/2 Sediment Kd
(1.13 ± 0.13) × 109 m3 1.1 ± 2.6 days 45 ± 15 days 70 ± 15 days 1.0 ± 0.1 days (2.7 ± 1.9) × 10−2 m3 kg−1
Winkles sub-model Tb1/2 for faecal matter Tb1/2 for shell Tb1/2 for storage organs Excretion T1/2 Delayed excretion T1/2 Uptake T1/2 for GI tract
Food sub-model Diatoms Tb1/2
1.6 ± 1.2 days
Diatoms CF Fraction ingested sediment Seaweed Tb1/2 Seaweed CF Seaweed to diatom food ratio Winkle ingestion rate
1.1 ± 0.1 m3 kg−1 0.1± 0.4 6.4 ± 1.0 days 0.28 ± 0.11 m3 kg−1 2.0 ± 0.1 (2.6 ± 0.6) × 10−5 kg d−1
Value 115 ± 12 days 279 ± 28 days 67 ± 7 days 0.44 ± 0.04 days 4.4 ± 0.4 days 23 ± 2 days
Uptake T1/2 for 211 ± 21 days shell from tissue Uptake T1/2 for 163 ± 16 days shell from water Uptake T1/2 for soft tissue 3.1 ± 0.3 days 33 ± 3 days Uptake T1/2 for storage organs Winkle shell mass (3.0 ± 0.2) × 10−4 kg Winkle soft parts mass (1.4 ± 0.7) × 10−4 kg Winkle whole mass (4.5 ± 0.3) × 10−4 kg
452
J. Vives i Batlle et al.
model is given in Fig. 1 where, for the purposes of intelligibility, the model comprises three main parts, linked by linear transfer rates and controlling influences: • The seawater sub-model, for simple calculation of 129 I concentrations in seawater in the vicinity of Sellafield from an input table of annual 129 I discharges. • The food sub-model, whose purpose is to calculate activity concentrations in the two principal food vectors for the winkle (micro-algae and seaweed). • The winkle sub-model, for calculating the distribution of activity, either ingested from food or absorbed directly from seawater, within different parts of the winkles themselves. 2.1. Seawater sub-model For the purposes of calculating the local seawater concentrations arising from 129 I authorised discharges, a simple hydrodynamic sub-model was developed, based on previous 99 Tc modelling work (Vives i Batlle et al., 2002; Olsen and Vives i Batlle, 2003). As a potential sink for radionuclides the model includes a sediment compartment, permitting uptake of sedimentbound iodine by winkles through ingestion. Iodine has a low affinity for sediment, with a Kd in the order of 2 × 10−2 m3 kg−1 (IAEA, 2004), suggesting that this mechanism is of little significance. The movement of water in the local area is represented by three compartments. The near field compartment (5 km wide × 15 km long × 15 m deep = 1.13 × 109 m3 ) represents the coastal strip where the winkles live, feed and are routinely harvested and sampled for monitoring purposes. From the near field, the water enters the “recirculation area”, representing hypothetical water hold-up south of Sellafield, as suggested in a technetium pulse study by Jackson et al. (2001). Recirculation was found to have little effect on modelled output, but this feature was retained for future use. Clearance of radioactivity towards the outer Irish Sea (“far field”) is regulated by a flushing time of 1.1 days, as calculated by MIKE21, a 2D engineering modelling tool for coastal hydrodynamics produced by the Danish Hydraulic Institute (DHI, 2001). 2.2. Food sub-model According to White and Keleshian (1994), Littorina littorea use seaweed directly for food. A typical diet would include green, brown and red algae, such as Ulva lactuca, Enteromorpha spp., Cladophora spp. and Ectocarpus spp. (Jackson, 2000). Investigation of winkles’ dietary preferences using algae extracts show that polyphenol anti-herbivore inhibitors, present in mature Fucus, deter the winkle from feeding on that species Watson and Norton (1985). In field trips across the Cumbrian area, we observed that winkles tended to graze on red seaweed (Chondrus crispus). Winkles are also known to feed on micro-algal biofilms from coated surfaces of rocks, containing diatoms (Davies and Beckwith, 1999). Hence, a food sub-model was designed based on the assumption that winkles feed principally on algae, whose concentration factor is more akin to red rather than green seaweed, as well as diatoms. Limited ingestion of sediment was also considered. A winkle diet variable combines these food inputs using ingestion rate data for winkles from Rickard (1994) and Vives i Batlle et al. (2002), in combination with our own laboratory-based observations (Vives i Batlle et al., 2005; Wilson et al., 2005).
Uptake and depuration of radioiodine by the edible periwinkle Littorina littorea
Fig. 1. Complete biokinetic model for 129 I uptake and turnover by winkles.
453
454
J. Vives i Batlle et al.
2.3. Winkle sub-model On representing the uptake of radioactivity by an organism, it is generally assumed that the intake is split into two fractions: one not directly retained (cleared through the gastro-intestinal tract) and another transferred to the rest of the body through the gut (to be cleared on a longer timescale). To this, the effect of radioactive decay must be added. The winkle sub-model (Fig. 1) is a more evolved version of this basic design. In this figure, a “gastro-intestinal tract” compartment receives 129 I from near-field seawater and connects with a “direct excretion” flow towards a “faecal matter” compartment, linked in turn to the near field seawater box (to represent decomposition). Additionally, a “shell” compartment covers direct adsorption and release of iodine from seawater onto the outer shell surface. In order to reproduce the multiphasic release pattern of iodine in winkles observed experimentally, two additional compartments have been included: “soft tissue” (receiving the input from the gastro-intestinal tract) and “storage organs”. It is not considered necessary to include tertiary release following uptake from seawater, as this process clears only a minor fraction of the iodine (Phase III Tb1/2 = 56 days, 11% loss – Vives i Batlle et al., 2005). In order to complete this semiempirical representation of the winkles, an additional process (uptake and release between soft tissue and shell) has been included, allowing for shell absorption of 129 I from inner tissue. It must be emphasised that the above representation is not physiologically based but guided by semi-empirical considerations; compartment names are indicative and not intended in any way to represent specific organs of the winkles. 2.4. Model data Three data sets were used in conjunction with the present model: • Input data, in the form of annual liquid 129 I discharges (TBq yr−1 ) from Sellafield for the period 1950–2030 (Fig. 2). Discharges for 1977–2002 were extracted from BNFL annual
Fig. 2. Sellafield annual 129 I liquid discharges 1950–2030 (BNFL, 1974–2002; Gray, 1998, pers. comm.; Morley, 2003, pers. comm.).
Uptake and depuration of radioiodine by the edible periwinkle Littorina littorea
455
reports (BNFL, 1974–2002). Data from 1952–1976 are from Gray (1998, pers. comm.). Future 129 I discharges for the period 2003–2030 were estimated from a projected modelling scenario proposed by BNFL for OSPAR submissions (Morley, 2003, pers. comm.). • The calibration dataset, consisting of direct experimental data on the uptake and depuration by winkles of 131 I from seawater and seaweed (Chondrus crispus) (Vives i Batlle et al., 2005; Wilson et al., 2005; Fig. 3). • The validation dataset. The principal difficulty in validating this model was the absence of a continuous monitoring record over the period 1950–present. However, the BNFL Environmental Monitoring Database contains 7 data for 129 I in winkles, corresponding to the brief period July 2002–April 2003. The datum for April 2003 was excluded in accordance with criteria for outliers (Dixon, 1953). The remaining data were combined with one additional measurement performed by the authors in late 1999 (Table 2).
Fig. 3. Comparison between experimentally measured (data points ±1σ ) and modelled (solid line) 129 I profiles in winkles: uptake and release from seawater (left) and seaweed (right).
Table 2 129 I activity concentration in winkles from the Sellafield coast (1999–2003)
Sampling date Activity (Bq kg−1 w.w.)
±1σ Sampling date Activity (Bq kg−1 w.w.)
±1σ
01-Dec-99 26-Jul-02 29-Jul-02 11-Sept-02
0.01 12-Sept-02 0.05 10-Dec-02 0.05 23-Jan-03 0.07
0.04 0.04 0.10
0.34 0.47 0.45 0.70
0.44 0.37 1.00
456
J. Vives i Batlle et al.
3. Model parameterisation 3.1. Parameterisation of the seawater and food sub-models Parameterisation of the seawater and food sub-models was based on previous modelling work for 99 Tc in lobsters (Olsen and Vives i Batlle, 2003) and winkles (Vives i Batlle et al., 2002). Information on Tb1/2 and CFs for iodine in diatoms was derived from Kuenzler (1967), Bowen, Coughtrey et al. (1984), IAEA (2004). Equivalent data for red seaweed were obtained from Polikarpov (1966), Sombrito et al. (1982), Coughtrey et al. (1984) and IAEA (2004). The ingestion rate used in Vives i Batlle et al. (2002) was updated using data from Wilson et al. (2005). In Table 3, data are given from which a mean ingestion rate of (8.3±1.6)×10−6 kg h−1 was derived. This rate was corrected to (2.6±0.6)×10−5 kg day−1 (consistent with Rickard, 1994) to take into account the fact that ingestion over 1 hour of feeding would not be sustained over a 24-h period, but rather it appears to end after 3 hours. 3.2. Parameterisation of the winkle sub-model In order to parameterise the winkles’ sub-model using laboratory data from Vives i Batlle et al. (2005) and Wilson et al. (2005), a stand-alone harness test was created, allowing separate calibration with a food or seawater input, as well as radioactive decay of the 131 I tracer used. In this test, the entry route for iodine from seawater was a constant concentration of 0.54 ± 0.03 Bq ml−1 (uptake) and 0 Bq ml−1 (depuration). The entry route for food was a variable supplying daily feeds of 131 I, each of 1 h duration, as per the experimental procedure. Each route could be activated or de-activated to model uptake from seawater and food separately. Test model parameters were calculated either by direct derivation from experimental data (>70% of the parameters) or by adjustment to reproduce the overall concentration profiles of the laboratory experiments. Additional parameter adjustment using the ModelMaker® optimisation algorithm was then carried out to improve the fit between modelled and experimental data, resulting in changes not exceeding 20%. Final values for all parameters are given in Table 1. 3.3. Model process testing The modelled profiles of 129 I entering winkles via seawater and seaweed are given in Fig. 3. The modelled uptake of 129 I from seawater was successful in the following respects: Table 3 Ingestion rate of Chondrus crispus for winkles, expressed as an average of 1–2 hour determinations over 11 days (based on Wilson et al., 2005) Batch
Feed duration (h)
Activity ingested (Bq)
Seaweed activity (Bq kg−1 )
Ingestion (kg per winkle)
Ingestion rate (kg h−1 per winkle)
Batch A Batch B Batch C
1.4 ± 0.3 1.6 ± 0.3 1.8 ± 0.4
41 ± 17 40 ± 23 38 ± 28
(3.7 ± 1.6) × 105 (3.7 ± 1.6) × 105 (3.7 ± 1.7) × 105
(1.3 ± 0.5) × 10−5 (1.3 ± 0.9) × 10−5 (1.2 ± 0.7) × 10−5
(9.8 ± 5.6) × 10−6 (8.2 ± 5.6) × 10−6 (6.7 ± 4.2) × 10−6
Uptake and depuration of radioiodine by the edible periwinkle Littorina littorea
457
• The experimental profiles for both the uptake and multiphasic release of 131 I from labelled seawater (and by extension, the relevant CFs and multiphasic Tb1/2 ’s) were reproduced by the model (R 2 = 0.99; Fig. 4). • The basic body part distribution of activity for L. littorea could be reproduced (Table 4)1 . The modelled uptake of 129 I from seaweed was successful in the following respects: • The feeding/depuration sequence (activity increments followed by depuration after each feed) was reasonably reproduced, although less well than the uptake and release from seawater (R 2 = 0.89; Fig. 4). There was a small time drift of about 8% for the modelled daily peaks of activity. 1 The following criteria were used to compare dissected parts and model variables: Digestive gland = Gastrointestinal (GI) tissue; operculum + remainder = organs + rest of the body; soft parts = GI tissue + organs + rest of the body; edible fraction = soft tissue + GI tissue.
Fig. 4. Model prediction for the uptake and turnover of 129 I compared with experimental results: Seawater exposure (left) and ingested seaweed exposure (right).
Table 4 Modelled and experimentally observed fractions of 131 I in winkles Experiment stage
Modelled (%)
Experimentally observed (%)
Digestive tissue
Other tissue
Shell
Digestive tissue
Other tissue
Shell
Uptake/release from seawater End of uptake (t = 21 days) End of experiment (t = 45 days)
12.86 0.00
20.88 10.68
66.26 89.32
7±1 ND
17 ± 11 ND
76 ± 16 76 ± 18
Uptake/release from seaweed End of uptake (t = 10 days) 24 h after uptake (t = 11 days) End of experiment (t = 21 days)
28.28 20.55 2.61
71.37 78.94 90.26
0.35 0.51 7.13
73 ± 24 48 ± 32 66 ± 7
18 ± 14 42 ± 34 26 ± 9
9±6 11 ± 5 7±1
458
J. Vives i Batlle et al.
• The final depuration sequence, considered separately (and by extension, the relevant CFs and multiphasic Tb1/2 ’s) was also well predicted by the model (R 2 = 0.97). • The activity distribution in the body of L. littorea could only be partially reproduced (Table 4). The model successfully predicted that, following ingestion from seaweed, radioiodine partitioned towards the soft parts in favour of shell (the opposite from uptake from seawater). However, the model does not predict the observation that the digestive gland holds the majority of the activity in comparison with the remaining flesh. We attribute this discrepancy to uncertainties in dissecting and γ -counting minute winkle parts post-ingestion, compounded with the accepted limitation of the model to provide a detailed representation of the winkles at a physiological level. From the above testing it is concluded that the model provides an integrated representation for the uptake and depuration of radioiodine from both seawater and food ingestion, using a single calibration. This provides confidence that the model as designed is fully functional in representing the processes modelled.
4. Model validation Having tested that the model functionality, the model was validated using an independent data set as given in Table 2. 4.1. Comparison of modelled and monitoring (field) results A validation model run was carried out using the Sellafield discharge data for the period 1950–2030 as input. Results are given in Fig. 5. Winkle activity concentrations (edible fraction) follow the general profile of the discharge (Figs. 2 and 5) as expected, with a plateau in the order of 1 Bq kg−1 being reached between 2003 and 2020. All the data available from the BNFL Environmental Monitoring Database correspond to the transitional period 1994–2003, where discharges increased by one order of magnitude. For this period, the model approximates the data reasonably well. A slight tendency for the model to over-predict could be coincidental, or could be due to the fact that the winkles as monitored were cooked before analysis (potential loss of activity in this way is not modelled). Measured and modelled 129 I activities in winkles expressed as an annual average are statistically compatible. For the three years for which data are available, some linear correlation exists (modelled 129 I = 0.78 × measured 129 I + 0.21, R 2 = 0.83). In future work this limited comparison will be improved with additional field measurements. The model predicts that the majority (∼90%) of the activity retained by winkles originates from seawater uptake alone (Table 4). This prediction involves the assumption that the ratio of seaweed ingestion to diatom ingestion is about 2; however model output is relatively insensitive to variations of this ratio (−14% when varying between 0 and 100). The model predicts that the edible winkle fraction is 18 ± 7% in terms of 129 I activity (remainder in shell and inedible parts), and that the whole soft parts fraction is 42 ± 2%. The seawater uptake experiment suggested that, in regularly fed winkles, the soft parts contained ∼35% of 131 I (Table 4). Uptake experiments from algae suggested a much higher partitioning. Modelled results therefore approach the analytical data for seawater uptake.
Uptake and depuration of radioiodine by the edible periwinkle Littorina littorea
459
Fig. 5. Model prediction of 129 I concentration in winkles 1950–2030 (solid line) and associated confidence interval (dashed line), in comparison with available monitoring data (data points ±1σ ).
4.2. Confidence interval and sensitivity analysis All model parameters have an associated statistical uncertainty, which was directly quantified or assumed to be ±10% if data were not available. Using this information, a confidence interval around the model prediction was calculated (displaying at ±1σ ). The flushing halftime and near-field volume were excluded from the calculation because they are dependent variables, as were the nominal mass values for sediments and bioindicators. The calculated confidence interval indicates that the model is stable, i.e. varying parameters within their uncertainty ranges does not affect model output excessively. The typical range of variation was ±13% (Fig. 5). Generally speaking, it is considered successful when model output can be tied to an uncertainty range that does not exceed the current order of magnitude. The key parameters for which the model was more sensitive to were: • Uptake half-time for shell from water (regulates how much iodine goes into the shell). • Biological half-life for shell (determines rate of return from shell to water and thus, with the above, influences the CF for shell uptake which is the main route). • Uptake half-time for GI tract (regulates how much iodine is retained by the winkle). • Excretion half-time (regulates how much iodine is released by the winkle). • Flushing half-time (regulates how the discharge is dispersed away from the local area). 4.3. Running the model in predictive mode: an OSPAR application Modelled 129 I activity concentrations in winkles for the period 2001–2030 under the OSPAR scenario are given in Fig. 5. The model predicts that winkles will maintain a 129 I activity concentration of about 1 Bq kg−1 in 2004, remaining at a similar level until after the year 2022,
460
J. Vives i Batlle et al.
Table 5 Modelled ranges of committed effective dose (CED) and internal absorbed dose rate to winkles (WID), unweighted and weighted by radiation quality Period
Activity (Bq kg−1 )
CED (µSv yr−1 )
uWID (µGy yr−1 )
wWID (µGy yr−1 )
1950–1968 1969–1992 1993–2002 2003–2023 2024–2030
0.006–0.015 0.04–0.16 0.12–0.73 0.6–1.1 0.04–0.10
0.001–0.004 0.01–0.22 0.08–0.57 0.4–0.8 0.03–0.08
0.002–0.005 0.01–0.06 0.04–0.25 0.2–0.4 0.01–0.03
0.003–0.006 0.02–0.07 0.05–0.31 0.2–0.4 0.02–0.04
followed by a sharp decrease of more than one order of magnitude in the ensuing 2 years. By 2030 129 I concentrations are predicted to stabilise at about 0.05 Bq kg−1 . 4.4. Exposure calculations Committed effective doses (CED) associated with the consumption of winkles by humans were calculated for all years from 1950 to 2030 using the ICRP-67 methodology (ICRP, 1993). Results are given in Table 5. BNFL annual reports contain consumption rates (CR ) for the marine critical group for the years between 1981 and 2002, which have been used for this calculation. Years prior to 1981 were assumed to have the same CR as 1981. Similarly years post-2002 were assumed to have the CR of 2002. In all cases, dose estimates are based on a dose per unit intake via ingestion factor of 1.10 × 10−7 Sv Bq−1 (BNFL, 2003). In BNFL reports prior to 2002, 129 I doses are given for consumption of molluscs in general. There is only one single annual dose reported for winkles, the 2002 value of 0.36 µSv yr−1 . This value is reasonably compatible with the model-predicted value of 0.57 µSv yr−1 for that year. In magnitude this dose, as well as the maximum of 0.83 µSv yr−1 modelled for 2014, have no radiological significance and are much lower than the dose constraint of 300 µSv yr−1 to the members of the public for a new facility (National Radiological Protection Board, 1993). Absorbed dose rates to winkles from 129 I, calculated using the FASSET methodology (Brown et al., 2003; Vives i Batlle et al., 2004) are also given in Table 5. This table contains absorbed dose rates, both unweighted and weighted by radiation quality (uWID and wWID, respectively). For winkles, it is the dose from internally incorporated 129 I that is of relevance, because external doses from exposure to 129 I in sediments are negligible. The highest predicted weighted dose rate was 0.45 µGy yr−1 in 2014. The lowest dose rate from chronic exposure that may produce changes in behaviour or development is about 10 mGy yr−1 (Rose, 1992). Hence, the model predicts that Cumbrian winkles are not at risk as far as 129 I exposure is concerned. 5. Conclusions A model for the prediction of 129 I concentrations in the edible periwinkle Littorina littorea following uptake from seawater and food has been successfully developed, calibrated and
Uptake and depuration of radioiodine by the edible periwinkle Littorina littorea
461
tested. The hypothesis that a relatively simple biokinetic model is capable of reproducing closely the behaviour of winkles in the laboratory and that, additionally, the model can provide a reasonable approximation to the available monitoring data for 129 I in winkles from the Sellafield area (spanning several years), has been demonstrated. Modelling of 129 I exposure to winkles and human consumers in the west Cumbrian coastal area reveals that doses have no radiological significance under either the current or predicted discharge regime. Despite this, the model is a useful tool for simulating short-term releases (such as accidents or post-closure scenarios) in which radiological exposures could potentially become more significant. The biokinetic model presented here provides a dynamic representation of how short-term discharge fluctuations influence variability at a local level. In this respect, the present methodology supersedes the concentration factor approach, which is generally valid only in environments under equilibrium or near-equilibrium conditions. Further work is recommended to calculate the proportion of diatoms, sediment and macroalgae that constitute the winkle’s diet in the natural environment, making the food sub-model more realistic. However, confidence interval calculations suggest that the model is reliable and the uncertainty affecting these parameters has only a limited effect. References BNFL (1974–2002). Discharges and Monitoring of the Environment in the UK. Annual Reports 1986 to 2003. BNFL, Risley, Warrington. BNFL (2003). Discharges and Monitoring of the Environment in the UK. Annual Report 2003. BNFL, Risley, Warrington. Brown, J., Gomez-Ros, J.-M., Jones, S.R., Pröhl, G., Taranenko, V., Thørring, H., Vives i Batlle, J., Woodhead, D. (2003). Dosimetric models and data for assessing radiation exposures to biota. G. Pröhl (Ed.), FASSET Deliverable 3 Report. Contract No. FIGE-CT-2000-00102. Coughtrey, P.J., Jackson, D.J., Jones, C.H., Thorne, M.C. (1984). Radionuclide Distribution and Transport in Terrestrial and Aquatic Ecosystems – A Critical Review of Data, vols. 1–6. Balkema, Rotterdam. Cuvin-Aralar, M.L.A., Umaly, R.C. (1988). Uptake and elimination of iodine-131 by the freshwater clam Corbicula manilensis Philippi from water. Nat. App. Sci. Bull. 40, 141–158. Cuvin-Aralar, M.L.A., Umaly, R.C. (1991). Accumulation and tissue distribution of radioiodine (131 I) from algal phytoplankton by the fresh-water clam Corbicula manilensis. Bulletin of Environmental Contamination and Toxicology 47 (6), 896–903. Davies, M.S., Beckwith, P. (1999). Role of mucus trails and trail-following in the behaviour and nutrition of the periwinkle Littorina littorea. Marine Ecology – Progress Series 179, 247–257. DHI (2001). MIKE21 Environmental Hydraulics. DHI Water and Environment, Denmark. Dixon, W.J. (1953). Processing data for outliers. Biometrics 9, 74–89. Gray, J. (1998). Personal communication, BNFL. Hess, T., Matthews, R., Quinton, J. (1999). Modelling Environmental Systems. Cranfield Univ., Bedford, 59 pp. IAEA (2004). Sediment Distribution Coefficients and Concentration Factors for Biota in the Marine Environment. IAEA Technical Report Series, vol. 422. IAEA, Vienna, 95 pp. ICRP, International Commission on Radiological Protection (1993). Age-Dependent Doses to Members of the Public from Intake of Radionuclides. Part 2: Ingestion Dose Coefficients. ICRP Publ. 67, Annal. ICRP, vol. 23 (3/4). Pergamon Press, Oxford, 167 pp. Jackson, A. (2000). Littorina littorea. Common periwinkle. Marine life information network: Biology and sensitivity key information sub-programme. Available at http://www.marlin.ac.uk. Jackson, D., Vives i Batlle, J., Parker, T.G., Whittall, A.J., McDonald, P. (2001). Considerations in optimising the marine discharge regime for 99 Tc. In: Proc. British Nuclear Energy Society Conf. Radiation Dose Management in the Nuclear Industry. Windermere, UK, 14–16 May 2001.
462
J. Vives i Batlle et al.
Kuenzler, E.J. (1967). Elimination of iodine, cobalt, iron, and zinc by marine zooplankton. In: Nelson, D.J., Evans, F.C. (Eds.), Proc. 2nd National Symposium on Radioecology. Ann Arbor, Michigan, May 15–17, 1967. USAEC Conf. No. 670503, pp. 462–473. Mayr, L., Moraes, R., Lopes, M.A., Vicente, C., Mauro, J.N. (1988). Transit and absorption of nuclear industry derivatives by marine biota. In: Proc. 2nd Gen. Conf. Nucl. Energy, Rio de Janeiro, vol. 3, pp. 281–293. McDonald, P., Cook, G.T., Baxter, M.S. (1992). Natural and anthropogenic radioactivity in coastal regions of the UK. Radiation Protection Dosimetry 45 (1–4), 707–710. McDonald, P., Baxter, M.S., Fowler, S.W. (1993). Distribution of radionuclides in mussels, winkles and prawns. 1. Study of organisms under environmental-conditions using conventional radio-analytical techniques. Journal of Environmental Radioactivity 18 (3), 181–202. Morley, R.G. (2003). Personal communication, data supplied by BNFL. National Radiological Protection Board (1993). Occupational, public and medical exposure: Guidance on the 1990 Recommendations of ICRP. Documents of the NRPB, vol. 4, No. 2. Olsen, Y.S., Vives i Batlle, J. (2003). A model for the bioaccumulation of 99 Tc in lobsters (Homarus gammarus) from the West Cumbrian coast. Journal of Environmental Radioactivity 67 (3), 219–233. Polikarpov, G.G. (1966). Radioecology of Aquatic Organisms. North-Holland, New York, 314 pp. Preston, A., Jefferies, D. (1969). Aquatic aspects in chronic and acute contamination situations. In: Environmental Contamination by Radioactive Materials. IAEA, Vienna, pp. 183–214. Rickard, A. (1994). The influence of body size on the uptake of radionuclides in Cumbrian winkles – implications for shellfish consumers. BSc thesis. King’s College, Univ. of London. Rose, K.S.B. (1992). Lower limits of radiosensitivity in organisms, excluding man. Journal of Environmental Radioactivity 15 (2), 113–134. Shunhua, C., Qiong, S., Xiaokui, K. (1997). Effects of body size on accumulation and distribution of 125 I in the green mussel (Perna viridis), Beijing, China. Nucl. Inf. Centre Report CNIC-01210. Sombrito, E.Z., Banzon, R.B., de la Mines, A.S., Bautista, E. (1982). Uptake of 131 I in mussel (Mytilus smaragdinus) and algae (Caulerpa racemosa). The Nucleus 22 (1), 83–89. Swift, D.J. (1989). The accumulation and retention of Tc-95m by the edible winkle (Littorina-Littorea L). Journal of Environmental Radioactivity 9 (1), 31–52. Swift, D.J., Smith, D.L., Allington, D.J., Winpenny, K. (1995). A laboratory-study and field-study of Po-210 depuration by edible winkles (Littorina-Littorea L.) from the Cumbrian Coast (North-eastern Irish Sea). Journal of Environmental Radioactivity 26 (2), 119–133. Vives i Batlle, J., McDonald, P., Parker, T.G. (2002). Biokinetic studies of 99 Tc in the edible winkle (Littorina littorea). In: Proc. Int. Conf. Radioactivity in the Environment. 1–5 Sept. 2002, Monte Carlo, Monaco. Vives i Batlle, J., Jones, S.R., Gómez-Ros, J.M. (2004). A method for calculation of dose per unit concentration values for aquatic biota. Journal of Radiological Protection 24, A1–A22. Vives i Batlle, J., Wilson, R.C., Mc Donald, P., Parker, T.G. (2005). Uptake and depuration of 131 I by the edible winkle Littorina littorea: Uptake from seawater. Journal of Environmental Radioactivity 78 (1), 51–67. Walker, A. (1997). ModelMaker 3 User Manual. Cherwell Scientific Ltd., Oxford, 362 pp. Watson, D.C., Norton, T.R. (1985). Dietary preferences of the common periwinkle, Littorina littorea (L.). Journal of Experimental Marine Biology and Ecology 88, 193–211. White, S., Keleshian, M. (1994). A field guide to economically important seaweeds of northern New England. New Hampshire Sea Grant Marine Advisory Program, MSG-E-93-16. University of Maine. Whitehead, N.E., Ballestra, S., Holm, E., Huynhngoc, L. (1988). Chernobyl Radionuclides in Shellfish. Journal of Environmental Radioactivity 7 (2), 107–121. Wilson, R.C., Vives i Batlle, J., McDonald, P., Parker, T.G. (2005). Uptake and depuration of 131 I by the edible periwinkle Littorina littorea: Uptake from labelled seaweed (Chondrus crispus). Journal of Environmental Radioactivity 80 (3), 259–271.
463
Environmental modelling: Modified approach for compartmental models M. Iosjpe* Norwegian Radiation Protection Authority, P.O. Box 55, N-1332 Oesteraas, Norway Abstract Compartment modelling is used for evaluation of radiological consequences and dispersion of radionuclides in the marine environment. One of the general features of the compartment modelling is an instantaneous mixing of contaminants in the ocean. The influence on radionuclide dispersion in the marine environment and radioecological assessment of the developed modified approach for compartment modelling with non-instantaneous mixing of contaminants in oceanic space is discussed. New modifications of the model (calculations of doses to biota and ice transport of radionuclides) underline the significance of the modified approach. Simulations provided by the modified approach indicate that the algorithm with non-instantaneous mixing in oceanic space can predict the different time trends for ice and water transport in the Arctic Ocean (for instance, more rapid transport of contaminants by ice through the Kara Sea to the Fram Srait in comparison to water transport). Keywords: Modelling, Non-instantaneous mixing, Radionuclides, Marine environment
1. Introduction An evaluation of radiological consequences after the release of radionuclides into the marine environment can include different aspects: contamination of the water and bottom sediments, bioaccumulation of radionuclides in biota, and dose assessments to marine organism and human populations. For such analysis it is often necessary to cover large distances and long time scales. Compartment models can be used for modelling radiological consequences with spatial and temporal scales of several thousand kilometres and millenniums, respectively. This approach covers whole processes such as dispersion of radionuclides in oceanic space, transfer of radioactivity between seawater and sediments, uptake of radionuclides by biota and, finally, dose calculations for man and biota, which are important for radiological assessment. Therefore, compartmental modelling has been recommended by the European Commission for radiological assessment (CEC, 1990, 1994; EC, 1995, 1997, 2000, 2003; Nielsen et al., 1997; IAEA, 2003). * Address: Norwegian Radiation Protection Authority, Grini naringspark 13, P.O. Box 55, N-1332 Oesteraas, Norway; phone: (+47) 67 16 26 02; fax: (+47) 67 14 74 07; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08037-X
© 2006 Elsevier Ltd. All rights reserved.
464
M. Iosjpe
Instantaneous mixing of radionuclides in each compartment is a principal feature of the compartmental modelling, but this general assumption leads, in practical calculations, to instantaneous mixing in the whole ocean space. This feature of traditional modelling is not liable to correctly describe the water motion in large oceanic systems and, therefore, inserts an additional systematic uncertainty into calculated results. The model developed by the Norwegian Radiation Protection Authority (NRPA) is based on a compartment modelling approach which includes terms describing the radionuclide dispersion into oceanic space with time (non-instantaneous mixing in oceanic space). Some estimations of how this approach influences dispersion of radionuclides and radiological assessment were conducted by Iosjpe et al. (2002). It showed that the modified approach for compartment modelling keeps all features of compartment modelling and, at the same time, gives a more realistic/physical approach compared to traditional compartment modelling. Simulations show that the modified approach is more sensitive to processes near the sources of contamination and, also, during the initial time of radionuclides dispersion. Results between modified and traditional approaches differ widely, especially for scenarios where redistribution of radionuclides between different marine areas was important. In this article the discussion concerning the modified approach for compartment modelling will be continued with regards to the new modifications of the approach and new effects for dispersion of contaminants in marine environment and radioecological assessment.
2. Model description The equations describing transfer of radionuclides between the compartments are of the form dAi kj i Aj − kij Ai γ (t Tj ) − ki Ai + Qi , = dt
t Ti ,
Ai = 0,
t < Ti ,
n
n
j =1
j =1
(1)
where kii = 0 for all i, Ai and Aj are activities (Bq) at time t in compartments i and j ; kij and kj i are transfer rates (yr−1 ) between compartments i and j ; ki is an effective activity transfer rate (yr−1 ) from compartment i taking into account loss of material from the compartment without transfer to another, for example radioactive decay; Qi is a source of input into compartment i (Bq/yr); n is the number of compartments in the system. Ti is the time of availability for compartment i (the first times when compartment i is open for dispersion of radionuclides) and γ is an unit function,
1, t Ti , γ (t Ti ) = 0, t < Ti . The availability times Ti =
min
μm (v0 ,vi )∈Mi
wj k
j,k
are calculated as a minimised sum of the weights for all paths μ0 (v0 , . . . , vi ) from the initial compartment (v0 ) with discharge of radionuclides to the compartment i on the oriented graph
Environmental modelling: Modified approach for compartmental models
465
Fig. 1. Graph elements and available paths.
G = (V , E) with a set V of nodes vj correspondent to compartments and a set E of arcs ej k correspondent to the transfer possibility between the compartments j and k (graph elements and examples of available paths from the node v0 to the node vi are illustrated by Fig. 1). Every arc ej k has a weight wj k which is defined as the time required before the transfer of radionuclides from compartment j to compartment k can begin (without any way through other compartments). Weight, wj k , is considered as a discrete function F of the water fluxes fj k , fkj between boxes j and k, geographical information gj k and expert evaluation Xj k . Mi is a set of feasible paths from the initial compartment (v0 ) to the compartment i (vi ). Traditional compartment modelling is a particular case of the present approach because the system of equations (1) becomes identical to the traditional system of equations dAi kj i Aj − kij Ai − ki Ai + Qi , = dt n
n
j =1
j =1
where all times of availability are zero. Figure 2 shows the surface structure of the model compartments for the Arctic Ocean, the Nordic Seas and the North Atlantic. The structure of the compartments for surface, mid-depth and deep waters (Fig. 3) is developed with regards to the improved description of Polar, Atlantic and Deep waters in the Arctic Ocean and the Northern Seas (Karcher and Harms, 2000) and site-specific information for description of the compartments. The volume of the water layers in each compartment has been calculated by using a detailed bathymetry in geographical information system (IBCAO, 2001; ETOPO-5, 2002). A schematic structure of the processes involved in modelling is shown in Fig. 3. The model includes the processes of advection of radioactivity between compartments, sedimentation, diffusivity of radioactivity through the pore water, resuspension, mixing due to bioturbation and partical mixing and a burial process of activity in deep sediment. Radioactive decay is included in all compartments. The contamination of biota is further calculated from the radionuclide concentrations in filtered seawater in the different water. Doses to man are calculated on the basis of data for the catch of seafood and assumptions concerning human diet.
466
M. Iosjpe
Fig. 2. The structure of the surface water boxes.
Fig. 3. A schematic structure of the processes involved in modelling.
Environmental modelling: Modified approach for compartmental models
467
Fig. 4. The linkage between ice, water and sediment boxes.
The present version of the model is improved to calculate doses to biota, which are calculated on the basis of radionuclide concentrations in marine organisms, in water and sediment phases and radionuclides dose conversion factors (Iosjpe et al., 2003). The ice module was developed for evaluation of the ice transport of radionuclides and includes processes of interactions of radioactivity between water, suspended sediment particles in water column, bottom sediment, ice, ice sediment and different (water, sediment and ice) model compartments (Fig. 4).
3. Results and discussions In the following, the Kara Sea region (Figs. 2 and 5) will be mainly used as a source for radionuclides dispersion in the marine environment. This region includes a shallow Arctic Sea with Siberian rivers (the Ob and the Yenisey) estuaries and with water and ice exchange with the Arctic Ocean. On the other hand, dumped radioactive waste in the Kara Sea and the possibility of radionuclides discharges into the Kara Sea through the Ob River from the Mayak facilities, were under consideration in recent years (IAEA, 2003; NRPA, 2004). 3.1. Radioecological module Results of the calculations of the dispersion after six months of 1 TBq of 60 Co discharged into the Ob Bay are shown in Fig. 5. The time of availability for the east part of the Kara Sea is
468
M. Iosjpe
Fig. 5. Simulations of the dispersion of 60 Co (dark grey colour) after six months of 1 TBq discharged into the Ob Bay.
TKSE = 0.57 years for this scenario. Therefore, dispersion of 60 Co is still limited by the Ob Bay boundaries six months after the discharge of the radionuclide. Hence the present model predicts compartments without contamination in the initial phase of dispersion from a point source. Without an algorithm, which provides non-instantaneous mixing in oceanic space (traditional modelling with TKSE = 0), the 60 Co dispersion will cover all compartments in Fig. 2 after the first time step of calculations. It should be pointed out that time of availability from the present model, which was founded as a simple estimation of residence time as ratio VKSE /fKSE (here VKSE and fKSE are compartment volume and water flux, correspondently) and residence time 0.55 year for the Ob Bay according to the two-dimensional finite-element model calculations (Paluszkiewicz et al., 1997) are in close agreement. As pointed out above, the present version of the model uses dose conversion factors conceived for the assessment of internal and external doses to the marine organisms (FASSET, 2003). An approach involves the selection of a set of reference organisms (Pentreath, 1999) that can be used as representatives of the marine environment in its entirety. The assessment of internal and external doses to the marine organisms is based on assumptions on uniform distributions of radionuclides in organisms as well as surrounding water and sediment compartments. The calculations of the total dose are based here on the expression D˙ = kW · DCF (E) · CW + kS · DCF (E) · CS + DCF (I) · CI , where D˙ I is a dose rate (µGy/h); CW , CS and CI are concentrations (Bq/kg) of radionuclide in water, sediment and marine organism, correspondently; DCF (E) and DCF (I) are dose conversion factors (µGy/h per Bq/kg) for internal and external doses (FASSET, 2003); kW and kS are weight coefficients for water and sediment factors. For example, for benthic fish it is considered here that water and sediment phases as semi-infinite systems and, therefore, kW = kS = 0.5.
Environmental modelling: Modified approach for compartmental models
469
Fig. 6. Internal, external and total doses to the benthic fish in the Ob Bay.
The present modelling approach is illustrated by calculations of the dose to the benthic fish in the Ob Bay (Fig. 6) according to the above-mentioned scenario. Figure 6 shows that during the initial time of the radionuclide dispersion influence of the external dose increases and after, approximately, one year the external dose dominates the internal dose. Figure 7 shows a comparison between present approach and modelling with instantaneous mixing in oceanic space. Results of calculations in Fig. 7 show that even for the initial compartment (the Ob Bay), when the process of dispersion starts at the same time for both approaches (time of availability for the Ob Bay compartment is zero), differences between results are up to 30 percentages. The present model predicts higher results for doses as well as for concentrations in water and sediment phases, because under the modified approach simulations 60 Co activity remains only in the Ob Bay compartments during the first half year of radionuclide dispersion. It is necessary to note that 60 Co concentration in fish was calculated here with concentration factor 103 (IAEA, 2004; Nielsen et al., 1997). 3.2. Ice module Let us consider another example of radiological assessment with regards to the other modification of the model, namely, transport of radionuclides by ice. Ice transport is a unique pathway in polar areas (Pfirman et al., 1995, 1997; Strand et al., 1996), which is not directly linked to the movement of water mass (Pfirman et al., 1997). A schematic map of the ice drift in the Arctic Ocean (ARCTICMAR, 2002) is shown in Fig. 8. The potential pathways for incorporation of radioactivity into ice are widely discussed in AMAP (1998). The present model includes description of the processes concerning the ice transport of radionuclides (Fig. 4). Transfer of radioactivity to the ice compartment from the liquid phase of the water column, from the suspended sediment particles in the water column and from the bottom sediment compartment is described assuming that transfer of radioactivity varies directly as the freezing process. Transfer of radioactivity through ice transport between different
470
M. Iosjpe
Fig. 7. Comparison of the dose calculations to benthic fish and 60 Co concentrations in water and sediment compartments in the Ob Bay.
Environmental modelling: Modified approach for compartmental models
471
Fig. 8. The schematic map of the ice drift in the Arctic Ocean.
sea areas is described on the basis of the ice fluxes. Transfer of radioactivity into sea water is described through melding processes assuming that transfer of radioactivity varies directly as the melding process. A more detailed description of the model ice module can be obtained from (Iosjpe, 2002). The results concerning the ice transport of radionuclides are illustrated by the transport of radionuclides from the Kara Sea to the Fram Strait through the Arctic Ocean after the releases of 1 TBq of radionuclides into the Ob Bay of the Kara Sea. Comparisons of the ice and water fluxes as well as sediment masses, which can be transported by water and ice correspondently from the Kara Sea, show that the annual content of radionuclides in the ice phase is significantly smaller (ANWAP, 1997). On the other hand, ice transport of contaminants through the Kara Sea to the Fram Strait is a more rapid pathway in comparison to water transport (AMAP, 1998). Therefore, simulations of the radionuclides dispersion with regards to noninstantaneous mixing in oceanic space are especially interesting for this case. Figure 9 shows the incorporation of 90 Sr and 60 Co in the ice compartment of the Kara Sea, transfer of the radionuclides from the Kara Sea to the Fram Strait through the Central Arctic and radionuclides concentrations in the ice and water compartments. A figure indicates clearly that concentration of 90 Sr in ice is significantly lower than the concentration in the seawater of the Kara Sea and the additional inventory of 90 Sr in the Frame Strait is insignificant in comparison to the seawater impact. Contrary to 90 Sr dispersion, Fig. 9 shows that the concentrations of 60 Co in the Fram Strait, provided by the ice and water transport, are relatively similar.
472
M. Iosjpe
Fig. 9. Comparison between concentrations of 90 Sr (left) and 60 Co (right) in the ice and seawater compartments.
Generally, simulations show a significant influence of the sediment distribution coefficient (Kd ) on results of calculations and high Kd values lead to relatively high concentrations of radionuclides in the seawater during the ice melting. The effect of radionuclides redistributing by ice transport is illustrated in Fig. 10 and Table 1 (calculations correspond to 1 TBq discharges of radionuclides into the Ob Bay of the Kara Sea). Figure 10 shows concentrations of 241 Am in lobsters in the Greenland Sea with regards to the dispersion of the radionuclide with and without ice transport. Results demonstrate that during the initial time of dispersion, concentrations differ dramatically. Calculations of the collective doses to man in Table 1 show that additional considerations of ice transport of radionuclides is insignificant for radionuclides with low sediment distribution coefficients, but even for the long time scale the ice transport is significant for the radionuclides with high Kd values. It is necessary to note that demonstrations of different time trends for ice and water transport in Figs. 9 and 10, as well as differences in Table 1, is possible only because the present approach uses an algorithm with non-instantaneous mixing in oceanic space, which is more sensitive to the redistribution of radionuclides in comparison to the traditional approach. The practical feasibility of the present approach for compartment modelling was demonstrated in (Iosjpe et al., 2002) by the comparison between model predictions and experimental data. The simulation results were generally in good agreement with observations. In the present model the remobilisation part of the water–sediment interaction algorithm described by Chartier et al. (1989) and Nielsen (1999) was changed with a more flexible algorithm described by MacKenzie and Nicholson (1987) and Simmonds et al. (2002), which was also supported by the evaluation of site-specific information to define the model parameters by Mitchell et al. (1999). Therefore, the comparison of the calculations and the experimental data for 239 Pu discharge from Sellafield and global fallout (Nielsen and Hou, 2002) for sediment compartments in the Irish Sea is shown in Fig. 11 and has to confirm feasibility of the model. The solid line in Fig. 11 corresponds to calculation results executed by the present model. Circles show the average values while the error bars show the minimal and maximal values of the experimental data. Figure 11 demonstrates that the model prediction has reasonable accuracy.
Environmental modelling: Modified approach for compartmental models
473
Fig. 10. Comparison of 241 Am concentrations in lobster in the Greenland Sea for scenarios with and without the ice transport.
Table 1 The effect of the ice transport of radionuclides for the doses to man. Scenarios A and B correspond to the calculations with and without the ice module, respectively. The effect of the ice transport is calculated as ratio between doses for sce(A) (B) narios A and B (Dt /Dt ) for 1000 years (A)
Radionuclide
Scenario A (man Sv)
Scenario B (man Sv)
Dt
90 Sr
2.01×10−4 1.41×10−3 2.30×10−1 8.14×10−2 2.22×10−5
2.01×10−4 1.41×10−3 2.28×10−1 5.84×10−2 7.97×10−6
1.0 1.0 1.0 1.4 2.8
137 Cs 239 Pu 241 Am 60 Co
(B)
/Dt
4. Conclusions New modifications (ice transport of radionuclides and dose assessment to biota) were provided to the compartment modelling approach with non-instantaneous mixing of radionuclides in oceanic space. These modifications underline that the present approach is available for new, more sensitive and realistic possibilities to evaluate the dispersion of radionuclides in marine environment and dose assessment. At the same time the approach keeps all features of the traditional compartment modelling.
474
M. Iosjpe
Fig. 11. Comparison of the results of the calculation and experimental data for discharge of 239 Pu from Sellafield and global fallout for the east (left) and west (right) parts of the Irish Sea.
Model simulations showed that ice transport of radionuclides can be a significant factor for radionuclides with high sediment distribution coefficients when ice transport of contaminants is a more rapid pathway in comparison to water transport. The remobilisation part of the water–sediment interaction algorithm was changed with a more flexible algorithm. Validation of the model confirms a satisfactory comparison with experimental data.
Acknowledgements This work has been support by the EC, contracts FIGE-CT-2000-00102 (FASSET), FIGECTT-2000-00085 (REMOTRANS) and ICA2-CT-2000-10032 (EPIC), and the Norwegian Ministry of the Environment.
References AMAP (1998). Arctic pollution issues. AMAP Assessment Report. Arctic Monitoring and Assessment Programme, Oslo. ANWAP (1997). Radionuclides in the Arctic Seas from the former Soviet Union: Potential health and ecological risks. Layton, D., Edson, R., Varela, M., Napier, B. (Eds.), Arctic Nuclear Waste Assessment Program (ANWAP), ONR. ARCTICMAR (2002). Radiological assessment of consequences from radioactive contamination of Arctic marine areas (ARCTICMAR). Final Report. Iosjpe, M., Strand, P. (Eds.), Contract IC 15-CT98-0209. NRPA, Østerås. CEC (1990). The radiological exposure of the population of the European community from radioactivity in North European marine waters. Project “Marina”. Commission of the European Communities. EUR 12483, Bruxelles. CEC (1994). The radiological exposure of the population of the European Community from radioactivity in the Mediterranean Sea. Project “MARINA-MED”. Commission of the European Communities. XI-094/93. Chartier, M., Durrieu de Madron, X., Poulin, M. (1989). In: Nyffeler, F., Simmons, W. (Eds.), In: Interim Oceanographic Description of the North-East Atlantic Site for the Disposal of Low-Level Radioactive Waste, vol. 3. NEA, pp. 291–307. EC (1995). Methodology for assessing the radiological consequences of routine releases of radionuclides to the environment. EUR 15760 EN. European Commission, Luxembourg.
Environmental modelling: Modified approach for compartmental models
475
EC (1997). Evolution of the radiological situation around the nuclear reactors with spent fuel which have been scuttled in the Kara Sea. EUR 17634 EN. European Commission, Luxembourg. EC (2000). The radiological exposure of the population of the European Community to radioactivity in the Barents Sea. MARINA-Balt Project. EC Radiation Protection Series 110. EUR 19200 EN. European Commission, Luxembourg. EC (2003). Update of the MARINA Project on the radiological exposure of the European Community from radioactivity in North European marine waters. EC Radiation Protection Series 132. European Commission, Luxembourg. ETOPO-5 (2002). “ETOPO-5” Five Minute Gridded World Elevation. U.S. National Geophysical Data Center (NGDC). Boulder, Colorado, USA. Available at http://www.grid.unep.ch/datasets/earth.html. FASSET (2003). Framework for assessment of environmental impact (FASSET). Deliverable 3. Dosimetric models and data for assessing radiation exposures to biota. EC 5th Framework Programme. Contract FIGE-CT-200000102. Available at http://www.fasset.org/ . IAEA (2003). Modelling of the radiological impact of radioactive waste dumping in the Arctic Seas. Report for the International Arctic Seas Assessment Project (IASAP). IAEA-TECDOC-1330. International Atomic Energy Agency, Vienna. IAEA (2004). Sediment distribution coefficients and concentration factors for biota in the marine environment. Technical Reports Series no. 422. International Atomic Energy Agency, Vienna. IBCAO (2001). International Bathymetric Chart of the Arctic Ocean (IBCAO). Available at http://www.ngdc.noaa.gov/mgg/bathymetry/arctic/arctic.html. Iosjpe, M. (2002). Ice module for “ARCTICMAR 2” model: Evaluation of some effects of the influence of ice transport of radionuclides for dispersion of radionuclides and dose assessment. In: Strand, P., Jølle, T., Sand. Å. (Eds.), Proceedings of the 5th International Conference on Environmental Radioactivity in the Arctic and Antarctic. St. Petersburg, Russia, 16–20 June 2002, pp. 159–162. Iosjpe, M., Brown, J., Strand, P. (2002). Modified approach for box modelling of radiological consequences from releases into marine environment. Journal of Environmental Radioactivity 60 (1/2), 91–103. Iosjpe, M., Brown, J., Strand, P. (2003). Modelling approach for environmental impact assessment from radioactive contamination of marine environment. In: International Conference on the Protection of the Environment from the Effects of Ionising Radiation. 6–10 October 2003, Stockholm, Sweden. IAEA-CN-109, pp. 212–215. Karcher, M.J., Harms, I.H. (2000). Estimation of water and ice fluxes in the Arctic for an improved box structure of the NRPA box model. In: Iosjpe, M. (Ed.), Transport and Fate of Contaminants in the Northern Seas. NRPA. MacKenzie, J., Nicholson, S. (1987). COLDOS – A computer code for the estimation of collective doses from radioactive releases to the sea. SRD R 389. UK Atomic Energy Authority. Mitchell, P.I., Condren, O.M., Leon Vintro, L., McMahon, C.A. (1999). Trends in plutonium, americium and radiocaesium accumulation and long-term bioavailability in the western Irish Sea mud basin. Journal of Environmental Radioactivity 44, 223–251. Nielsen, S. (1999). Resuspension of sediment to account for suspended load (not published). Nielsen, S.P., Hou, X. (2002). Environmental data, Marina II project. C6496/TR/004. Nielsen, S.P., Iosjpe, M., Strand, P. (1997). Collective doses to man from dumping of radioactive waste in the Arctic Seas. The Science of the Total Environment 202, 135–146. NRPA (2004). Impacts on man and the environment in northern areas from hypothetical accidents at “Mayak” PA, Urals, Russia. Joint Norwegian–Russian Expert Group for Investigation of Radioactive Contamination in the Northern Areas. Østerås, January 2004. Paluszkiewicz, T., Hibler, L.F., Becker, P., Mandych, A., Richmond, M.C., Thomas, S. (1997). An assessment of the flux of 90 Sr contamination through the Ob’ River and estuary to the Kara Sea. The Science of the Total Environment 202, 43–56. Pentreath, R.J. (1999). A system for radiological protection of the environment: Some initial thoughts and ideas. Journal of Radiological Protection 19 (2), 117–128. Pfirman, S.L., Eicken, H., Bauch, D., Weeks, W.F. (1995). The potential transport of pollutants by Arctic sea ice. The Science of the Total Environment 159 (N2-3), 129–156. Pfirman, S.L., Kogeler, J.W., Rigor, I. (1997). Potential for rapid transport of contaminants from the Kara Sea. The Science of the Total Environment 202, 111–122.
476
M. Iosjpe
Simmonds, J.R., Bexon, A.P., Lepicard, S., Jones, A.L., Harvey, M.P., Sihra, K., Nielsen, S.P. (2002). Radiological impact on EU member states of radioactivity in North European Waters. C6496/TR/004. Strand, P., Sickel, M., Aarkrog, A., Bewers, J.M., Tsaturov, Y., Magnusson, S. (1996). Radioactive contamination of the Arctic marine environment. In: Radionuclides in the Oceans: Inputs and Inventories. Les Edition de Physique, Les Ulis, pp. 96–119.
477
Assessment of 137Cs outspread in the Lithuanian part of the Baltic Sea L. Davuliene* , N. Tarasiuk, N. Spirkauskaite, G. Trinkunas, L. Valkunas Institute of Physics, LT-02300 Vilnius, Lithuania Abstract Results of analyses of 137 Cs in the Lithuanian part of the Baltic Sea and in the Curonian Lagoon in 1999–2001 are presented. According to these observations, the total 137 Cs activity concentration in Baltic Sea water has been the highest when compared with data for the world ocean. Sediment samples were collected in the Lithuanian part of the Baltic Sea for the first time after the Chernobyl accident. It is shown that the main 137 Cs accumulation zone in the area under consideration is located in the depth below 50-meters isobar. In the coastal water not influenced by the fresh water discharge from the Curonian Lagoon the 137 Cs was found mostly in the soluble form. The 137 Cs activity of the particulate phase was less than 10% of the total 137 Cs activity. Measurements revealed a linear relation between the 137 Cs concentration and salinity in the fresh water and seawater mixing zone. Validation studies of the circulation model for the Lithuanian marine were also carried out. The hydrodynamic model is able to assess spatial variations of 137 Cs concentrations with good accuracy (uncertainty of about 15%). Keywords: 137 Cs, Activity concentration, Fresh water, Seawater, Sediment, Hydrodynamic model, Curonian Lagoon, Baltic Sea
1. Introduction The Baltic Sea is one of the most contaminated seas with anthropogenic 137 Cs (NATO, 1998; Osvath et al., 2001; IAEA, 2005). The origin of this contamination is diverse. It is caused by global fallout as a result of nuclear weapons testing, by releases from nuclear reprocessing plants, by releases from nine nuclear power plants, which use seawater for reactor cooling in countries surrounding the sea, by the river run-off and the fallout right after the Chernobyl Nuclear Power Plant accident in 1986. As a result of the Chernobyl accident the long-lived radionuclide 137 Cs has had the most significant impact on the radioactive contamination of the sea. Measurements of 137 Cs activity concentration in the Baltic Sea near the Lithuanian coast revealed inhomogeneous distributions and remarkable fluctuations depending on meteorolog* Corresponding author. Address: Institute of Physics, Savanoriu 231, LT-02300 Vilnius, Lithuania; phone: (+370) 5 2661640; fax: (+370) 5 2602317; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08038-1
© 2006 Elsevier Ltd. All rights reserved.
478
L. Davuliene et al.
ical conditions (Styro et al., 2001). However, no correlation between 137 Cs and salinity was found in surface waters of the area. The Nemunas River is the third largest tributary of the Baltic Sea. Its drainage basin covers nearly the whole Lithuanian territory. The Nemunas River first flows into the Curonian Lagoon and then, through the Klaipeda Strait, it flows into the Baltic Sea. The influence of the water exchange between the Curonian Lagoon and the Baltic Sea on the 137 Cs activity distribution in the Lithuanian coastal waters has not been considered yet. Presence of the operating nuclear power reactor in Ignalina, Lithuania, which is similar to the Chernobyl-type RBMK reactor, makes the transport studies of radioactive substances via the Nemunas River in a case of severe accident of great importance. For this reason a circulation model of the North and Baltic Seas (known as BSHcmod) developed in the BSH (Bundesamt fur Seeschiffahrt und Hydrographie, Hamburg, Germany), has been applied to the south-eastern part of the Baltic Sea at the Lithuanian coast (Kleine, 1994; Dick et al., 2001; Davuliene et al., 2002a). This model with higher resolution has been extensively used in the Baltic Sea region, delivering daily forecast of water currents, water temperature and salinity, as well as water level and ice cover parameters, for combating the oil spill problems. It is the three-dimensional model of the Lithuanian marine waters covering both the Lithuanian part of the Baltic Sea and the Curonian Lagoon. This is also important for modelling the water exchange through the Klaipeda Strait. The applicability of this model to describe the distribution of 137 Cs, which is found in both soluble and particulate forms in the Lithuanian marine waters, is discussed in this paper. The model after successful validation can be used for the analysis of the heterogeneous distribution of 137 Cs in the Lithuanian waters, as well as for the assessment of the Nemunas River influence on the Lithuanian marine waters. In the years 1999–2000 and 2001–2002 two projects “Assessment of radionuclide migration in the Lithuanian part of the Baltic Sea environment” and “Assessment of radionuclide migration in the Lithuanian part of the Baltic Sea and Curonian Lagoon” were implemented. This paper will present the main outcomes of the projects concerning assessment of the 137 Cs outspread in the Lithuanian part of the Baltic Sea and in the Curonian Lagoon. Results of analyses of 137 Cs in the Lithuanian part of the Baltic Sea during 1999–2001 will also be presented. These data were used for validation of the hydrodynamic model of the Lithuanian marine waters using 137 Cs as a passive tracer.
2. Materials and methods 2.1. Study site The Curonian Lagoon represents a fresh-water shallow basin separated by the Curonian Spit from the Baltic Sea (Fig. 1). It is a transitory water basin collecting the run-off waters from the Nemunas (98%) and other minor rivers. Annually about 24 km3 of water from the Curonian Lagoon is transmitted into the sea. The interconnection between the Baltic Sea and the Curonian Lagoon occurs through the narrow Klaipeda Strait, which is artificially deepened to 14 m. Seawater is mostly observed in the northern part of the Curonian Lagoon. However, in some rare cases it penetrates 40 km into the lagoon, reaching its central part (Dubra and Dubra, 1998).
Assessment of 137 Cs outspread in the Lithuanian part of the Baltic Sea
479
Fig. 1. Location of sampling stations in the Baltic Sea and the Curonian Lagoon during investigations in 1999–2000 and 2001 (black rectangles).
The seawater mixes with fresh water of the Curonian Lagoon in the upper layer, less than 5–7 meters deep. Fresh water is mostly carried in the N-NW direction and can be traced as far as 7–10 nautical miles (nm) with salinity decreasing to 3h at the Klaipeda Strait (Dubra and Dubra, 1998). In the deeper waters (>60 m) of the study area, i.e. in the open sea, the salinity in the water column above thermocline is nearly uniform and reaches 7.5h. It should be admitted that the study area covering the sea is quite shallow, 30 m depth in average. Therefore, the formation of the sea-bottom profile and the structure of sedimentation material are highly affected by waves. The sedimentation material at the Lithuanian coast is mainly composed of sediments transported from the Curonian Lagoon, from the Sambian peninsula with along-shore current and of the corrosion products of submerged moraine deposit outcrops, accumulated mainly to the north from the Klaipeda Strait (Gavrilov et al., 1990). Coarse sand prevails near the coast. While the moraine loam, silt and alevrites prevail in deeper places and between boulders, in canyons and pits. The sea-bottom profile of shallow coastal water changes remarkably during storms. The transfer of resuspended sedimentation material offshore takes place, resulting in formation of radionuclide accumulation zones in bottom sediments in the deeper sea places. The Curonian Lagoon is a shallow water basin with the mean depth of 3.8 m. Therefore, sediments are periodically resuspended and mixed due to the wind-induced waves and water currents. The measurement data show variations in 137 Cs concentrations, characteristic for the whole Curonian Lagoon. They are site-dependent due to differences in sediment compo-
480
L. Davuliene et al.
sition and resuspension intensity (Tarasiuk et al., 1995). The 137 Cs concentration in the whole Curonian Lagoon depends on the seasonal and annual variations of the Nemunas River discharge and water exchange between the Curonian Lagoon and the Baltic Sea, as well as on meteorological conditions. Several 137 Cs accumulation zones in the Lithuanian marine waters can be indicated. 137 Cs accumulation zones of anthropogenic origin in the Baltic Sea at Lithuanian coast have developed near the deep-water pipe-lines. Here the coagulation process is important due to enforced interactions of radionuclides in seawater with the drainage water or the industrial water rich of organic substances (Tarasiuk and Spirkauskaite, 1994). 137 Cs accumulation zones of anthropogenic origin are also found in the Klaipeda Strait and the Curonian Lagoon (Tarasiuk et al., 1995; Lujaniene et al., 2005). According to the measurements, the Klaipeda Strait can also be indicated as a considerable 137 Cs accumulation zone. Due to continuous dredging the sediment distribution in the strait is changing. The dredged sediment material is brought to the dumping place in the vicinity of station No. 20a in the Baltic Sea (Fig. 2). The sedimentary material transport studies were done by Galkus and Jokšas (1997).
Fig. 2. Topography of the modelling area of the Baltic Sea and the measured 137 Cs massic activities in sediments.
Assessment of 137 Cs outspread in the Lithuanian part of the Baltic Sea
481
137 Cs
in seawater is found both in soluble and particulate forms. Suspended matter binds particle reactive radionuclides on their way from the Nemunas River mouth to the Curonian Lagoon, and in the coastal zone of the Baltic Sea. The first zone is the Curonian Lagoon itself which contains large amounts of organic (phytoplankton, algae, fish products, etc.) materials, that accumulate radiocaesium. The second zone is the fresh–seawater mixing area in the Baltic Sea coastal zone. 2.2. Experiment Investigations of the distribution of 137 Cs in the Lithuanian part of the Baltic Sea were accomplished in 1999–2001 (Tables 1–3). Water and bottom sediment samples were collected at the national and Helsinki Commission (HELCOM) monitoring stations in the Baltic Sea and the Curonian Lagoon. Measurements were carried out at the background stations in Preila and Juodkrant˙e during the campaigns on board of the R/V “V˙ejas” (at the sea) and of the cutter “Gintaras” (in the lagoon). In 2001, measurements were carried out at stations located on the 20 × 10 grid (Fig. 1). In Preila and Juodkrant˙e water samples were also collected on the beach. 3. Results 3.1. Seawater According to the field data obtained in May 1999, the 137 Cs activity concentration in surface water of the open sea was almost uniform. The measured 137 Cs concentration was 73– 79 Bq/m3 . In October 2001 it decreased to 60–67 Bq/m3 , i.e. about 20% (Fig. 3). The lowest 137 Cs concentrations in the sea, 39–53 Bq/m3 , were found in the samples collected at stations No. 4, 4a and 16, which are located in the close vicinity of the Klaipeda Strait. Thus mixing of the fresh and seawater is substantial. In May 1999 the highest 137 Cs concentrations (75–98 Bq/m3 ) were observed at the Preila background station. However, in October 2001 no increase was observed (Tables 1 and 2). The 137 Cs activity concentrations measured in the Baltic Sea were linearly correlated with the measured salinity of water (Fig. 4). In 1999, the increase in the salinity up to the mean value for the open sea surface water (about 7h) was followed by the increase in the 137 Cs activity concentration to 82 Bq/m3 , and in 2001 reaching up to 66 Bq/m3 . The correlation coefficients between the measured activity concentration and the salinity for the two periods were 0.94 and 0.90. During the field campaign in May 1999, the total 137 Cs activity concentration was measured. The activity concentration of 137 Cs bound with suspended matter varied from 0.9 to 9.7 Bq/m3 in the coastal zone of the Baltic Sea (Table 1). It comprised up to 10% of the total 137 Cs activity. The 137 Cs concentration in the open waters of the Baltic Sea was below the detection limit (Tables 1 and 2). Large variations in 137 Cs activity concentrations were observed in the northern and central parts of the Curonian Lagoon, including the Klaipeda Strait, during the storm in October 1999. The total 137 Cs activity concentration varied from 8 to 74 Bq/m3 with the highest values observed at the Klaipeda Strait, and then decreasing towards the central part of the Curonian
Date
Coordinates N
E
1999.05.17 1999.05.17
43 46a
56◦ 42 56◦ 03
19◦ 52 19◦ 24
1999.05.15
1b
56◦ 02
20◦ 50
1999.05.15
65
55◦ 53
20◦ 20
1999.05.15
64
55◦ 46
20◦ 54
1999.05.15
16
55◦ 45
21◦ 02
1999.05.15
4
55◦ 44
21◦ 03
1999.05.16
6b
55◦ 31
20◦ 34
1999.05.11 1999.05.13 1999.05.15 1999.05.17 1999.05.19 1999.05.21 1999.05.18 1999.10.27 2000.06.12 2000.06.12 2000.06.16 2000.06.16
Preila Preila Preila Preila Preila Preila 7 Preila Preila Preila Juodkrant˙e Juodkrant˙e
55◦ 21 55◦ 21 55◦ 21 55◦ 21 55◦ 21 55◦ 21 55◦ 19 55◦ 21 55◦ 21 55◦ 21 55◦ 32 55◦ 32
21◦ 01 21◦ 01 21◦ 01 21◦ 01 21◦ 01 21◦ 01 20◦ 57 21◦ 01 21◦ 01 21◦ 01 21◦ 06 21◦ 06
Activity concentration1 (Bq/m3 )
Sampling depth (m)
Volume (L)
Total
Dissolved
Particulate
Particulate loading1 (mg/L)
Surface Surface 73 Surface 25 Surface 47 Surface 30 Surface 14 Surface 12 Surface 65 Surface2 Surface2 Surface2 Surface2 Surface2 Surface2 Surface Surface2 Surface2 Surface2 Surface2 Surface2
50 50 25 50 30 50 30 50 30 50 30 50 30 100 40 50 47 52 50 50 50 50 50 50 50 50 50
79 ± 6 76 ± 7 77 ± 6 62 ± 6 69 ± 8 73 ± 2 67 ± 6 63 ± 6 72 ± 6 44 ± 4 59 ± 6 41 ± 3 52 ± 5 62 ± 3 73 ± 5 77 ± 6 98 ± 8 85 ± 5 82 ± 8 75 ± 7 84 ± 6 80 ± 6 89 ± 6 58 ± 3 59 ± 3 69 ± 3 63 ± 3
79 ± 6 76 ± 7 77 ± 6 62 ± 6 69 ± 8 73 ± 2 67 ± 6 63 ± 6 72 ± 6 44 ± 4 59 ± 6 41 ± 3 52 ± 5 62 ± 3 73 ± 5 71 ± 6 88 ± 8 84 ± 5 78 ± 8 73 ± 7 79 ± 6 80 ± 6 81 ± 5 56 ± 3 57 ± 3 67 ± 3 61 ± 3
b.d.l. b.d.l. b.d.l. b.d.l. b.d.l. b.d.l. b.d.l. b.d.l. b.d.l. b.d.l. b.d.l. b.d.l. b.d.l. 0.2 ± 0.1 b.d.l. 6.4 ± 0.7 9.7 ± 1.1 0.9 ± 0.3 3.6 ± 0.7 1.8 ± 0.4 4.9 ± 0.6 b.d.l. 8.0 ± 1.2 2.0 ± 0.2 2.0 ± 0.2 2.0 ± 0.2 2.0 ± 0.2
0.08 0.06 – 0.18 0.07 0.13 0.06 0.1 0.05 0.39 0.12 0.55 0.17 – – 1.61 2.17 3.21 1.14 1.12 3.89 0.1 22.6 13.8 15.2 – –
Notes: b.d.l. – below detection limits. 1 Radiochemical analyses and estimation of 137 Cs activity concentration by dr. Galina Lujaniene (Institute of Physics, Lithuania). 2 From the beach.
Massic activity in bottom sediments1 (Bq/kg)
– – 86 ± 5 – 12 ± 1 21 ± 2 64 ± 3
L. Davuliene et al.
Station, No.
482
Table 1 137 Cs activities of water and sediment samples collected in the Lithuanian part of the Baltic Sea (1999–2000)
Table 2 137 Cs activities of water and sediment samples collected in the Lithuanian part of the Baltic Sea in October 2001 Date
N
E
2001.10.16
4
55◦ 50
21◦ 00
2001.10.16
5
55◦ 50
20◦ 40
2001.10.16
6
55◦ 50
20◦ 20
2001.10.16 2001.10.16
5I 4a
55◦ 46 55◦ 44
20◦ 56 21◦ 03
2001.10.16
7
55◦ 40
21◦ 00
2001.10.16
8
55◦ 40
20◦ 40
2001.10.16
9
55◦ 40
20◦ 20
2001.10.16
20a
55◦ 39
20◦ 44
2001.10.16
10
55◦ 30
21◦ 00
2001.10.16
11
55◦ 29
20◦ 40
2001.10.16
12
55◦ 30
20◦ 20
2001.10.16 2001.10.16
12a (I) 12a(II)
55◦ 37 55◦ 37
20◦ 20 20◦ 24
Sampling depth (m)
Volume (L)
Activity concentration (solube fraction) (Bq/m3 )
Surface 18 Surface 37 Surface 20 45 29 Surface 15 Surface 26 Surface 20 45 Surface 20 40 55 Surface 20 42 Surface 25 Surface 59 Surface 20 47 74 74
55 47 53 51 52 53 49.5 – 53 53 53 55 52 30 56 56 22 30 56 53 30 51 52 53 54 30 53 26 49 – –
54 ± 2 60 ± 3 61 ± 3 60 ± 3 66 ± 3 62 ± 3 61 ± 3 – 39 ± 2 55 ± 2 63 ± 4 62 ± 3 62 ± 3 61 ± 3 59 ± 4 64 ± 3 67 ± 3 62 ± 3 59 ± 3 60 ± 4 62 ± 3 64 ± 3 60 ± 3 62 ± 3 60 ± 3 64 ± 3 61 ± 3 66 ± 3 65 ± 3 – –
Sediment sample thickness (cm)
Massic activity in bottom sediments (Bq/kg)
137 Cs load
(Bq/m2 )
0.8
19 ± 1
0.9
36 ± 1
0.9 0.8
90 ± 1 20 ± 1
859 253
0.9
14 ± 1
203
0.9
23 ± 1
314
0.9
31 ± 1
400
2.4
175 ± 2
1377
0.9
22 ± 1
333
0.9
265
57
– 1171 609
483
Coordinates
Assessment of 137 Cs outspread in the Lithuanian part of the Baltic Sea
Station, No.
4.43
4.0 ± 0.2
0.5
18 ± 1
– 2.6 1.3
– 171 ± 5 202 ± 5
–
484
L. Davuliene et al.
(a)
(b)
Fig. 3. Measured 137 Cs activity concentration and salinity along with the simulated salinity distribution (a) on 15–16 May 1999 and (b) on 16–17 October 2001. (Salinity in the Lagoon is less than 3.5h.)
Fig. 4. Relationship between the salinity and 137 Cs activity concentration.
Assessment of 137 Cs outspread in the Lithuanian part of the Baltic Sea
485
Lagoon (Table 3). The westerly and northwesterly wind of 12 m/s resulted in seawater flow into the Curonian Lagoon. It should be noted that a considerable variation of the 137 Cs activities in the Curonian Lagoon was found for water-soluble and particulate forms during this period, ranging from 7 to 67 Bq/m3 for water-soluble, and from 0.7 to 20 Bq/m3 for particulate caesium (Table 3). The percentage of the particulate 137 Cs in the Curonian Lagoon water samples varied from 4 to 27% of the total activity. During the field campaign in the Curonian Lagoon on 13–18 June 2000 the meteorological conditions were rather calm with dominant southeasterly wind of 5 m/s. The total 137 Cs activity concentration measured in the central part of the Curonian Lagoon ranged from 1.1 to 2.3 Bq/m3 . It should be noted that a considerable variation of the 137 Cs activities in the Curonian Lagoon was found for caesium in the particulate form. The percentage of the particulate 137 Cs reached 87% of the total activity (Table 3, Nida station). Variations of the particulate loading showed only a slight correlation with the particulate 137 Cs activity concentration measured in the Curonian Lagoon. During the stormy and calm weather periods, the particulate loading in the northern and central parts of the Curonian Lagoon was about 35 mg/L. This is considerably higher particulate loading as compared with the measured in the coastal zone (22.6 mg/L, Preila station), and in the open sea (less than 0.1 mg/L, stations No. 43, 46a, Tables 1 and 3). The particulate loading in the Lithuanian coastal zone varied from 0.06 to 0.55 mg/L because of the influence of the fresh water outflow from the Curonian Lagoon. It is noteworthy that the particulate loading in the Klaipeda Strait during the stormy period in October 1999 reached 102.2 mg/L in the near-bottom water layer. 3.2. Sediment Field data on the 137 Cs massic activity of sea sediments showed scarce zones of 137 Cs accumulation in the eastern part of the Baltic Sea. The sea bottom was mainly covered with different types of sand – from the coarse type to alevritic (densities vary from 1.73 to 1.37 g/cm3 ), or sand with silt admixtures (station No. 6, depth 45 m, density 0.91 g/cm3 ). In the latter case, the 137 Cs massic activity of surface sediments was rather high, up to 90 Bq/kg dry weight (d.w.) (Table 2). At the Lithuanian coast of the Baltic Sea the lowest 137 Cs activity (8–21 Bq/kg) was observed in sandy sediment sampled in shallow water right down to the 5-metre isobar. The fine fraction of suspended particles in sediments increases with increasing water depth (15–25 m), and therefore a gradual increase in the 137 Cs activity in sediments is expected. Generally, the 137 Cs activity in the bottom sediments varied from 4 to 90 Bq/kg d.w. (Tables 1 and 2). The main accumulation zone of 137 Cs in the area under investigation was found in the region of the old bed of the Nemunas River (stations No. 9, 12a, Table 2). Vertical 137 Cs profiles had maximum (175–202 Bq/kg d.w.) in surface layers of the sediment cores, and showed decreasing activity with the sediment depth. 137 Cs loads in these soft sediments varied in the range of 5120–6180 Bq/m2 . The 137 Cs massic activities in the bottom sediments taken in the Curonian Lagoon varied from 0.7 to 90 Bq/kg. The high 137 Cs activity (150 Bq/kg d.w.) was characteristic for black silt sediments found in the Klaipeda port (station No. 2a, at the Klaipeda town sewerage outlet (Fig. 1, Table 3). However, the 137 Cs activities in the sediments near the fairway (Klaipeda port, station No. 3) decreased to 22 Bq/kg d.w. A comparatively high 137 Cs activity in bottom sediments measured in the vicinity of the settlements in the central part of the Lagoon (in
486
Table 3 137 Cs activities of water and sediment samples collected in the Curonian Lagoon (1999–2000)∗ Date
Total
Dissolved
Particulate
21◦ 17 21◦ 17 21◦ 01 21◦ 01 21◦ 08 21◦ 08 21◦ 09 21◦ 09 21◦ 08 21◦ 06 21◦ 06 21◦ 01 21◦ 01 21◦ 04 21◦ 04 21◦ 09 21◦ 09
Surface 1.5 Surface 3.5 Surface 2.0 Surface 2.0 13.0 Surface 12.0 Surface 4.0 Surface 4.0 Surface 8.0
100 – 100 – 100 – 100 – – 100 58 100 – 150 – 100 –
8±1 – 23 ± 2 – 15 ± 2 – 74 ± 7 – – 52 ± 5 74 ± 7 2.3 ± 0.4 – 1.1 ± 0.2 – 53 ± 2 –
7±1 – 16 ± 1 – 11 ± 1 – 67 ± 6 – – 46 ± 4 54 ± 5 0.3 ± 0.1 – 0.5 ± 0.1 – 52 ± 2 –
0.7 ± 0.1 – 7±1 – 4±1 – 7±1 – – 6±1 20 ± 2 2.0 ± 0.3 – 0.6 ± 0.1
Coordinates N
Nemunas River mouth, 12 Nemunas River mouth, 12 Nida Nida 8 8 Klaipeda port, 3 Klaipeda port, 3 Klaipeda port, 2 Klaipeda port, 1 Klaipeda port, 1 Nida Nida Preila Preila Klaipeda port, 3 Klaipeda port, 3
55◦ 20 55◦ 20 55◦ 19 55◦ 19 55◦ 25 55◦ 25 55◦ 39 55◦ 39 55◦ 42 55◦ 44 55◦ 44 55◦ 19 55◦ 19 55◦ 21 55◦ 21 55◦ 44 55◦ 44
∗ Radiochemical analyses and estimation of 137 Cs activity concentration by Dr. Galina Lujaniene.
1.0 ± 0.2 –
Particulate loading (mg/L) 4.6 – 32.9 – 32.6 – 31.6 – – 48.1 102.0 37.9 – 36.7 – 9.00 –
Massic activity in bottom sediments (Bq/kg) – 0.7 ± 0.1 – 58 ± 6 – 47 ± 3 – 22 ± 1 67 ± 6 – 30 ± 3 – 90 ± 6 – 71 ± 7 – 150 ± 11
L. Davuliene et al.
1999.10.27 1999.10.27 1999.10.27 1999.10.27 1999.10.27 1999.10.27 1999.10.27 1999.10.27 1999.10.27 1999.10.27 1999.10.27 2000.06.13 2000.06.13 2000.06.14 2000.06.14 2000.06.18 2000.06.18
Volume (L)
Activity concentration (Bq/m3 )
E
Sampling depth (m)
Sampling station, No.
Assessment of 137 Cs outspread in the Lithuanian part of the Baltic Sea
487
Preila – up to 71 Bq/kg d.w.) appeared to be related to the local zones of the anthropogenic organic contamination. The thick layer of bottom sediments of black silt was also taken in the vicinity of Nida. The 137 Cs activity in sediments varied from 58 to 90 Bq/kg d.w. (Table 3). 4. Model 4.1. Model implementation A detailed description of the BSH model is given by Kleine (1994) and Dick et al. (2001), and will not be presented here. The adapted circulation model for the Baltic Sea (Davuliene et al., 2002a) was used for calculating currents in shallow surface waters influenced by bottom topography, wind and river discharge. The model enables to gain information about the water level, water temperature and salinity, as well as about dispersion of passive substances in a real time. The area selected for modelling (with the upper left corner at 56◦ 20 45 N, 19◦ 55 25 E) covered the Lithuanian part of the Baltic Sea including the Curonian Lagoon, part of Latvian and Kaliningrad (Russia) coasts (Fig. 2). The 1 nm grid (or 1 × 1 40 ) was chosen and the water column was divided into five layers. Only one river input (the Nemunas River splitting into two branches called Atmata and Skirvyt˙e) was taken into account. For the western and northern open boundaries the BSHcmod simulation data were used. The applicability of the circulation model to simulate the 137 Cs activity concentration distribution as a passive tracer in the Baltic Sea near the Lithuanian coast should also be examined before applying the model to the case under consideration. 4.2. Validation of the hydrodynamic model As already discussed in the experimental part, the experimental data show a linear relation between the salinity and the 137 Cs activity concentration (Fig. 4). The salinity and the passive tracer distribution for the periods of expeditions were simulated in order to assess the applicability of the circulation model to simulate the 137 Cs activity concentration distribution in the Lithuanian marine waters (Fig. 3). The standard deviation between simulated and measured salinity for the two periods in 1999 and 2001 are 0.34 and 0.19h, respectively. This comprises about 10% of the salinity variation (about 3.5h) in the surface water of the area. The 137 Cs activity concentration can be calculated from the measured or simulated salinity data using the linear relationship. Correlation between the simulated distribution of the tracer and measured 137 Cs activity concentration during the periods of field campaigns was 0.93 and 0.96, with standard deviations of 6 and 2 Bq/m3 , respectively (Fig. 5). This indicates the applicability of the hydrodynamic model of Lithuanian marine waters to simulate the 137 Cs activity concentration within the error gap of about 15%. 4.3. Modelling results The mean value of the 137 Cs activity concentration in the Nemunas River mouth water is less than 2 Bq/m3 (Tarasiuk et al., 1999), therefore this is a diminishing factor for the 137 Cs
488
L. Davuliene et al.
(a)
(b)
Fig. 5. Measured and simulated 137 Cs activity concentration (AC) (a) on 15–16 May 1999 and (b) on 16–17 October 2001.
activity concentration in the southeastern part of the Baltic Sea coastal zone. The observed fluctuation of the 137 Cs activity concentration could be clearly attributed to the distribution of fresh water along the Lithuanian coast (Fig. 3). The fresh water flow in the Lithuanian coastal waters can be well followed due to the salinity decrease. In the reduced salinity zone the relation between the 137 Cs activity concentration and the salinity is most appropriate. The data sets of the measured 137 Cs activity concentration during the two field campaigns show different data scattering. As follows from the simulations of these two periods, the variation in the 137 Cs activity concentration could be explained by the fresh water distribution in the coastal zone. Several days before the field campaign on 11–14 May 1999 the east-southeast wind was dominating. The fresh water from the Curonian Lagoon was spreading along the Lithuanian coast in the northern direction from the Klaipeda Strait. However, during the field campaign the wind changed to the north-northwest. Because of that the fresh water spreading from the Curonian Lagoon was moving in the southwest direction from the Klaipeda Strait. This caused a decrease in the salinity followed by a considerable decrease in the 137 Cs activity concentration in the large area of the Lithuanian coast of the Baltic Sea during the period of 15–16 May 1999 (Fig. 3(a)). However, this was not the case during the field campaign on 16–17 October 2001. Due to the strong (up to 15 m/s) westerly wind dominating from 11 October 2001, the affluent evolved at the Lithuanian coast and the water outflow through the Klaipeda Strait was nearly stemmed. Although during the field campaign the wind direction changed (south-easterly) and the water outflow through the Klaipeda Strait was resumed, the measured salinity as well as variation of the 137 Cs activity concentration at the Lithuanian coast were rather weak (Fig. 3(b)). The simulations of the salinity distribution for the period of 1999–2001 have shown that the fresh water flow from the Curonian Lagoon modifies the salinity in the Lithuanian coast of the Baltic Sea in the area of 15 km from the seashore (Davuliene et al., 2002b). Because of dominating northerly and south-easterly winds at the Lithuanian coast, the obtained averaged simulated salinity distribution has the most reduced salinity values along the seashore to the
Assessment of 137 Cs outspread in the Lithuanian part of the Baltic Sea
489
north of the Klaipeda Strait. The simulated variability of the salinity was found negligible in the area approximately 30 km off the coast. In the open sea the measured salinity and the 137 Cs activity concentration can be considered as constant. The salinity in this area has a constant value down to the halocline. The measured 137 Cs activity concentration is fluctuating around its mean value in the range of uncertainties. Therefore, the determined values could be used for the model calibration as the background values, typical for the south-eastern Baltic Sea.
5. Conclusions The measured total 137 Cs activity concentration in the open Baltic Sea is determined by the self-cleaning process, which still persists. At present the 137 Cs activity concentration in the water of the southeastern Baltic Sea reaches 60–70 Bq/m3 , that is together with Irish Sea waters the highest concentration found in the world ocean (IAEA, 2005). Although in previous studies no correlation between the 137 Cs activity concentration and the salinity was found in the Lithuanian marine waters (Styro et al., 2001), our measurements revealed the linear relationship in the mixing zone of fresh water and seawater. The coefficient of proportionality between the 137 Cs concentration and the salinity depends on the average 137 Cs concentration in the Baltic Sea. The applied hydrodynamic model enabled to simulate the salinity distribution in Lithuanian waters and to evaluate the extent of the mixing zone. Measurements carried out in the Lithuanian part of the Baltic Sea and in the Curonian Lagoon in different seasons indicated that the 137 Cs concentration in the water depended on the hydrological situation, as well as on the water exchange between the sea and the lagoon. The observed fluctuations in the 137 Cs concentration in the Lithuanian part of the Baltic Sea could be clearly attributed to the distribution of the fresh water along the Lithuanian coast (reliability uncertainty of about 15%). The measured 137 Cs activity concentration of the surface water in the central part of the Baltic Sea (stations No. 46a and 43) in spring 1999 was on average by 5 Bq/m3 higher than at the Lithuanian coast (the offshore station No. 65). The caesium was found mostly in the soluble form in the open sea, and in coastal waters far from the fresh water and seawater mixing zone. The particulate 137 Cs activity concentration in the coastal water was less than 10% of the total 137 Cs activity concentration. The sediment samples were taken for the first time in open waters of the Lithuanian part of the Baltic Sea after the Chernobyl accident. According to the experimental data, the main 137 Cs accumulation zone in the area under consideration is located in the region of the old bed of the Nemunas River (below 50-meters isobar), which is in the reduced hydrodynamic activity zone below the thermocline (Fig. 2). The 137 Cs massic activity here is about 10 times higher than at the coast. The 137 Cs activity of particulate matter in the northern and central part of the Curonian Lagoon might reach during the period of calm weather 90% of the total activity. However, as the total 137 Cs activity concentration in the fresh water of the Curonian Lagoon is less than 5% of the total 137 Cs activity in the open sea, this fact is of minor importance when modelling the 137 Cs distribution in the Lithuanian marine waters. During the stormy period with the dominating northwesterly and westerly winds, the salt water intrusion into the Curonian Lagoon can occur. This is more often observed during the
490
L. Davuliene et al.
autumn season (Dubra and Dubra, 1998). The influence of this process on the salinity and on the 137 Cs distribution in the Curonian Lagoon has not been studied in detail so far. During the study period in October 1999 the 137 Cs activity concentration in the northern part of the Lagoon was measured during the intrusion of salt water. During this event, the total 137 Cs concentration in the northern part of the Curonian Lagoon increased remarkably and varied in a wide range. The largest values characteristic for the open sea were measured in the Klaipeda Strait. The particulate 137 Cs activity measured in the surface water of the Curonian Lagoon reached about 6 Bq/m3 and was nearly 5 times higher compared to the total 137 Cs activity observed during the period of calm weather. The resuspension may become of crucial importance for the total 137 Cs concentration during period of stormy weather. However, the contribution of all 137 Cs accumulation zones located in Lithuanian marine waters to the total 137 Cs concentration (depending on the meteorological situation) has not been studied yet. This would be the considerable task for the application of the sediment transport model. The high correlation found between the simulated and measured 137 Cs concentrations in the Lithuanian part of the Baltic Sea shows that the hydrodynamic model is able to assess spatial variations of 137 Cs concentration with good accuracy (uncertainty of about 15%). The variability in salinity was found negligible in the area approximately 30 km off the coast. Therefore, the 137 Cs concentrations measured at this distance could be used for the model calibration as the background value, typical for the south-eastern part of the Baltic Sea.
Acknowledgements This research was partly supported by the IAEA projects “Assessment of radionuclide migration in the Lithuanian part of the Baltic Sea environment”, LIT/2/002 (1999–2000), “Assessment of radionuclide migration in the Lithuanian part of the Baltic Sea and Curonian Lagoon”, LIT/7/002 (2001–2002), and by the Lithuanian State Science and Studies Foundation.
References Davuliene, L., Dailidiene, I., Dick, S., Trinkunas, G., Valkunas, L. (2002a). Validation of circulation model for Lithuanian coastal waters. Environmental and Chemical Physics 24 (4), 226–231. Davuliene, L., Trinkunas, G., Valkunas, L. (2002b). Analysis of salinity variation in Lithuanian marine waters. Lithuanian Journal of Physics 42 (4), 309–314. Dick, S., Kleine, E., Müller-Navarra, S., Klein, H., Komo, H. (2001). The operational circulation model of BSH (BSHcmod). Berichte des BSH 29, 48. Dubra, J., Dubra, V. (1998). Calculation of the water changes in the Lithuanian nearshore of the Baltic Sea. Centre of Marine Research, Scientific and Information Publication, Klaipeda, pp. 49–56 (in Lithuanian). Galkus, A., Jokšas, K. (1997). Sedimentary Material in the Transitional Aquasystem. Institute of Geography, Vilnius, 198 pp. Gavrilov, V.M., Gritchenko, Z.G., Ivanova, L.M., Orlova, T.E., Tiskov, V.P., Tiskova, N.A. (1990). Strontium-90, caesium-134 and caesium-137 in the water basins of the Baltic States at Soviet Union (years 1986–1988). Radiochemistry 3, 171–179 (in Russian). IAEA (2005). Worldwide Marine Radioactivity Studies (WOMARS). IAEA TECDOC-1429, Vienna. Kleine, E. (1994). Das operationelle Modell des BSH für Nordsee und Ostsee. Konzeption und Übersicht. Technical Report. Bundesamt für Seeschiffahrt und Hydrographie, Hamburg.
Assessment of 137 Cs outspread in the Lithuanian part of the Baltic Sea
491
Lujaniene, G., Vilimaite-Silobritiene, B., Joksas, K. (2005). Accumulation of 137 Cs in bottom sediments of the Curonian Lagoon. Nukleonika 50 (1), 23–29. NATO (1998). Cross-border environmental problems emanating from defence-related installations and activities. Phase II: 1995–1998. Final report, vol. 2: Radioactive contamination of rivers and transport through rivers, deltas and estuaries to the sea. Report No. 225. Vienna, 105 pp. Osvath, I., Samiel, M., Valkunas, L., Zlatnansky, J. (2001). Dynamic waters of the Baltic Sea. IAEA Bulletin 43 (2), 9–15. Styro, D., Bumelien˙e, Ž., Lukinskien˙e, M., Mork¯unien˙e, R. (2001). 137 Cs and 90 Sr behavioural regularities in the southeastern part of the Baltic Sea. Journal of Environmental Radioactivity 53, 27–39. Tarasiuk, N., Spirkauskaite, N. (1994). Radioactive contamination of the coastal zone of the Baltic Sea of Lithuania. Atmospheric Physics 16, 85–89. Tarasiuk, N., Spirkauskaite, N., Stelingis, K., Lujaniene, G., Lujanas, V. (1995). Evaluation of the 137 Cs residence time in the Curonian gulf after the Chernobyl fallout. Atmospheric Physics 2, 22–33. Tarasiuk, N., Spirkauskaite, N., Lujaniene, G. (1999). The formation of 137 Cs concentrations in the Nemunas River accumulation zone. Environmental and Chemical Physics 21 (2), 27–32.
This page intentionally left blank
10. Radiometrics techniques
This page intentionally left blank
495
Low-level germanium gamma-ray spectrometry at the µBq/kg level and future developments towards higher sensitivity G. Heussera,* , M. Laubensteinb , H. Nedera,+ a Max-Planck-Institut für Kernphysik (MPI-K), Heidelberg, Germany b Laboratori Nazionali del Gran Sasso (LNGS), Assergi (AQ), Italy
Abstract GeMPI, a highly sensitive germanium gamma-spectrometer operated at the Gran Sasso underground laboratory is described. Its complete cryostat system and shield was made from highly selected low activity materials. Radon suppression, also during sample insertion is achieved by an air-lock system combined with an airtight steel casing around the shield, which is pressurised with nitrogen gas. The achieved background level in combination with a large sample capacity of up to 15 l around the 2.2-kg Ge crystal allows measuring concentrations of the gamma-active chain members down to 10−12 g/g of U/Th equivalent (12.3 µBq/kg 238 U, 4.06 µBq/kg 232 Th) and 10−9 g/g of K (31 µBq/kg 40 K). This is demonstrated with some key measurements of Cu and Pb shielding materials. Cosmogenic production rates for 8 radioisotopes have been determined in exposed Cu to range from 50 µBq/kg for 46 Sc to 2.1 mBq/kg for 60 Co. A possible further background reduction in Ge-spectrometry by operating naked crystals in liquid nitrogen is discussed. Keywords: Natural radionuclides, Cosmogenic radionuclides, Low-level Ge-spectrometry
1. Introduction A Ge-spectrometer (called GeMPI) designed for measurements at the µBq/kg level sensitivity level is operated at the underground laboratory LNGS (Laboratori Nazionali del Gran Sasso) near L’Aquila/Italy since 1997. First preliminary results of background and sample measurements have been published in (Neder et al., 2000). Primarily the spectrometer is used for material screening (radio-purity) measurements in connection with the solar neutrino experiment BOREXINO (Alimonti et al., 2001; Arpesella et al., 2002) and more recently for the * Corresponding author. Address: MPI-K, POB 103980, D-69029, Heidelberg, Germany; phone: (+49) (0) 6221 516 404; fax: (+49) (0) 6221 516 607; e-mail:
[email protected] + Present address: University of Heidelberg, Germany.
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08039-3
© 2006 Elsevier Ltd. All rights reserved.
496
G. Heusser et al.
double beta decay experiment GERDA (Abt et al., 2004). Its high sensitivity makes the spectrometer also very well suited for environmental studies, as e.g. in measurements of activation products from accidental neutron exposures (Gasparo et al., 2004). Special attention in the construction of the spectrometer was devoted to achieve low background rates at the lines of the primordial decay chains 238 U and 232 Th and of 40 K, the most prominent contaminants. The BOREXINO experiment aims to measure low-energy solar neutrinos in real time by elastic neutrino–electron scattering. The mono-energetic neutrinos from 7 Be at 862 keV are of most interest. The experiment, located also in LNGS will use 300 t of liquid scintillator to detect the scattered electrons. Since they are practically indistinguishable from other ionising events produced by natural radioactivity at the same energy, extremely high radio-purity standards must be met to detect the very low rate of a few tens of events per day in the 100 t fiducial volume. For example, the central liquid scintillator needs to contain less than about 10 nBq/kg of the U/Th decay-chain activities and about 0.3 nBq/kg of 40 K. The radio-purity requirements of the neighbouring construction materials are relaxed by 3–6 orders of magnitude, depending on their location within the detector. This corresponds to the sensitivity range of the spectrometer under discussion and also partly of other low-level Ge-spectrometers (Arpesella et al., 2002). Since the tolerable contamination level at these locations is defined by γ -activity, Ge-spectroscopy is the ideal screening tool. In case of the natural decay chains interest is focused on the most gamma active progenies 214 Pb/214 Bi (238 U-chain) and 228 Ac, 212 Pb/212 Bi, 208 Tl (232 Th-chain). The high resolution spectroscopy with its spectral information gives Ge-spectroscopy an almost visualising ability and is thus superior to other screening methods which measure only atomic concentrations as e.g. in mass spectrometry, atomic adsorption, X-ray excitation or neutron activation. Concerning the U/Th decay chains, these techniques are only applicable to the parent nuclides, which are rarely in equilibrium with their gamma-active progenies in chemically processed construction materials. Moreover Ge-spectroscopy is able, within certain limits, to observe deviations from secular equilibrium in the U/Th series. This is important, e.g. in the case of the 228 Ra/228 Th sub-chain, where the time for noticeable changes in activity ratios can occur within the lifetime of an experiment. In addition the measurement can mostly be performed in a non-destructive way without labour intensive sample pre-treatment. The 76 Ge double beta decay experiment GERDA (Abt et al., 2004) will use almost bare Ge-crystals immersed in liquid nitrogen or liquid argon. Germanium is enriched in the ββ active 76 and thus the source and the detector are the same device. The highly radio-pure cryogenic liquid serves as the innermost shield against external radiation and as coolant (first proposed in Heusser (1995)). Compared to conventional detector mountings with several cladding layers, which carry surface and bulk contamination, very little material is needed to hold and contact the crystals. Together with other suppression methods a background reduction of two to three orders of magnitude seems possible in comparison to the presently most sensitive 76 Ge double beta decay experiment Heidelberg–Moscow (HDM) (Guenther et al., 1997). However, even with material savings by almost four orders of magnitude in mass, radio-purity in the µBq/kg range will be required to reach the envisioned background. Consequently, material screening at least at the GeMPI level will be needed.
Low-level Ge gamma-ray spectrometry at the µBq/kg level
497
2. The low-level Ge-spectrometer GeMPI 2.1. Design criteria Experience collected with earlier Ge-spectrometer systems (Heusser, 1994) and with the detectors of the HDM experiment (Guenther et al., 1997) resulted in the following design criteria for the construction of GeMPI: (A) Ge-crystal minimise cosmic ray exposure by fast processing after zone refining and by surface transportation and by shielding under several meter of water equivalent whenever feasible. Crystal size optimised for high counting efficiency in Marinelli-type geometry. (B) Cryostat system next to the crystal made only from screened materials with a radiopurity level below the mBq/kg level, where possible use of NOSV grade copper1 , stored underground (minimum 15 m w.e.) shortly after electrolysis and between machining steps electron beam welding, no soldering, crimping of contacts, metal sealing for joints electro-polishing of metals respectively acid cleaning of all parts under clean room conditions assembly of detector under clean room conditions. (C) Shield large sample volume around detector for high sensitivity of specific activity innermost layer copper (same as above), acid cleaned, then lead with 6 Bq/kg 210 Pb (acid cleaned) and subsequent Pb layers with increasing 210 Pb-concentration fully airtight steel box around shield, pressurised with nitrogen gas for protection against Rn air lock system for sample insertion to keep Rn and particulate contamination out storage space in box to allow the decay of plated out Rn progeny of Rn itself (dissolved or adsorbed) before the sample is placed into the measuring position insertion possibility for calibration sources. 2.2. The cryostat system The concept of the cryostat system was elaborated in close collaboration with Canberra Semiconductors N.V., Olen, Belgium. Figure 1 depicts a cross section of the found solution. In the U-type mounting the isolating vacuum is common to the dewar (left to Fig. 1, not shown) and the cryostat. Normally it is maintained by molecular sieve as pumping adsorbent embedded in aluminised Mylar foil next to the cool inner wall of the dewar. In our case the Ra-rich molecular sieve was replaced by low 226 Ra charcoal made of coconut shells. The reason for this precaution is, that in temperature cycles, the crystal and its holder stays longer cool during the warming up phase than the rest of the system, so that 222 Rn emanated from the warmed up adsorbent might freeze out onto the cool parts and thus concentrates the final decay product 210 Pb close to or on the crystal. On the left side of Fig. 1 the crystal is mounted in an insulation can made from VESPEL® surrounded by a holder made from Cu. Also the part which fixes the holder to the warm part of the cryostat, normally called spider, as well as the insulation around the signal contact is made from VESPEL® . The Field Effect Transistor (FET) part of the electronic (provided by Canberra) is mounted on a thin Cu plate (connected to the cool-finger) underneath a piece of low-activity lead (<0.2 mBq/kg 210 Pb2 ). The lead shields the residual activity of the FET unit 1 Copper of 99.9975% purity, produced by Norddeutsche Affinerie, Hamburg, Germany. 2 Analysed by M. Wojcik, Institute of Physics, Jagellonian University, Cracow, Poland.
498
G. Heusser et al.
Fig. 1. Cross-section of the cryostat of GeMPI.
against the crystal, although the FET, the hardpaper–baseplate and all other parts of this unit have been checked in larger quantities to assure their radioactivity being below 10 mBq/kg. Only here a special solder made from low activity lead was used in combination with a special selected flux liquid – much lower in potassium than the normally used colophon, whereas all other wire-connections are crimped with Cu sleeves. The screws indicated on the right side are made out of steel originating from a First World War battle ship. A measurement with GeMPI of that steel showed a 226 Ra and 228 Th contamination of 150 µBq/kg and 460 µBq/kg (see also Section 3.3). The Cu cover cap around the holder is sealed to the bottom part by a lead O-ring made from the same low activity lead as above. All parts of the cryostat (shown in Fig. 1) have been machined at the workshop of the MPI-K taking great care to avoid cross contamination. Electron beam welding was applied to form the lower left part of the cryostat from single pieces and to connect it to the tube around the cooling finger. This tube, the cooling finger as well as all other Cu parts, around the crystal (except the braid connecting the cooling finger and the holder) have been fabricated from a special order of NOSV grade copper (99.9975% purity, from Norddeutsche Affinerie; Hamburg, Germany). Within a few days after the electrolysis this copper was processed (melted, cast and warm formed) and placed underground, first in a former ice cave of a brewery at Heidelberg and later in the low-level laboratory of the MPI-K. The latter is covered by about 15 m w.e. concrete and rock soil which is estimated to reduce cosmogenic activation via spallation reactions by roughly a factor 20. Also during the machining phase the Cu was placed underground during longer breaks. The hot forming by forging or rolling converts the coarse crystalline structure of the cast Cu in a much finer, less porous structure with a density of up to 8.92 g/cm3 (about 8.7 g/cm3 before). Thus the material becomes better suited for delicate mechanical machining, for high vacuum applications and also for acid cleaning. Acid that enters into pores of the cast material is difficult to remove. If in addition the acid solution is not radio-pure, as e.g. in electropolishing solutions, contamination may result from residues remaining in the pores. The finished parts of the cryostat have been cleaned (acid/deionised H2 O/alcohol) at the MPI-K in a clean room hood (class 100) and the components from copper were electropolished with subsequent ultrasonic cleaning, passivation and clean room packing at a com-
Low-level Ge gamma-ray spectrometry at the µBq/kg level
499
pany3 . The assembly of the cryostat was performed at Canberra Olen under clean room conditions. The p-type high purity germanium crystal provided by Canberra was shipped before from the United States to Belgium via boat in order to minimise cosmogenic activation. Only after the conversion to a diode (2.2 kg active volume, 77.5 × 88.5 mm2 , relative efficiency 102%) and successful testing in a test cryostat at Olen, the direct cladding parts, insulation can and holder, were machined at Heidelberg to match the size. During the production phase of the low-level cryostat the crystal was placed underground in HADES4 , close to Olen. Immediately after the transfer of that crystal into its final cryostat system and the positive performance test, the detector was transported to the low-level laboratory at Heidelberg. Here first background tests were performed in a provisional shield made of 15 cm lead and one large cosmic veto detector (multiwire proportional chamber) placed on top. The measured background spectrum is shown below (Section 3.1) in comparison with the starting background obtained at the much deeper Gran Sasso laboratory. Calibration measurements of the spectrometer have been carried out with various sources under different geometry and the detector was scanned horizontally and axially with collimated sources to create a data base for Monte Carlo simulations (Prokosch, 1996; Neder, 1998) based on the code GEANT 3.21 (CERN Program Library). 2.3. The shield and radon protection system In Fig. 2 different views of the GeMPI spectrometer are shown. The shield consists (from inside) of 5 cm NOSV Cu, 5 cm of Pb with 6 Bq/kg 210 Pb, 10 cm Pb with about 20 Bq/kg 210 Pb and 5 cm Pb with about 130 Bq/kg 210 Pb. The shield rests on a 5 cm thick slap of borated (10% by weight) polyethylene, to keep the option for completing a neutron shield by later addition of plates on the other sides. The top shield layers of Cu and Pb, with the same sequence, are vertically split into two asymmetric halves which can be moved sideways by the aid of slide bearings along the frame, indicated in the middle picture, so that the full sample chamber (effective space – 250 × 250 × 238 mm3 minus 94 × 106 mm2 detector) is accessible from above. A vacuum tight steel casing encloses the shield as part of the radon protection system. The other part is the (trapezoidal) sample storage/handling compartment mounted on top with the air lock box connected to it. This box has two doors. Samples to be measured are first placed inside the air lock box with the inner door still closed. Next the volume between the two doors is flushed with nitrogen gas, then the inner door is opened by the aid of gloves (round openings) from inside the sample storage/handling compartment and the sample is placed on a movable table. There is also such a table in the air lock box which can be moved through the opened inner door into the glove box compartment to transfer heavy loads. Here a small pulley can lift the sample to place it either on the inner table or to lower it into the sample chamber. Normally the sample rests in the low Rn nitrogen gas of the glove box until the plated out 222 Rn and 220 Rn progenies or attached/dissolved 222 Rn have decayed, before the two top Cu/Pb lids are slid aside and the sample is replaced against the former one. Experience 3 GCE-Druva, Eppelheim, Germany. 4 Underground Research Facility (URF) operated by SCK/CEN Nuclear Research centre at Mol, Belgium.
500
G. Heusser et al.
Fig. 2. Cross-sections and outside views of the GeMPI spectrometer (see text).
teaches us that this is important to assure the peak count rates of 214 Pb/214 Bi respectively and of 212 Pb/208 Tl with their high γ -abundance being interpreted as 226 Ra respectively 228 Th concentrations. Moreover, the precious counting time is thus more effectively used. The whole Rn protection system is continuously flushed at slight over-pressure with nitrogen gas from standard 50 l gas cylinders. It is directed into the sample chamber below the detector through a thin Teflon tube, so that Rn is prevented from diffusing in by the outward flow. The system with all its joints is sealed by the aid of O-ring material made from polyurethane, which was tested to have a low permeability for Rn (Rau and Heusser, 2000) as well as a low Rn emanation rate (Laubenstein et al., 2004). For the gloves butyl rubber was chosen for the same reasons. However, due to their large area and small thickness additional measures had been taken to keep Rn permeation into the sample area low. When not in use the gloves are pulled back into their fixing tube and both ends of that tube are closed by lids with O-ring sealing. A valve mounted on the outer lid avoids overpressure built up during the closing procedure. The two lids strongly reduce the direct airborne Rn transfer by permeation through the glove material, but also help to minimise the accumulation of Rn between the glove and the inner lid, which at the moment of the next opening is released into the interior. There are two possibilities to insert calibration sources. The main one uses a 1 mm diameter encapsulated 152 Eu source welded to a flexible wire, with which it can be pushed through a thin Teflon tube near to the detector bottom. A motor driven spinning wheel outside the shield carries the source between the calibration position and the outside-shielded position. For Rn protection the end of the Teflon tube is closed and the box of the spinning machine can be flushed with N2 . The other (Fig. 2 – source insertion system) uses a combination of a strong well-type magnet and a thin counter rod magnet fixed to a flexible wire carrying a 55 Fe/133 Ba source. They are separated by a Cu/Teflon tube, which forms an extension of the cryostat tube (Fig. 1 – source pipe). The Teflon tube guides the source along the cool finger through the signal contact into the well of the crystal. Here the almost wall-less p-contact surface allows extending the calibration energy range into the X-ray regime. Unfortunately the latter method turned out to be unpractical since the stiffness of the Teflon tube turned out to be much higher at low temperature than at room temperature during the final function test before the cryostat
Low-level Ge gamma-ray spectrometry at the µBq/kg level
501
assembly was completed. The risk that the source gets blocked in the bend part of the tube under the crystal holder does not allow pushing it across this point.
3. Measurements 3.1. Background measurements After a few months of testing at the low-level laboratory at Heidelberg, the shield was disassembled and GeMPI (Fig. 2) was installed (Neder, 1998) at the low background counting facility of the Gran Sasso underground laboratory LNGS. At this occasion all parts have been cleaned once more. Already at Heidelberg the heavy Cu plates of the inner chamber have been washed with diluted nitric acid, rinsed with distilled water and packed for shipping. The lead bricks were cleaned directly before installation, those for the inner layers with a mixture of acetic acid/H2 O2 and deionised water, the outer ones with alcohol. Clean room clothing was obligatory during the set-up phase of GeMPI. The about 3800 m w.e. rock shielding above the laboratory reduces the muons flux by 6 orders of magnitude compared to sea level, so that the muonic contribution to the background should be negligible compared to Heidelberg, where only about a factor of 3 is effective by the 15 m w.e. overburden. The integral background rate was immediately lower by more than two orders of magnitude compared to Heidelberg and continuously decreased with time due to the decay of cosmogenic radionuclides, mainly 57 Co, 58 Co in Cu and Ge, 65 Zn in Ge. After about 2 years it slowly levelled off to about 0.04 cpm (100–2730 keV), roughly 2–3 times lower than at start. This factor would have been much larger without all the shielding measures before the cryostat assembly and the preconditioning at Heidelberg. In Fig. 3 spectra are compared which have been measured at Heidelberg unshielded and with the provisional shielded of about 15 cm Pb and one veto detector on top and at Gran Sasso with the full shield. The latter background spectrum represents only a rather weak statistic of about 60 days counting time, collected between sample measurements during the first 1.5 years of operation at LNGS. A reduction of about 5 orders of magnitude compared to the unshielded mode at Heidelberg seems to be almost independent on energy, at least up to the Compton edge of the 2.615 MeV 208 Tl line. The shielded measurements at Heidelberg and at Gran Sasso differ by more than two orders of magnitude. A short unshielded measurement at Gran Sasso yielded a comparable continuous level as at Heidelberg, only lower by factors 2–5 depending on energy. This is mainly due to a lower concentration of primordial radionuclides at LNGS and less to the difference in muon flux. The contribution of muons to the total count rate of 8100 cpm (100–2690 keV) at Heidelberg is below 1%. On the bottom of Fig. 3 the GeMPI background (lowest spectrum, this time with 101 days statistic, collected after about 3 years at LNGS) is compared with the background of (shielded) low-level detectors from other European underground laboratories (Laubenstein et al., 2004) co-operating in CELLAR5 . The overburden of this laboratories varies from almost zero (ARC-S and upper spectrum), 110 m w.e. (VKTA-R), 500 m w.e. (JRC-IRMM) to 3800 m w.e. (LNGS). In (Laubenstein et al., 2004) it is recognized that the normalized integral counting 5 Collaboration of European Low-level Underground Laboratories, see e.g. http://www.ptb.de/org/6/63/udo/ cellar.html.
502
G. Heusser et al.
Fig. 3. Background spectra of GeMPI at Heidelberg and Gran Sasso (top) and in comparison with other detectors from CELLAR (footnote 5) laboratories (bottom), taken from (Laubenstein et al., 2004).
rate of the shown spectra decreases only up to about 500 m w.e. in accordance with the growing overburden and than levels off. Consequently, the background contribution from residual contamination becomes dominant in relation to that of cosmic muons beyond that depth. This also holds for GeMPI, but the difference in Fig. 3 reflects the higher radio-purity that has been achieved with GeMPI. In Table 1 background information is given for the most abundant lines of the primordial decay chains U/Th, of 40 K and of 137 Cs (anthropogenic), 60 Co (cosmogenic) as well as the integrated background between 100 and 2730 keV. For GeMPI they have been evaluated from a measuring time of 101 days. They are compared with those measured in the individual detectors of HDM (Dörr, 2002; Dietz, 1999), representing together a statistic of close to 50 kg yr and thus with the highest statistic ever measured in low-level Ge-spectroscopy. Also here an
Low-level Ge gamma-ray spectrometry at the µBq/kg level
503
Table 1 Background count rates for the main primordial lines, for 40 K, 137 Cs, 60 Co and integrated over the energy range 100–2730 keV for GeMPI and the HDM detectors (line background from (Dörr, 2002), integral from (Dietz, 1999)) Energy (keV)
352 609 583 2615 1461 662 1173 1332 100–2730
Chain/Nuclide
238 U/214 Pb 238 U/214 Bi 232 Th/208 Tl 232 Th/208 Tl 40 K 137 Cs 60 Co 60 Co
Peak (integral) background count rate (kg yr)−1 GeMPI
HDM detectors 1–5
24 25 21 18 ± 5 86 ± 12 57 ± 27 43 ± 10 35 ± 8 6840 ± 110
112–177 90–137 18–42 11–22 74–291 41–914 46–71 36–54 11400–18700
improvement is recognised at least for the 226 Ra sub-chain lines. The count rates for 40 K, 60 Co and 137 Cs are comparable to those of the cleanest HDM detectors. The difference in the total background count rate between GeMPI and the HDM detectors may partially reflect the higher contribution from the two neutrino ββ-decay of the isotope 76 Ge, which is higher enriched in the HDM detectors (∼68% versus 7.44% in natural Ge-detectors like GeMPI). On the other side the HDM experiment had a complete neutron shield around detectors #1, 2, 3, 5 and a cosmic veto shield for all detectors. However, the latter resulted in a reduction of only about 0.5% (Dörr, 2002) and the effect of the neutron shield was estimated to be also only in the sub-% range (Dietz, 1999). 3.2. Choice of optimal sample size The units applied so far – counts (kg keV yr)−1 and counts (kg yr)−1 – are useful for background comparisons. More relevant for the sensitivity in Ge-spectroscopy is the specific activity: A [Bq/kg] that can be detected after a certain measuring time. Besides the background B (average from both sides of the peak) it takes into account the counting efficiency εM for the sample with the mass M. In first approximation, by neglecting any proportional factors it can be expressed as A ≈ B 1/2 (εM Mt)−1 , 1/2 (εM Mt)−1 , A ≈ B + tb + t 2 σb2
if the peak background is zero, if the peak background rate b is not negligible,
with εM = efficiency for the sample with the mass M, t = measuring time of the sample, σb = statistical error of the peak background rate, see e.g. Table 1.
504
G. Heusser et al.
The product εM M increases until self-adsorption within the sample becomes dominant. Some orientation is available from the 1/e thickness for the highest energy of interest, in our case the 2.615 MeV line of 208 Tl. For example, it is 234 mm, 33.4 mm, 29.6 mm and 20.7 mm for water, iron, copper and lead, respectively. With this information and economic considerations we have designed the minimal distance between the detector and the inner Cu wall to be about 78 mm on the sides and 132 mm on top (238 mm maximal sample height besides the detector). The Marinelli-type geometry results in an effective volume of close to 15 l since the detector-cap is cylindrical and the sample chamber quadratic in cross section. If affordable, the sample chamber is always completely filled, even though the gain in sensitivity for dense materials like Pb is only minor, but the reliability for the background correction in data evaluation is higher. Uncertainties of minor background differences between the empty and the full chamber, not only from higher radio-impurities in the chamber wall, but also from Rn concentration variations6 in the nitrogen purging gas are thus minimised. For GeMPI extensive background measurements showed indication of this problem but the statistic significance is of course always limited to a few %. 3.3. Some key sample measurements Copper, lead, iron, and Teflon® are very common materials used for shielding or construction of rare event rate detectors in neutrino, double beta, and in dark matter experiments. Measurements of these materials performed with GeMPI resulted in the data given in Table 2. For the Low Background Facility of the LENS solar neutrino experiment (Motta et al., 2003) Cu was used as inner shielding material. Cosmic activation could be kept low by a close co-ordination of the Cu producer, the company to roll it and that of the final machining. Out of this production line two sets of Cu plates (33 mm thick) have been prepared to fit into the sample chamber of GeMPI. Old ships iron of a First World War battle ship was used since many years for different purpose of low-level projects including the fabrication of the screws for the HDM detectors and for GeMPI (Fig. 1). A roll of stainless steel foil (0.125 mm thick) was purchased in course of emanation studies for the BOREXINO experiment. Two lead samples (DowRun and Boliden quality according to the supplier J.L. Goslar) have been investigated for the GERDA experiment (Abt et al., 2004) and one lead sample made from ancient roman lead (Alessandrello et al., 1991, 1998) for the cryogenic double beta decay experiment CUORE (Ardito et al., 2005). The roman lead7 was measured in form of two openend cylinders and a top plate (first cylinder: i = 106 mm; a = 159 mm, h = 76.5 mm; second cylinder: i = 106 mm, a = 162 mm, h = 36 mm; top plate = 176 mm, h = 30 mm). A roll of Teflon foil (0.5 mm thick; i = 98 mm, a = 206 mm, h = 230 mm) was measured for general interest. This roll was also investigated by the emanation method at Heidelberg (Rau and Heusser, 2000) for 226 Ra surface contamination, respectively the 226 Ra supported emanation of the bulk material (Buck, 2001). 6 Variations in 222 Rn concentrations between about 200 µBq/m3 and 3 mBq/m3 have been measured in standard bottled nitrogen with clear indication of 226 Ra contamination in the cylinder from which the high activity has been extracted (Freudiger, 1998). 7 A sample of this lead (see also Alessandrello et al., 1991) has been measured for the experiment CUORE (Ardito et al., 2005).
Low-level Ge gamma-ray spectrometry at the µBq/kg level
505
Table 2 Measured radionuclide concentrations in typical shielding materials and in Teflon® . Uncertainties with coverage factor k = 2; upper limits are decision thresholds, both according to ISO (1995, 2000). The time denotes the life time of the measurement Material
Weigth (kg)
Time (days)
Lead (DowRun)
144.6
Lead (Boliden)
Specific activity (µBq/kg) 226 Ra (U)
228 Th (Th)
101.7
<29
<22
440 ± 140
98 ± 24 (207 Bi) 180 ± 20 (60 Co) (2.7 ± 0.4) × 107 (210 Pb)
144.3
75.0
<46
<31
460 ± 170
<13 (207 Bi) <11 (60 Co) (2.3 ± 0.4) × 107 (210 Pb)
Lead (Roman)
22.1
37.2
<45
<72
<270
Copper (LENS)
125.0
100.7
<16
<19
<88
Stainless steel foil
38.1
80.1
600 ± 200
200 ± 100
1800 ± 600
18000 ± 1000 (60 Co)
Old ships iron
47.29
42.4
150 ± 40
460 ± 140
1000 ± 400
230 ± 40 (54 Mn) <18 (60 Co) <30 (137 Cs) <20 (58 Co)
Teflon
12.35
20.9
1500 ± 240
<70 (137 Cs)
<160
<160
40 K
Various
<19 (207 Bi) <25 (60 Co) <1.3 106 (210 Pb) <10 (60 Co)
As mentioned in the Introduction we used the most prominent lines of the daughter nuclides and 212 Pb, 208 Tl to determine the concentrations of 226 Ra and 228 Th, since in all samples Rn emanation can be neglected. Also for Teflon this is ruled out by the measured emanation rate of 0.26 mBq/kg 222 Rn for the complete roll (Buck, 2001), which scales to about 23 µBq/kg8 by taking into account the 222 Rn diffusion properties for Teflon. The efficiencies have been estimated from Monte Carlo simulations (Neder et al., 2000) based on the code GEANT 3.21 (CERN Program Library). Uncertainties of the simulations have been tested with calibration makeups to be below 5%. Concentrations of 210 Pb in lead have been evaluated via the 803.1 keV line of 210 Po. The high radio-purity level found in Pb and Cu makes both shielding material well suited for rare event experiments, but especially for GERDA (Abt et al., 2004). It is assumed that the low concentration of 226 Ra and 228 Th is generally valid for good lead qualities and after further verification may be also for electrolytic Cu. The higher 207 Bi and 60 Co concentrations in the Pb of DowRun quality might be due to a contamination introduced at the company during the cutting of the recess in the plates to fit around the detector. For the Boliden lead quality the cutting was done at the workshop at MPI with the usual care to avoid cross contamination. It is astonishing that 40 K is lower in the ancient roman lead than in modern lead. Also for the LC2 lead (about 0.4 mBq/kg 210 Pb) used in the HDM experiment a 40 K concentration 214 Pb, 214 Bi
8 Zuzel, G., private communication, calculation based on measurements of the diffusion coefficient and of the solubility by W. Mueller, Master Thesis, University of Heidelberg (Mueller, 1978).
506
G. Heusser et al.
of about 0.31 ± 0.3 mBq/kg was found by neutron activation (Pernicka, 1991). The stainless steel has unusual low concentrations of 226 Ra and 228 Th (see e.g. Arpesella et al., 2002) but rather high 60 Co content. The old ships iron is even lower in 226 Ra, and only slightly higher in 228 Th. However the contribution of this material from which the screws of GeMPI (see Section 2.2) have been fabricated to the background of GeMPI is not negligible. The decision for the use of this material was based on a measurement with a less sensitive Ge-spectrometer at Heidelberg, which resulted only in upper limits at the mBq/kg level. Teflon® proved to be rather free of 226 Ra and 228 Th impurity, but 40 K is unusually high compared to a measurement on the same batch of Teflon material with neutron activation analysis9 . The preliminary result of 0.7 mBq/kg 40 K (22 ppb K) is in the same ballpark as that of other NAA of K in Teflon10 . Also earlier Ge-measurements of smaller more bulky samples of Teflon by the authors did not show indications of 40 K contamination above the sensitivity level of a few mBq/kg. An explanation could be that a K-rich substance was placed on the surface during the foil forming process, which was removed in the sample preparation (acid cleaning, etc.) procedure for NAA. Further investigations are planned to solve the contradiction between the γ -ray and the neutron activation analysis. Apart from the suspected local surface contamination of 60 Co and 207 Bi on the DowRun lead and of the 40 K on the Teflon foil there is no indication of surface contamination for the other investigated samples, even though only wipe cleaning with alcohol soaked paper was applied in case of the lead and copper sample preparation after machining. Their surface area corresponds to about 1.6 m2 and 1.1 m2 . The steel roll and the Teflon® roll were measured as delivered except for outer surface wipe cleaning. Rust from the old ships iron plates (about 45 mm thick) was removed with diluted nitric acid. 3.4. Cosmogenic production rates in copper One set of the two identical produced Cu plate samples was exposed for 270 days at a hall of LNGS (altitude 1038 m a.s.l., latitude 42◦ 27 N) under an averaged roof thickness of about 20 g/cm2 . The 33 mm thick slabs were placed side by side as a single layer on a pallet during the exposure. The average total count rate (100–2730 keV) of the activated sample at end of counting (106.2 days after the end of the exposure) was 447.3 day−1 , compared to the 41.2 day−1 of the normal background listed in Table 1. Table 3 lists the evaluated specific activity for 9 radionuclides from the spectrum measured over 103 days that has been scaled to saturation. The saturation activity corresponds to the cosmogenic production rate. Also the following upper limits for the primordials were obtained for this set of copper plates: <35 µBq/kg 226 Ra; <20 µBq/kg 228 Th; <110 µBq/kg 40 K. They are well in agreement with the result of the unexposed set (Table 2). The cosmogenic radionuclides are typical products of spallation reactions, since at least 4 nucleons and up to 18 nucleons have to be emitted after the neutron or the proton interacted with 63 Cu. 60 Co reaches with a production rate of 2.1 mBq/kg the highest saturation activity and thus strongly exceeds the activity of any other radio-impurity listed in Table 2 for Cu. 9 Cattadori, C., private communication. 10 Piepke, A., private communication.
Low-level Ge gamma-ray spectrometry at the µBq/kg level
507
Table 3 Cosmogenic production rates (saturation activity) in Cu. Uncertainties according to ISO (1995, 2000) with coverage factor k = 1 Radionuclide
Half-life
Saturation activity (µBq/kg)
56 Co
77.31 days 271.83 days 70.86 days 5.27 year 312.15 days 44.5 days 83.79 year 15.97 days
230 ± 30 1800 ± 400 1650 ± 90 2100 ± 190 215 ± 21 455 ± 120 53 ± 18 110 ± 40
57 Co 58 Co 60 Co 54 Mn 59 Fe 46 Sc 48 V
This demonstrates the importance of limiting the cosmic ray exposure if Cu is used as construction material for rare event experiments. Cosmogenic production rates deduced from a reconstruction of not well documented sea level exposure of Cu used for the HDM experiment (Heusser, 1994) yielded similar results: 1 mBq/kg 60 Co; 1.7 mBq/kg 58 Co; 1 mBq/kg 57 Co and 0.35 mBq/kg 54 Mn. The 60 Co activity in the unexposed set of Cu plates was undetectable at a limit of 10 µBq/kg (Table 2) and thus indicates a sea level exposure of less than about 14 days. In Fe the production rate for 54 Mn is even higher than in Cu (Heusser, 1994), therefore it was still detectable in the old ships iron (Table 2), although this sample was resting for many 54 Mn half-lives at a shielded position in the floor of the Heidelberg low-level laboratory. No activation was detectable in the lead samples, which however have not been exposed by purpose for a longer period. There were also no indications by earlier investigations of the authors, however performed at a lower sensitivity level. The measured production rates are useful to study and test model calculations for cosmogenic production, an issue of growing importance for upcoming rare event experiments.
4. Future developments The high statistics spectrum of the HDM ββ experiment has been modelled by Monte Carlo simulations in order to localise the residual contamination in the cryostat systems of the individual detectors (Dörr, 2002; Dörr and Klapdor-Kleingrothaus, 2003). This method is exploiting the different degrees of attenuation for γ -ray emissions of different energies from different materials and different locations in the experimental set up. The result of this analysis and earlier ones (Dietz, 1999; Maier, 1995) is that the most probable location of the primordial activities is the copper of the detector cryostats. In Table 4 the thus deduced concentrations are compared to the measured ones of the same NOSV copper quality with GeMPI. The most plausible explanation of the large spread in primordial concentrations among the HDM cryostats and their much higher contamination (one exception – 228 Th in ANG 2) with respect to the measured bulk material is that the activity is surface correlated. The model for the simulation was based on the assumption that the activity is homogeneously distributed. A variation
508
G. Heusser et al. Table 4 Contamination in Cu according to Monte Carlo calculation for the HDM detectors (Dörr and Klapdor-Kleingrothaus, 2003) and measured with GeMPI Location of copper
Activity (µBq/kg) 226 Ra
Cu of cryostat No. ANG 1 ANG 2 ANG 3 ANG 4 ANG 5
168 ± 8 91 ± 4 105 ± 5 115 ± 3 100 ± 4
Cu measured with GeMPI
16
228 Th
84 ± 7 10 ± 3 84 ± 5 87 ± 4 26 ± 4 12
40 K
236 ± 61 78 ± 22 927 ± 46 199 ± 4 1632 ± 49 88
in surface cleanliness is also plausible from the fact that different people have assembled the HDM detectors over a period of 4 years under not always identical conditions. A clear indication follows also from count rate variations by about a factor 30 of the 210 Po α-peak observed in the high energy spectra of the 5 detectors (Dietz, 1999). The location of this contamination must be the area around the core or in it, since only there alphas are detectable. From these findings it is obvious that a strong reduction of surface and mass of the several layers of cladding material around the crystal (see Fig. 1) may open the possibility for further background reduction. This idea (Heusser, 1995) is now being realised in the GERDA experiment, a further development of a former proposal (Klapdor-Kleingrothaus et al., 1999; Baudis et al., 1999). The basic design is to contact and hold the Ge-crystals with a minimum of highly radiopure material, to deploy them in liquid nitrogen or liquid argon for cooling and to shield them against external radiation in a deep underground location. It has been demonstrated that liquid nitrogen can be made highly radio-pure (Heusser et al., 2000; Zuzel et al., 2004). The high insulation property of the liquid gases allows installing a larger array of Ge-diodes in a compact way with only little distance to each other. GERDA aims to reduce the background in the energy region around the Qββ -value of 76 Ge at 2.039 MeV by two to three orders of magnitude compared to what has been reached in HDM (Klapdor-Kleingrothaus et al., 2004). It is hoped that the same reduction factor can be obtained also for the whole energy range from 50–2800 keV. A further prerequisite is that cosmogenic activation of the crystals from sea level exposure needs more control than achieved so far. If once cryogenic liquid-type Ge-spectroscopy is established, a further step in screening sensitivity would be reachable, may be even down to the nBq/kg range. Limitations will arise from the increasing measuring periods needed to arrive at useful counting statistics from such low activity concentrations, if not compensated by larger sample masses and multi-Ge detector arrays. Therefore, an application in life sciences seems to be rather limited, except may be for some very special key measurements. For example cosmogenic activation of materials exposed at surface or shallow depth could be used as tracer for their residence time. Copper samples of about 100 g could be sufficient to measure cosmogenic radio-nuclides with dif-
Low-level Ge gamma-ray spectrometry at the µBq/kg level
509
ferent half-lives (see Table 4). Gamma-active radioisotopes produced by accidental neutron exposures could be an issue in radiation protection.
5. Conclusions It could be demonstrated that underground low-level germanium spectrometry can reach a sensitivity level in the range of 10 µBq/kg, corresponding to about 10−12 g/g 238 U/232 Th and about 10−9 g/g K. This has been achieved by a combination of background reduction, mainly as a result of extensive material screening of the construction materials and a favourable large sample capacity. Some key measurements of materials frequently used in rare event projects as copper and lead are helpful for the design of upcoming second generation experiments. To overcome the relative slow sample throughput due to the long counting time there are two additional GeMPI-type Ge-spectrometers under construction, of which one is already in the testing phase at LNGS (Rugel et al., 2005).
Acknowledgements The authors gratefully acknowledge the technical staff of the MPI workshop guided by V. Mallinger for the excellent work done in realising GeMPI, but especially O. Lackner for the skilful machining of the parts next to the crystal. Canberra Semiconductors N.V., Olen, Belgium is thanked for the very good co-operation in the design and assembly of the cryostat. We are indebted to B. Prokosch for his help during the starting phase of this project. The colleagues of the collaborations GALLEX/GNO, BOREXINO, HEIDELBERG–MOSCOW and GERDA are thanked for valuable discussions.
References Abt , I., et al. (2004). A new 76 Ge double beta decay experiment at LNGS, hep-ex/04040. Alessandrello, A., et al. (1991). Measurements on radioactivity of ancient roman lead to be used as shield in searches for rare events. Nuclear Instruments & Methods in Physics Research B 61, 106–117. Alessandrello, A., et al. (1998). Measurements of internal radioactive contamination in samples of Roman lead to be used in experiments on rare events. Nuclear Instruments & Methods in Physics Research B 142, 163–172. Alimonti, G., et al. (2001). Borexino Collaboration, Science and Technology of Borexino: A real time detector for low energy solar neutrinos. Astroparticle Physics 16, 205–234. Ardito, R., et al. (2005). CUORE: A cryogenic underground observatory for rare events, hep-ex/0501010. Arpesella, C., et al. (2002). Borexino Collaboration, Measurements of extremely low radioactivity levels in Borexino. Astroparticle Physics 18, 1–25. Baudis, L., et al. (1999). Nuclear Instruments & Methods in Physics Research A 426, 425–435. Buck, Ch. (2001). Diploma thesis. University of Heidelberg. Dietz, A. (1999). Diploma thesis. University of Heidelberg. Dörr, Ch. (2002). Diploma thesis. University of Heidelberg. Dörr, C., Klapdor-Kleingrothaus, H.V. (2003). New Monte Carlo simulation of the HEIDELBERG–MOSCOW double beta decay experiment. Nuclear Instruments & Methods in Physics Research A 513, 596–621. Freudiger, B. (1998). Diploma thesis. University of Heidelberg.
510
G. Heusser et al.
Gasparo, J., et al. (2004). Measurements of 60 Co in spoons activated by neutrons during the JCO criticality accident at Tokai-Mura in 1999. Journal of Environmental Radioactivity 73, 307–321. Guenther, M., et al. (1997). Heidelberg–Moscow ββ experiment with 76 Ge: Full set up with five detectors. Physics Review D 55, 54. Heusser, G. (1994). Background in ionizing radiation detection. In: Garcia-Leon, M., Garcia-Tenorio, R. (Eds.), LowLevel Measurements of Radioactivity in the Environment. Proc. 3rd Intern. Summer School, Huelva, 1993. World Scientific, Singapore, pp. 69–112. Heusser, G. (1995). Low-radioactivity background techniques. Annual Review of Nuclear and Particle Science 45, 543–590. Heusser, G., et al. (2000). 222 Rn detection at the µBq/m3 range in nitrogen gas and new 222 Rn purification technique for liquid nitrogen. Applied Radiation and Isotopes 52, 691–695. ISO, International Organization for Standardization (1995). Guide to the Expression of Uncertainty in Measurement (First corrected edition). Geneva, Switzerland. ISO, International Organization for Standardization (2000). Determination of the Detection Limit and Decision Threshold for Ionizing Radiation Measurements – Part 3: Fundamentals and Application to Counting Measurements by High Resolution Gamma Spectrometry, without the Influence of Sample Treatment. First edition. International Standard ISO 11929-3. Geneva, Switzerland. Klapdor-Kleingrothaus, H.V., et al. (1999). MPI-Report, MPI-H-V 26. Klapdor-Kleingrothaus, H.V., et al. (2004). Data acquisition and analysis of the 76 Ge double beta experiment in Gran Sasso 1990–2003. Nuclear Instruments & Methods in Physics Research A 522, 371–407. Laubenstein, M., et al. (2004). Underground measurement of radioactivity. Applied Radiation and Isotopes 61, 167– 172. Maier, B., (1995). PhD thesis. University of Heidelberg. Motta, C., et al. (2003). LLBF: LENS prototype at Gran Sasso. Nuclear Physics B (Proc. Suppl.) 118, 451. Mueller, W. (1978). Master thesis. University of Heidelberg. Neder, H. (1998). Diploma thesis. University of Heidelberg. Neder, H., et al. (2000). Low-level germanium-spectrometer to measure very low primordial radio-nuclide concentrations. Applied Radiation and Isotopes 53, 191–195. Pernicka, E. (1991). Private communication. Prokosch, B. (1996). Diploma thesis. University of Heidelberg. Rau, W., Heusser, G. (2000). 222 Rn emanation measurements at extremely low activity. Applied Radiation and Isotopes 53, 371–375. Rugel, G., et al. (2005). Nuclear Physics B (Proc. Suppl.) 143, 564. Zuzel, G., et al. (2004). Ar and Kr concentrations in nitrogen as measure of the 39 Ar and 85 Kr activity in connection with the solar neutrino experiment BOREXINO. Applied Radiation and Isotopes 61, 197–201.
511
Depth profiles of environmental neutron fluxes in water and lead Y. Hamajima* , K. Komura Low-Level Radioactivity Laboratory, Kanazawa University, Wake, Tatsunokuchi, Japan Abstract Depth profiles of environmental neutron fluxes were measured by the activation method of gold targets in 0–200 g cm−2 of seawater, fresh water and lead, using ten ultra-low-background Ge detectors in Ogoya Underground Laboratory. The intensity of thermal neutrons in gold has been decreasing exponentially, the mass attenuation coefficient μ in gold was determined to be 0.70 ± 0.17 cm−2 g−1 . The neutron profiles in seawater and fresh water were almost the same. A maximum at ∼5 g cm−2 , a saddle between 20 and 30 g cm−2 , and an exponential decrease after 50 for fresh water and 100 g cm−2 for seawater were observed. The mass attenuation coefficient μ of fresh water was determined to be 0.0090 ± 0.0011 cm−2 g−1 . The depth profile of neutrons in lead after having a maximum at 5–10 g cm−2 was almost constant. Keywords: Environmental neutrons, Neutron self-absorption, Activation method, Gold, Water, Lead, Underground laboratory
1. Introduction It is well known that environmental neutrons have two sources of origin: (i) secondary cosmic-ray neutrons, originating as a result of interactions of primary cosmicray particles with the atmosphere, and (ii) radiogenic neutrons, originating in radioactive decay of heavy elements in the Earth crust. The characteristics of neutrons such as the energy distribution, the depth profiles, and the temporal variations have been studied using BF3 and 3 He counters for many years (e.g. NCRP, 1975), however, it has been difficult to declare the characteristics with high depth resolution in materials, e.g. in water and metals, because of difficulties of counter settings and low neutron fluxes. Canet et al. (2000) measured recently the ratio of the activity induced by environmental neutrons in a gold target placed at certain depths in water to that measured at the ground level. Radioactive products of neutron spallation and capture in rocks and soil have been used in geo-sciences for dating and erosion studies. As a depth profile of neutron flux in a given * Corresponding author. Address: Low-Level Radioactivity Laboratory, Kanazawa University, Tatsunokuchi, Ishikawa 923-1224, Japan; phone: (+81) 761 51 4440; fax: (+81) 761 51 5528; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08040-X
© 2006 Elsevier Ltd. All rights reserved.
512
Y. Hamajima and K. Komura
medium has not been measured yet, the production rates of cosmogenic radionuclides have been estimated by model calculation (Dep et al., 1994; Dunne, 2003). Small amounts of radionuclides produced by neutrons were used for example for the reassessment of the neutron dose in Hiroshima by measuring their residual activity in various samples (Komura and Hamajima, 2004), or for the assessment of doses from a recent accident at the fuel plant in Tokai-mura (Komura, 2001). In low-level gamma-ray measurements, photopeaks of neutron induced and scattered reaction products in the Ge detectors and in surrounding shielding materials are found both in ground and shallow underground facilities (Heusser, 1995). Special anti-cosmic shielding of detectors, and/or their operation deep underground is required to reduce the cosmic-ray component of the detector’s background. Recently, we constructed an ultra low-level measurement system in the Ogoya Underground Laboratory (OUL) (Hamajima and Komura, 2004), where 6 well-type, 4 planar, and 1 coaxialtype HPGe detectors with high efficiency were set up. The operating depth of the OUL laboratory is 270 m w.e. The integrated background counting rates of these detectors in the energy range from 100 to 2000 keV were less than 2 min−1 . The effects of cosmic-rays and secondary neutrons are approximately 100 times weaker than in a surface laboratory. Therefore, it is now possible to measure neutron depth profiles using these detectors at the same time without a particular normalization. In this work, depth profiles of environmental thermal neutron fluxes in seawater, fresh water and lead were measured using the activation technique of gold foils. The effect of selfshielding in the gold target and production rates during transportation from the exposure site to OUL were also measured and the results were corrected to the former. 2. Experimental 2.1. Nuclear data and reactions The environmental neutron flux was estimated from the reaction 197 Au(n, γ )198 Au using a gold target. The half-life of 198 Au is 2.69 day, and the activation cross section for thermal neutrons is 98.8 barn (Lederer and Shirley, 1978). Since the energy distribution of neutrons in a target material is not known, it was assumed that all 198 Au activity was induced by thermal neutrons only, although contributions of epi-thermal and the fast neutrons need not be negligible. 2.2. Target and exposure The targets were 15–25 g gold foils (40 × 300 mm2 ) and plates (40 × 40 mm2 ) of 0.19 g cm−2 (0.1 mm) and 1.35 g cm−2 (0.7 mm) in thickness, respectively. Ten of the targets were horizontally hung from a small float at 0, 5, 10, 15, 20, 30, 50, 100, 150 and 200 cm depth, all at the same time. The target was exposed for 19–22 days in seawater at a pier head of the Tsukumo Bay (almost an infinite surface area and depth over 5 m), and in fresh water at Wake pond (160 × 60 m2 and 7 m deep). Other nine targets were sandwiched in the both sides using a lead absorber with the same thickness of 0, 0.5, 1, 1.5, 2, 5, 10, 15 and 20 cm plates (40 × 50 cm2 ) and/or blocks
Depth profiles of environmental neutron fluxes
513
(50 × 50 cm2 ) and exposed until saturation at 10 cm above ground at the Low-Level Radioactivity Laboratory (LLRL). For the estimation of self-absorption of neutrons in the targets, a thin gold target of 10 mm in diameter and 0.039 g cm−2 (0.02 mm) in thickness was stacked at 0, 0.039, 0.096, 0.39, 0.77 and 1.35 g cm−2 of the gold absorber. As the backside of the stack was shielded by 1-mm Cd plate, neutrons were registered only from the topside surface. The exposure was carried out for 9–29 days outside of a 252 Cf neutron source (10 MBq at March 16, 2000) and an Am–Be neutron source (11.1 GBq at December 21, 1964), surrounded by 30 cm of paraffin. The backside of the target stack was covered with 10 cm of paraffin as well. The energy distribution of neutrons at this exposure site was not measured, but it was assumed to be approximately the same as for environmental neutrons. The measured neutron flux at this site was 0.48 cm−2 sec−1 , which was two orders of magnitude larger than the environmental neutron flux. 2.3. Production rates during transportation It took less than 4 hours by car to reach the OUL from the exposure site. The activity produced during transportation was taken into consideration. All targets during transportation were stored in a small container with 1 mm of Cd and 5 cm of paraffin as neutron shielding. The environmental neutron flux in this container was 1/10 of that above ground. Therefore, the activity produced during the transportation during 4 h was less than 0.4% and 2.8% of the surface and the minimum activity produced during the exposure period, respectively. 2.4. Gamma-ray spectrometry and calculations The photopeak of 412 keV was measured for 4–5 days using ten of the ultra low-background HPGe detectors in OUL at the same time. The small foil was left as it was, and the long foil target was rolled and then inserted into the well of the detector. The plate target was put onto the planar or coaxial-type detector head. The net counts of the photopeak were in the range 50–100 counts. The counting efficiency of each detector was experimentally estimated taking into account the thickness of the gold target. The total uncertainty of the efficiency measurement was estimated to be within 10%. The neutron flux was calculated from the activation reaction, taking into account corrections for saturation and decay during the exposure, transportation, counting period, and the effect of neutron self-shielding in the target. The production rate during transportation (less than 2.8%) and the differences in the self-absorption of 412-keV photons between the targets (less than 3.4%) were neglected. The reported uncertainties of the neutron fluxes include only the measurement uncertainties.
3. Results and discussion The results of the neutron flux measurements in thin gold target, seawater, fresh water, and lead are listed in Tables 1–4, respectively. In order to directly compare the neutron depth profiles in different materials, the depth or absorber thickness are shown in units of g cm−2 .
514
Y. Hamajima and K. Komura Table 1 Neutron flux in a thin gold foil Thickness of the absorber (g cm−2 )
Neutron flux (cm−2 sec−1 )
0.00 0.04 0.10 0.39 0.77 1.35
0.563 ± 0.050 0.416 ± 0.042 0.466 ± 0.046 0.376 ± 0.033 0.288 ± 0.034 0.230 ± 0.022
Table 2 Environmental neutron fluxes in seawater Depth (g cm−2 )
0.0 5.1 10.2 15.4 20.5 30.7 51.2 102.3 153.5 204.6
Neutron flux (10−4 cm−2 sec−1 ) Jun. 11, 2004– Jul. 2, 2004
Jun. 20, 2004– Aug. 11, 2004
7.47 ± 1.18 7.00 ± 1.38 4.80 ± 0.93 3.84 ± 0.84 5.14 ± 1.21 3.60 ± 0.70 4.57 ± 0.78 4.22 ± 1.31 2.54 ± 0.47 1.13 ± 0.29
5.14 ± 1.11 13.4 ± 1.34 8.34 ± 1.38 5.54 ± 0.70 4.34 ± 0.55 4.72 ± 0.48 3.36 ± 0.43
Table 3 Environmental neutron fluxes in fresh water Depth (g cm−2 )
Neutron flux (10−4 cm−2 sec−1 ) May 6, 2004– May 26, 2004
0.0 5.0 10.0 15.0 20.0 30.0 50.0 100.0 150.0 200.0
9.74 ± 1.01 8.66 ± 1.74 6.94 ± 1.12 6.51 ± 1.03 6.88 ± 1.61 6.31 ± 0.94 6.03 ± 0.95 3.21 ± 0.83 2.13 ± 0.49 1.48 ± 0.37
Depth profiles of environmental neutron fluxes
515
Table 4 Environmental neutron fluxes in lead Depth (g cm−2 )
0.0 5.7 11.3 17.0 22.7 56.7 113.4 170.1 226.8
Neutron flux (10−4 cm−2 sec−1 ) May 14, 2004– Jun. 10, 2004∗
Jun. 16, 2004– Jul. 29, 2004
21.8 ± 2.8 31.5 ± 3.2 31.1 ± 3.0
21.8 ± 1.9
19.5 ± 2.6 16.9 ± 1.9
15.5 ± 1.0 19.4 ± 1.0 16.6 ± 1.2 11.7 ± 0.9
∗ The fluxes in this column were normalized using the surface flux given in the right column.
3.1. Thermal neutron fluxes in the gold target and correction for the neutron self-shielding The depth profile of thermal neutron flux in the gold target is shown in Fig. 1. The intensity is decreasing with the depth exponentially. In a thick target, the attenuation of a parallel neutron beam can be expressed as (Friedlander et al., 1981) I = I0 e−nσt x , where I is the number of incident neutrons per unit time, n is the number of target nuclei per cm3 of the target, σt is the total cross section in cm−2 , and x is the target thickness in cm. The experimental results are consistent with this equation. The mass attenuation coefficient (μ) in gold estimated from the figure is 0.70 ± 0.17 cm2 g−1 , and that calculated from the cross section of 98.8 barn is 0.30 cm2 g−1 . As the incident angle of neutrons is considered to be isotropic in the present experimental set-up, the larger attenuation coefficient obtained from the curve seems to be reasonable. As environmental neutrons irradiate both surfaces of the target exposed in seawater, fresh water, and lead, the distribution of the neutron flux (f ) as a function of thickness of the material (x) can be expressed as f (x) = eμx + e−μx . Therefore, we need to integrate this function between both surfaces (from –t to t), normalize by the uninfluential case, and get the following expression R=
1 eμt − e−μt , μt eμt + e−μt
where R is the ratio of the shielding effects, and t is the half-thickness of the gold target in g cm−2 . The estimated correction coefficient (1 − R) is 0.002 and 0.07 for 0.19 g cm−2 and
516
Y. Hamajima and K. Komura
Fig. 1. Depth profile of neutron fluxes in gold foil.
1.35 g cm−2 of gold target, respectively. All results were corrected using these correction coefficients. 3.2. Depth profiles of environmental neutron fluxes in seawater, fresh water and lead The environmental neutron fluxes in seawater, fresh water, and lead are plotted as a function of depth in Figs. 2, 3 and 4, respectively. The difference between the absolute fluxes at the surface (at 0 g cm−2 ) is derived from the difference of the environmental neutron fluence during exposure period in the atmosphere. The observed neutron profiles in seawater and fresh water were almost the same. The mass attenuation coefficient μ for fresh water is 0.0090 ± 0.0011 cm2 g−1 , smaller than the one reported by Canet et al. (2000). There seems to be a small maximum at ∼5 g cm−2 (visible only in seawater), and a saddle-like structure between 20 and 30 g cm−2 . It may be deduced that part of environmental fast and epi-thermal neutrons are thermalized around this depth, so that the neutron energy distribution is shifted to thermal neutrons. The flux is exponentially decreasing in depths deeper than 50 g cm−2 . The shape of the depth profiles in seawater and fresh water is in a reasonable agreement with calculations of Kastner et al. (1970), however, the absolute values differ by about a factor of ten. This disagreement may come from the fact that neutron energies in the calculation are in the range from thermal to 0.25 eV, while the targets are assumed to be activated only by thermal neutrons, as the epi-thermal and the fast neutron absorption cross-sections in the present work are neglected. Moreover, the cut-off rigidity of cosmic-radiation depends on the latitude, which is reported to be 10–12 GeV in Japan, but 1–4 GeV in the North Europe and the North American continent (Kodama, 1968). Since the low energy primary cosmic radiation is cut in Japan, it is expected that the environmental neutron flux will be smaller in Japan than at northern latitudes. The neutron depth profile in lead was at the investigated range of depths between 0 and 250 g cm−2 almost constant, a small maximum was found between 5 and 10 g cm−2 . Dep
Depth profiles of environmental neutron fluxes
517
Fig. 2. Depth profiles of environmental neutron fluxes in seawater.
Fig. 3. Depth profiles of environmental neutron fluxes in fresh water. The solid line is the simulation result of Kastner et al. (1970).
518
Y. Hamajima and K. Komura
Fig. 4. Depth profiles of environmental neutron fluxes in lead.
et al. (1994) simulated neutron flux in concrete and they found a maximum between 30 and 50 g cm−2 . The difference may be due to the neutron energy shift in different target materials, however, this can be confirmed only when simulations for lead will be available. 4. Conclusions Depth profiles of environmental neutron fluxes in seawater, fresh water and lead were measured by the activation method of gold foils. The neutron profiles in seawater and fresh water were almost the same, having a maximum at 5–10 g cm−2 and a saddle at 20–30 g cm−2 , and then decreasing exponentially for the depths greater than 100 and 50 g cm−2 , respectively. The mass attenuation coefficient μ of fresh water was measured to be 0.0090 ± 0.0011 cm2 g−1 . The depth profile in lead is scarcely decreasing, a maximum was found at 5 to 10 g cm−2 . In order to clearly explain these profiles, it is necessary to carry out Monte Carlo simulations for different target materials, and to measure an energy distribution of neutrons with depth. Acknowledgements The authors wish to acknowledge the officials of Tatsunokuchi-machi for their authorization to carry out the exposure experiment at Wake pond. This work was partly supported by a Grant-in-Aid for Scientific Research, and also the 21st COE Project from the Ministry of Education, Culture, Sports, Science and Technology of Japan. References Canet, M.J.M., Hult, M., Köhler, M., Johnston, P.N. (2000). Measurements of activation induced by environmental neutrons using ultra low-level γ -ray spectrometry. Applied Radiation and Isotopes 52, 711–716. Dep, L., Elmore, D., Fabryka-Martin, J., Masarik, J., Reedy, R.C. (1994). Production rate systematics of in-situproduced cosmogenic nuclides in terrestrial rocks: Monte Carlo approach of investigating 35 Cl(n, γ )36 Cl. Nuclear Instruments and Methods in Physics Reseach B 92, 321–325.
Depth profiles of environmental neutron fluxes
519
Dunne, J.A. (2003). Monte Carlo simulations of low-energy cosmogenic neutron fluxes near the bottom of cliff faces. Earth and Planetary Science Letters 206, 43–49. Friedlander, G., Kennedy, J.W., Macias, E.S., Miller, J.M. (1981). Nuclear and Radiochemistry, Third edition. Wiley, New York. Hamajima, Y., Komura, K. (2004). Background components of Ge detectors in Ogoya Underground Laboratory. Applied Radiation and Isotopes 61, 179–183. Heusser, G. (1995). Low radioactivity background techniques. Annual Review of Nuclear and Particle Science 45, 543–590. Kastner, J., Oltman, B.G., Feige, Y., Gold, R., Congel, F. (1970). Nuclear radiation detection for the natural environment. IEEE Nuclear Science 17, 144–150. Kodama, M. (1968). Geomagnetic and solar modulation effects of sea-level cosmic ray intensity: Summary of cosmic ray latitude surveys aboard the expedition ship SOYA during 1956–1962 (AERONOMY). In: JARE Scientific Reports. Series A, Aeronomy. Japanese Antarctic Research Expedition, vol. 5, pp. 1–61. Komura, K. (2001). Radiochemical approach to the JCO criticality accident in Tokai-mura, 1999 – An overview of the radiochemistry group. Journal of Radiation Research 42, S17–S29. Komura, K., Hamajima, Y. (2004). Ogoya Underground Laboratory for the measurement of extremely low levels of environmental radioactivity: Review of recent projects carried out at OUL. Applied Radiation and Isotopes 61, 185–189. Lederer, C.M., Shirley, V.S. (1978). Table of Isotopes, 7th edition. Wiley, New York. National Council on Radiation Protection and Measurements (NCRP) (1975). Natural background radiation in the United States: Recommendations of the National Council on Radiation Protection and Measurements. Report No. 45. NCRP, Washington, DC.
520
Radiocarbon measurement by liquid scintillation spectrometry at the Gran Sasso National Laboratory Wolfango Plastinoa,b,* , Lauri Kaiholac a Department of Physics, University of Roma Tre, Rome, Italy b National Institute of Nuclear Physics, Section of Roma Tre, Rome, Italy c PerkinElmer Life and Analytical Sciences, Turku, Finland
Abstract The background of a liquid scintillation spectrometer for radiocarbon measurements was investigated in the Gran Sasso National Laboratory (National Institute of Nuclear Physics), where 1400 m rock overburden (3800 m w.e.) absorbs most of the cosmic radiation. Comparisons were made with measurements carried out in surface laboratories (Radiocarbon Laboratory of ENEA, Bologna, Italy; PerkinElmer Life and Analytical Sciences, Turku, Finland) using Teflon and optimised copper–Teflon vials. The residual radioactivity of phototubes and the role of radon were a larger fraction of the total background signal in the Gran Sasso underground laboratory. Keywords: Liquid scintillation spectrometry, Radiocarbon, Cosmic background, Environmental radioactivity background, Underground laboratory
1. Introduction Carbon dating by radiometric detection of 14 C decay using liquid scintillation counters (LSC) is still widely used, although accelerator mass spectrometry (AMS) has been available since 1970. The low-level liquid scintillation spectrometer Quantulus™1 was introduced in 1984 to reach such background count rates and stability that radiometric detection of very old environmental and archaeological samples became possible (Polach et al., 1984). In routine dating background levels down to 55 kyr have recently been reported, which is better than in routine AMS dating (Hogg, 2004). As maximum measurable age is AON tmax = 8033 ln √ , 8B/T
(1)
* Corresponding author. Address: Department of Physics, University of Roma Tre, Via della Vasca Navale 84, I-00144 Roma, Italy; phone: (+39) 06 55177277; fax: (+39) 06 5579303; e-mail:
[email protected] 1 Quantulus is a trademark of PerkinElmer, Inc.
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08041-1
© 2006 Elsevier Ltd. All rights reserved.
Radiocarbon measurement at the Gran Sasso National Laboratory
521
where AON is the modern sample count rate corresponding to 13.56 dpm/g carbon, i.e. concentration of 14 C in the atmosphere non-disturbed by recent human activities (fossil fuel combustion, nuclear tests and nuclear power effluents), B is the background count rate, and T is the counting time. Sometimes a radiocarbon figure of merit is used, defined as aV E AON FOM = √ = √ , (2) B B where a = 11 dpm/ml benzene corresponds to 13.6 dpm/g carbon (multiplied by the density of benzene {0.88} and contribution of carbon in C6 H6 {0.92}). V is the sample volume and E is the counting efficiency (cpm/dpm). There are several ways to achieve high performance in counting: (a) to improve counting efficiency, e.g. low-quenching benzene samples provide high counting efficiency; (b) to have as many radiocarbon atoms in the sample as possible. Benzene molecule contains 6 carbon atoms, is aromatic scintillator and furthermore no scintillation cocktail is needed. The cocktail is replaced by fluors added in powder form; (c) to increase the sample volume. The total sample volume is limited to 20 ml in LSC, and usually it is difficult to obtain even this sample volume; (d) to have as little inherent radioactivity in the instrument as possible. This is not straightforward, as some counter components contain radioactive materials. Phototubes for instance contain some 40 K, U and Th; (e) to decrease the instrument background by various means. Anti-cosmic guard was selected as the main active background reduction device in Quantulus with the massive lead shield being the main passive background reduction component. Other forms of background reduction include a pulse amplitude/shape discrimination, high bias, adjustable multichannel windows, removal of electromagnetic interferences and static electricity. Laboratories, which are shielded with special low activity materials and/or are placed underground, offer additional advantages, as small fraction of external radiation can penetrate the lead shield; (f ) to increase the counting time. This requires a very stable instrument operating on weeks/sample basis, however, the output of the instrument will be low. The items (e) and (f ) affect tmax via the square root term of Equation (1). The instrument’s background is in direct proportion to the sample volume in masked vials (Kaihola et al., 1992). The throughput of the instrument will, however, improve with the lower background. Long term stability is the most important requirement in radiocarbon dating and in any other low activity radiometric measurement as counting times will be long. Higher backgrounds may be accepted as they can be compensated by longer counting times. This study introduces results of the work carried out in the Gran Sasso National Laboratory (Italy), where cosmic background is almost fully eliminated. 2. Quantulus with anti-cosmic guard detector and sample-guard coincidences Quantulus passive lead shielding is designed to stop soft cosmic radiation, electrons, positrons and photons, mainly attacking the detector from the direction of the zenith. The hard compo-
522
W. Plastino and L. Kaihola
Fig. 1. Quantulus detector with its passive lead shield and active guard detector.
nent on the Earth’s surface at sea level is mostly composed of muons, 170 particles/(m2 s) (Povinec, 1994). The Quantulus sample counting chamber is inside an active, asymmetric anti-cosmic guard detector made of Oxygen-free High Conductivity (OFHC) copper, filled with mineral oil scintillation cocktail. The guard has two phototubes working in coincidence, which detect a fraction of the inherent tube radioactivity. The guard is a true cosmic event detector with no optical contact with the sample. Its performance does therefore not depend on the sample matrix, as in designs where single phototube pair is used for both sample and guard. The copper lining together with the guard copper walls attenuates Pb X-rays generated by the small amount of 210 Pb contamination in the lead. The sample is lifted into the counting chamber with a massive copper piston, which acts as a passive shield downwards (Fig. 1). The anti-cosmic guard reduces the full window background by a factor of five. It detects cosmic muons to almost 100%. This fact can be verified by recording the sample events, which occur simultaneously with guard events (i.e. spectrum 12 in 14 C counting mode) in high channels, Ch 850 upwards. The multichannel analyser collects up to four simultaneous spectra by Boolean logic control (Polach et al., 1984). Rejected events may be recorded and this feature is used below in the estimation of the guard detector efficiency. The count rates above the low energy noise peak were 680, 380 and 120 cpm for Bologna, Turku and Gran Sasso laboratories, respectively. The full count rate was 140 cpm for Gran Sasso. The muon peak count rates were 400, 320 and 1 cpm for Bologna, Turku and Gran Sasso, respectively (Fig. 2). Figure 3 shows cosmic sample coincident events with guard, i.e. the background events, which are removed by the guard. The muon coincidences are clearly seen in high channels, and there is also a dependence on the benzene volume. The muon path length and resulting number of photons is related to the thickness of the benzene layer and its volume. The removal efficiency of the sample events, which were in coincidence with the guard detector, was calculated in the sample spectrum energy range channel by channel as fractional guard efficiency. In 14 C counting mode with pulse amplitude comparator PAC = 1 (inactive): Guard efficiency = 100 ×
SP12 . SP11 + SP12
Radiocarbon measurement at the Gran Sasso National Laboratory
523
Fig. 2. Guard spectra of a Quantulus in Bologna, Turku and Gran Sasso. The distinctive muon peak is in high channels. The low energy component depends on the radioactivity of the building materials and bedrock.
Fig. 3. Sample background events in coincidence with guard recorded for 1, 3, 5 and 7 ml benzene in 9-ml Teflon vial in Bologna (i.e. spectrum 12).
The spectrum SP12 contains the guard rejected sample counts or the events in coincidence with the guard. The spectrum SP11 contains net accepted, or the sample counts in anticoincidence with the guard. The efficiency is compared for background sample events obtained in the surface laboratory (Fig. 4) and in the underground laboratory (Fig. 5). The fraction of sample background events removed by the guard in a surface laboratory (Bologna) was about 89%, while in the Gran Sasso underground laboratory it was only about 16%. Gran Sasso had thus almost zero counts in the high energy cosmic muon range, and very small number of coincidence events in the low energy range which should be removed. The bias threshold has been high in the measurements, cutting the sample signal below channel 100. The Bologna measurements presented in Fig. 4 showed a strange peak at the spectrum beginning, which may be due to high luminescence in channels below 110, caused probably by the epoxy sealing of the Bologna designed vials. The resulting high random coincidence count rate appear as increased guard efficiency.
524
W. Plastino and L. Kaihola
Fig. 4. Fraction of counts removed by the guard in Bologna from 7 ml benzene sample in 9-ml Teflon vial.
Fig. 5. Fraction of counts removed by the guard in Gran Sasso from 7 ml benzene sample in 9-ml Teflon vial.
The guard efficiency in cosmic muon channels 750–1024 was 99.0% in the surface laboratory. The full sample window Ch 1–1024 efficiency was 86.9%, and the efficiency in Ch 1–749 was 82.3%. The drop of guard efficiency below 80% (PAC = 1 or inactive in 14 C counting mode) always means some extra activity in the vial, in the sample (which also includes cocktail) (Kaihola, 1993) and/or in the counting chamber (a contamination or a radon presence). The drop was also observed in the Gran Sasso measurements, where inherent sample phototube activity was a greater proportion to the total background because of the absence of cosmic radiation.
3. Background dependence on environmental gamma-radiation The sample background varies with environmental gamma-background. It is possible to give an estimate of the expected 14 C background when guard continuum count rate is known (Figs. 2 and 6). Sample background variations between laboratories are due to different environmental gamma-fluxes in each laboratory (Kaihola et al., 1986). Backgrounds are stable in each laboratory, when the environmental conditions remain steady.
Radiocarbon measurement at the Gran Sasso National Laboratory
525
Fig. 6. Background 14 C. Quantulus background count rate in optimum 14 C window versus guard rate in 15-ml copper–Teflon vials in 27 laboratories. Sample is benzene with 15 mg/ml butyl-PBD as fluor.
4. Volume dependence of sample background Special vial designs eliminate cross-talk events, as in copper–Teflon vials by Wallac, where the unused volume is replaced by a massive, non-transparent cap and base (Polach et al., 1983). This masked vial design leads to the linear dependency of sample 14 C background on its volume. Figure 7 shows a graph down to 0.3 ml masked vial size with extrapolated 0.05 cpm empty vial background in surface laboratories (Kaihola et al., 1992). When a large 9-ml Teflon vial is used for variable volumes, the background dependence is again linear but additional contribution remains due to the unfilled empty volume, which is source of air scintillations (Kaihola, 1996). The count rate of 9-ml Teflon vials with benzene volumes of 1, 3 and 5 ml related to surface (Bologna) and underground (Gran Sasso) laboratories (Plastino et al., 2001) is shown in Table 1. The surface Bologna laboratory has extrapolated 0.26 cpm background at zero benzene volume in 9 ml unmasked vial, while the underground Gran Sasso has 0.04 cpm (Plastino et al., 2001; Plastino and Kaihola, 2004). A masked vial filled with benzene approaches the same zero volume background in surface laboratories as an unmasked vial with equivalent benzene volumes in Gran Sasso. Absolute background figures in Gran Sasso are about a half of the surface figures of the ideal masked copper–Teflon vials, which have not been tested in Gran Sasso yet.
5. Discussion Excellent performance for liquid scintillation spectrometry of radiocarbon has been achieved underground in the Gran Sasso National Laboratory. Unmasked vials in Gran Sasso show better background than masked vials in surface laboratories – absolute backgrounds are about a half of the surface figures. As the cosmic flux is negligible in the Gran Sasso laboratory, we can attribute the sample background to be fully derived from the instrument’s internal radioactivity and external
526
W. Plastino and L. Kaihola
Fig. 7. 14 C background and removed background dependence on benzene volume in a 9-ml unmasked Teflon vial in Gran Sasso National Laboratory (cpm = 0.0422V + 0.0422); 9-ml unmasked Teflon vial in Radiocarbon Laboratory ENEA–Bologna (cpm = 0.0908V + 0.2598); 3- and 7-ml masked copper–Teflon vials and 0.3-ml Teflon vial in a black masking Delrin adapter in normal and low-level surface laboratories in Turku (cpm = 0.0850V + 0.0474).
Table 1 The count rate of the set A of 9-ml Teflon vials with benzene volumes of 1, 3 and 5 ml related to surface and underground (italic) laboratories (Plastino et al., 2001). The labels L and H indicate background and active samples, respectively. FM = E 2 /B Sample
Count rate (cpm)
Counting uncertainty (cpm)
L1A H1A L1A H1A L3A H3A L3A H3A L5A H5A L5A H5A
0.278 12.949 0.059 12.282 0.398 39.140 0.150 38.235 0.655 65.206 0.235 63.874
0.022 0.148 0.010 0.144 0.026 0.257 0.016 0.254 0.033 0.332 0.020 0.328
∗ BP = Before Present.
Modern activity uncertainty (cpm)
E (%)
FM
FOM
tmax (BP∗ )
8.853
0.088
80.93
25,530
17
48,200
8.540
0.094
76.76
99,730
35
54,000
27.068
0.161
83.28
17,420
43
55,900
26.609
0.166
81.35
44,050
69
59,600
45.101
0.209
83.60
10,680
57
58,000
44.464
0.215
81.89
25,580
92
61,900
Modern activity (cpm)
Radiocarbon measurement at the Gran Sasso National Laboratory
527
gamma-radiation. The gamma-flux in the laboratory is not negligible and radon is present in the water that flows into the laboratory (Plastino and Bella, 2001). The laboratory is well ventilated and stable in temperature (9◦ C). No cooling unit is therefore used in the Quantulus. The instrument is placed in a steel container, which is a good Faraday cage. The results agree with the previous findings that the Compton scattered gamma-radiation measured by the guard detector gives a good indication of environmental radioactivity in the laboratory and will be correlated with the instrument’s performance in beta-particle counting. The role of radon remains to be examined in the Gran Sasso environment, where its presence should be eliminated.
6. Conclusions The difference in the efficiency of the guard detector between surface and underground laboratories, such as the Gran Sasso National Laboratory, is related to the absence of cosmic and associated lower energy radiation in the underground laboratory. The residual radioactivity of phototubes and the role of radon were a larger fraction of the total background signal in Gran Sasso.
Acknowledgements We wish to thank Prof. Eugenio Coccia, Director of the Gran Sasso National Laboratory, for his kind collaboration. The work was funded by EU-LNGS (INFN) (HPRT-CT-2001-00149).
References Hogg, A. (2004). Towards achieving low background levels in routine dating by liquid scintillation spectrometry. Radiocarbon 46, 123–131. Kaihola, L. (1993). Glass vial background reduction in liquid scintillation counting. The Science of the Total Environment 130/131, 297–304. Kaihola, L. (1996). Direct detection of radon gas in air using a liquid scintillation counter. In: Proc. of International Conference on Technologically Enhanced Natural Radiation by Non-Uranium Mining. Central Mining Institute, Katowice, Poland, pp. 169–175. Kaihola, L., Kojola, H., Kananen, R. (1986). Low level liquid scintillation counter performance in a Low Level Surface Laboratory. Nuclear Instruments and Methods in Physics Research B 17, 509–510. Kaihola, L., Kojola, H., Heinonen, A. (1992). A minivial for small sample 14 C dating. Radiocarbon 34, 402–405. Plastino, W., Bella, F. (2001). Radon groundwater monitoring at underground laboratories of Gran Sasso (Italy). Geophysical Research Letters 28, 2675–2678. Plastino, W., Kaihola, L. (2004). Surface and underground ultra low-level liquid scintillation spectrometry. Radiocarbon 46, 97–104. Plastino, W., Kaihola, L., Bartolomei, P., Bella, F. (2001). Cosmic background reduction in the radiocarbon measurements by liquid scintillation spectrometry at the underground laboratory of Gran Sasso. Radiocarbon 43 (2A), 157–161. Polach, H., Gover, J., Kojola, H., Heinonen, A. (1983). An ideal vial and cocktail for low-level scintillation counting: Copper-shielded PTFE (Teflon) and butyl-PBD. In: McQuarrie, S.A., Ediss, C., Wiebe, L.I. (Eds.), Advances in Scintillation Counting. University of Alberta Press, Edmonton, pp. 508–525.
528
W. Plastino and L. Kaihola
Polach, H., Kojola, H., Nurmi, J., Soini, E. (1984). Multiparameter liquid scintillation spectrometry. Nuclear Instruments and Methods in Physics Research B 5, 439–442. Povinec, P. (1994). Underground low-level counting. In: Garcia-Leon, M., Garcia-Tenorio, R. (Eds.), Proc of 3rd International Summer School Low-Level Measurements of Radioactivity in the Environment Techniques and Applications. Huelva, Spain, 1993. World Scientific, Singapore, pp. 113–139.
529
Monte Carlo simulation of the muon-induced background of an anti-Compton gamma-ray spectrometer placed in a surface and underground laboratory Pavol Vojtylaa,* , Pavel P. Povinecb a European Organization for Nuclear Research (CERN), Geneva, Switzerland b International Atomic Energy Agency, Marine Environment Laboratory, MC 98000, Monaco
Abstract Simulations of the cosmic ray muon induced background of a HPGe detector placed inside an anti-Compton shield on the surface and shallow underground is described. Investigation of several model set-ups revealed some useful trends for design of low-level gamma-ray spectrometers. It has been found that the background spectrum of a HPGe detector can be scaled down with the shielding depth. No important difference is observed when the same set-up of the anti-Compton spectrometer is positioned horizontally or vertically. A cosmic-muon rejection factor of at least 40 (at around 1 MeV) can be reached when the anti-Compton suppression is operational. The cosmic-muon background can be reduced to such a level that other background components prevail, like those from the residual contamination of the detector and shield materials and/or from radon, especially for the underground facilities. Keywords: HPGe detector, Anti-Compton spectrometer, Cosmic ray background, Underground laboratory, Monte Carlo simulation, GEANT
1. Introduction The background of low-level HPGe spectrometers placed on the surface or in shallow shielding depths is dominated by the cosmic-muon induced component (Heusser, 1995). Present Monte Carlo detector simulation tools, such as GEANT developed at CERN (1990), are capable of dealing with the simulation of particle passage and detectors in complex systems, and the necessary computing power is available. Hence, there has been an opportunity to examine the influence of various parameters on the detector background, to investigate the trends in model systems or even to compute the cosmic-muon induced background before a low-level system was constructed. * Corresponding author. Address: CERN, Safety Commission, CH-1211 Geneva 23, Switzerland; phone: (+41) 227673893; fax: (+41) 227679360; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08042-3
© 2006 Elsevier Ltd. All rights reserved.
530
P. Vojtyla and P.P. Povinec
Vojtyla et al. (1994), Vojtyla (1995) developed and validated Monte Carlo codes based on GEANT that were suitable for simulating the cosmic-muon background component in typical HPGe gamma-ray spectrometers placed on the surface. The core of the code was a standard cosmic-muon particle generator simulating the actual momentum and direction distributions of cosmic-muons at sea level. With this tool, the physics governing the background induction process could be understood quantitatively and the influence of various parameters on the background, mainly the effects of the shield material and the shield dimensions, could be investigated easily. Lower cosmic-muon background is obtained in tighter shields made of lead lined from the inner side with as thin layers of low-Z materials as possible (Vojtyla, 1996). It was a logical step to extend the models to shallow shielding depths, not deeper than about 50 meters water equivalent (m w.e.) that was done by Vojtyla and Povinec (2000). This work focuses on even more complex systems – anti-Compton spectrometers – and describes investigations carried out for the underground facility of the IAEA’s Marine Environment Laboratory (IAEA-MEL) in Monaco. A model detector was considered and the work aimed at revealing important trends in the background characteristics.
2. Simulation codes 2.1. Surface laboratories The details of the simulation Monte Carlo code for surface laboratories have already been described by Vojtyla (1995), here we recall just some essential points. HPGe detectors are described in GEANT (CERN, 1990) complemented with a standard sea-level cosmic-muon random initial-particle generator. The set-ups are placed entirely in fictive volumes of cylindrical or rectangular shapes, which fit them as closely as possible. Surface areas visible from above are bombarded with cosmic-muons that have suitable momentum, charge and angular distributions. It is assumed that the differential muon flux j (ϑ, p), where ϑ is the zenith angle and p is the muon momentum, can be expressed as j (ϑ, p) = j (0, p) cosn(p) ϑ.
(1)
The data for constructing the functions j (0, p) and n(p) can be found in the literature (Allkofer, 1967, 1975; Allkofer et al., 1971; Allkofer and Grieder, 1984; Moroney and Parry, 1954). It was shown that finer dependencies such as the East–West muon flux asymmetry, the dependence of the muon charge ratio on the muon momentum, latitudinal and seasonal variations can be neglected (Vojtyla, 1995). The simulation processes for the calculation of the HPGe detector background have included (in the order of importance for the background induction): (i) production of δ-electrons by muons, and their bremsstrahlung; (ii) direct electron–positron pair production by muons; (iii) muon decay; (iv) muon bremsstrahlung; (v) direct ionization by muons in the sensitive volume of the HPGe detector; (vi) others.
Monte Carlo simulation of the muon-induced background of an anti-Compton gamma-ray spectrometer
531
2.2. Underground laboratories The energy losses of muons below about 500 GeV c−1 are mostly due to quasi-continuous ionization. For muons of higher energies, discrete processes dominate (bremsstrahlung, fast muon nuclear interactions). The transport of muons in the lower energy range is simpler, as the muons lose their energy almost continuously and they do not change their directions significantly. In the higher energy range, significant fluctuations of both energy and direction can occur. If the prevailing part of the muon spectrum observed underground stems mostly from muons with original energies below a few hundreds of GeV, the attenuation of the muon flux can be taken into account simply by accounting for mean ranges and integrated energy losses of muons. Deep underground, where the muon spectrum originates from sea-level muons of 1 TeV c−1 or more, this approach is not possible. The case of shallow shielding depths can be treated therefore in a simpler way. For generic set-ups when the accurate topography and the material definition of the overburden rock are not specified, we have to limit ourselves only to coarse estimates of the background. This allows us to accept two approximations: (i) there are no fluctuations in muon ranges and energy losses in the rock; (ii) there is no deviation from the original muon direction due to multiple scattering to small angles. Although the effects (i) and (ii) take place in reality, they can be ignored in calculations of the muon transport underground, as they are partly smoothed-out due to continuous energy and angle spectra of cosmic-muons. The muon generator in the codes for the simulating muon-induced background under an overburden is similar to that at the sea level, but the muon rejection and momentum reduction algorithms are applied. After the generation of the initial kinematics of the muon has been completed, the range of a muon in rock is calculated and compared with the slanted depth of the counting system X = D/ cos ϑ, where D is the vertical depth and ϑ is the muon zenith angle. If the range is shorter than the slanted depth, the event is rejected. Otherwise the original muon momentum p0 is reduced according to the formula pred = p0 − R −1 (X),
(2)
where R −1 (X) is the inverse function of a function describing range of muons in dependence on the muon momentum R(p). It is assumed that the rock has the ‘standard rock’ composition with the density of 2.65 g cm−3 , the average mass number A = 22 and the average atomic number Z = 11 (Heusser, 1995). The transport of muons through the standard rock was simulated using GEANT and data on mean ranges for various muon momenta have been in agreement with observations (Vojtyla and Povinec, 2000). The rejection method is not time consuming, as most of the total computing time is spent for particle tracking. It wastes only random number generator. For example, for an overburden of 30 m w.e., about 80% of muons are rejected. The method is in principle applicable to any complex shapes above the underground laboratory, such as hilly terrain or buildings, however, the rejection routine becomes more complicated. The assumed muon momentum range of 0.2–100 GeV c−1 is sufficient for shallow shielding depths.
532
P. Vojtyla and P.P. Povinec
3. Model detector system An n-type coaxial 100% relative efficiency (in comparison with 7.5 cm × 7.5 cm NaI(Tl) detector) HPGe detector was assumed to be surrounded by a model NaI(Tl) anti-Compton shield and the whole assembly was placed inside a passive shield on the surface or 30 m w.e. underground. The HPGe detector corresponded to a real design. The anti-Compton shield consisted of an annular NaI(Tl) crystal with the inner diameter of 11 cm, the outer diameter of 30 cm and the height of 40 cm, closed from one side by another cylindrical NaI(Tl) crystal 7.5 cm diameter × 7.5 cm height. The whole detector set-up was placed inside a cylindrical shield made of 15 cm thick lead with a diameter of 70 cm and length of 90 cm. The HPGe detector head was so deep in the anti-Compton shield that only free space of 11 cm in diameter and 5 cm in height was allowed for samples. Obviously, the anti-Compton shield reduces not only the Compton continuum but also the cosmic-muon induced background. The reduction is very efficient due to the high detection efficiency of NaI(Tl) crystals for various radiation types. Practically there are two possibilities to position the whole anti-Compton system: either vertically (the NaI(Tl) stopcock is on the top), or horizontally. A drawing of the set-up positioned horizontally is shown in Fig. 1. It cannot be assessed simply what configuration would be more efficient in respect to the background reduction by the anti-Compton shield. Therefore simulations of the vertical and horizontal set-ups using the developed code were carried out. The code for the horizontal set-up has a more complicated muon generator than the standard codes described earlier. A box containing the passive shield is bombarded by cosmic muons instead of horizontally positioned cylinder, allowing to use the already written and debugged functions for kinematics distributions. Because the basis of the box is not a square, bombardment of sides of different areas has to be correctly sampled so that the impact rates are proportional to the respective areas.
Fig. 1. Drawing of the anti-Compton set-up positioned horizontally.
Monte Carlo simulation of the muon-induced background of an anti-Compton gamma-ray spectrometer
533
4. Results and discussion The results for the vertical and horizontal anti-Compton shield arrangements at sea level and underground are summarized in Tables 1 and 2, respectively. The corresponding lower energy spectra are shown in Figs. 2 and 3. The simulated background spectra 30 m w.e. underground show the same features as those at sea level, but all absolute background characteristics are reduced by a general factor of about 3 that was observed also in passive set-ups in the previous work of Vojtyla and Povinec (2000). The background spectra in the energy region up to 100 MeV are shown in Fig. 4. A striking feature of the background spectra is that the difference between the horizontal and vertical arrangements is negligible in the lower energy region. This can be caused by equilibration of the lower-energy photon flux inside the shield cavity and anti-Compton NaI(Tl) crystals. The vertical arrangement seems to be slightly better than the horizontal one below about 1000 keV. A mistake during writing or running the codes is excluded because Table 1 Background characteristics of HPGe detectors with anti-Compton shields in horizontal and vertical positions simulated at sea level
Anti-Compton off 0–3000 keV (min−1 ) 511 keV (day−1 )
Horizontal
Vertical
89.5 ± 0.1 3090 ± 30
90.1 ± 0.1 3010 ± 30
Anti-Compton on 0–3000 keV (min−1 ) 511 keV (day−1 )
1.04 ± 0.01 31 ± 3
Reduction factor 0–3000 keV 511 keV
86 ± 1 99 ± 10
0.914 ± 0.012 25 ± 2 99 ± 1 119 ± 12
Table 2 Background characteristics of HPGe systems with anti-Compton shields in the horizontal and vertical positions simulated 30 m w.e. underground
Anti-Compton off 0–3000 keV (min−1 ) 511 keV (day−1 ) Anti-Compton on 0–3000 keV (min−1 ) 511 keV (day−1 ) Reduction factor 0–3000 keV 511 keV
Horizontal
Vertical
29.7 ± 0.1 1050 ± 30
30.1 ± 0.1 1030 ± 30
0.33 ± 0.01 8.7 ± 2.5 90 ± 3 120 ± 35
0.29 ± 0.01 5.1 ± 1.9 102 ± 4 200 ± 70
534
P. Vojtyla and P.P. Povinec
Fig. 2. Lower-energy background spectra of the anti-Compton gamma-ray spectrometer in the horizontal and vertical positions at sea level.
Fig. 3. Lower-energy background spectra of the anti-Compton gamma-ray spectrometer in the horizontal and vertical positions 30 m w.e. underground.
a comparison of the spectra in the higher energy region in Fig. 4 shows that the spectra differ significantly in the region above ∼10 MeV where the contribution of the secondary processes in the shields is negligible, and the direct ionization by muons prevails. The wide peaks have maxima at different energies according to the track distributions inside the sensitive volume of the HPGe diode. With the anti-Compton rejection off, the maximum is reached at 55 MeV for the horizontal position, corresponding to the most probable track length of 7.5 cm (7.3 MeV/cm ionization losses of minimum ionizing particles in Ge (Montanet et al., 1994), which is close to the crystal diameter of 7.6 cm. The muons moving in directions close to the vertical are likely to move along the HPGe crystal diameter. On the other hand, the maximum for the vertical position is reached at about 63 MeV corresponding to the track length of 8.6 cm, which is close to the height of the HPGe crystal of 8.2 cm. The muons move
Monte Carlo simulation of the muon-induced background of an anti-Compton gamma-ray spectrometer
535
Fig. 4. Higher-energy background spectra of the anti-Compton gamma-ray spectrometer in the horizontal and vertical positions at sea level.
more likely through the crystal height in this case. The apparent peak at 61 MeV (8.4 cm) observed for the vertical position with the anti-Compton rejection switched on stems from the “leakage” of muons through the space between the annular NaI(Tl) crystal and the stopcock NaI(Tl) crystal. A large number of muons must move in the direction parallel to the axis of the set-up to induce such a peak. This condition is well fulfilled in the vertical arrangement of the set-up when many low-energy muons move along the vertical direction. Note that the presence of thick layers of NaI close to the detector crystal has a detrimental effect on the background continuum if the anti-Compton rejection is switched off. Although the arrangement could be considered as a small shield, the maximum in the continuum reaches about 200 day−1 keV−1 on the surface to be compared, for example, with about 100 day−1 keV−1 (at around 200 keV) for a small cylindrical descending-Z lining shield (Vojtyla and Povinec, 2000), despite of the relatively high effective Z and A of NaI (46.6 and 111, respectively). But this also implies that a thin housing of the NaI crystals, which was not assumed in the model set-up, would not have a considerable effect on the results. The background reduction factor by the anti-Compton rejection depends on the energy in the spectrum. Figure 5 shows the background reduction factors calculated for various energies up to 3000 keV. It reaches values of about 200 for the lowest energies (50–100 keV) and decreases down to about 40 at 1500 keV. Then it slowly increases up to about 100 at 3000 keV. 5. Conclusions Monte Carlo simulations of cosmic-muon induced background in HPGe counting systems using strong background reduction techniques (underground, anti-Compton shield) are feasible at present. Simulations of a model HPGe detector placed inside an anti-Compton shield on the surface and in shallow underground revealed some trends useful in design of such systems. First the scaling practice was confirmed. Given the detector set-up, its background spectra induced by cosmic-muons can be scaled down by a factor corresponding to the shielding depth.
536
P. Vojtyla and P.P. Povinec
Fig. 5. Dependence of the background induction factors of the anti-Compton suppression on the energy for the horizontal and vertical positions at sea level.
No important difference was observed when the same set-up was positioned horizontally or vertically. An NaI(Tl) anti-Compton gamma-ray spectrometer inside a passive shield has a detrimental effect on the cosmic-muon induced background if the anti-Compton rejection is switched off because of a thick layer of a relatively low-Z material, compared with standard shielding materials like lead. On the other hand, a cosmic-muon rejection factor of at least 40 (at around 1 MeV) can be reached when the anti-Compton rejection is operational. In such situation, the cosmic-muon background is reduced to such a level that other background components prevail, like those from the residual contamination of the detector and anti-Compton shield materials, or from radon, especially for the underground facilities. However, such high cosmic-muon rejection factors can be reached only if construction materials with negligible radionuclide contamination have been used for the construction of the HPGe detector, its cryostat, the NaI(Tl) detectors and the passive shielding.
Acknowledgement The IAEA is grateful for the support provided to its Marine Environment Laboratory by the Government of the Principality of Monaco.
References Allkofer, O.C. (1967). Die Höhen- und Zenitwinkelabhängigkeit der Energiespektren von Utrastrahlungsteilchen. Fortschritte der Physik 15, 113–196. Allkofer, O.C. (1975). Introduction to Cosmic Radiation. Thiemig, München. Allkofer, O.C., Grieder, P.K.F. (1984). Cosmic rays on Earth. Fachinformationszentrum, Karlsruhe.
Monte Carlo simulation of the muon-induced background of an anti-Compton gamma-ray spectrometer
537
Allkofer, O.C., Karstensen, K., Dau, W.D. (1971). The absolute cosmic ray muon spectrum at sea level. Physics Letters B 36, 425–427. CERN (1990). GEANT Detector Description and Simulation Tool. CERN Program Library Office, CERN, Geneva, Switzerland. Heusser, G. (1995). Low-radioactivity background techniques. Annual Review of Nuclear and Particle Science 45, 543–590. Montanet, L., et al. (1994). Review of particle properties. Physical Review D 50, 1173. Moroney, I.R., Parry, I.K. (1954). Momentum distribution and charge ratio of μ-mesons at zenith angles in the east– west plane. Australian Journal of Physics 7, 423–428. Vojtyla, P. (1995). A computer simulation of the cosmic-muon background induction in a Ge γ -spectrometer using GEANT. Nuclear Instruments & Methods in Physics Research B 100, 87–97. Vojtyla, P. (1996). Influence of shield parameters on cosmic-muon induced background of Ge γ -spectrometers. Nuclear Instruments & Methods in Physics Research B 111, 163–170. Vojtyla, P., Povinec, P.P. (2000). A Monte Carlo simulation of background characteristics of low-level HPGe detectors. Applied Radiation and Isotopes 53, 185–190. Vojtyla, P., Beer, J., Št’avina, P. (1994). Experimental and simulated cosmic muon induced background of a Ge spectrometer equipped with a top side anticoincidence proportional chamber. Nuclear Instruments & Methods in Physics Research B 86, 380–386.
538
IAEA-MEL’s underground counting laboratory – The design and main characteristics P.P. Povinec* , J.-F. Comanducci, I. Levy-Palomo, F. Avaullee Marine Environment Laboratory, International Atomic Energy Agency, Monaco Abstract An underground counting laboratory designed for low-level gamma- and beta-ray spectrometry of anthropogenic and natural radionuclides in environmental samples has been constructed at the depth of 35 m water equivalent (w.e.) at the IAEA’s Marine Environment Laboratory in Monaco. The laboratory is equipped with 4 large volume HPGe detectors placed in a common lead shield with anti-cosmic plastic scintillator shielding, a coincidence–anticoincidence gammaray spectrometer, comprising an HPGe detector and NaI(Tl) shielding, a HPGe–HPGe telescope (sandwich) detector system, and a Quantulus liquid scintillation spectrometer. The laboratory design is presented and performance of the gamma-ray spectrometers under different coincidence–anticoincidence modes of operation are discussed. The anticosmic shielding has proved to be a valuable investment as the backgrounds of HPGe detectors operating at shallow depths are comparable with backgrounds of detectors placed in laboratories operating at moderate depths of about 250 m w.e. Keywords: Underground laboratory, Gamma-ray spectrometry, HPGe detector, NaI(Tl) detector, Si beta-ray detector, Plastic scintillation detector, Coincidence–anticoincidence spectrometer, Anti-Compton spectrometer, Background
1. Introduction The most important recent development in the radiometric methods of analysis of radionuclides in the environment has been the availability of large volume HPGe detectors for low-level gamma-ray spectrometry (about 200% relative efficiency compared to a 76×76 mm2 NaI(Tl) detector). The large volume and mass of HPGe detectors require to apply specific shielding arrangements, as the interactions of cosmic ray particles with the detectors affect considerably their background characteristics. An anti-cosmic shielding and/or underground operation of detectors has become therefore important for their optimum utilization for analysis of environmental radionuclides (Heusser, 1994, 1995; Povinec, 1994, 2004; Vojtyla et al., 1994; Reyss et al., 1995; Niese et al., 1998; Neder et al., 2000; Neumaier et al., 2000; Vojtyla and Povinec, 2000; Semkow et al., 2002; Povinec et al., 2004; Schroettnert et al.). * Corresponding author. Address: IAEA-MEL, 4 Quai Antoine 1er, MC-98000, Monaco; phone: (+377) 97977272; fax: (+377) 97977273; e-mail:
[email protected],
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08043-5
© 2006 Elsevier Ltd. All rights reserved.
IAEA-MEL’s underground counting laboratory
539
The main background components of a typical low-level HPGe detector, not situated deep underground, are: • cosmic radiation, represented by secondary cosmic ray particles such as muons, neutrons, protons, electrons, positrons and photons; • natural radionuclides present in the detector and construction materials; • radon and its progenies; • activation products formed by interactions of cosmic nucleons with the detector and surrounding construction materials; • anthropogenic radionuclides present as contaminants in the construction materials. For a present-day, carefully designed low-level HPGe spectrometer using construction materials with minimum radioactive contamination, the dominating background component is cosmic radiation, mainly cosmic muons (Heusser, 1995; Theodorsson, 1996; Vojtyla and Povinec, 2000; Laubenstein et al., 2004; Povinec, 2004). Therefore a low-level gamma-ray spectrometer with an anti-cosmic shielding will greatly reduce its background. The anti-cosmic shielding is usually made of gas or scintillation detectors, which surround the HPGe detector (Heusser, 1994, 1995; Vojtyla et al., 1994; Semkow et al., 2002; Povinec et al., 2004; Schroettnert et al.). Another possibility is to use an anti-Compton spectrometer, which is a powerful tool for reducing the detector’s background as it combines both anti-cosmic and anti-Compton background suppression (e.g. Debertin and Helmer, 1998; Povinec, 1982; Zvara et al., 1994; Povinec et al., 2004). In the present paper, a brief description of the construction features of the International Atomic Energy Agency Marine Environmental Laboratory’s (IAEA-MEL) underground Counting lAboratory for enVironmental radionuclidEs (CAVE) is given and the main counting characteristics of HPGe gamma-ray spectrometers are presented and discussed.
2. The CAVE laboratory The CAVE laboratory is situated in an underground cellar in a car parking area at a depth of 35 m of water equivalent (w.e.). At such a shallow depth the cosmic muon flux is reduced only by a factor of 6. However, the nucleonic component of cosmic rays has been attenuated by more than 4 orders of magnitude. The laboratory is situated about 300 m from the main IAEA-MEL premises, with easy access by car for the weekly supply of liquid nitrogen. The operation of the laboratory is very convenient as it requires little extra effort. The laboratory is equipped with: (i) A ventilation and air conditioning system maintaining overpressure, stable humidity and temperature levels in the laboratory. (ii) A common lead shielding housing four large volume HPGe detectors (Fig. 1). (iii) An anti-cosmic shielding made of plastic scintillation detectors, surrounding the lead shielding, which protects the HPGe detectors against cosmic radiation (Fig. 2). (iv) A coincidence–anticoincidence gamma-ray spectrometer, comprising an n-type HPGe detector, NaI(Tl) detector, Si beta-ray detector and NaI(Tl) shielding.
540
P.P. Povinec et al.
Fig. 1. Construction of the lead shield and the anti-cosmic plastic scintillator shielding for HPGe detectors.
(v) A telescope (sandwich) HPGe–HPGe gamma-ray spectrometer working in single or antiCompton mode. (vi) A low-background liquid scintillation spectrometer (Quantulus 1220), mainly used for 3 H and 241 Pu analysis of marine samples. 2.1. HPGe detectors with anti-cosmic shielding 2.1.1. Lead shielding with plastic scintillation detectors The lead shield is comprised of two layers. The outer layer is made of low activity lead 7.5 cm thick, and the internal layer, which is 5 cm thick, is made of very low activity lead which was specially ordered for the underground laboratory (the 210 Pb massic activity is below 0.1 Bq kg−1 ). The construction features of the lead shielding with HPGe detectors are shown in Fig. 1. The lead shielding is surrounded on sides and from the top by plastic scintillation sheets 7 cm thick, which are viewed by 5 cm diameter photomultipliers (Figs. 1 and 2). The plastic scintillation detectors are connected with the HPGe detectors in anticoincidence. They work
IAEA-MEL’s underground counting laboratory
541
Fig. 2. HPGe gamma-ray spectrometers with anti-cosmic shielding and the electronics.
as a guard for HPGe detectors, eliminating effects of nucleonic component, and partially reducing the effects of the muonic component of cosmic rays with their secondaries on the detectors background. Such a multi-detector arrangement of HPGe detectors situated in a common lead shield protected against cosmic radiation with a guard detector is described in the literature for the first time. During all measurements in the CAVE laboratory radon is expelled from the detector chambers by the evaporation of nitrogen from the detector’s Dewar containers (Fig. 2), thus keeping stable background during measurements (Sykora et al., 1992). The CAVE laboratory is used only for the analysis of low activity samples. Samples of higher activities are measured in the counting laboratory situated on the 3rd floor at IAEAMEL. 2.1.2. HPGe detectors All HPGe detectors (coaxial p-type) were specifically designed for low-level gamma-ray spectrometry in an underground laboratory. They are of U-type with preamplifiers housed outside of the lead shield. However, the Field Effect Transistor (FET) is mounted on a Cu plate connected to the cold-finger. Only materials with minimum radionuclide contamination were used for the detector construction. Four types of HPGe detectors are housed in the lead shield with the anti-cosmic guard: (i) A 100% relative efficiency coaxial detector (CANBERRA) with cryostat made of electrolytic copper with carbon window. The detector resolution at 1.33 MeV (60 C line) is 2.09 keV, the peak/Compton ratio is 80. (ii) A 150% relative efficiency well-type detector (CANBERRA) with cryostat made of electrolytic copper (well dimensions – 32.5 mm diameter, depth 73 mm), well walls made of
542
P.P. Povinec et al.
thin copper. The detector resolution at 1.33 MeV is 2.48 keV, the peak/Compton ratio is 51. (iii) A 170% relative efficiency coaxial detector (EURISYS) with cryostat and window made of pure aluminium. The detector resolution at 1.33 MeV is 2.08 keV, the peak/Compton ratio is 101. (iv) A 200% relative efficiency well-type detector (CANBERRA) with cryostat made of electrolytic copper (well dimensions – 25 mm diameter, depth 60 mm), well walls made of thin copper. The detector resolution at 1.33 MeV is 2.31 keV, the peak/Compton ratio is 120. 2.1.3. Electronics for HPGe detectors with anti-cosmic shielding ORTEC NIM modular electronics with ORTEC software has been used for signal processing, data acquisition and spectra evaluation. The block schema of the electronics used for the HPGe detectors with anti-cosmic shielding is presented in Fig. 3. The signals after amplification are fed separately for timing and energy evaluation. Each HPGe detector is connected in anticoincidence with the plastic scintillation guard detector, therefore events caused by penetrating cosmic muons which have been registered in the guard detector are excluded from gamma-ray spectra of the HPGe detectors. Time amplitude converters have been used for time analysis of signals from the HPGe detectors and the plastic scintillation detectors. Using proper timing characteristics of pulses (delay and width), it has been possible to reach high
Fig. 3. Block schema of the electronics for the HPGe gamma-ray spectrometers with anti-cosmic shielding.
IAEA-MEL’s underground counting laboratory
543
efficiency of the guard plastic scintillation detector with negligible losses in the counting rate of HPGe detectors. 2.1.4. Coincidence–anticoincidence gamma-ray spectrometer The coincidence–anticoincidence gamma-ray spectrometer (Fig. 4) has as the main gammaray detector an n-type HPGe detector (ORTEC) of 100% relative efficiency (the resolution at 1.33 MeV is 2.35 keV, the peak/Compton ratio is 64). The detector arrangement is U-type with a preamplifier situated outside of the lead shielding, however the FET is mounted on Cu plate connected with cooling finger. The detector cryostat is made of electrolytic copper and the window is made of high purity aluminium. The HPGe detector is surrounded by NaI(Tl) shielding (annulus of 30 cm in diameter and 40 cm long; top NaI(Tl) detector is of 7.6 cm in diameter and 7.6 cm long). A double side Si beta-ray detector (ORTEC) of 450 mm2 active area and of 96 µm sensitive thickness, located between the HPGe detector and the top NaI(Tl) detector, has been used in beta–gamma coincidence studies as well. All detectors are housed in a shield of 10 cm thick, made of ordinary lead. The coincidence– anticoincidence gamma-ray spectrometer is a versatile detection system that can operate in several detection modes: (i) A single HPGe or NaI(Tl) gamma-ray spectrometer. (ii) An anti-Compton gamma-ray spectrometer when the HPGe detector is connected in anticoincidence with the NaI(Tl) detectors, thus suppressing cosmic ray events registered
Fig. 4. Construction of the anti-Compton HPGe gamma-ray spectrometer.
544
P.P. Povinec et al.
by the NaI(Tl) detectors, as well as Compton scattered gamma-quanta registered by both the HPGe and NaI(Tl) detectors. (iii) A double coincidence gamma-ray spectrometer – gamma–gamma coincidences between the HPGe detector and the top NaI(Tl) detector; the same with the NaI(Tl) annulus either in coincidence or anticoincidence (e.g. in the case of analysis of 60 Co) (Stanicek and Povinec, 1986). (iv) A triple coincidence gamma-ray spectrometer – all three detectors in coincidence (e.g. in the case of analysis of positron emitters (22 Na and others) when the annihilation quanta are registered by the HPGe and the top NaI(Tl) detector, and the characteristic gammarays by the NaI(Tl) annulus) (e.g. Sykora and Povinec, 1986). (v) A sum coincidence gamma-ray spectrometer – coincidence summing of cascade (or annihilation) gamma-rays, enabling to “clean” the gamma-ray spectrum by registration of the full energy absorption peaks only. Because of higher efficiency and lower background, free of any interferences, this arrangement enables to reach best detection limits for emitters with cascade or annihilation photons in large volume samples (Povinec, 1982; Sykora and Povinec, 1990). (vi) A beta–gamma coincidence spectrometer – coincidences between the beta-ray (Si or a gas detector) and HPGe/NaI(Tl) detectors are registered. This spectrometer is suitable for analysis of small volume samples, which are mounted on both sides of the beta-ray detector (Hlinka et al., 1977). (vii) A beta–gamma–gamma coincidence spectrometer – coincidences between the Si, HPGe and the top NaI(Tl) detector. This arrangement, because of the extremely low background, enables to reach best detection limits for beta–gamma(–gamma) emitters in small volume samples. The electronics (ORTEC) used for the anti-Compton spectrometer and various coincidence– anticoincidence modes is very complex and has already been described (Povinec et al., 2004). The arrangement of the detectors and electronics in the CAVE laboratory is shown in Fig. 5.
Fig. 5. A general view of the CAVE laboratory – anti-Compton HPGe spectrometer with electronics (left); liquid scintillation spectrometer, Quantulus (right); 4 × HPGe spectrometers with anti-cosmic shielding and electronics (back).
IAEA-MEL’s underground counting laboratory
545
3. Results and discussion 3.1. HPGe detectors with anti-cosmic shielding The background characteristics of the HPGe detectors used in the CAVE laboratory operating with anti-cosmic shielding are compared in Fig. 6. It can be seen that the lowest background per kg of Ge has been reached with the 150% detector. On the contrary, the 170% detector with “pure” aluminium cap has a relatively higher background due to its contamination by natural radionuclides. Several peaks from natural radionuclides present in construction materials can be identified in the spectra, e.g. 40 K (1.46 MeV) and 214 Bi (609 keV). The best visible is 40 K, and its contribution to the background of 100%, 150%, 170% and 200% detectors is 7.0, 1.4, 5.6 and 5.2%, respectively. The highest background observed in the 170% detectors is, however, not due to 40 K but to contamination with low-energy gamma-ray emitters. The construction materials should be therefore carefully tested by the manufacturing companies for the presence of radionuclides before the detector construction, especially for the detector’s cryostat and its window. It can be seen from Fig. 6 that the annihilation peak at 511 keV is still visible in all background spectra obtained with active anti-cosmic shielding. This peak is produced by the annihilation of electrons and positrons in the shield and in the detector itself. Although one would expect that its suppression by the anti-cosmic shield could depend on the detector’s volume, this has not been, however, confirmed. The range of observed suppression factors from 4 (for the 200% detector) to 10 (for the 170% detector) does not confirm this assumption. The highest contribution to the total detector’s background from the annihilation peak has been observed for the 200% detector (6.9%), 150% detector (4.9%), 100% (3.5%) and 170% (2.6%). However, this has to be taken in relation with contamination of detectors, as the largest radionuclide contamination has been observed for the 170 and 100% detectors. The second important finding from Fig. 6 is a large background continuum, which has not been completely removed by the anti-cosmic shielding. This background continuum, having the maximum at around 200 keV, is produced by the interactions of cosmic muons with the lead shield (Vojtyla and Povinec, 2000). When comparing the integral backgrounds of HPGe detectors with the anti-cosmic shielding (from 40 to 2700 keV), divided by the mass of the Ge crystal, the lowest background has been obtained for the 150% well detector (35 h−1 kg−1 ), then for the 200% well (42 h−1 kg−1 ), 100% (57 h−1 kg−1 ) and 170% (68 h−1 kg−1 ) detectors. The total background of HPGe detectors in the energy interval 40–2700 keV has decreased with the anticosmic shielding by a factor of 6, 8, 4 and 5, for the 100%, 150%, 170% and 200% detectors, respectively. The lowest suppression factor observed again for the 170% detector has been caused by its highest contamination by low energy gamma-ray emitters. Laubenstein et al. (2004) compared background of HPGe detectors operating in different underground laboratories. When comparing the integral background of HPGe detectors with the anti-cosmic shielding operating in the CAVE laboratory at 35 m w.e. (Povinec et al., 2004), it has been found that their background is comparable with background of HPGe detectors operating in underground laboratories at around 250 m w.e. depths (with passive shielding only) (Pagava et al., 1992; Neumaier et al., 2000; Komura and Hamajima, 2004). Therefore,
546
P.P. Povinec et al.
Fig. 6. Background spectra of the HPGe detectors measured with active anti-cosmic shielding (counting time ∼60,000 s).
IAEA-MEL’s underground counting laboratory
547
an anti-cosmic shielding in an underground laboratory operating at a shallow depth is very important for reducing the detectors background and should be widely used. 3.2. Coincidence–anticoincidence gamma-ray spectrometry 3.2.1. Anti-Compton gamma-ray spectrometer The characteristics of the anti-Compton gamma-ray spectrometer were investigated using 65 Zn and 137 Cs sources. The anti-Compton suppression in both cases is documented in Fig. 7. While in the direct mode (single gamma-ray spectrometer) the peak to Compton ratio is 86 and 78 for 65 Zn and 137 Cs, respectively, in the anticoincidence mode the peak/Compton ratios have increased to 960 and 610, representing an improvement by a factor of 11.5 and 7.8, respectively. Figure 8 shows the Compton suppression of the background of the 100% HPGe detector. The background in the energy interval 30–2500 keV has decreased from 0.5 s−1 to 0.1 s−1 with anti-Compton shielding. However, when comparing the background spectra of the HPGe detectors with anti-cosmic shielding (Fig. 6) with the background spectrum of the anti-Compton spectrometer (Fig. 8) it can be noticed that the 100% HPGe detector has in the background spectrum several gamma-lines from natural radionuclides, probably due to contamination of the ordinary lead shielding, as the internal low activity lead shielding has not been installed yet. 3.2.2. Coincidence gamma-ray spectrometer The performance of the coincidence gamma-ray spectrometer for a mixture of radionuclides with a single photon emission (137 Cs) and with two cascade photons (60 Co) operating in different modes is plotted in Fig. 9. In the single gamma-ray spectrum of 137 Cs and 60 Co, the peak/Compton ratio for 60 Co is 56. The gamma-ray spectrum with anticoincidence shielding shows a suppression of cascade gamma-quanta of 1.17 and 1.33 MeV (as expected), however, the peak/Compton ratio has increased to 144. For the gamma-ray emitters with cascade photons (such as 60 Co), the anticoincidence mode is not advantageous because of loses due to registration of scattered cascade photons by the NaI(Tl) annulus, causing a lower detection efficiency. The gamma–gamma coincidence spectrum measured with the active anticoincidence shielding shows a clean spectrum without presence of 137 Cs, with peak/Compton ratio of 244. The integral background measured in this mode is negligible (∼2 × 10−4 s−1 ). The coincidence gamma-ray spectrometry (i.e. a double gamma, a triple gamma, and a sum coincidence spectrometer) is advantageous to use for analysis of cascade photon emitters when there is no restriction of sample size. However, for a limited sample size (down to a mg scale), it is more advantageous to use beta–gamma, or beta–gamma–gamma coincidence spectrometry, if there is a suitable gamma-ray emitter to be analyzed (Hlinka et al., 1977; Sykora and Povinec, 1990). Figure 10 compares the gamma-ray spectra of a source with a 137 Cs and 60 Co mixture. In the beta–gamma–gamma mode the 60 Co peaks became dominant and the 137 Cs peak almost disappeared. The background under the 1.33 MeV peak decreased from 48 × 10−2 s−1 (direct mode) to 3.05 × 10−2 s−1 (without window setting), then to 0.7 × 10−2 s−1 with 1.17 + 1.33 MeV window in the NaI(Tl) line, and finally to 1 count per day in the full energy region. Of course, there is also a big decrease in the detection efficiency, from 1.3% in the direct
548
P.P. Povinec et al.
Fig. 7. 65 Zn and 137 Cs gamma-ray spectra without and with the anti-Compton suppression.
IAEA-MEL’s underground counting laboratory
549
Fig. 8. Compton background suppression of the 100% HPGe detector (counting time ∼60,000 s).
mode to 0.002% in the beta–gamma–gamma mode (for 60 Co). However, the factor of merit √ 60 for the detection of Co (the ratio E/ B, where E is the detection efficiency and B is the background) has increased by about a factor of 20. The coincidence–anticoincidence gamma-ray spectrometry has decreased the detection limits for most of the radionuclides down to 1 mBq. 3.2.3. HPGe telescope as an anti-Compton gamma-ray spectrometer The telescope (sandwich) gamma-ray spectrometer consists of a planar HPGe detector (20 mm thickness, 1.73 keV resolution at 1.33 MeV), housed together with a HPGe coaxial detector (60 mm thickness, 1.76 keV resolution at 1.33 MeV) in the same cryostat. Both detectors use their individual preamplifiers so they can operate in different modes as: (i) A single planar gamma-ray spectrometer – only the planar detector is used for registration of low-energy gamma-ray emitters. (ii) An anti-Compton spectrometer – a more sophisticated arrangement when the planar and the coaxial detectors are working in anticoincidence, discriminating thus gamma-rays scattered from the planar detector which are registered in the coaxial detector. (iii) A coincidence gamma–gamma spectrometer – the planar and the coaxial detectors are working in coincidence. Figure 11 compares a direct background measurement of the planar detector with that of the HPGe–HPGe anti-Compton spectrometer placed in the lead shield. The advantage of the telescope detector for analysis of low-energy gamma-ray emitters is clearly visible, as the background in the low-energy part of the spectrum has decreased by about a factor of five. The telescope gamma-ray spectrometer has been designed for the analysis of low-energy gamma-ray and/or X-ray emitters (such as 210 Pb, U and Pu isotopes, 241 Am and others) in environmental samples. Its main use is in the environmental research (e.g. 210 Pb dating of
550
P.P. Povinec et al.
Fig. 9. 137 Cs and 60 Co gamma-ray spectra without and with the anti-Compton suppression, and in gamma–gamma coincidence mode.
IAEA-MEL’s underground counting laboratory
551
Fig. 10. A single (direct) gamma-ray spectrum of a source with 137 Cs and 60 Co mixture (top), and a beta–gamma–gamma coincidence spectrum (bottom).
Fig. 11. A background spectrum of the telescope HPGe–HPGe spectrometer operating as a single spectrometer (top), and as an anti-Compton spectrometer (bottom).
552
P.P. Povinec et al.
sediments), with possible applications in emergency situations when screening of large number of samples for the presence of e.g. U and Pu isotopes would be required.
4. Conclusions The CAVE underground counting laboratory with anti-cosmic shielding of detectors has permitted us to effectively decrease the background of large volume HPGe detectors, especially when operating in anti-cosmic, anti-Compton and/or coincidence modes. The anti-cosmic shielding has proved to be a valuable investment as the background of HPGe detectors operating at shallow depths is comparable with underground laboratories operating at moderate depths of about 250 m w.e. Coincidence–anticoincidence modes of operation (including the anti-Compton spectrometer) have further improved the detection capabilities of the low-level gamma-ray spectrometry, reaching detection limits of the order of 1 mBq.
Acknowledgements Financial support provided by the Government of Japan for the construction of the underground laboratory is highly acknowledged. The Agency is grateful for the support provided to its Marine Environment Laboratory by the Government of the Principality of Monaco.
References Debertin, K., Helmer, R.G. (1998). Gamma and X-Ray Spectrometry with Semiconductor Detectors. Elsevier/NorthHolland, Amsterdam. Heusser, G. (1994). Background in ionising radiation detection. In: Garcia-Leon, M., Garcia-Tenorio, R. (Eds.), Low-Level Measurements of Radioactivity in the Environment. World Scientific, Singapore, pp. 69–112. Heusser, G. (1995). Low-radioactivity background techniques. Annual Review of Nuclear and Particle Science 45, 543–590. Hlinka, V., Usacev, S., Povinec, P., Chudý, M. (1977). Die Untersuchung der Koinzidenzmethoden bei der Identifikation einiger Radioisotopen. Acta Physica Universitatis Comeniae 18, 109–116. Komura, K., Hamajima, Y. (2004). Y. Ogoya underground laboratory for the measurement of extremely low levels of environmental radioactivity: Review of recent projects carried out at OUL. Applied Radiation and Isotopes 61, 185–190. Laubenstein, M., Hult, M., Gasparro, J., Arnold, D., Neumaier, S., Heusser, G., Kohler, M., Povinec, P., Reyss, J.-L., Schwaiger, M., Theodorsson, P. (2004). Underground measurements of radioactivity. Applied Radiation and Isotopes 61, 167–172. Neder, H., Heusser, G., Laubenstein, M. (2000). Low-level gamma-ray germanium spectrometer to measure very low primordial radionuclide concentrations. Applied Radiation and Isotopes 53, 191–195. Neumaier, S., Arnold, D., Böhm, J., Funck, E. (2000). The PTB underground laboratory for dosimetry and spectrometry. Applied Radiation and Isotopes 53, 173–178. Niese, S., Köhler, M., Gleisberg, B. (1998). Low-level counting techniques in the underground laboratory Felsenkeller, Dresden. Journal of Radioanalytical and Nuclear Chemistry 233, 167–172. Pagava, S., Burchuladze, A., Robakidze, T., Rusetski, L., Tsintsadze, D., Povinec, P., Chudy, M., Stanicek, J. (1992). Tbilisi underground laboratory. In: Povinec, P. (Ed.), Rare Nuclear Processes. World Scientific, Singapore, pp. 300–304.
IAEA-MEL’s underground counting laboratory
553
Povinec, P.P. (1982). Dual parameter gamma-ray spectrometer for low-level counting. Isotopenpraxis 18, 423. Povinec, P.P. (1994). Underground low-level counting. In: Garcia-Leon, M., Garcia-Tenorio, M. (Eds.), Low-Level Measurements of Radioactivity in the Environment. World Scientific, Singapore, pp. 113–140. Povinec, P.P. (2004). Developments in analytical technologies for marine radionuclide studies. In: Livingston, H.D. (Ed.), Marine Radioactivity. Elsevier, Amsterdam, pp. 237–294. Povinec, P.P., Comanducci, J.F., Levy-Palomo, I. (2004). IAEA-MEL’s underground counting laboratory in Monaco – Background characteristics of HPGe detectors with anti-cosmic shielding. Applied Radiation and Isotopes 61, 85–93. Reyss, J.-L., Schmidt, S., Legeleux, F., Bonté, P. (1995). Nuclear Instruments & Methods in Physics Research A 357, 391. Semkow, T.M., Parekh, P.P., Schwenker, C.D., Khan, A.J., Bari, A., Colaresi, J.F., Tench, O.K., David, G., Guryn, W. (2002). Low-background gamma spectrometry for environmental radioactivity. Applied Radiation and Isotopes 57, 213–223. Stanicek, J., Povinec, P.P. (1986). Internal pair production in alpha-decaying nuclei and gamma-ray intensities of 241 Am. Nuclear Instruments & Methods in Physics Research B 17, 462–466. Sykora, I., Povinec, P.P. (1986). Measurement of electron capture to positron emission ratios in light and medium nuclides. Nuclear Instruments Methods in Physics Research B 17, 467–471. Sykora, I., Povinec, P.P. (1990). Low-level gamma-ray spectrometry. Acta Physica Universitatis Comeniae 31, 83– 120. Sykora, I., Durcik, M., Stanicek, J., Povinec, P.P. (1992). Radon problem in low-level gamma-ray spectrometry. In: Povinec, P.P. (Ed.), Rare Nuclear Processes. World Scientific, Singapore, pp. 321–326. Theodorsson, P. (1996). Measurement of Weak Radioactivity. World Scientific, Singapore, 333 pp. Vojtyla, P., Povinec, P.P. (2000). A Monte Carlo simulation of background characteristics of low-level HPGe detectors. Applied Radiation and Isotopes 53, 185–190. Vojtyla, P., Beer, J., Stavina, P. (1994). Experimental and simulated cosmic muon induced background of a Ge spectrometer equipped with a top side anticoincidence proportional chamber. Nuclear Instruments & Methods in Physics Research B 86, 380–386. Zvara, I., Povinec, P.P., Sykora, I. (1994). Determination of very low levels of radioactivity. Pure and Applied Chemistry 66 (12), 2537–2586.
554
Levels of airborne radionuclides at Hegura Island, Japan K. Komura* , N. Muguntha Manikandan, Y. Yamaguchi, M. Inoue, T. Abe, Y. Murata Low-Level Radioactivity Laboratory, Institute of Nature and Environmental Technology, Kanazawa University, Wake, Tatsunokuchi, Japan Abstract Activity concentrations of airborne 222 Rn, 210 Pb, 7 Be, 137 Cs and 22 Na have been measured at Hegura Island (37.8◦ N, 136.9◦ E) located 50 km from Noto Peninsula, and their variations have been studied and compared with those found at Tatsunokuchi, Ishikawa Prefecture, Japan. Activity concentrations of 210 Pb and 7 Be varied hourly even in calm weather, and predominantly during drastic changes in meteorological conditions such as the passage of a typhoon or cold front. Activity concentrations and seasonal variations of both 210 Pb and 7 Be at Hegura Island were found to be similar to those measured at Tatsunokuchi, with decreasing levels from late spring to summer and increasing levels from autumn to winter season. These variations are opposite to those obtained for 222 Rn in the corresponding areas. Activity concentrations of airborne 137 Cs and 22 Na at Hegura Island varied from 0.1 to ∼1.5 µBq m−3 and showed seasonal variations similar to those of 210 Pb and 7 Be. Keywords: Airborne nuclides, Anthropogenic radionuclides, Natural radionuclides, 222 Rn, 210 Pb, 210 Po, 212 Pb, 7 Be, 22 Na, 137 C, Low-level gamma-ray spectrometry, Underground laboratory, Seasonal variation, Diurnal variation, Hegura Island, Japan
1. Introduction Variations of 210 Pb (T1/2 = 22.3 yr) and 7 Be (T1/2 = 53.3 day) in the atmosphere are expected to occur on a time scale of about hours or less, as has also been observed for 222 Rn (Yamada et al., 1997). These variations enable us to obtain valuable information on the transport and mixing of air-masses (Yamaguchi et al., 2004). However, long-term measurements have not been performed mainly due to the measuring difficulties because of the very low activity concentrations, usually in the range of 0.1–10 mBq m−3 . However, this problem could be overcome if ultra low-background HPGe detectors are operating in an underground laboratory. * Corresponding author. Address: LLRL, K-INET, Kanazawa University, Wake, Tatsunokuchi, Ishikawa 923-1224, Japan; phone: (+81) 761 51 4440; fax: (+81) 761 51 5528; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08044-7
© 2006 Elsevier Ltd. All rights reserved.
Levels of airborne radionuclides at Hegura Island, Japan
555
Fig. 1. Location map of sampling points and the Ogoya Underground Laboratory (OUL).
The work presented here has been carried out in the framework of the project “Environmental Monitoring and Prediction of Long and Short-Term Dynamics of Pan-Japan Sea Area”, which was adopted in “The 21st Century COE (Center Of Excellence) Program”. Monitoring of airborne radionuclides started in 2002 at three locations, namely, at the Low-Level Radioactivity Laboratory (LLRL) in Tatsunokuchi, on the Shishiku Plateau (640 m above sea level), and at the Hegura Island (0.75 km2 ) located 50 km from Noto Peninsula, all in Ishikawa Prefecture (Fig.1). In this paper we report diurnal and seasonal variations of airborne 210 Pb, 212 Pb and 7 Be, and seasonal variations of artificial 137 Cs and cosmogenic 22 Na (T1/2 = 2.60 yr) observed at Hegura Island.
2. Sampling and measurements 2.1. Sampling of airborne radionuclides Airborne radionuclides were collected on a silica fiber filter (Advantec QR-100, 25.4 cm × 20.3 cm) by using a high volume air sampler (Sibata HV-1000F) operated at a flow rate of 700 L min−1 . Silica fiber filter was chosen because massic activities of 238 U- and 232 Th-series nuclides and 40 K were several orders of magnitude lower than those in glass fiber filter (Yamaguchi et al., 2004), as shown in Table 1. Two kinds of sampling were conducted in this study. The first one was aimed to investigate diurnal variations of 210 Pb and 7 Be. The second one aimed for the analysis of seasonal variations of 210 Pb and 7 Be by 1–2 days of sampling at LLRL and Shishiku and 1 week sampling at Hegura Island. The short samplings were performed at Tatsunokuchi or simultaneously at two or three points, i.e. Tatsunokuchi– Hegura, Tatsunokuchi–Shishiku, and Tatsunokuchi–Shishiku–Hegura. Activity concentration of airborne 222 Rn was monitored at one hour intervals using high sensitive electrostatic radon monitor (ERM) developed by Iida et al. (1991). Wind direction and speed data were obtained from Prefectural Institute of Public Health and Environmental Sciences (Tatsunokuchi), Japan Coast Guard (Hegura Island) and Ishikawa Fliers Association (Shishiku).
556
K. Komura et al. Table 1 Activities of natural radionuclides in silica fiber and glass fiber filters Activity (mBq sheet−1 )
Nuclide
238 U (234 Th) 226 Ra (214 Pb) 210 Pb 228 Ra (228 Ac) 228 Th (212 Pb) 40 K
Silica fiber filter
Glass fiber filter
Acivity ratio (Silica/Glass)
13 7.7 8 <2 1.2 <17
150.0 80 120 65 50 4980
0.087 0.096 0.067 <0.03 0.024 <0.003
2.2. Measurements After sampling, a 1/2 or 1/3 portion of each filter was taken for gamma-ray spectrometry of 210 Pb, 212 Pb, 7 Be, 137 Cs and 22 Na, and 1/4 portion for 210 Po analysis by alpha-ray spectrometry, and the rest of the filter has been stored for future use. Counting sources of 35 mm or 21 mm in diameter were prepared using a 10 ton hydraulic press and subjected to gammaray spectrometry as soon as possible after sampling in order to measure the short-lived 212 Pb (T1/2 = 10.6 h). The samples were analyzed in the Ogoya Underground Laboratory (OUL), located at the depth of 270 m of water equivalent in Ishikawa prefecture, Japan (Hamajima and Komura, 2004). Eleven ultra low-background HPGe detectors (6 well-type, 4 planar and 1 coaxial) (Hamajima and Komura, 2004) were used to analyze simultaneously several samples collected at 1–2 hours of intervals. Each sample was analyzed two times, i.e. a shorter measurement for analysis of 212 Pb, and a longer measurement for analysis of 210 Pb, 7 Be, 137 Cs and 22 Na.
3. Results and discussion 3.1. High-resolution measurement of airborne 210 Pb and 7 Be Examples of temporal variations of airborne 210 Pb and 7 Be measured for two-hour intervals are shown in Fig. 2. If samples would be measured over one day intervals, such short time variations might be overlooked, as is documented in the step-graphs. It is interesting to note that the observed 210 Pb and 7 Be variations are different from the well-known diurnal variation of 222 Rn, i.e. with maximum values in early morning and minimum values in the afternoon. 3.2. Variations of
210 Pb
and 7 Be during the passage of a cold front at Hegura Island
Variation of airborne 210 Pb and 7 Be activities for the sampling period from June 23–30, 2003 is shown in Fig. 3. During this sampling period at Hegura Island, a cold front was passing over the island on June 28. During the passage of the cold front, the 210 Pb and 7 Be levels decreased by 40%–90%. It is interesting to note that 7 Be was more sensitive to this weather change than
Levels of airborne radionuclides at Hegura Island, Japan
557
Fig. 2. Temporal variations of 210 Pb and 7 Be observed by high-resolution analysis at two-hour sampling intervals. Step-graph shows the variations simulated by one-day intervals of sampling.
Fig. 3. Temporal variations of 210 Pb and 7 Be during the approach of a cold front. 210 Pb,
namely, the decrease of 7 Be began one day before the arrival of the cold front and was deeper than that of 210 Pb. A similar decrease in activity concentrations was observed also during the passage of a typhoon (Komura and Hamajima, 2004). These decreases may be explained by the fact that airborne radionuclides were scavenged from the air by rainfall. Larger decreases and faster recoveries of 7 Be compared with those of 210 Pb, may be explained by its shorter half-life and more rapid diffusion of 7 Be due to the mixing of air at high altitudes, where it is produced by interactions of cosmic-ray particles with the atmosphere. 3.3. Seasonal variations of
210 Pb
and 7 Be at Hegura Island
Seasonal variations of 210 Pb and 7 Be observed at Hegura Island and at Tatsunokuchi are compared in Fig. 4, together with those of 222 Rn. The observed activity concentrations of 210 Pb and 7 Be at Hegura Island are 0.1–3 mBq m−3 and 0.1–8 mBq m−3 , respectively. They are a little lower than the levels at Tatsunokuchi, showing a decrease from late spring (May) to June and then increase gradually from autumn to the winter season. Lower values in July (just after
558
K. Komura et al.
Fig. 4. Seasonal variations of 210 Pb, 7 Be and 222 Rn at Hegura Island and at Tatsunokuchi.
the rainy season) may be explained by the fact that the air-masses covering this area come from the Pacific Ocean, which contains lower activity concentrations of 210 Pb. The variation pattern of both 210 Pb and 7 Be is just opposite to that of 222 Rn. The 222 Rn activities in Hegura Island are about 1/4 of that at Tatsunokuchi because of the low contribution from exhalation over the island, which comprise only 0.75 km2 of land surface. 3.4. Simultaneous measurements of temporal variations of Tatsunokuchi
212 Pb
at Hegura Island and at
Temporal variations of 212 Pb, which is a progeny of the short-lived 220 Rn, were measured simultaneously at Hegura Island and at Tatsunokuchi at 3-hour intervals. Air filter samples collected at Tatsunokuchi were measured at LLRL immediately after sampling. However, most of the Hegura samples were measured 1 day after sampling and 3 days in the odd cases
Levels of airborne radionuclides at Hegura Island, Japan
559
Fig. 5. Temporal variation of short-lived 212 Pb at Hegura Island and at Tatsunokuchi.
due to the stoppage of the ferry service because a cold front passed over the island during this experiment. Results are plotted in Fig. 5. Here, the concentration of 212 Pb was assumed to be constant during 3 hours of sampling and decay during the sampling and transportation was corrected before the calculation of the activity concentration. The ratio of 212 Pb activity at Hegura Island and at Tatsunokuchi ranged from 0.1 to 1, which can be well understood because the production of 212 Pb within the island is expected to be very low. A drastic decrease of 212 Pb activity was observed both at Hegura Island and at Tatsunokuchi during the passage of the cold front, as was also observed for 210 Pb and 7 Be, and explained by the scavenging of airborne particles due to rainfall. It is also noteworthy that the variation patterns at Hegura Island and at Tatsunokuchi seemed rather similar, but the peak time is 2–4 hours in advance at the Hegura area. This may be explained by the movement of air masses which passed over Hegura Island and the Tatsunokuchi site. 3.5. Levels of airborne 137 Cs and cosmogenic 22 Na at Hegura Island Levels of airborne 137 Cs have decreased considerably since its peak concentration in the 1960s, so it can be detected only in large volume air samples or using an ultra lowbackground Ge detector. In this study, each filter sample corresponding to volume of about 3000–4000 m3 of air was measured for a long time in order to determine 137 Cs and cosmogenic 22 Na, after the measurements of 210 Pb and 7 Be were completed. Figure 6 shows an example of a gamma-ray spectrum with weak 137 Cs and 22 Na peaks. The results are plotted in Fig. 7 for the period from May to November, 2003. Activity concentrations of 137 Cs and 22 Na were lower than ∼1 and ∼1.5 µBq m−3 , respectively, and showed similar patterns, i.e. decrease from May to July to levels between 0.1–0.5 µBq m−3 . The higher value of 22 Na in May is due to a tailing of the spring peak, which begins in March. The lower values during summer to autumn may be explained by air masses arriving from the south-east, which originated at
560
K. Komura et al.
Fig. 6. Example of a gamma-ray spectrum showing anthropogenic 137 Cs and cosmogenic 22 Na measured by a 28 cm2 × 2 cm planar-type Ge detector in OUL. Sampling volume and measuring time were 4033 m3 and 5620 min, respectively.
Fig. 7. Activity concentrations of airborne 137 Cs and cosmogenic 22 Na at Hegura Island.
low latitudes over the Pacific Ocean, where production rates of cosmogenic radionuclides are low. As already mentioned, the Hegura Island is very small and located 50 km from Noto Peninsula, therefore contributions of re-suspended 137 Cs from the surface of the Island as well as that from the main island of Japan are considered to be very low. Therefore, the observed values show typical levels of airborne 137 Cs without an influence from re-suspension in nearby areas. On the other hand, attention was paid to cosmic ray induced 22 Na as possible geo-
Levels of airborne radionuclides at Hegura Island, Japan
561
chemical tracer in hydrology (Sakaguchi et al., 2003). Reliable measurements of this radionuclide are therefore required in order to open new research studies in hydrology, meteorology and cosmic-ray physics.
Acknowledgements The authors express their deep thanks to colleagues, students and school teachers at the Hegura Island for their assistance during simultaneous samplings at Tatsunouchi, Hegura Island and Shishiku Highland.
References Hamajima, Y., Komura, K. (2004). Background components of Ge detectors in Ogoya underground laboratory. Applied Radiation and Isotopes 61, 185–189. Iida, T., Ikebe, Y., Toji, K. (1991). An electrostatic radon monitor for measurements of environmental radon. Res. Lett. Atmos. Electr. 11, 55–59. Komura, K., Hamajima, Y. (2004). Ogoya Underground Laboratory for the measurement of extremely low levels of environmental radioactivity: Review of recent projects carried out at OUL. Applied Radiation and Isotopes 61, 164–189. Sakaguchi, A., Yamamoto, M., Otsuka, Y., Sasaki, K., Yokota, K., Komura, K. (2003). Low-level measurement of the cosmogenic 22 Na radionuclide in fresh water by ultra low-background gamma-ray spectrometry after simple radiochemical separation. Journal of Radioanalytical and Nuclear Chemistry 258 (1), 101–105. Yamada, K., Iida, Y., Ikebe, Y., Miyachi, H., Nagao, I., Komura, K. (1997). Time series analysis of atmospheric 222 Rn concentrations and meteorological factors in Japan. In: Katase, A., Shimo, M. (Eds.), Proc. 7th Tohwa University International Symposium “Radon and Thoron in the Human Environment”. Fukuoka, Japan 23–25 Oct., 1997. World Scientific, Singapore, pp. 255–271. Yamaguchi, Y., Komura, K., Murata, Y., Muguntha Manikandan, N., Iida, T., Moriizumi, J., Hirao, S. (2004). High resolution and simultaneous monitoring of variations of airborne Rn-222, Pb-210, Po-210, Pb-212 and Be-7 at Tatsunokuchi, Shishiku and Hegura Island. In: Proc. 5th Workshop on Environmental Radioactivity. KEK Proceedings 2004-8, pp. 137–144.
562
The use of sodium iodide detectors to locate buried radioactive particles in the seabed off Dounreay nuclear facility, Scotland J. Toolea,* , S.C. Innesa , M. Liddiarda , J. Cassidyb , S. Russb a Environmental Projects Department, UKAEA Dounreay, Thurso, Caithness, KW14 7TZ, UK b Fathoms Ltd, The Harbour, Wick, Caithness, KW1 5HA, UK
Abstract As part of a wider programme dealing with the monitoring of radioactive particles in the marine environment around Dounreay, UKAEA and our contractor Fathoms Ltd tested the viability of using remotely-operated sodium iodide detectors, as a potential alternative to the more risky method of diver-operated hand-held monitors. In 2003, a small frame-mounted detector was deployed successfully underwater and showed that buried particles could be detected out to 70 cm offset distance. In order to survey large areas underwater, however, a more sophisticated system was required. A remotely-operated vehicle (ROV) equipped with a much larger detector was designed, built and commissioned within nine months of conception. The system was tested in static mode in a laboratory environment and in dynamic mode offshore to verify the detection capabilities prior to undertaking area surveys underwater. The ROV design and particle detection methodology are described and the results of initial seabed surveys are presented. An area of 30,000 m2 of seabed has been surveyed and over 120 contacts were obtained and mapped. These preliminary measurements are consistent with diver-recovered finds in a north-easterly direction from the Dounreay effluent discharge point. They also show the absence of particles in other areas. Further ROV surveys are required to delineate the extent of seabed contamination. Keywords: Radioactive particles, Gamma-ray spectrometry, 137 C, Dounreay, Remotely-operated vehicle
1. Introduction The radioactive particles found in the local environment around the Dounreay nuclear site, Caithness, UK, are fragments of irradiated nuclear fuel elements, which have been released accidentally into the environment from the site in the late 1960s and early 1970s. The most prevalent type were produced when research reactor fuel was being milled prior to reprocessing operations and consist of small uranium inclusions (UAl4 ) within an aluminium matrix. A second type are thought to have originated from high-temperature reactions in fuel dissolver baskets, which led to the production of uranium–niobium alloy containing fission products * Corresponding author. Address: UKAEA Dounreay, Building D2003, Thurso, Caithness, KW14 7TZ, UK; phone: (+44) 1847 806092; fax: (+44) 1847 806901; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08045-9
© 2006 Elsevier Ltd. All rights reserved.
Radioactive particles in the seabed off Dounreay nuclear facility, Scotland
563
and 94 Nb activation product. Particles have been considered an important environmental issue by the United Kingdom Atomic Energy Authority (UKAEA) since the first recorded particle find on the Dounreay Foreshore in 1983. These early particles were located during routine environmental surveys using handheld probes. Numerous study programmes have been undertaken to understand the origins of the particles, their pathways to the environment, and the risks associated with them (Koskelainen, 2005). The contractor Fathoms Ltd has provided diver services since 1997 for the location and retrieval of buried particles and a total of 846 particles have been retrieved up to October 2004 using handheld probes. Particles are detected mainly via their 137 Cs gamma-ray emission. Since they also contain beta-emitting 90 Sr/90 Y and small amounts of other fission products and activation products, they are sometimes detectable in the near-surface terrestrial environment using handheld beta probes (Toole, 2005). UKAEA’s view is that the risks associated with these diving operations far outweighs those from the buried particles and a requirement for remote detection systems was identified. In addition, there was a limitation on the water depth in which the divers could safely operate, and deeper water zones needed to be surveyed. Therefore, as part of a wider programme dealing with the monitoring and retrieval of radioactive particles in the environment around Dounreay, UKAEA contracted Fathoms Ltd to trial the use of a sodium iodide detector at a variety of locations in nearshore regions. There are several reported studies of in-situ gamma spectrometry in the marine environment (e.g. Osvath and Povinec, 2001; Povinec et al., 1997; Jones, 2001). These deal with monitoring of natural or artificial radioactivity in the bulk sediments or in the water column. We are unaware of any underwater gamma spectrometry system which has attempted to contend with hundreds or thousands of buried radioactive sources, so the present work is the first published account of an attempt to meet this technical challenge. Fathoms marinised a 76 mm diameter × 76 mm sodium iodide detector in a stainless steel jacket for underwater use and selected a SAM-935 Radiation Detection, Surveillance and Measurement system, supplied by Berkeley Nucleonics Corporation, as the signal processing package. This system provides gamma spectra in either linear mode (through PC-mounted Quantum Gold spectrum processing software) or in QCC mode (Quadratic Compression Conversion, within the SAM-935 firmware). Details of the SAM-935 operation and its capability in real-time identification of radionuclides are provided elsewhere (McQuaid, 2002). Fifteen days of offshore trials were achieved in September 2003. Following the experience with this small detector and its simple frame mounting, a more sophisticated system was envisaged which would be capable of moving a larger detector over the sediment surface at a controlled speed and with a low seabed-detector clearance. A twintracked remotely-operated vehicle (ROV) equipped with a marinised 102 mm × 102 mm × 406 mm sodium iodide detector was designed, built and tested within 9 months of its conception. This particle detection system, which used the same SAM-935 signal processing software as in the 2003 trial, was deployed for 22 days in September/October 2004. Over 130 contacts with radioactive particles were obtained in a total survey area of 30,000 m2 . This paper summarises the results of the 2003 deployment, describes the ROV system deployed in 2004 and presents and discusses the results derived from these ROV surveys undertaken in various areas off Dounreay.
564
J. Toole et al.
2. Field trial of small sodium iodide detector, September 2003 2.1. Sodium iodide detector configuration A 76 mm diameter × 76 mm height NaI(Tl) detector was used to detect the gamma-rays from buried radioactive particles. The detector and associated photomultiplier tube were housed in a 3.2 mm thick sealed stainless steel jacket for use underwater. Stainless steel 304L was the material chosen for the watertight container within which the detector would be housed. A water ingress alarm was also fitted internally with an audible and LED warning, so that power could be shut down immediately, minimising damage to components. The marinised detector unit was secured within a tubular frame, with tubular skids to allow easier transit between successive spectral gathering locations. The detector in its frame is shown in Fig. 1. The design allows the deployment of other user selectable probes with different sensors if required. When the frame is deployed on a flat seabed, the end of the stainless steel jacket is approximately 2 cm from the sediment surface; this places the front edge of the NaI crystal less than 5 cm from the sediment surface. A video camera was permanently attached to the frame to monitor underwater, single-point energy calibration (using a 137 Cs standard source, 370 kBq on 24/9/01), lowering of the system through the water column onto the seabed and to check positioning status during spectrum acquisition. Various aspects of the operation were recorded on DVD. 2.2. Deployment and positioning The frame and detector were lowered to the seabed via a drum winch, the descent being visible to the winch operator on a live video screen. Once deployed on the seabed, spectra were acquired within an area, typically for one minute, followed by relocation to a new
Fig. 1. Frame-mounted sodium iodide detector at the stern of MV Teal. The video camera and positioning transponder are also visible.
Radioactive particles in the seabed off Dounreay nuclear facility, Scotland
565
site. Over 1,200 drops were achieved, even on days with marginal operating conditions. That the marinised detector withstood these repeated jolts during placement on the seabed and on retrieval to the stern of the vessel is testament to its design and functionality in a hostile environment. Surface positioning was by differential GPS using a Trimble DGPS DSM212L receiver. This equipment coupled with the Navigation Software, Tower Hydrographics v3.2.9 has a proven track record of reliable accuracy over several years’ operations off Dounreay. Typical accuracy of ship positioning is 1 metre. Positioning of the detector frame was via ultra short baseline transponder (USBL), which was then logged as a fixed offset to the vessel. The overall positioning accuracy for the frame underwater was approximately 4 metres (in water depths of 20–40 m). 2.3. Signal processing The SAM-935 detection system has real-time isotope recognition capability. A patented process (quadratic compression conversion) is utilised which compresses the spectrum in a few microseconds via an algorithm related to the square root of the gamma energy. The compression is time-sliced and updated each second. A recent background spectrum is continuously subtracted from the live, processed spectrum, while a fast deconvolution algorithm removes gamma continuum also each second (McQuaid, 2002). Small increases in count rates due to the presence of gamma emitters such as 137 Cs, either in the photopeak region or in the Compton region are then readily detectable. As long as the ambient background is not changing from the reference background, the system has high sensitivity (low detection limit). If the local reference background, taken in the survey area from influence of particles, is higher than the background in the study area, there will be some loss of sensitivity and the risk of false-negative results increases. Conversely, if the reference background is lower than the background in the study area, the system will trigger too often (false-positives). Regular monitoring data (e.g. see Fig. 2) has shown that the natural background in the sediment over most areas of interest in this study is rather constant. This constant spectrum improves the reliability of the whole data analysis. The firmware allows several triggers to be set at a userselectable multiple of standard deviations (sigmas) above background. In this work, triggers were set for the Full spectrum, for the 137 Cs window and for the 60 Co window (about 1% of particles found contain 60 Co). Normally the settings used were at 3σ , but occasionally the
Fig. 2. Examples of consistent background count rate, gross cps, from offshore sediment.
566
J. Toole et al.
Full trigger was raised to reduce the number of false-positive triggers due to the detector becoming buried in the sediment as large sand ripples were encountered, either when placing the detector frame or during forward movement or turns of the ROV. When any of the pre-set sigma thresholds are exceeded, a spectrum is acquired for the duration of the trigger. These real time alarm events are captured and stored for further analysis and review. 2.4. Some results of the small detector trial Between the 1st and 16th September 2003, 15 days of offshore surveys were achieved, with only one full day lost due to weather conditions. In this period, a total of 1278 gamma spectra were obtained with the detector static on the seabed: 291 linear spectra, 928 QCC spectra and 59 alarm spectra (alarm spectra are acquired automatically whenever one of several preset count rates are exceeded). Linear spectra were acquired for presentational and comparison purposes as this is the standard type provided by other gamma spectrometry systems. Some examples of the total gamma count rates (cps) observed over several survey areas are given in Fig. 2. It is evident that the sandy sediment is quite homogeneous over much of the nearshore seabed, showing background count rates of about 70 cps for the detector geometry used. Other areas in deeper water did show more variable patterns of background count rates of up to a factor of four (Toole, 2004). The coordinates of a close approach to a buried particle on the 11th September survey were logged on the ship’s computer. A decision was made to go back to this site as soon as a particle recovery crew and procedure could be made ready. On 15th September, the MV Teal returned to this logged position and moored on four anchors with the support boat on station. The detector was put into the water, calibrated with a standard 137 Cs source and deployed on the bottom sediment. The detector assembly was lowered to the seabed, but no alarms were triggered. However, the particle was located by a diver, less than about 2–3 metres away from the detector. The diver found the particle using a handheld plastic scintillator probe and then placed the NaI(Tl) detector in its frame directly over the particle. With the SAM-935 in surveillance (alarm) mode, alarm readings were taken at 0, 15, 30 and 45 cm offsets from the particle using a marked-up aluminium rod to set the distances. On each alarm, the readings do fluctuate and the maximum readings for the various alarm categories were recorded. Readings from the handheld probe at the same points as the detector was placed were also recorded. Figure 3 shows the various readings at the measured distances from the particle, normalised to local background readings. Figure 4 shows the spectra obtained at the various offset distances. The approximate distances for the readings to reach local background levels are provided in Table 1. The peak alarm is usually set in case there is a photopeak present for a radionuclide which is not pre-set in the trigger list. In the present case, it responds to the 137 Cs present and effectively mirrors the 137 Cs trigger. The SAM-935 also records isotope-specific dose rate in nSv/h; this is not a useful measure of particle presence compared to the other methods. In this example of seabed measurements in the location of a buried particle (subsequently measured in the laboratory as 440 kBq 137 Cs and with a calculated burial depth 3 cm), and with the detection system set up in surveillance mode, the Full alarm provides the longest range for responding to the presence of the particle, out to 70 cm in this example. It is more sensitive than the 137 Cs alarm because it includes the much larger count rates from 137 Cs found in the
Radioactive particles in the seabed off Dounreay nuclear facility, Scotland
567
Fig. 3. Alarm readings and handheld probe count rate at various distances from a particle at 3 cm depth. Readings are relative to local background values.
Fig. 4. Gamma spectrum of buried particle at various offsets: 0, 15, 30, 45 cm and comparison with background spectrum. Log scaling. Counts normalised to 180 seconds acquisition.
Compton region of the gamma spectrum (scattered radiation) rather than the counts being restricted to a narrow 137 Cs window. This is consistent with the observation on board ship that a triggered alarm was almost always caused by the Full spectrum trigger, whether due to a particle or varying natural radioactivity (U and Th series and 40 K). In addition, whenever the Full spectrum alarm was accompanied by a 137 Cs alarm, this would almost certainly signify that a particle was within range, i.e. within about 50 cm for this detector and geometry.
568
J. Toole et al. Table 1 Sensitivity of various detection methods to a particle buried at 3 cm depth in sediment Detection method
Approximate range (cm)
Full alarm 137 Cs alarm Handheld probe Full dose
70 50 50 40
Both the Full alarm and the 137 Cs alarm appeared to be more sensitive to the presence of this particle than the diver-held plastic detector probes. However the sodium iodide detector in its mounting frame lacks the flexibility of movement that a diver has and the diver will be more successful at homing in on a particle. Particle searches using the drop-and-count mode were very inefficient; no more than 5 particle contacts were achieved in over 1000 drops, even though the search areas included those known to have a high density of particles.
3. Trials of large NaI detector fitted to ROV 3.1. Laboratory trials On 30th June and 1st July 2004, a large marinised NaI detector and SAM-935 system were tested at Dounreay in a submerged geometry using a purpose-built crate, bags of offshore sediment (pre-screened for radioactivity) and offshore seawater. Figure 5 shows the detector geometry over a 10 cm layer of sediment in the crate. Information on the calibration sources used for the trial are provided in Table 2. The detector was submerged under 30 cm of seawater in preparation for each counting period and spectrum acquisition. Prior to acquisition commencing, the SAM-935 system was energy calibrated. In addition, a background spectrum was also taken manually and at regular intervals thereafter. The detector was placed on two polystyrene supports to raise it approximately five centimetres above the sediment. This corresponds with the detector configuration that is expected to be used offshore – water is a good shield for gamma-rays and a close detector–sediment spacing is needed. The larger detector provides an increase in volume of over 12 times compared to the 76 mm × 76 mm detector described above, and so provides a much improved signal intensity and therefore detection range. Depths of sediment of 10, 30, 50 and 70 cm were successively built up within the crate. At each depth, the sources were moved under the box at predetermined horizontal offset distances. In the case of zero offset distances, the detector was placed with the NaI crystal in the centre of the box, to try and replicate as closely as possible the scattering characteristics of the sediment below as would be experienced offshore. In the case of other offset distances, the detector was moved to the edge of the box to enable offset distances of 70 cm or more to be achieved. In a final test, the sources were doublewrapped inside rubber closures and placed on the sediment surface underwater at various
Radioactive particles in the seabed off Dounreay nuclear facility, Scotland
569
Fig. 5. Large, marinised NaI detector under seawater and above a 10 cm layer of sediment. Radioactive sources were moved under the box to test the response at various offset distances. Table 2 Radioactivity content of sources used to test the large NaI detector Source
Dounreay code
Activity at June 2004 (kBq)
137 Cs
HO216 FN157 HO436
25,000 361 79.3
137 Cs 60 Co
offset distances from the detector. This would determine how far away the detector would register a particle located on the surface of the sediment. Details of the experimental set up and results of count rates, spectra and trigger values obtained at each depth/offset are given in a separate report (Innes, 2004). The main findings of the sensitivity of the detector underwater are provided in Table 3. It was also observed that the detection system would register the presence of a particle on the surface of the sediment at a much greater distance than if buried under the sediment. This is due to the lower density and lower total mass attenuation coefficient of water compared to sand. For example, a small 137 Cs source was detectable on Full trigger, with a static detector as described, to a depth of about 37 cm, but was detectable to an offset distance of about 75 cm through seawater. This was a practical demonstration of the same observation by Maucec et al. (2004) who used Monte Carlo modelling to determine detection depth limits and acquisition times for detection of offshore particles using a bismuth germanate (BGO) and other detectors. From the detection envelope generated from the laboratory results, it was calculated that a 5 cm sediment–detector clearance and an 80 cm swathe width for the ROV would provide
570
J. Toole et al. Table 3 Extrapolated depths for detection and identification of radioactive point sources in a simulated underwater environment Source
Activity (kBq)
Approximate maximum depth for detection, using Full trigger (cm)
Approximate maximum depth for identification, using Cs or Co trigger (cm)
137 Cs
25,000 361 79.3
67 37 30
50 25 22
137 Cs 60 Co
approximately 95% volume coverage on Full trigger down to a sediment depth of 65 cm for a 20 MBq particle, i.e. 95% efficiency for a static detector. However, it was considered that detection capability would be different during the field trials, as the detector would be dynamic, and have less time to evaluate a spectrum. In addition, Full trigger thresholds may need to be raised above the 3σ value so as to raise confidence of detection of a particle and to cope with varying natural background. Also, since ambiguity is removed by identifying 137 Cs by means of capturing its corresponding alarm, rather than relying on Full triggers only, some mechanism for interrogating Full alarms on the seabed was required. The features of the ROV system which were designed to address these points are described in the next section. 3.2. Description of offshore ROV Fathoms’ tracked ROV was designed and built to provide a reliable and stable platform for the SAM-935 detector system. To achieve maximum reliability various design criteria were identified in the early stages of development. As the ROV would be working for up to ten hours per day continuously, it was designed to operate well within the specifications of various components of the system. Furthermore, in the event of system failures, to minimise downtime as far as was possible, the ROV employed well-proven industry standard components. The ROV is fitted with two high definition colour cameras, one on a pan and tilt unit, high power lighting and a positioning transponder. The electronics housing has a water ingress alarm, so in the event of water leaking into the bottle, power can be shut down immediately, minimising the damage to components within. Hydraulic power for the tracks and rams is supplied by an 11 kW power pack. The three phase 415 volt supply, derived from the ship’s onboard generator is fed to the ROV via the umbilical. All possible safety features are employed. The soft-start system provides over current, short circuit and thermal protection, and an RCB on the ship’s distribution panel provides earth leakage protection. The soft-starter also ensures minimal loading on the ship’s generator and prolongs the life of the three phase motor within the power pack. As the ROV is usually operating at a speed of 0.35 m s−1 , the whole hydraulic system is operating well within the 1 m s−1 maximum speed possible. To minimise the size of the umbilical all control functions to the ROV are encoded and transmitted to the ROV by a single twisted pair within the umbilical. The system used is a
Radioactive particles in the seabed off Dounreay nuclear facility, Scotland
571
Fig. 6. ROV in operation underwater, detector deployed in monitoring geometry. The plastic scintillator probe is visible behind the large NaI probe.
standard off-the-shelf, modular, industrial controller. There is no need for any programming and if further functions are required they can be very simply accommodated. If there is a failure within the telemetry system, the fault should be easily identified and a replacement module inserted minimising downtime. Throughout all offshore operations, there was never a failure of the electronics or hydraulics on the ROV. Figure 6 shows the ROV operating underwater, with the large NaI detector in the survey position. 3.3. Offshore validation trials with ROV Between the 3rd and 10th September 2004, a series of ROV/detector trials were held in the offshore environment near Dounreay. Divers were tasked with locating a number of buried particles using their standard plastic scintillator probes – see Fig. 7. The ROV equipped with the large NaI detector was then run past the particle locations, marked with metal pins, at a variety of offset distances: 0, 0.5, 1.0 and 1.5 metres. After these run, the divers retrieved the particles by coring, determined the burial depth onboard ship and the retrieved particles were transported to the Dounreay radiometric laboratory for measurement of 137 Cs content. In total, twelve particles were recovered, of which 9 were retrieved satisfactorily by coring. The maximum recorded values for Full trigger at 1 metre offset from the diver-located particles are summarised in Table 4. The coring procedure is known to slightly compress the core, so the depths given are slight underestimates. It can be seen that particles of activity at least down to 270 kBq and at depths of about 45 cm will trigger the system and can be detected by the Full trigger out to offset distances of 1 metre, with the ROV moving at a speed of approximately 0.35 m s−1 .
572
J. Toole et al.
Fig. 7. Diver searching for buried particles using a plastic scintillator probe.
Table 4 Illustration of sensitivity of dynamic ROV to particles buried at different depths 137 Cs
Date particle retrieved
Pin no.
Depth (cm)
(kBq)
Maximum Full trigger value at 1.0 m offset
9-Sept-04 8-Sept-04 10-Sept-04 8-Sept-04 8-Sept-04 8-Sept-04 9-Sept-04 8-Sept-04 9-Sept-04
3 5 6 7 8 9 10 11 12
35 25 20 45 40 37 15 28 10
2300 2300 270 2300 280 3300 500 3400 740
7.7 8.2 9.8 8.9 17.7 6.7 7.1 No ROV runs 7.6
Although the evidence from the lab trials and the offshore trials showed that particles such as those shown in Table 4 were detectable at offset distances of over 1 metre, a cautious approach was taken for the actual ROV surveys: the swathe width was reduced to 0.6 metres (30 cm offset distance) and the ROV was operated at a speed of 0.35 m s−1 rather than the originally intended 0.5 m s−1 . A swathe width of 0.6 m was conveniently achieved by ensuring via video link that the centre line of the NaI detector, when on the return track, ran just inside the nearest rubber track mark from the previous track line. This would provide more
Radioactive particles in the seabed off Dounreay nuclear facility, Scotland
573
sensitivity for particles buried deeper than 50 cm, although the exact detection depth for these operating conditions is not known. We can say that particles with activities of about 1000 kBq 137 Cs are readily detected to 50 cm depth, while lower-activity particles will be detected at shallower depths and higher-activity particles will be detected to greater depths. No hardware detection system will be able to detect all particles at all of the depths to which they may be mixed. If small particles exist at depth they will not be readily detected by the present or other systems. It is not known if smaller (ca. 100 kBq 137 Cs or less) particles exist at depth. A correlation of increasing particle activity with depth has been noted from an assessment of 101 particles located by divers between 1997 and 1999 (Scirea, 2002). Also, Atkinson (2002) has reported the existence of two populations of particles, with the higher activity particles (>2000 kBq 137 Cs) only occurring buried in the sediment below 20 cm depth. Since there was evidence for spectrum drift in both the laboratory trials and in the offshore trials, the capability for regular energy calibration was required. A 137 Cs check source was purchased from AEA Technology QSA GmbH. The sealed source contained 43.5 kBq and was pressure rated to 1700 bar. It was suitably encapsulated and collimated and fixed to a metal rod, which was de-mountable from the ROV. This source provided a beam of 662 keV photons to which the NaI detector would be exposed when raised from its horizontal operating position to the vertical position. The source did not register on the SAM-935 when the detector was in its operating position. Regular energy calibrations were performed during the offshore surveys to ensure photopeaks were not missed, although the Full alarms and Peak alarms would still have responded to particles. A limitation of the SAM-935 system surveillance screen (showing trigger levels) was that it did not provide a continuous reading of counts per second; count rate was obtainable on another screen, but we needed to monitor the trigger thresholds during survey. One of the plastic scintillator probes as used by the divers was therefore mounted on the ROV, behind and 20 cm offset from the NaI detector. This provided a gross count rate each second, which could be monitored. Apart from recording this probe’s own response to buried particles, it provided an important record of the background count rate continuously. If the background was increasing or decreasing, a new reference background could be taken to maintain the detection sensitivity of the NaI system. A trigger of the SAM-935 system on Full trigger only is not unambiguous evidence for the presence of a particle, as this can be caused by a variation (increase) of the natural radiation background over the reference background stored in the firmware. The preference is to obtain unambiguous evidence in the form of a 137 Cs trigger and its associated energy loss spectrum. The capability was therefore included in the ROV design to move the detector laterally as well as vertically. Whenever a Full alarm was triggered during an ROV survey, but not accompanied by a 137 Cs (or 60 Co) alarm, the operator stopped the vehicle, reversed until the Full alarm was maximised and moved the detector left or right if necessary to develop the 137 Cs peak and the trigger value. An example of the real-time data obtained during the survey is drawn together in Fig. 8. This shows the second by second count rate data (cps) recorded by the ROV-mounted plastic scintillator probe, the alarm spectrum obtained by the NaI/SAM-935 system and the trigger data (trigger duration, Full value, 137 Cs value and cps recorded in the NaI probe). Such traces and trigger data are available for each ROV survey, together with positioning coordinates
574
J. Toole et al.
Fig. 8. Example of a survey trace showing response in cps of plastic scintillator probe, and the trigger data and trigger spectrum from the SAM-935/NaI system. The arrow marks the time of the NaI alarming; the other probe responds a few seconds later.
logged every 10 seconds and continuous video footage of the SAM-935 monitoring screen (recorded on ship). 3.4. Results of offshore trials with ROV and large detector Between 11th September and 27th October 2004, weather conditions permitted 22 days or part-days of ROV surveying in the offshore environment. Over 120 particle contacts were confirmed and located to about 4 m resolution. These were mapped using GIS software, alongside diver-recovered particles following diver surveys in pre-defined circular areas. The positions of the diver-recovered particles and the ROV-located particles for an area in which the main particle plume is presumed to lie is shown in Fig. 9. Information on the number of finds in the diver areas and in the ROV areas surveyed is presented in Table 5. The ROV numbers have been corrected for double-counting i.e. situations where the same particle was identified in two adjacent ROV survey lines. The particle densities, number per square metre, range from 0 to 0.0192 for diver finds and from 0 to 0.0128 for ROV finds. It is to be expected that particle densities for the ROV will be lower than those for divers since (1) divers have retrieved over 800 particles from the seabed since 1997 and (2) areas which have been cleared are being repopulated by a mobile cache of surface and near-surface particles. Thus the areas over which the ROV is traversing have a reduced particle density compared to earlier years. The ROV was sent into repop 4 area (see Fig. 9) to see if it could detect any of these mobile particles. These are known to be at the lower end of the range of particle activities (and probably also physical size). Ten
Radioactive particles in the seabed off Dounreay nuclear facility, Scotland
575
Fig. 9. Map of particle finds near the discharge diffuser. Diver finds (dots) in repopulation circles are from first visits to these areas. ROV finds in shaded survey areas are denoted by star symbols.
576
J. Toole et al.
Table 5 Summary of particle finds by divers and by ROV in various areas offshore. The figures for diver finds in the repop zones are for first visits Diver surveys Particle density (m−2 )
Year
Strathy remote Melvich remote Redpoint remote Sandside inshore Repop 1 Cell C6 Repop 2 Repop 2 outer Repop 10 Repop 9 C2 Repop 3 Repop 3 outer Repop 9 C1 Repop 8 Repop 4 Repop 4 outer Cell K22 Repop 5 Repop 5 outer Repop 6 Repop 11 Repop 7 Repop 12 Crosskirk remote Brims remote Brims east
2003 2003 2003 2003 2000 2003 2000 2002 2004 2002 2000 2002 2002 2002 2000 2002 2002 2000 2002 2000 2004 2000 2004 2003 2003 2004
2000 2000 2000 2000 2500 2000 2500 5355 2500 2000 2500 5355 2500 2500 2500 5355 1958 2500 5355 2500 2500 2500 2500 2000 2000 2000
0 0 0 3 0 1 6 12 9 26 15 15 48 10 38 78 12 19 31 15 9 6 8 4 5 0
0 0 0 0.0015 0 0.0005 0.0024 0.0022 0.0036 0.0130 0.0060 0.0028 0.0192 0.0040 0.0152 0.0146 0.0061 0.0076 0.0058 0.0060 0.0036 0.0024 0.0032 0.0020 0.0025 0
ROV surveys 1 2 3 Dunnet Bay 4 5 Dunnet Bay Dunnet Bay 6 7 8 9 (deep water) 10 (deep water) Dunnet Bay 11 Repop 4 12 (deep water)
2004 2004 2004 2004 2004 2004 2004 2004 2004 2004 2004 2004 2004 2004 2004 2004 2004
2557 2505 1936 1763 2506 1979 244 789 1151 1979 2206 1074 682 2688 2131 3677 156
13 16 10 0 16 3 0 0 11 12 5 0 0 0 27 10 0
0.0051 0.0064 0.0052 0 0.0064 0.0015 0 0 0.0096 0.0061 0.0023 0 0 0 0.0128 0.0027 0
30784
123
ROV Total
Area (m2 )
No. of finds
Survey area
Radioactive particles in the seabed off Dounreay nuclear facility, Scotland
577
particles were found in this area which had been cleared previously by divers in May 2003, some 17 months earlier. Also of note in Table 5 is the absence of particles in two areas surveyed further offshore in deeper water (35–40 m water depth) and in four surveys some 20 km further east in Dunnet Bay. While the present ROV trial has been limited in duration and geographical extent, the total area surveyed is quite extensive, over 30,000 m2 . The survey results obtained thus far are consistent with an offshore particle plume which is oriented north-east of the Dounreay site effluent chamber, located 600 metres offshore in 23 metres water depth. Further offshore particle surveys are planned for summer 2005 using the same or an improved remotely operated system to map particle locations in key areas of interest. Particle maps will provide essential information on the extent of transport of radioactive particles, will allow decisions to be made on related programmes of beach monitoring and will provide validation data for particle transport models.
4. Conclusions The use of sodium iodide radiation detectors for the underwater detection and identification of radioactive particles buried in marine sediment has been successfully illustrated in two field trials near Dounreay, northern Scotland. Using a large sodium iodide crystal within a stainless steel jacket, coupled to proprietary data processing software and with the detector mounted on a tracked ROV has permitted the locations of buried particles to be identified and mapped. Further use of the ROV-detector system within the assumed plume and wider afield will provide an opportunity to establish the geographical spread of particles and, consequently, to assess the likelihood of contamination of public beaches in the area.
References Atkinson, T.C. (2002). Depth and size distributions of off-shore radioactive particles at Dounreay, northern Scotland. Dounreay Particles Advisory Group paper no. DPAG/2002/008. Innes, S.C. (2004). Report on a laboratory test of a marinised 4 × 4 × 16 sodium iodide detector coupled to a SAM-935 surveillance and measurement system. UKAEA Report EPD(04)P178. Jones, D.J. (2001). Development and applications of offshore gamma-ray spectrometry measurements: A review. Journal of Environmental Radioactivity 53, 313–333. Koskelainen, M. (2005). The Dounreay particles resume. UKAEA Report EPD(03)P107. Maucec, M., de Meijr, R.J., van der Klis, M.M., Hendriks, P.H., Jones, D.G. (2004). Detection of radioactive particles offshore by gamma-ray spectrometry – II: Monte Carlo assessment of acquisition times. Nuclear Instruments & Methods in Physics Research A 525, 610–622. McQuaid, J.H. (2002). SAM’s birth and milestones to maturity. American Radiation Safety Conference and Exposition, June 16–20, 2002. Osvath, I., Povinec, P.P. (2001). Seabed gamma-ray spectrometry: Applications at IAEA-MEL. Journal of Environmental Radioactivity 53, 335–349. Povinec, P.P., Osvath, I., Baxter, M.S., Ballestra, S., Carroll, J., Gastaud, J., Harms, I., Huynh-Ngoc, L., Liong Wee Kwong, L., Pettersson, H. (1997). Summary of IAEA-MEL’s investigation of Kara Sea radioactivity and radiological assessment. Marine Pollution Bulletin 35, 235–241. Scirea, M. (2002). Dounreay offshore particles: Burial depth in the sediment distribution analysis. UKAEA Report EPMD(00)P16.
578
J. Toole et al.
Toole, J. (2004). Field trials of a marinised sodium iodide detector off Dounreay in September 2003. UKAEA Report EPD(04)P143. Toole, J. (2005). Review of instrumentation used for on-site radioactive monitoring at Dounreay. UKAEA Report EPD(04)P160.
11. Mass spectrometry techniques
This page intentionally left blank
581
Isotope selective ultratrace analysis of plutonium by resonance ionisation mass spectrometry Stefan Bürgera , Razvan Aurel Budaa , Horst Geckeisb , Gerhard Huberc , Jens Volker Kratza,* , Peter Kunzc , Christoph Lierse von Gostomskid , Gerd Passlerc , Ariane Remmertd , Norbert Trautmanna a Institut für Kernchemie, Universität Mainz, D-55099 Mainz, Germany b Institut für Nukleare Entsorgung, Forschungszentrum Karlsruhe, D-76021 Karlsruhe, Germany c Institut für Physik, Universität Mainz, D-55099 Mainz, Germany d Institut für Radiochemie, Technische Universität München, D-85748 Garching, Germany
Abstract Resonance ionisation mass spectrometry (RIMS) is a sensitive and selective method for isotopically resolved ultratrace analysis of long-lived radionuclides. For the routine analysis of plutonium three titanium–sapphire lasers pumped by a pulsed Nd:YAG laser in combination with a time-of-flight mass spectrometer are used. The detection limit of this system is as low as 106 –107 atoms for the plutonium isotopes 238 Pu to 244 Pu. The RIMS technique was applied to investigate the isotopic composition and the content of plutonium in a depleted uranium penetrator as used during the Balkan conflict delivering important information on the origin of the depleted uranium in this type of ammunition. Furthermore, groundwater samples from an in-situ experiment performed at the Grimsel Test Site, Switzerland, have been analysed with the aim to study the migration behaviour of plutonium under natural conditions. Keywords: Resonance ionisation mass spectrometry, Plutonium isotopes, Depleted uranium penetrator, Migration, Groundwater
1. Introduction Many radioactive elements of natural origin are relatively common in the Earth’s crust, like thorium or uranium, while others are mostly or entirely man-made, like technetium, neptunium, plutonium, or americium. Plutonium, which was discovered in 1941 by Seaborg and co-workers (Seaborg et al., 1946) was and is still released into the environment mainly as a result of nuclear weapons tests, releases from nuclear facilities, and from accidents. Beginning with the first man-made nuclear detonation at Alamogordo, New Mexico, 1945, the atmospheric weapon’s testing programmes introduced the major amount of plutonium onto the Earth’s surface (Hanson, 1980; Pentreath, 1995). Another world-wide source was * Corresponding author. Address: Institut für Kernchemie, Universität Mainz, D-55099 Mainz, Germany; phone: (+49) 6131 39 25704; fax: (+49) 6131 39 24510; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08046-0
© 2006 Elsevier Ltd. All rights reserved.
582
S. Bürger et al.
the burn-up of satellites equipped with 238 Pu-nuclide batteries (Hanson, 1980). Furthermore, plutonium was released by accidents, e.g., the loss of atomic weapons at Thule, Greenland, and Palomares, Spain (Mitchell et al., 1997), by sunken submarines with nuclear reactors, or during the Chernobyl accident in 1986. Releases from nuclear reprocessing plants, nuclear power facilities and, handling of waste from nuclear reactors, transport and storage of spent fuel and highly radioactive waste, have further increased the amount of plutonium in the environment. 239 Pu produced by neutron capture in uranium and extremely small amounts of primordial 244 Pu (Hoffman et al., 1971) have been found in nature. The isotopic composition of plutonium is characteristic for its origin (Hanson, 1980; IAEA, 1998). Thus, plutonium from nuclear power plants can be distinguished from global fallout or weapons-grade plutonium if the isotopic composition is known. Therefore, sensitive and isotope selective methods for plutonium detection are required for nuclear forensic and safeguard analysis, low-level environmental monitoring, personnel dose monitoring, or studies on the migration behaviour of plutonium, especially in the context of nuclear waste disposal. Alpha-spectrometry conventionally used for ultratrace analysis of plutonium is not well suited for isotopically resolved detection. Due to their very similar alpha-energies, 239 Pu and 240 Pu can hardly be distinguished and the beta-emitter 241 Pu cannot be directly detected by alpha-spectrometry. The detection limit depends on the half-life of the isotope, yielding a detection limit of, e.g. 4 × 108 atoms 239 Pu in 1000 min counting time (Peuser et al., 1981). This detection limit can be improved by a longer counting time. The detection limits of mass spectrometric methods, like inductively coupled plasma mass spectrometry (ICP-MS) (Becker and Dietze, 1999; Huber et al., 2003), are independent of the half-life of the isotope, but the analysis may suffer from isobaric interferences. In the case of 238 Pu and 241 Pu, the interferences are mainly caused by 238 U and 241 Am, but the formation of cluster ions such as 238 UH+ may disturb the analysis of 239 Pu (Truscott et al., 2001). Very recently, isotopic ratio measurements with a detection limit of less than 109 atoms of plutonium using a multi collector sector-field ICP-MS systems have been achieved (Truscott et al., 2001; Boulyga et al., 2001; Taylor et al., 2003). Accelerator mass spectrometry (AMS) as an alternative technique has a detection limit of down to 104 atoms of plutonium (Fifield et al., 1996; McAninch et al., 2000; Oughton et al., 2004) with extremely high isotopic selectivity, but the experimental equipment is rather complicated and costly. Very recently table-top AMS systems with low terminal voltages yielded detection limits down to 106 atoms of plutonium (Fifield et al., 2004) have been introduced. Other groups performed nuclear forensic studies of plutonium by SIMS and TIMS (Betti et al., 1999; Wallenius and Mayer, 2000). Resonance ionisation mass spectrometry (RIMS) was proposed for trace analysis almost three decades ago (Hurst and Payne, 1988; Lethokov, 1987). In the meantime, this technique has developed to an excellent method for the isotopically selective ultratrace analysis of longlived radionuclides (Huber et al., 2003; Wendt et al., 2000) with detection limits down to 106 atoms. Among others, RIMS was applied for ultratrace analysis of plutonium and other actinide isotopes (Passler et al., 1997; Nunnemann et al., 1998; Ham and Harrison, 2000; Grüning et al., 2004; Trautmann et al., 2004) and for the spectroscopic determination of the first ionisation potential of various actinides (Köhler et al., 1997; Erdmann et al., 1998).
Analysis of plutonium by resonance ionisation mass spectrometry
583
In this paper, examples for isotopically selective plutonium analysis in nuclear forensic and radionuclide migration studies will be presented.
2. Materials and methods 2.1. Resonance ionisation mass spectrometry The atomic energy levels are typical for each element and even for a single isotope of an element. By using a multi-step resonant laser excitation and ionisation of the element of interest in the atomic state with subsequent mass analysis, resonance ionisation mass spectrometry (RIMS) provides an excellent element and isotope selectivity, due to a complete suppression of isobaric or molecular interferences, as well as a good detection limit and a high sensitivity. For the ultratrace analysis of plutonium in environmental samples by RIMS, different steps are required. A simple chemical procedure separates the plutonium from the sample matrix. After electrodeposition on a filament, the plutonium is atomised by thermal evaporation into vacuum. The atomic beam is crossed by the laser light of the titanium–sapphire lasers and the plutonium atoms are excited into a Rydberg state by a three-step three-colour resonant excitation and are subsequently field ionised. The plutonium ions are mass-selectively detected after passing a reflectron time-of-flight (TOF) mass spectrometer by a channel plate detector (Grüning et al., 2004). 2.2. Sample preparation 2.2.1. Chemical treatment of a depleted uranium penetrator In collaboration with the Institut für Radiochemie, Technische Universität München, samples of depleted uranium (DU) ammunition as used during the Balkan conflict (Kosovo) were investigated with respect to their isotopic composition and content of plutonium. The sample treatment and the chemical separation of plutonium from a dissolved penetrator was performed at the TU München (Remmert, 2002). The surface of the penetrator (96.11 g) was cleaned by rinsing with concentrated HNO3 and washing with water. After drying, the residual uranium (92.57 g) was dissolved in concentrated HNO3 and concentrated HF. The solution was evaporated to dryness and dissolved in 2 M HNO3 . Insoluble TiO2 was removed by centrifugation. The solution was divided in two fractions. 236 Pu tracer was added to one fraction for the determination of the chemical yield. The further chemical treatment of both fractions was identical. Lanthanum nitrate and hydroxylamine hydrochloride to reduce the plutonium to Pu(III) were added to the solution, and the plutonium was co-precipitated with LaF3 . The precipitate was separated by centrifugation and dissolved with warm, saturated H3 BO3 and concentrated HCl in an ultrasonic bath. The solution was passed through an ion exchange column (Dowex 1 ×2) and washed with concentrated. HNO3 . Plutonium was eluted from the column using 8 M HCl plus NH4 I. The eluate was evaporated to dryness and iodine was fumed off with concentrated. HNO3 . The residue was dissolved in 2 M HNO3 . 244 Pu was added as yield tracer for the RIMS measurements. The solutions were transferred to electrolytic cells for the preparation of the filaments.
584
S. Bürger et al.
2.2.2. Chemical preparation of groundwater samples from the Grimsel Test Site Groundwater samples were collected at the Grimsel Test Site and stored in polyethylene bottles after acidification with ultrapure HNO3 to a final pH <1. All samples were analysed by ICP-MS. For RIMS analysis, seven samples from the CRR#31 experiment (containing 242 Pu) and six samples from the CRR#32 experiment (containing 244 Pu) were available. 50 ml aliquots of the CRR#31 and CRR#32 samples were spiked with highly enriched 244 Pu (1.4 × 1010 atoms) and 240 Pu (1.67 × 1010 atoms), respectively, as yield tracers for RIMS. 236 Pu was added to determine the chemical yield. The solutions were evaporated to dryness and dissolved in 4 M HNO3 . These solutions were passed through columns filled with TEVA·SPEC SPS resin (Aliquat 336N). After washing with 4 M HNO3 , plutonium was eluted from the column with 0.5 M HCl. The eluates were used for preparing the filaments. 2.2.3. Preparation of filaments After chemical separation, the plutonium was deposited electrolytically as Pu(OH)4 onto a tantalum backing (3 mm spot in diameter) from a 20% (NH4 )2 SO4 solution of pH = 1.5, using a current of 3.7 A/cm2 and a deposition time of 60–90 min. The filament was covered with a ≈ 1 µm thick layer of titanium metal produced by sputtering. Inside the TOF spectrometer, the filament was heated from 1150 to 1400 K. At this temperatures, Pu(OH)4 is converted to plutonium oxide (Eichler et al., 1997). During diffusion through the titanium layer, plutonium oxide is reduced to Pu, and an atomic beam is evaporated from the filament surface into vacuum. 2.3. Experimental set-up The RIMS set-up is shown in Fig. 1 (Grüning et al., 2004). It consists of a reliable high repetition rate solid state laser system and a reflectron time-of-flight (TOF) mass spectrometer. Three titanium–sapphire lasers pumped by a pulsed Nd:YAG laser provide up to 2 W of laser light in a wavelength range from 740 to 880 nm at a repetition rate of 6.6 kHz. As a pump source, a frequency-doubled Nd:YAG laser (Clark–MXR ORC-1000) with an average output power of up to 50 W at a wavelength of 532 nm is used. The linewidth of the titanium– sapphire lasers is about 3 GHz. In order to excite and ionise plutonium (ionisation potential IP = 6.02 eV) in a three-step three-colour excitation scheme, the frequency of the titanium– sapphire laser light for the first excitation step is doubled in a special crystal (BBO). To obtain a good spatial overlap with the atomic beam, the three laser beams are simultaneously guided in a quartz fibre to the interaction zone. The plutonium atoms are excited into a Rydberg state by three-step resonant excitation with λ1 = 420.76 nm, λ2 = 847.28 nm and λ3 = 767.53 nm and are subsequently field ionised. The isotope shifts for the isotopes 238 Pu through 244 Pu are between 5 to 20 GHz for the first and the second step, i.e. the wavelengths of the laser 1 and 2 have to be adjusted for each isotope. No isotope shift for the third step is observed. The ions are mass-selectively separated in a reflectron time-of-flight (TOF) mass spectrometer and counted by channel plates. The TOF mass resolution of m/m ≈ 600 is sufficient to discriminate the different isotopes of plutonium. The overall efficiency of the RIMS apparatus is 10−5 , resulting in the detection limit of −6 10 –107 atoms.
Analysis of plutonium by resonance ionisation mass spectrometry
585
Fig. 1. Experimental set-up for RIMS measurements with a pulsed Nd:YAG pump laser, three titanium–sapphire lasers, and a reflectron type time-of-flight mass spectrometer.
3. Results and discussion 3.1. RIMS for nuclear forensic studies: determination of isotopic composition and amount of plutonium in a depleted uranium penetrator as used during the Balkan conflict The discussion on the health risks of depleted uranium (DU) ammunition caused a series of analytical investigations, e.g. by the United Nation Environmental Programme (UNEP, 2001). DU was used as armour-piercing ammunition in several military conflicts, for example in Iraq and Kuwait 1991 (about 321 t of DU), Bosnia–Herzegovina (about 3 t DU) and Kosovo 1999 (about 10 t DU) (Bleise et al., 2003). There is no acute external health risk due to DU ammunition, because the specific radioactivity is rather low, except for long-lasting contact of a penetrator with the skin, which may cause a notable dose (Pöllänen et al., 2001). When hitting a hard target, the impact of the penetrator can generate DU dust, which may lead to exposure of the lung and other organs due to inhalation. Several studies showed, that 239+240 Pu and also 236 U were present in the DU ammunition, indicating that the depleted uranium used may stem from a batch that had been neutronirradiated and then reprocessed (Danesi et al., 2003; McLaughlin et al., 2003; Pöllänen et al., 2001; Burger and Schmid, 2001; Boulyga et al., 2001). Figure 2 shows the mass spectrum of the Pu-fraction from the depleted uranium penetrator with the emphasis on 242 Pu and 244 Pu. The other isotopes 239 Pu, 240 Pu, and 241 Pu are partly suppressed due to the isotopic shifts, but still visible in the mass spectrum. The com-
586
S. Bürger et al.
Fig. 2. Mass spectrum of the isotopic ratio measurement for 242 Pu/244 Pu. The other isotopes are suppressed due to their isotope shifts, but still visible in the mass spectrum. Table 1 Isotopic ratios of plutonium in DU ammunition as measured by RIMS, compared with published values for weapons-grade plutonium
239 Pu 240 Pu 241 Pu 242 Pu
Fraction of Pu as measured by RIMS (%) (3σ confidence level)
Fraction in weapons-grade plutonium (%) according to Fetter et al. (1990)
93.6 ± 5.9 6.2 ± 0.7 0.09 ± 0.03 0.06 ± 0.01
93.5 6.0 0.44 0.015
position has been determined by RIMS to 239 Pu: (93.6 ± 5.9)%, 240 Pu: (6.2 ± 0.7)%, 241 Pu: (0.09 ± 0.03)%, and 242 Pu: (0.06 ± 0.01)%. The isotope ratios correspond very well with those published by Fetter et al. (1990) for weapons-grade plutonium (Table 1). Assuming that the difference between the observed 0.09% and the published 0.44% for 241 Pu results from the radioactive decay (T 1/2 = 14.35 yr), the plutonium must have been 33 years old (dated 2002). The average age of weapon grade plutonium in the US stockpile is 33 years (dated 2002) (Fetter et al., 1990), in good agreement with our calculated value. According to the ADHOC Committee on Depleted Uranium (ADHOC, 2001): “The machinery used for the enrichment process was also used in the 1950–1970s to enrich uranium extracted from recycled reactor fuel. This resulted in the contamination of those facilities with amounts of transuranics, uranium-236 and technetium. These trace amounts were picked up in the DU processed in the facility. In addition, a small fraction of the raw material used for producing our DU came from the uranium extracted from reactor fuel”.
Analysis of plutonium by resonance ionisation mass spectrometry
587
The amount of 239 Pu determined in the penetrator is 3.5 × 1012 atoms for 92.57 g uranium or 15 pg 239 Pu/g DU, i.e. 15 ppt. The amount of 239+240 Pu corresponds to 43 mBq/g. McLaughlin et al. (2003) report a plutonium analysis carried out on a depleted uranium penetrator recovered from a target site in southern Serbia. Here, the activity concentration of 239+240 Pu is (45.4 ± 0.7) mBq/g. 3.2. RIMS for radionuclide migration studies: colloid-mediated plutonium migration in a granitic shear zone It is expected that plutonium introduced into the environment exists mainly in the tri- and tetravalent state and thus is immobile due to its low solubility in groundwater and strong sorption onto mineral surfaces (Allard et al., 1984). But mobile colloids – suspended particles in the sub-micrometre size onto which plutonium is sorbed – can enhance the transport of plutonium in groundwater (Allard et al., 1984; Honeyman, 1999). This has been observed at the Nevada Test Site, USA, where quite a number of underground nuclear tests have been carried out. The 240 Pu/239 Pu isotope ratio found in ground water samples 1.3 km north of an underground nuclear test site indicates that the plutonium in the water stems from the respective site and ultrafiltration studies reveal that colloids have played an important role in transporting the plutonium (Kersting et al., 1999). Detailed knowledge on the migration behaviour of plutonium in aquifer systems is of special interest and concern for the long-term safety assessment of nuclear waste disposal concepts. Therefore, in-situ migration experiments have been performed in the frame of an international cooperation at NAGRA’s (Nationale Genossenschaft für die Lagerung radioaktiver Abfälle) underground laboratory in Grimsel, Switzerland. The main aim of the “colloid and radionuclide retardation” experiment (CRR) was to investigate the migration behaviour of actinide ions such as plutonium in a granite fracture under natural conditions and notably to study the impact of aquatic colloids. More detailed information on those experiments is available from Hauser et al. (2002), Möri et al. (2003), and Geckeis et al. (2004). 242 Pu tracer (CRR#31 experiment without bentonite colloids) and 244 Pu tracer (CRR#32 experiment with 20 mg/L clay colloids) dissolved in groundwater were injected as Pu(IV) into a well characterised granitic shear zone with an induced water flow. Samples have been collected at different times in the outflow of the experimental fracture. The plutonium concentration was measured by ICP-MS at FZK and for samples at or below the detection limit of ICP-MS (∼0.1 ng/L) with RIMS, down to the detection limit of 107 atoms (∼8 × 10−5 ng/L). Figure 3 shows the tracer breakthrough curves for 242 Pu migration in absence and for 244 Pu transport in presence of bentonite colloids. The clay colloids (20 mg/L) have been prepared from a natural bentonite that is discussed as an appropriate backfill material to encapsulate nuclear waste in a repository located in crystalline host rock. The plutonium concentration in the samples is plotted versus the retention time. The maximum of the tracer breakthrough curve obtained with bentonite colloids is higher than the one without bentonite colloids. The congruence of bentonite colloid and plutonium breakthrough curves (Möri et al., 2003) clearly indicates that the plutonium transport under given conditions is enhanced in the presence of such colloids. But even without adding bentonite colloids, part of Pu(IV) appears to migrate in colloidal form (Möri et al., 2003). In this case, groundwater colloids are assumed to be responsible for the observed unretarded plu-
588
S. Bürger et al.
Fig. 3. Tracer breakthrough curves of Pu(IV) in a granitic fracture without and with bentonite colloids.
tonium migration (Geckeis et al., 2004). The clear effect of colloidal species on plutonium migration is due to the Grimsel groundwater composition (high pH, low ionic strength) leading to the strong stabilisation of aquatic colloids and the relatively high induced groundwater flow velocity applied in those studies. Only by RIMS analysis, it was possible to observe an increased plutonium level at later retention times in the CRR#31 experiment (absence of bentonite colloids) as compared to that in the CRR#32 experiment. Such behaviour indicates the release of strongly retarded plutonium previously absorbed on the surface of the rock. In the presence of the bentonite colloids, a permanent decrease of the plutonium concentration down to the detection limit of RIMS has been found. A quantitative interpretation of those data is currently underway. It should be noted that actinide migration experiments under relevant natural conditions are necessary in order to validate results of laboratory experiments and the predictions made by geochemical modelling. Realisation of such type of studies in the field calls for: (i) compliance with radiation protection regulations, requiring actinide concentrations below permitted exemption levels; (ii) the use of geochemically relevant actinide concentrations below the respective solubility limits of 10−8 mol/L depending on groundwater composition. ICP-MS and notably the ultrasensitive RIMS are considered as valuable sensitive isotope analysis techniques of high importance for such type of studies.
4. Conclusions RIMS is a remarkable tool for ultratrace analysis of plutonium in various samples due to its low detection limit in the femtogram region and its extreme element and isotope selectivity. The reliable and routine application of RIMS to environmental samples has been demonstrated.
Analysis of plutonium by resonance ionisation mass spectrometry
589
Acknowledgement This work was partly funded by the Graduiertenkolleg “Spurenanalytik von Elementspezies: Methodenentwicklungen und Anwendungen”, GRK 826/1, Deutsche Forschungsgemeinschaft, Universität Mainz.
References AHCDU, ADHOC Committee on Depleted Uranium (2001). U.S. information paper on depleted uranium. 23 January 2001. Allard, B., Olofsson, U., Torstenfelt, B. (1984). Environmental actinide chemistry. Inorganic Chimica Acta 94, 205– 221. Becker, S.J., Dietze, H.-J. (1999). Precise isotope ratio measurements for uranium, thorium and plutonium by quadrupole-based inductively coupled plasma mass spectrometry. Fresenius Journal of Analytical Chemistry 364, 482–488. Betti, M., Tamborini, G., Koch, L. (1999). Use of secondary ion mass spectrometry in nuclear forensic for the characterization of plutonium and highly enriched uranium particles. Analytical Chemistry 71, 2616–2622. Bleise, A., Danesi, P.R., Burkart, W. (2003). Properties, use and health effects of depleted uranium (DU): A general overview. Journal of Environmental Radioactivity 64, 93–112. Boulyga, S.F., Testa, C., Desideri, D., Becker, J.S. (2001). Optimisation and application of ICP-MS and alphaspectrometry for determination of isotopic ratios of depleted uranium and plutonium in samples collected in Kosovo. Journal of Analytical Atomic Spectrometry (JAAS) 16, 1283–1289. Burger, M., Schmid, E. (2001). Report of the Swiss Members of UNEP-Team. Spiez Laboratory. Danesi, P.R., Bleise, A., Burkart, W., Cabianca, T., Campbell, M.J., Makarewicz, M., Moreno, J., Tuniz, C., Hotchkis, M. (2003). Isotopic composition and origin of uranium and plutonium in selected samples collected in Kosovo. Journal of Environmental Radioactivity 64, 121–131. Eichler, B., Hübener, S., Erdmann, N., Eberhardt, K., Funk, H., Herrmann, G., Köhler, S., Trautmann, N., Passler, G., Urban, F.-J. (1997). An atomic beam source for actinide elements: Concepts and realization. Radiochimica Acta 79, 221–233. Erdmann, N., Nunnemann, M., Eberhardt, K., Herrmann, G., Huber, G., Köhler, S., Kratz, J.V., Passler, G., Peterson, J.R., Trautmann, N., Waldek, A. (1998). Determination of the first ionization potential of nine actinide elements by resonance ionization mass spectrometry. Journal of Alloys and Compounds 271–273, 837–840. Fetter, S., Frolov, V.A., Prilutsky, O.F., Sagdeev, R.Z. (1990). Fissile materials and weapon design. Science & Global Security 1, 225–302. Fifield, L.K., Cresswell, R.G., di Tada, M.L., Ophel, T.R., Day, J.P., Clacher, A.P., King, S.J., Priest, N.D. (1996). Accelerator mass spectrometry of plutonium isotopes. Nuclear Instruments and Methods in Physics Research B 117, 295–303. Fifield, L.K., Synal, H.-A., Suter, M. (2004). Accelerator mass spectrometry of plutonium at 300 kV. Nuclear Instruments and Methods in Physics Research B 223–224, 802–806. Geckeis, H., Schäfer, T., Hauser, W., Rabung, Th., Missana, T., Degueldre, C., Möri, A., Eikenberg, J., Fierz, Th., Alexander, W.R. (2004). Results of the colloid and radionuclide retention experiment (CRR) at the Grimsel Test Site (GTS), Switzerland – Impact of reaction kinetics and speciation on radionuclide migration. Radiochimica Acta 92, 765–774. Grüning, C., Huber, G., Klopp, P., Kratz, J.V., Kunz, P., Passler, G., Trautmann, N., Waldek, A., Wendt, K. (2004). Resonance ionisation mass spectrometry for ultratrace analysis of plutonium with a new solid state laser system. International Journal of Mass Spectrometry 235, 171–178. Ham, G.J., Harrison, J.D. (2000). The gastrointestinal absorption and urinary excretion of plutonium in male volunteers. Radiation Protection Dosimetry 87, 267–272. Hanson, W.C. (Ed.) (1980). Transuranic elements in the environment. DOE/TIC-22800. Technical Information Centre, U.S. Department of Energy, Washington, DC.
590
S. Bürger et al.
Hauser, W., Geckeis, H., Kim, J.I., Fierz, Th. (2002). A mobile laser-induced breakdown detection system and its application for the in situ-monitoring of colloid migration. Colloids and Surfaces A: Physicochemical Engineering Aspects 203, 37–45. Hoffman, D.C., Lawrence, F.O., Mewherter, J.L., Rourke, F.M. (1971). Detection of plutonium-244 in nature. Nature 234, 132–134. Honeyman, B.D. (1999). Colloidal culprits in contamination. Nature 397, 23–24. Huber, G., Passler, G., Wendt, K., Kratz, J.V., Trautmann, N. (2003). In: L’Annunziata, M.F. (Ed.), Handbook of Radioactivity Analysis. Second edition. Academic Press, San Diego, pp. 799–843. Hurst, G.S., Payne, M.G. (1988). Principles and Applications of Resonance Ionisation Spectroscopy. Adam Hilger, Bristol. IAEA (1998). Safe handling and storage of plutonium. Safety Reports Series no. 9. International Atomic Energy. Kersting, A.B., Efurd, D.W., Finnegan, D.L., Rokop, D.J., Smith, D.K., Thompson, J.L. (1999). Migration of plutonium in ground water at the Nevada Test Site. Nature 397, 56–59. Köhler, S., Deißenberger, R., Eberhardt, K., Erdmann, N., Herrmann, G., Huber, G., Kratz, J.V., Nunnemann, M., Passler, G., Rao, P.M., Riegel, J., Trautmann, N., Wendt, K. (1997). Determination of the first ionization potential of actinide elements by resonance ionization mass spectrometry. Spectrochimica Acta B 52, 717–726. Lethokov, V.S. (Ed.) (1987). Laser Photoionization Spectroscopy. Academic Press, Orlando. McAninch, J.E., Hamilton, T.F., Brown, T.A., Jokela, T.A., Knezovich, J.P., Ognibene, T.J., Proctor, I.D., Roberts, M.L., Sideras-Haddad, E., Southon, J.R., Vogel, J.S. (2000). Plutonium measurements by accelerator mass spectrometry at LLNL. Nuclear Instruments and Methods in Physics Research B 172, 711–716. McLaughlin, J.P., León Vintró, L., Smith, K.J., Mitchell, P.I., Žuni´c, Z.S. (2003). Actinide analysis of depleted uranium penetrator from a 1999 target site in southern Serbia. Journal of Environmental Radioactivity 64, 155–165. Mitchell, P.I., León Vintró, L., Dahlgaard, H., Gascó, C., Sánchez-Cabeza, J.A. (1997). Perturbation in the 240 Pu/239 Pu global fallout ratio in local sediments following the nuclear accidents at Thule (Greenland) and Palomares (Spain). The Science of the Total Environment 202, 147–153. Möri, A., Alexander, W.R., Geckeis, H., Hauser, W., Schäfer, T., Eikenberg, J., Fierz, Th., Degueldre, C., Missana , T. (2003). The colloid and radionuclide retardation experiment at the Grimsel Test Site: Influence of bentonite colloids on radionuclide migration in a fractured rock. Colloids and Surfaces A 217, 33–47. Nunnemann, M., Erdmann, N., Hasse, H.-U., Huber, G., Kratz, J.V., Kunz, P., Mansel, A., Passler, G., Stetzer, O., Trautmann, N., Waldek, A. (1998). Trace analysis of plutonium in environmental samples by resonance ionisation mass spectrometry (RIMS). Journal of Alloys and Compounds 271–273, 45–48. Oughton, D.H., Skipperud, L., Fifield, L.K., Cresswell, R.G., Salbu, B., Day, P. (2004). Accelerator mass spectrometry measurement of 240 Pu/239 Pu isotope ratios in Novaya Zemlya and Kara Sea sediments. Applied Radiation and Isotopes 61, 249–253. Passler, G., Erdmann, N., Hasse, H.U., Hermann, G., Huber, G., Köhler, S., Kratz, J.V., Mansel, A., Nunnemann, M., Trautmann, N., Waldek, A. (1997). Application of laser mass spectrometry for trace analysis of plutonium and technetium. Kerntechnik 62, 85–90. Pentreath, R.J. (1995). The analysis of Pu in environmental samples: A brief historical perspective. Applied Radiation and Isotopes 46, 1279–1285. Peuser, P., Gabelmann, H., Lerch, M., Sohnius, B., Trautmann, N., Weber, M., Herrmann, G., Denschlag, H.O., Ruster, W., Bonn, J. (1981). Detection methods for trace amounts of plutonium. IAEA-SM-252/40. IAEA, Vienna, pp. 257–262. Pöllänen, R., Ikäheimonen, T.K., Klemola, S., Vartti, V.-P., Vesterbacka, K., Ristonmaa, S., Honkamaa, T., Sipilä, P., Jokelainen, I., Kosunena, A., Zilliacus, R., Kettunen, M., Hokkanen, M. (2001). Characterisation of projectiles composed of depleted uranium. Journal of Environmental Radioactivity 64, 133–142. Remmert, A. (2002). Untersuchung von Kosovo–Uran mittels radiometrischer und massenspektroskopischer Verfahren. Diplomarbeit. Fachhochschule NTA, Isny. Seaborg, G.T., McMillan, E.M., Kennedy, J.W., Wahl, A.C. (1946). Radioactive element 94 from deuterons on uranium. Physical Review 69, 367. Taylor, R.N., Warneke, T., Milton, J.A., Croudace, I.W., Warwick, P.E., Nesbitt, R.W. (2003). Multiple ion counting determination of plutonium isotope ratios using multi-collector ICP-MS. Journal of Analytical Atomic Spectrometry (JAAS) 18, 480–484. Trautmann, N., Passler, G., Wendt, K.D.A. (2004). Ultratrace analysis and isotope ration measurements of long-lived radioisotopes by resonance ionization mass spectrometry (RIMS). Analytical and Bioanalytical Chemistry 378, 348–355.
Analysis of plutonium by resonance ionisation mass spectrometry
591
Truscott, J.B., Jones, P., Fairman, B.E., Evans, E.H. (2001). Determination of actinide elements at femtogram per gram levels in environmental samples by on-line solid phase extraction and sector-field-inductively coupled plasma-mass spectrometry. Analytica Chimica Acta 433, 245–253. UNEP (2001). Depleted uranium in Kosovo – post-conflict environmental assessment. United Nations Environmental Programme. Wallenius, M., Mayer, K. (2000). Age determination of plutonium material in nuclear forensic by thermal ionisation mass spectrometry. Fresenius Journal of Analytical Chemistry 366, 234–238. Wendt, K., Trautmann, N., Bushaw, B.A. (2000). Resonant laser ionization mass spectrometry: An alternative to AMS? Nuclear Instruments and Methods in Physics Research B 172, 162–169.
592
Two 60-year records of 129I from coral skeletons in the South Pacific Ocean D.L. Biddulph* , J.W. Beck, G.S. Burr, D.J. Donahue Department of Physics and NSF-Arizona AMS Facility, The University of Arizona, Tucson, AZ 85721, USA Abstract 129 I is an important radionuclide tracer for certain natural and anthropogenic nuclear processes. 129 I has a half-life of 15.7 Myr and can be measured by accelerator mass spectrometry (AMS). This paper presents 129 I results made at
the University of Arizona with a NEC 3 MV Pelletron accelerator. For this study, we selected living corals from the Solomon Islands and Easter Island to monitor increases in anthropogenic 129 I in the surface waters of the Pacific Ocean. 129 I/127 I values were measured in cores taken from massive Porites head coral skeletons. Typical sample sizes for this study were 10 g for the Solomon Islands corals, and 30 g for the Easter Island corals. Temporal resolution was semi-annual at the Solomon Islands, and annual at Easter Island. Iodine was extracted from the corals without the use of carrier iodine. Results of our study produced records at both sites from roughly 1935 to 1996. While 129 I/127 I values have increased at both locations since the beginning of atmospheric nuclear weapons testing, different bomb-pulse curves at these sites suggest different transport mechanisms and/or 129 I inputs at the two sites. The implications of these measurements are discussed below. Keywords: 129 I, Accelerator mass spectrometry, Corals
1. Introduction Coral skeletons provide an archive of the chemical and physical conditions present in the surface waters of the ocean that existed when the material was deposited. Reef building corals live in shallow waters, generally within the upper 50 to 75 meters, and border islands or continental margins that bracket the latitudes 35◦ N to 30◦ S (Druffel, 1997). Coral records have been used extensively to study past changes in ocean circulation, ocean chemistry and climate. The analysis of uranium, thorium and 14 C in corals has been used to extend the atmospheric calibration curve used for 14 C dating beyond the dendrochronological limit of ∼ =12,400 years before present (Stuiver et al., 1998). The use of coral skeletons as a record of radioiodine has some advantages over other environmental iodine reservoirs, including: enhanced time resolution due to relatively rapid deposition (10 to 20 mm/year), the absence of mixing processes that commonly affect sediments * Corresponding author. Address: 1118 E. 4th St. PAS 81, Tucson, AZ 85721, USA; phone: (+520) 621 6825; fax: (+520) 626 9348; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08047-2
© 2006 Elsevier Ltd. All rights reserved.
Two 60-year records of 129 I from coral skeletons in the South Pacific Ocean
593
(such as bioturbation), and the existence of dating techniques (Edwards et al., 1987), to establish reliable chronologies for particular specimens. For living corals, annual bands preserved in their skeletons can be counted for age identification. Ages of older corals can be determined by measuring the parent/daughter ratio of 234 U and 230 Th (Edwards et al., 1987). The purpose of this study is to assess the possibility of using coral skeletons as reliable recorders of the 129 I content of surface ocean waters during the time of skeletal formation. In order for this to work, the iodine must exist in substantial quantities (at least the ppb level) and be immobile within the coral skeleton once incorporated.
2. Site selection and sampling Both sampling sites were originally chosen as part of radiocarbon studies that looked at dissolved inorganic carbon (DIC) in surface ocean waters (Schmidt et al., 2004; Beck, 2004). The Solomon Islands skeletal core was retrieved in September 1994 from Marau Sound on the eastern tip of Guadalcanal, latitude 9.5◦ S and longitude 162◦ E. This location proves interesting for 129 I studies as it is presumably far from any point sources of anthropogenic iodine. In addition, it is located in the Southern Hemisphere where few anthropogenic iodine studies have been done to date. Easter Island is located at the center of the South Pacific gyre, latitude 27◦ S and longitude 109◦ W. Also far removed from anthropogenic sources of 129 I, this location was sampled as a comparative study site for the core drilled from the Solomons. This skeletal core was retrieved while the coral colony was still active, in the fall of 1996. The coral samples were obtained with an underwater hydraulic drill that was equipped with an 8 cm diameter bit. Cores were drilled perpendicular to the coral growth surface. A 5 mm slab was cut from each core and X-rayed to identify annual growth bands. These bands were marked for 129 I sampling in order to obtain a continuous, integrated record. The Solomon Islands core was sampled along growth bands for semiannual resolution, while larger samples were used for the Easter Island core to obtain annual resolution. The chronology for the Easter Island core was obtained by counting annual density bands where they were clearly identifiable in the X-ray record (the latest 15 years or so). Afterward, a linear growth rate was assumed. This chronology was then checked to ensure agreement with that obtained by counting seasonal cycles in the δ 18 O record from the same core (Beck, 2004; Mucciarone and Dunbar, 2003). The Solomon Islands chronology was initially determined by counting annual density bands only (Biddulph, 2004), but was subsequently adjusted to fit both the δ 18 O record and the 129 I data that was obtained from Easter Island.
3. Sample preparation To remove external contamination, the coral samples were acid washed and sonicated for 45 minutes, then dissolved in a five to one solution of distilled water and 85% H3 PO4 , with an acid to coral ratio of 2.1 ml per gram of sample. This leaves the solution with a low pH, which is required to oxidize existing I− ions to I2 molecules. Particulates were removed by filtration through a 0.22 µm cellulose acetate membrane. After filtration, the solution was poured into
594
D.L. Biddulph et al.
a separatory funnel with the addition of 5–10 ml of CHCl3 and one drop of 1 M NaNO2 . The solution was shaken for one minute, and time was allotted for the CHCl3 plus I2 mixture to settle to the bottom of the funnel. This oxidation and extraction process was repeated three times, or until no color change due to dissolved I2 was noticeable in the CHCl3 . At this point, we took advantage of the fact that silver powder has an affinity for I2 molecules (Yiou et al., 2004). The CHCl3 plus I2 mixture was drained into a glass vial containing 10–15 mg of 120 mesh Ag powder. Within a couple of hours the I2 is transferred from the CHCl3 to the Ag powder, and this transfer is evident as the CHCl3 loses its purple hue and the Ag powder darkens in color. Upon completion, the CHCl3 was allowed to evaporate in a fume hood, and the Ag plus I2 mixture was rinsed three times in distilled H2 O and allowed to dry. This material was then pressed into an aluminum cathode for AMS measurement. Carrier iodine was not used for target preparation. A 50 ml aliquot of solution was saved for each coral sample for stable iodine measurements with an iodide selective electrode. The measured iodine content in the Solomon Islands coral samples varied little between individual samples from a mean of (3.4 ± 0.1) ppm. The biophile nature of iodine is evident in these samples, as this iodine concentration is elevated nearly 100 times with respect to the iodine content of surface seawater (50 ppb). Stable iodine measurements of the Easter Island samples are pending, but results from chemical extraction and ion source current yields suggest that the concentrations are similar to those measured from the Solomon Islands.
4. AMS measurements The AMS measurements were performed at The University of Arizona with a NEC 3 MV Pelletron accelerator. Nominal 127 I− ion source currents produced from the coral samples were in the 500–1000 nA range. Measurements were made in the +5 charge state, at a terminal voltage of 2.75 MV with a transmission of 4.0%. Each sample was measured six times, and the weighted average of the measurements was calculated. Errors include statistical uncertainties and a random machine error of 4.3%. A blank correction factor of 129 I/127 I = (1.0 ± 0.1) × 10−12 due to the chemical extraction and filtration process was applied to all coral samples. A more detailed description of the AMS measurements performed at this facility has been published previously (Biddulph et al., 2000; Biddulph, 2004).
5. Results and discussion The results of the radioiodine measurements for both skeletal cores are plotted together in Fig. 1. Both records show equivalent pre-bomb pulse levels of 129 I/127 I ≈ 5 × 10−13 , confirming the expectation of ocean homogeneity given the very long half-life of 129 I (15.7 Ma) relative to an ocean mixing time of roughly 1000 years (Fabryka-Martin et al., 1984). Approximately half of this natural 129 I comes from atmospheric spallation reactions involving cosmic rays and 129 Xe, with the rest coming from fission of natural uranium. 129 I increases at similar rates in both records from the late 1950s to about 1970 in consequence of fallout from atmospheric nuclear weapons testing which disseminated large
Two 60-year records of 129 I from coral skeletons in the South Pacific Ocean
595
Fig. 1. 129 I/127 I values from coral skeletons in the South Pacific Ocean.
amounts of fissiogenic 129 I throughout the atmosphere. Whereas atmospheric nuclear testing virtually ceased in 1963 as a result of the Nuclear Test–Ban Treaty, 129 I levels in both records increase continuously at the same linear rate between 1963 and 1970. Beginning in 1970 129 I/127 I values remained nearly unchanged in both records, until 1975, when the Solomon Islands record began to rise at a rapid rate. In contrast, 129 I/127 I in the Easter Island record has remained nearly constant, or has risen only slightly since 1975. The response asymmetry between the Solomon Islands site and Easter Island suggests that the large signal observed at Guadalcanal after 1975 resulted from one or more point source releases of radioiodine into the ocean mixed layer. We believe that at least some of the rapid rise observed after 1975 in the Solomon Islands record may be due to contaminated wastewater releases into the Columbia River from the Hanford nuclear fuels reprocessing center in the state of Washington (Heeb et al., 1996). The signal could have arrived at the Solomons by propagating southward via the coastal California Current system into the North Equatorial Current which in turn helps to feed the Pacific Equatorial Countercurrent, some of which passes the vicinity of the Solomon Islands. Weak cross-equatorial advection of these water masses would explain the much-attenuated post-1975 radioiodine signal reaching Easter Island, which is located near the center of the South Pacific gyre. This being stated, we must emphasize the difficulties involved in locating specific iodine sources responsible for the signals seen in the South Pacific. Other possible sources may include nuclear test remnants from regional island testing in the 1970s, or point discharge sources from the eastern Asian coastline. In contrast to the behavior exhibited after 1975, both records exhibit virtually identical behavior between 1958 and 1970. Because these two sites have strongly dissimilar hydrographic settings, we could not explain this coherence if 129 I were being transported from a point source via ocean water masses. Instead an atmospheric transport pathway must be invoked, which can produce a fairly homogeneous 129 I distribution. The residence time of iodine in the troposphere is quite short, however (ca. two weeks) (Kocher, 1981), which
596
D.L. Biddulph et al.
would not only lead to a heterogeneous deposition pattern, but we could not also explain the decade long period of coherence exhibited in these records between 1960 and 1970. Thus, most of the 129 I deposited in the Pacific Ocean mixed-layer during this time interval may have been stratospheric bomb pulse fallout, where iodine could have a much longer residence time. Using 36 Cl as an analogue for bomb 129 I, we would expect the stratospheric residence time to be on the order of two to four years (Synal et al., 1990; Elmore et al., 1982). However, such a short residence time would not produce the nearly linear rate of increase in 129 I/127 I observed for the period 1963 to 1970. On the other hand, if the residence time were longer, we would not be able to explain the period between 1970 and 1975 during which 129 I/127 I was not increasing in either record. This somewhat puzzling finding suggests that there was likely more than one source of atmospheric 129 I during this period. One of these must have been stratospheric fallout from bomb testing, but the other is less certain. Tropospheric 129 I emissions from Hanford, Sellafield and La Hague were substantial during this time interval (Heeb et al., 1996; Wagner et al., 1996; Lopez-Gutierrez et al., 2004), and it is possible that these releases conspired to mask the exponential decay in the stratospheric fallout signal expected from bomb 129 I. Additional records may help clarify this situation. It is instructive to compare these plots with their radiocarbon records, Fig. 2 (Schmidt et al., 2004; Beck, 2004). In striking contrast to the 129 I records, the 14 C values for the waters around Easter Island are elevated roughly 25% as compared to the Solomon Island waters. These results are consistent with ocean surface water measurements (Linick, 1980; Key et al., 2004) of 14 C, which show a north–south 14 C gradient with maxima at roughly 30◦ N and 30◦ S, and a minimum in the equatorial regions. The contrast between the radiocarbon and radioiodine curves at these two sites is a clear indication of different sources and transport mechanisms for these two anthropogenic isotopes.
Fig. 2. 14 C values from the same corals as in Fig. 1.
Two 60-year records of 129 I from coral skeletons in the South Pacific Ocean
597
6. Conclusion Coral skeletons preserve an iodine record in the tropical surface waters of the ocean. Once deposited within the skeleton, the iodine is fixed and may be retrieved with straightforward chemical extraction techniques. Comparison of anthropogenic iodine concentrations in skeletons from the Solomon Islands and Easter Island suggests different input functions and transport mechanisms for these two locations. Additional radioiodine records of corals at alternate locations should aid in the understanding of these transport phenomena, and may help pinpoint the locations of the most important point sources of the signals being observed. Coral skeletons, which are common throughout the Pacific, represent a valuable resource for understanding anthropogenic inputs and the transport of iodine in surface ocean currents of the world’s oceans.
References Beck, J.W. (2004). Unpublished data, personal correspondence. Physics Department, The University of Arizona. Biddulph, D.L. (2004). PhD dissertation. The University of Arizona. Biddulph, D.L., Beck, J.W., Burr, G.S., Donahue, D.J., Hatheway, A.L., Jull, A.J.T. (2000). Measurement of the radioisotope 129 I at the NSF-Arizona AMS laboratory. Nuclear Instruments and Methods in Physics Research B 172, 693–698. Druffel, E.R.M. (1997). Geochemistry of corals: Proxies of past ocean chemistry, ocean circulation, and climate. Proceedings of the National Academy of Sciences USA 94, 8354–8361. Edwards, R.L., Chen, J.H., Wasserburg, G.J. (1987). 238 U–234 U–230 Th–232 Th systematics and the precise measurement of time over the past 500,000 years. Earth and Planetary Science Letters 81, 175–192. Elmore, D., Tubbs, L.E., Newman, D., Ma, X.Z., Finkle, R.C., Nishiizumi, K., Beer, J., Oeschger, H., Andree, M. (1982). The 36 Cl bomb pulse measured in a shallow ice core from Dye 3, Greenland. Nature 300, 735–737. Fabryka-Martin, J., Bentley, H., Elmore, D., Airey, P.L. (1984). Natural iodine-129 as an environmental tracer. Geochimica et Cosmochimica Acta 49, 337–347. Heeb, C.M., Gydesen, S.P., Simpson, J.C., Bates, D.J. (1996). Reconstruction of radionuclide releases from the Hanford Site, 1944–1972. Health Physics 71, 545–555. Key, R.M., Kozyr, A., Sabine, C.L., Lee, K., Wanninkhof, R., Bullister, J.L., Feely, R.A., Millero, F.J., Mordy, C., Peng, T.-H. (2004). A global ocean carbon climatology: Results from GLODAP. Global Biogeochemical Cycles, in press. Kocher, D.C. (1981). A dynamic model of the global iodine cycle and estimation of dose to the world population from releases of iodine-129 to the environment. Environmental International 5, 15–31. Linick, T.W. (1980). Radiocarbon 22 (3), 599. Lopez-Gutierrez, J.M., Santos, F.J., Garcia-Leon, M., Schnabel, Ch., Synal, H.-A., Ernst, Th., Szidat, S. (2004). Levels and temporal variability of 129 I concentrations and 129 /127 I isotopic ratios in atmospheric samples from southern Spain. Nuclear Instruments and Methods in Physics Research B 223/224, 495–500. Mucciarone, D., Dunbar, H. (2003). Stable isotope record of El Niño-Southern Oscillation events from Easter Island. In: Loret, J., Tanacredi, J.T. (Eds.), Easter Island, Scientific Exploration into the World’s Environmental Problems in Microcosm. Kluwer Academic/Plenum Publishers, New York, NY, pp. 113–132. Schmidt, A., Burr, G.S., Taylor, F.W., O’Malley, J., Beck, J.W. (2004). A semiannual radiocarbon record of a modern coral from the Solomon Islands. Nuclear Instruments and Methods in Physics Research B 223/224, 420–427. Stuiver, M., Reimer, P.J., Bard, E., Beck, J.W., Burr, G.S., Hughen, K.A., Kromer, B., McCormac, G., Van Der Plicht, J., Spurk, M. (1998). Intcal 98 radiocarbon age calibration, 24,000–0 cal BP. Radiocarbon 40 (3), 1041–1083. Synal, H.A., Beer, J., Bonani, G., Suter, M., Wölfli, W. (1990). Atmospheric transport of bomb-produced 36 Cl. Nuclear Instruments and Methods in Physics Research B 52, 483–488.
598
D.L. Biddulph et al.
Wagner, M.J.M., Dittrich-Hannen, B., Synal, H.-A., Suter, M., Schotterer, U. (1996). Increase of 129 I in the environment. Nuclear Instruments and Methods in Physics Research B 113, 490–494. Yiou, F., Raisbeck, G., Imbaud, H. (2004). Extraction and AMS measurement of carrier free 129 I/127 I from seawater. Nuclear Instruments and Methods in Physics Research B 223/224, 412–415.
599
Factors influencing the determination of ultra low levels of Pu-isotopes by sector field ICP-MS Per Roos* Risoe National Laboratory, Department of Radiation research, Roskilde, Denmark Abstract Factors influencing the determination of ultra low levels of Pu-isotopes by sector field ICP-MS were investigated. Of main concern are the poor abundance sensitivity, hydride formation, polyatomic interferences and blank contribution from laboratory and instrument. Selecting instrument settings and sample introduction systems is important in minimising hydride formation and abundance sensitivity and may together with careful separation chemistry remove interfering signals from uranium and polyatomic species to allow for Pu analysis in the fg range. For best absolute detection limit, optimisation of count rate relative to the amount of analyte delivered to the plasma should be considered and not only count rate relative to sample uptake. Keywords: Mass spectrometry, ICP-MS, Pu, Abundance sensitivity, Uranium hydride
1. Introduction The interest in determining plutonium isotopes in environmental samples may be caused by dosimetric reasons in the case of accidents or releases, by interest in its biogeochemical behaviour in the environment or by tracing the source of the plutonium using its isotopic composition as a fingerprint. There are several different techniques available for the determination of trace amounts of plutonium. Of these the most widely used for environmental samples has traditionally been alpha spectrometry even though complementary techniques like thermal ionisation mass spectrometry (TIMS) also has been used for high precision isotopic analysis (Buesseler and Halversen, 1987). To a somewhat lesser extent low-energy characteristic X-ray measurements (Pu LX-rays ) and fission track analysis (FTA) (Johansson and Holm, 1996) have been used. Their limited application may be because of lack of sensitivity (Pu LX-rays measurements) as well as its inability to provide isotopic information (FTA only measures the 239 Pu isotope). The use of ICP-MS has during the last decade gained increased interest as an alternative to alpha spectrometry because of the good sensitivity, short analysis time and 239 Pu–240 Pu–242 Pu * Address: Risoe National Laboratory, Department of Radiation research, NUK-204, DK-4000, Roskilde, Denmark; phone: (+45) 4677 5340; fax: (+45) 4677 5347; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08048-4
© 2006 Elsevier Ltd. All rights reserved.
600
P. Roos
isotopic information otherwise difficult to obtain through alpha spectrometry. The technique should be seen as a complement rather than a replacement to ordinary alpha spectrometry since the latter technique still is much better suited for the analysis of the 238 Pu/239 + 240 Pu ratio, which in some cases is more informative than the 240 Pu/239 Pu ratio obtained by mass spectrometry. The 238 Pu isotope is very difficult to analyse in environmental materials by mass spectrometry due to presence of the much more abundant 238 U. Mass spectrometric techniques are also usually insufficient to accurately analyse the short-lived 241 Pu (14.4 yr) in normal environmental samples unless the 241 Pu/239 Pu ratio is large like in the Chernobyl accident, or the sample size is large. The use of the ingrowths method (241 Am) or analysis by low-level liquid scintillation counting is more sensitive with respect to 241 Pu but may necessarily not be more accurate due to problems with cross-calibration of tracers (242 Pu–243 Am), imprecise detector efficiency or unknown blank contribution and calibration problems (LSC). Analysing Pu-isotopes in the mBq to Bq range (i.e., pg to ng of 239 Pu) usually poses few problems with ICP-MS. Even with relatively poor radiochemistry and an ICP-MS not optimised for the purpose, determination of the Pu-isotopes is fairly straightforward. Due to the high sensitivity of many ICP-MS instruments it is however also possible to analyse Pu-isotopes at a far lower level. Traditionally, sample collection and preparation based on alpha spectrometry has typically meant hundreds of litre of water and several grams of sediment or soil. Due to the fact that the sensitivity is good enough for the analysis down to a few femtograms (some microBq) sample size may similarly be reduced to a few litres of seawater or milligrammes of sediment. This, theoretically, opens up completely new types of investigations of the biogeochemical behaviour of Pu, e.g. interstitial waters in soils and sediments, mineral surfaces or biologically incorporated Pu. However, very few of these kind of studies have been published. The type of studies where ICP-MS has been used deviates little from traditional studies where alpha spectrometry has been used. Furthermore, sample sizes, even though they may be reduced, still are of the same magnitude as has been used for alpha spectrometry. The major difference in the use of ICP-MS compared to alpha spectrometry for Pu-analysis has thus not been studies in a lower concentration range but rather to use the ICPMS to provide 239 Pu/240 Pu isotopic ratios. Even though the reason for this may partly be due to lack of ideas and funding, a major obstacle is the ability to reliably measure Pu-isotopes at very low and ultra low levels. This may especially be a problem at laboratories where the instrumentation is shared between many research groups and the possibility to elaborate on the equipment and associated items is very limited. There are a number of problems to address and deal with before being able to trust the weak signals provided by the instrument. This paper is not meant to cover all these problems but merely will act to summarise the experience at our laboratory. Some of the problems when measuring ultra low levels of Pu-isotopes with commercial ICP-MS instruments are the relatively low abundance sensitivity, the risk of interferences from polyatomic species, blanks, background and sensitivity. In order to minimise the problems, both instrumental settings as well as sample conditions may be optimised. Tailing from interfering peaks may be somewhat restricted by attempting to change the peak shape by hardware settings, polyatomic species may be partly removed by desolvating equipment and proper chemical separations. Sensitivity may be improved by a suitable sample inlet system. These factors were investigated with respect to analysis of ultra low levels of Pu isotopes using a sector field ICP-MS.
Determination of ultra low levels of Pu-isotopes by sector field ICP-MS
601
2. Materials and methods All measurements were performed on a Plasma Trace 2 sector field ICP-MS at Risø National laboratory. The instrument operates in a reverse Nier-Johnson geometry with adjustable mass resolution between 300 and 10,000. Two sample introduction systems were used in this study: concentric polypropylene nebuliser (CPI Instruments, The Netherlands) connected to a water cooled (5◦ C) glass cyclonic spray chamber designed for flow rates of 0.2–3 ml min−1 and an ultrasonic nebuliser without desolvator (U-5000AT+, CETAC Technologies Omaha, NE, USA) with cooling and heating temperature set at 3 and 140◦ C, respectively. RF-power was set to 1350 W and gas flows as well as instrumental optical settings and torch position were optimised with respect to sensitivity of 238 U (cps per ppb in feed solution) before each experiment. High purity nitric acid was prepared from 65% pro analysis by sub-boiling (BSB-939-IR, Berghof, Eningen, Germany) while other acids and chemicals used were of pro analysis quality. Ultrapure water (>18.2 M cm−1 ) was obtained from an Elgastat Maxima Analytica system (Elga, High Wycombe, Bucks., UK). Uranium and plutonium (242 Pu) solutions were prepared from stock solutions by dilution with 1–5% nitric acid. 2.1. Abundance sensitivity and UH+ A 5 ppb NBL 112a uranium solution and the concentric nebuliser with cyclonic spray chamber were used for the abundance sensitivity measurements while also the USN was used for the UH+ /U experiments. For the abundance tests, spectra were obtained by scanning over the mass region m/z 236.5–242.5 using both fully open slits and a partly closed source slit. For the Plasma Trace 2 instrument a partly closed source slit is necessary in order to obtain flattopped peaks. The reduction in source slit width also caused a reduction of pressure in the flight tube of about 50% relative to when slits were fully opened. The mass range was divided in 150 channels with a dwell time per channel of 50 ms, scanned 50 times and individual spectra was added to each other. The contribution of the 238 U tailing to neighbouring masses were calculated both using peak top position ratios as well as integrated peak area ratios. The same 5 ppb NBL 112a uranium solution was used for the measurements of UH+ /U. In order to reduce the 238 U tailing into the mass 239 window, resolution was increased to 1500 where no visible disturbance due to tailing could be seen. A dwell time per channel of 10 ms was used for mass 238 and 50 ms for mass 239, a total of 200 sweeps were accumulated and the UH+ /U calculated on a peak area basis. The measurements were performed using both the concentric and the ultrasonic nebuliser. 2.2. Polyatomic interferences A solution of 1 ppm lead in 5% HCl was used to qualitatively observe the potential interferences created. The test was conducted following a long rinse time using 0.5 M HCl to be sure that no uranium or plutonium left in the tubing or sample inlet system remained. The mass region 238–243 was scanned 50 times in low-resolution mode using a dwell time of 20 ms per channel. The integrated dwell time per mass unit was about the same as when analysing for ordinary low levels of Pu-isotopes.
602
P. Roos
2.3. Sample introduction A 1 ppb uranium solution was used (for the USN a 0.2 ppb solution) and the pumping rate of solution to the nebuliser was varied while measuring uptake and drain rate. Between changes in pump rate the spray chamber and tubing was completely drained. After allowing some time for conditioning the tubing (about 10 minutes) the drained solution was collected and measured as well in order to determine the amount of uranium actually transported to the plasma. In order to be able to exclude any possible changes in signal variation induced by lack of tuning, gas flows (auxiliary and nebuliser) was changed to see if this was needed during the experiment. As expected it was observed that re-tuning was not necessary during changes in flow rate to the nebuliser. 2.4. Sensitivity for Pu and U analysis Tuning of ICP-MS instruments for Pu-analysis is frequently done by using a weak uranium solution instead of Pu in order to keep blank levels down. For this reason it was of interest to compare the sensitivity of the two elements. A solution made up of 0.1 ppb 233 U and 242 Pu in 5% HNO3 was used. In order to determine if the element ratios were invariant with measurement conditions the auxiliary gas flow was changed from 1.1 to 1.8 lpm while monitoring the U and Pu intensity. Since an increase in auxiliary gas flow approximately means pushing the plasma forward (or reducing the distance between torch and cones) the change in auxiliary gas flow affects the sampling depth in the plasma. In order to assure that the observations were not time dependent due to other factors, the procedure was repeated twice with about an hour in between. 2.5. Memory effects, blank contribution and radiochemical analysis The contributions to the m/z 238–242 mass range from instrument background, the 242 Pu tracer and the laboratory blank were measured. The contribution from instrument background was measured by sweeping over the mass range during 30 minutes (200 seconds effective measurement time per peak) with nebuliser gas set to zero and sample tubing blocked. The USN was used in all experiments. Contribution from the tracer was measured for an activity amount of 0.55 mBq ml−1 (3.8 ppt) 242 Pu in 5% ultra pure HNO3 . Contribution from the acid was measured separately. Laboratory blanks were measured for 4 samples run in parallel consisting only of 242 Pu (1 mBq). The blank samples were approximately treated as ordinary samples. FeCl3 (50 mg Fe) and 100 ml DI water was added, ammonia added to precipitate the iron where after samples were dissolved in 8 M HNO3 and passed through a column (5 ml AG 1 × 4), washed with 9 M HCl and Pu eluted with 9 M HCl–0.1 M NH4 I. The final purification step was made up of a 0.5 ml AG 1 × 4 column where the samples were passed through in 2.4 M HCl–60% EtOH after Pu first have been reduced by NH4 I. After passing the column the samples were evaporated, excess iodine removed by nitric acid and the samples made ready for ICP-MS analysis by dissolving in 10 ml 5% ultra pure HNO3 . Although real samples may be treated with more comprehensive separation steps before the final clean-up step the procedure used here may be indicative of blank contributions from general radiochemical analysis in the lab used. Memory effects of U and Pu were studied
Determination of ultra low levels of Pu-isotopes by sector field ICP-MS
603
using a 242 Pu–238 U standard in 1% HNO3 . After aspirating the standard for about 20 minutes the system was washed with 1% HNO3 for about 2 minutes (typical washing time between samples in an ordinary sample run) and then a separate 1% HNO3 and a 5% HNO3 solution respectively was used to measure the amount of 242 Pu and 238 U that still could be washed out. Each of the solutions was aspirated for about one minute and measured twice with no delay in-between.
3. Results and discussion 3.1. Abundance sensitivity and UH+ The spectra obtained in this experiment is shown in Fig. 1. The significant tailing from the large 238 U peak stretches out over a large m/z range. It is especially pronounced at the high mass side, which is due to the UH+ (and to a much less degree UH+ 2 ). In Table 1 the contribution from the 238 U peak is given to mass 237, 239 and 240 calculated either on an integrated peak area basis or on peak top position basis where only the top counts has been used. Since no real peak was present at mass 239 the position of the 239 peak was determined from the mass calibration. From the results in the table it is clear that the abundance sensitivity depends both on resolution and on the way it is calculated. Normally abundance sensitivity should consider the mass 237–238 signal ratio in order to avoid the influence of UH+ on mass 239 provided the peaks can be assumed to be symmetrical. However, the ways in which values found in the literature are defined are rarely presented and it is obvious that the 239–238
Fig. 1. A 5 ppb CRM 112a solution obtained with fully open slits (rounded peak) and with source slit partly closed (flat top peak).
604
P. Roos Table 1 Contribution from 238 U to masses 237, 239 and 240 calculated on an integrated peak area basis and on peak top position basis. Measurements were performed at fully open slits and partly closed source slit, respectively 237/238
239/238
240/238
Open slits, R = 320, peak intensity = 100% Peak top Peak area
35 ppm 86 ppm
85 ppm 176 ppm
11 ppm 26 ppm
Narrow source slit, R = 440, peak intensity = 47% Peak top Peak area
19 ppm 48 ppm
70 ppm 140 ppm
3.5 ppm 14 ppm
signal ratio sometimes is used instead, producing a mixed tail + UH+ contribution. If peak top position values are used rather than integrated peak areas the results are of course generally better. The reason for the higher abundance sensitivity when using the whole peak area is due to a more significant contribution of the 238 U tail being included than when using the more distant peak position. The better abundance sensitivity when increasing the resolution is mostly due to the more restricted path ions have to travel to reach the detector but also the better vacuum may play a minor role. Correction for tailing from adjacent peaks onto a given mass is commonly done either by subtracting values interpolated from signals measured at half-mass positions from the peak to be corrected (Chen et al., 1986) or by applying a predetermined correction value. Disadvantages with the first method is that the peak is subtracted for its own tailing and that the tailing is assumed to be linear which it is not. Also with the second method care needs to be taken as how to apply the obtained value. For instance, if it is a value obtained over the integrated peak it is important to use the same mass interval during the correction as during the separate measurement of the abundance sensitivity. Thus, it is difficult to accurately perform corrections for peak tailing which is necessary at ultra low-level measurements unless uranium concentrations may be held sufficiently low. Once the abundance sensitivity has been determined it is important that the peak tail profile is fairly constant over the narrow mass range of interest (e.g. between mass 238 and 242 for instance) and that it is independent of the ion beam intensity. Deschamps et al. (2003) investigated this and found that it was reasonably true. The problem of poor abundance sensitivity and how to correct for it has been addressed in several papers (e.g. Deschamps et al., 2003; Baglan et al., 2004; Thirlwall, 2001). As seen from Table 1, one possible way to improve the abundance sensitivity in ICP-MS instruments is by increasing the resolution of the system. This however also reduces the sensitivity. Improving resolution by decreasing the source slit width however also improves the peak shape since flat-topped peaks are obtained on the Plasma Trace 2 instrument. This is an advantage in the analysis of low concentrations simply because peak hopping means less time spent on low intensity parts of the spectrum and jumping between the flat regions of the peaks is safer than a similar manoeuvre for rounded peaks (as in quadrupole ICP-MS) if drift in mass calibration occurs with time. The relatively poor abundance sensitivity is a problem of particular importance when using ICP-MS for ultra low-level measurements of any element. The origin of this problem
Determination of ultra low levels of Pu-isotopes by sector field ICP-MS
605
is rather complicated but is partly due to the rather large spread of the energy of ions produced in the plasma (10–20 eV). Other important factors contributing to the peak tails are collisions between ions and residual gas molecules in the flight tube (depends on the vacuum and the length of the flight path), scattering of slits and on other material. The energy spread produced by the plasma instability may be reduced by grounding the load coil or by inserting a metal shield (guard electrode) between plasma and load coil to reduce the capacitive coupling between plasma and load coil. The energy spread of the ions may also be reduced by using a collision cell where ions are thermalised or, more commonly, by using an electrostatic analyser (ESA) which is standard equipment in nearly all sector field ICP-MS instruments. Even so, the abundance sensitivity rarely goes below about 10 ppm in single detector SF-ICP-MS although Ketterer et al. (2004) reported a 5 ppm abundance sensitivity for a VG Axiom instrument at low (R = 410) resolution. With TIMS, this problem is less important due to the much lower energy spread by the thermal-generated ion source (around 0.2 eV). Adding energy filters (ESA, Retarding Potential Quadripole (RPQ) or Wide Aperture Retarding Potential (WARP)) (Chen et al., 1992; Cheng et al., 2000; Rubin, 2001) to TIMS instruments may result in abundance sensitivity far below 1 ppm. Such restrictions in peak tailing are necessary when performing high precision 230 Th/232 Th ratio measurements where atom ratios are around 10−5 –10−6 . Some of these filters (e.g. WARP) is designed to eliminate scattered ions, thus having lower energy, and would therefore only reduce tailing on the low mass side of the peaks. The improvement on the high mass side (e.g. for 239 Pu relative 238 U) would be limited. Even better abundance sensitivities than provided by TIMS are required when measuring isotopes like 36 Cl, 26 Al or 129 I where ratios 36 Cl/35 Cl, 26 Al/27 Al and 129 I/127 I are 10−10 or less which therefore requires accelerator mass spectrometry (AMS). This technique has also been used for the analysis of low levels of Pu isotopes (Oughton et al., 2004) with the great advantage in the ultra low abundance sensitivity which means that uranium removal need not to be performed at the same level as for ICP-MS. Apart from the peak tailing, which mainly is of instrumental origin, the UH+ interference at mass 239 also must be considered. The mass resolution needed to resolve this peak from 239 Pu is in the order of 40,000 which is not practically possible. During ultra low-level measurements it is also not desirable to increase resolution due to loss in transmission. Since the UH+ interference depends on the amount of hydrogen (water) present in the plasma the magnitude of the peak depends on several factors such as nebuliser flow rate, sample uptake rate and sample introduction methods. Among these the sample introduction method is a very important factor in affecting the hydride generation. The results of the comparison between the concentric nebuliser and the USN demonstrate this clearly (Table 2). The UH+ is significantly reduced when the water loading is lowered, as is the case with the USN. Comparing with values in Table 1 the magnitude of interference from peak tailing and UH+ are about of equal size. Previously reported values of the UH+ /U+ ratio have ranged from high 130 ppm (Zoriy et al., 2004) using DIHEN and 100 ppm (Becker et al., 2004) using a micro concentric nebuliser to 5.5 ppm (Taylor et al., 2001) using an MCN-6000. There is some tendency that UH+ /U+ ratios reported for quadrupole ICP-MS are somewhat lower than for sector field ICP-MS using similar nebulisers (e.g. Becker and Dietze, 1999; Pointurier et al., 2004). Improvement in the UH+ /U+ ratio is most of all achieved by selecting an appropriate nebuliser. Kim et al. (2000), using a Plasma Trace 2 as in this study, reported a five-fold reduction in the ratio when changing from an ordinary pneumatic nebu-
606
P. Roos Table 2 Comparing the UH+ /U ratio for two different sample introduction systems. A pure uranium solution (NBL U112a) and an instrument resolution of about 1500 were used in the study Sample introduction system
UH+ /U
Concentric nebuliser + cooled cyclonic spray chamber CETAC USN 5000+
63 ppm 23 ppm
liser to a microconcentric nebuliser with desolvation (MCN-6000). They also used the same type of USN as was included here and reported a UH+ /U+ ratio of 40 ppm. A value identical to what was reported for this instrument earlier, using the same USN as was used now (Sturup et al., 1998). There are also other ways of improving the UH+ /U+ ratio. Boulyga et al. (2002) reported an improvement by a factor of two in the UH+ /U+ ratio by optimising gas flow and RF-power. An interesting way of reducing the UH+ interference for the mass 239 is to use heavy water, D2 O. By entirely replacing the sample H2 O with D2 O Vais et al. (2004) succeeded in reducing the UH+ /U from about 13 ppm to 0.3 ppm. This may be an attractive option if only 239 Pu is of interest. The method will of course mean more problems for mass 240 (UD+ ) but this may be of less importance in many studies concerning the environmental behaviour of plutonium. 3.2. Polyatomic interferences There are a number of potential interferences apart from the UH+ that may be of importance at trace level determination of plutonium isotopes. These may originate from combinations of elements present in the plasma gas, sample and solute atoms. Of major concern are combinations of Pb, Tl, and Hg isotopes with isotopes of Cl and Ar since they all may contribute to the mass range 239–242. Even though these interferences may not be of great importance when analysing gram amounts of soil and sediments, where total Pu levels are in the pg to ng range, care must be taken either when ultra low levels are analysed and/or when high precision isotopic ratio information is of importance. Metals such as Pb, Hg and Tl are also frequently found in multielement tuning solutions used for ICP-MS calibrations. Since it is difficult to remove elements originating from the plasma gas (Ar and trace rare gases) focus is therefore mainly directed to remove constituents in the sample and solute. Considering the range of concentrations found in environmental samples and the impact of polyatomic interferences on the minor Pu-isotopes (239 Pu, 240 Pu and 241 Pu) analysed in ICP-MS, lead would be the most important element to remove. Figure 2 illustrates the background due to polyatomic species likely to be originating from Pb, Cl and Ar isotopes during a typical Pu-analysis sequence with a sample 1 ppm in lead (environmental concentrations are typically in the range of 1–200 ppm for uncontaminated soils or sediments) and 5% HCl. There are also a range of other potential interferences that may form depending on sample type and processing, e.g. rare-earth elements (Nygren et al., 2003) or complex organic matter such as urine (Ting et al., 2003) and this necessitates careful analysis of the risks given during any sample and sample processing procedure including blank samples to be analysed in an identical manner as real
Determination of ultra low levels of Pu-isotopes by sector field ICP-MS
607
Fig. 2. Mass spectrum of the range 239–243 m/z for a 1 ppm Pb + 5% HCl solution obtained in a similar manner as during a standard Pu-analysis procedure.
samples. Another alternative in identifying the influence of polyatomic interferences in a sample is to use a higher resolution or to mass calibrate very carefully and observe potential peak shifts (Wyse et al., 2001) but this is seldom possible with ultra trace Pu analysis since the peak shape does not permit accurate mass determination. Using higher resolution is usually not a viable alternative due to the loss in transmission. 3.3. Optimum use of sample Since the analysis of very low levels of plutonium often means having limited amount of sample it is important to make the best use of what is available. Routine use of ICP-MS instruments require optimisation of torch position, gas flows and some optical settings to gain maximum sensitivity with respect to counts per second per concentration unit in tuning solution, irrespectively of sample consumption rate. Less frequent does the tuning protocol address the sample introduction efficiency, which is more important when a limited amount of sample is available. The experiments conducted with the two nebulisers used shows that about 90% of the sample volumes are wasted. The concentration of uranium found in the drain solution from the cyclonic spray chamber was approximately the same (±10%) as the feed solution, thus 90% of the sample 238 U atoms was wasted. Bulk drain from the USN (drain from non-nebulised liquid, condensed solution from heater and cool steps) also contained the same concentration as the feed solution although it may be expected that the heater/cooler drains may differ in concentrations from the feed solution (being lower). The main volume of the drain is however made up by the non-nebulised liquid from the spray chamber. As judged from Fig. 3 the optimum sensitivity of the concentric nebuliser is obtained by using a pump flow resulting in an uptake rate of about 0.4–0.5 ml per minute. If also considering the fraction of sample actually reaching the torch (Fig. 4) it is clear that the higher the feed
608
P. Roos
Fig. 3. Measured 238 U count rate as a function of sample uptake rate for a concentric nebuliser and cooled cyclonic spray chamber.
Fig. 4. Fraction of sample 238 U reaching the torch as a function of sample uptake rate. The fraction reaching the plasma was calculated as net mass of solution to plasma (mass of feed solution minus mass of drain solution) divided by feed solution. Concentration of uranium was found to be about the same in feed and drain solutions.
rate of the solution the smaller fraction of the sample is actually being used. This of course is important when it comes to optimum use of a limited amount of sample and from Fig. 4 it is obvious that the lower the flow rate of sample to the nebuliser the better it is used as long as instrument background is not preventing longer counting times. The better instrument response at lower uptake rates (Fig. 5) is probably a combination of better aerosol generation (as seen from Fig. 4) and transport through the spray chamber as well as more efficient ionisation of uranium in the plasma (less energy is spent per time unit on the water entering the plasma). Another effect reducing the efficiency at higher feed rates (although of less importance) is the increased rate of oxide formation (Fig. 6), which probably
Determination of ultra low levels of Pu-isotopes by sector field ICP-MS
609
Fig. 5. Instrument response versus analyte (238 U) feed rate reaching the plasma. The analyte feed rate was calculated from feed minus drain multiplied by concentration of 238 U in feed solution.
Fig. 6. UO+ /U ratio for different delivery rates of sample liquid to plasma using a concentric nebuliser with cooled (5◦ C) cyclonic spray chamber. The delivery rate was calculated as the difference in uptake and drain rates.
is due to lowering of the plasma temperature at increased feed rates. The comparison between the concentric nebuliser and the USN (Table 3) show that the USN produced a signal roughly 15–20 times more intense than the concentric nebuliser for a similar sample feed rate. This is partly due to a better transport of sample to the plasma (50% better) but mainly due to the better use of the injected material in the plasma (about a factor of 10), which is due to less plasma energy being wasted on evaporating water when comparatively more dry aerosols are injected. If the concentric nebuliser had been used with a sample uptake rate based on its maximum sensitivity (cps ppb−1 ) of around 0.5 g min−1 , rather than based on maximum use of fraction of sample reaching the torch, the difference between the two nebulisers would have
610
P. Roos Table 3 Sample introduction characteristics for two different sample introduction systems using a 1 ppb (0.2 ppb for the USN) 238 U solution
Uptake rate (g min−1 ) Drain rate (g min−1 ) Used liquid (g min−1 ) Fraction of feed used Feed rate of U to plasma (pg s−1 ) Cps per ppb consumed of sample Cps per pg U to plasma UO+ /U
Concentric nebuliser cyclonic spray chamber
CETAC U-5000AT+
0.14 0.126 0.0133 9.5% 0.19 3.8 × 105 2.1 × 106 0.0038
0.14 0.119 0.0203 14.5% 0.07 7.5 × 106 22.2 × 106 0.00085
been even greater. The count rate per unit mass of uranium reaching the plasma at 0.5 g min−1 was 1.5 × 106 cps per pg U instead of 2.1 × 106 at an uptake rate of 0.14 g min−1 used in the comparison example. Although not investigated here, the analyte transport efficiency of the two nebulisers is also affected by parameters such as nitric acid concentrations (e.g. Stewart and Olesik, 1998) and salt load but the transport efficiency may not necessarily change in a similar manner with differences in concentration for the two nebulisers due to different physical aerosol generation processes. We may finally conclude that very little is gained in trying to reduce the UO+ formation for the concentric nebuliser since this constitute a very small fraction for both sample introduction systems used. 3.4. Sensitivity for Pu and U As expected, the relative sensitivity (cps ppb−1 ) between the two elements varied considerably with auxiliary gas flow rate. From Fig. 7 the maximum sensitivity for uranium can be found at a flow rate of about 1.4 lpm. Maximum sensitivity for Pu occurred almost at the same flow rate. The ratio Pu/U however differed about a factor of two over the range of flow rates used. At increasing flow rate the sensitivity of uranium drops faster than it does for plutonium. Although not an attractive alternative in ultra low-level analysis, the possibility of suppressing the uranium signal by optimising the Pu/U ratio exists. At the optimum Pu/U ratio the uranium sensitivity has been cut by a factor of five while the Pu sensitivity only has been reduced by about a factor of two. In absolute terms the sensitivity (cps ppb−1 ) of Pu was found to be about a factor of two higher than for uranium. Combining this with data for uranium sensitivity using the USN (Table 3) of 7.5 Mcps ppb−1 yields a Pu sensitivity of about 15 Mcps ppb−1 . Results from normal sample runs yields similar or somewhat lower sensitivities but variations are large and sensitivities may be as low as 1–2 Mcps ppb−1 in some cases. 3.5. Memory effects, blank contribution radiochemical analysis The contributions to background from instrument, nitric acid, 242 Pu tracer and blank samples are summarised in Table 4. The overall blank contribution to masses 239–241 is generally
Determination of ultra low levels of Pu-isotopes by sector field ICP-MS
611
Fig. 7. Uranium sensitivity and the Pu/U ratio for different auxiliary gas flows. Each data set is normalised to the maximum.
Table 4 Blank contribution to masses 238–242 from instrument, nitric acid, 242 Pu tracer and laboratory. Memory effect of U and Pu when using 1 and 5% washing solutions Experiment
m/z 238 (cps)
m/z 239 (cps)
m/z 240 (cps)
m/z 241 (cps)
m/z 242 (cps)
Dark current 5% HNO3 3.8 ppt 242 Pu tracer Blank Pu-1 Blank Pu-2 Blank Pu-3 Blank Pu-4 242 Pu–238 U standard 1% HNO 3 Wash-1, 1% HNO3 Wash-2, 1% HNO3 Wash-3, 5% HNO3 Wash-4, 5% HNO3
0.25 75 1547 66 48 252 87 143,000 650 312 577 18
0.30 0.15 1.30 0.9 0.6 0.7 0.5 – – – – –
0.17 0.16 0.45 0.3 0.2 0.5 0.4 – – – – –
0.15 0.11 0.91 NM NM NM NM – – – – –
0.55 0.92 8350 1678 1566 1490 1710 13,000 77 75 641 33
low in all cases although contributions from tracer or blank samples should be considered when performing Pu-analysis in the 1–10 fg range (2–20 µBq 239 Pu). The typical equivalent Pu-amounts in the blank samples are around 1–2 fg and are thus not negligible. The origin of the somewhat higher background in the blank samples is uncertain and may not necessary be Pu-related. The glassware/PFA and other laboratory equipment used in the blank runs were of ordinary type and used for various sorts of samples and activity levels. Memory effects in instrument and tubings may also be of importance. Washing of the sample inlet system may often be based on the degree to which uranium is removed. Analysing for Pu and U in some of the tests conducted above revealed that wash out times for Pu are considerably longer than for U. While little extra uranium signal appears when increasing acid strength of the washing solution from 1 to 5% HNO3 (about a factor of two) the corresponding increase for Pu
612
P. Roos
was almost a factor of ten. The absolute limit for the Pu analysis is governed by the process blank and ultimately by the instrument background. If detection limits should be pushed further down instrument efficiency needs to be at optimum, tracer amount added should be at minimum and carefully washed lab ware (all plastic) should be used.
4. Conclusions Measurement of Pu-isotopes at femtogram levels should be done with great care considering influencing factors such as abundance sensitivity, uranium hydride, polyatomic interferences and blank contributions from instrument and laboratory. Although ultimate detection limits may be set by instrument and laboratory blank levels most practical detection limits are determined by the degree of uranium removal. A combination of good radiochemistry with carefully selected instrument settings to minimise tailing and hydride formation therefore remains the primary goal of Pu analysis by ICP-MS.
Acknowledgements This work was partially funded by the NKS-B ‘Norcmass’ project. The author wishes to acknowledge the help and technical support from Lis Vinther Kristensen.
References Baglan, N., Hemet, P., Pointurier, F., Chiappini, R. (2004). Evaluation of a single collector, double focusing sector field inductively coupled plasma mass spectrometer for the determination of U and Pu concentrations and isotopic compositions at trace level. Journal of Radioanalytical and Nuclear Chemistry 261 (3), 609–617. Becker, J.S., Dietze, H.-J. (1999). Precise isotope ratio measurements of uranium, thorium and plutonium by quadrupole-based inductively coupled plasma mass spectrometry. Fresenius Journal of Analytical Chemistry 364, 482–488. Becker, J.S., Zoriy, M., Halicz, L., Teplyakov, N., Muller, C., Segal, I., Pickhardt, C., Platzner, I.T. (2004). Environmental monitoring of plutonium at ultratrace level in natural water (Sea of Galilee – Israel) by ICP-SFMS and MC-ICP-MS. Journal of Analytical Atomic Spectrometry 19, 1257–1261. Boulyga, S.F., Matusevich, J.L., Mironov, V.P., Kudrjasov, V.P., Halicz, L., Segal, I., McLean, J.A., Montaser, A., Becker, J.S. (2002). Determination of 236 U/238 U isotopic ratio in contaminated environmental samples using different ICP-MS instruments. Journal of Analytical Atomic Spectrometry 17, 958–964. Buesseler, K.O., Halversen, J. (1987). The mass spectrometric analysis of fallout 239 Pu and 240 Pu in marine samples. Journal of Environmental Radioactivity 5 (6), 425–444. Chen, J.H., Edwards, R.L., Wasserburg, G.J. (1986). 238 U, 234 U and 232 Th in seawater. Earth and Planetary Science Letters 80 (3/4), 241–251. Chen, J.H., Edwards, R.L., Wasserburg, G.J. (1992). Mass spectrometry and applications to uranium-series disequilibrium. In: Ivanovich, M., Harmon, R.S. (Eds.), Uranium-Series Disequilibrium. Applications to Earth, Marine, and Environmental Sciences. Clarendon Press, Oxford (Chapter 6). Cheng, H., Edwards, R.L., Hoff, J., Gallup, C.D., Richards, D.A., Asmerom, Y. (2000). The half-lives of Uranium234 and Thorium-230. Chemical Geology 169 (1/2), 17–33. Deschamps, P., Doucelance, R., Ghaleb, B., Michelot, J.-L. (2003). Further investigations on optimized tail correction and high-precision measurement of uranium isotopic ratios using multi-collector ICP-MS. Chemical Geology 201, 141–160.
Determination of ultra low levels of Pu-isotopes by sector field ICP-MS
613
Johansson, L., Holm, E. (1996). Determination of trace-amounts plutonium using fission track analysis. Nuclear Instruments and Methods A 376, 242–247. Ketterer, M.E., Hafer, K.M., Link, C.L., Kolwaite, D., Wilson, J., Mietelski, J.W. (2004). Resolving global versus local/regional Pu sources in the environment using sector ICP-MS. Journal of Analytical Atomic Spectrometry 19, 241–245. Kim, C.-S., Kim, C.-K., Lee, J.-I., Lee, K.-J. (2000). Rapid determination of Pu isotopes and atom ratios in small amounts of environmental samples by an on-line sample pre-treatment system and isotope dilution high resolution inductively coupled plasma mass spectrometry. Journal of Analytical Atomic Spectrometry 15, 247–255. Nygren, U., Rodushkin, I., Nilsson, C., Baxter, D. (2003). Separation of plutonium from soil and sediment prior to determination by inductively coupled plasma mass spectrometry. Journal of Analytical Atomic Spectrometry 18, 1426–1434. Oughton, D.H., Skipperud, L., Fifield, L.K., Cresswell, R.G., Salbu, B., Day, P. (2004). Accelerator mass spectrometry measurement of 240 Pu/239 Pu isotope ratios in Novaya Zemlya and Kara Sea sediments. Applied Radiation and Isotopes 61, 249–253. Pointurier, F., Baglan, N., Hemet, P. (2004). Ultra low-level measurements of actinides by sector field ICP-MS. Applied Radiation and Isotopes 60, 561–566. Rubin, K.H. (2001). Analysis of 232 Th/230 Th in volcanic rocks: A comparison of thermal ionization mass spectrometry and other methodologies. Chemical Geology 175 (3–4), 723–750. Stewart, I.I., Olesik, J.W. (1998). The effect of nitric acid concentration and nebulizer gas flow rates on aerosol properties and transport rates in inductively coupled plasma sample introduction. Journal of Analytical Atomic Spectrometry 13, 1249–1256. Sturup, S., Dahlgaard, H., Nielsen, S.C. (1998). High resolution inductively coupled plasma mass spectrometry for the trace determination of plutonium isotopes and isotope ratios in environmental samples. Journal of Analytical Atomic Spectrometry 13, 1321–1326. Taylor, R.N., Warneke, T., Milton, J.A., Croudace, I.W., Warwick, P.E., Nesbitt, R.W. (2001). Plutonium isotope ratio analysis at femtogram to nanogram levels by multicollector ICP-MS. Journal of Analytical Atomic Spectrometry 16, 279–284. Thirlwall, M.F. (2001). Inappropriate tail corrections can cause large inaccuracy in isotope ratio determination by MC-ICP-MS. Journal of Analytical Atomic Spectrometry 16, 1121–1125. Ting, B.G., Pappas, R.S., Paschal, D.C. (2003). Rapid analysis for plutonium-239 in urine by magnetic sector inductively coupled plasma-mass spectrometry using Aridus desolvation introduction system. Journal of Analytical Atomic Spectrometry 18, 795–797. Vais, V., Li, C., Cornett, J. (2004). Preventing uranium hydride formation in standard uranium samples for determination of 239 Pu by ICP-MS. Journal of Analytical Atomic Spectrometry 19, 1281–1283. Wyse, E.J., Lee, S.H., La Rosa, J., Povinec, P., de Mora, S.J. (2001). ICP-sector field mass spectrometry analysis of plutonium isotopes: Recognizing and resolving potential interferences. Journal of Analytical Atomic Spectrometry 16, 1107–1111. Zoriy, M.V., Pickhardt, C., Ostapczuk, P., Hille, R., Becker, J.S. (2004). Determination of Pu in urine at ultratrace level by sector field inductively coupled plasma mass spectrometry. International Journal of Mass Spectrometry 232, 217–224.
This page intentionally left blank
12. Management of data quality
This page intentionally left blank
617
Recent IAEA reference materials and intercomparison exercises for radionuclides in the marine environment M.K. Pham* , J. Gastaud, J. La Rosa, S.-H. Lee, I. Levy-Palomo, B. Oregioni, P.P. Povinec International Atomic Energy Agency, Marine Environment Laboratory, Monaco Abstract Results of the recent intercomparison exercises on Irish Sea sediment (IAEA-385) and a mixed fish sample from the Irish and North Seas (IAEA-414) are compared and discussed. Both materials have been designed for the determination of anthropogenic and natural radionuclides in marine samples. Several anthropogenic and natural radionuclides were analyzed by over 100 laboratories. The data were evaluated and recommended and information values for several radionuclides have been derived. Both materials will be issued as IAEA reference materials (RMs). IAEA-385 will replace IAEA-135 (Irish Sea sediment), which is out of stock. Ongoing and planned intercomparison exercises for determination of radionuclides in seawater (IAEA-418), marine biota (IAEA-437, IAEA-415, IAEA-438 and IAEA-442) and sediment (IAEA-410 and IAEA-412) are presented as well. Keywords: Marine radioactivity, Anthropogenic radionuclides, Natural radionuclides, Marine biota, Sediment, Reference Material, Intercomparison exercise, IAEA-385, IAEA-414
1. Introduction The accurate and precise determinations of radionuclide concentrations in marine samples have been important aspects in assessments of radionuclides in the marine environment and the use of radionuclides in the studies of oceanographic processes. To address the problem of data quality, and to assist IAEA’s Member States in verifying the performance of their laboratories, the Radiometrics Laboratory (RML) of the IAEA Marine Environment Laboratory (IAEA-MEL) in Monaco has conducted intercomparison exercises on radionuclides in marine samples for many years as part of its contribution to the IAEA’s programme of Analytical Quality Control Services (AQCS) (Povinec et al., 1999a). An important part of the IAEA’s AQCS programme has been the production of Reference Materials (RMs), and their provision to laboratories in required matrices and data quality. * Corresponding author. Address: IAEA-MEL, 4 Quai Antoine 1er, MC-98000, Monaco; phone: (+377) 97977272; fax: (+377) 97977273; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08049-6
© 2006 Elsevier Ltd. All rights reserved.
618
M.K. Pham et al.
RMs represent samples of well-established properties used for the assessment of analytical methods. With the aims of increasing the importance of the quality of the data, improving the accuracy and precision of marine laboratories, and providing the required traceability to the SI, the RML has also been developing Certified Reference Materials (CRMs) (Povinec, 2002). The CRMs represent materials of high quality with well-defined characteristics, which were obtained by independent methods (ISO, 2003). The RMs and CRMs should be available for different marine matrices (sediment, water, biota), with accuracy and precision required for the present state of the art of radiometrics and mass spectrometry methods. Both RMs and CRMs have a great impact on the development of the methods with a known accuracy. They represent an important benchmark in quality management of laboratories, identify weak methodologies, detecting training needs, upgrading the quality of laboratories’ performance and assessing the validity of analytical methods. The RML has produced RMs as the final products of world-wide intercomparison exercises during the last 30 years. A total of 44 intercomparison exercises were undertaken and 38 RMs and 1 CRM were produced for radionuclides in the marine environment. Table 1 lists the main characteristics of recently produced RMs for radionuclides in the marine environment. All required matrices (seawater, biota, sediment) have been covered with radionuclide concentrations ranging from typical environmental levels to elevated levels affected by discharges from nuclear reprocessing plants. The required long-term availability of RMs (over 10 years) necessitates their long-term stability, but also the collection and the preparation of large volume samples (over 100 kg), which requires the use of specific methods for their pre-treatment. The production of an RM is therefore a long process, covering the identification of needs, sample collection, pre-treatment, homogenization, bottling, distribution to laboratories, evaluation of data, preliminary report-
Table 1 RMs recently produced by the Radiometrics Laboratory of IAEA-MEL for radionuclides in the marine environment Code
Matrix, place of the origin
Radionuclide values Recommended
Information
IAEA-381
Seawater, the Irish Sea
40 K, 90 Sr, 137 Cs, 237 Np, 238 Pu, 239 Pu, 240 Pu, 239,240 Pu, 241 Am
3 H, 235 Sb, 234 U, 235 U, 236 U, 238 U, 241 Pu, 244 Cm
IAEA-384
Sediment, Fangataufa Lagoon
40 K, 60 Co, 155 Eu, 210 Pb, 230 Th, 234 U, 235 U, 238 U, 238 Pu, 239,240 Pu, 241 Am
90 Sr, 137 Cs, 214 Bi, 214 Pb, 226 Ra, 228 Ac, 232 Th, 237 Np, 239 Pu, 240 Pu, 241 Pu
IAEA-414
Fish flesh, the Irish and North Seas
40 K, 137 Cs, 226 Ra, 232 Th, 235 U, 238 U, 238 Pu, 239,240 Pu, 241 Am
90 Sr, 208 Tl, 210 Pb, 228 Ra, 230 Th, 234 Th, 234 U, 239 Pu, 240 Pu
IAEA-385∗
Sediment, the Irish Sea
40 K, 137 Cs, 226 Ra, 232 Th, 235 U, 238 U, 238 Pu, 239+240 Pu, 241 Am
90 Sr, 228 Ra, 230 Th, 234 U, 239 Pu, 240 Pu
∗ Preliminary values.
IAEA intercomparison exercises for radionuclides in the marine environment
619
ing, additional analyses in expert laboratories, certification of the material, and finally issuing the RM. We compare here the results of the recent intercomparison exercises on an Irish Sea sediment sample (IAEA-385) with a mixed fish sample from the Irish and North Seas (IAEA414). Both materials have been designed for the determination of anthropogenic and natural radionuclides in biota and sediment with elevated radionuclide levels due to the discharges from the Sellafield reprocessing plant.
2. Material description 2.1. IAEA-385 About 250 kg of sediment was collected from the Irish Sea by the Centre for Environment, Fisheries and Aquaculture Science (CEFAS), Lowestoft, UK, in 1995. Because of the large sample size the sediment pre-treatment was carried out in a commercial laboratory. The samples were packed in 100 g quantities into brown glass bottles under a nitrogen atmosphere and sterilized at 10 kGy in an irradiation facility in accordance with ISO (2003) recommendations. A particle size analysis has shown that about 34% of the sediment was less than 63 µm, and 70% less than 125 µm. The sediment is mainly composed of Si (160 mg g−1 ), Ca (55 mg g−1 ), Al (45 mg g−1 ), Fe (31 mg g−1 ) and K (18 mg g−1 ). The percentage of organic carbon in the sediment is 0.95%, the total carbon content is 2.75%. The density of the sediment is 2.56 g cm−3 . The average moisture content of the lyophilized sample after bottling, determined by drying several aliquots in an oven at 80◦ C to a constant weight (1–2 days), and stabilized for half-hour in the desiccator before weighing, was 1.14%. Since moisture content can vary with ambient humidity and temperature, it is recommended to check the water content prior to an analysis, and to report all results on a dry-weight (dw) basis. 2.2. IAEA-414 About 350 kg of mixed fish species were collected in the eastern Irish Sea by the Centre for Environment, Fisheries and Aquaculture (CEFAS, Lowestoft, UK) in 1966. The skin was removed, the fish filleted, freeze-dried and processed in the laboratory. As the quantity of material obtained was small (about 10 kg) and the measured activities of some anthropogenic radionuclides were relatively high, the sample was mixed (using the jet pulverizing technique in a nitrogen atmosphere) with 100 kg of North Sea fish powder (a fraction below 250 µm). The samples were then packed into brown glass bottles (100 g per bottle) under nitrogen gas, and sterilized at 10 kGy in an irradiation facility. The average moisture content of the lyophilized sample after bottling, determined by drying several aliquots in an oven at 80◦ C to a constant weight (1–2 days), was found to be approximately 3%. Since moisture content can vary with an ambient humidity and temperature, it is recommended to check the water content prior to an analysis, and to report all results on a dry-weight (dw) basis.
620
M.K. Pham et al.
2.3. Homogeneity tests The homogeneity of samples was checked by measuring the activities of 40 K, 137 Cs, 210 Po, 235 U, 238 U, 238 Pu, 239+240 Pu and 241 Am on 10–44 bottles taken at random. Non-destructive gamma-ray spectrometric (GRS) measurements were performed on 10–100 g samples using HPGe-detectors. 210 Po, 235 U, 238 U, 238 Pu, 239+240 Pu and 241 Am were determined by semiconductor alpha-ray spectrometry (SAS) of 0.1–10 g samples. The “between samples” variances showed no significant differences from the “within sample” variances for all radionuclides tested. Both materials were thus considered sufficiently homogeneous for the tested radionuclides at the weights used.
3. Results and discussion The samples were distributed to 110 laboratories, which were requested to determine as many radionuclides as possible by radiometrics as well as by mass spectrometry methods, e.g. ICPMS (Inductively Coupled Plasma Mass Spectrometry) and AMS (Accelerator Mass Spectrometry). The reference date for data reporting was set for 1 January 1996 for IAEA-385, and 1 January 1997 for IAEA-414. Over 100 sets of results were received from participants for each sample which were included in the data evaluation (Pham et al., 2004, 2005). 3.1. Certification criteria and data evaluation For data sets comprising 5 or more accepted laboratory means, median values and confidence intervals were calculated following previously developed procedures (Povinec and Pham, 2000). The median values were considered as the recommended values when (i) at least 5 laboratory means were available, calculated from at least 3 different laboratories; (ii) the relative uncertainty of the median did not exceed ±5% for activities higher than 100 Bq kg−1 , ±10% for activities from 1 to 100 Bq kg−1 and ±20% for activities lower than 1 Bq kg−1 . An activity value was classified as an information value when at least 5 laboratory means (within the same order of magnitude) calculated from the results of at least 2 different laboratories were available. Laboratory means were calculated from individual results either as arithmetic means with corresponding uncertainties when more than two partial results were reported, or as weighted means with weighted uncertainties in the case of only two results being reported. All the values have been rounded off to the most significant figure. The data treatment consisted of identifying and eliminating the outlying values, calculating the median and the confidence intervals. The calculations were based on the assumption of a non-parametric distribution of the data to which a distribution-free statistics is applicable. The “less than” values were segregated from the results and the remaining values were checked for the presence of outliers using a box and whisker plot test. Confidence intervals were determined from a non-parametric sample population and expressed as a two-sided interval, representing 95% confidence limits.
IAEA intercomparison exercises for radionuclides in the marine environment
621
The performance of a laboratory is considered to be acceptable if the difference between the robust mean of the laboratory and the assigned value (defined as Z-score) is less than or equal to 2 (Povinec et al., 1999b). The analysis is regarded as being out of control when |Z| > 3. The Z-score evaluation represents a simple method, which provides the participating laboratories with a normalized performance score for bias. Results of analysis of anthropogenic 90 Sr, 137 Cs, 238 Pu, 239+240 Pu, 239 Pu, 240 Pu, 241 Am and natural 40 K, 210 Pb, 226 Ra, 228 Ra, 230 Th, 232 Th, 234 U, 235 U, 238 U in IAEA-385 and IAEA-414, as supplied by the laboratories, have been treated separately. Due to the space limitations, we can present here only a few typical results. 3.2. Cesium-137 IAEA-385: 85 laboratories reported 137 Cs results (Fig. 1(a)), obtained mainly by a nondestructive GRS. Five results were identified as outliers. The rest of the data was homogeneous within two standard deviations of the distribution mean. Z-score values were below 2.6, thus showing good performances of the laboratories (Fig. 2(a)). The median, given as the recommended value, is 33.7 Bq kg−1 dw (the 95% confidence interval is 32.4–34.7 Bq kg−1 dw). IAEA-414: 81 laboratories reported 137 Cs results (Fig. 1(b)), 19 results were identified as outliers. The rest of the data was homogeneous within two standard deviations of the distribution mean, Z-score values were below 2.2 (Fig. 2(b)). The median, given as the recommended value, is 5.14 Bq kg−1 dw (the 95% confidence interval is 5.00–5.27 Bq kg−1 dw). The results confirmed that laboratories can analyze 137 Cs with good precision, although some laboratories are still having problems with efficiency calibration (for different sample matrices) and background corrections (for lower activity samples). 3.3. Plutonium-239+240 IAEA-385: Chemical treatments followed by SAS (239+240 Pu), or ICPMS and AMS were used for analysis of 239 Pu and 240 Pu from which 239+240 Pu activity was calculated. Forty-three data sets (combined SAS, ICPMS and AMS results) were reported (Fig. 3(a)). Seven results were identified as outliers. The rest of the data was homogeneous, within two standard deviations of the distribution mean. Z-score values were below 2.1, thus showing good performances by the laboratories. The median, given as the recommended value is 2.98 Bq kg−1 dw (the 95% confidence interval is 2.81–3.13 Bq kg−1 dw). 239 Pu and 240 Pu results were also treated separately, using data provided by ICPMS and AMS laboratories. However, as only a few results were available, only information values can be given at present. The obtained medians are 1.92 and 1.18 Bq kg−1 dw, with corresponding 95% confidence intervals of 1.30–2.07 Bq kg−1 dw and 0.97–1.32 Bq kg−1 dw for 239 Pu and 240 Pu, respectively. IAEA-414: 42 combined data sets (SAS, ICPMS and AMS) were reported (Fig. 3(b)), 9 results were identified as outliers. The rest of the data was homogeneous, within two standard deviations of the distribution mean. Z-score values were below 2.0. The median, given as the recommended value is 0.12 Bq kg−1 dw (the 95% confidence interval is 0.113–0.125 Bq kg−1 dw).
622
M.K. Pham et al.
(a)
(b)
Fig. 1. Data evaluation of 137 Cs in (a) IAEA-385 and (b) IAEA-414. Vertical lines separate outliers.
IAEA intercomparison exercises for radionuclides in the marine environment
623
(a)
(b)
Fig. 2. Z-score values for 137 Cs in (a) IAEA-385 and (b) IAEA-414.
Only four laboratories reported 239 Pu and 240 Pu results obtained by ICPMS. The data were homogeneous, from 0.064 to 0.082 Bq kg−1 dw for 239 Pu, and from 0.047 to 0.065 Bq kg−1 dw for 240 Pu. Generally there was a reasonable agreement between results reported by SAS, ICPMS and AMS, however, there are still only a few laboratories using mass spectrometry techniques for Pu analysis. There has been clear progress in radiochemical separation techniques, mainly thanks to new gas chromatography resins (UTEVA, TRU). 3.4. Americium-241 IAEA-385: Both a non-destructive GRS and destructive/chemical separations followed by SAS were used for analysis. Sixty results (41 obtained by GRS and 19 by SAS) were re-
624
M.K. Pham et al.
(a)
(b)
Fig. 3. Data evaluation of 239+240 Pu in (a) IAEA-385 and (b) IAEA-414. Vertical lines separate outliers.
IAEA intercomparison exercises for radionuclides in the marine environment
625
ported, from which 56 (40 GRS and 16 SAS results) were available for statistical treatment. Three SAS results and 1 GSR result did not pass the outlier test. The treated results were less than two standard deviations from the distribution mean. The data evaluation was done separately for the GSR and the SAS results to check the performance of each method. The wider spread of data and higher uncertainties obtained by GRS than those obtained by SAS document that the laboratories still have difficulties in analyzing low-energy gamma-ray emitters. The obtained results were, however, homogeneous, within two standard deviations of the distribution mean. Therefore the results obtained by both the GRS and SAS were used in the data evaluation process. The Z-score values for the combined data set are below 2.0, thus showing good performance of the laboratories. The median, given as the recommended value, is 3.90 Bq kg−1 dw (the 95% confidence interval is 3.60–4.10 Bq kg−1 dw). IAEA-414: 34 combined results (13 obtained by GRS and 21 by SAS) were reported, from which 22 (8 GRS and 14 SAS results) were available for statistical treatment. Five SAS results and 1 GSR result did not pass the outlier test. The rest fall less than 2.2. standard deviations from the distribution mean. Z-score values were below 2.2. The median, given as the recommended value is 0.196 Bq kg−1 dw (the 95% confidence interval is 0.189–0.199 Bq kg−1 dw). 241 Am is a good test example for the analysis of low-energy gamma-ray emitters with GRS. Higher activity present in IAEA-385 allowed several laboratories to take part in this exercise, however, high precision data were only produced by SAS. 3.5. Uranium-238 IAEA-385: Twenty-five results were reported (Fig. 4(a)) using GRS (6 results failed the outlier test), and 23 results were obtained by SAS and ICPMS (17 were accepted for data evaluation). The results obtained by GRS were more widespread than the data obtained by SAS/ICPMS, demonstrating again that many laboratories had difficulties with the GRS measurements. Results obtained by ICPMS were in the 95% confidence interval range. The Z-score values (Fig. 4(a)) were below 2.2. The median given as the recommended value is 29.4 Bq kg−1 dw (the 95% confidence interval is 28.0–30.5 Bq kg−1 dw). IAEA-414: 27 results obtained by GRS, SAS and ICPMS were reported (Fig. 4(b)). Most results obtained by GRS were reported as detection limits only or as outlying values. 20 laboratories reported results obtained by SAS (14 were accepted). Two of 3 results obtained by ICPMS were in the 95% confidence interval. Z-score values for accepted data were below 1.8. The median given as the information value is 1.112 Bq kg−1 dw (the 95% confidence interval is 1.065–1.170 Bq kg−1 dw). 3.6. Radium-226 IAEA-385: Twenty-three laboratory means (13 accepted, i.e. almost half were outliers), obtained mainly by GRS, were reported. Most outlier results were overestimating the median value. The rest of the data was homogeneous between two standard deviations. The Z-score values of the accepted data were below 1.5. The median given as the information value is 32.9 Bq kg−1 dw (the 95% confidence interval is 30.7–35.0 Bq kg−1 dw). IAEA-414: Eighteen laboratory means mostly obtained by GRS were reported (13 accepted). Z-score values of accepted data were below 1.7. The median given as the information value is 1.40 Bq kg−1 dw (the 95% confidence interval is 0.59–1.76 Bq kg−1 dw).
626
M.K. Pham et al.
Fig. 4. Data evaluation of 238 U in (a) IAEA-385 (outliers are grouped together) and (b) IAEA-414.
IAEA intercomparison exercises for radionuclides in the marine environment
627
Although GRS is the most widely used radiometrics technique for radionuclide analysis, the results presented both for 238 U and 226 Ra document that laboratories need to improve their calibration procedures and spectra evaluation. The full data sets for all evaluated anthropogenic and natural radionuclides (see also Table 1) are available in the intercomparison reports (Pham et al., 2004, 2005).
4. Future intercomparison exercises Decreasing concentrations of anthropogenic radionuclides in the marine and terrestrial environments and the development of new improved technologies for radionuclide determinations present new challenges for the AQCS programme. Therefore, new, more precisely characterized RMs with lower radionuclide concentrations will be necessary to meet these new requirements. The on-going and planned intercomparison exercises for different matrices (seawater, marine biota, sediments) and origin (the Mediterranean Sea, the Northeast Atlantic Ocean, the Baltic Sea, the North Sea and the Pacific Ocean) of marine samples are listed in Table 2. Samples with typical environmental radionuclide levels will be analyzed, e.g. in the on-going exercise IAEA-437 (Mediterranean mussel), and in IAEA-412 (Pacific sediment, in preparation), as well as samples having very low radionuclide concentrations, specifically designed for mass spectrometry techniques, e.g. in the on-going IAEA-418 (Mediterranean seawater), and in IAEA-415 (Northeast Atlantic fish, in preparation).
5. Conclusions Many laboratories have benefited from regular participation in world-wide and regional intercomparison exercises, from supply of RMs developed to cover their needs, long and short-term training programmes and from assistance in developing quality assurance programmes. The AQCS programme has contributed to the implementation and/or improvements of nuclear analytical methods, and RMs have been recognized as an important component of quality Table 2 On-going and planned intercomparison exercises for radionuclides in the marine environment Code
Matrix
Place of origin
IAEA-437 IAEA-418 IAEA-410 IAEA-412 IAEA-415 IAEA-438 IAEA-442
Mussel Seawater Sediment Sediment Fish flesh Fish flesh Fish flesh
Mediterranean Sea Mediterranean Sea Bikini Atoll Pacific Ocean North Atlantic Sea Baltic Sea North Sea
628
M.K. Pham et al.
management in analytical laboratories. Available RMs are listed in the IAEA’s AQCS biennial catalogue, as well as on the IAEA website http://www.iaea.org/programmes/naal/, and can be purchased at a reasonable price. As an example of intercomparison exercises carried out in the framework of the AQCS programme, we presented and compared here results obtained for IAEA-414 (mixed fish flesh from the Irish and North Seas) and IAEA-385 (Irish Sea sediment). The materials have been characterized for several natural and anthropogenic radionuclides, and recommended and information values for several radionuclides have been derived. The IAEA-414 material has already been certified as an RM and it is available to laboratories. It is expected that IAEA-385 after successful certification will be issued as the IAEA reference material as well and it will replace IAEA-135 (Irish Sea sediment).
Acknowledgements The laboratories participating in the intercomparison exercises organized by IAEA-MEL are highly acknowledged. Special acknowledgement is given to the Centre for Environment, Fisheries and Aquaculture Science (CEFAS), Lowestoft, UK, for providing Irish Sea fish and sediment samples. The Agency is grateful for the support provided to its Marine Environment Laboratory by the Government of the Principality of Monaco.
References ISO (2003). Certification of reference materials – General and statistical principles. ISO Guide 35, 61 pp. Pham, M.K., La Rosa, J., Lee, S.-H., Povinec, P.P. (2004). Radionuclides in mixed fish from the Irish Sea and the North Sea. Report on the worldwide intercomparison IAEA-414. IAEA/MEL/73. IAEA, Vienna, 80 pp. Pham, M.K., Sanchez-Cabeza, J.-A., Povinec, P.P. (2005). Radionuclides in Irish Sea sediment. Report on the worldwide intercomparison exercise IAEA-385. IAEA, Vienna. Povinec, P.P., Gastaud, J., Pham, M.K. (1999a). IAEA-MEL’s AQCS programme for marine radioactivity measurements. IAEA-TECDOC-1094. IAEA, Vienna, pp. 87–93. Povinec, P.P., Pham, M.K., Ballestra, S. (1999b). Radionuclides in Irish Sea water. Report on the intercomparison run and certified reference material IAEA-381. IAEA/MEL/66. IAEA, Vienna, 43 pp. Povinec, P.P., Pham, M.K. (2000). Radionuclides in Fangataufa Lagoon sediment. Report on the intercomparison run IAEA-384. IAEA/MEL/68. IAEA, Vienna, 44 pp. Povinec, P.P., et al. (2002). Certified reference material for radionuclides in seawater IAEA-381 (Irish Sea water). Journal of Radioanalytical and Nuclear Chemistry 251, 369–374.
Further reading Lee, S.-H., Povinec, P.P., Wyse, E., Pham, M.K., Hong, G.-H., Chung, C.-S., Kim, S.-H., Lee, H.-J. (2005). Distribution and inventories of 90 Sr, 137 Cs, 241 Am and Pu isotopes in sediments of the Northwest Pacific Ocean. Marine Geology 216, 249–263. Pham, M.K., Povinec, P.P. (2003). Matériaux de référence de l’AIEA pour les radionucléides dans le milieu marin. Bulletin de BNM 127, 210–214. Povinec, P.P., Pham, M.K. (2001). IAEA reference materials for quality assurance of marine radioactivity measurements. Journal of Radioanalytical and Nuclear Chemistry 248, 211–216.
629
Towards quality excellence in radioanalytical laboratories at STUK, Finland Tarja K. Ikäheimonen* , Seppo Klemola, Pia Vesterbacka STUK – Radiation and Nuclear Safety Authority, Helsinki, Finland Abstract The development of a uniform and modern quality system for the whole of the Finnish Radiation and Nuclear Safety Authority (STUK) began in 1997. The principle of continuous improvement is widely applied by performing annual self-assessments as well as internal and external audits. One of the most important steps in a quality assurance programme is the proficiency of the personnel. Good initiation, non-stop training during work and knowledge of the physical and chemical properties that affect analytical procedures are the requirements for first-rate and motivated working. Working procedures have been planned and followed through in order to prevent any contamination in laboratories. Considerable emphasis has also been placed on the standardisation of uncertainty estimations. This demands good knowledge of the sources of errors in sample preparation, analysis and measurement. Participating in intercomparison exercises and proficiency tests is one way to demonstrate the validity and quality of the methods used. However, a lack of intercomparison exercises and reference material on certain natural radionuclides occurring in waters and certain artificial radionuclides has complicated the validation and proficiency testing of these measurements. Moreover, the arrangement of intercomparisons has not always been optimal. Keywords: Quality assurance, Radiochemical analyses, Radiometrics, Gamma-spectrometry
1. Introduction STUK is, besides a regulatory authority and research centre, also an expert organisation providing contracted services to a wide variety of customers including Finnish citizens and companies and also international organisations. All the activities of STUK are managed by using a quality system based on the Total Quality Management (TQM) and ISO (International Organization for Standardization) standard 9004 (ISO, 2000). The development of a uniform and modern system started in 1997 and regulation was completed in 1999. At the top of our quality hierarchy is the quality policy of overall objectives, while the second level is the quality manual of the whole of STUK, following the manuals of each department. All these documents are available to the staff via the STUK intranet. The base level in the hierarchy comprises the laboratory manuals and guides (Fig. 1). The guidelines for all work in the laboratory are standards mentioned above. * Corresponding author. Address: STUK, P.O. Box 14, FIN-00881 Helsinki, Finland; phone: (+358) 9 7598 8596; fax: (+358) 9 7598 8589; e-mail:
[email protected]
RADIOACTIVITY IN THE ENVIRONMENT VOLUME 8 ISSN 1569-4860/DOI 10.1016/S1569-4860(05)08050-2
© 2006 Elsevier Ltd. All rights reserved.
630
T.K. Ikäheimonen et al.
Fig. 1. The structure of the quality system at STUK.
Accreditation was awarded by FINAS (the Finnish Accreditation Service) in 1999 and was renewed in 2003 according the European Standard EN ISO/IEC 17025 (2000). Accredited fields comprise tests on radiation safety and related environmental sampling, namely gammaspectrometric analyses, radiochemical analyses of tritium, strontium-90 and transuranic elements in environmental, foodstuff and biological samples, as well as natural radionuclides in water, indoor radon measurements, direct measurement of contamination people, and chromosome analysis. In this paper, quality assurance in radiochemical and gamma-spectrometric laboratories is examined.
2. Proficiency of personnel One part of the Quality Assurance (QA) programme concerns the resources used for analysis. Here, the most important are the qualified personnel. Skilled and motivated personnel can carry out high quality work with rather poor instrumentation and equipment, but the most modern and costly tools do not guarantee reliable results if the personnel are not proficient and do not have motivation or good work ethics. Laboratory personnel must be competent to perform technical procedures. This is assured not only with the right level of education but also with a careful and documented initiation programme. The training programme consists of thorough familiarisation with the instruments and procedures. Technical personnel must also understand all the chemical and physical phenomena that affect the analysis and measurements. This is the only way they can recognise and react to any deviations from the expected behaviour of the determined radionuclide. After initiation training, personnel must demonstrate that they are capable of performing the procedure with an acceptable level of competence by, for instance, analysing a reference material. Access to the different areas of STUK is automatically controlled and passes are given according to the work assignment of the employee. In addition, in the laboratory room, part of the cabinet and fume cupboards is equipped with a lock in order to ensure the reliable handling of samples. This learning is a continuous process and must support the overall goals of STUK. The circulation of personnel between different kinds of laboratory work increases their versatility and can also improve motivation. At STUK we have pointed out that, in addition to the proficiency of personnel, there is a need to preserve and increase motivation and the work ethics. Every person has a clear job
Towards quality excellence in radioanalytical laboratories at STUK, Finland
631
description in which authorities and responsibilities are defined and documented. Empowerment gives an opportunity to achieve the demanded results and to have an effect on making them. The opinions of the personnel are considered when developing working procedures and methods. This advances creativity, which is a source of motivation and also one characteristic of a dynamic organisation. A good flow of information is an important part of a working culture. Nobody must repeat the same mistakes and everybody has to know good practices. Furthermore, everybody has a right to know how well the laboratory has succeeded in intercomparisons and other exercises. Finally supervisors, who are chemists or physicists, have overall responsibility for the laboratory procedures and the reliability of final results. This requires good experience not only of chemistry, physics and measurement techniques, but also of all factors having an influence on the quality of the work.
3. Working procedures Carrying out analytical work for environmental analyses implies that the analyst must be aware of good laboratory practices. These include safety at work, cleanliness, precautions for handling chemicals and equipment, and work organisation. All instructions in the laboratory are given in a written form and they are stored in the laboratory manuals. During the analysis all details and observations concerning the samples, analysis and analyst need to be noted (EN ISO/IEC 17025, 2000). These details are recorded in the follow-up forms, which are filled in during the analysis. Systematic cleaning is arranged for laboratory and measuring rooms to ensure cleanliness, which checked by surface contamination measurements and analysing swipe samples. Laboratories, various glassware and equipment are classified for different purposes according to the activity concentration and dustiness or chemical toxicity of the sample. Separate rooms are also organised for pre-treatment and measuring equipment. Measuring equipment and detectors are separated according to the measured radionuclide and its activity in the sample. This guarantees that cross-contamination of samples in the laboratory is avoided. The corresponding classification is also followed in washing. Laboratory containers are first kept in an acid or detergent solution in order to prevent the adsorption of radionuclides before they are washed in the dishwasher. In gamma-spectrometric measurements, mainly disposable measuring beakers are used in order to avoid cross-contamination. Only Marinelli beakers are washed and reused. In waste management, STUK is following its own instructions that take into account national waste management regulations. The wastes that cannot be discarded into the environment, are collected. Radioactive waste, chemical waste and solid waste are separately collected and treated according to instructions. Additionally, a separate sewer network is installed in the laboratory where the wastewater is neutralised before it is discarded into the common sewerage system. The environmental conditions are carefully planned so that they do not invalidate the results or adversely affect the measurements. The measurement rooms are constructed from special concrete in which the concentration of natural radionuclides is low. This is especially important for gamma-spectrometry. In addition, the rooms are equipped with constant air
632
T.K. Ikäheimonen et al.
conditioning in order to decrease background radiation due to indoor radon. Continuous monitoring of temperature and humidity in the measurement rooms is also arranged in order to evaluate changes in background counts and operation of the measurement equipment. The equipment in measurement rooms is equipped with uninterruptible power supplies (UPS) in order to guarantee a continuous supply of power. Laboratory Information Management System (LIMS) software has been taken implemented at STUK. All analytical data, including basic information on sampling and samples, raw measuring data and results, are stored in the database. LIMS is a password-protected system and access is limited to authorised persons. With the help of LIMS, various input and calculation errors are prevented. The system also supports improved traceability of the results. However, it is necessary to ensure that basic sample data entered into the database are correct. LIMS offers the means to arrange analysis flow control and to optimise sample throughput. Contracted results are reported to customers in a test report or a calibration certificate. These are accurately, clearly, unambiguously and objectively prepared. This means that a long list of factors must be included in the test report, a detailed list of which can be found from the ISO 17025 standard.
4. Control of quality and stability of equipment Confidence in the performance of equipment is maintained by regular calibration and servicing, routine checks and control charts (EN ISO/IEC 17025, 2000; Adams et al., 1998). The most important control charts, namely background, energy resolution and detection efficiency, are based on regular Quality Control (QC) measurements. In some analytical methods there are inherent, method-specific performance meters that can be used to monitor the status of the measuring system. A summary of the various quality control measurements for different equipment and standard solutions is presented in Table 1. In alpha- and gamma-spectrometry, computerised techniques have been developed to record the performance characteristics of spectrometers and software. These include peak widths and positions. The peak width or energy resolution is the key parameter of spectrometer health. In gamma-spectrometry, these are evaluated during the spectrum analysis of every sample measurement, thus decreasing the frequency of regular QC measurements. In alpha-spectrometry the peak width is evaluated less frequently, such as once per month. Environmental samples always contain some natural radionuclides. The position of background peaks in the gammaspectrum can be used as a gain stability gauge of a spectrometer. All the obtained values are recorded in a log-file. The confidence of liquid scintillation counters is manually controlled by the measurement of samples together with background and efficiency measurements using background, 3 H and 14 C standards. The control charts are plotted using the data attained from the measurements. Corresponding warning limits and action limits are set, taking into account the different characteristics of spectrometers. The acceptance criteria in the charts are normally set as warning limits of 2σ and an action limit of 3σ above and below the long-term average data. Figures 2 and 3 give examples of these control limits set for the background count-rate and detection efficiency of a gamma-spectrometer.
Towards quality excellence in radioanalytical laboratories at STUK, Finland
633
Table 1 Quality control measurements and checking at STUK Equipment for radioactivity measurements
Controlled parameter
Frequency
Gamma-spectrometer
Background Energy calibration and resolution Detection efficiency Background FWHM, area and energy calibration Leak current Total calibration Background Efficiency by 3 H and 14 C Control samples Calibration for different nuclides
At least 3–4 times per year In every measurement Once a month 3–4 times per year Once a month Every week 2–3 times in two years In every measurement batch Once a month In every measurement batch Depends on measured nuclides, 2–3 times in two years In every measurement batch In every measurement batch 4 times per year 2–3 times in two years
Alpha-spectrometer
Liquid scintillation spectrometer
Proportional counter
Background Stability Control samples Calibration for 89,90 Sr
Examples of other quality controls Standard solutions Control of activity concentration Visual checking Pipets Checking Calibration Scales Internal checking of the scale Checking with calibration weight Total calibration Measurement of known standard Atomic absorption spectrometer (AAS, Calibration with known standards used in yield det.)
Once per year during the period of validity When used Once a month Once per year In every weighing batch Once in two months Once per year Before measurement Before measurement
Fig. 2. Example of warning and action limits set for the background count-rate of 40 K of a gamma-spectrometer. Uncertainties are given at the 68% confidence level. (UAL – Upper Action Limit, UWL – Upper Warning Limit, LWL – Lower Warning Limit, LAL – Lower Action Limit.)
634
T.K. Ikäheimonen et al.
Fig. 3. Example of control limits set for detection efficiency measurements of a gamma-spectrometer. Uncertainties are given at the 68% confidence level.
Fig. 4. Examples of background components (210 Pb and 214 Bi) differing in temporal behaviour. Uncertainties are given at the 68% confidence level.
However, there are a number of exceptions to these defaults. For example, background components might have differing behaviour and deserve special consideration. While 40 K is considered as stable (Fig. 2), the count rates of the peaks due to the ambient radon daughters might vary either periodically or irregularly (Fig. 4). 210 Pb in detector material and shielding is a unique source of background, since it decays at a constant rate, and can even be calculated for the future. Each of these components might require separate update frequency, also taking into account the contribution of the background to the total uncertainty of a particular result. 5. Uncertainty estimations Much attention has been given to the best way of performing uncertainty estimations. The uncertainty of a result is determined not only by considering single measurements and their standard deviations, but also by examining the overall results of a test (Ellison et al., 2000; IAEA, 2004). Uncertainty in analytical measurement incorporates all components of a test. When no standard method exists, as usually in analysis of many radionuclides, uncertainty estimation demands a thorough understanding of sources of uncertainties in sample preparation, analysis and measurement. When using a standard method, estimation of uncertainties is
Towards quality excellence in radioanalytical laboratories at STUK, Finland
635
easy to perform: estimation guides are usually given in a method, and little more is expected than following the evaluation procedure as given in a standard. At STUK we have used stepwise implementation of the uncertainty concept (Ellison et al., 2000). This means that factors contributing to the overall uncertainty must be identified and the size of each uncertainty component estimated. After rejecting negligible components and cancelling duplicates, combined uncertainty is finally calculated. To quantify the size of a single uncertainty component, statistical analysis of a series of observations can be used. If this is not possible, evaluations by means of other than statistical analysis are used. This requires good experience and state of the art understanding. There are no rules for the acceptable level of uncertainty, but it has to be decided separately. Sometimes a large uncertainty is acceptable, while in other cases a small one is required, depending on analysis, the client’s wishes, research, and so on. However, the uncertainty of the result always needs to be defined and expressed (ISO, 1995). In radiochemical analysis, including alpha or beta-spectrometric measurements, typical sources of uncertainty are sample weighing, yield determinations (e.g. activity of tracers, AAS measurements, pipetting), count rates and uncertainties arising from intra- and interlaboratory analyses. Figure 5 illustrates the sources and size of different uncertainties in 210 Pb analysis determined via 210 Po deposition and alpha-spectrometry. The dominant uncertainty contributions in gamma-spectrometry of environmental samples are usually those of count rate, full-energy peak efficiency and source–detector geometry. In special circumstances other sources of uncertainty will have a major contribution. Examples of these are the same or higher background count rate of a nuclide compared to sample activity, low energy gamma-rays from high density samples, and a cascade-summing correction of the poorly-defined decay scheme of a very rarely analysed nuclide. It is not always appreciated that gamma-emission probabilities of some nuclides have a relatively high uncertainty, e.g. for 234 Th it varies from 5% to 10%. Uncertainty contributions in gamma-spectrometry are listed in Table 2 (International Standard IEC 1452, 1995).
Fig. 5. Source, size and variations in uncertainties in 210 Pb analysis determined via 210 Po deposition and alpha-spectrometry. Uncertainties are given at the 95% confidence level.
636
T.K. Ikäheimonen et al. Table 2 Uncertainty contributions in gamma-spectrometry (International Standard IEC 1452, 1995)
Peak-area evaluation Background peak contribution Source–detector geometry Full-energy peak efficiency Standard source activity Source aliquots Sample attenuation variation Cascade summing Decay of nuclide Gamma-ray emission probability ADC live timer Pulse pile-up (at high count rates)
Typical magnitude (% 1σ )
Depends on
0.1–30 0–100 0.1–10 0.5–15 0.1–2 0.1–3 0–50 0–100 0–5 0.1–20 0.1–0.5 0–30
count rate → activity nuclide sample size energy
energy nuclide, efficiency nuclide nuclide activity
A large, ongoing process is the estimation of uncertainty related to sampling. Especially when concerning the aquatic environment, where sampling is not an easy process (e.g. sampling of sediment, bottom-water or aquatic plants), the estimation of uncertainties is complicated. Performing replicates or intercomparisons on sampling is difficult and expensive. Some intercomparisons have been arranged by the HELCOM/MORS group in the Baltic Sea (Ilus et al., 2000). In an estimation process, long-term experience and a good understanding of aquatic processes is very important. At present, the uncertainties of sampling are not yet included in the overall uncertainty at STUK, which is mentioned in test reports.
6. Validation of methods Developed or adopted methods are used after validation (EN ISO/IEC 17025, 2000; Ellison et al., 2003; Quevauviller, 2002; Adams et al., 1998). The values obtainable from a validated method should be relevant to the requirements for accuracy and precision, sensitivity and detection limits, selectivity and specificity, range and linearity, repeatability and reproducibility, and stability and robustness against external influences. Which of these parameters should be considered and determined depends on the use of a particular method. The validation procedure includes specification of the requirements, determination of the relevant performance characteristics and comparison of the requirements with the obtained values. The different steps of validation of method at STUK are summarised in Fig. 6. After confirmation of the method characteristics, the statistical control of the analysis is assured and the precision evaluated. Evaluation of the accuracy is performed using certified reference materials or spiked samples, by comparison against results from other methods or via inter-laboratory comparison. The validation procedure is recorded as a whole, including the original data and a statement of the validity. As noted in standard ISO 17025, validation is always a balance between costs, risks and technical possibilities.
Towards quality excellence in radioanalytical laboratories at STUK, Finland
637
Fig. 6. Steps of method validation at STUK.
7. Intercomparisons Successful participation in intercomparisons has been highlighted as the best way of proving good quality. It is also one of the ways to validate a new method. An example of the intercomparison results of STUK for 137 Cs together with uranium and transuranic elements during about 10 years is presented in Fig. 7. Since it is not possible to have intercomparison exercises for all the gamma-emitting nuclides, it is important to show quality of the analyses of a few nuclides representing a variety of situations, especially including the range of gamma-ray energies. Typical nuclides representing the low and high energies are 241 Am and 60 Co, correspondingly. Figure 8 presents an intercomparison history for these nuclides. Low 241 Am results indicate insufficient sample attenuation correction in some sample materials. However, intercomparison exercises have some problems. Normally, the time and replicates used for analysis in intercomparisons are larger than can be used for routine analysis. The results are thus not comparable with routine work. The results of intercomparisons are often only available after a long time, usually years. The quality situation in a laboratory may then have changed. The time period within which analysis should be performed could be shortened and the reporting of results of intercomparison must be no longer than a maximum of one year. Sometimes, the matrix or the level of activity of intercomparison samples is not suitable for using routine analytical procedures, and then again the results obtained with modified procedures do not represent the quality of routine work. There has been a lack of intercomparison exercises and reference materials for certain natural radionuclides such as 222 Rn, 210 Po and 210 Pb in water or certain artificial nuclides, such as 99 Tc and 241 Pu in environmental samples. The results of comparison are always dependent on the quality of the participating organisations.
638
T.K. Ikäheimonen et al.
Fig. 7. The results of intercomparisons for 137 Cs, 238 U, 239,240 Pu and 241 Am at STUK, as the ratio of STUK value to the target (reference) value. Uncertainties are given at the 95% confidence level.
Fig. 8. The results of intercomparisons for 60 Co and 241 Am at STUK, as the ratio of STUK value to the target (reference) value. Uncertainties are given at the 68% confidence level.
8. Conclusions and future work on quality assurance An issue of increasing importance is software quality. When evaluating the quality of various software products, the following characteristics should be considered: usability, reliability, efficiency, functionality, maintainability and portability. Accreditation is an excellent way to bring clarity to routine work. Written instructions and practices help people to work more carefully and less uncertain situations occur. However, some improvements are still needed. For example, intercomparison exercises should be
Towards quality excellence in radioanalytical laboratories at STUK, Finland
639
arranged to represent more routine work, including timetables for analyses and final reports of intercomparison results. Since the accreditation has been adopted, the whole working culture has needed to be developed continuously. This means, for example, that good practices need be to identified and documented. Moreover, any problems and all deviations must be documented and corrected actions must be performed and followed up. Internal audits are a good tool for following the effects of actions on work; they can also provide ideas for future development. Selfassessments and reviews of guidelines are also regularly arranged for work development and to ensure that the practice and guidelines are in good balance (European Foundation for Quality Management, 2005). Our accreditation is fixed to certain detailed methods. In the future, we plan to extend the accreditation model from fixed to flexible scope (i.e. to cover radioactivity in the environment and foodstuffs) at STUK (Finas, 2004).
References Adams, C., Cammann, K., Deckers, H.A., Dobkovski, Z., Holcombe, D., LaFleur, P.D., Radvila, P., Rohrer, C., Steck, W., Vermaercke, P. (1998). Quality assurance for research and development and non-routine analysis. Eurachem/Citac Guide. Ellison, S., Rosslein, M., Williams, A. (2000). Quantifying uncertainty in analytical measurement, Second edition. Eurachem/Citac Guide, CG4. Ellison, S., King, B., Rösslein, M., Salit, M., Williams, A. (2003). Traceability in chemical measurements. A guide to achieving comparable results in chemical measurement. Eurachem/Citac Guide. European Foundation for Quality Management (EFQM) (2005). The EFQM Excellence Model. Brussels. Available at http://www.efqm.org/ (15.3.2005). European Standard EN ISO/IEC 17025 (2000). General requirements for the competence of testing and calibration laboratories (ISO/IEC 17025:1999). FINAS (2004). FINAS Bulletin 3/2004. IAEA, International Atomic Energy Agency (2004). Quantifying uncertainty in nuclear analytical measurements. TECDOC-1401. IAEA, Vienna. Ilus, E., Ilus, T., Ikäheimonen, T.K., Niemistö, L., Herrmann, J., Suplinska, M., Panteleev, J., Ivanova, L., Gritschenko, G., Neumann, G. (2000). Intercomparison of sediment sampling devices using artificial radionuclides in Baltic Sea sediments. The MOSSIE Report. Baltic Sea Environment Proceedings No. 80. HELCOM, Helsinki. International Organization for Standardization (1995). Guide to the expression of uncertainty in measurement. BIPM, IEC, IFCC, ISO, IUPAC, IUPAP, OIML, Corrected edition (First edition, 1993). Geneva, Switzerland. International Organization for Standardization (2000). International Standard ISO 9004: 2000. Quality management systems – Guidelines for performance improvements, Second edition. Geneva, Switzerland. International Standard IEC 1452 (1995). Nuclear instrumentation – Measurement of gamma-ray emission rates of radionuclides – Calibration and use of germanium spectrometers, First edition. Geneva, Switzerland. Quevauviller, P. (2002). Quality Assurance for Water Analysis. Wiley, New York.
This page intentionally left blank
641
Author Index
Abe, T., 554 Ahmad, E., 382 Ahmad, N., 382 Airey, P., 307 Aközcan, S., 272 Aldridge, J., 183 Allen, M., 345 Aoyama, M., 53, 67 Aranami, K., 24 Avaullee, F., 538 Ayalon, A., 297 Bar-Matthews, M., 297 Barci, V., 215 Barci-Funel, G., 215 Baskaran, M., 96 Beck, J.W., 3, 592 Benbrahim, S., 148 Benkrid, M., 156 Benmansour, M., 148 Bexon, A., 183 Biddulph, D.L., 3, 592 Boaretto, E., 297 Bode, A., 281 Boudjenoun, R., 156 Boulahdid, M., 137, 156 Bürger, S., 581 Buda, R.A., 581 Burr, G.S., 3, 592 Carmi, I., 297 Carrera, P., 281 Carroll, J., 245 Cassidy, J., 562 Chafik, A., 148 Charette, M., 345 Chouba, L., 137 Chung, C.S., 96
Comanducci, J.-F., 373, 538 Cunha, I.I.L., 419 Davuliene, L., 477 Dehnert, J., 326 de Oliveira, J., 345, 373 de Santis Braga, E., 345 Di Lullo, A., 223 Donahue, D.J., 592 Færevik, I., 193 Figueira, R.C.L., 419 Florou, H., 235 Fukasawa, M., 53 Furtado, V.V., 345 Gäfvert, T., 193 Gann, J., 3 Gasa, S., 83 Gastaud, J., 255, 617 Gaudino, S., 223 Gaye, C.B., 382 Geckeis, H., 581 Göktepe, B.G., 265 Güngör, E., 265 Güngör, N., 265 von Gostomski, C.L., 581 Hamajima, Y., 511 Hatheway, A.L., 3 Haug, T., 245 Henderson-Sellers, A., 39, 307 Herut, B., 360 Heusser, G., 495 Hirose, K., 53, 67 Hodgins, G.W.L., 3 Hong, G.H., 96 Huber, G., 581
Ibn Majah, M., 148 Ikäheimonen, T.K., 629 Ikeuchi, Y., 397 Innes, S.C., 562 Innocenzi, P., 223 Inoue, M., 554 Iosjpe, M., 463 Iseda, K., 83 Javed, T., 382 Jia, G., 223 Jokšas, K., 165 Jull, A.J.T., 3 Kadi-hanifi, M., 156 Kafri, U., 360 Kaihola, L., 520 Kanduˇc, T., 317 Kawamura, H., 83 Kehagia, K., 235 Kim, C.K., 67, 96 Kim, C.S., 53, 67 Kim, S.H., 96 Kim, Y.I., 96 Klemola, S., 629 Knoeller, K., 326 Köksal, G., 265 Komura, K., 53, 511, 554 Kratz, J.V., 581 Kronfeld, J., 297 Kunz, P., 581 La Rosa, J., 617 Laissaoui, A., 148 Lange, T.E., 3 Laubenstein, M., 495 Lavrova, T., 206 Lee, H.M., 96 Lee, S.-H., 96, 137, 156, 617
642 Less, G., 360 Levy-Palomo, I., 373, 538, 617 Liddiard, M., 562 Lifton, N.A., 3 Lujanien˙e, G., 165 Mahjoub, A., 137 Manikandan, N.M., 554 Mantoura, F.R., 137 Mantoura, R.F.C., 53 Mashiatullah, A., 382 McDonald, P., 449 McGuffie, K., 39, 307 Menacer, M., 156 Michel, H., 215 Mork¯unien˙e, R., 165 Murata, Y., 554 Nakano, M., 433 Nakayama, T., 83 Neder, H., 495 Noureddine, A., 137, 156 Ogrinc, N., 317 Omoto, K., 406 Oregioni, B., 255, 373, 617 Özden, B., 272 Pagava, S.V., 255 Parker, T.G., 449 Passler, G., 581 Pham, M.K., 255, 617
Author Index Plastino, W., 335, 520 Pontis, J.-L., 137 Porteiro, C., 281 Povinec, P.P., 53, 67, 137, 148, 156, 255, 373, 529, 538, 617 Privitera, A.M.G., 373 Qureshi, R.M., 382 Readman, J.W., 24 Reguigui, N., 137 Remmert, A., 581 Rissanen, K., 245 Roos, P., 599 Rosamilia, S., 223 Rowland, S.J., 24 Rudjord, A.L., 193 Russ, S., 562 Saito, R.T., 419 Samaali, M., 137 Sanchez-Cabeza, J.A., 137 Sansone, U., 223 Savidou, A., 235 Schertz, M., 215 Schubert, M., 326 Shima, S., 83 Sihra, K., 183 Šilobritien˙e, B., 165 Soroka, Y., 206 Spirkauskaite, N., 477 Stiller, M., 297
Stone, D.J.M., 307 Sunlu, U., 272 Tarasiuk, N., 477 Tasneem, M.A., 382 Tessler, M.G., 419 Toole, J., 562 Topcuo˘glu, S., 272 Torri, G., 223 Trabidou, G., 235 Trautmann, N., 581 Treutler, H.-C., 326 Trinkunas, G., 477 U˘gur, A., 272 Valkunas, L., 477 Vaupotiˇc, J., 317 Vesterbacka, P., 629 Vives i Batlle, J., 449 Voitsekhovitch, O., 206 Vojtyla, P., 529 Weinstein, Y., 360 Weiss, H., 326 Wilson, R.C., 449 Yamaguchi, Y., 554 Yang, D.B., 96 Yechieli, Y., 297 Yener, G., 272
643
Keyword Index
A abundance sensitivity 599 accelerator mass spectrometry 3, 592 activation method 511 activity concentration 477 advection 433 Aegean Sea 272 airborne nuclides 554 Alboran Sea 148 α-ray spectrometry 223 algae 255, 265 Algerian Basin 137 Alpine wetland 215 Amazon 39 americium-241 (241 Am) 137, 148, 215, 255 anthropogenic radionuclides 53, 83, 137, 255, 554, 617 anti-Compton spectrometer 529, 538 aquifer 326, 335, 360 Atlantic Ocean 53, 148, 373 atmosphere 96 B background 538 Baltic Sea 165, 477 beryllium-7 (7 Be) 554 beryllium-10 (10 Be) 3 beach rock 406 sand 265 bioaccumulation 245, 272 biokinetic model 449 bismuth-214 (214 Bi) 373 Black Sea 265 bottom sediments 165 Brazil 345, 373, 419
C caesium-137 (137 Cs) 53, 96, 137, 148, 156, 165, 215, 255, 397, 419, 433, 477, 554, 562 carbon-13 (13 C) 317, 382, 406 carbon-14 (14 C) 3, 83, 297, 406, 520 Caspian Sea 255 caviar 255 Chernobyl accident 156 chlorinated organic compounds (COCs) 24 chlorine-37 (37 Cl) 24 close-in fallout 67 coastal sediment 419 water 83 exchange rates 345 zone 373 coincidence–anticoincidence spectrometer 538 compound-specific isotopic analysis (CSIA) 24 radiocarbon analysis (CSRA) 24 corals 592 cosmic-ray background 520, 529 cosmogenic radionuclides 495 CSERAM 183 Curonian Lagoon 477 D distribution coefficient (Kd ) 165 deep well 382 deforestation 39 depleted uranium penetrator 581 depuration 449 deuterium (2 H) 39, 307, 317, 382 DIC 297 diffusion 433 diurnal variation 554 Dounreay 562
644 E earthquake 335 East China Sea 96 Elbe River 326 environmental neutrons 511 radioactivity 397 background 520 estuary 419 F fish 255, 265 food web 281 fresh water 477 G gamma-ray spectrometry 373, 538, 562, 629 GCMs 307 GEANT 529 geophysics 335 Germany 326 global fallout 67, 156 gold 511 Greece 235 greenhouse 39 Greenland 245 groundwater 297, 326, 335, 373, 382, 581 re-charge 307 H heavy metals 419 Hegura Island 554 Holocene 406 HPGe detector 529, 538 hydrodynamic model 477 hydrogen-2 (2 H) see deuterium hydrogen-3 (3 H) see tritium I iodine-129 (129 I) 3, 83, 449, 592 IAEA-385 617 IAEA-414 617 ICP-MS 599 Indian Ocean 53 infiltration 326 intercomparison exercise 617 inventory 67 isotope fractionation 406 J Japan 406, 554 coast 397 Sea 433
Keyword Index K Karachi Sea 382 L lake sediment 215 lead-210 (210 Pb) 148, 156, 193, 419, 511, 554 dating 419 lead-212 (212 Pb) 554 liquid scintillation spectrometry 520 Littorina littorea 449 low-level gamma-ray spectrometry 554 Ge-spectrometry 495 M MARINA II 183 marine biota 255, 617 environment 345, 463 modelling 183 organisms 265 radioactivity 156, 617 mass spectrometry 599 mean residence time 137 migration 326, 581 Mediterranean Sea 137, 148, 156, 360, 373 model comparison 183 modelling 463 Monte Carlo simulation 529 Morocco 148 mussel 265, 272 N NaI(Tl) detector 538 Nansei Islands 406 natrium-22 (22 Na) 554 natural radionuclides 223, 255, 495, 554, 617 neutron self-absorption 511 nitrogen-15 (15 N) 281 non-instantaneous mixing 463 NORM 193 North Pacific 67 North Sea 193 Northwest Pacific 83, 96 O oceanic observation 83 OGCM 433 oil and gas industry 193 OSPAR 183 oxygen-18 (18 O) 39, 307, 317, 382 P Pacific Ocean 53, 67, 83, 96 Pakistan 382
Keyword Index paleoclimate studies 3 pelagic fish 281 periwinkle 449 plankton 281 plastic scintillation detector 538 plutonium isotopes 53, 67, 83, 96, 137, 148, 215, 255, 433, 581, 599 plutonium-238 (238 Pu) 215 plutonium-239 (239 Pu) 67, 599 plutonium-240 (240 Pu) 67, 599 plutonium-239,240 (239,240 Pu) 83, 96, 148, 215, 255, 433 plutonium-240/plutonium-239 atom ratio 67, 599 polonium-210 (210 Pu) 255, 265, 272, 554 precipitation 397 produced water 193 profile 67 Q quality assurance
629
R radiation dose 223, 235 radioactive particles 562 radiocaesium see caesium-137 radiocarbon see carbon-14 radiocarbon dating 3, 406 radiochemical analyses 629 separation 215 radiological risk 206 radiometrics 629 radionuclide inventory 156 radionuclides 96, 148, 206, 215, 235, 373, 397, 419, 433, 463 radium isotopes 345 radium-226 (226 Ra) 193, 235, 360 radium-228 (228 Ra) 193, 360 radon (222 Rn) 235, 326, 335, 360, 373, 554 radon daughters 335 re-evaporation 39 reference material 617 remotely-operated vehicle 562 reprocessing plant 83 resonance ionisation mass spectrometry 581 river models 307 water 96, 206, 326 Rokkasho 83 runoff 39 S salinity 360 Sardinia Channel
137
645
Sava River 317 scavenging 433 sea level change 406 seals 245 seasonal variation 554 seawater 53, 83, 96, 148, 156, 165, 193, 345, 360, 373, 397, 477 shallow well 382 SHOTS 53 Si beta-ray detector 538 Sicily 373 Channel 137 sediment 96, 148, 156, 272, 397, 477, 617 sedimentation model 419 rate 148, 156 soil 215, 397 Soreq Cave 297 sorption–desorption 165 South Mediterranean 137 Southern Ocean 53 Southwest Atlantic 373 speciation 165 spring water 235 stable isotopes 281, 317, 382 carbon-13 (13 C) 24, 382 deuterium (2 H) 39, 307, 382 chlorine-37 (37 Cl) 24 nitrogen-15 (15 N) 281 oxygen-18 (18 O) 39, 307, 317, 382 water 39, 307 stalactite 297 Strait of Gibraltar 148 strontium-90 (90 Sr) 96, 137, 215, 255, 397 sturgeon 255 submarine groundwater discharge 360, 373 sulphur-34 (34 S) 382 surface water 317, 382 suspended particles 165 T temporal variation 397 tracer 326 transpiration 39 tritium (3 H) 83, 297, 317 trophic position 281 Turkish coast 265 U Ukraine 206 underground laboratory uptake 449 uranium hydride 599
511, 520, 529, 538, 554
646 uranium isotopes 206, 223 uranium-234 (234 U) 206, 223 uranium-235 (235 U) 223 uranium-238 (238 U) 206, 223 W water 223, 511
Keyword Index column 53, 67, 83, 137, 433 masses 83 resources 307 Y Yellow Sea 96