ACID RAIN RESEARCH: DO W E HAVE ENOUGH ANSWERS?
This Page Intentionally Left Blank
Studies in Environmental Science 64
ACID RAIN RESEARCH: DO WE HAVE ENOUGH ANSWERS? Proceedings of a Speciality Conference, ‘s-Hertogenbosch, The Netherlands, 10-12 October 1994
Edited by:
G.J. Heij and J.W. Erisman National Institute of Public Health and the Environment, F!O. Box 1,3720 BA Bilthoven, The Netherlands
ELSEVIER Amsterdam Lausanne N e w York Oxford. Shannon Tokyo 1995
-
-
-
ELSEVIER SCIENCE B.V. Sarah Burgerhartstraat 25 P.O. Box 21 1, 1 0 0 0 AE Amsterdam, The Netherlands
ISBN 0-444-82038-8
0 1995 Elsevier Science B.V. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the Publisher, Elsevier Science B.V., Copyright & Permissions Department, P.O. Box 521, 1000 A M Amsterdam, The Netherlands. Special regulations for readers in the U.S.A. This publication has been registered with the Copyright Clearance Center Inc. (CCC), Salem, Massachusetts. Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the U.S.A. All copyright questions, including photocopying outside the U.S.A., should be referred t o the copyright owner, Elsevier Science B.V., unless otherwise specified.
No responsibility is assumed by the Publisher for any injury and/or damage t o 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. Printed in The Netherlands on acid-free paper
V
FOREWORD "his book represents the Proceedings of the International Specialty Conference, "AcidRain Research; do we have enough answers",held for about 120 scientists from 15 countries 10 - 12 October 1994 in 's-Hertogenbosch in the Netherlands. The conference proved a valuable conclusion to the co-ordinated research on acidification in the Netherlands, lasting from the beginning of 1985 to the end of 1994. Directly following the conference, an international team of experts in the field reviewed the research of the third and last phase of the Dutch Priority Programme on Acidification. The main results of the first two phases including a scientific review were published in the Elsevier series on Studies in Environmental Sciences, no. 46 (Heij and Schneider, 1991), while the results of the third phase of the programme, including the review team's report, will also be published in the same series. "he Specialty Conference focused on: Atmospheric deposition Effects of acid deposition on forest ecosystems in the Netherlands Future of acidification research. Atmospheric deposition has been a major research issue in several national and international research programmes. The aim of the Dutch Priority Programme on Acidification in this field was to assess acid, nitrogen and base-cation deposition loads to forest and heathland, and to compare these loads with critical deposition values to determine exceedances. As the critical loads concept is applied to ecosystems, deposition fluxes must also be assessed a t the ecosystem level. During the conference, special attention was given to the following subjects: trace gases, chaired by David Fowler (Institute of Terrestrial Ecology, UK); ammonia, chaired by Willem Asman (National Environmental Research Institute, Denmark) and particle deposition, chaired by J a n Willem Erisman (National Institute of Public Health and Environmental Protection, the Netherlands). Other topics, such as wet deposition, fog and cloud-water deposition, important for obtaining a n overall assessment of deposition loads to ecosystems and soils, were discussed in a session on generalisation chaired by Bruce Hicks (National Oceanic and Atmospheric Administration, USA). At the end of a long-term research programme the question usually arising is: Do we have enough answers, or are we generating new problems to keep our research going .......? Final results and conclusions of the Dutch research on forest stands and forest soils were presented and discussed in that light in a session chaired by BertJan Heij (National Institute of Public Health and Environmental Proteciton, the Netherlands). The session on "Futureof acidification research" on the last day of the conference brought up the question of whether present day knowledge and research trends have attracted sufficient support for decision-making purposes. This session was chaired by Ellis Cowling (College of Forest Resources, North Carolina State University, USA). Future acidification research has to be combined with research on other environmental topics, such as climate change, landuse changes or ecosystem dynamics, incorporating all relevant stress factors. A special session on these topics was chaired by Tomas Paces (Czech Geological Survey, Czech Republic).
vi
Each chairman summarized the main conclusions of his session. These conclusions, and an answer to the question: Do we have enough answers? are listed in the chapter titled Conclusions. The proceedings of the Conference start with the opening statements by Andre van Alphen and are followed by the papers presented during the different sessions along with about 30 posters for explanation to the visiting scientists. Posters were divided into the following topics: critical loads I exceedances, wet deposition I throughfall, dry deposition I concentrations and a miscellaneous session.
ACKNOWLEDGEMENTS The editors would like to express their gratitude for the outstanding effort of all the chairmen and, in particular, the organizing chairman, Toni Schneider. Ottelien van Steenis not only carried out the organisation of the Conference but also operated the Registration and Information desk, with the excellent help of Marianne Vonk. Last but not least, Ottelien also took care of all the preparatory work for the Proceedings. Ottelien and Marianne are therefore greately acknowledged for their contributions. We hope that although the Conference focused on the three topics: atmospheric input, summary of Dutch Acidification research results and the future of acidification research, the excellent calibre of work and new initiatives on acidification research as a whole, reflected in the proceedings, will be of value to both research scientists and policy makers.
vii
CONCLUSIONS Deposition 1. Trace gases - Fluxes of the acidifying compounds NO2, NH3, SO2 and aerosols to forests and short vegetation have not only been measured directly, overcoming important uncertainties in methods and interpretation, but also have heen monitored over long periods. This work provides the basis for greatly improved accuracy of input estimates of pollutants to forests and the landscape in The Netherlands and across Europe. 2. Ammonia / ammonium
- The highest uncertainty in estimates of NH3 deposition is caused by uncertainties in temporal and spatial variations in NH3 emissions.
- The conversion rate of NH3 to N H 4 + aerosol is not known accurately. It is likely that it shows temporal and spatial variations, that e.g. depend on the concentration of acidic compounds in the atmosphere. This information should be known as it determines where NH, will be deposited, because the dry deposition velocity of NH3 is much larger than the dry deposition velocity of aerosol. For that reason reduction of emissions of acidic compounds in the air only could lead to a change in the dry deposition pattern of NH3. The concentration of NH3 at the surface of vegetation and seawater determines partly the flux of NH3 to or from the surface. It should be taken into consideration in transport modelling. NH3 emissions from agricultural crops could be potentially important in the growing season.
m+
-
3. Particle deposition - Dry deposition of particles to forests has often been underestimated until now. Furthermore, the role of particles in regulating water layer (chemistry) on vegetation and thus influencing gaseous dry deposition is important.
4. Generalization - Deposition should be determined at a scale that enables the estimation of risk for ecosystem damage. Furthermore, most important factors determining deposition (edge effects, slopes, topography, roughness transition zones, etc.) should be taken into account in estimating input to sensitive ecosystems. For model development it is necessary to obtain key parameters by field experiments and validate the models by further field measurements. Effects of acid deposition 1. No direct relationship exists between tree health and acid deposition. 2. Atmospheric deposition of N and S compounds on forests leads to:
viii
- changes in vegetation composition into the direction of nitrogen-loving species and monocultures;
- high concentration of Al and NO3 in soil solution and groundwater and to loss of biodiversity in non forest ecosystems. 3. Ozone has a significant adverse impact on plants. Not only crops, but also forest trees can be affected. The impact on natural vegetation is largely unknown as yet. In The Netherlands the contribution of NO, to the total nitrogen deposition is currently less than 20%.But its adverse impact through formation of ozone must not be neglected. 4. The impact of atmospheric deposition on forest trees should be evaluated in terms of risk rather than in terms of visible effects. The future 1. Global climate change and land use change will influence acidification processes; 2. A shift is necessary from effect oriented to system oriented research; 3. Ecologists, studying acidification effects, have to include climate factors; 4. Scientific uncertainties have to be reported explicitly. 5. Long term monitoring programmes are necessary to evaluate effects of acidification and of policy actions. 6. "Local" processes are largely unknown (especially for N). Knowledge on "local" processes will improve knowledge on causal relations.
ix
CONTENTS
Foreword
V
Acknowledgements
vi
Conclusions
vii
OPENING SESSION
1
Opening Remarks A. van Alphen, Chairman Steering Committee Acidification Research
3
ATMOSPHERIC DEPOSITION Session I Trace aases Long term measurements of SO, dry deposition over vegetation and soil and comparisons with models D. Fowler, C. Flechard, R.L. Storeton-West, M.A. Sutton, K.J. Hargreaves and R.I. Smith
7 9
The use of the gradient method to monitor trace gas fluxes over forest: flux-profile functions for ozone and heat J. Duyzer and H. Weststrate
21
Deposition of nitrogen oxides and ozone to Danish forest sites K. Pilegaard, N.O. Jensen and P. Hummelshj
31
Monitoring dry deposition fluxes of SO, and NO,: analysis of errors M.G. Mermen, J.E.M. Hogenkamp, H.J.M.A. Zwart and J.W. Erisman
41
Session 11 Ammonia /Ammonium Ammonia and ammonium in the atmosphere: present knowledge and recommendations for further research W.A.H. Asman
53
Measurement and modelling of ammonia exchange over arable surfaces M.A. Sutton, J.K. Burkhardt, D. Guerin and D. Fowler
71
Preliminary Validation of ammonia emission data using a combination of monitoring and modelling J.M.M. Aben, P.S.C. Heuberger, R.C. Acharya, and A.L.M. Dekkers
81
55
X
Deposition network of the Federal Environmental Agency (UBA) Results and trends D. Kallweit
91
The influence of ammonium nitrate equilibrium on the measurement of exchange fluxes of ammonia and nitric acid Y. Zhang, H. ten Brink, S. Slanina and P. Wyers
103
Session 111 Particle deposition Particle deposition to forests J.W. Erisman, G. Draaijers, J. Duyzer, P. Hofschreuder, N. van Leeuwen, F. Romer, W. Ruijgrok and P. Wyers
113 115
Deposition of aerosol to coniferous forest G.P. Wyers, A.C. Veltkamp, M, Geusebroek, A. Wayers and J.J. Mols
127
Microscopic processes governing the deposition of trace gases and particles to vegetation surfaces J. Burkhardt
139
The atmospheric input of inorganic nitrogen and sulphur by dry deposition of aerosol particles to a spruce stand K. Peters and G. Bruckner-Schatt
149
Session IV
Generalization: total atmospheric deuosition and soil loads: measurements and models On the determination of total deposition to remote areas B.B. Hicks
161
Quantifying the scale dependence in estimates of wet and dry deposition and the implications for critical load exceedances R.I. Smith, D. Fowler, and K.R. Bull
175
Uncertainties associated with the inferential modelling of trace gas dry deposition: A comparison of four models with observations from four surface types J.R. Brook and J. Padro
187
EDACS: European deposition maps of acidifying components on a small scale W.A.J. van Pul, C.J.M. Potma, E.P. van Leeuwen, G.P.J. Draaijers and J.W. Erisman
197
163
xi EFFECTS OF ACID DEPOSITION ON FOREST ECOSYSTEMS IN THE NETHERLANDS Effects of acid deuosition on forest ecosystems in The Netherlands Assessment and evaluation of critical levels for 0, and NH, E. Steingrover, T. Dueck and L.G. van der Eerden
21 1
Experimental manipulations: forest ecosystem responses to changes in water, nutrients and atmospheric loads A.W. Boxman, P.H.B. de Visser and J.G.M. Roelofs
223
Effects of acid deposition on forest ecosystems in The Netherlands: analysis of the Speuld Douglas fir site H. van Grinsven, B.J. Groenenberg, K. van Heerden, H. Kros, F. Mohren, C. van der Salm, E. Steingrover, A. Tiktak and J.R. van de Veen
245
Large scale impacts of acid deposition on forests and forest soils in The Netherlands W. de Vries, E.E.J.M. Leeters, C.M.A. Hendriks, H. van Dobben, J. van den Burg and L.J.M. Bouwmans
26 1
Ecological effects of atmospheric deposition on non-forest ecosystems in Western Europe R. Bobbink and J.G.M. Roelofs
279
Evaluation; integration H.F. Van Dobben
293
Session V
213
FUTURE OF ACIDIFICATION RESEARCH Todav’s knowledge; is it sufficientfor tomorrow’s decision makina uuruoses Lessons learned in acidification research: Implications for future environmental research and assessments E. Cowling
305
Exceedence, damage and area minimisation approaches to integrated acidic deposition modelling C. Gough, J.J. Kuylensstierna, P. Bailey and M.J. Chadwick
321
Reliability of environmental information obtained by modelling and monitoring J.A. Hoekstra, J.C.H. van Eijkeren, A.L.M. Dekkers, B.J. de Haan, P.S.C. Heuberger, P.H.M. Janssen, A.U.C.J. van Beurden, A.A.M. Kusse and M.J.H. Pastoors
333
Session VI
307
xii Session VII Future research: combination with other environmental touics Future of acidification research T. Paces
34 1 343
Potential ecological risk due to acidification of heavy industrialized areas the Upper Silesia case A. Worsztynowicz and W. Mill
353
Acidification and metal mobilization: effects of land use changes on Cd mobility P.F. Rornkens and W. de Vries
367
Acidification interacting with global changes: research to manage drifting systems R.S.A.R. van Rompaey
381
POSTERS Session VIII Critical loads / Exceedances Critical loads of heavy metals for European forest soils G.J. Reinds, J. Bril, W. de Vries, J.E. Groenenberg and A. Breeuwsma
385 3 87
Setting critical loads of acidity for dystrophic peat - a new approach E.J. Wilson, R.A. Skeffington, C.J. Downer, E. Maltby, P. Immirzi and C. Swanson
3 89
A comparison of models for the assessment of critical loads on different scales of observation R.J.M. Lenz, S. Mendler and R. Stary
393
Session LY Wet deaosition / Throuahfall Deposition and leaching at forest stands in Sumava Mts. J. Kubiznhkovii, J. Kubiznhk and 0. Rauch
395 397
Precipitation input of inorganic chemicals in the S. Vitale pine stand of Ravenna (Italy) T. Georgiadis, F. Fortezza, L. Alberti, P. Rossini and V. Strocchi
399
Trends of some components of wet deposition in East Germany after the unification E. Briiggeman and W. Rolle
403
Eight years studying bulk and wet deposition in Spanish Basque country D. Encinas and H. Casado
407
Scavenging of gases during growth of ice crystals G. Santachiara, F. Prodi and F. Vivarelli
41 1
...
Xlll
Contribution of root-derived sulphur to sulphate in throughfall in a Douglas fir forest A.C. Veltkamp, M. Geusebroek and G.P. Wyers
413
Fate of nitrogen in Spruce and Pine ecosystems T. Staszweski, S. Godzik and J. Szdzuj
417
Acid deposition: data from the Swiss Alps S. Braun and W. Fluckiger
419
Dry deposition to bulk samplers underneath a roof in a spruce (picea abies Karst.) forest M. Bredemeier and M. Meyer
423
Deposition on air pollutants to forest ecosystems along pollution and climatic gradients in Poland S. Godzik, T. Staszweksi and J. Szdzuj
425
Session X Dn,deDosition / Concentrations Immission and dry deposition of SO, and ozone lee side of the conurbation of Leipzig in Eastern Europe G. Spindler and W. Rolle
427 429
Dry deposition of acidifying and alkaline particles to forests: model and experimental results compared W. Ruijgrok
43 1
The measurement of ammonia in the National Air Quality Monitoring Network (LML) (1) Instrumentation and network set-up B.G. van Elzakker, E. Buijsman, G.P. Wyers and R.P. Otjes
439
The measurement of ammonia in the National Air Quality Monitoring Network CML) (2) Results and performance B.G. van Elzakker, J.T. Schippers, J. Stuiver and G.J.B.M. van Uden
443
Fine resolution modelling of ammonia dry deposition in Great Britain R.J. Singles, M.A. Sutton and K.J. Weston
449
Fog deposition measurements on Douglas Fir forest A.T. Vermeulen, G.P. Wyers, F.G. Romer, G.P.J. Draaijers, N.F.M. van Leeuwen and J.W. Erisman
453
The contribution of canopy exchange to differences observed between atmospheric deposition and throughfall fluxes G.P.J. Draaijers, J.W. Erisman, N.F.M. van Leeuwen, F.G. Romer, B.H. te Winker and G.P. Wyers
455
xiv Dry deposition monitoring SO,, NH, and NO, over a coniferous forest J. Hogenkamp, J.W. Erisman, M. Mermen, E. Kemkers, A. van Pul, G. Draaijers, J. Duyzer and P. Wyers
457
Gas deposition of sulphur dioxide on the territory of the Czech Republic in 1991 M. Zapletal
459
Session X I Miscellaneous Forest condition in Europe and North America: What have we leant over the past ten years? H. Visser
463 465
Stomata1 regulation in field-grown Douglas-fir W.W.P. Jans and E.G. Steingrover
473
Carbon partitioning in Douglas-fir E.G. Steingrover, M. Posma and W.W.P. Jans
475
Decreasing concentration of air pollutants and the rate of dry and wet acidic deposition at three forestry monitoring stations in Hungary L. Horvhth, Gy. Baranka and E.Gy. Fiihrer
477
The characteristics of acid precipitation in Southern China Y. Bai and X. Tang
483
The response of peat wetland methane emissions to temperature, water table and sulphate deposition D. Fowler, J. MacDonald, I.D. Leith, K.J. Hargreaves and R. Martynoga
485
Annex I
List of participants
489
OPENING SESSION
This Page Intentionally Left Blank
3
ACID RAIN RESEARCH CONFERENCE, October 10-12, 1994 Opening remarks by Andr6 van Alphen Deputy director Air and Energy, Ministry of Housing, Spatial Planning and Environment, P.O. Box 30945, 2500 GX, The Hague, The Netherlands Ladies and Gentlemen, Addressing this conference, while deputizing for Joris Al, the Chairman of the Steering Committee on Acidification Research, is a rather confusing experience. Of course I feel privileged to do so, but I realise that Joris Al, who chaired the entire third phase of the Dutch acidification research programme, is far more experienced in this matter than I. However, I have sufficient knowledge of the subject to know that there is another, more factual, reason for these mixed feelings. This Symposium marks the completion of a ten-year acidification research programme in the Netherlands, at a moment when the call for hard and conclusive scientific evidence coincides with the most drastic cut in research budgets ever. That is why I really feel confused. In the next ten minutes I intend to focus on three elements of the problem: -the desire for hard evidence; -the end of a research programme of long standing; -future acidification research.
To start with the first, I must say that the sub-title of the conference "Do we have enough answers?" is a perfect description of the policymakers' dilemma: we know a lot about acidification, but is it all the knowledge we need for policy purposes and would more knowledge lead to policy measures that are not only easier but also better? When looking into that question there are two points which have to be stressed. Firstly, results from scientific research can never be a substitute for policy decisions. Both scientists and policymakers may regard this as a truism, but it is still worthwhile repeating it now and then. Secondly, one should not forget that environmental problems like acidification, with a great lapse of time between the onset of the effects and evidence of damage, can only be dealt with on the basis of a risk approach. Action should be taken on the basis of the risk that acidification results in harmful effects which, if we postpone action until damage is apparent, will probably be irreversible. Since the start of acidification abatement the approach has been to start by gradually reducing emissions while intensifying research into acidification to consolidate the scientific basis for action. Of course the intention was to ensure that the timing of more drastic measures coincided with the development of further scientific substantiation of the acidification issue. In real life however, scientific knowledge develops more capriciously: not only does our
4
understanding of a great number of issues increase in the course of time, but other issues, which seemed clear in the beginning, pose new problems. This is what is happening at present with respect to ammonia. These questions are not necessarily fundamentale, but they require clarification to prevent the main issue from being confused. It is entirely understandable that when the social and economic consequences of abatement measures become extremely serious, those who have to pay for these consequences will have difficully accepting any validation that is unclear. Nevertheless, I do not expect new research programmes to result in a break-through in our knowledge about the risks of acidification. As I mentioned before, science cannot choose between protection of the environment and nature on the one hand and social and economic consequences of emission reduction measures on the other. It is the task of the government to weigh these interests. To enable government to come to a balanced decision, it is of great importance to be as clear as possible about what is known with respect to the risks of acidification and the reliability of that knowledge, using the tremendous amount of scientific information now available. The project team has the important and difficult task to create such clearness after this conference and the subsequent international scientific review process. And then the second element of the problem that causes the confusion: the end of the coordinated acidification research programme. A tremendous amount of knowledge has been generated in a unique cooperative venture involving numerous scientific institutes and scientists. Scientific cooperation on a research programme financed all these years by an equally unique form of cooperation between government and industry. An almost countless number of publications in scientific periodicals and theses has resulted from this acidification research programme over the years. I am sure that the international review will confirm that the research in the third phase of the research programme was of high quality, as it was in the first and second phases. Dutch scientists play an importante role in improving scientific understanding of acidifica-tion at a European level, and Dutch research has made an important contribution to the development of national and international acidification control strategies. Its role is exemplified by the development of the scientific basis for the recently signed UN-ECE Second Sulphur Protocol. Although this acidification research programme is now coming to an end, there are still questions unanswered and undoubtedly new questions on specific aspects of the acidification process will arise when our national acidification policy is implemented further. In addition, acidification abatement at a European level, preparing "second generation protocols", will require further scientific support. That brings me to the third element of the problem: the future of acidification research. To generate answers to remaining questions at a national and international level, further investigations are certainly required. However, there is no specific need to incorporate relevant studies in a new coordinated research programme. It is considered sufficient that institutes which investigate acidification further can apply for government funding, competing with other air pollution research projects. In principle there is nothing wrong with such a development. However, it is alarming that at
5
the time the third phase of the acidification research programme comes to an end, the budget for acidification research of my ministry is drastically cut. Not much financial support for further acidification studies can be given when budgets considered necessary for the years to come are cut by 50% or more; as this is the probable outlook for 1995 and after. An important area of environmental research has been developed. After a decade of financial support by government and industry, it seems that acidification research is to be left to sink or swim on its own. Will it be possible to carry on or do we have to fear for the decline of research facilities in the Netherlands? I don't know. But I do not think it necessary to end on a gloomy note. Because of the position the acidification issue has acquired in scientific research, because of the enthusiasm and dedication of the scientists and because I believe that in the end its importance will once more be recognized at a political level, but above all, because scientists are extremely creative in raising funds, I have confidence in a positive future for acidification research and in the continuation of cooperation between scientists and policymakers.
I wish you all a successful conference here in Den Bosch, where on earlier occasions other memorable meetings on acidification have been hosted. But, I hope especially that your stay in this charming city and in the Netherlands proves to be a very pleasant one.
This Page Intentionally Left Blank
ATMOSPHERIC DEPOSITION SESSION I TRACE GASES
This Page Intentionally Left Blank
G.J.Heij and J.W Erisman (Editors). Acid Rain Research: Do we have enough answers? 0 1995 Elsevier Science BCI All rights reserved.
9
LONG TERM MEASUREMENTSOF SO, DRY DEPOSITION OVER VEGETATION AND SOIL AND COMPARISONS WITH MODELS
D. Fowler, C. Flechard, R.L. Storeton-West, M.A. Sutton, K.J.Hargreaves and R.I. Smith Institute of Terrestrial Ecology, Bush Estate, Penicuik, Midlothian, EH26 OQB, Scotland.
Abstract A semi-continuous series of measurements of SO, fluxes above soil, wheat and sugar beet has been used to quantify the major components of canopy and surface resistance in a wide range of conditions. The data show that over dry cereal canopy, the marked diurnal cycle in canopy resistance is regulated primarily by changes in stomata1 resistance. Good agreement between a process based model of SO, deposition and the field data in dry conditions has been obtained. For a dew and rain wetted crop, the canopy resistance is decreased from 80 s m-l to 55 s rn-'. Longterm (4 month) median deposition velocity (V,) was 7.2 mm s" for wheat. For sugar beet, similar results are obtained with median vd of 5.7 mm s-'. There is also a clear dependence of r, on SO, concentrationwith r, decreasing from 100 s m-' at 4 pg m-3 SO, to 40 s m-l at 20 pg m-3for both crops. For bare soil, canopy r, is small with a median value of 15 s rn-' for dry soil and 5 s m-' for wet soil. The widespread assumption of very small r, for wet surfaces is clearly an oversimplificationin current models. The influence of these new findings for annual dry deposition estimates will be discussed. For both crop surfaces vd increased in the presence of surface water for dew precipitation by typically 30%. The increased affinity of wet leaf surfaces for SO, uptake w a s equivalent to a 40% decrease in the canopy resistance r,.
-
INTRODUCTION The turbulent deposition of SO, to terrestrial surfaces represents the major removal process for boundary layer SO, over large areas of Europe. Even in the high rainfall regions dry deposition may contribute 20-30% of deposited sulphur. While wet deposition is routinely monitored throughout Europe and North America to provide estimates of wet deposition, the majority of countries rely on simple parameterization of dry deposition using monitored SO, concentrations and deposition velocities from the literature. In North America, the dry deposition monitoring network is in fact a series of monitoring stations for meteorological variables and SO, concentration for which deposition rates are inferred, the measurements do not provide flux to the surface directly.
10
Our understanding of dry deposition process in the field was provided by micrometeorological studies of fluxes of SO, to natural surfaces. The early work in the 1970's by Garland (1977) and Fowler and Unsworth (1979) amongst others, quantified the relative importance of surface and atmospheric processes in regulating deposition rates for a range of natural surfaces and atmospheric conditions. These analyses have been extended to provide annual dry deposition estimates on the basis of multiple resistance models Hicks et al. (1987), and similar approaches have been widely applied to estimate annual SO, inputs to catchments, regions or countries (Fowler, 1980; RGAR, 1990). Although it was widely considered that rates of dry deposition could not be continuously monitored the recent work by Erisman et al. (1993) shows that instrumentation has developed t o a point where simple flux-gradient approaches may be used to estimate dry deposition continuously at sites suitable for application of micrometeorologicalmethods. The measurements by Erisman et al. demonstrated the importance of surface water on foliage on rates of SO, deposition and provide an excellent basis for the temporal extrapolation of fluxes from measured SO, concentration fields with appropriate meteorological and land use information. This paper describes an SO, flux measurement station for continuous operation and reports results of a year of measurements of fluxes at an agricultural site in the English Midlands. The results are used to demonstrate the relative importance of Werent sinks at the terrestrial surfaces in a range of conditions, and the results of measurements are compared with model estimates. THEORY AND METHODS The measurements were made at a field site in arable agricultural land close to Sutton Bonington, a University of Nottingham field station. The equipment was placed close to a field boundary and provided a fetch of between 200 and 300 m in most wind directions, excepting the 45" sector centred on 180" within which a n instrumentation cabin was placed to house the monitoring and logging equipment. The site provided winter wheat, bare soil and sugar beet according to sector and time of the year during the measurement period of April 1993 until June 1994. Concentrations of SO, were monitored using a high sensitivity pulsed fluorescence monitor (TECO 43s) which sampled filtered air along heated sample lines at two heights in the surface layer and up to a height of 2.3 m. Wind speeds were measured at three heights and air temperature at two heights using sensitive cup-anemometers (Vector Instruments) and miniature thermocouples (Campbell Scientific) respectively. Precipitation, surface wetness, wind direction and total radiation were also monitored continuously at the site. All of the instrumentation was controlled and logged using a Campbell 21X micrologger. In addition to the long-term continuous measurements, 2 campaigns each of 2 weeks duration of measurement at the site were also made using a 5-point gradient system for SO,, temperature and windspeed and a Campbell instruments Bowen ratio system. The field data were analysed to provide fluxes of SO, (F,) calculated from
11
where k i s Von Karmans constant, u is windspeed, z height, x is SO, concentration, d the zero place displacement and ($,&J1 is a stability factor to correct fluxes for the effects of thermal stratification of the surface layers of the atmosphere. A detailed account o f the standard micrometeorological theory is outside the scope of this short paper a n d may be found in m o m (1976) or Monteith & Unsworth (1990). The sequential sampling system provided a measure of vertical gradients during periods w i t h constant SO, concentration. However, non-stationarity in SO, concentration introduced an important source of uncertainty in the sequential sampling system, and contributed an average of 18% of the uncertainties in V,. To overcome these effects the results were detrended and running median values were obtained for SO, concentration ( x ) , deposition velocity (Vd),the flux F, and r, with a time constant of 3 hours. For the continuous monitoring of fluxes there are a range of conditions which result in erroneous fluxes. It was necessary therefore to filter out data obtained when the wind direction was within the sector occupied by the instrument cabin, when atmospheric stability effects were too large to be corrected by standard approaches, and in this case data were rejected when Monin-Obukhov lengths were < 5m. For periods during which SO, concentration changed rapidly, non-stationarity effects were important, but these also provided important data for the investigation of processes and whenever possible, corrections to the data set to overcome sequential sampling errors were made and the data were accepted.
SO, concentrations at the site The half-hour mean values of SO, concentrations for one year of monitoring are, as is usual, log normally distributed. The mean concentration is 12.8 yg m-3, geometric mean 6.1 yg m-3 a n d geometric standard deviation is 3.5 as shown in Fig. 1. There is a clear diurnal cycle in SO, concentration at the site with a rninimum at F ~ ( K 5) 0300 GMT and at 1300 maximum GMT. Similarly, windspeed shows a -2 -1 0 1 2 3 marked diurnal cycle =f-=@a lb with a minimum at 0500 G M T and a maximum at 1400 Figure 1. The log-normal distributionof SO, concentrations(April 1993 - April 1994).
12
GMT. The SO, data show a pronounced sector dependence with concentrations being largest (- 15 pg m-3)with northerly winds and smallest with southerly winds 5 pg m-3,as shown in Fig. 2. The measurements of SO, deposition may be divided into three groups, according to the surfaces present during the year within the upwind fetch, as wheat, sugar beet and bare soil.
w c w h0
-
. . .
,
243.
'.
, - , '
- . ,
'
-
?I0
=I< E
. ,
,
.
- ,
'
1IU
.
150
'
HO
m Figure 2. Wind sector dependence of sulphur dioxide air concentrations at Sunon Bonington (April 1993 - April 1994).
Dry deposition on to wheat
The measurements over winter wheat were made during the period AprilJuly 1993 during which the canopy height and leaf area index increased from 30cm to -1OOcm and 1.5 to 4.5 respectively. Considering the whole data set for the period, the median of the population of hourly average data for deposition velocity was 7.2 mm 5.' whereas that of the maximum rates of deposition (Vmm)possible was 17.1 mm s-l showing that canopy resistance exerted a major control
Figure 3. FrequencydistributionofV-and V, Im. above a wheat canopy (April - July 1993). Negative values denote SO, emission - or more likely release b the atmosphere of pre-depositedSO,.
13
over rates of deposition under moist conditions (Fig. 3). The median canopy resistance rewas 81 s m-' for the whole data set. The leaf wetness sensor data were used t o show the average influence of leaf wetness on canopy resistance and deposition velocity as shown in Fig. 3. These data show, as expected, that re is significantlysmaller in wet conditions averaging 55 s m-' than in dry conditions (83 s m-') but the wet surfaces were not generally a perfect sink for the SOz. The generalized data from Fig. 3 conceal large diurnal variations in deposition velocity which result from both changes in both r,, r b and r,. Typical diurnal variations in vd for the vegetative phase of the crop growth are shown in Fig. 4 with afternoon mixima in V, of 10 to 15 mm s-' and nocturnal minima of 2 to 5 mm s-'. It is important to recognise that a significant fraction of the diurnal variation 0 fI in v d results from the decrease in ----v, windspeed and the increase in atmospheric stability during the night 1UOUl993 W30 0230 0430 0630 a30 1030 1190 I430 1630 1 8 3 2030 P:30 which lead to a marked diurnal cycle M Figure 4. Measured depositionvelocities of SO, over a wheat canopy. i n r, and rb. The effect is clear even .averaged over the April-July data set as I d shown for the wheat crop in Fig. 5. The small effects of liquid water on leaf surfaces on r, relative to those 0 : : : : : : : . : : : : . : : : : : reported by Erisman et al. (1993) may result from solution and oxidation 'b (..nil) processes in the liquid film not being sufficiently rapid to maintain or oxidise m:w, 0230w30 0 6 : a:w ~ iow 1130 I*:N 16:wIWO a130 0:)o GMr the S'" in solution as Figure 5. Mean diumalc ~ l ofe afmospheric aerodynamic (r. (Im.)) and viscous sub-layer (r,) resistances to SO, transfer above a wheat canopy (April - July 1993). Values are the water evaporates. geometric means and 95 % confidence intervals. An example of this
- I
5
:
'
:
:
14
effect is shown in Fig. 6 in which the wetting of the wheat canopy by rain initially results in a large rate of deposition but as the canopy dries after a brief shower a peak in SO, emission is observed. Clearly, the emission is consistent with loss of dissolved S'" back to the gas phase as water evaporates in warm sunny conditions, and the sequence of events is repeated following a further shower later i n the day. These effects are not the general rule but ~3002:30M.301:301:3010301230 143016:30 1P3020302230 to illustrate the do serve Qbfr point that the assumed Figure 6. SO2exchange over a wheat canopy (23/07/1993): influence of precipitation reduction in rc to and leaf W ~ Q I ~ S Son the dry&position velocity of SO,. Precipitationevents that close to zero require the occur at 11:OO and between 13:30 and 1430 result in a wetted canopy and of the chefid increased deposition velocity at 11:00-11:30 and 13:3014:00, soon after, emission occurs (negative v,) as the result of desorptionof unoxidized SO, S'" in solution to proceed during evaporation. t o SO: and for the acidity generated in the aqueous film to be neutralized by soil derived base cations or by ammonia to prevent r, increasing as the aqueous and gas phase SO, achieve an equilibrium. The process is discussed in some detail by Brimblecome (1978) and by Chameides (1987) largely on theoretical grounds but in neither case are there field data for canopy exchange of water and SO, to confirm the details of the kinetics of the processes. Such field data are necessary for a range of conditions to provide a satisfactory basis for modelling the effects of surface water on the values of rcand V, and hence long term rates of exchange. Until then, the application of average r, values obtained by experiment are the only alternative, and clear differences are shown in the data presented here and that by Erisman et al. (1993). The mean diurnal changes in V, for dry canopies are those driven mainly by changes in stomatal resistance with daytime minima for r, entirely consistent with measured stornatal resistance (rJ. These findings are consistent with the data for grassland by Erisman et al. (19931, Garland (19771, and for wheat by Fowler and Unsworth (1979). The collection of a substantial data set allows the variation in r, with a range of variables to be examined. In particular it is important to show whether the deposition velocity is strictly independent of SO, concentration. By sorting the data ~~
k(d)
! a 200
loo 1% 50
m
-* V.(ImP)
18 16
14
--
12. 10
1 -
so. c
.
( rl=*)
Figure 7. Dry deposition of SO, onto a wheat canopy (April-July 1993). Variations of canopy resistance and deposition velocity with SO, concentmtion. Reference height is lm.above the zero-plane.
-10
0
10
sq
30
u) . .'
40
50
(om?
Egure 8. Frequencydkautionof V, and V, Im. above a sugar beet canopy (May-July 1994).
by concentration the relationship between V,, r, and atmospheric resistances with SO, concentration has been investigated. The analysis shows in Fig. 7 that the atmospheric resistances are almost constant with SO, concentration in the range 1 to 20 p g m-3. However there is a clear reduction in r, with increasing SO, concentration over the range 2 to 12 g so, from 150 s rn-' at 2 p g m-3 to 50 s m-' at 12 pg SO,m'3. Dry deposition of SO, on to sugar beet An extensive data set obtained for measurements of SO, fluxes over sugar beet were obtained during the period May to July 1994. The median deposition velocity was 5.7 mm s-' whereas 1 hour of V,, was 14.6 mm s-'. Median deposition velocities to dry and wet canopies of sugar beet were 5.6 and 6.3 mm s" respectively as shown in Fig. 8. An example of the differences in behaviour of V, and r, in these measurements with those of Erisman may be seen in the data for two days over the sugar beet crop. The first day (8/6/1994)a warm
16
-Vd
I-.
....v-
.......
,..a.
10
,A -
Figure 9. Diurnal Mliations m canopyre&ance and deposition velocity of SO, over a dry sugar beet canopy (OWM/1994). z l
I
sunny early summer day, shows the common diurnal cycle in vd with mid-day maxima of 10 mm s-land small nocturnal minima of 2 m m s" in calm, strongly stable conditions (Fig. 9). By contrast, for a day during the preceding October (13/10/93) the deposition velocity shows V, very close to V,, all day, for a canopy in cool wet conditions (Fig.10). On this day there was no evidence of r, increasing with time in the presence of a wet canopy. The reason r, does not increase, and that the surface remains a 'perfect sink' for SO, is a matter of speculation but must be closely linked to the chemical processing of the dissolved SOz.
SOILS Following the harvest of the winter wheat crop the field was cultivated and provided a long period of smooth bare soil over which SO, fluxes were measured. (-.) 10 The median vd was 13 mm s-l and was quite close to the median V,, of 15 mm s-'. The underlying canopy resistance was therefore very small and averaged just 12 s m-'. The surface was behaving almost as a perfect sink for SO, and while the presence of liquid water did reduce r, as shown in Fig. 11to 4.9 s m 1the increase in deposition velocity was small (to 15 mm s-l)and was then almost equal to vmu. The very large rates of deposition to base soil were larger than those reported by Garland )1977) and by Payrissat and Beilke (1975). A consequence of the small r,
17
is that vd for the bare soil increased almost linearly with wind speed (Fig. 12). The r, values widely applied in models of SO2 dry deposition are generally much larger than those reported here and, if such values as those reported here are applicable generally, the uptake by arable land is winter will be much larger than has been assumed and landscape values for V, in winter should be significantly larger.
---
Comparison of modelled and measured deposition velocities -100 w m w w m m m w m i l a The application of resistance 1’ 11. ~requ~cydicttibutpn &tograms for measured deposition models to estimate rural SO, dry velocities (V,) and maximum deposition velocities allowed by deposition is now a standard technique turbulence (v-) Over bare soil at Bonington (JanUatY-MaY by Hicks et (1987), Fowler (1980) 1994). Figures in brackets are 95 % confidence intervals. and by Sandness (1993) and Erisman (1994). The application and methods in all cases differ although the underlying principles are common. In the method applied for the UK (RGAR 1990) the landscape is subdivided at 5 categories (arable, forest, grassland, moorland and urban). The stomatal response to light and temperature for species representative of such land uses are used to calculate the canopy resistance for water vapour, which is corrected for difisivity differences between SOz and HzO t o provide stomatal uptake for SOz. The leaf surface resistance is assumed constant and quantified by experiment. Climatological, or measured meteorological data (radiation air temperature, wind velocity and canopy heights) are then used to calculate vd which are combined with monitored SO, concentration to yield the flux. One of the most valuable products of the SO, deposition monitoring study is the large .-, data set t o check against 31 A model production. comparison of the UK dry 1 deposition model (Smith & Fowler, 1994) with the measurements reported here is presented in Fig. 13 for a dry wheat canopy in its vegetative phase. The agreement between measurements and the model Figure 12. Mean variationswith wind speed of deposition velocity for SO, over bare soil (January-May1994). Reference height is lm. above the zero-plane
18
--
0 M
00m mm um
V d WdaaaO
06.00
01.00
iom
12m 14:oo 16m 11:oo
2a.m Dim
carrr Figue 13. Dry deposition of SO, onto a wheat canopy (16/04/1993). Measurements are compared to model predictions using halfhourlyme&omb@~ and mean deposition velocities computed for a 'wet surface'day. Reference height is 1 m. above the zeroplane.
25 20
-ModdlcdVd
MtarmedVd
wetdke
----------
15 10 5 0
WOO 02:OO 04:OO 06:OO 08:OO 1O:OO 1290 1490 1600 18:OO 2000 2200
GMT 14. Dry deposition of S0,onto sugar beet (07/06/19!34). Measurements are compared to model predictions using half-hourly meteorobgical data and mean deposition velocities computed for a 'wet surface' day. Reference height is 1 m. above the zero-plane.
is excellent on average although some of the 30 minute measurements differ significantly from the model, largely as a result of non-stationarity effects. For wet surfaces, as shown earlier, the assumption of constant and small value for r, provides a poor estimate of SO, deposition onto the sugar beet canopy to wet conditions (Fig. 14). The data show consistently smaller rates of deposition when the canopy was wet. This, as described earlier, is a consequence of oversimplistic assumptions about the chemical behaviour of the S'" in the liquid film on the sugar beet canopy. The overall results of the model and measurement comparisons for the two canopies vegetation show that 1. Diurnal and seasonal cycles in V,, are simulated very well by the model for dry canopies of either sugar beet or wheat.
19
2. Wet canopies are not perfect sinks and to model the r, and V, for those more knowledge of the processes in liquid films is necessary. A fix for the model can be provided by setting r, for wet canopies at this site to 60 s m-'.
ACKNOWLEDGEMENTS The authors gratefully acknowledge the UK Department of the Environment for funding this study and Miss Fiona Greenwood of the University of Nottingham for assistance with the field measurements.
REFERENCES Brimblecombe, P. (1978). Dew as a sink for sulphur dioxide. Tellus 3 0 151-157. Chameides, W.L. (1987). Acid dew and the role of chemistry in the dry deposition of reactive gases to wetted surfaces. J. Geophys. Res. 92: 11,895-11,908. Erisman, J.W. (1993a). Monitoring the dry deposition of SO, in the Netherlands. Atmospheric Environment 27A pp 1153-1161. Erisman, J.W. (1984). Evaluation of a surface resistance parameterization of sulphur dioxide. Atmos. Environ. 28(16): pp 2583-2594. Erisman, J.W. and Van Pal, A. (1994). Parameterization of surface resistance for the quantification of atmospheric deposition of acidifying pollutants and ozone. Atmos. Environ. 28(16): pp 2595-2607. Fowler, D. and Unsworth, M.H. (1979). Turbulent transfer of sulphur dioxide to a wheat crop. Quart. J. Roy. Meteor. SOC. 105 767-783. Fowler, D. (1980). Removal of sulphur and nitrogen compounds from the atmosphere in rain and by dry deposition. Proceedings of the conference on ecological impact of acid precipitation, Norway 1980 (eds: Drablm, D. and Tollan, A.) pp 22-32, S.N.S.F. Publishers. Garland, J.A. (1977). The dry deposition of sulphur dioxide to land and water London. A364: 245-268. surfaces. Proc. Roy. SOC. Hicks, B.B., Baldocchi, D.D., Meyers, T.P., Hosker, R.D. and Matt, D.R. (1987). A preliminary multiple resistance routine for driving dry deposition velocities from measured quantities. Water, Air and Soil Pollution 36: 311-330. Monteith, J.L. and Unsworth, M.H. (1990). principles of Environmental Physics. Arnold, London. Payrissat, M. and Beilke, S. (1975). Laboratory measurements of the uptake of sulphur dioxide by different European soils. Atmos. Environ. 9: 211-217. RGAR (1990). Acidic deposition in the United Kingdom. The third report of the Review Group on acid deposition. UK Department of the Environment. Sandnes, H. (1993). Calculated budgets for airborne acidifying components in Europe, 1985, 1987, 1989, 1990, 1991 and 1992. Det Norske Meteorologiske Institutt. EMEP/MS C-W Report 1/93. Thom, A.S. (1975). Momentum, mass and heat exchange of plant canopies. In: Vegetation and the atmosphere Vol. 1principles. Ed. Monteith, J.L. pp 57-109. Academic press, London.
This Page Intentionally Left Blank
G.J. Heij and J.U! Erisman (Editors). Acid Rain Research: Do we have enough answers?
0 1995 Elsevier Science BV A11 rights reserved.
21
The use of the gradient method to monitor trace gas fluxes over forest: Flux-profile functions for ozone and heat. Jan Duyzer and Hilbrand Weststrate IMW TNO, P.O. Box 6011,2600 JA Delft, The Netherlands Abstract
This study aims to assess the relation between fluxes and profiles above forest for trace gases. To this purpose the flux of ozone to a Douglas fir forest was measured continuously by eddy correlation for seven months in 1993. During the same period vertical profiles of air temperature and ozone concentration were determined over the forest. In addition several turbulent parameters were recorded. From the observed temperature profiles and sensible heatfluxes flux-profile functions could be derived. Due to the scatter in the vertical profiles, flux-profile functions for ozone could not be derived with the same confidence. However, a significant difference with the observed flux-profile functions for heat could not be detected. Fluxes of ozone calculated according to the gradient method using the derived functions for heat showed very good agreement with the fluxes observed by the eddy correlation method. This result shows that the gradient method can be used with good results over forest when local flux-profile functions are used. 1. INTRODUCTION
To asses the achievement of environmental policy goals regarding the input of acidifying species by atmospheric deposition a monitoring system needs to be available. Such a monitoring system is now being developed for sulphur dioxide and ammonia in the Speulderbos, a Douglas fir stand in the centre of the Netherlands [ 11. The deposition will be monitored by the micro-meteorological gradient method. This indirect method is used because monitors fast enough to be used in the eddy correlation method are not available. The deposition flux is determined from measurements of the concentration of gases at several heights above the forest and a turbulent diffusion coefficient k2 Over low vegetation this diffusion coefficient can be determined from empirical flux-profile functions given in the literature. Because of the large roughness and the use of towers barely extending above the canopy the situation is more complex over forest. Measurements may take place in the so-called roughness layer where large deviations from classical flux-profile functions may be expected. To support calculations of the flux from the concentration gradient observed in the Speulderbos, local fluxprofile functions are required.
22 To determine the flux-profile functions over Speulderbos an automatic system capable of measuring the flux and the concentration gradient of ozone was operated continuously in a 36 m tower for nearly eight months in 1993. From these measurements flux-profile functions for heat and ozone were derived. In this paper the experimental set up is described and the results of the measurements are reported. 2. THEORY
The eddy correlation method is considered a reference method to measure the fluxes of trace gases to the surface. The average flux is equal to the covariance of the vertical component of the wind velocity ( w ) and the air concentration (c):
In order to measure the contribution of all eddies to the flux, fast response sensors are required. These are often not available for important trace gases such as sulphur dioxide and ammonia. Therefore gradient methods are often applied. Principle to the gradient method is the flux-gradient assumption. Similar to Ficks law the flux Fi of a component i is calculated from:
The deposition velocity vd is equal to: -Fi/ci. ci is the concentration at a reference height z=h-d. Where h is the height above ground and d is the so-called zero displacement height. k is von Karman’s constant (taken equal to 0.4), u* the so called friction velocity and @C(dL) is the dimensionless flux-profile relation. L is the Monin Obukhov length scale defined in [Z] When gas fluxes are measured the flux-profile functions are often taken equal to the fluxprofile function for heat @h. Empirical values for @ j are given in the literature [Z]. In principle the functions for trace gases could be different to those of heat although field experiments over grassland [3] showed reasonable agreement between turbulent exchange coefficients for heat and ozone. Over forest such measurements have not been reported. Especially the displacement height for gas could be different from the one for heat. Therefore the objective of this study was to determine @functions for a trace gas in Speulderbos. In an earlier study carried out in 1988 and 1989 in the framework of the acidification programme local flux-profile functions ( @ j ) were used together with a displacement height of 11.5 m. These functions were derived from measurements carried out at the site [4]. A height dependent correction factor a was used to correct the flux-profile functions @ for heat given by Dyer and Hicks [ 5 ] :
23
Since then the forest has grown by a few metres. As a consequence the values for the zero plane displacement height d will have changed.
a and
3. METHODS 3.1.
Descriptionof the site
The measurements were carried out in a roughly 30 year old Douglas fir stand [2]. The stand is homogeneous of an area of 2.5 ha. It is surrounded by oak and larch. The stem density is nearly 800 stems per hectare. The height of the trees was about 18 to 20 metre. The one-sided leaf area index varies over the years between 10 and 17 [6]. A diagram of the site is given in [2]. 3.2.
Instruments
The experimental set up is schematically presented in Figure 1. Central to the instrumentation is a sonic anemometer, a Kayo Denki DAT 310 with TR61 probe mounted at the 30 metre level on a boom extending some 3 metre from the tower on the South-West side. The fast response ozone monitor (7) is mounted on a smaller boom in a way that the air inlet was located 25 cm from the sonic centre. At 24, 26.5, 30 and 35 metre above the ground high accuracy temperature sensors (8) are mounted on smaller booms. Relative humidity of air is obtained from a Vaisala instrument with a capacitive sensor. Radiation instruments for net radiation (Schenk 8110), Global radiation (Li-Cor) and Skye SKP 215 sensor for Photosynthetically Active Radiation (PAR) are mounted on the 30 metre level. Air is drawn from the inlets at the same heights as the temperature sensors to a central manifold located near the instruments at the 28 m level. Using a computer controlled valve system the height from which air is drawn through the manifold can be selected. The ozone concentration is determined using a Bendix 8002 with ozone detection based upon the chemiluminescent reaction of ozone with ethylene. Using this set up the air concentration can be determined four times at each height within each cycle of 20 minutes. In [9] and [lo] several tests with the set-up are described. An experiment with all tubes sampling from the same height showed that systematic differences in the concentration observed with the different tubes were not detectable and less than 0.2%. Typically this leads to maximum errors in the deposition velocity of 0.35 m d s . Random fluctuations in V d however can be as high as a few W s .
24 3.7,rr
P I 3u rn 76 5
4
A.1
11'
4 I
1 '
Blm
n l
lq
--
-
._.-.
-.----
Re'erence fillerhead I;rau,rllt f ~ i i e rr k d r i
PT100
EMsOnielrmanrlw ptFA5 3fvrie rr,vniIvr
h-
Nel radla! mi
PAR Glcoal 'adlallcn Valvss system
a DATA
f.IUI1IiUIC
Rmp
AGJISITIOV
3.3.
Calculations of the flux-profile functions from data
For every 20 minute interval the flux-profile functions @ for heights 1 and 2 were calculated from the eddy correlation fluxes and gradients.
with 0, the potential temperature calculated from the observed air temperature T as 0, = T + ywith ythe dry adiabatic lapse rate. In order to use only good quality data measurements in the wind sector 3 15" to 45" were excluded because the measured turbulent parametres may be affected by the tower. Measurements during sunrise and sundown were also excluded. Only cases with a monotonously increasing or decreasing temperature gradient and with the absolute value of the heatflux larger 18 W/m2 were considered. The ozone flux observed using the eddy correlation method was corrected for spectral response according to [l I]. Using an experimentally determined response time for the ozone monitor of 0.1 sec the average correction was only 4% with maxima up to 8%. The flux-profile function for ozone was only calculated when the following criteria were met: the flux of 0 3 was larger than 0.1 ppb.m.s-1. the standard deviation in the 0 3 concentration measured at one height was smaller than 5%. the concentration of 0 3 is larger than 15 ppb. the error in V d due to instationarity in the 0 3 concentration is smaller than 5 r r d s [see 121.
25 The purpose of applying these additional criteria is to reduce the noise level in the calculated flux-profile functions. There is no indication in the results suggesting that the above criteria have a systematic influence on the results. 4. RESULTS AND DISCUSSION
All instruments were operated from December 1992 to September 1993. For several different reasons a large fraction of the data obtained in the wintertime appeared to be unreliable. For the analysis presented here only the data from the period April 1993 to September 1993 were used. Roughly 2500 successful twenty minute measurements of the eddy correlation flux were available. The gradient measurements yielded nearly 3500 successful runs. 4.1.
General
Figure 2 shows a typical result of the eddy correlation measurements on July 9 and 10, 1993. A diurnal cycle of the canopy resistance R,, calculated as in [4], is clearly detectable with values going down to around 70 s/m during the day and values of 500 s/m at night. The actual results are plotted without any smoothing in order to give an impression of the good quality of the data. This diurnal cycle is probably linked with uptake of 0 3 by stomata.
800
t I
0.00
I
I
0.00
I
I
I 0.00
Time (h)
Figure 2. The canopy resistance to uptake of 0 3 calculated from eddy correlation measurements at Speulderbos (July 9 and 10, 1993). The dependency of the canopy resistance on PAR is illustrated in Figure 3a. Figure 3b shows the dependency of the canopy resistance of the vapour pressure deficit. Especially the dependency of the vapour pressure deficit is quite strong, although it is important to realize
26 that several cross-correlations between the air temperature, PAR and vapour pressure deficit exist. Therefore, a more detailed interpretation is required. In an earlier study carried out in Speulderbos the effect of several parameters on the stomata1 resistance to water vapour, calculated from eddy correlation measurements, was investigated. A strong influence of vapour pressure deficit and radiation and hardly an effect of canopy temperature was also observed [ 191. -1 Rc03(s.m )
R,
o3 (s.m")
300
300
250
250-
200 150 100- fl
.' 500
rn
. .... . . .
150-
loo-,
-
50 . . ' I . . . , . . . 1 ' ' . , ' ' . , . . .
.
-
200
0
..
.m
m .
. . . . . . . . . . . . . . . . . . . . . . .
The average ozone flux was around 0.15 ppb m/s with a standard deviation of 0.12 ppb m/s. The average 0 3 concentration was equal to 35 ppb. The average deposition velocity was
7 mm/s. The median canopy resistance calculated from the measurements was 150 s/m. A more detailed statistical treatment of the date is given in [lo]. These results fit quite well with literature data. Greenhut [ 131 reports aircraft measurements over coniferous forest in southern New Jersey with an average canopy resistance of 50 to 400 s/m. Lenshow [14] reports a canopy resistance of 50 s/m. Using the eddy correlation method Wesely [15] found values varying between 150 and 400 s/m with values up to 1500 s/m at night. 4.2.
Flux-profile functions
A specific averaging procedure was used to calculate flux-profile functions. After all selection procedures around 500 twenty minute cases were left. The arithmetic average of
27 these 0 functions for ozone in a certain stability interval iYZ. shows too much scatter. The large scatter is caused by the magnitude of the concentration gradient as well as the magnitude of the fluxes. Especially at (near) neutral conditions (windy, overcast) when the ozone concentration is low the deposition flux becomes very small. Only between 24 and 35 metre the difference can be observed well above the detection limit. In order to reduce the large effect of outliers a robust statistical treatment is required. Therefore several averaging procedures were compared [lo]. The larger temperature data set showed less scatter and was used as a database to test these procedures. Best results were obtained using a robust statistical method proposed by the Analytical Methods Committee [17]. With this method the average is calculated on the basis of the 50 and 75 percentile values thereby minimising the influence of outliers. For heat the difference in the results obtained using the various procedures appears to be small. For 0 3 only the robust methods gave useful results. Based on the similarity in transport mechanisms it was assumed that the procedures used for heat could be applied for 0 3 with confidence as well. Figure 4 shows the flux-profile functions calculated for heat for the heights 30 to 35 m. A displacement height of 15 m was assumed as 75% of the height of the trees [16]. It is important to realise that the zero displacement height was now chosen to be 15 metre rather than the 11.5 metre chosen earlier. This difference is related to the growth of the forest over this period the forest which was between 0.6 to 0.9 m per year and therefore some 3 m over these years [6]. It appears that the functions can be described quite well with small corrections to the existing flux-profile functions. The values of the @h function calculated using classical equations (3) and (4)with a equal to 0.9 are also plotted. It was noted that a slightly better comparison between the observations and the functions could be obtained when the coefficient@ in the flux-profile functions was taken to be 7 rather than 5. Similar results are also reported by Bush [2]. Table 1 shows the data for the other height intervals as well. The correction factors compare quite well with the results found in the earlier study [4]. As was observed earlier the deviation from the original functions increases when the canopy is approached. Table 1 Values of a to correct flux-profile functions as in equations (3) and (4)for heat for different . a displacement hei ht of 15 m the effective height t e f f height intervals hl and h ~Assuming for which the
d G g
Interval h?-h f
Zeff
Q
24 -26.5 26.5 - 30 30 -35 24 -35
10.2 13.1 17.3 13.4
0.75 0.8 0.9 0.75
28 @h
- a =0.9 observed (30-35)
'7 1
4 ' 0.25
Figure 4. Flux-profile functions @h observed over Speulderbos. The lines indicate the classical functions according to equations ( 3 ) and (4) using a value of 0.9 for the correction factor a.
2.5
-a=1 2- __._... a=0.8 - cC=O.6
'~5-
observed (30-35)
1-
0.5.... ............,_......
0 -0.75
-0.5
-0.25
0
0.25
z/L
Figure 5. Flux-profile functions Qc over Speulderbos. The drawn line indicates classical functions according to equations ( 3 ) and (4) using indicated values for a.
The results for ozone for the 30-35 m interval are displayed in Figure 5. The results for this height interval compare reasonably well with the results for this interval for heat. The uncertainty due to scatter however is much larger. For the other height intervals the scatter is even worse. For the interval 24-35 metre the comparison between heat and ozone is quite good. For both 0 3 and heat an a factor of 0.75 gives a reasonable fit to the data. It is therefore concluded that there is no significant difference between the functions for heat and those for ozone. In order to test this assumption further the ozone flux was calculated using the gradient method and compared with fluxes measured by the eddy correlation method. To calculate the gradient flux the flux-profile functions derived for heat (given in Table 1) were used for each of the three height intervals. The results of this comparison are presented in Figure 6 where the average of the three height intervals is compared with the eddy correlation flux. The observed fluxes compare quite well over a range of a factor of ten. This gives confidence in the flux-profile functions chosen for the analysis. The flux calculated according to the so called modified Bowen ratio method was also compared with the eddy correlation flux. The fluxes calculated for several height intervals also compared very well with the eddy correlation flux [ 101.
29 Flux (ppb.m.s-l)
=
gradient flux 1:l
-0.2
-0.5 y -0.5
I
-0.4
I
-0.3
I
I
-0.2
-0.1
EC-flux (ppb.m.s-’)
Figure 6. The flux of 0 3 to the Speulderbos calculated using the gradient method against the flux observed using the eddy correlation flux. 6. CONCLUDING REMARKS
The Speulderbos area is relatively inhomogeneous. To test the validity of the constant flux assumption over the forest several additional measurements were carried out in a three week campaign in June 1993. The ozone flux was measured by eddy correlation at the 25 m and at 35 m level. No significant differences could be detected between the fluxes of momentum, heat and ozone observed at two levels [18]. During this campaign the ozone flux was recorded simultaneously at a second tower located some 50 m away from the one described here. The fluxes at both towers did not deviate either. Both observations endorse the assumption that a fairly homogeneous flux field exists over Speulderbos. In this experiment the effect of chemical reactions of ozone with nitrogen oxides was also studied. Some of the observed uptake of ozone may be caused by chemical reactions in air between ozone and nitric oxide emitted from the forest floor. This effect will be investigated in more detail. The experiments shown here form evidence that the gradient (and the Bowen ratio method) can be used over forest for ozone. It seems reasonable to assume that this conclusion is also valid for other trace gases with sinks in the canopy such as sulphur dioxide or ammonia. This conclusion however may not be valid for gases with sources and sinks at the forest floor such as nitrogen oxides. The flux profile function found in this study can only be applied in Speulderbos. Generalisation of the flux profile functions observed in this study in more general terms such as tree height and density will be attempted in the near future.
30 Acknowledgements - This work was funded by the Dutch Ministry of Housing Physical Planning and the Environment, the National Institute of Public Health and Environmetal Protection and the Commission of the European Communities. 7. REFERENCES
1
2 3 4 5 6 7 8 9 10 11 12
13 14 15 16 17 18
19
Erisman J.W., M.G.Mennen, J.E.M.Hogenkamp, E.Kemkers, D.Goedhart, W.A.J. van Pul, G.M.F. Boermans, J.H. Duyzer, G.P. Wyers, 1993, Report nr. 722108002, RIVM, Bilthoven, The Netherlands. Bush N.E. (1973) In: Workshop on micrometeorology, D. A. Haugen Editor American Meteorological Society, Boston, 1-65. Droppo (1985) J. of Geophys. Res., 90, D1,2111-2118 Duyzer, J.H., H.L.M. Verhagen, J.H. Weststrate, F.C. Bosveld (1992) Env. Poll., 75, 313. Dyer A.J., Hicks B.B., ( 1970), Quart. J. of the Royal Met. SOC.9 6 ,7 15. Jans, W.W.P., G.M. van Roekel, W.H. van Orden, E.G. Steingrover (1994) Institute for Forestry and Nature Research, IBN Research Report 94/1, Wageningen, The Netherlands Gusten H.,G Heinrich, R.W.H. Schmidt, U. Schurath, 1992, J. of Atmos. Chem. 14.7384 Slob, W.H. (1978), Scientific report W.R. 78-1, KNMI, De Bilt, The Netherlands. Duyzer (1994), IMW-TNO report R94/060, Delft, The Netherlands Weststrate en Duyzer (1994) IMW-TNO report R94/104, Delft, The Netherlands Moore C.J. (1986) Bound. Layer Met. 37, 17-35. Fowler, D., J.H. Duyzer (1989) In: Exchange of trace gases between terrestrial ecosystems and the atmosphere. M.O. Andreae, S.D. Shimel, (eds.), John Wiley and Sons Ltd. D., 189-207 Greenhut G.K. (1983), Bound. Layer Met. 27,387-391. Lenschow D.H., Pearson R., Stankov B.B., (1982) J. of Geophys. Res. 87, c l l , 88338837. Wesely M.L., (1983), in: Trace atmospheric constituents, properties, transformations & fates., Schwartz S. editor, John Wiley and Sons Inc., 345-370 Jarvis, P.G., G.B. James, J.J. Landsberg (1976). In: Vegetation and the Atmosphere Volume 2, Monteith Editor Academic Press London, 171-240 Analytical Methods Committee (1989) Part 1. Basic Concepts Part 2. Inter-laboratory Trials, Analyst. December 1989 vol. 114, 1693-1702 Choularton, T.W., H. Coe, S. Walton, M.W. Gallagher, K.M. Beswick, C. Dore, J. Duyzer, H. Weststrate, K. Pilegaard, N.O. Jensen, P. Hummelshoj (1994) Proceedings of Eurotrac Symposium '94, P. Borrel, editor SPB Academic Publishing, the Hague, The Netherlands Bosveld, F., Bouten, W., Noppert, F., E. Steingrover (1991). KNMI report FM-91-02, De Bilt, The Netherlands
G.J.Heij and J. W Erisman (Editors). Acid Rain Research: Do we have enough answers? 0 1995 Elsevier Science BY All rights reserved.
31
Deposition of nitrogen oxides and ozone to Danish forest sites K. Pilegaard", N.O. Jensenband P. Hummelshaj* "Environmental Science and Technology Department, Risa National Laboratory, DK-4000 Roskilde, Denmark *Meteorologyand Wind Energy Department, Risa National Laboratory, DK-4000 Roskilde, Denmark
Abstract Preliminary results of eddy correlation measurements of fluxes of NOz, and 0 3 made over a coniferous and a deciduous forest site in Denmark are presented. The total resistance to deposition are calculated and subdivided into aerodynamic, viscous sub-layer and surface resistance for investigation of the influence of meteorological factors. The viscous sub-layer resistance is derived by a new theory, taking the bluff roughness elements of the forest and the dimension of the needleslleaves as well as the LAI into account. The fluxes of nitrogen dioxide and ozone are related to the fluxes of water vapour and carbon dioxide. The results from the coniferous forest site (Norway spruce) show a diurnal variation in the deposition velocities and surface resistances during the growth period, which is consistent with a stomatal uptake of the gases. However, a substantial deposition is also found at night and in winter indicating a significant role of atmospheric chemistry and surface reactions. The experiment a t the deciduous forest site (beech) shows the difference in deposition to the site before and after bud burst, thus describing the influence of the stomatal activity of the leaves on the uptake of gases in the forest ecosystem.
1. INTRODUCTION The dry deposition of pollutants to forest ecosystems plays an important role in the cycling of as well nutrients (e.g. nitrogen) as harmful substances (e.g. ozone). Estimation of the flux and examination of the governing processes are important in order to model the deposition and in order to evaluate critical loads. This paper presents preliminary results of two recent field experiments in Danish forests (one in a coniferous forest and one in a deciduous forest before and after bud burst). The measurements of concentrations and fluxes were used to calculate deposition velocities and subsequently by means of measured meteorological parameters to split the total resistance to deposition in aerodynamic, viscous sub-layer and surface resistances. The role of stomatal uptake was evaluated.
32 2. MATERIALS AND METHODS
2.1 Experimental sites and instrumentation
The instrumentation for the eddy correlation measurements consisted of a 3D Gill Sonic Anemometer (wind fluctuations), a GFAS OS-G-2 Ozone Sonde, a Scintrex LOZ-3 O3 analyzer (chemiluminescence with Eosin Y) ,a Scintrex LMA-3 NO2 analyzer (chemiluminescence with luminol) and an Advanet E009A infrared COZ and H 2 0 fluctuation meter. A large number of meteorological measurements were made simultaneously, the most important being wind speed (cup anemometers in different heights), wind direction (wind vane), temperature in different heights, and water vapour flux (Ophir Infrared Hygrometer IR-2000). Flux measurements were made in Ulborg (a forest in a remote rural area of western Jutland) during the period 7-17 June, 1994. The site is described in Andersen et al. [ 13. The measurements were carried out from a 36 m tall mast placed in a Norway spruce (Piceaabies) plantation with trees of a height of approximately 12 m and a good fetch in most directions except from a small sector towards SW. The flux measurements were made a t 21 m. The flux measurements in the deciduous forest were made in Corselitze Forest on the island of Falster in southeast Denmark during April and May 1994. The measurements were made from a 57 m tall mast in a stand of 24 m tall beech (Fagus sylvatica) trees. The instruments for eddy correlation measurements were placed a t 41 m. There was a good (500 m) fetch in the sector from ENE to SSE. There were no local sources of pollution in this direction since the forest is bordering the Baltic Sea. The incoming air-masses thus carry pollutants from distant sources in Poland and eastern Germany. 2.2 Calculations
Fluxes were measured by the eddy correlation method and calculated from the equation:
where F, is the flux of the compound in question, w the vertical wind velocity and C ( z ) the concentration of the compound a t the measurement height (z). The prime indicates instantaneous deviation from the mean and the over-bar indicates the time average (0.5 h). The measured fluxes were corrected for errors due to changes in atmospheric density caused by heat and water vapour flux [21. Deposition velocities (&) were calculated from the equation:
Resistances were calculated according to the model: 1
- = rt = ra i-rb -t r,
vd
(3)
33
where rt is the total resistance, r, the aerodynamic resistance, rb the viscous sub-layer resistance and r, the surface or canopy resistance. r , and rb can be calculated from meteorological observations; r, is then calculated from equation 3 as the residual resistance. The aerodynamic resistance (r,) was calculated by the expression analogous to equation 2: 21
r, = -
4
(4)
where i~ is the mean wind speed (m s-l) and u, the friction velocity, calculated -1 as -u/IL+*. For forests we have developed the following parameterization of rb 131
m2 s-'), D the diffusion where u is the kinematic viscosity of air (=15 . coefficient of gas (for O3 and NO, = 14 . m2 s - l ) , P the leaf dimension , LAIe is a modified leaf area index ( L A I )representing the effective surface area and c is a correction factor, determined empirically from the temperature profile and aerodynamic roughness. The parameter c is set to 3 for both spruce forest and deciduous forest. The leaf dimension (0 is set to 0.005 m for the spruce forest and 0.01 m for the deciduous forest. What the exact meaning of LAI, is, depends on the actual chemical compound under consideration. Thus for COz it is strictly the green leaves that counts (ignoring respiration from the forest floor) and in this case LAI, = LAI. For a deciduous forest LA1 is zero after fall and until the leaves come out again, and hence rb is infinitely large (see equation 5 ) . For O3 on the other hand, the flux also goes to other parts of the canopy (especially to small branches and twigs due to the rb dependence on PI and to the forest floor. For this compound the minimum LAI, (on a bare smooth surface) will be 1. In a deciduous forest without leaves it will have a larger value, perhaps 5. When the leaves come out it will increase with about 5, which is the LAI efficient for CO,. The surface resistance to water vapour was calculated according to equations 2 and 3 after subtraction of the saturated water vapour concentration a t the surface temperature. The canopy stomata1 resistance was calculated by conversion of the resistance to water vapour to resistance to carbon dioxide according to the relation:
where D, is the molecular diffusivity of COZand D,,, the molecular diffusivity of water vapour.
34
3. RESULTS AND DISCUSSION
3.1 Coniferous forest
The diurnal variation of the concentration, flux and deposition velocity of 0 3 in the period 7-17 June, 1994 is shown in figure 1. The concentrations show a diurnal pattern with a minimum in the early morning and a maximum in the late afternoon. The difference between the minimum and maximum is about 15 ppb, indicating that the air is only moderately polluted. The dominating wind-direction during the experiment was Nw; this wind brings relatively clean air from the North Sea. The fluxes of ozone ranged from about -0.25 pg m-2 s-' to -0.6 pg m-2 s-l with a maximum around noon. The deposition velocities ranged from 3.5 mm s-* to 7 mm s-l with a sharp rise a t dawn and a maximum already in the morning. This pattern indicates that stomatal processes are important for the deposition of 0 3 . However, the rather high deposition velocities during the night indicate that other processes for removal of O3 also play a significant role. The results of the calculation of resistances are shown in figure 2 and table 1, where data from earlier experiments in the spruce forest 141 are included for comparison. The resistances show a diurnal pattern with the lowest values during daytime. The surface resistance is highly dominating during daytime and the influence of meteorology low. During nighttime the surface resistance to O3 was substantially lower than the canopy stomatal resistance, which again shows that stomatal uptake is not the only sink for 03.Table 1shows that low surface resistances to 0 3 also can be found during winter time, when the activity of the trees is low.
June 94 day night 8 31 ra 9 14 rb 151 288 Tc(03) r,(NO?) 347 1023
Norway spruce July 92 day night 6 68 9 20 368 800
-
February 93 day night 16 16 208 1153
27 17 179 428
Beech April 94 May 94 day night day night 17 69 7 11 21 7 8 14 330
692
269
602
Table 1: Average day (06:OO - 18:OO) and night (18:OO - 06:OO) values of resistances ( s m-* from Ulborg (Norway spruce) and Corselitze (beech).
3.2 Deciduous forest
Two periods are selected for presentation from the data collected in Corselitze forest. The first period (April 22-25) represents the forest without leaves on the beech trees, and the second period (May 11-14) represents a time when the leaves were fully unfolded. Results for the fluxes of 03,NOz and COz are given in figure
35 Ozone concentration
00.00
06:OO
12:oo
18.00
24:OO
18:oo
24:OO
hour
Ozone flux
0o:oo
moo
12:oo hour
Ozone deposition velocity
0o:oo
06:OO
12:oo
1e:oo
24:OO
hour
Figure 1: Ozone concentration, flux and deposition velocity over a Norway spruce stand at Ulborg Plantation 7-17 June, 1994. The plots show means f standard error and a smoothed line.
36
B! 0
0
(0 0
m
0 - L0 O
'E
* 0 0 0 N 0 0
0 0O:OO
0O:OO
12:OO
24:OO
0O:OO
12:OO
24:OO
hour
hour
rb
Canopy stornatal resistance
1200
hour
24:OO
0O:OO
1200
24:OO
hour
Figure 2: Diurnal variation of aerodynamic resistance (ra>,viscous sub-layer resistance ( r J , surface resistance to O3 (T,) and canopy stornatal resistance over a Norway spruce stand at Ulborg Plantation 7-17 June, 1994. The plots show means f standard error and a smoothed line.
37
Ozone flux April 22-25, 1994
Ozone flux May 11 - 14, 1994 o. o
o (5 'm
'on
#t i .........
::L
00:00
06:00
12:00
18:00
::L
c?
24:00
00:00
06:00
12:00
18:00
hour
hour
Nitrogen dioxide flux April 22-25, 1994
Nitrogen dioxide flux May 11 - 14, 1994
24:00
t m.
=L o. o
"
-:~
--
,
,
;
,
f
00:00
06:00
12:00
18:00
24:00
00:00
06:00
12:00
18:00
hour
hour
Carbon dioxide flux April 22-25, 1994
Carbon dioxide flux May 11-14, 1994
.......
O4 (5
o
~ = oo.
~'m ~
E o,"
E e
.
.
.
24:00
.
"t. 9 w
;
|
!
!
'l'
|
!
i
i
00:00
06:00
12:00
18:00
24:00
00:00
06:00
12:00
18:00
24:00
hour
hour
Figure 3: Fluxes of 03, NO2 and CO2 over a beech stand in Corselitze forest before and after bud burst. The plots show means -}- s t a n d a r d error and a smoothed line.
38 3. The m e a n concentration of 03 was 48 ppb in the first period and 41 ppb in the second; for NO2 the values were 10 ppb and 5 ppb, respectively. The concentrations of CO2 were quite similar in the two periods. These differences should be noted when comparing the fluxes. The flux of 03 was almost always downwards, whereas the flux of NO2 was mostly upwards. Both compounds showed diurnal patterns with the largest fluxes during daytime. There were no significant differences in the deposition velocities between the two periods for these two compounds. The flux of CO2 was much larger in the second period and showed a clear diurnal pattern with downward flux during daytime and upward flux during nighttime. In the first period the diurnal pattern was less pronounced although the downward flux peaked in the late morning. The m e a n daytime deposition velocity of O3 was around 4 m m s -1 in both periods and the m e a n nighttime deposition velocity was 1 mm s -1 in April and 2 m m s -~ in May. A m a x i m u m of around 7 mm s -1 was found in both periods. The surface resistance to 03 show a diurnal pattern in both periods, although it is more pronounced in the second period (figure 4 and table 1). The daytime m i n i m u m lies around 200 s m -~. The effect of the beech leaves is seen most clearly in the fluxes of CO2, whereas the effects on the fluxes of 03 and NO2 are much less than expected. The diurnal pattern of the surface resistance to 03 indicates an influence of stomata, especially in the second period. The reason for the absence of a strong difference in deposition of 03 before and after bud burst might be that O3 is taken up by vegetation on the forest floor (mainly A n e m o n e n e m o r o s a in the April period) and removed by other means t h a n stomatal uptake, such as destruction on leaf surfaces, deposition to soil ([5]) and reaction with NO. The upward flux of NO2 is most likely explained by the reaction of O3 with NO, emitted as a result of bacterial activity in the forest soil ([6]). The a m o u n t of 03 removed in this way is, however, only a small part of the total 03 flux. There seems to be a larger NO2 flux in the first period than in the second. This might be either a result of lower bacterial activity in the second period due to the shadowing effect of the beech leaves, which leads to lower temperatures, or the result of u p t a k e in the canopy. 4. C O N C L U S I O N S The fluxes of 03 over both a spruce forest and a beech forest exhibited a diurnal pattern. When the resistances of the atmospheric boundary layer and the viscous sub-layer were subtracted from the total resistance to deposition, the remaining canopy resistance still showed a diurnal pattern, which indicates t h a t stomatal activity is important for 03. However, a substantial flux was found during nighttime and other sinks for 03 like destruction at surfaces, uptake in the soil and reaction with NO m u s t be taken into consideration. Measurements over a beech forest before and after bud burst showed t h a t the influence of the beech leaves on the O3 flux was small, probably due to u p t a k e by
39
rc(O3), April
rc(O3), May
o
O
9
]
o
O
-
!
l
l
I
I
i
!
i
i
i
00:00
06:00
12:00
18:00
24:00
00:00
06:00
12:00
18:00
24:00
hour
hour
Figure 4" Surface resistance to 03, in a beech stand in Corselitze forest before and after bud burst. The plots show means + standard error and a smoothed line. the forest floor vegetation and destruction at other surfaces. The upward flux of NO2 in the beech forest is attributed to emission of bacterial NO from the soil and rapid chemical reaction with O3. 5. A C K N O W L E D G E M E N T S
The field experiments were supported by the Danish Strategic Environmental Research Programme and the European Union (contract no. EV5V-CT92-0060). The National Forest and Nature Agency funded the permanent installations for meteorological observations in Ulborg for use in the "Ion balance project". The instruments for flux measurements were funded by the Danish Science Research Council and Ris~s Integrated Environmental Project (RIMI).
40
References [1] H.V. Andersen, M.F. Hovmand, P. Hummelshoj, and N.O. Jensen. Measurements of the NH3 flux to a spruce stand in Denmark. Atmospheric Environment, 27A(2):189-202, 1993. [2] E. K. Webb, G. I. Pearman, and R. Leuning. Correction of flux measurements for density effects due to heat and water vapour transfer. Quarterly Journal of the Royal Meteorological Society, 106:85-100, 1980. [3] N.O. Jensen and P. Hummelsh0j. Derivation of canopy resistance for water vapour fluxes over a spruce forest, using a new technique for the viscous sublayer resistance. Agricultural and Forest Meteorology, (in press), 1994. [4] K. Pilegaard, P. Hummelsh0j, and N.O. Jensen. Deposition of ozone and nitrogen dioxide to open land and forest. In J. Slanina, G. Angeletti, and S. Beilke, editors,
General assessment of biogenic emissions and deposition of nitrogen compounds, sulphur compounds and oxidants in Europe. Air Pollution Research Report 47, pages 157-164. Commission of the European Communities, 1993. [5] W.J. Massman. Partitioning ozone fluxes to sparse grass and soil and the inferred resistances to dry deposition. Atmospheric Environment, 27A(2):167174, 1993. [6] W.A. Kaplan, S.C. Wofsy, M. Keller, and J.M. Da Costa. Emission of NO and deposition of 03 in a tropical forest system. Journal of Geophysical Research, 93(D2): 1389-1395, 1988.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
Monitoring dry deposition fluxes of errors
SO 2
41
and N O 2" Analysis of
M.G. Mennen, J.E.M. Hogenkamp, H.LM.A. Zwart and J.W. Erisman Air Research Laboratory, National Institute of Public Health and Environmental Protection, P.O. Box 1, 3720 BA Bilthoven, The Netherlands
Abstract Dry deposition fluxes of acidifying trace gases are continuously determined at two locations in the Netherlands, a low vegetation site and a forest. In this work the errors in determined fluxes and deposition velocities are analysed in detail. The analysis is based on the results of both laboratory and field experiments. The average random error in SO2 deposition fluxes is 50% for 30-rain measuring cycles. Systematic errors are generally limited to 20% (underestimation) and these can be partly reduced by correction procedures. NO2 fluxes are systematically in error, varying from -50 to +100%, mainly due to interference by 03 and H20 and meteorological errors. The random error in the NO2 fluxes is large (25% to 1000% for 30-rain measuring cycles). Despite the large errors, NO2 deposition velocities determined at a test site showed approximately the same diurnal course as values obtained from a parameterization, which is based on the assumption that the surface exchange of NO2 is mainly determined by stomatal behaviour.
1. INTRODUCTION Dry deposition of trace gases and aerosols, particularly sulphur and nitrogen components, leads to soil acidification and eutrophication [1,2]. Despite many studies on dry deposition (e.g. [3,4]) questions still remain with regard to the underlying processes and effects. Systematic and continuous measurement (monitoring) of dry deposition fluxes of acidifying compounds is required to gain more insight and to validate and improve models used to calculate land use specific deposition parameters. Further, trends may be observed and effects of emission reduction programmes can be evaluated. In the Netherlands, two automatic systems are operated to monitor dry deposition fluxes of SO2, NI-I3 and NO2, one at a low vegetation site and the other at a forest location. The systems are based on the gradient method. They produce a restricted amount of continuous data and need little maintenance and calibration. The two systems are similar, but not identical. Differences are given in section 2. Detailed descriptions are given in [5,6]. Dry deposition measurements are liable to several errors, which are important to be known in order to derive sufficiently accurate fluxes from measurements. Businger gives
42 a good survey of the different error sources and some correction procedures [7]. Error sources may be instrumental errors (e.g. drift, noise, sensitivity to interfering compounds), errors caused by equipment and system configuration (e.g. adsorption effects in sampling tubes, upward wind distortion) and methodical errors (occurring when requirements with respect to fetch and measuring heights, steady state conditions, chemical reactions, etc., are not met). The latter are difficult to determine. Therefore, only data selected according to criteria based on these requirements are used [4,7,8]. However, elimination of data may also lead to errors in long term average dry deposition fluxes, since the selected periods may not be representative. Erisman et al. have developed a procedure to derive general dry deposition parameters from a limited set of selected data [3,8]. He has also estimated the errors caused by this procedure. In this paper we only deal with errors caused by instruments, equipment and system configuration. We have estimated these errors for our systems, based on results of laboratory tests and field experiments. In the laboratory, the specifications of the trace gas monitors used in the systems were determined. Field measurements were conducted during a test period (December 1991 until August 1992) at the heathland Elspeetsche Veld with the two systems operated side by side. The performance and reliability of both systems were tested. SO2 fluxes were mutually compared (note: NO2 fluxes were determined with only one system). Effects of flow distortion by obstacles on wind velocity and momentum flux were studied. Further, eddy correlation measurements of latent heat and NO2 fluxes were carried out during a few days. These will not be discussed in this work. The analysis given in this paper is focused on the Elpseetsche Veld experiments and on SO2 and NO2. Analogous procedures can be applied for other gases or other sites.
2. THE MONITORING SYSTEMS
2.1 Instruments and equipment Fig. 1 shows a schematic of one of the systems, when operated during the test period at the heathland Elspeetsche Veld. (Later this system was installed at Speulder forest, where it is still operated nowadays). The low vegetation system is similar, but not identical. Differences are given at the end of this section. A sonic anemometer (Kayo Denki DAT-310, Japan) is used to measure horizontal and vertical wind velocity, wind direction, friction velocity and heat flux. Its probe is mounted on the top of a slim mast (I) at 5.5 m height. The probe is provided with an inclinometer to measure the deviation of its position to horizontal alignment, and with a rotor to turn its open side towards the wind direction every hour. A net radiation meter (Thies 8110, Germany) and a temperature/relative humidity sensor (Vaisala HMP121B, Finland) are fixed on a small mast (IH) at a height of 1.5 and 2 m, respectively. Air is sampled at a flow rate of 2 1 min 1 through five 7 m long, isolated FEP Teflon sampling tubes (4 mm i.d.) mounted on 50 cm long outriggers fixed to a slim telescopic mast (H) at a height of 4 m (reference tube) and 4, 2, 1 and 0.5 m (other tubes). The tubes are connected to a valve system in box B, with which the gas monitors are supplied either with zero air (ambient air led over an active charcoal filter and a dust filter) for regular zero calibration or with ambient air from the inlets. All tubes are continuously
43 sucked through with a bypass pump, thus maintaining equilibrium inside. Box A houses two SO2 and two NO2 monitors. The SO2 monitors are Model 43S Pulsed Fluorescence analysers (Thermo Environmental Instruments Inc., USA) and the NO2 monitors are LMA-3 Luminox analysers (Scintrex Ltd., Canada). For both gases one monitor (called the reference monitor) continuously measures the concentration at the reference level (4 m). In this way changes of the gas concentration during a measuring cycle can be measured and corrected for. The other monitor (called the scanning monitor) samples consecutively at the various heights to measure the gas concentration gradient. The applied measuring scheme is given in section 2.2. At Elspeetsche Veld the monitors were calibrated at least every two weeks using a Model V3M portable calibrator (Environnement, France) provided with one permeation device for SO2 and one for NO2. Zero calibration is performed by the system, as described above. The luminol for the NO2 monitors was refreshed at least every 2 weeks. Box C houses the computer, used to control the monitoring system, to collect and store data and to perform necessary calculations The low vegetation system resembles the forest system except that NO2 is not measured and that a wind vane and a three-cup anemometer are used instead of a sonic anemometer. These are mounted at 5 m height on the mast carrying the sampling tubes (mast I is not used). Friction velocity and sensible heat flux are derived from wind velocity, standard deviation of the wind direction,
sonic anemometer ~ reference 4m rel. humidity temperature IJ 2rn
net
lm
ra~ati'cln
B m
C
A I
Figure 1. Schematic of the monitoring system at Elspeetsche Veld.
III
44 temperature and net radiation using a parameterization [8,9]. Vertical wind velocity is not measured. Further, the low vegetation system contains a surface wetness sensor mounted in the canopy. During the tests at Elspeetsche Veld masts and boxes were placed in such a way that the ideal fetch was not significantly disturbed and errors introduced by possible inhomogeneities were minimal.
2.2 Measuring scheme The scheme described below was applied during the Elspeetsche Veld experiments. Measurements are performed in consecutive hourly cycles, consisting of four periods each. During the first period (6 min) zero air is supplied to the monitors in order to determine the zero signal and minimize errors due to zero drift. During the following two periods, each taking 4x6=24 min, gradients of SO2 and NO2 are measured. In both periods the scanning monitor measures the concentrations at four levels successively (4, 2, 1 and 0.5 m), whereas the reference monitor measures simultaneously concentrations at the reference level (4 m). In all measurements the monitors are flushed for 5 min with the supplied air, whereafter the signal is sampled for 1 min. During the fourth period (6 min) data are written to disk, the prevailing wind direction is determined and the rotor carrying the sonic anemometer probe is set. Meteorological parameters are continuously measured during the first three periods (= 54 min) of the cycle; 6-min average values of horizontal and vertical wind velocity, wind direction, friction velocity, heat flux, net radiation, temperature and relative humidity are calculated from the raw data and stored.
3. ERROR ESTIMATIONS The equations used to calculate dry deposition fluxes from measured data are derived from the work of Hicks et al. [10] and are given in [6,8,11]. In this work we only give the essential equations. For simplicity we use two measuring heights (the highest and the lowest) instead of four assuming that gradients are linear. In practice, this assumption must always be checked by using more than two heights, because e.g. the lowest measuring height might be chosen just inside the roughness layer. For NO2 gradients are often non-linear due to reactions involving NO and 03. To first order this does not harm the results of our estimations. The dry deposition flux F (in ppb m sl) is calculated from the concentration gradient 8c/8Z according to -u
F=
ac Ku,Ac J;-~ = - A---'-"-Z-
(I)
where u, is the friction velocity (in m sl), c the gas concentration at 4 m (in ppb) and the von Karman constant (= 0.41). Because we use two measuring heights, 8c/8Z can be approximated by Ac/AZ, in which Ac is the difference in concentration between 4 and 0.5 m (in ppb), and AZ is given by
45
A Z = in( zl-d zl-d d z2_d ) -w ( L ) +W (z2--~)
(2)
where z~ is the measuring height, z0 the roughness length, d the displacement height and L the Monin-Obukhov length (all in m). The dimensionless stability functions V are given in [8]. For neutral conditions V = 0. Since errors in u, and AT. are usually limited to 10 to 20%, as we will show in section 3.3, and Ac is small, the error in F is dominated by the error in Ac. From eq.(1) and F
(3)
= -VdC
where vd is the deposition velocity at 4 m (in m s"), we arrive at Ac
(4)
,,va-'Z^
C
K/Z,
With this equation and suitable values for d, z0, vd, u, and c, values of Ac can be calculated. For Elspeetsche Veld the displacement height is 10 cm and the roughness length 4 cm (for wind directions, where the fetch is adequate). Typical values of u. and vd, obtained from the measurements during the test period, are given in Table 1. Five classes of stability conditions were defined based on the value of (z-d)/L. Their relative occurrence is also given in Table 1. Further, this table shows the calculated values of Ac for the 50-percentiles of c (1.2 ppb for SO2 and 10 ppb for NOz). We used these values to estimate the errors resulting from the behaviour of the gas monitors.
3.1 Gas monitors Drift, linearity, noise, temperature dependence, sensitivity to interfering components, etc., of the SO2 and NO2 monitors were determined in a calibration chamber according to a standard Performance Characteristics procedure [ 12]. For the most important error
Table 1 Values of relevant parameters used for error estimations Parameter
very unstable
unstable
neutral
stable
very stable
<-0.5
-0.5 - -0.1
-0.1 - 0.1
0.1 - 0.5
>0.5
5
11
66
14
4
u. (m s1)
0.15
0.25
0.4
0.15
0.1
AZ (-)
0.94
1.42
2.28
3.68
6.94
Va SO2 (cm s"1)
0.1-0.7
0.2-1
0.3-3
0.15-0.7
0.05-0.3
Vd NO2 (cm s"1)
0.02-0.3
0.02-0.3
0.01-0.5
0.005-0.1
0.002-0.05
Ac SO2 (ppt)
20-130
30-170
50-500
100-500
100-600
Ac NO2 (ppt)
30-400
30-400
20-700
30-600
40-800
(z-d)/L rel. occurrence (%)
46 sources the results and the corresponding errors in Ac are briefly discussed below. Detailed results and other error sources are given in [6]. Table 2 (see section 4) gives a summary. We note that in all calculations presented below the concentration of a gas is assumed to be constant during one measuring cycle. Periods with rapidly changing concentration are eliminated because of the steady state condition requirement [4,7,8].
Temperature dependence When sampled air enters the monitors, its temperature quickly takes the temperature inside the box housing the monitors. This implies, that significant errors might be expected, when the temperature inside the box suddenly increases or decreases, which happens sometimes as a result of rapid change in sun radiation at sunset or sunrise. When the temperature dependence of both the scanning and the reference monitor is exactly the same, the error in Ac is 0. Unfortunately, from the experiments in the calibration chamber the dependence of the four SO2 monitors was found to vary from 0.03 to 0.12% K 1 in the range of 5 to 40~ The error in Ac per K temperature change can be estimated from errAc
=
C (St,S--~ r,r) (I-~-~)
(5)
where 5r~ and 5r~ are the temperature dependence of the scanning and the reference monitor, respectively. The temperature change within one measuring cycle is usually limited to 3 K. Taking the most extreme values for 5r~ and 5r~ found in the laboratory experiments (0.03 and 0.12% K'I), the error in Ac varies from 0.3% to 16% for the cases given in Table 1. According to the manual, a temperature compensation circuit should correct the signal of the Luminox NO2 monitors for temperature influence. However, we found values for 5r~ and 5r~ of-0.3 and -0.15% K 1, respectively, in the range of 5 to 40~ Hence, the error in Ac varies from 5% to 150%. These large errors can be avoided by keeping the temperature inside the box at a constant value, as is done nowadays at Speulder forest. We note that for longer periods the average error due to temperature influence is nearly 0, since then temperature increases and decreases counterbalance. However, when studying temperature related processes, systematic errors might be introduced.
Interferences Sensitivity to other components (interferents) may lead to systematic errors. In general, the concentration measured by a monitor can be expressed as c' = c + (~ c + ~ ) cx
(6)
where c is the real concentration, c I is the concentration of the interfering component and r and 1] are interference parameters. The first refers to effects, which depend on the concentration of both gas and interferent, such as the quenching effect of humidity in chemiluminescence NO, monitors. The second refers to effects which only depend on the interferent concentration, such as the influence of all S-containing components on the signal of flame photometers used for measurement of SO2. Using eq.(6) for both c and c-Ac and eq.(4) for both the gas and the interferent, we derived
47
r
e r r A c = a c z + ( a + C~ ' )
vdZcz vd
(7)
where the t e r m ( ~ c I has been neglected. In most cases a = 0 and the error reduces to the product of 13 and the ratio of VdC of the interferent and the gas, respectively. Only NO and humidity interfere with the signal of the SO 2 monitors. For N O o~ = 0 and 13 = 0.024. NO is sometimes transported towards the soil and sometimes emitted. In both cases the absolute value of Vd for NO is usually much smaller than for SO2. NO concentrations are generally below 1 ppb, so that the error in Ac due to NO interference is generally smaller than 1%. The humidity interference appeared to be rather complicated. To first order eq.(7) is a good approximation for ambient concentrations of SO2 and humidity, if a is assumed 0 and 13 = -0.013 ppb per g m 3 H20. "Deposition velocities" (quotation-marks are used, since the humidity flux is usually directed upwards) of humidity were estimated from net radiation, sensible heat flux, temperature and relative humidity. Values varied from -0.2 to -1.2 cm s~ for unstable conditions, from -0.05 to -1 cm s~ for neutral conditions and from -~0 to -0.1 cm s 1 for stable conditions. Using these values and a humidity concentration of 10 g m 3 (~50-percentile), the error in Ac becomes 2 to 20%. The NO2 monitors are liable to interference of PAN, 03 and humidity. The PAN interference was determined by others [13,14] to be 25% (0~ = 0, 13 = 0.25). PAN concentrations and its v d values are usually very low [15] and the error in Ac is therefore less than 2%. For the humidity interference we found a = 0.004 and 13 = 0. With the Vd values and concentration of I-I20 given above, the error in Ac becomes 10 to 25%. From our interference experiments with 03 we found t~-0.001 and J3~4).01. These values are considerably higher than found by others [13,14]. With our experimental values the error in Ac is highly dependent on NO2 and O3 concentrations and varies between -25% to +100%. However, if results of others are used for a and 13, much smaller errors are obtained. This must be further investigated.
Noise Noise leads to random errors in measured concentrations. Suppose that the random error in c due to noise is t~. Because of its random character the error in Ac is about ~/2 times as large. Since the concentrations measured by the scanning monitor are corrected by those measure with the reference monitor, the error in Ac will be given by aA~ = 2 o ~
(8)
Noise was determined as the standard deviation in the measured signal (1-min average values), when a constant concentration was supplied to the instruments. For the SO2 monitors aAc = 70 ppt at the 50-percentile (c = 1.2 ppb). Hence, the error in Ac varies from ca. 10% at very stable conditions with high Vd tO 300% at very unstable conditions with low Yd. At neutral conditions, which occur most of the time, with Vd = 1 crn S1 (average value), the error is nearly 50%. For NO2 aAc = 180 ppt at the 50percentile (c = 10 ppb) and the error in Ac varies from about 25% at very stable conditions with high Vd tO nearly 1000% at conditions with very low Vd, USUally occurring at night, when the stomata are closed. At neutral conditions with Vd = 0.1 cm S1 (average value), the error is about 130%.
48 O t h e r errors
Errors due to drift, alinearity and response time were small. Drift is <0.1% day" or <0.1 ppt cycle' for the SO2 monitors and even less for NO2. The SO2 monitors were linear within 1% in the range used in the system (0-50 ppb). The NO2 monitors show strong alinear behaviour below 4 ppb and above ca. 20 ppb. Between these concentrations linearity is better than 1.5%. Below 4 ppb the calibration curve shows a dip; at c = 1.6 ppb deviation from linearity is 35%. Above 20 ppb the curve shows a downward bend and deviation from linearity rapidly increases with increasing concentration. For practical reasons the whole calibration curve cannot be regularly checked in the field. Therefore, only data between 4 and 20 ppb are selected for calculation of gradients. We note that this range covers nearly 70% of the entire range of concentrations at Elspeetsche Veld. The response times (100 s for the SO2 monitors and 1 s for the NO2 monitors at 95% of the total concentration change) are sufficient, since during all measurements the monitors are flushed for 5 min before signal sampling (see section 2.2). Summary Table 2 gives a summary of errors in gradients resulting from the behaviour of the gas monitors and the sampling equipment. The latter will be discussed in the next section.
3.2 Errors caused by sampling equipment Because all parts of the sampling equipment (tubes, valves, couplers) are made of FEP Teflon, reactions of SO2 and NO2 with the tube walls are not expected. Forming of water layers on the tube walls, in which the gases may dissolve, as a result of condensation is
Table 2 Summary of estimated errors in Ac values for SO2 and NO2 (in %) Parameter
error type
(so9
error in Ac (NO9
error in Ac
linearity
systematic
<2
<3
drift
systematic
<0.5
<0.5
temperature dependence
syste m atlc " ')
0 . 3 - 16
5 - 150
interference by H20
systematic
2 - 20
10 - 25
interference by NO
systematic
<1
-
interference by 03
systematic
-
-25 - +100
interference by PAN
systematic
-
<2
random
10 - 300
25 - 1000
systematic
=0
<5
noise sampling tubes
l) Worst ease situation. Errors are usually smaller. For long term average flutes the errors are nearly 0 (see section 3.1).
49 prevented as much as possible by isolation of the tubes. We checked this by several field experiments, e.g. by comparing concentrations measured by both SO2 and NO2 monitors connected to the same sampling tube (using a FEP Teflon tee just in front of the monitor inlets), and connected to different tubes with their inlets both at 4 m height. No differences were found, implying that adsorption effects are negligibly small. The photochemical equilibrium involving NO, NO2 and 0 3 will be disturbed inside the isolated, dark sampling tubes. There, NO 2 is still formed from NO and 03, whereas the photochemical conversion of NO2 into NO and 03 stops. This results in a net NO2 production, which can be estimated with the following equation derived by Beier en Schneewind [ 16] for short residence times in tubes A [NO 2] = k[NO]
CO3]At
(9)
in which A[NOz] is the net NO2-production (in ppb), k the reaction constant for the production of NO2 from NO and 03, At the residence time in the tube (2.6 s at Elspeetsche Veld) and [NO] en [03] the equilibrium concentrations in the air (in ppb). The reaction constant varies from 2.5x104 tot 4.5x10 -4 ppb ~ s1 depending on air temperature and pressure. We estimated A[NOz] using 50-percentiles of NO and 03 concentrations measured at nearby locations of the National Air Quality Monitoring Network (ca. 1 and 30 ppb, respectively) to be 35 ppt. Because NO and O3 show complementary behaviour, higher values will scarcely occur. The error due to the difference in net NO2 production at both measuring heights was determined analogously to the interference calculations. It was found to be less than 5%. 3.3 Errors in meteorological parameters According to eqs.(1) to (3) F depends on u. and AZ, which on its turn depends on u,, T and the sensible heat flux H 0. The uncertainty in u, and H 0 was determined by comparing values obtained with the two systems. For u, the linear regression slope was 1.05 and the intercept did not significantly deviate from 0; the correlation coefficient was 0.983 and the standard deviation of points around the regression line was 0.03 m sI. These results fairly agree with those of Erisman and Duyzer [17]. From the comparison we concluded that the random error in u, is less than 0.03 m s"I or 10%. However, systematic errors cannot be detected in this way. From the work of Moore, Zeller et al., Dillmann [18-20] and our own flow distortion experiments [6], it appears that u, values measured with the sonic anemometer in our system are underestimated by 8 to 12% due to sampling errors, response characteristics and flow distortion. Since the parameterisation yields approximately the same u. values, their systematic error is about the same. For H0 the agreement between measured and parameterized values was good for fluxes between -100 and 150 W m 2, as had also been found by Erisman and Duyzer [17]. Random differences are about 15%. At higher Ho, the parameterized values are underestimated by 20%. Depending on atmospheric stability, the systematic error in H 0 values measured with a sonic anemometer may be -I0 to 10% due to sampling errors, response characteristics [18,19] as well as flow distortion [6] and neglect.ion of terms in the calculation of temperature fluctuations [21]. Errors in u,/~Z were estimated using data measured at Elspeetsche Veld for several stability conditions and the errors in u, and H0 just given (the influence of T was neglected). These estimations yielded systematic underestimations of 5% at unstable, 10%
50 at neutral and 15% at stable conditions. Random errors of 10% (unstable, neutral conditions) and 2 0 % (stable conditions) must bc added to these values. Other meteorological parameters, which arc measured in the systems (net radiation, relative humidity, temperature, wind velocity and wind direction) arc used either as input for the paramctcrization or for process study and data selection. For that purpose the specified accuracy of the instruments is sufficient [6]. For all parameters values measured by similar instruments of both systems at Elspcctschc Vcld showed good agreement and specified accuracies wcrc conflnned.
3.4 Field intercomparison SO2 deposition fluxes and velocities measured simultaneously with both systems at Elspcctschc Vcld wcrc mutuaUy compared. Fig. 2 shows the course of Vd obtained with both systems during 4 days in April 1992, when nearly all measurements standed the selection criteria (scc section 1). Except for a few cycles the agreement is good. The average vd values, both for this 4-days period and for all cycles selected from the whole test period at Elspcctschc Vcld, did not systematically differ. The random uncertainty in vd was estimated from the root mean square difference between Vd values to bc ca. 75% for a 30-min cycle. This value is comparable to those obtained from the estimations in sections 3.1 and 3.3 with respect to random error sources. Estimated systematic errors could not bc verified in this way.
1(] (.n
E ID
5-
.
-
,
CM
0(./-j
,.."
.
""
J
:~
.
O
-----
-0
-5 23
0:~ 24 0:~ 25 day (April 1992)
0:~
26
Figure 2. Course of the SO2 dry deposition velocity (cm s1) measured with the two monitoring systems (indicated with full and dotted line, rcsp.) at Elspe~tschc Vcld.
51 NO2 fluxes determined from the gradient measurements at Elspeetsche Veld were compared to those obtained from a parameterisation, in which the NO2 surface resistance is assumed to be equal to the stomatal resistance. Despite the large errors in NO2 gradients, the agreement was fair. At night, the measured values were somewhat higher (measured and parameterised vd values were 0.07 and =0 cm s"1, respectively) and during daytime they were slightly lower (0.2 and 0.3 cm sl).
4. CONCLUSIONS Errors in SO2 and NO2 dry deposition fluxes determined with a monitoring system based on the gradient method were analysed in detail. Instrumental noise is the dominating random error source. The estimated errors in concentration gradients due to noise were 10 to 300% for SO2 and 25 to 1000% for NO2, depending on meteorological and surface conditions. At neutral conditions, which occur more than 60% of the time, these errors are ca. 50 (SO2) and 130% (NO2), respectively, for average vj values. Random errors in the ratio of u. and the stability corrected logarithmic height difference AZ are 10 (neutral, unstable) to 20% (stable conditions). For SO2 estimated errors were comparable to the root-mean-square difference between Vd values simultaneously determined with both systems at the heathland Elspeetsche Veld. We note that due to its random character the error in long term average fluxes will be small. However, for process studies (e.g. determination of daily course; study of relationship between fluxes and other parameters) errors may still be large, if the number of selected measurements is too small. This implies that such studies must be done carefully. At rapidly increasing temperature inside the box housing the gas monitors, significant errors occur due to temperature dependence of the monitors. Such errors can be minimized by keeping the temperature inside the box at a constant level, as is done nowadays in the system at Speulder forest. Significant systematic errors are caused by some interfering components. For the SO 2 monitors humidity interference causes errors of 2 to 20% (underestimation) at unstable to stable conditions, respectively. For the NO2 monitors O3 and humidity interference cause errors o f - 2 5 to +100% and -10 to -25%, respectively. Systematic underestimations in F resulting from errors in meteorological parameters (u, and H.) are 5% at unstable, 10% at neutral and 15% at stable conditions. Other systematic errors (due to drift, alinearity, other interfering components, sampling equipment) are smaller than 5%. Because systematic errors depend on several conditions, such as atmospheric stability, concentrations of interfering components and correlations between them, the impact on long term average SO2 and NO2 fluxes is difficult to determine. However, for SO2 we may conclude that systematic errors are generally lower than 20%. For NO 2 such a general statement cannot be made. It is better to correct fluxes for systematic errors. This can be partly done for errors in u, and H 0 (and hence AZ) by applying procedures proposed by Moore [18] and Zeller et al. [19]. Corrections for interferences can be made by applying the equations presented in this work and concentrations and deposition velocities of the interfering components. The latter should preferably be measured, but this is not always possible. However, estimations of these values are sufficient for first-
52 order corrections. For example, humidity concentrations and vd values can be well estimated from temperature, relative humidity, net radiation and sensible heat flux, which are all measured by our systems.
REFERENCES 1 2 3 4
5 6 7 8 9 10 11 12
13 14 15 16 17 18 19 20
21
G.J. Heij and T. Schneider (eds.), Acidification research in the Netherlands, Studies in Environmental Science 46, Elsevier, Amsterdam, 1991. B. Ulrich, Interaction of indirect and direct effects of air pollutants in forests. In: C. Troyanowsky, ed., Air pollution and plants, Wennheim, Germany, 149-181, 1985. J.W. Erisman, Water Air Soil Pollut., 71 (1993) 51. B.B. Hicks et al., Atmospheric processes research and procecess model development, State of Science/Technology report no. 2, National Acid Precipitation Assessment Programme, USA, 1989. M.G. Mennen, J.W. Erisman, B.G. van Elzakker and H.J.M.A. Zwart, In: G. Angeletti, S. Beilke and J. Slanina, eds., EC Air Pollution Report 39, (1992) 225. H.J.M.A. Zwart, J.E.M. Hogenkamp, M.G. Mennen and J.W. Erisman, Report no. 722108001, RIVM, Bilthoven, The Netherlands, 1994. J.A. Businger, J. Clim. Appl. Meteor., 25 (1986) 1100. J.W. Erisman, A.H. Versluis, T.A.J.W. Verplanke, D. de Haan, D. Anink, B.G. van Elzakker, M.G. Mennen and R.M. van Aalst, Atmos. Environ., 27A (1993) 1153. A.C.M. Beljaars, A.A.M. Holtslag and R.M. van Westrhenen, Technical Report no. TR-112, Royal Netherlands Meteorological Institute, de Bilt, The Netherlands, 1987. B.B. Hicks, D.D. Baldocchi, T.P Meyers, R.P. Hosker Jr. and D.R. Matt, Water, Air and Soil Poll., 36 (1987) 311. J.W. Erisman et al., Report no. 722108002, RIVM, Bilthoven, The Netherlands, 1994. H.J. van de Wiel and H.J.Th. Bloemen, Test procedures and minimum performance requirements to demonstrate equivalency of sulphur dioxide measurement methods for EC compliance monitoring, European Community catalogue number CD-NA11143-EN-C, 1987. Th.J. Kelly, C.W. Spicer and G.F. Ward, Atmos. Environ., 24A (1990) 2397. F.C. Fehsenfeld et al., J. Geophys. Res. D, 95 (1990) 3579. G.J. Dollard, B.M.R. Jones and T.J. Davies, A.E.R.E. Report R13780, Harwell, Oxfordshire, 1990. N. Beier and R. Schneewind, Ann. Geoph., 9 (1991) 703. J.W. Erisman and J.H. Duyzer, Boundary-Layer Meteor., 57 (1991) 115. C.J. Moore, Boundary-Layer Meteor., 37 (1986) 17. K.F. Zeller, W.J. Massman, D. Stocker, D.G. Fox, D. Stedman and D. Hazlett, Research Paper RM-282, USDA Forest Service, Fort Collins, Colorado, USA, 1989. M. Dillmann, Untersuchungen zur Genauigkeit bei der Bestimmung turbulenter Fliisse mit der Eddy-Korrelationsmethode, Meteorological Institute of the LudwigMaximilians University of Miinchen, Mtinich, Germany, 1991. P. Schotanus, F.T.M. Nieuwstadt and H.A.R. de Bruin, Boundary-Layer Meteor., 26 (1983) 81.
ATMOSPHERIC DEPOSITION SESSION H AMMONIA~AMMONIUM
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
55
A m m o n i a and a m m o n i u m in the atmosphere: Present k n o w l e d g e and r e c o m mendations for further research Willem A.H. Asman National Environmental Research Institute, Frederiksborgvej 399, DK-4000 Roskilde, Denmark
Abstract The atmospheric behaviour of ammonia (NH3)and ammonium (NI-I4+) including emission, surface exchange and wet deposition, is discussed. Some modelling aspects are also described. Although much more is known about the atmospheric behaviour of NH3 and NH4+ aerosol than ten years ago, many essential aspects are still poorly known.
1. INTRODUCTION Ammonia (NH3) and ammonium (NH4§ are important atmospheric components, the former being the most abundant alkaline component in the atmosphere. A substantial part of the acid generated in the atmosphere by the oxidation of sulphur dioxide (SO2) and nitrogen oxides (NOx) is neutralized by NH 3. As a result NH4§ is a major component of atmospheric aerosols and precipitation. NH3 and NH4§ also play important roles in biological cycles. Nitrogen (N) can be a limiting factor for growth in oligotrophic ecosystems. In many ecosystems a substantial part of the N input is caused by the deposition of NH3 and NH4§ from the atmosphere. In some areas these two components dominate the N input. Very high concentrations of NH3 (annual average of about 75 lag m 3) can cause direct damage to vegetation (van der Eerden, 1982). When NH3 and NIL § are deposited and enter the soil as NH4+, nitrification can occur by Nitrosomas and Nitrobacter leading to the overall reaction (Van Breemen et al., 1982): NH4+ + 2 02 --->2 H § + NO3 + H20. As a result, not only is acid formed by the oxidation, but also the acid formed in the atmosphere is no longer neutralized by NH3. In this way NH3 and NH4+ can cause acidification of the soil which may lead to adverse effects on vegetation. In forest ecosystems high inputs of NH3 and NH4§ lead to the leaching of K § Mg 2§ and Ca 2+ from the soil, often resulting in increased ratios of NH4§ to K +, Mg 2+ and/or A13+ to Ca 2§ in the soil solution (Roelofs et al., 1985). As a result, a net flux of Mg 2+, Ca 2+ and K § from the root system to the soil solution occurs. Moreover, coniferous trees can take up NH4+ by the needles and compensate for this by excreting K + and/or Mg 2§ This combination of effects results in K § and/or Mg 2§ deficiencies, severe nitrogen stress and as a consequence premature shedding of needles.
56 Another effect is that large deposition fluxes of N H 3 and NH4+ cause nitrogen-poor species to disappear, because they are no longer able to compete with nitrophilous species. An example of this is found in the lowland heaths in The Netherlands. Bobbink et al. (1992) report that over 35% of the former heathland has now been replaced by grassland. (For an overview of effects see Sutton et al., 1993c). In central Europe about 65-80% of the threatened vascular plants are adapted to low nutrient conditions (Nilsson and Grennfelt, 1988). In this paper the symbol NI-I~ is used, standing for the sum of NH 3 and NH4§ In large parts of Europe the critical load of nitrogen to ecosystems is exceeded. The critical load is defined by "a quantitative estimate of an exposure to one or more pollutants below which no harmful effects may occur" (Nilsson and Grennfelt, 1988). The critical load for forests and heathlands is 15-20 kg N ha ~ year ~ and is even lower in some more sensitive regions (Hettelingh et al., 1991). Asman (1994a) presents a detailed review of NH x. In this paper a short review is given of all processes involved as well as atmospheric transport models. Needs for further research are also discussed.
2. EMISSION
Most of the 4534 ktonne N year 1 NH 3 emission in Europe (excl. the former USSR, Figure 1), is caused by agricultural activities (Buijsman et al., 1987; Asman, 1992). The emission from livestock is dominant, but that from the application of fertilizers is also important. The emission of NH 3 in Europe is comparable to that of nitrogen oxides (5094 ktonne N year 1 excl. the former USSR; Pacyna et al., 1991). No significant emission to the atmosphere of either NH4§ or organic nitrogen components takes place. The NH 3 emission from livestock depends on many factors (Isermann, 1990): the nitrogen content of the food and the relative share of amino acids, the housing and storage system, the farmer's way of working, properties of the manure, fertilizer and the soil, the method of application of manure and fertilizer, the time between spreading and ploughing (for arable land), the duration of the grazing period and meteorological conditions. Information on global NH 3 emissions can be found in Schlesinger (1992). Emission factors for livestock categories representative for The Netherlands are presented are: 19.0, 4.4, 0.2, 10.0 and 1.4 kg N animal -~ year a for respectively cattle, pigs, poultry, horses and sheep (Asman, 1992). The emission factors for fertilizers vary from about 2 to 15% of the nitrogen content, depending on the type. The average emission factor for fertilizers in Europe is about 5% of the nitrogen content (Asman, 1992). The emission per animal in the stable can vary up to a factor 10, even for the same stable type. If the animals are grazing in the meadows, the manure is not stored, but deposited directly. After deposition it is therefore immediately exposed to loss processes other than volatilization of NH 3 to the atmosphere. These loss processes are: uptake by the grass, wetting by precipitation (leading to dilution and penetration of the soil with diluted manure) and nitrification. The NH 3 emission rate during the grazing period is for this reason less than it would be if the animals were in the stable, including the contribution during storage and subsequent spreading. The total emission from animals therefore depends on the fraction of the time they are in the meadows. This fraction is not known for all European countries. Van der Hoek (National Institute of Public Health and Environmental Protection, RIVM, pers.
57 comm., 1991) estimates that if the cows in The Netherlands were the stable all year round, the emission factor would be about 59 kg NH 3 animal 1 year 1, but if all the cows were in the meadows during half a year the emission would then be 31 kg NH 3 animal -~ year 1. Table 1 NH 3 emission in European countries excluding the former USSR (ktonne N year 1) (Asman 1992). Category
Emission
%
Cattle Pigs Poultry Horses Sheep Fertilizer application
2391 819 282 57 231 754
53 18 6 1 5 17
Total
4534
100
These facts illustrate that emission factors are not constant and that large geographical differences will occur. At the moment, however, insufficient information is available to calculate local emission factors for all areas and/or farm types in Europe. The uncertainty in the annually averaged European NH 3 emissions is at least 30-40%. The fact that geographical distributions of airborne NH 3 and NH4§ as well as wet deposition can be reproduced by atmospheric transport models using the current emission inventory, indicates that the current estimate and its geographical distribution is reasonable. It should be noticed here that even if all other conditions were constant, the annually averaged NH 3 emissions would show interannual variations that are caused by variations in meteorological conditions. The emissions are often not equally distributed. The average emission density in The Netherlands is about 50 kg N ha ~ year ~, but the highest emission density on a 5x5 km 2 grid element is over 200 kg N ha ~ year ~. For a sparsely populated country like Sweden the emission density is much less (1 kg N ha 1 yearl). The emission densities refer to the average emission for the whole territory of a country, i.e. not only agricultural areas. The NH 3 emission in Europe has doubled since 1950 (Asman et al., 1988). The average NH 3 emission rate is likely to show a large diurnal variation with a peak during the early afternoon (Asman, 1992). This peak can be explained by a peak in turbulence and one in temperature, both favouring evaporation of NH 3. Moreover, many agricultural activities like manure spreading take place mainly during daytime, which also will lead to increased emissions during daytime. But the emission rate for any particular day can be quite different from this average pattern. The NH 3 emission rate will also show a significant seasonal variation. This seasonal variation is caused by the spreading of manure and fertilizer in spring and autumn and by the practice of keeping cattle in the meadows in summertime in a large part of Europe. A seasonal variation of the NH3 emission, derived from measurements of NH4§ in air and
58 precipitation, does not show such large variations. This could indicate that additional NH 3 sources exist, which are important during part of the season (Asman, 1992). Apart from the major sources, minor sources of NH 3 emission are present. Unfortunately, much less is known about these sources. Buijsman et al. (1987) indicate that these emissions are about 10% of those from animal manure and fertilizer. They include emissions from undisturbed land and are not included in Table 1. There are recent indications that evaporation of NH 3 from agricultural crops may also be an important source during part of the season (Schj~rring, 1991; Schj~rring et al., 1992; see also section 3).
60 55 50-
.....
1;:1::::1:::1::::11111
~i:~i~ii!iiiiiii~i::::::~
45-
i:i:~!:!ii:ii 4035-
~
30-
..
:1"
...." .
25
,==,=,,,,,,
lemwuurs-e-lrmmmQ
.
.
.
, ~
.... .
. . . . . . . .
15-
~
"
~
....
""
~
9
::::;::m I
20
.......
~:..
" ....
~~
~
obove
"
.
'
r%,
~_/'~"
r
":~iiiii!iiiii
10-
6.00 -
4.00
1.00-
2.00
0.50-
I ....
20
I ....
25
I ....
30
I ....
35
I ....
40
I ....
45
I ....
50
I ....
55
I ....
60
I ....
65
I ....
70
I ....
75
6.00
2.00 -
1.00
0.25
-
0.50
0.00
-
0.25
I ' ' '
80
E-coordinates
Figure 1. Emission density NH 3 emission on IE grid (75x75 km 2) in Europe without Russia (tonne NH 3 km 2 year1). (Reprinted from Asman and van der Hoek, submitted to Atmospheric Environment, 1994. Copyright 1994, with kind permission from Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK). There are also indications that under certain circumstances the sea can act as a source of NH 3 (Quinn, et al. 1988 a,b; Quinn et al., 1990; Asman et al., 1994a). The influence of emissions over the sea is not unimportant for land areas in those parts of Europe that have very low emission densities. Information on emission sources in cities is almost non-existent. Cass et al. (1982) give a list of possible NH 3 sources for California, which could also be very useful
59 for regions in other parts of the world and Lee et al. (1992) investigated emission from sewage treatment plants in the UK. In general, it can be concluded that there is a lack of knowledge on the geographical distribution of factors that influence the emissions, on emissions from vegetation, emissions over the sea and in cities, on diurnal and seasonal variations in the emission rate, and on what is called "minor sources", which may be very important in more remote areas, such as north Sweden. Moreover, emission inventories with a high spatial resolution (e.g. 5x5 km 2) are needed, at least for areas where dry deposition of NH 3 is relatively important (for areas with a high or medium emission density, see section 6).
3. SURFACE EXCHANGE An extensive overview of the surface exchange of NI-I~ from a biological point of view is given by Sutton (1993a,c). By surface exchange is meant here the exchange attributed to atmospheric turbulence. This includes both dry deposition to the surface as well as emission from the surface, but not wet deposition, which is treated in the next section. In contrast to wet deposition, dry deposition is very difficult to measure (Fowler and Duyzer, 1989; Davidson and Wu, 1990). This means that dry deposition is usually not monitored continuously. It is mostly computed from continuously monitored concentrations and meteorological parameters and dry deposition velocities or surface resistances measured by micrometeorological methods during short field campaigns. The exchange velocity ve is a measure of the exchange caused by turbulence. The simplest model to describe the surface exchange is the "big leaf model". In this model the transport from an arbitrary point in the atmosphere (reference height) to the surface takes place in three subsequent steps. The first step occurs in the turbulent layer from the reference height to a laminar boundary layer very close (- 1 mm) to the surface. The second step is through the laminar boundary layer to the surface. The third step is the uptake by the surface. These steps can be represented by the resistances to transport. If emission occurs, the order of the resistances to be overcome is opposite to the one for deposition. The exchange velocity ve can be expressed as: v. -
1 ra + rb +
(1) re
where: v, is the exchange velocity (m sl), r, is the aerodynamic resistance (s ml), r b is the laminar boundary layer resistance (s m ~) and rc is the surface resistance (s m ~) The "big leaf model" is a simple, but quite useful model. It is, however, possible to make more complicated models, e.g. by splitting resistances into those for transport into different vegetation surfaces. But then all these resistances have to be known. They usually are not because it difficult to infer them from measurements. Under the same meteorological conditions, the aerodynamic resistance is the same for all gases and in fact also for aerosols. Only for aerosols with radii > 5 lam does the additional contribution of gravitational settling become significant. The transport through the laminar boundary layer takes place for gases by molecular diffusion. This is a rather efficient process. Transport of particles is caused by
60 several processes, none of which are efficient for those with radii < 5 lam. The large difference in exchange/deposition velocity between NH 3 and particulate NH4§ is caused mainly by the difference in laminar boundary layer resistance. The surface resistance for uptake of NH3 by the surface is often low for water-containing media. This is caused by the high solubility of NH 3 in water. The surface resistance for uptake/release of NH 3 is negligible for manure, freshwater and seawater. NH 3 is absorbed well by plants by transport through the stomata (Hutchinson et al., 1972; van Hove et al., 1987). The surface resistance of plants is, however, also low at night, indicating that absorption occurs on other plant surfaces as well (Sutton et al., 1992a). For vegetation ro will be of the order of 0-100 s m 1. In atmospheric transport models a value of 30 s m 1 is often used. The surface resistance for NH4§ which is associated with particles with radii < 2 lam, is assumed to be zero, i.e. the surface is assumed to be a perfect sink. Due to the low diffusivity of the particles, transport through the stomata is negligible and they will be deposited mainly on the outer parts. The average exchange velocity of NH 3 will be about 20 mm s ~ for moorland and grassland (Duyzer et al., 1987; Sutton et al., 1992a, b; Erisman and Wyers, 1994), about 3040 mm s ~ for forest (Duyzer et al., 1992; Wyers et al., 1992; Andersen et al., 1993) and about 8 mm s-' for (sea)water (Lindfors et al. 1991; Asman et al., 1994b). These differences in exchange velocity are caused mainly by differences in roughness for momentum between these surfaces. The exchange/dry deposition velocity of particles is very difficult to measure, because it is usually rather low. In atmospheric transport models a dry deposition velocity of about 1 mm s1 is used for particulate NH4+. Duyzer et al (1987) found a value of 1.8 mm s~ for heather/purple moor grassland. H6fken et al (1983) derived indirectly a dry deposition velocity of 5-15 mm s 1 for a forest, which is rather high. Experiments (Larsen et al., 1994) have shown that seaspray does not enhance the dry deposition of particles, as has been suggested by Williams (1982). This means that the dry deposition velocity of particulate NH4 § at sea is less than 1 m m s -1. The flux to the surface is given by: F " -vo (c a - cs')
(2)
where: F = Flux of the component (mol m 2 s l ) . In this case the flux is by definition negative when material leaves the atmosphere. ca = The concentration of the component in the atmosphere at reference height (mol m 3). c~*
= The theoretical concentration of the component in air (mol m-3), which would be in equilibrium with the concentration in the surface. For NH3 c~* can be important. For particulate NH4§ this is not the case, because particles are not re-emitted and c~* can therefore be set to 0 in the equation, c~* is called "compensation point" and is for NH4 § a function of the pH, temperature, NH4 § concentration and ionic strength of the surface (Asman et al., 1994a).
The flux can be split up in the deposition flux F d = -VeCa and emission flux F e = vecs*. The direction of the flux depends entirely on the concentration difference (ca - cs*).
61 The dry deposition velocity v d (m s~) is defined as: Vd ~- F / c
(3)
In this equation v a and c, refer to the same reference height. By comparing (3) with (2) it can be seen that v a = v e (1 - c~*/c~) and that v a = ve when c~* ~ c,. It is common practice to measure F and c, and to derive Vn from it. Then the surface resistance is often computed by estimating r, and r b from meteorological measurements and subtracting r, + r b from 1/v d (rc' derived from measurements is equal to 1/Vd - r~ - rb). The computation of Va is correct, but re' is equal only to r~ if c*~ = 0. If r~ = 0, but c*~ ~ 0, then r~' would actually exceed 0. In practice though, this will pose problems only when c*Jc, > 0.2 ; in other words, a value of r~', which is significantly larger than 0 could not only indicate that the surface resistance is larger than 0, but also that the component could be present in the surface. NH 3 is present in many surfaces. In manure or recently fertilized soil large concentrations occur and c~* will be much larger than c,, resulting in a net emission. Also in the sea c~* is not zero and can sometimes be larger than c, (Quinn et al., 1988a,b; Quinn et al., 1990; Asman et al., 1994a). In agricultural crops there NH3 is also present in the surface (see Farquhar et al., 1983 and Schj0rring, 1991 for a review) and emission can occur. NH 3 is an important intermediate in the photorespiratory N cycle, in the conversion of nitrate to amino acids, and in the breakdown of proteins. The compensation point may vary considerably during the life cycle of plants. It seems to be especially high during senescence (Harper et al., 1987; Parton et al., 1988; O'Deen, 1989; Morgan and Parton, 1989). High concentrations can also occur during grainfilling and after anthesis. Compensation points of 1-5 pg NH 3 m 3 in air are not uncommon for agricultural crops in northwestern Europe (Schjcrring, 1991). Semi-natural ecosystems often show very low compensation points (0.01 lag NH3 m 3 in air; Sutton et al., 1992a,b) Langford and Fehsenfeld (1992) found a somewhat higher compensation point of 0.1 lag NH 3 m 3 in air in a forest in Colorado. Quinn et al. (1990) found a surface concentration ranging from 0.05 - 0.36 lag NH 3 m 3 (average 0.17) for seawater in the North Pacific. Asman et al. (1994a) found an average surface concentration for the North Sea of about 0.27 lag NH 3 m 3. The compensation point in manure or intensively grazed pastures is often so high, that the airborne concentration has no influence on the flux. Most of the NH 3 emission originates from low-level sources (ground-level, stables), which results in rather high concentrations close to the earth's surface. The concentration decreases rapidly with distance due to atmospheric mixing. The flux to the earth's surface close to the source is often high. In this way more than 20% of the NH 3 emission can be dry deposited within a few kilometres from the source (Asman and van Jaarsveld, 1992). This situation differs from that for SO 2, which originates mostly from high-level sources for which dry deposition cannot take place until atmospheric turbulence has mixed the emissions down to the surface - this takes a considerable distance to achieve. A reasonable estimate of the exchange velocity of NH 3 can be found by assuming a surface roughness for momentum of 0.3 m and a surface resistance of 30 s m 1. This would give a dry deposition velocity of 22 mm s1 for a windspeed of 5 m s ~ and a surface concentration of 0. This would lead to a removal rate due to dry deposition of about 7% h -~ if the NH 3 is homogeneously mixed over a 1000 m high mixing layer. Under the same conditions, assuming a laminar boundary layer resistance of 600 s m -1 a removal rate of 0.6% h -1 can be found for NH4 +.
62 It would be a good idea to model the surface exchange of NH 3 from the exchange velocity and a surface concentration. In this way both emission and dry deposition could be modelled. This is, however, possible only if the surface concentration were known. The past ten years has given us much information on the exchange velocity of NH 3 on land. More information is needed on the exchange velocity at sea, the surface concentrations of NH 3 (vegetation, sea) and its temporal variation, and on the dry deposition velocity of particulate NH4§
4. WET DEPOSITION Components can be removed by different wet deposition processes. There exist removal processes within clouds (in-cloud scavenging) and removal processes below the cloud base (below-cloud scavenging), where components are removed by falling raindrops and snowflakes. Cloud- and precipitation water are usually acidic. Consequently, most of the NH 3 taken up by the drops reacts with H § to form NH4§ It is therefore possible to distinguish between the contribution of NH 3 and particulate NH4§ only if models are used. NH 3 is a highly soluble gas. Cloud droplets are so small (about 10 tam) that they take up NH 3 rapidly. Almost all NH 3 is found in the cloud droplets after the few seconds it takes to achieve equilibrium with NH 3 in the surrounding air. The NH 3 concentration in the interstitial air has then become very low. NH4§ aerosol acts as condensation nucleus, i.e. that water vapour condenses onto aerosols when the air becomes saturated with water vapour. In this way almost all NH4§ in clouds will become part of cloud droplets. NH 3 and NH4§ are transferred rapidly to the cloud droplets, but this does not necessarily lead to their removal from the atmosphere. The removal of NH x from the atmosphere by in-cloud scavenging is therefore more determined by the dynamical and physical processes that result in precipitation formation. Raindrops of typically a radius 500 lam and are much larger than cloud drops. The time they need to fall from the cloud base to the surface is relatively short (a few minutes). Raindrops are so large that the transport of airborne NH 3 to the drops is not fast enough for equilibrium with the surrounding air to be reached before they strike the surface. Consequently they will take up NH 3 after collection, unless contact with the surrounding air is avoided. For this reason the NH4§ concentration in precipitation in agricultural areas is often too high. This is caused by uptake (dry deposition) of NH 3 to the wetted funnel of bulk collectors, which are not closed during dry periods. NH4§ aerosol is not captured very well by falling raindrops. Usually not enough information is known to model dynamical and physical processes in clouds in atmospheric transport models. Moreover, if such information were available it would often take too much cpu-time to perform the necessary calculations. For these reasons scavenging is often modelled by using so called "scavenging coefficients". The change in airborne concentration due to scavenging is then described by: c, -- c,.o e -xt
(4)
where %0 is the concentration in the air at the onset of the precipitation (mol m-3), ~, is the scavenging coefficient (s 1) and t=time (s).
63 Scavenging coefficients are used to describe the effect of in- and below-cloud scavenging separately, or sometimes an overall scavenging coefficient is used to describe the effect of both processes together. Scavenging coefficients apply to gases as well as aerosols. A general function for the scavenging coefficient is: ~, = a I b
(5)
where I is the precipitation rate (mm hl). The scavenging coefficient increases with precipitation rate. For the cloud volume (in-cloud scavenging) a is about 4x10 4 and b about 0.64 for both NH 3 and NH4§ For the below-cloud volume (below-cloud scavenging) a is about 9.9x10 5 and b about 0.62 for NH 3 (Asman, 1994b). The below-cloud scavenging coefficient of particulate NH4+ is not well known, but is less than 10.5 s1 (Pruppacher and Klett, 1978). Below-cloud scavenging occurs in the lowest few hundred metres of the atmosphere (the average cloud base height during precipitation in northwestern Europe is about 300-400 m), whereas in-cloud scavenging takes place in a much larger volume. Below-cloud scavenging is less efficient than in-cloud scavenging, both for NH3 and NH4§ aerosol. Despite the much larger concentration of NH 3 near the surface in areas where net-emission is occurring (F_~sman et al., 1988), in-cloud scavenging contributes the largest fraction to the NH4§ concentration in precipitation. Computations for Denmark show that the contributions of the different processes to the NH4§ concentration in precipitation is." in-cloud scavenging of NH 3 15%, in-cloud scavenging of NH4§ 77%, below-cloud scavenging of NH 3 6% and belowcloud scavenging of NH4§ 2% (Asman and Jensen, 1993). During precipitation periods NH 3 and NH4 + are removed very efficiently (on the order of 75% h l) from the atmosphere, much more so than the removal due to dry deposition. But as precipitation occurs only 5-10% of the time in northwestern Europe, the total amount of NI-t~ wet deposited is not necessarily larger than the total amount dry deposited. Maps with the annual wet deposition of NH~ in Europe are presented in Buijsman and Erisman (1988) and Schaug et al. (1993). The wet removal of NI-Ix in statistical transport models can be modelled well if information is known on precipitation statistics (length of dry and wet periods). In other transport models it is necessary to know the same type of information, i.e. the fraction of the area that is exposed to precipitation if the whole area is supposed to be wet when precipitation is collected in one sampler. Usually only part of an air mass is exposed to precipitation. In this part NH x will be removed almost entirely after one hour of exposure. The concentration in the dry part will not change much and can be transported to other areas. More information on this type of precipitation statistics is needed.
5. REACTION In northwestem Europe most NH 3 reacts with acid aerosols that contain sulphuric acid (H2SO4). This reaction has been investigated in the laboratory (Robbins and Cadle, 1958; Baldwin and Golden, 1979; Huntzicker et al., 1980; McMurry et al., 1983). The reaction proceeds rapidly at high relative humidity and will take only a few seconds. The reaction rate is limited by the rate at which NH 3 diffuses to the acidic particle. At low relative
64 humidity the reaction is slower because only 10-40% of the collisions of NH 3 with a particle lead to a reaction (Huntzicker et al. 1980; McMurry et al., 1983). The reaction proceeds faster when the particles involved are small, because the diffusion is then faster. The pseudofirst-order reaction rate decreases with the degree of neutralization of the particle. Moreover, the reactions by which H2SO4-containing aerosols are formed could also limit the uptake of NH 3 if all of the acid is already neutralized. Not enough is known at present about these possibilities to quantify the reaction rate that actually takes place. A minor part of NH 3 reacts with gaseous nitric acid (HNO3) and gaseous hydrochloric, acid (HC1) to form particulate NH4NO 3 or NH4C1, which is part of the aerosols that contain other components (Stelson et al., 1979; Stelson and Seinfeld, 1982a,b,c; Pio and Harrison, 1987; Allen et al., 1989). Unlike the reaction with H2SOn-containing aerosol, which is a one-way reaction, these reactions can occur in both directions" N H 3 + H N O 3 ~ NH4NO 3
and
NH~ + HC1 ~ NH4C1
(6)
The fact that these reactions are in equilibrium means that if the concentration of one or two of the gaseous components becomes very low the component in aerosol form will dissociate. NH 3 can also react with OH, O and O(~D). Levine et al. (1980) found that the reaction with OH was most important. When adopting a constant and relatively high OHconcentration of 4x 106 molecules cm 3 (Logan et al., 1981) a pseudo-first-order reaction rate of 5.4x10 7 s ~ for NH 3 is found. This value is much lower than the pseudo-first-order rate for the reaction with H2SO4-containing aerosol, HNO3 and HC1 (see below). These reactions are therefore usually neglected in regional transport models. It is also possible to estimate the reaction rates from field measurements, but it is then necessary to make several assumptions, which may not always be valid. Lenhard and Gravenhorst (1980), Erisman et al. (1988) and Harrison and Kitto (1992) measured a pseudofirst order reaction rate of NH 3 between 10 4 and 10 .6 s1. The reaction rate can also be estimated by changing the rate in atmospheric transport models until the best agreement is obtained with measured concentrations of NH 3 and NH4+ in air and NH 4 in precipitation. Asman and Janssen (1987) found in this way a pseudo-first-order reaction rate of 8x10 s s ~. A value of this order gives good results even for annual average concentrations on a global scale according to Dentener and Crutzen (1993), who include all relevant processes in a more sophisticated way. This reaction rate is about 30% h ~, which is much greater than the oxidation rates of sulphur dioxide (SO2; of the order 1% h 1) or nitrogen dioxide (NO2; of the order 4% hl). It should be noted that the rate derived in this way is an annual average value. The rate may well show diurnal and seasonal variations which may depend on the local meteorological and chemical conditions. As a consequence, even the annual average reaction rate may show geographical variations. The dry deposition velocity of NH 3 is large compared to that of particulate NH4§ A high reaction rate, such as has been found, favours particulate NH4+ over NH 3. As a consequence, the dry deposition velocity of NH x as a whole is lower than if the reaction rate were lower. The lower dry deposition velocity of NH4§ promotes also long-range transport of NH x. It is rather unsatisfactory that no more information is known on the reaction rate of NH 3. More information on the reaction mechanisms and the temporal and spatial variability of the reaction rate is needed.
65 6. M O D E L L I N G A historical overview of the modelling of NI-I~ is given by Asman (1994a). Most of the NH3 is emitted from a large number of scattered low-level sources (ground-level, stables). As a consequence the NH3 concentration shows an extremely high spatial variability. Thousands of stations are needed to get a reliable average concentration for a country. This of course is not possible and that is why models are urgently needed in this case. But a good spatial resolution in a model can be achieved only if there is an emission inventory with a high spatial resolution is available. The results of a model can then be verified with measurements at a limited number of stations in areas with different emission densities (Asman and van Jaarsveld, 1992).
::..
"~..i:~i-~-.-...
'.':
. . . . '
.
:
"..':'
:.'"
9
iii~i.~iii~ii!!~ii:iii'iiiii
_
.:::::-
9
" : ;
:!'I!~!!~!!M .......... ........... : -..-::::: 9. . . . . . . . 9
9
.
.
.
.
.
.
.::.:~.:! . . . . . . . . . . . . . . . .
.
i
.
.
....
..
- .
~9. : : : .....
..
:
8|
iiiiiiiii !iiiiiiiii !iiiiiiiJi utiJiiiil l/ (
200
(
400
(
800
(
1600
(
3200
Figure 2. Modelled total deposition of NI-I~ in Europe (mol ha 1 year1); 1000 mol ha 1 a1 = 14 kg N ha ~ year a. (Reprinted from Asman and van Jaarsveld, 1992. Copyright 1992, with kind permission from Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK).
66 Models are also needed to calculate import/export balances for areas. The reaction product NH4+-containing aerosol is transported over long distances. This results in concentrations in air and precipitation (the contribution of NH4§ aerosol to precipitation is larger than that of NH 3) which do not show a high spatial variability. This means that a good impression of the concentration field can be obtained from measurements. Normally atmospheric transport models are developed either for short ranges (< 10 kin) or for long ranges (500 - 2000 kin). A transport model for NH~ should be capable of both, at least if realistic NH 3 concentrations and dry depositions are needed. The average value of the dry deposition of NH 3 in a grid element in a model (e.g. a 150x150 krn 2 area in the EMEP model) is not incorrect. It is, however, not representative of the deposition of a nature area that covers only part of the grid element. Some attempts have been made to correct the EMEP model outputs for differences in deposition to ecosystems within the 150x150 km 2 grid element used (Hettelingh et al., 1991). This is done by redistributing the deposition within the grid element in such a way that the total deposition to the whole grid element does not change. In this way mass is conserved (see also discussions in Sutton et al., 1993b), but due to the local character of NH 3, this approach is unlikely to give correct results near important emission areas, i.e. those areas where dry deposition of NH 3 is larger than wet deposition of NI-Ir Models can also be used to compute the import or export of NH~ to or from countries. It is almost impossible to infer such transboundary fluxes from measurement data. They can also be used to estimate the contribution of in- and below-cloud scavenging of NH 3 and NH4§ aerosol to the wet deposition of NH~. This cannot be derived from measurements. Moreover, models can be used to estimate historical or future depositions by using emissions representative for these periods. In this case one should check that the emissions of other components like SO 2, NO~ and hydrocarbons have not changed so much that the reaction rate of NH3 to particulate NH4§ is altered considerably (Asman et al., 1988). Most models are able to reproduce measured concentrations and depositions reasonably well, except for NH 3, which can be handled by only a few models and then not even very well. Figure 2 shows the total NI-I~ deposition in Europe (sum of wet and dry deposition of NH3 and NH4§ Model results show that in northwestern Europe 44% of the emitted NH 3 is dry deposited as NH3, 6% is wet deposited as the contribution of NH3 to the wet deposition of NI-I~, 14% is dry deposited as NH4§ aerosol and 36% is wet deposited as the contribution of NH4§ aerosol to the wet deposition of NI-I~ (Asman and van Jaarsveld, 1992). There is a need for atmospheric transport models for NI-I~ that have a spatial resolution that is sufficient to calculate realistic NH 3 concentrations. Such models should also be able to take into account that the dry deposition velocity differs from one surface to another, as otherwise e.g. no realistic dry deposition of NH 3 to forests can be calculated.
7. CONCLUSIONS The most important conclusion which can be drawn from the information presented here is that deposition of NHx takes place mainly in two forms, namely dry deposition of NH3 close to the source and wet deposition of NH~ at larger distances from the source contributed by the NH4§ aerosol. This indicates that the possibility exists of reducing the deposition of NH~ to nature areas close to those with a high NH3 emission density. This can be done to
67 some extent by selectively reducing NH 3 emissions close to these areas. This strategy does not work at long distances from important source areas, however. In these areas deposition of NH x can be reduced only by cutting down all the emissions in a much larger area. A policy to reduce the deposition of NI-I~ should also take into account the contribution of NOx and its reaction products to the total nitrogen deposition, as well as protection of other parts of the environment (soil, ground water) from emissions when atmospheric emissions are reduced. There exist good technical possibilities for reducing NH 3 emissions. A reduction can be obtained by altering the factors which lead to high emissions: a. Reducing the nitrogen content of the animal food in such a way that optimal nutrition is obtained for the stage of development of the animal under consideration. Nowadays the same food is often used for animals of almost all ages. b. Adding some amino acids to the animal food for non-ruminants (pigs etc.), so that no overdoses of other amino acids are needed. c. Prescribing housing and storage systems which give the lowest losses. There exist considerable differences in the emission rate per animal for different housing and storage systems. d. Ploughing the manure under as soon as possible after spreading or injecting the manure in the ground in the case of grassland. f. Spreading of manure and application of fertilizers under meteorological conditions (low temperatures, just prior to the onset of precipitation) which favour low emissions. Apart from these measures, emissions of NH 3 can be reduced by taking more technical measures, such as biofiltring of the air coming from stables. These measures are more expensive, however. But it is most important that the nitrogen cycle for a country, region or farm be as balanced as possible. This means a sharply reduced import of nitrogen-containing animal food, limitations to the number of animals ha 1 and in general, eliminating waste of such a precious element as nitrogen, leading to fine-tuning of the nitrogen supply to match the nitrogen demand of the crops. The conversion of nitrogen from manure and fertilizers into plant products is rather efficient (70+10%; Isermann, 1993). This efficiency can be increased by the adding of amino acids to the food, as mentioned earlier. The fact that humans in western Europe eat (animal) protein far in excess of their requirements than they need, leads also to a high production of animal proteins with all the ill consequences for the environment. Reduction of the share of animal proteins in the human diet could therefore also have beneficial environmental consequences. Although much more is known about the atmospheric behaviour of NI-I~ than 10 years ago, far from all essential information is known. Good progress has been made on the exchange velocity of NH 3 and also on the development of methods for continuously measuring NH 3. No models are, however, presently available that give good results for dry deposition of NH 3 to forests. Further information is needed on: a. Geographical differences in emission factors. b. Diurnal and seasonal variation in the NH 3 emission. c. Emission inventories for some countries with a high or medium emission density are needed. d. Emissions from plants and cities and other "minor sources". e. Exchange velocity of NH3 at sea. f. Surface concentration of NH 3 for vegetation and seawater and its temporal variation.
68 g. The dry deposition velocity of particulate NH4§ h. Precipitation statistics needed in atmospheric transport models. i. The reaction of NH 3 to NH4+: mechanisms, rate and its temporal and spatial variation. Moreover, atmospheric transport models for NI-I~ should be developed that have a spatial resolution that is sufficient to calculate realistic NH 3 concentrations and the possibility of having different dry deposition velocities for different surfaces.
8. REFERENCES
Allen, A.G., Harrison, R.M. and Erisman, J.W. (1989) Atmospheric Environment 23, 15911599. Andersen, H.V., Hovmand, M.F., HummelshOj, P. and Jensen, N.O. (1993) Atmospheric Environment 27A, 189-202 (1993). Asman, W.A.H. (1992) Report no. 228471008, National Institute of Public Health and Environmental Protection (RIVM), Bilthoven, The Netherlands. Asman, W.A.H. (1994a) Nova Acta Leopoldina NF 70, No. 288, 263-297. Asman, W.A.H. (1994b) Parameterization of below-cloud scavenging of highly soluble gases under convective conditions. Paper submitted to Atmsopheric Environment. Asman, W.A.H., Drukker, B. and Janssen, A.J. (1988) Atmospheric Environment 22, 725-735. Asman, W.A.H., Harrison, R.M. and Ottley, C.J. (1994a) Estimation of the net air-sea flux of ammonia over the southern bight of the North Sea. Paper accepted for publication in Atmospheric Environment. Asman, W.A.H. and Janssen, A.J. (1987) Atmospheric Environment 21, 2099-2119. Asman, W.A.H. and Jensen, P.K. (1993) Report 26 Danish Sea Research Programme 90, Danish Environmental Protection Agency, Copenhagen, Denmark. Asman, W.A.H., Sr L.L., Berkowicz, R., Granby, K., Nielsen, H., Jensen, B. and Runge, E. (1994b) Report 35 Danish Sea Research Programme 90, Danish Environmental Protection Agency, Copenhagen. Asman, W.A.H. and Van Jaarsveld, J.A. (1992) Atmospheric Environment 26A, 445-464. Baldwin, A.C. and Golden D.M. (1979) Science 206, 562-563. Bobbink, R., Boxman, D., Fremstad, E., Heil, G., Houdijk, A. and Roelofs, J. (1992) In: Grennfelt, P. and Th6mel6f, E.: Critical loads for nitrogen, Report Nord 1992:41, Nordic Council of Ministers, Copenhagen. Buijsman, E. and Erisman, J.W. (1988) J. Atmos Chem. 6, 265-280. Buijsman, E., Maas, J.F.M. and Asman, W.A.H. (1987) Atmospheric Environment 21, 10091022. Cass, G.R., Gharib, S., Peterson, M. and Tilden, J.W. (1982) Open File Report 82-6. Environmental Quality Laboratory, California Institute of Technology, U.S.A. Davidson, C.I. and Wu, Y.-L. (1990) In: Lindberg, S.E., Page, A.L. and Norton, S.A. (Eds.) Acidic precipitation. Vol. 3. Sources, deposition, and canopy interactions, Springer, New York, U.S.A., 103-215. Dentener, F.J. and Crutzen, P.J. (1993) A three dimensional model of the global ammonia cycle. Paper submitted to J. Atmos. Chem.
69 Duyzer, J.H., Bouman, A.M.H., Diederen, H.S.M.A. and Van Aalst, R.M. (1987) Report R 87/273, TNO Division for Society, Delft, The Netherlands. Duyzer, J.H., Verhagen, H.L.M., Westrate, J.H. and Bosveld, F.C. (1992) Environ. Pollut. 75, 3-13. Erisman, J.W., Vermetten, A.W.M., Asman, W.A.H., Waijers-Ypelaan, A. and Slanina, J. (1988) Atmospheric Environment 22, 1153-1160. Erisman, J.W. and Wyers, G.P. (1993) Atmospheric Environment 27A, 1937-1949. Farquhar, G.D., Wetselaar, R. and Weir, B. (1983) Gaseous nitrogen losses from plants. In: Freney, J.R. and Simpson, J.R." Gaseous loss of nitrogen from plant-soil systems. Nijhoff, The Hague, The Netherlands, 159-180. Fowler, D. and Duyzer, J.H. (1989) In: Andrae, M.O. and Schimel, D.S. (Eds.): Exchange of trace gases between terrestrial ecosystems and the atmosphere, John Wiley, New York, U.S.A., 763-773. Harper, L.A., Sharpe, R.R., Langdale, G.E. and Giddens, J.E. (1987) Agron. J. 79, 965-973. Harrison, R.M. and Kitto, A.-M.N. (1992) J. Atm. Chem. 15, 133-143. Hettelingh, J.-P., Downing, R.J. and De Smet, P.A.M. (1991) Technical Report No. 1, Report 259101001, National Institute of Public Health and Environmental Protection, Bilthoven, The Netherlands. Huntzicker, J.J., Cary, R.A. and Ling, C.-S. (1980) Environ. Sci. Technol. 14, 819-824. Hutchinson G.L., Millington R.J. and Peters D.B. (1972) Science 175, 771-772. Htifken, K.D., Meixner, F.X. and Ehhalt, D.H. (1983) In: Pruppacher, H.R., Semonin, R.G. and Slinn, W.G.N. (Eds.) Precipitation scavenging, dry deposition, and resuspension, Elsevier, New York, U.S.A., 825-835. Isermann, K. (1990) In: Ammoniak in der Umwelt, KTBL, Darmstadt-Kranichstein, F.R.G., 1-1 to 1-76. Isermann, K. (1993) Nahrstoffbilanzen und aktuelle N~.hrstoffversorgung der BSden. RobertBosch-Stiftung GmbH, 5. Koloquium zur Bodennutzung und Bodenfruchtbarkeit: "N~flarstoffhaushalt - Kenntnisstand und Forschungsliicken", Schw~ibisch Hall, 21-22 November 1991, Sonderband Berichte iiber Landwirtschaft. Langford, A.O. and Fehsenfeld,F.C. (1992) Science 255, 581-583. Larsen, S.E., HummelshCj,P. Jensen, N.O., Edson,J.B., De Leeuw, G. and Mestayer, P.G. (1994) Report 47 Danish Sea Research Programme 90, Danish Environmental Protection Agency, Copenhagen, Denmark. Lee, D.S., Nason, P.D. and Bennett, S.L. (1992) Report AEA-EE-0328, AEA Environment and Energy, Harwell Laboratory, U.K. Lenhard, U. and Gravenhorst, G. (1980) Tellus 32, 48-55. Levine, J.S., Augustsson, T.R. and Hoell, J.M. (1980) Geophys. Res. Lett. 7, 317-320. Lindfors, V., Joffre, S.M. and Damski, J. (1991) FMI Contribution No. 4, Finnish Meteorological Insitute, Helsinki, Finland. Logan, J.A., Prather, M.J., Wofsy, S.C. and McElroy, M.B. (1981) J. Geophys. Res. 86, 72107254. McMurry, P.H., Takano, H. and Anderson, G.R. (1983) Environ. Sci. Technol. 17, 347-352. Morgan, J. A. and Parton, W.J. (1989) Crop. Sci. 29, 726-731. Nilsson, J. and Grennfelt, P. (1988) Report 1988:15, Nordic Council of Ministers, Copenhagen, Denmark. O'Deen, W.A. (1989) Agron. J. 81,980-985. Pacyna, J.M., Larssen, S. and Semb, A. (1991) Atmospheric Environment 25A, 425-439.
70
Patton, W.J., Morgan, J.A., Altenhofen, J.M. and Harper, L.A. (1988) Agron. J. 80, 419-425. Pio, C.A. and Harrison, R.M. (1987) Atmospheric Environment 21, 1243-1246. Pruppacher, H.R. and Klett, J.D. (1978) Microphysics of clouds and precipitation. Reidel, Dordrecht, The Netherlands. Quinn, P.K., Bates, T.S., Johnson, J.E., Covert, D.S and Charlson, R.J. (1990) J. Geophys. Res. 95, 16405-16416. Quinn, P.K., Charlson, R.J. and Bates, T.S. (1988a) Nature 335, 336-338. Quinn, P.K., Charlson, R.J. and Zoller, W.H. (1988b) Tellus 39B, 413-425. Robbins, R.C. and Cadle, R.D. (1958) Phys. Chem. 62, 469-471. Roelofs, J.G.M., Kempers, A.J., Houdijk, A.L.F.M. and Jansen, J. (1985) Plant and Soil 84, 45-56. Schaug, J., Pedersen, U., Skjelmoen, J.E. and Kvalv~tgnes, I. (1993) Data report 1991. EMEP/CCC-report 4/93, Norwegian Institute for Air Research, LillestrCm, Norway. Schjcrring, J.K. (1991) In: Sharkey T.D., Mooney, H.A. and Hollamd, E.A. (Eds.): Trace gas emissions by plants. Academic Press, New York, U.S.A., 267-292. Schj0rring, J.K. Kyllingsb~ek, A., Mortenssen, J.V. and Byskov-Nielsen, S. (1993) Plant, Cell and Environ. 16, 161-167. Schlesinger, W.H. and Hartley, A.E. (1992) Biogeochemistry 15, 191-211. Stelson, A.W., Friedlander, S.K. and Seinfeld, J.H. (1979) Atmospheric Environment 13, 369371. Stelson, A.W. and Seinfeld, J.H. (1982a) Atmospheric Environment 16, 903-922. Stelson, A.W. and Seinfeld, J.H. (1982b) Atmospheric Environment 16, 993-1000. Stelson, A.W. and Seinfeld, J.H. (1982c) Atmospheric Environment 16, 2507-2514. Sutton, M.A., Asman, W.A.H. and Schjorring, J.K. (1993a) In: L6vblad G., Erisman, J.W. and Fowler, D. (Eds.): Report 1993:573, Nordic Council of Ministers, Copenhagen, Denmark, 127143. Sutton, M.A., Fowler, D., Hargreaved, K.J. and Storeton-West, R.L. (1992b) In: Angeletti, G., Beilke, S. and Slanina, J. (Eds.) Field measurements and interpretation of species related to acid deposition. Air Pollution report 39, Commission for the European Communities, Brussels, Belgium, 211-217. Sutton, M.A., Fowler, D., Smith, R.I., Eager, M. Place, C.J. and Asman, W.A.H. (1993b) Proceedings of the joint CEC/BIATEX workshop, Aveiro, Portugal, May 1993. Commission for the European Communities, Brussels, Belgium, 117-131. Sutton, M.A., Moncrieff, J.B. and Fowler (1992a) Environ. Pollut. 75, 15-24. Sutton, M.A., Pitcaim, C.E.R. and Fowler, D. (1993c) Adv. Ecol. Res. 24, 301-393. Van Breemen, N., Burrough, P.A., Velthorst, E.J., Van Dobben, H.F., De Wit, T., Ridder, T.B. and Reijnders, H.F. (1982) Nature 299, 548-550. Van der Eerden, L.J.M. (1982) Agric. Envir. 7, 223-235. Van Hove, L.W.A., Koops, A.J., Adema, E.H., Vredenburg, W.J. and Pieters, G.A. (1987) Atmospheric Environment 21, 1759-1763. Williams, R.M. (1982) Atmospheric Environment 16, 1933-1938. Wyers, G.P., Vermeulen, A.T. and Slanina, J. (1992) Environ. Pollut. 75, 25-28.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
71
Measurement and modelling of ammonia exchange over arable croplands M.A. Sutton, J.K. Burkhardt, D. Guerin and D. Fowler Institute of Terrestrial Ecology, Bush Estate, Penicuick Midlothian, EH26 0QB, Scotland, UK. Abstract
Micrometeorological measurements of the exchange of atmospheric ammonia over arable land are reported. Measurements were made over bare soil and wheat canopies at early canopy closure, approaching anthesis and during early senescence. Bare soil showed a slow rate of deposition with a canopy resistance (Rc) of around 300 s m -1, while fluxes over young wheat were still dominated by fertilizer urea soil emissions (up to 170 ng m -2 s -1) nearly a month after fertilization. Summer measurements showed that vegetation processes dominated exchange, apart from immediately after fertilization with ammonium nitrate, which contributed around 5 ng m -2 s -1 to the net flux 1-2 days after application. The Summer NH 3 fluxes have been interpreted using a new micrometeorological modelling approach. The R c model is unable to simulate the mechanisms of bi-directional exchange, and a recently developed model that quantifies the 'canopy compensation point' (gc) for NH 3 was applied. In this model net fluxes are predicted as the resolution of competing leaf surface deposition and bi-directional stomatal fluxes. The model provides a simple approach to predict net NH 3 fluxes with the atmosphere, though sometimes underestimates morning NH 3 emissions. A possible explanation of this effect is that the leaf surface behaves as a capacitor for N H 3 adsorption. A revised capacitance model is developed that extends the analysis of the Xc model. The capacitance model is able to reproduce the behaviour of morning emission, though further development of both models is required to provide a description valid over longer periods. 1. I N T R O D U C T I O N
Deposition of atmospheric ammonia (NH3) and ammonium (NH4+) (collectively NHx) is an important contributor to ecosystem acidification and eutrophication (e.g. Grennfelt and Thtirneltif 1992; Sutton et al. 1993d). It is necessary to quantify the inputs of these species to sensitive ecosystems, to be able to assess their environmental impact as well as the contribution of NH x relative to other acidifying inputs. One of the main uncertainties in quantifying total inputs of NH x is the magnitude of the gaseous N H 3 dry deposition term. The uncertainty is emphasized because NH 3 may be both emitted from and deposited to land surfaces. The atmospheric nitrogen inputs to arable land may be small compared with agricultural practice. However, because of the large area extent of arable land in Europe it becomes important to quantify the net fluxes (emission and deposition) in calculating atmospheric budgets and parametrizing atmospheric transport models (e.g. Sandnes and Styve 1992; Asman and van Jaarsveld 1992; Singles et al. 1995). In this way knowledge of crop-atmosphere exchange becomes important in understanding the N inputs to semi-natural ecosystems. Measurements of NH 3 exchange over cereals have often implied that a 'compensation point' exists within plants andis important in controlling net fluxes (e.g. Farquhar et al. 1980; Schjorring 1991; Sutton et al. 1993c). It is known that ammonia plays a major role within the biochemical pathways of plants, so that for a particular intercellular NH4+ concentration, there is an equilibrium atmospheric gaseous N H 3 concentration. The term 'compensation point' is used to reflect the interpretation that this is the concentration at which metabolic consumption processes balance production, while the exchange is viewed as operating via stomata. This is referred to here as thej 'stomatal compensation point' (Zs)- Against this physiological background, it is also known that ammonia is a very soluble gas and is frequently found in micrometeorological experiments to deposit rapidly to semi-natural vegetation and this is believed to be the result of leaf surface sorption processes (Duyzer et al. 1994; Sutton et al. 1993b; Erisman and Wyers 1993). Ammonia emission from a compensation point by therefore be short circuited by deposition to leaf cuticles.
72 Arable croplands to show evidence of both processes operating, resulting in the direction of the net atmospheric flux changing with environmental and plant conditions. In part, this may be related to the higher nitrogen status of these ecosystems, resulting in a larger compensation point and potential to overcome cuticular uptake. Until recently, model descriptions of these processes had taken one of two lines: either to assume a solely stomatal exchange with a compensation point (mostly the chamber experiments); or to treat the exchange with a deposition velocity (Vd) and canopy resistance (Rc) model, originally designed for parametrizing deposition processes (mostly the micrometeorological experiments). A combined model to reconcile these different interpretations was proposed by Sutton and Fowler (1993). In this model, calculation of R c is replaced by its concentration analogue (Zc), referred to as the 'canopy compensation point', the magnitude of which is defined by competition for exchange between the atmosphere, cuticle and stomata. The present paper reports measurements of ammonia surface-atmosphere exchange, made using the aerodynamic gradient technique. Fluxes were measured at several stages during the year, including over bare soil and over different wheat canopies at different stages. In addition to these main measurements, a short chamber study was made to provide an indication of the contribution of soil to the net atmospheric emission from a wheat canopy. Where the plant canopy was established as being the main site of ammonia surface-atmosphere exchange, particular attention was given to investigating the processes controlling the net ammonia flux. The results are interpreted using the canopy compensation point model of Sutton and Fowler (1993) and constraints of the model identified. In particular the possibility that the leaf surface may behave as a capacitor for NH 3 adsorption/dissolution is investigated and an initial framework to parametrize this effect suggested. 2. MICROMETEOROLOGICAL THEORY
A full description of the micrometeorology applied in this study and its restrictions has been given by Sutton et al. (1993a, b), so only an outline description is given here. The flux measurements were made using the aerodynamic gradient method. In this approach the net flux (Ft) (negative fluxes denote deposition) is determined from wind (u) and concentration (Z) profiles with height above a uniform surface with extensive fetch: Ft = _ k 2
du d[ln(z-d)-~M]
dz d[ln(z-d)-~tu]
(1)
where z is height above ground, d is the displacement of ground level due vegetation, k is the von Karman constant (0.41) and ~tM, XCHare corrections for atmospheric stability. As noted in the introduction it is usual to interpret measured fluxes by calculating the deposition velocity (Vd) and canopy resistance (Rc). In principle this resistance approach is designed to describe deposition processes, since it assumes that the concentration at the absorbing surface (~surface) is zero. In this case, by analogy with Ohm's law:
Vd{z-d} = 1/Rt{z-d} = Ft/(~surfac e - ~{z-d}) = -Fg/~{z-d}
(2)
Rt{z-d} = Ra{z-d} + R b + R c
(3)
where R t is the total resistance to deposition, R a the turbulent atmosphere resistance, R b the quasi-laminar diffusion resistance. This approach is convenient since Vd is assumed to be independent of Z, and the flux may be modelled given estimates of R c and Z, together with meteorological information. Nevertheless, it is not possible without manual switches to use the model to predict bi-directional exchange. The alternative stomatal model noted above describes bi-directional fluxes satisfactorily, but ignores any parallel deposition to leaf surfaces. Here Zs is the stomatal compensation point and R s the stomatal resistance: Ft = (ks - Z{z-d})/(Ra{z-d} + Rb + Rs)
(4)
73 180" chamber over bare soil between plants
150" 120"
Flux NH3
90
r
(ng m-2 s-1) 60 30 0 -30
,:'~
160
,o
ii
140
8
!~
12
Air conc (1 m) NH3 (ug m-3)
6
t
i
12o Air conc (1 m) lOO SO2 (ug m-3) so
',
:: i i
NH3
4
i~:
6O
i
40 20 0
' " "'"
0
00:00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00
Time (GMT) 2 April 1993
Figure 1. Micrometeorologicalammonia flux measurements and air concentrations of NH3 and SO2 over a young wheat canopy (14 cm). Measurementsmade 1 month after fertilization with 35 kg N ha-1 as urea. Open squares are chamber measurementsof soil emission. Given the limitations of both these approaches, in later sections the results are used to develop revised resistance models that treat both a compensation point and cuticular uptake. 3. SITE AND M E A S U R E M E N T S The measurements were made over bare soil and canopies of wheat at the field station of the University of Nottingham at Sutton-Bonington, Loughborough, UK. The site used provided extensive fetch of >200 m in most wind sectors and up to 500 m. Measurements were made in the surface layer up to 2 m above the wheat canopy. In May 1992, measurements were made above a wheat canopy (60-70 cm) approaching senescence. On May 20, 35 kg N ha -1 as ammonium nitrate was applied. A second campaign was made in March 1993, on the boundary between a bare field (harrowed soil) and a young wheat canopy (14 cm high). This wheat canopy had been fertilized with 35 kg N ha -1 as urea about 4 weeks prior to the measurements. A third campaign was made in June - July 1993, at the same site, when the wheat was approaching senescence (80 cm high), and sugar beet growing on the previously bare soil. Ammonia concentration profiles were determined using a continuous wet annular denuder system described by Wyers et al. (1993). Measurements were made at two heights, and, to reduce possible bias between denuders, the collected NH 3 was brought (as NH4+ in the stripping solution) to a common detector for analysis. Analysis of the NH4 + in this system is by membrane diffusion of NH 3 at high pH into a counter flow of deionized water, with subsequent measurement by conductivity. Sulphur dioxide air concentrations were also determined though fluxes of SO 2 are not reported here. Windspeed was measured using 6 sensitive cup anemometers (Vector instruments), and temperature and water vapour profiles measured with a fine thermocouple and dewpoint meter system. Further details of the measurement methods and data treatment have been provided by Sutton et al. (1993a). 4. RESULTS Figure 1 shows an example of the emissions detected during the Spring 1993 measurements, with fluxes up to 160 ng NH 3 m -2 s-1 and only limited net deposition occurring at night. Such large emissions are probably related to the fertilization of the soil with urea 1 month prior to the measurements. It is well known that substantial ammonia emissions occur from urea compared with other fertilizers (e.g. Whitehead and Raistrick 1990), however, it is generally
74 25 20 15
Flux NH3 (ng m-2 s-l)
10 5 0 -5 -10 -1520O 180 5
160
Air conc (lm) 4 NH3 3 (ug m-3)
14o Air conc
12o ~oo
(1 m)
SO2 (ug m-3)
6O 4O
"'!.,...':...~ 0 12:00
2O 0
16:00
20:00
O0:00
04:00
08:00
12:00
16:00
20:00
Time (GMT) 21-22 May 1992
Figure 2. Micrometeorological ammonia flux measurements and air concentrations of NH3 and $ O 2 o v e r a wheat canopy prior to anthesis. The soil was fertilized with 35 kg N ha-1 as ammonium nitrate on 20 May. generally assumed that this soil emission is complete within 1-2 weeks of fertilization, after which vegetation processes dominate exchange. The hypothesis that the emissions derived from the soil was tested using a simple chamber placed over bare soil between the plants. The flux was derived from the difference of the inlet and outlet concentrations with a flow rate of 30 1 min -1. Given the small area of the chamber (0.07 m2), possible patchiness in emissions over the field, and potential adsorption of NH 3 to chamber walls, this test was only expected to give a broad indication of the soil flux. The results, presented in Figure 1, show a very close agreement with the micrometeorological estimates. This high level of similarity is almost certainly fortuitous, though clearly demonstrates that the soil was the main source of emission. During the same campaign, measurements were made of the deposition rate to bare agricultural soil (pH 6-7). This showed a reasonably consistent pattern of deposition at a slow rate, with good data indicating an R c of around 300 s m-1. The measurements in May 1992 were made before and after fertilization with ammonium nitrate. It is known from laboratory studies that emissions from this fertilizer are expected to be small (e.g. 1 % applied N). Measurements before fertilization (1 month after previous fertilization with ammonium nitrate) showed no detectable soil flux, with vegetation dominating the exchange process. In contrast to the April 1992 measurements, only a very small soil emission was detected following fertilization. The results are shown in Figure 2. In addition to day time emissions, they show a small but significant (5 ng m -2 s-1) net emission during night time. Since stomata are closed during the night this is attributed to soil emission (supported by within-canopy concentration profiles). In both Figures 1. and 2., SO 2 concentrations are plotted against the NH 3 concentrations. As previously shown by Sutton et al. (1993a), there is an inverse relationship between the concentrations of these 2 gases. This may be related to formation of ammonium sulphate aerosol depleting the concentration of one of the gases in the presence of an excess of the other. 5. I N T E R P R E T A T I O N O F R E S U L T S . In order to simulate explain the bi-directional fluxes observed over the wheat canopy a modelling framework is required that can treat both NH 3 sources (compensation point, soil emission) and sinks (cuticular uptake, stomatal uptake) and their interaction. In particular,
75
II
II
R~
R~
F~
F~
LT Rw
IT
Rs
Kr~FQd _-z"
>0
A.
Cd
--
Rs
s>0
B.
Figure 3. Model diagrams to describe the bi-directional exchange of ammonia between a plant canopy and the atmosphere. The net flux is described as the result of the exchange between the air concentration (Z) and the 'canopy compensation point' (Zc,), which is the result of competition between exchange to the leaf surface and transfer through stomata (Rs) with a stomatal compensation point (Zs)- Deposition to the leaf surface is treated in different ways: A, as a simple resistance (Rw); B. as an adsorption capacitance (Cd) of given charge (Qd) with adsorption resistance (Rd) and a reaction rate (Kr). R a and Rb are the resistances to atmospheric turbulence. attention is given here to addressing the interaction between a stomatal compensation point, leaf cuticle uptake processes and atmospheric NH 3 concentrations. 5.1. Canopy compensation p o i n t , cuticular resistance model A resistance diagram of the model of ammonia canopy exchange proposed by Sutton and Fowler (1993) is shown in Figure 3a. The model may be used as a tool either to interpret measured net fluxes or to infer the net flux given particular conditions. The difference between this and the of model Eq. (3) is that here the canopy compensation point (2c) is calculated rather than the canopy resistance (Rc). As with R c (Eq. 3), Zc may be directly estimated from micrometeorological measurements:
~,c = ~,{ z-d} + Ft(R a{ z-d} + Rb)
(5)
In principle the total flux F t must be conserved in its component fluxes to the cuticle/water layers (Fw) and with stomata (Fs) so that: Ft = Fw + Fs Fw =-Zc/Rw;
(6) Fs = (Zs - )~c)/Rs
(7a, 7b)
Hence, having determined F t and Xc from measurements, and from measured (or modelled) R s, the value of R w may be determined given different assumed values of )~s- Alternatively, given an estimate of R w, ~s may be found. The presence of two unknowns points to the need for independent laboratory determinations of these parameters (e.g.)~s). In contrast, the model may be used in an inferential manner to explicitly estimate Zc and F t. Substitution of Eqs. (7a,b) into Eq. (6) and rearrangement provides: F t = (Zs-Zc)/Rs - Zc/Rw This may then be combined with Eq. (5) to eliminate Ft:
(8)
76 20 10
.,~176
0 Flux NH3
-10
obs
(ng m-2 s-l)
. . . . . . . model
-20 -30
7
-40 19 May 1992 -50 00:00'
30
04:00
'
lz':oo
08:00
'
16':oo
zo:~o
oo:oo
20 10 Flux NH3 (ng m-2 's-l) 0
. . . .
-10
IIli
-20 ~30
iI ] 21-22 May 1992 I
-40 12:00
16:00
20:00
00:00
,++:I'-
~
obs model (with soil emission)
....
model (with no soil emission)
04:00
08:00
12:00
16:00
20:00
Time (GMT)
Figure 4. Comparison of measured ammonia flux over wheat and that predicted by the simple (Rw) canopy compensation point model. Rw calculated according to Eq. (12) and assuming plant intercellular pH 6.8 and 100 ~rnol NH4+ 1-1. Model soil emission in the Zc calculation (Eq. 13) for 21-22 May set at 10ng m-2 s-1. %s/%c - 1 Rs
1 Rw
Zs
1
1
ZcRs
Rs
Rw
1-Z{z-d}/Zc R a {z - d } + R b
1 Ra{z-d}+R b
(9)
Z{z-d} Zc(Ra{z-d}+Rb)
(10)
From which the canopy compensation point may be given as:
[Z{z-di/(Ra {z-di+Rb)+Zs/Rs] Zc - [ ( R a { z _ a i + R b ) _ 1 + R s - 1 +Rw_l]
(11)
and the flux found from Eq. (5). Examples of the application of this model are shown in Figure 4. On the basis of comparison with laboratory studies of the relative humidity (RH) response of adsorption, a simple parametrization of R w (s m -1) was used (Sutton and Fowler 1993)" R w = 2 exp ([ 100-RH]/12)
(12)
The stomatal compensation point was calculated based on the temperature dependent Henry equilibrium using pH of 6.8 (Farquhar et al. 1980) and assuming 100 gmol NH4 + 1-1 in the leaf intercellular fluid. The latter was chosen to fit the data in Figure 4a, and is consistent with other published estimates (Farquhar et al. 1980; Sutton et al. 1993c). Applying these values to the measurements after fertilization with ammonium nitrate (Figure 4b) significantly underestimated the flux. This may be due to daily variations in R w but is also likely to be related to the presence of an additional soil emission, as indicated by the night time
77 measurements. It is possible to provide to provide a simple treatment of soil emission in the canopy model:
Zc = [)~{z-di/(Ra { z - d } +Rb)+ )Cs/Rs + Fsoil ] [ ~ 7 { z - d } + Rb)-l+Rs-l+Rw-1]
(13)
The effect of adding a soil flux in this manner is shown in Figure 4b, indicating that the daytime emissions would be consistent with an additional emission into the canopy space of 10 ng m -2 s -1. The difference between the two model estimates is not constant because of varying recapture by leaf surfaces and stomata. It should be noted, however, that introducing a soil emission in this simple way is unlikely to give precise results, because of the different physical location of the soil to the vegetation canopy. As a single layer model, it is assumed that all the exchange occurs at a single hypothetical height, which would be expected to provide errors for soil exchange processes.
5.2. Canopy compensation point- cutieular capacitance model Although it is possible to provide reasonable agreement with the measured data using the simple canopy compensation point model, the measured results sometimes show much larger emission in the morning than predicted from a temperature dependent compensation point, while emission in the evening may be smaller than be expected. Two effects that may explain this are that the concentration of NH4+ in leaves is not constant throughout the day, and/or that the leaf cuticle acts more like a capacitor for adsorption of ammonia. There is some recent support from laboratory studies that the concentration of NH4+ may vary diurnally, with larger values in the early morning (Schjc~rring, 1994, pers. comm.), and further work is required to examine this possibility and its cause. However, it is equally possible that during increasing humidity conditions the leaf surface will be able to absorb more NH 3 than when the surface is drying. In the latter case, if deposited NH4+ is not 'fixed' by reaction to form salts with low vapour pressure (e.g. [NH412SO4), it may be released back as NH 3 and contribute to net emission. An model to explore this behaviour ammonia on leaf surfaces is shown in Figure 3b. A number of studies have investigated the link between relative humidity and thickness of notional 'water-films' on leaf surfaces (Van Hove et al. 1988; Benner et al. 1992; Burkhardt and Eiden 1994), which, for a given pH, would be expected to have a defined capacitance according to Henry's law. The relation defining the capacitance (Cd) may be expressed as:
Cd = Qdl)~d
(14)
where Qd is the adsorption charge (lag m -2) and )~d the adsorption concentration (l.tg m -3) associated with the capacitor. By analogy to this relationship, an estimate of C d may be found from the Henry equilibrium constant and an equivalent water-film thickness (MH20). Using the solubility equilibria provided by Sutton et al. (1993d) gives:
Here Cd and MH20 are given in metres and T in Kelvin. On the basis of measurements on polluted leaves it is estimated that a typical value of MH20 would be 20 nm at 60% (Burkhardt, unpublished data). Using a similar humidity response to Eq. (12) and accounting for the canopy leaf area index (LA/), an initial estimate of MH20 was found as: MH20 = LM * 20 exp ([RH-60]/10)
(16)
Unlike straightforward resistance models, treating the leaf surface exchange process as a capacitance results in the flux at a given time being time dependent on previous fluxes. Hence calculation of the modelled is linked over different model time steps. It is necessary to set an initial value of either Qd or )(;d, so that for an initial time (i): )~d{i} = Qdlii/Cd
(17)
78 The flux into or out of the adsorption capacitor (Fd) is then: F d = (Zd{i}-Zc{i})/Rd
(18)
where R d is the charging resistance of the capacitor. The value of Zc{i] may be found by applying the canopy compensation point equation (Eq. 11) in slightly modified form:
[~{z-d}/(Ra {z-d} Zc =
+ R b) + Zs/Rs + zd/Ra] [(Ra{z-dI+Rb) -1 +Rs -1 +Rd -1]
(19)
The new capacitance charge (Qd{i+t}) after t seconds is then found as: Qd{i+t} = Qd{i} - (Fd.t) (20) Given the new value of Qd a revised value of Zd is found according to Eq. (17), and the process repeated for the next time step. While this parametrization will treat adsorption and desorption to cuticular water-layers, it does not provide for any net removal of NH 3 from the air by leaf surfaces. This may be accounted for by proposing a reaction flux (F r) of the stored NH4+ (e.g. to form ammonium sulphates) with a rate constant Kr (s-l):
Fr = ad. Xr where a negative value of Kr indicates deposition. Inclusion of F r into Eq. (20) provides: Qd{i+t} = Qd{i} - (Fd.t) + Fr (21) In running this model it is found that the value of R d defines the rate of charging of Qd and provides the time constant of the adsorption/desorption process. It is anticipated that R d should be larger in dry conditions (reduced access to the water layer), and this is also required to make the model run, since the model requires smaller time steps to remain stable with decreasing values of R d and C d. The initial results presented here are calculated using R d (s m -1) = 5000/C d (m). This is equivalent to a time constant of 83 minutes. An example of the application of this model is shown in Figure 5, for the measurements on 3-4 July 1993 over mature wheat (early leaf senescence). The measured fluxes on the 4 July showed a larger emission in the morning and less in the evening than would be expected by a compensation point emission through stomata. In this case the capacitance model is run assuming a leaf surface pH of 4.5 and an adsorbed NH4+ reaction rate (Kr) of-0.01 s-1. The effect of including K r is shown on the value of Qd" Estimates of the modelled flux are provided, one just accounting for the capacitance effect, and the second linking this with the compensation point exchange. The model is able to predict successfully the peak of emission on the morning of 4 July, though daily differences related to untreated factors also occur since the flux on 3 July is not well represented. The comparison between the two model flux estimates in Figure 5, shows that stomatal uptake of N H 3 desorbed by the cuticle reduces the peak emission, while later on in the day stomatal exchange contributes to the modelled net emission. 6. DISCUSSION AND CONCLUSIONS Measurements of NH 3 surface atmosphere-exchange over arable croplands show that each of stomatal, leaf surface and soil exchange processes are important in def'ming net fluxes. Where fertilizer nitrogen is added as ammonium nitrate, soil emissions are small and contributed here about 5 ng m -2 s -1 to the net flux 1-2 days after fertilization. In contrast, fertilization with urea provides much larger ammonia emissions. Chamber measurements made here supported the interpretation that the emissions were soil rather than vegetation related. An important finding was that the enhanced ammonia flux following urea fertilization may continue long after fertilization (4 weeks). Previous studies have often considered soil emissions complete after 12 weeks and may have underestimated urea emissions. Emission of ammonia from urea is a result of its hydrolysis to by urease producing ammonia and raising solution pH. In the example reported there had been tittle rain between fertilization and the measurements reported, suggesting that hydrolysis was slow. In contrast to these results, measurements over
79
2500
700 ,.,-, ..," ",
2000 Adsorption 15oo Capacitance Cd (m)
""
,<'-'C','"
1000
]~
\':
~."
l~ ~ II ! I .'".
t:
.tlo.....,l" ;
I"
::k" . . . . J/ .~'_'~1 - - \
~"
500
~,.,,^
_J \
V,
.s
ef
I 600 500 Qd (Kr = -0.01 s-1)l J 4oo
Cd . . . . . . . Qd (Kr = 0 s-l)
J
....
3oo
\"
. 200
~',
I
I
~;
k '<,,
I
Adsorption Charge Qd (ug m-2)
100 0
0 40 20 Flux NH3
0
(ng m-2 s-l) -20 -40 -60 -80 00:00 04:00 08:00 12:00 16:00 20:00 00:00 04:00 08:00 12:00 16:00 20:00 00:00 Time (GMr) 3-4 July 1993
Figure 5. Comparison of measured ammonia flux over a ripening wheat canopy (78 cm, early senescence) with model estimates parametrized using the capacitance only (Cd) description of NH3 uptake onto leaf surfaces, and the full model allowing for both Cd and stomatal exchange (R s, Zs). The model flux assumes leaf intercellular pH 6.8 and 50 lxmolNH4+ 11, leaf surface pH 4.5, with an equivalent water-film thickness and Cd according to Eqs. (15-16). The model fluxes are calculated applying an adsorption reaction rate (Kr) of-0.01 s-1, the effect of which is shown on the adsorption charge (Qd), alongside the modelled value of Cd. bare soil.(pH 6-7) suggested a slow rate of dry deposition with canopy resistances of the order 300 s m- I. Apart from immediately after fertilization with ammonium nitrate, Summer measurements over a growing wheat crop, , showed no significant soil emissions. In these cases a recently developed model for estimating NH 3 fluxes was tested. In contrast to usual models, which calculate canopy resistances (Rc), the present model calculates the canopy compensation point (Zc), which is found from the resolution of competing leaf cuticle and stomatal emission fluxes. Using a simple function of humidity to describe cuticular uptake, and assuming a stomatal compensation point dependent on leaf NH4 +, provided a reasonable estimate of the net flux. Inclusion of a soil emission in the model provided a better fit to the measurements made after ammonium nitrate fertilization, though in principle such a singlelayer model would be expected to be uncertain here because source-sink locations differ. One limitation of the simple canopy compensation point model is that it sometimes underestimates emissions in the morning and overestimates emission in the evening. This may be due to either temporal variability in the stomatal compensation point, or to the leaf surface behaving as a capacitor for NH 3 adsorption. A revised capacitance model was developed to provide an initial investigation into the second possibility. Modelled fluxes in the capacitance model are dependent on previous fluxes, which regulate the value of the adsorption charge (pg m -2) of NH 3. By itself the capacitance treatment does not result in a net removal of ammonia from the atmosphere, though this effect may be treated by allowing for reaction of the adsorbed NH4+ to form fixed salts. The model is only at a preliminary stage, and yet is able to predict accurately the elevated morning emission of ammonia. Other subtleties of the chemical processes and physiological controls will require further development of these approaches. The simole canoov compensation ]9oint model has fewer unknowns than the capacitance model, and is therefore to be preferred in principle. However, an examination of the fluxes in
80
more detail using the capacitance model provides a useful aid to understanding the processes which regulate the net exchange of ammonia between soil/vegetation and the atmosphere. 7. A C K N O W L E D G E M E N T S The authors are grateful for financial support from the U K Department of the Environment, Air Quality Research Programme and the Commission for the European Communities.
8. REFERENCES Asman W.A.H. and Jaarsveld J.A. van (1992) A variable-resolution transport model applied for NHx for Europe. Atmos. Environ. 26A, 445-464. Benner W.H., Ogorevc B. and Novakov T., 1992. Oxidation of SO2 in thin water films containing NH 3. Atmos. Environ. 26A, 1713-1723. Burkhardt J. and Eiden R. (1994) Thin water films on coniferous needles. Atmos. Environ. 28A, 2002-1019. Duzyer J.H., Verhagen H.L.M., Westrate J.H., Bosveld F.C. and Vermetten A.W.M. (1994) The dry deposition of ammonia onto a douglas fir forest in the Netherlands. Atmos. Environ. 28, 1241-1253. Erisman J.W. and Wyers G.P. (1993) Continuous measurements of surface exchange of SO2 and NH3: implications for their possible interaction in the deposition process. Atmos. Environ. 27A, 1937-1949. Farquhar G.D., Firth P.M., Wetselaar R. and Wier B. (1980) On the gaseous exchange of ammonia between leaves and the environment: determination of the ammonia compensation point. Plant Physiol. 66, 710-714. Grennfelt P. and ThOrnelOf E. (1992) Critical loads for nitrogen - a workshop report. ('Eds) Nord 1992:41, Nordic Council of Ministers, Copenhagen. Hove L.W.A. van, Adema E.H. and Vredenberg W.J. (1988) The uptake of atmospheric ammonia by leaves. In: Air pollution and ecosystems, fEd: Mathy P.) pp 734-738. CEC, Brussels/D. Reidel, Dordrecht. Sandnes H. and Styve H. (1992) Calculated budgets for airborne acidifying components in Europe, 1985, 1987, 1988, 1988, 1989, 1990 and 1991. EMEP report 1/92. Norwegian Meteor. Inst. Oslo, Norway. SchjCrring J.K. (1991) Ammonia emission from the foliage of growing plants. In: Trace gas emissions by plants. (Eds: Sharkey T.D., Holland E.A. and Mooney H.A.) pp 267-292. Academic Press, San Diego. Singles R., Sutton M.A. and Weston K.J. (1995) Fine resolution modelling of ammonia dry deposition in Great Britain. (This volume). Sutton M.A. and Fowler D. (1993) A model for inferring bi-directional fluxes of ammonia over plant canopies. In: Proceedings of the WMO conference on the measurement and modelling of atmospheric composition changes including pollutant transport. WMO/GAW-91. (Sofia 4-8 October 1993). WMO, Geneva. 179-182. Sutton M.A., Fowler D., Hargreaves K.J. and Storeton-West R.L. (1993a) Interactions of NH 3 and SO2 exchange inferred from simultaneous flux measurements over a wheat canopy. In: General Assessment of biogenic emissions and deposition of nitrogen compounds, sulphur compounds and oxidants in Europe. fEd: Slanina J., Angeletti G., and Beilke S.), pp 165-182. Air Pollut. Res. Report 47, CEC, Brussels. Sutton M.A., Fowler D. and Moncrieff J.B. (1993b) The exchange of atmospheric ammonia with vegetated surfaces. I: Unfertilized vegetation. Quart. J. Roy. Meteor. Soc. 119, 1023-1045. Sutton M.A., Fowler D., Moncrieff J.B. and Storeton-West R.L. (1993c) The exchange of atmospheric ammonia with vegetated surfaces. II: Fertilized vegetation. Quart. J. Roy. Meteor. Soc. 119, 1047-1070. Sutton M.A., Pitcairn C.E.R. and Fowler D. (1993d) The exchange of ammonia between the atmosphere and plant communities. Adv. Ecol. Research. 24, 301-393. Whitehead D.C. and Raistrick N. (1990) Ammonia volatilization from five nitrogen compounds used as fertilizers following surface application to soils of differing characteristics. J. Soil Sci. 41, 387-394. Wyers G.P., Otjes R.P. and Slanina J. (1993) A continuous-flow denuder for the measurement of ambient concentrations and surface-exchange of ammonia. Atmos. Environ. 27A.
G.J. He(j and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BE All rights reserved.
81
Preliminary validation of ammonia emission data using a combination of monitoring and modelling J.M.M. Aben a, P.S.C. Heuberger b, R.C.Acharya b and A.L.M.Dekkers b aAir Quality Research Laboratory, bCentre for Mathematical Methods, National Institute of Public Health and Environmental Protection, Postbox 1, 3720 BA Bilthoven, Netherlands
Abstract A method is presented to validate the ammonia emission data used as input in model calculations of ammonia concentrations. The method takes into account the uncertainty in model parameters and uses measured ammonia and ammonium concentration and wet deposition data. The calibrated emissions are subsequently used in the calculation of the spatial distribution of ammonia concentrations.
1. INTRODUCTION Acidification is regarded as one of the major environmental problems in the Netherlands. To protect vegetation and groundwater from adverse effect, the Dutch Government has set a limit for potential acid deposition. In 2010 this should not exceed 1400 mol ha -1 a -1. For forested areas 400 mol ha-l a-1 is the target[ 1]. The major contributing substances to potential acid deposition are oxidised sulphur and nitrogen compounds (SO x, NOy) and reduced nitrogen compounds (NHx). In the Netherlands with its intensive livestock breeding the latter contributes to a rather high degree. For 1993 it was estimated that the deposition of NH x contributed about 45% to the potential acid deposition in that year, being 4000 mol ha-1 a-1. The DEADM model [2] is used for the determination of the potential acid deposition. The distribution of wet deposition of SO x, NOy and NH x is determined by interpolation of the wet deposition values determined for the 15 fixed stations of the Dutch Precipitation Chemistry Network. The dry deposition is determined by inference; for each 2-hour period the concentration field is multiplied with the appropriate deposition velocity which varies with space and time because of its dependence on surface characteristics and meteorological conditions. The distribution of the yearly dry deposition is then obtained by integrating the 2hourly deposition fields. For the oxidised sulphur and nitrogen compounds the required concentration fields are obtained from continuous measurements at some thirty fixed stations of the Dutch Air Quality Monitoring Network. For NH 3 however a different route is followed. A very dense network of measuring points would be required for the accurate assessment of the spatial distribution of NH 3 by interpolation. This is due to the many local sources of NH 3 emission, the low emission height and the atmospheric chemical behaviour of NH 3. Moreover, only recently an operational method for monitoring NH 3 concentrations became available. Consequently, the spatial distribution is calculated with an atmospheric dispersion and deposition model (OPS; [3]) using spatial detailed emissions of NH 3 as input. The annual mean NH 3 and NH 4
82 concentrations calculated with the OPS-model are subsequently used in the DEADM model to calculate depositions. Though the application of the OPS model for NH 3 was validated by Asman & Van Jaarsveld [4] with data available those days, there has been no possibility until recently for routine validation of the model results because of the before mentioned lack of an operational monitoring method. Such a method has recently become available resulting in the establishment of an (interim) network for NH 3 mid 1992. In this paper measured and calculated values of NH 3 and NH 4 concentrations and of wet NH x deposition are compared. Subsequently, it is investigated whether the observed discrepancy between measured and calculated values of NH 3 can be explained by uncertainties in measured and/or calculated data. Lastly, a calibration procedure is applied in which the emissions are adapted in such a way that the deviations between measured and calculated values are minimised. Herewith the uncertainty in model parameters is taken into account. In this way information is obtained about the validity of the emissions.
2. MATERIALS AND METHODS 2.1. The monitoring network for NH 3 and NH 4
The monitoring network for NH 3 consists of 8 stations where the concentration is measured continuously with an annular denuder system [5]. Four of these stations are located in areas with high emission densities whereas the rest of the stations is located in regions with low to moderate emissions. Together with the establishment of the NH 3 network the number of stations where NH 4 aerosol is measured was increased from 2 to 5. Daily values of NH 4 aerosol are obtained by low volume sampling. In addition to concentration measurements wet deposition of NH 3 and NH 4 (together NHx) is determined at 15 stations in the Dutch Precipitation Chemistry Network. For all 3 components data for 1993 were used in this study. The location of the measurement stations is shown in Figure 1. The spatial distribution of emissions is also indicated in this figure. 2.2. Measurements on spatial representativity
Due to the very good characteristics of the measurement method employed [5] and the continuous monitoring, the annual mean values of NH 3 are regarded as very accurate values for the location where the samples are taken. However, when measurements are compared with model calculations the measured values should be representative for the grid cell surrounding the fixed station because the OPS model predicts averaged values for the 5 by 5 km area surrounding the receptor point. In order to determine the spatial representativity of the (annual) NH 3 concentrations, measurements with a van were carried out around the fixed points. The number of reference points chosen is dependent on the emission strength in the surrounding area, decreasing from 8 in strong emission areas, via 6 in moderate emission areas to 4 in background areas. All reference points belonging to a certain fixed point were sampled at the same day during 15-20 minutes. For each fixed station the reference measurements were repeated several times
83
4 Vm
Figure 1. Locations of stations and spatial distribution of emissions. emission (in tonnes): 0 -50 (617) 11150-100(293) I 100-150 (308) m 150-300 (377) 11>300(56) stations: [] NH3,NH4, wet deposition O NH3, wet deposition NH3 only /~ NH4 only V wet deposition only
throughout the year, the number of repeats being dependent on the emission strength in the surrounding area. 2.3. Calculation of NH 3 and NH 4 concentrations and wet deposition with OPS
OPS is a Lagrangian dispersion, conversion and transport model which calculates the concentration and deposition of primary and secondary components in a receptor point due to each of the emission sources separately. It is thus linear with respect to the emissions. The model uses national mean values for roughness length, deposition velocity and conversion rate. Normally, also national mean values for meteorological conditions are used but in this study regionalised meteorological conditions were used because of the substantial influence on predicted concentrations [6,7]. The diurnal emission pattern used here was that according to Acharya [7]. The emission data for the Netherlands used in this study are the official emission data for 1992 as described by van der Hoek [8]. The emission data for the European countries were obtained from Asman [9]. 2.4. Determination of the uncertainty of model predictions
One of the sources for uncertainty of the model results is the uncertainty in the model parameters. The uncertainty of the predicted NH 3 concentration was determined by simultaneous variation of parameters using the UNCSAM package [10]. The three parameters studied are the conversion rate of ammonia to ammonium (a), the surface resistance for ammonia (rc) and the scavenging efficiency for ammonia and ammonium (s).For this study it was assumed that the value of each of these parameters was distributed homogeneously over the uncertainty region. The default values applied in the model and the boundaries of the uncertainty regions are listed in Table 1.
84 Table 1 Default value and uncertainty boundaries of some model parameters parameter a (% per hour) rc
default
boundaries
28.8
[10,50]
30
[10,100]
1.10 6
[ 105,2.10 6]
(s m -1)
s (-)
2.5. The calibration method
Foreign emissions and industrial emissions were left out of the calibration procedure. The field of the remaining agricultural and household emissions was subsequently divided into 5 regions, 4 emission regions and the rest of the Netherlands. The determination of the emission regions was based on the amount of emission from manure per 5 by 5 km grid cell. Grid cells with emissions higher than 150 000 kg were selected. This resulted in 4 clusters of grid cells being the 4 emission regions (see Figure 2). Due to the linearity of the model with respect to the emissions the following equation applies to the model outcomes: Vcp = ~
"fl ) + ~
"f2 )+ ...... +~
"f5 ) + ~
) + ops(Ef )
(1)
where Vcv is the vector with model outcomes (NH 3 and NH 4 concentrations, NH x deposition) for each of the measurement locations and parameter set p, Eah,n is the vector of grid cell emissions for region n and fn is the emission calibration factor for region n with value 1 in the default situation. The calibration now consists of finding the values of fl to f5 which minimise:
= 51w.C - m/l
(2)
where V m is the vector with measured values and w is a vector with weighting factors which correct for the different order of magnitude between ammonia, ammonium and wet deposition (set at 1, 3 and 0.015 respectively in this study). This problem is solved analytically. If one of the values fl to f4 becomes negative, the corresponding region is added to the rest of the Netherlands and the analysis is started again. Because the model parameters are also uncertain, it is subsequently searched iteratively for the values of p which minimise C p. The calibration procedure is carried out in 2 ways. The first method uses all stations in the analysis (for wet deposition only those stations where also the ammonium and/or the ammonia concentrations are measured). The second method uses all stations but one and the calibration is repeated as many times as there are stations (m), each time leaving out another station (jack-knife method). This procedure yields m values for the calibration factors and the model parameters. The inner m - 2 values are averaged and the standard deviation of the m - 2
85 Figure 2. Partitioning of the Netherlands into 4 emission regions. The partitioning is based on the emission from manure per grid cell. r-1 Eibergen region [] Zegveld region I Vredepeel region I Lunteren region
values is calculated. The jack-knife procedure was used as a way to 'validate' the results of the 'all stations' procedure.
3. RESULTS
3.1. Comparison between measured and calculated values Figure 3 shows the comparison between measured and calculated values for the ammonium concentration, the ammonia concentration and for the wet deposition of NH x. Though the correlation between calculated and measured values of NH 3 is reasonable, the calculated values for the stations in emission areas are much lower than the annual means from the measurements. For NH 4 aerosol there is hardly any correlation between calculated and
Figure 3. Comparison between measured and calculated values before calibration. predicted (pg/m3) 25
NH3
8
20 .
10
'o0
~*, 5
9
9
42
9
;0
9
1200-
9
9
.i 9
NH4
6
.
15
predicted (pg/m3)
predicted (pg/m3)
9
;5
2'0
2'5
measured (pg/m3)
0
.
. o."
,
2
.,,
.
.. o~
wetdepositi0n
900-
~
~
~
measured (pg/m3)
600-
. "~o
300-
9
o
9
0
,
,
3~0 600 900 12'00 measured (pg/m3)
86 measured values and calculated values are much higher than the measured values. For wet deposition of NH x there is a fairly good correlation between calculated and measured values. However, the calculated values are somewhat lower than the measured values over the entire range of measured values. It should be mentioned here that the poor correlation for NH 4 aerosol and the apparent underestimation of wet deposition of NH x may be caused by erroneous values for the measured data. Unlike NH 3 (next section) there has been no investigation of the reliability of measured values of NH 4 and wet NH x deposition.
3.2. Spatial representativity of measured values of NH 3 Figure 4 shows for each station the mean value of all reference measurements (averaged over space and time) and for comparison the mean value of all simultaneous measurements at the fixed point. Also indicated are the standard errors of the mean. From these measurements there is no indication of significant local influences on the fixed point. However, especially for the stations in areas with high emission density (131, 722 and 734) conditions with enhanced concentrations seem to be under represented in the reference measurements. This becomes evident when the mean value of the reference measurements at the fixed point is compared with the annual mean value of the continuous measurements at the fixed point. using the same time window (11.00-19.00 h) as with the reference measurements to avoid diurnal variation effects.
3.3 Uncertainty in the model calculations of NH 3 The results of the UNCSAM analysis are shown in Figure 5. The dots represent the complete range of the model outcomes. The outer values of this range are less probable than the inner values. It can be concluded that although the uncertainties in the three model parameters studied give rise to an appreciable uncertainty in the predicted NH 3 concentration, the gap between predicted and measured values can not be closed by the uncertainty in the parameters only.
Figure 4. Comparison between mean value of measurements made at reference points and mean value of simultaneous measurements at fixed point. The annual mean at the fixed point is also indicated.
concentration (pg/m3) 20reference
16-
~
fixed (sameperiodsas reference) fixed (wholeyear/11.00-19.00h)
128-
0
i
!
!
540
235
928
/
538
!
|
|
|
633
722
131
734
station number
87
)redicted NH3 concentration (pg/m3) 252015-
"1 I
t .
10-
.111|
o o
1'0
l's
2'0
Figure 5. Uncertainty in the calculated NH 3 concentrations due to uncertainty in 3 model parameters (conversion rate, surface resistance and scavenging efficiency.
2's
measured NH3 concentration (IJg/m3)
3.4. Calibration results The first calibration runs in which the model parameters were also calibrated showed that the value of the conversion rate a was consistently adjusted to its lowest possible value and that the value of the scavenging efficiency s was always adjusted to its highest possible value. Therefore a and s were fixed at these boundary values in order to save computer time. Because the results of the jack-knife method compare very well with the results of the 'all stations' method only the results of the jack-knife method will be presented here. In Table 2 the results of the calibration analysis are listed. When the model parameters are kept at their default values, the emissions of all four emission regions are increased by the calibration procedure. The highest adjustments are
Table 2 Calibration results with default and adapted parameter values. with default parameter values
with adapted parameter values
a
28.8
10
rc
30
62 + 11
s
106
2" 106
...........................................................................................................................................................................................................................
fl (Eib)
1.18 + 0.03
0.80 + 0.05
f2 (Lun)
1.99 + 0.04
1.56 + 0.04
f3 (Vre)
1.20 + 0.02
1.13 + 0.04
f4 (Zeg)
1.79 + 2.69
1.35 + 0.32
f.s (rest)
0.27 + 0.56
1.06 + 0.13
123
178
Etot (ktonnes)
88 Figure 6. Comparison between measured and predicted values after calibration. predicted (pg/m3)
25
predicted (pg/m3)
predicted (pg/m3)
NH4
O"
8
NH4
.
1200-
wetdeposiUon
-O
9001
1o
,r
"
oO " .
,.,
,.
o
~
600-
..
1'o 1's 2'o 2's measured (pg/m3)
O~
o
~
300-
~.
~
~
measured (pg/m3)
0
.,* "
o
."
*
3~o 6~o 9~o 12'oo measured (pg/m3)
needed for the Lunteren and the Zegveld regions (regions 2 and 4 respectively). However, simultaneously the emission of 'the rest of the Netherlands' is strongly decreased. Consequently, the national total NH 3 emission decreases from 166 ktonnes to 123 ktonnes. This result is regarded as very unlikely. When a and s are set at their lowest respectively highest value and r c is allowed to change within certain limits, the calibration procedure doubles the value of r c in order to increase the ammonia concentration at the stations in the high-emission areas. Consequently, the emissions need not be increased that much as with the default parameter values. For the Lunteren region and the Zegveld region calibration factors are in the range 1.4-1.6 now. For the Eibergen region (region 1) the reported emissions are decreased with about 20%. The national total NH 3 emission increases from 166 ktonnes to 178 ktonnes with this calibration 'scenario'. Figure 6 shows the comparison between measured and calculated values after calibration with parameter values allowed to change. Obviously, the calibration has resulted in a much better agreement for all components (NH 3, NH 4, wet deposition of NHx). Table 3 gives the national mean values for NH 3 concentration, NH 4 aerosol concentration, dry, wet and total deposition before and after calibration with adapted parameter values. As a result of increasing the emissions, decreasing the conversion rate and increasing the surface resistance, the ammonia concentration has increased by about 50% and the ammonium
Table 3 National mean values before and after calibration with adapted parameter values before calibration
after calibration
NH3 (lag m -3)
4.36
6.60 (+51%)
NH4 (lag m -3)
5.25
3.42 (-35%)
wet deposition (mol ha -1 a -1)
499
583 (+ 17%)
dry deposition (mol ha- 1 a- 1)
1004
1024 (+2%)
total deposition (mol ha- 1 a- l)
1503
1607 (+7%)
89 concentration has decreased by about 35%. Wet deposition has increased due to the increase in scavenging efficiency. Also the increased r c and therefore the increased ammonia concentration contributes to the increase in wet deposition. This becomes evident when the wet deposition data are not used in the calibration. In that case r c stays at its default value. Dry deposition is hardly influenced by the change in emissions and parameters. The increase in ammonium concentration is counteracted by the decrease in ammonia concentration and the increase in surface resistance. Consequently, also the total deposition has not changed much.
4. SUMMARY AND DISCUSSION The calculated values of ammonia and ammonium concentrations using the emissions as reported by van der Hoek [8] and the default model parameters show considerable disagreement with the measured values. The calculated values for the concentrations of ammonia are lower than the measured ones, whereas for ammonium the opposite is valid. For wet deposition of NH x the calculated values are lower than the measured values. In spite of the use in this study of the measured data as they are now available, further investigation of the reliability of the ammonium concentrations and also the wet deposition of NH x is needed. The deviations between measured and predicted values for ammonia cannot be explained by the uncertainty in either of these. Measurements of ammonia concentrations in the surroundings of the fixed point do not prevail local influences on the fixed point. However, conditions with high emissions are probably under represented. Therefore, local influences cannot be excluded. Considering the fact that the comparison measurements with a van are very time consuming, it is suggested to develop low-cost passive methods which integrate the ammonia concentrations over a certain time span and which should be applied a year long. The uncertainties in the model parameters result in a quite large uncertainty in the calculated ammonia concentration, but not large enough to account for the difference in measured and predicted concentrations. The parameters studied are the conversion rate of ammonia to ammonium, the surface resistance for ammonia and the scavenging ratio for ammonia and ammonium. All three parameters describe the removal process and represent the main uncertainties. Not covered in this study are some additional uncertainties caused by (meteorological) parameters describing the dispersal process. Calibration with the model parameters kept at their default value leads to a very unlikely best fit for the emission distribution. The emissions for the Zegveld and Lunteren area increase (up to a factor 2) whereas the emissions for 'the rest of the Netherlands' are substantially reduced. As a consequence, the total Dutch emission decreases from 166 ktonnes to 123 ktonnes. When the uncertainty in the model parameters is taken into account the required corrections of the emissions are smaller but still considerable (up to 1.6 for the Lunteren area). In that case the values for the conversion rate and the scavenging efficiency are invariably set at their lowest respectively highest possible value and the surface resistance is about doubled. Because the regions with the high corrections do not contribute that much to the total Dutch emission, the latter increases only slightly from 166 ktonnes to 178 ktonnes. This increase is within the 'normal' uncertainty of an emission estimate. As a consequence of the adapted emission and model parameters the modelled national mean ammonia concentration increases with about 50%. However, the dry deposition of ammonia and ammonium is hardly affected because of the simultaneous decrease in the
90 ammonium concentration (about 35%) and the increase of the surface resistance. The wet deposition increases lightly (17%) due to the increase in the scavenging efficiency but also resulting from the higher ammonia concentration. Following this, the total deposition of NH x increases with about 7%. The results from this study are (probably) not independent of the way the emission field is partitioned. In a further study not the emissions of distinguished regions but the emission factors of the activities leading to emission of NH 3 will be calibrated.
5. REFERENCES
1. Bestrijdingsplan verzuring, Report nr. VROM 90213/8-89 (in Dutch), Ministry of Housing, Spatial Planning and Environment (1989) 2. Erisman J.-W., Water, Air, and Soil Pollution 71 (1993) 51. 3. Van Jaarsveld J.A., Report nr. 222501002, RIVM, Bilthoven, 1990. 4. Asman W.A.H. & Van Jaarsveld J.A., Atmospheric Environment, 26A (1992) 445. 5. Van Elzakker B.G., Buijsman E., Wyers G.P. & Otjes B., this volume. 6. Boermans G.M.F. & Erisman J.-W., Report nr. 222105002, RIVM, Bilthoven, 1993. 7. Acharia R.C., Rep. nr. H.H. 203 (M.Sc. Thesis), IHE, Delft, 1994. 8. Van der Hoek K.W., Report nr. 773004003 (in Dutch), RIVM, Bilthoven, 1994. 9. Asman W.A.H., Report nr. 228471008, RIVM, Bilthoven, 1992. 10. Janssen P.H.M., Heuberger P.S.C. & Sanders R., Environmental Software 9 (1994) 1.
Acknowledgements: the authors wish to express their thanks to J.A.van Jaarsveld for fruitful discussion of the topic, to H.S.M.A. Diederen for critically reviewing the draft version, and to J. Burn for editorial assistance.
G.J. Heij and J. W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
91
D e p o s i t i o n N e t w o r k of t h e F e d e r a l E n v i r o n m e n t a l A g e n c y (UBA) - R e s u l t s and Trends D. Kallweit Umweltbundesamt, Bismarckplatz 1, 14191 Berlin, Germany Abstract The UBA network is based on longterm standardised meteorological immision and deposition measurements at national level. Deposition data of Eastern (EG) and Western (WG) Germany were compared for 1986-1993, especially at times characterised by a dramatic decrease in air pollution. Important results so far have been the strong decrease of base cations, but the decrease in SO4 deposition has not reached such a high level in EG. The current slow increase in N-deposition (nitrate, ammonia) is accompanied by a slow increase in acidification in EG. Generally, the deposition level in EG and WG both slowly reached the same level. Introduction A Commission of Experts on Environmental Matters and the German Union of Water M a n a g e m e n t and Natural Resources (DVWK) found the deposition measurement operations too diverse to reach uniform findings on deposition for Germany. These findings are based on the fact that too many factors are involved in this process: for example, different deposition collectors, sampling frequencies and chemical analytical methods. The distribution of sampling sites is irregular and regionally confined. This leads to minimising the comparison of data. Using this as an argument, the UBA network has been extended to include the wet-only deposition measurement. The special quality of this network is the combined standardised meteorological immission and deposition long-term measurements at national level using automatically operated container stations and manned stations. B a c k g r o u n d , S a m p l i n g Sites and M e t h o d s Since the beginning of the 1980s, daily-bulk measurements have been carried out at UBA-manned stations. These measurements are indicated in Table 1. With the unification of Germany, 10 stations from a total of 30 belonging to the wet-only deposition network of the Meteorological Service of the GDR were joined up to the network of the UBA. From the end of 1991, this has led to the gradual development of a national wet-only measuring programme as stated above (Figure 1).
92
container station %~ ~ measurement site 0~ wet-only site _&~F q ,
/
~ ,~r~.~ ~ ~
~ ~ .... t
~
~" Hohenwestedt t.~
~)Twixlum
- -~
O,.um
~,
I./"-jNeug~br.s~;i ~.
~,
i'
:-~_~
~ '
.~'-:c,
, ~.~Solling
Hilchenba
,A~
...... ~ ~
Waldhof
-%:
~
~,
Wies,enburg ......
'/
Lindenberg
Leinefelde Lehn ' J J Regnitzlo~:
uselbach
8
Kehlheim Ansbach Brotjacklriegel
~chauinsland
Murnauer Moo
2 Figure 1. Immission and deposition sites within the UBA Monitoring Network
93 At the moment, measurements are being carried out at 26 stations; the entire network will be completed by the end of this year. The measuring programme is intended to estimate the quantity of rain, pH, conductivity, main ions and selected heavy metals. First, the precipitation sampling was achieved by means of an ANTAS wet-only collector produced by the Meteorological Service. After a favourable comparison of deposition collectors of the former GDR and the FRG (mean deviation < ca. 5 %) the collector was replaced by NSA 181 KD produced by the Eigenbrodt firm. This wet-only collector contains a permanent cooling and sample-changing system, allowing samples to be changed every two weeks instead of weekly. The samples are analysed in the laboratory of the Institute of Energetics in Leipzig. The bulk samples are analysed at the UBA site in Schauinsland (see Table I for methods). Table 1 UBA programme of deposition measurements and analytical methods p r o g r a m m e type wet-only programme: number of stations: 30 8 manned stations 16 a u t o m a t i c a l l y operated c o n t a i n e r stations 6 stations in cooperation with the federal states sampler: wet-only NSA 181 KD Fa. Eigenbrodt frequency of sampling: weekly (Tuesday-Tuesday), 8:00 a.m. (7:00 UTC) precipitation volume, conductivity, pH 8042", NO3-, NH4+, C1Na+, Mg2+, Ca2+, K+ heavy metals: Pb, Cd, Cu, Zn, Mn bulk programme: 8 manned stations number of stations: 8 bulk ARS 721 Fa. Eigenbrodt sampler: frequency of sampling: daily, 9:00 a.m. (8:00 UTC) precipitation volume, conductivity, pH SO42-, NO3-, NH4+, C1Na+, Mg2+, Ca2+, K+ heavy metals: Pb, Cd, Cu, Zn, Mn, Fe analytical m e t h o d s wet-only/bulk programme:" parameter method detection limit 8042ion chromatography 10 ~tg 0,21~teq/1 NO3on chromatography 10 ~tg 0,16 ~eq/l NH4 + ion chromatography 10 ~g 0,55 ~teq/l NH4 + flow injection 10 ~tg 0,55 ~eq/l C1ion chromatography 10 ~tg 0,28 ~teq/l Na+ ion chromatography 10 ~g 0,43 ~teq/1 Mg 2+ ion chromatography 10 ~tg 0,25 ~eq/1 Ca2+ ion chromatography 10 ~tg 0,50 ~eq/l K+ ion chromatography 10 ~tg 0,80 ~teq/l
94 wet-only deposition: laboratory of the Institute of Energetics, Leipzig bulk deposition: laboratory at Schauinsland/UBA site The wet deposition measurements are inspected at regular intervals to improve quality (QA/QC). For instance, spotchecks are carried out at the sites. We participate in national and international intercomparisons of collectors and analytical methods. The conditions for choosing the deposition measurement site must represent: - the background level of air pollution and deposition. They must include as many types of ecosystems as possible. The sites not owned by the UBA must be suitable for long-term measurement/observation. Another aspect was to measure on sites where scientific research installations and institutions of the various L~inder (federal states) already existed. The intent was for them to complement each other. In addition, at UBA automatic stations the following m e a s u r e m e n t s were combined: - wet-only/weekly- UBA; -wet-only/daily or four-hourly s a m p l i n g - Institute of Tropospheric Research, Leipzig; -micrometeorological/inferential m e a s u r e m e n t s - F r a u n h o f e r I n s t i t u t e of Environmental Research, Garmisch Partenkirchen; - measurements of TSP total and component amounts - UBA. This cooperation is presently being conducted through the SANA Research Project (re-development of Eastern Germany). The aim of this cooperation is to develop, if possible, a method to estimate the entire deposition input for Germany using measurements and modelling and also considering the chemical processes in the atmosphere.
Results
and
discussion
For the initial evaluation, data is available from eight stations in E a s t e r n Germany (EG) and five manned stations in the Western Sector of the Federal Republic of Germany (WG) for the period 1986-1993. The deposition data of the bulk deposition UBA network (WG) was compared with the wet-only deposition data from the Meteorological Service (since 1988 wet-only) up to 1991, when UBA data (EG) was used. In fact, the bulk samples on sites with a low pollution level are analysed only after a rainfall, so that they are more comparable to wet-only samples (see Table 2). The change in emission, and in its wake the immission matrix, was carried out at short notice in the quickest possible time, especially in EG, (i.e. in the period between 1986-1993). This was accelerated by the collapse of industry and agriculture in the former GDR, which is now experiencing an upward swing.
95 Table 2 Mean values for wet deposition for 1986-1993 year RR
WG 86 87 88 89 90 91 92 93
86 87 88 89 90 91 92 93
Na+
mm
kg~a ~a
kg/ha kg/ha ~ a
1.018 1.082 1.052 751 862 736 920 906
30.59 31.09 31.65 24.54 25.94 21.16 21.10 21.11
7.52 8.04 8.43 6.79 6.51 5.61 6.07 6.08
year RR
EG
SO42- NO3- NH4 + C1-
22.20 26.22 25.61 20.70 19.27 17.49 19.54 19.21
20.39 15.92 27.21 16.26 42.00 26.65 24.62 20.80
SO4 2" NO3- NH4 + C1-
Mg2++ Ca2+ ~a
11.32 1.59 8.37 1.28 14.56 2.11 8.16 1.41 22.212.70 13.601.94 13.581.75 10.861.41 Na+
K+
H+
kg/ha ~ a
kg/ha
3.38 3.84 3.80 3.57 3.75 3.16 3.86 3.69
1.49 1.30 1.56 1.20 1.75 3.38 1.22 1.03
0.32 0.35 0.31 0.24 0.22 0.21 0.20 0.20
K+
H+
MG2++ Ca2+
mm
kg~a ~a
kg/ha kg/ha ~ a
~a
kg/ha ~ a
~a
542 576 515 396 520
63.13 66.24 64.24 41.44 37.44
7.68 8.82 8.14 6.43 6.81
1.83 2.13 2.15 1.11 1.10
12.26 14.65 14.21 7.21 5.73
2.29 2.02 2.46 1.09 0.86
0.20 0.22 0.21 0.16 0.16
437 549
17.67 13.12 3.84 7.55 22.31 16.49 5.92 6.73
0.82 0.74
1.28 1.27
0.28 0.20
19.53 19.57 20.28 14.95 15.44
8.95 10.17 13.52 8.77 8.99
4.89 4.82 6.74 4.18 4.06
4.21 0.82 3.51 0.74
W G - mean values for the deposition sites of Westerland, Waldhof, Deuselbach, Schauinsland, Brotjacklriegel E G - m e a n values for the deposition sites of Schwerin, Teterow, Neuglobsow, Lindenberg, Leipzig, Leinefelde, Schmiicke The decrease in SO2 emission (see Figure 2) is accompanied by a s i m u l t a n e o u s decrease of particulated m a t t e r emission (flying ash with Ca and Mg) and is in accordance with the mean value of SO2 immission (see Figure 3) in rural areas (ca. 35 % lower) during 1985-1988. For the same period, in urban areas this value rose to 40 to 65 %. Nevertheless, the SO2 immission in EG is three times higher t h a n in WG and leads to exceedances of permitted a n n u a l values (140 Ilg/m3 SO2) of the Technical I n s t r u c t i o n on Air QualityfrA-Luft) in Germany. The p a r t i c u l a t e d m a t t e r immission dropped by 30 % on average.
95 The annual wet deposition (see Table 2) depict a slower decrease in comparison to the immission. This is reflected especially in the case of WG, where a marked fall in immission took place between 1987-1988. But the SO4 deposition showed a decline only from the beginning of 1989. The same can be said of EG, although the decline began here in 1989/1990. This could be related to the delayed (1990) decrease of particulated matter immission.
(~o
SO2
~oo
~oo ...... ...........
"
NO2
............................................................
2500 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
~ = .................................
.
I ~
o,
l ........
,
,
,
;'
,
,
,
partlcultecl m a t t e r
NH3
25OO
290O
. . . . . . .
~
. . . . . . . . . . . . . . . . . . . . . . .
4O0-
ISO0
lOGO
,
86
,
e7
,
e~
,
89
,
9Q
gl
O"
"
e7 ~ / 7 ~ " . G .
wG eg
j
go
.1
[~KQ i w ~ ]
Figure 2
Trend of SO2, NO2, NH3 and particulated matter emission in Germany 1986-1991 (Data from UBA)
97 At the beginning of the period in question, the SO4 deposition decreased twofold in comparison to WG (see Figure 4). In EG the base Ca cations achieved a fourfold decrease and Mg also decreased somewhat more than in WG (see Figure 5). At the end of 1993 the situation was different. The SO4 deposition in EG decreased by two-or threefold, the Ca deposition likewise. Within this same period the SO4 deposition decreased by a third and the Ca deposition in WG remained unchanged. The SO 4Ca-ratio for EG is characterised by strong deviations caused by irregular decreases in SO2 and Ca emissions.
S02 36"
30 .............
!~
26 ..............
~
15 ..............
~
..............................................................................................................
.....
I
.............................................................................
~.............. 86
.................... 87
$$
e0
I 91
90
.
.
.
.
.
.
.
.
.
.
.
.
92
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
93
J
[p'-/-~eQ ~wQ
TSP 45 40 . . . . . . . . . 353025 .............. 2015 ..............
10 .............. 5 ..............
0
$6
87
v
$$
~)
q
90
IrT/-~eo ~wo
Figure 3
|
91
92
93
]
Trend in 802 and TSP immission in Germany (yearly mean concentration based on daily means)
98 The following developments have been recorded for the nitrogen compounds. The NO3 deposition decreased slightly in WG between 1988-1991 and has since either remained constant or shown a minor increase. In EG this tendency has increased only since 1993. The NH4 deposition shown a downward trend in WG. In EG the values were the same at the beginning of this period. In 1992 these deposition values were reduced by 50% and with the revitalisation of agriculture there has been a further increase. It can be noted that the acidifying capacity is in principle characterised by a slow increase in acidifying species (SO42-, NO3-, CI-, NH4+, H +) and the more decreasing input of base cations. This development indicates a slow increase in acidification (see Figure 7) and represents a danger for the ecosystems. The ration of the main parameters of wet-deposition in EG and WG is increasing to the same level (see Figure 8). This development is a result of the active environmental policy and application of measures for air purification of national level.
Conclusions
- Acidification is increasing, especially in Eastern Germany. With the long-term and standardised wet-only network demonstrated we are able to follow the trend more accurately. - I n order to estimate the total input by deposition and its effects we need measurements on fog and cloud deposition, also on dry deposition over low vegetation and forests. -
It will be necessary to combine the measurement activities of several institutes to get information on deposition in Germany, including application using model calculations.
99
EG 1,4 , 1,2-1
0,8-1 "~ 0,6-1 0,4-1 0,2-t ! 19/~
1907
1988
1989
I mm so, ~
1990
.o3 ~
1991
| 1992
1993
c,
WG 1,4 1,31,21,110,9i
0,80,7-
0,60,50,40,30,20,10
198(!
i
1987
!
19eS
,
[ m e so, ~
1989
|
1990
.03 ~
1991
I
1992
I
1993
c,
Figure 4 Yearly development of the m e a n deposition values (anions) in EG a n d WG. There is no sea-salt correction for Westerland, so the m e a n value of C1 is influenced by this factor
100
EG 1,4 1,2-
0,80,60,40,2-
1986
i
1987
i
1988
I
1989
(3,
~
H
~Mg
i
NH4
1990
i
1991
~
N.
I
I K
i
1992
i
1993
WG 1,4 1,2-
0,8-~ 0,60,40,2-
1986
!
1987
~H
!
1988
!
1989
~Mg
i
1990
i
!
1991
!
1992
!
1993
IK
Figure 5 Annual development of mean deposition value cations in EG and WG
101 S 0 4 / C a ratio 4
3,5 - ................................................................................................................................ 3"~ 2,5 -~. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
~/ ..............................................
..........................................
~ .................................
1,S-
..................................
~ .................................
1-
..................................
~ ............................
................................
~j~/...........................
2-
0,5
-
,
1988
,
19:B7
,
198:8
,
1989
,
1990
~
1991
,
,
1992
1993
Figure 6 Yearly development of SO4Ca deposition
EG/WG 4 3,6-
32,521,5I0,50
,
1986
,
1987
,
1~
,
1989 1990 Jehr
,
1991
1992
1993
Figure 8 Yearly development of ratio of main ions in deposition EG/WG
102
EG
:)(5 2 15 1
H H H~
05 i 19~(5
[~
! 1987
|
19(~
198g
aoidifying ion. ~
|
1990
i 1991
btme oldiomi
~
1992
lgg3
aoidily
WG
25
2
Z
15
i~
1 057
, 198e
L~7~ - 9
'~ / ~ ~ t ~ i ~
i 1957
| lgB$
ion- ~
| lg89
!
!
lggO
lggl
b-ee oallon.
[~]
, lgg2
aoidily
1993
I
Figure 7 Yearly development of acidity (A-acidifying ions (SO4 2-, NO3-, NH4 § Cl-, H+), B-base cations (Na+, K+, Mg2+, Ca2+) acidity = A-B)
G.J. Heij and J. W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
103
The influence of ammonium nitrate equilibrium on the measurement of exchange fluxes of ammonia and nitric acid Yuanhang Zhang*, Harry ten Brink, Sjaak Slanina and Paul Wyers Netherlands Energy Research Foundation, PO box 1, 1755 ZG Petten, The Netherlands
Abstract Three sets of data with a time resolution of 40 minutes at Leende, Zegveld and Speuld in the Netherlands are used to evaluate atmospheric equilibrium of ammonia and nitric acid with ammonium nitrate. The comparison between the measured concentration products Km and theoretical values Ke reveals that gaseous ammonia and nitric acid are not in simultaneous equilibrium with ammonium nitrate aerosol, because the ratios of Km to Ke are far from unity, varying from less than 0.01 to above 100. Kinetic constraints upon the attainment of equilibrium are emphasized strongly by the results obtained at the three sites. The disequilibrium phenomena can be used to explain why HNO3 can show upward gradients and to obtain indications of the errors, using gradient methods in the measurement of deposition fluxes.
1. INTRODUCTION In Europe, exchange of NH3 and HNO3 between atmosphere and biosphere is of increasing interests. However, it is very difficult to derive their representative dry deposition velocities, especially for HNO3, because of the chemical reactions involving HNO3, NH3 and NH4NO3 aerosol. Sutton et al. [ 1] found consistent negative Rc for HNO 3 in Essex, England, while Huebert et al. [2] found a steeper aerosol nitrate gradient than that of its vapor and sometimes an apparent emission ofHNO 3 in the USA. According to Kramm and Dlugi [3], if no equilibrium exists, the fluxes can not be calculated by micrometeorological methods without appropriate corrections. Therefore, it is necessary to test whether the chemical equilibrium among HNO 3, NH3 and NH4NO3 really occurs in the atmosphere. The hypothesis of atmospheric equilibrium among gaseous HNO3, NH3 and NH4NO3 aerosol was first proposed by Stelson et al. [4] in 1979, who tried to analyze field experimental data by application of the equilibrium relationship. Since then, extensive work has been done regarding thermodynamic theory and field experiments [5-17]. Field measurements in the United States, Japan and Europe showed that NH4NO3 aerosol was generally in equilibrium with its gaseous precursors [8-17], especially at temperatures above 5~ and a relative humidity less than 80%. Departure from equilibrium was mostly found under conditions of low temperature and high humidity [9,10,15,16], when the concentration product was depressed to values less than 1 ppbv2 [ 18].
104 Meanwhile the validity of this thermodynamic approach was questioned for a long time[ 18-19]. The large uncertainty in the equilibrium constant under ambient conditions, as it is derived from laboratory experiments, complicates any assessment of its importance. Numerical simulations indicated that significant deviations from equilibrium, both positive and negative, might occur as the reactions are not sufficiently fast to maintain equilibrium under conditions of changes in meteorological, emission and chemical reaction parameters. However, after Jaffe's criticism [ 19], Mozukewich [7] argued that the available thermodynamic data on this system was much better than was implied by a factor of two range and that the dissociation constant may be determined within + 12% for dry aerosol. In the Netherlands, equilibrium between HNO3, NH3 and NH4NO3 aerosol is not extensively investigated yet. Erisman et al. [ 14] and Allen et al. [ 16] reported that the concentration products of the gaseous acids and base were generally in good agreement with theoretical prediction at temperatures higher than 0~ and relative humidity less than 80%. However, large deviations from theoretical values can be observed from the data of Erisman et al. [ 14] and Allen et al. [ 16]. Recently, various kinds of denuder techniques were developed to measure NH3 and HNO3 with high accuracy [20]. Many good data sets on simultaneously measured gaseous NH3 and HNO3 concentrations as well as meteorological data were obtained in recent years with a time resolution of 40 minutes by ECN. Three sites with different characteristics in location, emission and meteorology are chosen to test whether gaseous HNO3 and NH3 are in equilibrium with NH4NO3 aerosol in the Netherlands. The results obtained indicate complications in the measurement of dry deposition fluxes of NH3 and HNO3.
2. EXPERIMENTAL Ammonium nitrate NH4NO3 is present either in solid phase or as a solution, dependent upon temperature and relative humidity in atmosphere. NH4NO3 is assumed to be in reversible equilibrium with its gas precursors NH3 and HNO3. The equilibrium constant of the reversible reaction is defined as the product of NH3 and HNO3 vapor pressure above NH4NO3 aerosol. For solid aerosol, temperature dependence of the dissociation constant Kp of ammonium nitrate is given as [7]: Ln Kp = 118.87 - 24084/T - 6.025 Ln T
(1)
with T in K and Kp in ppbv2 with an accuracy of + 12%. When relative humidity is above the deliquescence point, NH4NO3 will exist as droplets. The equilibrium constant Kp' above the droplet is given as [7]: Kp' = [P 1-P2 (1-aw) Ln P1 =-135.94 Ln P2 = -122.65 Ln P3 = -182.61
+ P3 (1-aw)2] (1 +aw) 1.73 Kp + 8763/T + 19.12 Ln T + 9969/T + 16.22 Ln T + 13875/T + 24.46 Ln T
(2) (3) (4) (5)
Where aw is water activity and Kp is the equilibrium constant of solid NH4NO3.
105 Simultaneous measurements of NH3 and HNO3 with a time resolution of 40 minutes were made at three sites in the Netherlands: Speuld (coniferous forest) for a whole year in 1989; Leende (heathland) between April 25 and May 10, 1991; and Zegveld (grass pasture) from July 13 to July 20, 1993. A detailed description of the experiments at the three sites is given elsewhere [20-22]. A brief introduction to the experiments is given below. NH3 and HNO3 were measured at two heights at Leende and Zegveld and at one height in Speuld by wet denuders. Combining chemical and meteorological data, in total 240, 400 and 894 samples were obtained at Leende, Zegveld and Speuld respectively. NH3 concentration showed a diurnal variation with a maximum at nighttime and a minimum in daytime and HNO3 concentration showed a weak diurnal variation with slightly higher values in the late a~ernoon at the three sites, shown in figure 1. NH3 fluxes were generally directed to the surface at Leende and Speuld, while NH3 showed a bi-directional flux at Zegveld. For HNO3, its fluxes were observed to be directed away from the surface at Leende and Zegveld most of the time, as shown in figure 2. As it is believed that any surface is a perfect sink for
so-~ 2o-I
/+", "+"
+ +'++ + +~
4+, . . . .
:/ooo
/+\
-o-~ L
/ x;'•
--+--Nm-H
\
.~ooo_oo,
~
"+
/
9
.~+.+
0 6 .",
3
.,,,,
a
,,,,
• •215 x /
\•
• •
• x ~
"• ~ • • ""
""•
x • • • • • x' ' x • • • • • • • • ^ • 2 1 5 •
x
0
t
\
•215 \•215215215215
116.0
116.5
•
• 2 1 5 2 1 5 2 1 5 2 1 5 2 1 5 2• 21 15 5 2 1 5 •
117.0
117.5
118.0
Julian day
Figure 1. Concentration variations ofNH 3 and HNO 3 on April 26-27, 1991 at Leende in the Netherlands (L = lowest level, H = highest level)
0.5
/,,,,
0.0
,~
\d
-,,
\?
-0.5 -
\~ -1.0
'
0
I
4
'
--+--~o3 I
8
'
I
12
'
I
16
'
I
20
'
I
24
Time Figure 2. NH 3 and HNO 3 fluxes on April 26, 1991 at Leende in the Netherlands
106 HNO3, the upward fluxes of HNO3 are probably caused by the changes in the NH3/HNO3/ NH4NO 3 system.
3. RESULTS AND DISCUSSION Figure 3 presents the measured concentration products [NH3][HNO3] in Leende, together with theoretical curves calculated for equilibrium between aerosol and gaseous compounds, expressed as a function of temperature and relative humidity. The measured products in Leende show a tendency to exceed the equilibrium condition generally (figure 3). Below a relative humidity Rh of 60%, the measured products fit theoretical prediction for solid aerosol qualitatively with a slightly systematic positive bias (positive and negative bias are defined as measured products higher or lower than the equilibrium value), while they greatly exceed theoretical predictions for droplets. In contrast to Leende, the measured products for [NH3 ][HNO3 ] in Zegveld are generally lower than theoretical prediction when Rh is below 60%. As Rh increases, the measured concentration products exceed the theoretical prediction increasingly. In Speuld, the measured concentration products can be higher or lower than
100 - -
-
---SOLID
-
--3W---95
o
+
--X--90 --V-- g5
10 t~
>
,.Q
O
<6O
-~A
61-70 71-80
o
81-90
D
>90
9 []
~o~ E! o
II
0.1
(a) Leende 0.01
[
3.4
'
I
'
3.5
I
3.6
'
I
3.7
1000/T Figure 3. Measured concentration products [NH3][HNO3] and theoretical equilibrium curves as a function of temperature and humidity at Leende.
107 theoretical predictions regardless of the fact that the aerosol was in solid or in aqueous phase. The significant characteristic in Speuld is that the measured concentration products tend to be smaller than theoretical predictions at lower relative humidity. The discrepancies from equilibrium are still large if the data with relative humidity > 80% and temperature < 5~ are rejected. Lewin et al. [ 13] obtained a good agreement between measured concentration products and theoretical equilibrium constants between HNO 3, NH3 and NH4NO3 at a temperature higher than 0~ at a rural site in the Northeastern U.S. on a 24 hours basis, although their data set was relatively small. In Speuld, a rural site in the Netherlands, disequilibrium was generally observed. Recently, Mozukewich [4] offered a model to calculate the dissociation constant of NH4NO3 aerosol as an explicit function of temperature and relative humidity, allowing a more precise evaluation of the equilibrium in different temperature and relative humidity. The ratio of measured concentration product Km=[NH3 ][HNO3] to the value Ke predicted by theory is defined as a measure of whether the system is in equilibrium. If the system is in equilibrium, the ratio should be near unity. Any ratio larger or smaller than unity means that system is not in equilibrium. Figure 4 depicts relative humidity and temperature dependence of this ratio Km/Ke at Leende. The ratios Km/Ke are mostly larger than unity, but some are slightly lower than unity at low humidity and high temperature. As Rh increases, the ratio increases and discrepancy from theory also increases (figure 4a). As temperature increases, the ratio decreases with a strong linear relationship (figure 4b). At Rh < 80 %, the ratios are ranging generally between 0.1 and 100. Figure 4 clearly shows that maximum disequilibrium occurs under conditions of high humidity and low temperature, while the concentration products are slightly under equilibrium at high temperature and low humidity. The same relations between humidity, temperature and the ratio Km/Ke are obtained in Zegveld and Speuld.
1000 -
~
lOO
~
1
~
- BB
~]
0.1
' 20
I 40
'
1000 -.
lOO
r'lffi
B o
~u
I 60
I
I 80
'
I 100
(a) Relative humidity (%)
0.1
i 270
280
I 290
I 300
(b) Temperature (K)
Figure 4. The ratio of measured concentration product to equilibrium value predicted by thermodynamics as a function of relative humidity (a) and temperature (b) at Leende.
108 1000
40 "" .
'
I
'
- I ~D~
'
I ' ---e---Km/Ke
I ~~
/%
--D--~
1
100
,,g ~
g
"~
10
o
eeee
c~ .
f 9
9e
N 0
'"~
116.0
I
I
I
116.5
117.0
117.5
4-
t~ 1
0.1 118.0
Julian day Figure 5. Diurnal variation of the ratio of measured concentration product Km to theoretical prediction Ke at Leende from April 26-27, 1991.
The ratio Km/Ke shows strong diurnal variation at Leende, as shown in figure 5. The ratios are much higher than unity in the night and the early morning and slightly lower than unity in the late afternoon. This variation is consistent with NH3 concentration changes (figure 1), showing the major role of NH3 in the disequilibrium. If the accuracy of the theoretical prediction is taken into account, the conclusion is that the system ofNH3, HNO3 and NH4NO3 is strongly above equilibrium in the night and nearly in equilibrium or under equilibrium at daytime in Leende. The results obtained here are slightly different from those ofErisman et al. [ 14] who found that a reasonable agreement existed between measured products and theoretical calculation for a temperature > 0 ~ and Rh < 80% at Cabauw in the Netherlands. The explanation for the disequilibrium between aerosol and gaseous compounds is believed to be kinetic constraints preventing rapid attainment of equilibrium. Gas-particle transformation processes are not governed by thermodynamic law only and kinetic constraints play an important role on the transformation. At the three sites, NH3 concentrations are higher at night and lower at daytime with a peak in the early morning while HNO3 concentrations shows a weak diurnal variation. Thus the concentrations products of NH3 and HNO3 do not show a strong diurnal variation with a peak in the afternoon, which is predicted by thermodynamic theory based on the variation of temperature and relative humidity, as shown in figure 5. The diurnal changes in temperature and humidity alter the dissociation constant at a rate which is sufficiently rapid to prevent equilibrium between NH4NO3 aerosol and its gas precursors. Similarly, changes in precursor gas concentrations can not be accommodated sufficiently rapid to maintain equilibrium conditions [ 18]. As shown in figure 5, it seems that kinetic constraints for gaseous reactions at low temperature and for aerosol evaporation at high temperature offer a reasonable explanation for the disequilibrium. In addition to kinetic constraints, high NH3 concentrations reduce HNO3 concentrations to such a low level that large measurement uncertainties may occur, resulting in departure from theory especially at high humidity. However, the measurement uncertainties are not the main reason for the disequilibrium because the measured products do not fit the theory well at
109 temperatures above 5~ relative humidity less than 80% and measured concentration products larger than 1 ppbv2. In Zegveld and Speuld, interference by dissociation followed by deposition of the products could cause a negative departure, because the sampling height is low and the wet surface of the canopy is a strong sink for HNO3 and NH3. The formation of hydrated gases, which are not accounted for in the theoretical model, may result in underestimation of the HNO 3 concentration especially at low temperatures and high humidity.
4. A FIRST ORDER ESTIMATION OF THE INFLUENCE OF THE NH3/HNO3/ NH4NO 3 SYSTEM ON THE FLUXES OF HNO 3 AND NH 3 In Leende, the interference on NH 3 and HNO3 fluxes by dissociation ofNH4NO 3 is obvious and large, especially for HNO 3, shown in figure 2. NH3 is deposited, while HNO3 fluxes are directed away from the surface. Since HNO3 shows a very weak peak in the late afternoon, the photochemical formation of HNO3 is not a reasonable explanation to the upward gradient of HNO 3. In Zegveld, NH4NO 3 and HNO 3 were measured by thermodenuder in July, 1992. Although the precision of the data can not allow calculation of aerosol flux, they reveal that the HNO 3 gradient and NH4NO 3 concentration show an anticorrelation, see figure 6. At high NH4NO 3 concentrations, apparent HNO 3 emissions are observed and at low NH4NO 3 concentrations, HNO 3 depositions are observed. In Manndorf, Germany where high ammonium sulfate and low NH4NO 3 concentration were observed, HNO 3 usually showed a deposition flux. In Leende, nitrate concentration was very high and about 70% of the nitrate existed as NH4NO 3. As figure 5 shows, in daytime, especially in the late afternoon, the system is nearly in equilibrium or under equilibrium. Decreasing nitrate aerosol concentrations were observed in Leende, as shown in figure 7. In daytime, NH4NO 3 aerosol dissociation is favored at low level compared to higher levels due to the influence of the dissociation followed by deposition. Equal amounts of NH 3 and HNO 3 are produced this way. The HNO 3 gradient is easily changed from downward to upward direction by this mechanism because HNO3 concentrations and gradients are small. At the night, the NH3, HNO3 and NH4NO3 are strongly above equilibrium and the system has shifted to aerosol formation. HNO3 upward gradient usually did not change very much, implying that aerosol formation rate was slow, otherwise, the strong NH3 deposition gradient would change the 20
O.8
/
/ .---4-- HNO3 | --o--NH4NO3
o ~ ~
Ti!
- 0.4 - 0.0
r,,2,
/ -
0 195
200
-w; i
I ~ 205 Julian day
&
I 210
-1.2 215
Figure 6. Anti-correlation between NH4NO 3 concentration and HNO 3 gradient in July, 1992 at Zegveld in the Netherlands
110 6 I I~! NIB/SO4 I-I NO3/SO4 I
O 2
0
i
10:00
i
i
t
11:30
12 38
i
i
13:37
17:05
Time
Figure 7. Aerosol NH4+/SO42- and NO3-/SO42" changes on April 26, 1991 in Leende
(a) April 26, 8:00 4
0
D
Measured
~
Corrected ~
(21
2 O
~
Corrected
~
0
D
;
OO
;
~
;
Measured
4
,'o
,;
(b) April 26, 13:00 4 0
0
0
~
~:
4
Measured
Cot,rected
O
Corrected .
--~
OO
/ O
Measured
o
0
HNO3 concentration (ug/m3)
NH3 concentration, ug/m3
Figure 8. Diagram of the influence ofNH4NO 3 dissociation on NH 3 and HNO 3 fluxes on April 26, 1991 at Leende in the Netherlands
HNO3 gradient greatly. At present, the kinetics of the NH 3, HNO 3 and NH4NO 3 equilibrium system are not well understood and NH4NO3 aerosol deposition is not measured precisely. It is very difficult to analyze quantitatively how the equilibrium influences the fluxes of the species involved in the system. Even though large uncertainties exist, the measurements of HNO3 and NH3 concentrations do indicate that the NH3/HNO3/NH4NO 3 system influences NH 3 and HNO 3 fluxes and the upward fluxes of HNO 3 are probably caused by NH4NO 3 dissociation. Based on the assumption that the HNO 3 upward gradient is caused by NH4NO 3 dissociation, a first order estimate can be made of the influence of the NH4NO 3 dissociation to NH 3 flux. The NH 3 and HNO 3 fluxes on April 26, 1991 in Leende are used as an example. The assumption is that the background HNO 3 concentration at 3.47 m is 1 Bg/m3, which is consistent with the HNO 3 concentration on May 3-4, 1991 when HNO 3 showed a deposition flux. The dry deposition velocity Vd ofHNO 3 at 3.47 m can be obtained by the inverse of the
111 sum of Ra and Rb, assuming Rc to be zero. From Vd and the HNO 3 concentration at 3.47 m, the corrected concentration ofHNO 3 at 0.93 m can be estimated, as shown in figure 8. The corrected NH 3 concentrations at both levels can be estimated from the measured and corrected concentration of HNO 3. Figure 8 shows two extreme cases: (a) high NH 3 deposition flux and (b) high HNO 3 upward flux. In both cases, NH4NO 3 dissociation has an enormous influence on the HNO 3 gradient and sometimes changes its direction. When the NH 3 gradient is large, NH4NO 3 dissociation has little influence on the NH 3 flux. In figure 8a, where the NH 3 flux is high, the measured NH 3 flux is estimated only about 5% lower than without the correction for NH4NO 3 dissociation. In figure 8b where the HNO 3 upward flux is high and the NH 3 gradient is small, NH4NO 3 dissociation has a greater influence on both HNO 3 and NH 3 fluxes. The measured NH 3 deposition flux is estimated 76% lower than without this correction. Figure 9 shows diurnal variations of HNO 3 and NH 3 fluxes before and after the correction for NH4NO 3 dissociation. Since HNO 3 usually shows an upward gradients at Leende, NH 3 deposition fluxes are underestimated or NH 3 emission fluxes are overestimated in the late afternoon because NH4NO 3 dissociation influences the NH 3 gradient to a large extent during that period.
D
l i t
0.5 -fj~
- - D - - NH3_M - - o - - NI-t3 C
/
9
~
[]
Dn--12_
-0.5
-
-1.0
-
I
116.0
116.5
'
I
117.0
'
I
117 9
'
I
118.0
Julian day Figure 9. The influence ofNH4NO 3 dissociation on NH 3 flux on April 26-27, 1991 in Leende (NH3_M: measured NH3 flux, NH3_C" corrected NH3 flux)
5. CONCLUSION The three data sets collected in 1989, 1991 and 1993 in Speuld, Leende and Zegveld in the Netherlands are investigated to evaluate the atmospheric equilibrium ofNH3, HNO3 and NH4NO3 aerosol. A very poor agreement between measured gaseous concentration products Km=[NH3][HNO3 ] and theoretical predicted equilibrium vales Ke is found at low relative humidity 9The comparison between the measured Km and theoretical Ke reveals that NH3 and HNO3 are generally not in equilibrium with NH4NO3 aerosol, because the ratios of Km to Ke are far from unity, varying from less than 0.01 to above 100. The ratio increase with increasing relative humidity and decreasing temperature. The maximum positive departure from equilibrium usually occurs at low temperature and high relative humidity, which is consistent with most field measurement 9The maximum negative deviation usually happens at
112 high temperature and low relative humidity. Kinetic constraints upon the attainment of equilibrium for the NH3-HNO3-NH4NO3 system are emphasized by the results obtained at the three sites. Disequilibrium between NH3, HNO3 and NH4NO3 can be used to explain why HNO3 upward gradients in Leende are observed. The dissociation ofNH4NO 3 is responsible for the upward gradient of HNO 3 and results in a slight underestimation of the NH 3 deposition flux during the night. Large errors in the estimation ofNH 3 emission or deposition fluxes are caused by the influence ofNH4NO 3 dissociation in the late afternoon when NH 3 gradients are small.
Acknowledgment. The authors gratefully acknowledge Geoffrey DoUard, AEA technology, Harwell, UK for providing aerosol data at Leende in the Netherlands.
Permanent address: Center of Environmental Sciences, Peking University, Beijing 100871, P.R. China
6. REFERENCES
[1]
[21 [3] [4]
[5] [6]
[7] [8] [9] [ 10] [11] [12] [13] [14] [15] [ 16] [ 17] [18] [ 19] [20] [21 ] [22]
Sutton M. et al. (1993) Exchange of ammonia and sulphur dioxide with vegetation. EUROTRAC/BIATEX, annual report 1993, part4, Biatex, 213-224. Huebert B. J. et al., J. Geophys. Res., 93D(6) 7127-7136 (1988). Kramm G. and Dlugi R., J. Atmospheric Chemistry. (in press) (1993) Stelson A. W. et al., Atmospheric Environment 13, 369-3 71 (1979). Stelson A. W. and Seinfeld J. H., Atmospheric Environment 16, 983-992 (1982). Stelson A. W. and Seinfeld J. H., Atmospheric Environment 16, 2507-2514 (1982). Mozurkewich M., Atmospheric Environment 27A, 261-270 (1993). Doyle G. A. et al., Environmental science & technology 13, 1416-1419 (1979). Tanner R., Atmospheric Environment 16, 2935-2942 (1982). Cadle S. H. et al., Atmospheric Environment 16, 2501-2506 (1982). Harrison R. M. and Pio C. A., Tellus 35B, 155-159 (1983). Hildeman L. M. et al., Atmospheric Environment 18, 1737-1750 (1984). Lewin E. E. et al., Atmospheric Environment 20, 59-70 (1986). Erisman J. W. et al., Atmospheric Environment 22, 1153-1160 (1988). Allen A. G. and Harrison R. M., Atmospheric Environment 23, 1591-1599 (1989). Pio, C. A. and Numes, T. V., Atmospheric Environment 26A,505-512 (1992). Chang Y. S. et al., Atmospheric Environment 20, 1969-1977 (1986). Harrison R. M. and Mackenzie A. R., Atmospheric Environment 24A, 91-102 (1990). Jeffe D. (1988) Accuracy of measured ammonium nitrate equilibrium values. Atmospheric Environment 22, 2329-2330. Slanina J. et al. (1990) Acidification research at ECN. Report ECN-C-90-064, Netherlands Energy Research Foundation, Petten, The Netherlands. Duyzer, J. H. et al. (1992) The joint experiment on surface exchange of trace gases over the Leende heathland. Report TNO PU 92/015. Plantaz M. A. et al. (1993) Continuous monitoring of deposition fluxes of nitrogen compounds on low vegetation, in: Eurotrac Annual Report 1992, part 4, B IATEX.
ATMOSPHERIC DEPOSITION S E S S I O N III PARTICLE DEPOSITION
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
115
Particle deposition to forests Jan WiUem Erisman', Geert Draaijers', Jan Duyzerb, Peter Hofschreuderc, Niek van Leeuwen d, Ferdinand Rtimerc, Walter Ruijgrok*, Paul Wyers f
"RIVM, P.O.Box 1, 3720 BA Bilthoven, the Netherlands. b TNO, P.O.Box 6011, 2600 JA Delft, the Netherlands c Department of Air Quality, Agricultural University, WAU, P.O. Box 8129, 6700 EV Wageningen, the Netherlands d Department of Physical Geography, University of Utrecht, P.O.Box 80.115, 3508 TB Utrecht, the Netherlands c KEMA, P.O.Box 9035, 6800 ET Arnhem the Netherlands f ECN, P.O.Box 1, 1755 ZG Petten, the Netherlands
Abstract Particle deposition to forest was studied using experimental and modelling results. Results show that the deposition of particles to forests has been underestimated until now. Particle deposition makes out reasonable contribution to the total deposition of acidifying components and base cations to forests. It was estimated that at Speulder forest the contribution of dry particle deposition to the total deposition was 18% for SO4, 38% for NO3, 23% for NH4, 56% for Na, 47% for K, 69% for Ca and 65% for Mg.. Deposition of compounds via fog at the Speulder forest was estimated to be small (<5%). Results of the canopy exchange experiments at Speulder forest and of the comparison of atmospheric deposition estimates and throughfall estimates at Speulder show that net-throughfall fluxes and atmospheric dry deposition estimates are reasonably in agreement for all components, except nitrate. There is still large uncertainty in canopy exchange processes of oxidised nitrogen and in deposition estimates of the different gases contributing to the total oxidised nitrogen flux.
1. INTRODUCTION In 1991, the results of the second phase of the Dutch Priority Programme on Acidification were reported in [ 1]. Although much knowledge had been gained during the approximately six years of research, some serious gaps remained. Therefore, a third limited phase was initiated to fill in the most important gaps [2]. One of the main uncertainties was recognized to be the estimation of atmospheric input of acidifying aerosols and base cations to ecosystems. Together with wet deposition, particle dry deposition is responsible for the atmospheric load to
116 ecosystems of compounds such as sulphate, nitrate, chloride and ammonium, base cations such as calcium, magnesium, sodium and potassium. Deposition of particles containing SO4, NO3, C1 and NH4 contribute to the potential acidification and eutrophication (nitrogen components) of ecosystems. Compared to gaseous deposition of acidifying compounds, particle deposition velocities and fluxes are usually found to be small. However, it is believed that the dry deposition velocity of small particles and therewith the fluxes is currently underestimated for very rough surfaces, like forests [3-5]. Current knowledge is therefore insufficient to give an adequate assessment of the dry deposition of particulate sulphur and nitrogen over the Netherlands and Europe. There is need for quantification to evaluate critical load exceedances and abatement strategies for atmospheric pollution. Base cation deposition may be of importance for nutrient cycling in softs and ecosystems and may also neutralise acid input. Base cation input is therefore important in the determination of critical loads and/or critical load exceedances for ecosystems. Ecosystems receiving a high atmospheric input of base cations have higher critical loads than those receiving smaller inputs [6]. Base cations are usually found in the coarse fraction of ambient aerosols. Their deposition velocities are therefore large. In several studies where throughfall fluxes are compared with atmospheric deposition estimates, large differences between the two estimates for deposition have been found [7-9]. Establishing a link between the two is useful because it provides a way to estimate soil loads from atmospheric deposition estimates on the one hand, and it allows the use of the relatively simple and cheap throughfall method to determine atmospheric deposition on the other. The link between atmospheric deposition and soil loads is important because critical loads refer to soil loads and because atmospheric deposition estimates provide a link with emissions. Thus if critical load exceedances are used to estimate emission reductions, the relation between atmospheric deposition and soil load should be known. It was suggested in [8] that beside canopy exchange processes, aerosol and/or fog and cloud water deposition might be important processes contributing to the observed differences between atmospheric deposition estimates and throughfaU measurements. The three main research questions that have been addressed in this project are: what is the contribution of "acidifying" aerosols to the total acid input of nature areas? what is the relation between atmospheric deposition estimates and throughfaU measurements and what is the contribution of aerosol deposition to the difference between the two estimates? - how important is the coarse particle flux (base cations) for the nutrient cycle in nature areas? The research needed to get satisfying answers to the three research questions was defined by a project group, in which research institutes and universities in the Netherlands participated with experience in the field of aerosol research. The work in the aerosol project is a joint initiative of the National Institute for Public Health and Environmental Protection (RIVM), KEMA -
-
(Laboratory for Environmental Research), TNO (Institute of Environmental Sciences), ECN (Netherlands Energy Research Foundation), RUU-FG (Utrecht University, Department of Physical Geography) and WA U-AQ (Wageningen Agricultural University, Department of Air Quality). It was anticipated in the project that by executing all available experiments in combination with a large modelling effort, a more accurate estimation of particle dry deposition velocities for rough surfaces would be obtained. A model was selected from existing models of aerosol deposition to forests [ 10]. The model had to be representative for the Dutch situation (pollution climate). From model simulations, insight was gained to the most
117 important processes involved in aerosol deposition. The main processes were tested by experimental research at the Speulder forest research site. The results of the experiments led to a verification of the model description and a basis for a parametrisation of Vd in terms of routinely available data. The parametrisation is used for the generalisation of aerosol deposition to other nature areas in the Netherlands. In this paper, the main results of the project will be summarised. Several publications form the basis of the results reported here, these are listed in the reference list.
2. EXPERIMENTAL The Speulder forest (52 ~ 151 N, 5 ~ 411 E) is located at the Veluwe, a large undulating area with forests and heathlands in the central part of the Netherlands. The measuring site covers an area of 2.5 ha, planted with Douglas fir of the provenance Arlington. By the end of 1993 the trees were 33 years old. The canopy is well closed with the maximum leaf area density at a height of 10-14 m. The one-sided LAI was between 149 and 10 for the measuring years 19871993 [11]. The stem density varies between 765 and 1216 trees per ha. In 1993 the trees were approximately 22 m high [12]. The stand is surrounded by a larger forested area of approximately 50 kmZ; the nearest edge is at as distance of 1.5 km. The stand itself is surrounded by Larix, Betula, Quercus, Pinus and Pseudotsuga trees, with mean tree heights varying between 12 and 25 m. A small clearing of 1 ha is situated to the North of the stand. Large source areas of SO2 and NOx are located about 200 km to the south-east (industrial Ruhr area) and about 100 km to the south-west (Rotterdam port) of the 'Veluwe'. The forested area is surrounded by large agricultural areas with intensive livestock farming acting as sources of NH3 due to volatilisation from animal manure. The location Speuld was equipped with two towers and measuring facilities (Figure 1). One tower is used for gas deposition measurements [ 13-15]. A sonic anemometer mounted on the top of the scaffolding at 36.5 m height is used to measure the horizontal and vertical wind velocity, wind direction and friction velocity and the sensible heat flux. A net radiation meter and a temperature/relative humidity sensor are mounted 1.5 m outside the scaffolding towards the south, at 35 and 33 m height, respectively. The two boxes house the gas monitors, two pulsed fluorescence SO2 analysers and two Luminox NO2 analysers, these are used to measure the vertical concentration gradient at 36, 32, 28 and 24 m height. Ammonia concentration measurements are made at three levels (34.5, 32 and 28 m height) using continuous flow denuders [16]. Ozone concentrations were measured at four heights (35, 32, 28 and 24 m) using chemoluminescence monitors. At 35 m the eddy correlation flux of ozone was measured using Gtisten ozone sampler and a Kaijo Denki sonic anemometer [15]. The second tower was completely used for the Aerosol project [ 17]. The tower is 36 m high and has a rectangle cross-section width a length and with of 2 and 1.5 m, respectively. During the experiments, meteorological data and dry deposition data for SO2, NH3 and NO2 were available from the concurrent project on trace gas deposition [14,16]. The experiments conducted within the framework of this project are listed in Table 1. In addition to the measurements used for estimating deposition, concentrations of SO4, NO3, NH4, C1, Na, K, Ca and Mg were measured in two size classes as 24 hr averages. (< 2.5 lam and > 2.5 ~m) during a period of nine months [ 18]. Concentrations of HNO2, HNO3, HC1, SO4, NO3, and NH4 were determined during several days with annular denuders [19] and once every week as hourly
118 values with wet rotating denuders [20]. Canopy exchange research using labelled S compounds (S 35) has been performed by ECN [20]. Fog composition, fog droplet diameter and liquid water content were measured by KEMA and ECN during several measuring campaigns [ 18,20,21]. In this way all ingredients determining throughfall and atmospheric deposition were measured. Aerosol deposition:
~=n~canemometer xo~,
etc.
Gas deposition:
) ~
s~
anemometer probe
'11
I ~.c~
qC:~r--.............. 35 m: tunnelsampler; PM-IO; cyclone; filterpack, DFM and a wet-only sampler, ADS 34 m: S04 denuder ................ 32 m: ADS
I1 =
,,~ JT/RH "t ~
36 m, reference height:
so2.
NHS.
eddy correlation NO2 32 m, SO2, NO2
r valve system 29 m: tunnelsampler; PM-10; cyclone; filterpack, DFM sonic anemometer probe~ liquid water content sensor
FSSP
"= < ~ ................. 28 m: filterpack; mist sampler
(L.
ASASP-X, sonic anemometq
-
-= -~- ........ 28 m0 SO2, NO2, NH3
26 m: SO4 denuder, aerosol concentrations "= ~
-
................. 24 m: filterpack, wet denuder (HNO3, HNO2, HCI, SO2)
-.,
24 m, SO2, NO2, NH3
22 rn: SO4 denuder; ADS; DFM __ ASASP-X, sonic anemom~
~I~--: ............... 21 m: tunnels,ampler; PM-10; cyclone 19 m: Pb-214:High Volume sampler, (artificial) branches
J<"T"
A
l
throughfall
l
l
A
A
scaffoldingwith SO2, NO2 monitors:
,
manipulation experimen~bin
transport hoist
L--a:leposition filter method (DFM) ,x~xanches and artificial branches: rinsing experiments
Figure 1. Schematic of the experimental set-up at Speulder forest. The campaigns with the experiments were executed in December 7-12 1992 and during April 18 to May 10 1993 9ThroughfaU, gas deposition and aerosol concentrations were measured continuously during a half year (26 November 1992 - 10 May 1993). The model used was that published by [25] and slightly modified by KEMA [ 10]. Parametrisations derived from model and measurement results were implemented in the RIVM models for estimating deposition of acidifying compounds at the Speulder forest and to nature areas in the Netherlands [5]. The experimental results were used to test and develop aspects of the model.
119 Table 1. Overview of experiments performed at the Speulder forest to quantify particle dry deposition. Technique
Species
NO3/SO4 (NI-h)2SO4 Na, K, Ca, Mg Eddy correlation LWC Fog droplets Particles accumulation on 214pb branches SO4 ... Mg
Size range Time (ttm) resolution < .9 2 hrs < .9 30 rain <2.5; <10; tot. 48 - 72 hrs 3 -45 1s 1 - 95 15 rain 0.1 - 3 0.1 s < 1 3 hrs
Sampling heights (m) 24, 30, 35 22, 26, 34 21, 29, 35 28 28 18, 25 18
0-*r
days
DFM Throughfall
0 - *r 0 - *r
7 days max. 1 week 1 hr
11, 15, 17, 13 19 21, 29, 35 14 1.5 30 24
< 2.5
24 hrs
<2.5, 2.5-15
12 hrs 24 hrs
Gradient
Na, K, Ca, Mg SO4 ... Mg
wet denuder
HNO3, HNO2, HCI, SO2 ADS SO4 ... Mg I-INO3, HNO2, HCI, SO2 aerosol sampler S04 ... Mg
3. EXPERIMENTAL
RESULTS
COMPARED
TO
Successful measurements 39 202 30 1637 1637 .9 26
Error estimate flux (%) 75 100 20 50 50 - 150 60
Reference
60
[18]
>25 30
[19] [24]
1000
15
[20]
34
35
40
[ 19]
21, 34 26
? -280
30 5
[18]
MODEL
RESULTS
[22] [20] [19] [20] [20] [23] [20]
For each experiment model calculations were made using the modified Slinn model with forests characteristics estimated for the Speulder forest, size distributions from Table 2, actual meteorology and concentrations measured above the forest and the exact periods the experiments lasted. The individual model results are compared to measurements in [ 10]. The experimental work was aimed at evaluating different aspects or process descriptions of the model. Most processes involved in particle deposition differ in their dependency on particle size. The best way to evaluate individual processes is to use measurements representative for different particle size classes. However, as discussed in [10], the size distribution for which experiments are representative are uncertain. Furthermore, uncertainty in measurement results is usually so large that it is difficult to draw conclusions from a comparison. The experiments made here for evaluation of the modelled deposition mechanisms can be divided into four categories: - 214pb measurements and eddy correlation measurements of particle deposition, representative for small particles with diameters smaller than 1 pm - NO3 and SO4 measurements which are representative for particles with a bimodal distribution, with most of the total mass in the size below 1 I.tm - measurements of total base cation deposition representative for large particles - fog deposition measurements, representative for large particles. Even though the MMD for which the four categories are representative differ much (Table 2), none of these is representative for a single deposition mechanism. All four are mainly determined by impaction and interception, and sedimentation in addition for fog deposition.
120 Table 2. Component specific size distributions (mass median diameter, MMD, and geometrical standard deviation t~g) derived from measurements in the Netherlands. Component MMD ~ ) og .... 214pb 0.35 2.0 Ca 7.73 3.47 Mg 5.92 2.73 K 2.64 1.84 Na 5.12 2.64 SO4 < 2.5 lain 0.6 2.2 SO4 > 2.5 lun 4.5 1.6 NH4 < 2.5 lain 0.6 2.2 NH4 > 2.5 Inn 4.0 1.6 NO3 < 2.5 lain 0.6 2.3 NO3 > 2.5 lain 4.5 1.6 fog droplets (December) 19.4 fog droplets (February) 7.4 The factor which is important for particle deposition, next to the size distribution of particles, is the friction velocity. If the size dependency and the u, dependence of the dry particle deposition velocity is similar for the measurements and the model results, this might serve as some sort of validation of the most determining processes. Figures 2 and 3 show the u, dependence of different experimental determined Vd and modelled Vo, respectively. In general the experimental and model results show a distinct influence of the size distribution and u, on the dry deposition velocity of particles. Both figures show that Vd (2~4pb) < Vd (SO4) < Vo (NO3) < Va (fog). Furthermore, the figures show similar relations between Va and u, for three of the categories distinguished above, although the variation in measured Va per u, class can be very high [10]. The category with base cations is not considered because no u, dependent measurements were made. The deposition velocity of fog is proportional to u, 2 indicating that impaction is the most important process determining Vd. Furthermore, sedimentation is important. The Vd values of other compounds are proportional to u, or a weak function of u, (~u,~a), indicating no distinct process is dominating rather a mixture of processes is occurring simultaneously [26]. Results are in line with other investigations, showing that particle deposition to forests can be considerable with Vd of several cm s-~ [27-29]. The relationships with u, found for different particle size classes provide confidence in the experimental results. This can only be considered as indication, because errors in deposition estimates usually show a high correlation with u,, because factors influencing the error in the deposition estimates highly correlated to u,. Evaluation of the integral model results can only be done with statistical parameters, including the uncertainty in modelled and measured values. The uncertainty in model results and the sensitivity of model results on input and model parameters has extensively been described in [10]. The overall error in modelled Vd integrated over the size distribution representative for acidifying aerosols is estimated about 65% [ 10]. For base cations this error is somewhat smaller (60%) because of the contribution of the relatively well parametrised description of sedimentation. The uncertainty in model estimates is lower than or about equal to the uncertainty in measurement results, with the exception of the fog deposition measurements which are estimated to have smaller errors (20%). The fractional bias of the
121 means (the relative difference between the mean calculated and observed value) falls within these limits. The relatively large sensitivity of the model and an inaccuracy of the same order in measuring results cause that a perfect 1"1 correspondence between both cannot be expected. Statistical testing of the difference between modelled and measured values is done using the non-paramatric sign and Wilcoxon tests for paired samples. Both test revealed no significant differences between the mean values of modelled and measured fluxes or Va'S [ 10]. This is of course mainly the result of the large standard deviations in measuring and modelling results. 0.1 0.08
/
.~ 0.06 ... 0.04 >
7
0.02 0 -0.02 0
0.2
0.4
0.6
0.8
=
fog
[3
NO3 (TNO)
*
SO4 (TNO)
<)
SO4 (ECN)
1
9
214Pb
u* (m/s)
Figure 2. Experimental determined Va versus u,. 0.1 0.08 o.o6 0.04
S
E no > 0.02
ol
-0.02 [ 0
0.2
0.4
0.6
0.8
U* (m/s)
=
fog
C
NO3 (TNO)
9
SO4 (TNO)
<>
SO4 (ECN) 214Pb
Figure 3. Summary of modelled Vd versus u,. It can be concluded that there are no strong indications for a significant underestimation or overestimation of the modelled Va compared to measured values. The fact that the model is reasonably capable of describing a similar response of Va to u. as the measurements for different particle diameter ranges provides confidence in the process descriptions. Furthermore, it is concluded that particle Va to forests and probably other rough surfaces are high. Average Vd values for half a year for fine particles at Speulder forest range from 1 to 2 cm s -~ (SO4, NO3, and NH4) with daytime values being 1.3 + 1.2 cm s~ and night-time values being 1.0 + 1.4 cm s~ (SO4). Vd values for coarse particles are about 5 cm s~ with daytime values of 5.1 :!: 3.9 cm s ~ and night-time values of 4.8 + 4.0 cm s~. In comparison, for the same period, Vd values for SO2, NH3 and NO2 were 1.5, 2.5 and 0.1 cm s l, respectively. This means that aerosol Va to forest canopies in The Netherlands is underestimated until now with a factor of 2 to 3. For forests in Europe this is even higher, taking the EMEP model results [30]. It was estimated that at Speulder forest the contribution of dry particle deposition to the total deposition was 18% for SO4, 38% for NO3, 23% for NH4, 56% for Na, 47% for K, 69% for Ca and 65% for Mg.. Deposition of compounds via fog at the Speulder forest was estimated to be small (<5%).
122 3.1 Parametrisation The dry deposition velocity for particles was parametrised using the modified Slinn [25] model, for routinely available data. The general form for Vd at 50 m height is:
Vd = 1
v.
(1), where V,~ can be estimated from:
Vd, m . u'2 , E
(2)
Uh
+ Ra(50)
Uh is the wind speed at canopy height, E is given for different components and conditions in [ 17]. Figure 4 shows parametrised and modelled values for the different Zo values at Speulder forest. The comparison between modelled and parametrised values is very good, indicating that the most important parameters used for generalisation to other forests are well represented in the parametrisation. 2.42.?. ~ 2.0 ~ 1.8 ~ 1.6 ~
.o4.2 -
~1.00.80.6 -~ 0.4 0.2 0.0
1 1.0
ZO
Ira]
Figure 4. Parametdsed (open circles) and modelled (line) Vd values (cm sa) as a function of Zo for the whole dataset obtained at Speulder forest.
4. C O M P A R I S O N W I T H T H R O U G H F A L L
Net-throughfall fluxes were estimated using wet deposition and throughfaU measurements during November 1992 to May 1993 [24,33]. Dry deposition for different components was estimated using the inferential technique with daily averaged measured concentrations and hourly averaged Vd calculated with the parametrisation given in section 3.2 for particles and those presented in [35] for gases. Comparison of net-throughfall estimates and dry deposition estimates are shown in Figure 5. Net-throughfall estimates were corrected for the influence of canopy exchange processes by applying the van der Maas/Ulrich model [31-33].
123 2500 2000
_~ 15oo I000 n-,
500
ntf SO4
m
dep SOx
net
ntf NO3
I--'Igas
dep NOy
ntf NH4
I~ aerosol
dep NHx
m
fog
th ro u g h fa II (A)
o
700 600 500
~. 400 300 _~ 2 0 0 Id.
IO0 0
nft Na
dep Na
I~
ntf Ca
net throughfall
dep Ca
ntf K
dep K
ntf Mg
r--1 p a r t i c l e d e p o s i t i o n
dep Mg
[
~) Figure 5. Net-throughfall estimates compared with dry and fog deposition estimates for acidifying components (A) and base cations 03) (tool ha "l al). Figure 5 (A) and 03) show that net-throughfall estimates, corrected for canopy exchange, are in good agreement with dry and fog deposition estimates. The only exception to this is nitrate: these estimates are significantly different. It has been shown in [33] that most of the experiments at Speulder forest point in similar directions regarding canopy exchange. Although the estimates of the absolute amounts of components retained or leached in the canopy may differ depending on experiment, the average values give a good picture of the situation at Speulder forest. It shows that H § is taken up by the canopy, which is accompanied by leaching of K and to a smaller extent Ca and Mg. SO2 taken up by stomata is eventually leached again, whereas NH3 taken up via stomata is almost completely retained by the canopy. Oxidised nitrogen components are taken up by the canopy, especially NO2. Results of the experiments show that NO3 uptake is negligible small, whereas NH4 is taken up, in exchange with K, Mg or Ca. Na and C1 are considered inert. Highest uncertainty in canopy exchange estimates relates to estimates of the nitrogen components [33]. Observed differences between dry and fog deposition estimates from micrometeorological measurements and inferential modelling on the one hand and net throughfaU fluxes on the other hand can not be regarded exclusively due to canopy exchange. Dry deposition estimates from micrometeorological measurements and inferential modelling are uncertain through errors in the air concentration measurements [34], their sometimes small time coverage and the uncertainties associated with the parametrisation of the dry deposition velocities [10,17]. Fog deposition
124 estimates are uncertain due to uncertainties associated with the estimation of water fluxes and the measurement of the average chemical composition of the fog droplets [21]. Uncertainties associated with the throughfall method when used for estimating dry and fog deposition include the dry deposition to the forest floor and understorey vegetation, dry deposition directly onto the throughfall gutters, the representativety of the throughfall sampling, the wet deposition estimate, the stemflow contribution and canopy exchange processes [8]. With canopy exchange processes being the only exception, mentioned factors probably contributed only to a very small extent to the uncertainty in the throughfall dry and fog deposition estimates in this study. Thus the combination of throughfall measurements and the empirical model lead to deposition estimates and soil load estimates which are at least as accurate as deposition estimates using other techniques. There is one drawback that is that the basic assumptions in the empirical model are not properly evaluated under different conditions (pollution climates, etc.).
6. CONCLUSIONS The Slinn [25] model was selected from a number of models able to estimate particle deposition to forests. The model was used to determine the most important processes involved in the deposition of acidifying aerosols and base cations. The model formulation was slightly modified based on a comparison of different process descriptions used in the different models. Model results were compared with results of several experiments, comprising eddy correlation fog and particle measurements, vertical gradient measurements of sulphate, nitrate and base cations; and different accumulation experiments, such as 2~4pb measurements, leaf washing experiments, deposition plate measurements and throughfall measurements. Both model results and experimental results showed a strong dependence of the dry deposition velocity of particles on particle size and friction velocity. It was found that Vd (214pb) < Va (SO4) < Va (NO3) < Va (base cations) < Vd (fog), in line with the size distributions. The deposition velocity of fog and base cations is proportional to u. 2 indicating that impaction is the most important process determining Vd. Furthermore, sedimentation is important. The Va values of other compounds are proportional to u, or u. ~-, indicating no distinct process being most important rather than a mixture of processes. It was concluded that the model results and measurement results are in good agreement. The model was used to derive a parametrisation of the deposition velocity of aerosols in terms of routinely available data. The parametrisation was used to generalize aerosol deposition to other nature areas in the Netherlands. It was shown that deposition of fine particles is an important pathway for acid input to forests. It was confurned that dry deposition velocities of particles to forests and probably other rough surfaces are high. Half year average Vd values for fine particles at Speulder forest ranged from 1 to 2 cm s ~ (SOn, NO3, and NH4) with daytime values being 1.3 + 1.2 cm s"~ and night-time values 1.0 + 1.4 cm s~ (SO4). Vd values for coarse particles were about 5 cm s"~ with daytime values of 5.1 + 3.9 cm s~ and night-time values of 4.8 + 4.0 cm s~. In comparison, for the same period, Va values for SO2, NH3 and NO2 were 1.5, 2.5 and 0.1 cm s 1, respectively. These results led to the conclusion that the deposition of aerosols to forest canopies in The Netherlands is underestimated until now with a factor of 2 to 3. Several field experiments were used to quantify canopy exchange. Using the new aerosol deposition estimates, differences observed between atmospheric deposition and throughfall fluxes of
125 acidifying compounds can almost completely be explained by canopy exchange processes. Uptake of H § and NH4 is compensated for by leaching of Mg, Ca and most of all K. SO2 taken up by stomata is eventually leached again, whereas NH3 taken up via stomata is not leached from the canopy. Oxidised nitrogen components are taken up by the stomata in the canopy, especially NO2. Whether NO3 is taken up is uncertain. Na and C1 are considered as inert. Highest uncertainty is found in the estimates of the nitrogen components.
7. REFERENCES
[ 1] Heij G.J. and Schneider T. (1991) Studies in Environmental Science 46. Elsevier, Amsterdam. [2] Heij G,J. and Schneider T. (1992) report no. 300-01, RIVM, Bilthoven, the Netherlands [3] Wiman, B.L.B., Unsworth, M.H., Lindberg, S.E., Bergkwist, B., Jaenicke, R., Hansson, H.C. (1990) J. Aerosol Sci.,21313-338. [4] Erisman J.W. 1992. Ph.D. Thesis, University of Utrecht, Utrecht, the Netherlands. [5] Erisman J.W. (1993a) Water Soil Air Pollut.,71,51-80. [6] Hettelingh J.P., Downing R.J. and Smet P.A.M. de (1991), RIVM, Bilthoven, The Netherlands. [7] Ivens W.P.M.F. (1990), Ph.D. Thesis, University of Utrecht, Utrecht, the Netherlands. [8] Draaijers G.P.J. and Erisman J.W. (1993),Atmospheric Environment,27A,43-55. [9] Erisman J.W. (1993b), Water Soil Air Pollut.71,81-99. [10] Ruijgrok W., Tieben, H., Eisinga, P. (1994) Report no. 20159-KES 94-xxxx, KEMA, Amhem, The Netherlands. [11] SteingrSver E.G. and W.W.P. Jans (1994, IBN-DLO Research Report no 94/3, ISSN:09286896, ? pp. [12] Jans W.W.P., G.M. van Roekel, W.H. van Orden and E.G. SteingrSver, 1994. IBN-DLO Research Report no 94/1, ISSN:0928-6896, 58 pp. [13] Zwart H.J.M.A., Hogenkamp J.E.M., Mennen M.G. (1993) Report no. 722108001, RIVM, Bilthoven, the Netherlands. [ 14] Erisman J.W., Mennen M., Hogenkamp J., Kemkers E., Goedhart D., Pul A. van, Boermans J. Duyzer, J.H., Wyers, G.P. (1994), Report no. 722108002, RIVM, Bilthoven, The Netherlands. [ 15] Westrate and Duyzer, 1994 [ 16] Wyers G.P., Otjes R.P., Slanina J. (1993a) A continuous-flow denuder for the measurement of ambient concentrations and surface exchange fluxes of ammonia. Atmospheric Environment, 27A,2085-2090. [17] Erisman, J.W., Draaijers, G.J.P., Duyzer, J.H., Hofschreuder, P., Leeuwen, N. van, R0mer, F.G., Ruijgrok, W., Wyers, G.P. (1994), Report no. 722108005, RIVM, Bilthoven, The Netherlands. [18] RSmer, F.G. and te Winkel, B.W. (1994), report 63591-KES/MLU 93-3243, KEMA, Arnhem, The Netherlands. [19] ] Hofschreuder, P, Bogaard, A.J., Hartog, K.D. (1994). Report no. R 637, Agricultural University Wageningen, Wageningen, The Netherlands. [20] Wyers, G.P., Veltkamp, A.C., Vermeulen, A.T., Geusebroek, M., Wayers, A., M~31s,J.J. (1994), Report no. ECN-C--94-051, ECN, Petten, the Netherlands.
126 [21] Vermeulen, A.T., Wyers, G.P., RSmer, F.G., Draaijers, G.P.J., van Leeuwen, N.P.M., Erisman, J.W. (1994) Report no. ECN-C--94-0xx, ECN, Petten, the Netherlands. [22] Duyzer, J.H., J.H. Weststrate, K. Beswick, M. GaUagher (1994) IMW-TNO report, in prep. [23] Beswick, K.M., Gallagher, M.W., Hummelshoj, P., Pilegaard, K., Jensen, N.O., Duyzer, J. (1994). In: Proc. EUROTRAC Symposium 94, Garmisch-Partenkirchen, Germany. [24] Leeuwen, N.P.M. van, Draaijers G.P.J., Bleuten, W., Hansen K. (1994), report no. xxxxx Department of Physical Geography, University of Utrecht, the Netherlands. [25] Slinn, W.G.N. (1982), Atmospheric Environment,16,1785-1794. [26] Aalst, R.M. van (1986). In: Aerosols (Lee, S.D., Schneider, T., Grant, L.D., Verkerk, P.J., eds.), Lewis Publ. Chelsea (MI), 933-949. [27] Hicks, B.B., Wesely, M.L., Durham, J.L., Brown, M.A. (1982). Atmospheric Environment,16,2899-2903. [28] Waraghai, A. and Gravenhorst, G. (1989), In: Mechanisms and effects of air pollutant transfer into forests (Georgii, H.W., ed), Kluwer Academic Publishers, Dordrecht, The Netherlands. [29] Sievering, H., Enders, G., Kins, L., Kramm, G., Ruoss, K., Roider, G., Zlger, M., Anderson, L., Dlugi, R. (1994). Atmospheric Environment,28,311-315. [30] Iversen T., Halvorsen, N., Mylona, S., Sandnes, H. (1991. MSC-West, the Norwegian Meteorological Institute, Oslo, Norway [31] Maas, M.P. van, Breemen, N. van, Langenvelde, I. van (1991) Department of soil science and geology, Agricultural University of Wageningen, The Netherlands. [32] Ulrich, B. 1983. In: Effects of Accumulation of Air Pollutants in Forest Ecosystems. (Ulrich, B. and Pankrath, J. Eds.). D. Reichel Publ. Co. 33-45. [33] Draaijers G.P.J., Erisman, J.W., Leeuwen, N.F.M. van, Rt~mer, F.G., Winkel, B.H. te, Wyers, G.P. (1994) Report no. 722108004, RIVM, Bilthoven, The Netherlands [34] Arends, B.G., Wyers, G.P., Mennen, M.G., Erisman, J.W., Rt~mer, F.G., Hofschreuder, P., Duyzer, J.H. (1994) Report no. ECN-C--94-058, ECN, Petten, The Netherlands. [35] Ersiman J.W., van Pul A., Wyers P. (1994)Atmospheric Environment,28,2595-2607.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BE All rights reserved.
127
Deposition of aerosol to coniferous forest G.P. Wyers, A.C. Veltkamp, M. Geusebroek, A. Wayers, J.J. M61s Netherlands Energy Research Foundation ECN, P.O. Box 1, 1755 ZG Petten, The Netherlands
Abstract The deposition velocity of ammonium(bi)sulphate (MMD 0.8 pm, og 2) was derived from vertical concentration gradients using flux-profile relationships. The deposition velocity has a near-linear dependency on u, and an average value of 0.012-0.015 m s I. The dry deposition of 214pb was investigated by measuring airborne activity and activity accumulated on Douglas fir branches. This radioisotope is a natural tracer for submicron aerosol (AMD 0.4 pm, og 3). The correlation of V~ with u, is poor. The average deposition velocity ranges from 0.0061-0.01 I0 m s-~, with a best estimate of 0.0073 m s -I, a factor 1.6-2.1 smaller than the deposition velocity for sulphate. The difference is regarded to be due to the difference in size distributions of214pb and sulphates.
1. INTRODUCTION Wind tunnel experiments have shown that the dry deposition of aerosol is strongly dependent on particle size [1]. Because these studies indicated very small deposition velocities less than 1 mm s 1 for particles in the range 0.1-1 pm, the deposition of submicron aerosol has been considered relatively unimportant, as compared to deposition of gaseous compounds. It has been reported however by Hicks et al. [2] that deposition velocities for submicron aerosol can be considerably larger than 1 mm s 1 over vegetative surfaces in the field. Few measurements of aerosol deposition have been made over forests. Hicks et al. [3] measured sulphate fluxes over a deciduous forest by eddy correlation. They found that V~ ranged from virtually 0 at night to a maximum of 1 cm s 1 at daytime with a long-term average value of 0.6 cm s -1. Similar values are reported by Wesely et al. [4] for a pine forest: the deposition velocity ranges from 0.90 cm s 1 for moderately unstable conditions to 0.48 for a neutral atmosphere. In order to estimate the deposition velocity of submicron aerosol over forests, two experiments were performed at the site Speuld, located in the center of The Netherlands. This site is a 2.5 ha monoculture of Douglas fir with a mean tree height of 20m. The canopy is dense and had in 1992 a maximum one-sided Leaf Area Index (LAI) of 10.7 (Jans and Steingr6ver, pers.comm.). The deposition velocity of ammonium(bi)sulphate was determined from vertical concentration gradients measured above the canopy. The natural
128 radio-isotope 214pb which occurs attached to aerosol surfaces was used as a tracer for
submicron aerosol. The deposition velocity of this radon daughter was determined by counting the airborne activity and the activity deposited on the needle surfaces.
2. DRY DEPOSITION OF SULPHATES 2.1. Experimental Vertical concentration gradients of ammonium(bi)sulphate and sulphuric acid over the forest were measured by CuO-coated thermodenuders with a time resolution of two hours. Three instruments were placed on a scaffold at heights of 22m, 26m and 34m above the forest floor. Before starting the measurements over the forest, three thermodenuders were run simultaneously in Petten for a period of two weeks to examine systematic and random errors in the determination. The results indicate systematic differences up to 6%. All measurements were corrected for the detected bias on the basis of this intercomparison. After correction, the average of the relative standard deviation of the triplicate measurements amounted to 4%. Turbulence intensity and sensible heat flux were measured by eddy correlation (Erisman, pers. comm.). 2.2. Results From 17 May to 23 August 1993 measurements of the vertical concentration gradient were performed every two hours. The average concentration of ammonium(bi)sulphate in the period May - August 1993 was 4.8 _+ 3.2 pgSO4 m 3. The sulphuric acid concentration was very low, mostly below 0.2 pgSO4 m 3 and never exceeding 0.8 pgSO4 m a, indicating a very high degree of neutralization of sulphate by ammonia. Only the measurements collected in the period 23 July - 23 August were used to calculate deposition velocities for sulphate aerosol, since these were collected while all equipment was operating properly. For the calculation of c, modified flux-profile relations were applied, which have been derived for this site from a comparison between heat fluxes directly measured by eddy correlation and calculated from temperature gradients by Bosveld [5]. A more recent study by Weststrate and Duyzer [6] involving a comparison of eddy correlation measurements of fluxes of ozone and sensible heat with vertical profiles of ozone and temperature yielded similar results. C, was calculated by linear regression of c(z) versus the logarithm of the stability-corrected height, with the correction factor ct calculated from the effective height zar [5], using a displacement of 15m and a reference height Zr of 23.5m [6]. The deposition velocity Vd and the surface deposition velocity Vds are calculated for the lowest measurement height z = 22m. The surface deposition velocity Vds is obtained from Vd by subtraction of the contribution of turbulent transport and is defined as (Vdl-ra) -1 [4]. 2.3. Discussion A time series of Vd, Vd~ and u, is given in Figure 1 for the period 23 July - 2 August 1993. Despite the large scatter in Vd, a weak positive correlation can be observed with the friction velocity. The scatter in Vds is considerably larger than in Vd.
129
.2
Vd, Vds ( m / s
u" ( m / s ) "
.1
"
v,
1, 1.5
v
v v V
V V
V
V V
v
V
V
V
V V
V
V
0
I
-.1 i
"~
-.2
Figure 1.
v
v
f -~Vd
v Vds
--u*
0
Time series of deposition velocity (Vd), surface deposition velocity (VdQ and friction velocity (u,) above the Speulder Forest, 23 July - 2 August 1993.
A random error of 4% in the individual concentration measurements may, under neutral atmospheric conditions, correspond with an uncertainty in Vd of 1.5 cm S1. Assuming that the nature of the errors is stochastic, the uncertainty can be reduced by averaging the deposition velocities for ranges of u,, provided u, is the determining factor. As the uncertainty in the individual gradient measurements is high, all measurements were sorted on u,, and for each u,-range a 50-percentile and an average Vd were calculated. Initial selection of the data on sufficient homogeneous fetch and u > 1 m s" yielded 339 measurements for further processing (Table 1). Both Vd(50) and average Vo display a positive correlation with u,-class, but the average is a factor two larger than the 50-percentile of Vd, which is caused by a few extreme values. This is obvious from the large standard deviation which is mostly exceeding the value of the average deposition velocity. The true V d is therefore better approximated by the median than by the average. The Vd(50) increases with increasing friction velocity and can be approximated by the power function: Vd(50) = 0.0526 U,T M
with r 2 - 0.988
(1)
At low values of u, (<0.4 m s") however approximately 25% of the measurements yield apparent upward fluxes, which probably cause an underestimation of Vd at low friction velocities. Sorting of the measurements was also attempted using rh or z o as a key. Sorting on rh yielded a poor, inverse relationship with Vd(50), which is due to the fact that high relative humidities occur at night, when turbulence intensity is low. A similar problem was
130 encountered with sorting on Zo, since the roughness length is dependent on wind direction, which is at this site strongly correlated with wind speed. It can be concluded that u, is the dominant parameter determining the magnitude of the deposition velocity.
Table 1.
Deposition velocities for different m-ranges. Selection on u ( > 1 m s -1) and sufficient fetch E
u,-range m sx
u,
Va(50) m s -~
Va
S"1
0.0-0.1 0.1-0.2 0.2-0.3 0.3-0.4 0.4-0.5 0.5-0.7 0.7-1.0 all data
0.058 0.153 0.245 0.348 0.449 0.610 0.822 0.357
0.0005 0.0021 0.0059 0.0092 0.0184 0.0196 0.0365 0.0062
0.0009 0.0049 0.0139 0.0300 0.0261 0.0306 0.0584 0.0216
m
m
s S1
m
S"1
0.0012 0.0016 0.0437 0.0600 0.0349 0.0462 0.0827 0.0480
28 64 71 50 53 42 31 339
Changes in concentration with time may result in artefact upward or downward fluxes. A further selection was made by setting the maximum error due to storage changes at 10% of the flux (5c/5t (z2-z~) F ~ < 0.1) [7]. This selection also served to eliminate spurious concentration measurements. After this selection 217 data remained for further evaluation. A further selection was aimed at removal of unrealistically high deposition velocities by limiting the maximum value of IVal to 2 ra~, i.e. two times the maximum deposition velocity for momentum. This led to disqualification of only 15 more measurements. The results are listed in Table 2. The median values for u, <_ 0.4 m s1 have become higher compared to the first selection which was based on windspeed and fetch only (Table 1). This increase in Va is partly the result of elimination of negative values resulting from poorly measured concentration gradients. The number of measurements resulting in apparent upward fluxes is now reduced to 15%. A possible bias towards higher deposition velocities could have been caused by removing measurements with storage errors exceeding 10% of the flux, since measurements representing small fluxes are preferentially removed. Both the elemination of apparent upward fluxes and the selection on storage errors lead therefore to higher values of V~(50), especially at low to moderate friction velocities (< 0.4 m s-1, see Tables 1 and 2). Therefore, the true value of the deposition velocity for u, < 0.4 m s1 lies probably in between the values listed in Tables 1 and 2.
131 Table 2.
Deposition velocities for different u,-ranges. Same selection criteria as in Table 1, with 5c/~t (z2-z~) F ~ < 0.1 and IVal < 2 r,,4
u~ rn s ~
u~ m s~
Va(50) m S"1
V~ m s~
s rn s -a
n
V~s(50) m s -a
V~ m s~
0.0-0.2 0.2-0.3 0.3-0.4 0.4-0.5 0.5-0.7 0.7-1.0 all data
0.156 0.248 0.345 0.450 0.615 0.834 0.400
0.0044 0.0088 0.0178 0.0211 0.0196 0.0344 0.0158
0.0071 0.0132 0.0216 0.0270 0.0248 0.0399 0.0227
0.0140 0.0219 0.0276 0.0313 0.0344 0.0557 0.0341
20 40 32 46 38 26 202
0.0034 0.0108 0.0092 0.0217 0.0239 0.0292 0.0167
0.0057 0.0114 -0.0062 0.0340 0.0370 0.0697 0.0256
s m s~
0.0403 0.0988 0.2259 0.2799 0.1353 0.1261 0.1865
Average and 50-percentile of the deposition velocities are now in reasonable agreement. The standard deviations have decreased and are of approximately the same magnitude as the average deposition velocities. The 50-percentile of the surface deposition velocity V~s is similar to that of the total deposition velocity Va, but the average Vas is highly variable and has a standard deviation several times its value. V~(50) and V~s(50) both show a high positive correlation with u. (Fig. 2). Linear regression as a function of average u, for each range yields: Vas = 0.0380 u~ - 0.0004 V~ = 0.0400 u~
with r2 = 0.895 with r2 = 0.903
(2) (3)
This near-linear relationship between deposition velocity and friction velocity is somewhat unexpected. When impaction is the only mechanism responsible for passing the viscous sub-layer, V~ should increase stronger than linearly with u. [8]. The observed linear relationship suggests that (a combination of) other mechanisms is responsible for this dependency on u.. Model calculations yielded an analogous near-linear relationship between Vd and u, [9], which support the validity of (3). For Va - 0-0.05 m s 1 much better agreement between observed and predicted deposition velocities was obtained using relationship (3) rather than (2). The results given in Figure 3 show a reasonable correlation for Vd between 0 and 0.05 m s 1, but also indicate large positive and negative values up to _+ 0.1 m s 1 which deviate strongly from the predicted behaviour and are very likely the result of erroneous concentration gradient measurements. Despite the possibility that deposition velocities at u. < 0.4 m s ~ are slightly overpredicted (see above), equation (3) is considered the most accurate that can be derived from these measurements. Also shown in Figure 3 are deposition velocities predicted using parametrizations by Erisman [10] and Wesely et al. [4]. The latter was derived for sulphate
132
.04
VdSO, Vds50 (m/s)
0
J
.03 /
.02 /
J
[]
/
J
J .01
0
i
0
Figure 2.
l
i
,
I
.2
I
'
i
,
I
i
I
I
I
I
.4 .6 average u* (m/s)
'
'
'
w
I
.8
[
l
'
~
'
'I
l l
1
Median deposition velocity (squares) and median surface deposition velocity (triangles) for S04 at different u,-ranges as a function of average u,. Selection criteria as listed in Table 2.
deposition on grassland in the U.S.A.; the parametrization by Erisman is currently used to estimate aerosol deposition in the Netherlands. These parametrizations lead to an underprediction of the deposition velocities by approximately a factor two (Erisman) to ten (Wesely et al.). Equation (3) was used to calculate average flux and deposition velocity from measured u, and SO4 concentration for the entire measurement period May-August 1993. This yielded an average deposition velocity of 0.0148 +_ 0.0082 m s 1 and an average flux of-0.0738 + 0.0618 lagSO4 m 2 sx, corresponding to approximately 240 mol SO4 ha ~ a 1. For reasons given earlier this can be considered an upper estimate. A lower estimate can be obtained by applying equation (1), which yields an average deposition velocity of 0.0120 + 0.0103 m s 1, only 20% lower than that derived from (3), corresponding with an annual flux of approximately 200 mol SO4 ha 4 a 1. The deposition velocities observed for sulphate aerosol over this forest are much higher than deposition velocities for submicron particles predicted from wind tunnel studies and are also high when compared to results from previous field studies over forests. For a deciduous forest with a similar tree height and a LAI at 50% of its summer maximum, Hicks et al. [3] found a long-term average deposition velocity of 0.6 cm s -1 with VJu,
133 I 0 Wyern el mL (this rel~Ortl
+ Wue~y et sl. (M86J
x Ednnm (111;031
Vdpar(22m) .05
.04
O
[]
_
O
[]
0
_
0
[]
O
]
.03
--
O
f-I
[]
O
-
[]
x
x
x
x
x
x
-
x~<~
_
•215
o
+
X
_
4"+
~
-.05
x
+ + ~
F
x
xx xo o o [] ~ o o~ ~.
++
.~.~_ +.
O
O
I-I
0 [] 0
o
~O ~x~
O
[]
o
[] 0
O []
+.,_~[.~+~-~ 7
0
oo
o ~
,,
boo Xoxo x ~ ~<Xx x x ,• ~ x o• . ~ x ~ ~'< x x ,~x~x xxX. ~ ~x x i x ix,~..+,~ ~y.~ v-~ .,X~ + X .i
,
[]
O
0
:]Z-"t ~a x~ [] o
~
I~
0~{
o o
Ox
[]
O O
o
~o~ []
[]
O
[]
0
O
%
x
-
[]
[]
01-I
O
[]
m
[] []
X
X
_]-
0
[]
O
.02
.01
O
O
[]
_ _ _
[]
n
--
X
•
x
x
x
[]
• X x
+,-
[]
0 o
x
T
.05
x
xo
x x x x
x
~<
+++
-~-,,.
1
T
l
.1
Vd(22m)
Figure 3.
Comparison of observed and predicted deposition velocities. Parametrizations from this report (equation (3)), Erisman [10] and Wesely et al. [4].
ranging from 0.001 at stable conditions to 0.02 for a moderately unstable atmosphere. The average wind speed however was only 2 m s 1. For the site Speuld an average of 0.04 is found for VJu, at an average wind speed of 3.6 m s 1 (for the period May - August 1993). Dry deposition of aerosol in The Netherlands is currently estimated using a parametrization derived by Erisman [10], which calculates the deposition velocity as a function of friction velocity and atmospheric stability. Application of this parametrization to the measurement period May - August 1993 results in an average deposition velocity of 0.0062 +_ 0.0051 m s-1. The results from the present study indicate that dry deposition of sulphate on coniferous forest is underestimated by more than a factor two. The deposition velocity is strongly dependent on particle size distribution, and it is possible that a few large particles lead to such high values. However, size distributions for antropogenic sulphate show that little mass is present as supermicron particles. Particle size distributions have not been measured during this experiment. In 1983 and 1984 size distributions of sulphate were determined with Andersen impactors during air pollution episodes in 1983 and 1984. The average can be represented by a log-normal distribution with a MMD of 0.8 pm and a Og of 2. Moreover, large particles > 5-10 pm will not pass
134 the air inlet of the thermodenuders. A second point worth mentioning is the very dense canopy of this forest. To avoid interference with the measurements, thinning of this forest was postponed, and the one-sided LAI now amounts to approximately 11 (Jans and Steingrover, pers. comm.). The canopy may therefore have become a very efficient receptor surface for fine aerosol. A third possible reason for the high observed deposition velocities is hygroscopic growth of the particles in the humidity gradient above the surface. Growth is under most conditions very fast compared to vertical transport times [11] and may shift the size distribution near the receptor surface to particle diameters which deposit very efficiently.
3. DRY DEPOSITION OF 214pb 3.1. Experimental 2~4pb is a daughter isotope of/=Rn, which itself belongs to the decay series of/3sU. After formation by decay of =6Ra in the subsurface, a2Rn may enter the atmospere, where it will decay according to the scheme: =2Rn (3.8d) ~ 218p0 (3.05min) ---> 214pb (26.8min) ---> 214Bi After decay of radon, the daughter isotopes will react rapidly with trace gases and vapours forming clusters or will attach to existing aerosol surfaces within 1-100 s [12]. The attachment rate is a function of particle size. The daughter isotopes of radon can thus be used as a selective tracer to examine the deposition of aerosol onto surfaces such as vegetation canopies. Fresh branches of Douglas Fir, collected near the measurement site, were mounted horizontally in a tower at 19 m height near the top of the canopy. After exposure for at least 3 h, twigs from the branches are rapidly (within 10 min) collected for radioactivity counting by gamma ray spectrometry. Airborne radioactivity is collected by sampling at 60 m3/h on a 20x25 cm glass fiber filter using a high-volume sampler. The high-volume sampler is installed at the tower near the exposed branches mentioned above. The inlet of the sampler is positioned at 19.5 m height. After approximately two hours of sampling, steady state of 214pb on the filter is reached, i.e. the 214Pb content on the filter is constant since collection and decay of this isotope on the filter are in equilibrium, assuming constant radioactivity levels in air during the sampling period. After steady state is reached, the filter is rapidly (within 10 min) prepared for radioactivity counting by gamma ray spectrometry. From simultaneous measurements of the airborne activity collected on filters by highvolume sampling and the activity accumulated on needles at the top of the canopy, a "local" deposition velocity Vo(loc), representative for deposition on 1 m 2 of needle surface at 19 m near the top of the canopy, is derived for submicron aerosol following Bondietti et al. [13]: Vd(1OC) = {L Vai r Eai r Aneedles}/ {S W Eneedles Aair} [m/(s m 2 needle area @ 19m)]
135 L
Vair Eair Eneedles mair mneedles S W
: : : : : : : :
decay constant 214pb sampled air volume counting efficiency gamma-measurement filter on 352 keV counting efficiency gamma-measurement vegetation on 352 keV counts on filter, recalculated to t-0 counts on vegetation, recalculated to t-0 specific leaf area (i.c. 6.0 m2/kg dry weight) dry weight vegetation sample
3.2. Results The emission of radon from the earth's surface is a function of the radium content of the subsurface. Emanations of radon from the subsurface in the Netherlands are very low. Sufficient deposited activity for a reliable measurement was only obtained in air masses which had travelled a long distance over land. Measurements were thus mainly performed during easterly and southerly winds. In the period December 1992 - May 1994 42 measurements of airborne and deposited activity were performed at the site Speuld. Of these, 26 measurements corresponded with wind directions from the sector 90-180. These measurements were used for further evaluation. In Figure 4 Vd(1OC) is shown as a function of friction velocity, measured above the canopy by eddy correlation.
Three measurements of the size distribution of 214pb, made at the site Speuld during easterly winds with a Sierra high-volume cascade impactor, confirmed that this isotope is associated with submicron particles. Assuming a log-normal size distribution, the Activity Median Diameter (AMD) and the geometric standard deviation og were estimated at 0.4 lam and 3 respectively. According to Porstend0rfer [12] the activity distribution of 2~4pb is unimodal and can be well approximated by a log-normal size distribution. Measurements of the activity distribution compiled by PorstendOrfer indicate an AMD of 0.15-0.39 pm and a Og of 2.3-2.8. A comparison with size distribution measurements of antropogenic sulphate (MMD of 0.8 pm) indicates that the size distribution for 214pb is characterized by smaller particle diameters. 3.3. Discussion In clean air masses from the North Sea or the Scandinavian countries anomalously high values were found for Vd(1OC) (0.0068+0.0055 rn/(s m 2 needle area). This was attributed to the presence of a large fraction of 218p0 and possibly 214pb in cluster mode (i.e., unattached to submicron particles) which deposits at a higher rate than accumulation mode aerosol [12, 14]. The attachment rate of radon daughters is related to the amount of aerosol surface present. In clean air masses the attachment rate will be considerably slower, resulting in a higher fraction of activity in cluster mode. Therefore only measurements in polluted air masses with (predicted) 36h-back-trajectories originating in the sector 90-180 were considered for studying the deposition velocity of fine aerosol.
The deposition velocities thus measured have to be corrected for the contribution by unattached activity. This correction is based on the following assumptions: an accommodation factor of unity, losses from recoil of 50%, a size distribution for sulphate aerosol as measured during air pollution episodes in 1982-1984 but with a reduction in
136 Vd(Ioc) (m/(s.m2 leaf area))
.01
.0075
.005
[]
[] []
0
.0025 [] []
l
0
Figure 4.
[] [] [] w
I
.3
[]
[]
[] [] 1
I
I
.6 u" (m/s)
i
,
I
.9
w
i
1.2
The local deposition velocity of 214pb (representative for 1 m 2 needle area @ 19 m) as a function of u, above the canopy.
number concentrations of a factor two, and a diffusion coefficient for clusters in air of 0.05 cm 2 s". This yields an attachment rate of 8 10 -3 s -]. Furthermore taking a minimum mixing layer height of 500 m and a maximum deposition velocity for cluster mode aerosol of 0.05 m s 1 (the average reciprocal aerodynamic resistance for this site) leads to a fraction of unattached activity of 1.25% or less. The deposition velocities for attached activity have been determined by subtracting the contribution due to 1% of unattached activity from the total deposition velocity. The Vd for cluster mode aerosol was calculated from (ra+rb)-'. A diffusion coefficient of 0.05 cm 2 s ~ was used to calculate r b. Brownian diffusion was assumed as the rate-limiting mechanism for transport of clusters through the viscous sublayer, so possible electrical interactions were ignored. By this correction the deposition velocity for attached activity is decreased by 6% on average relative to the total deposition velocity. The average V~(loc) for attached activity of 26 measurements obtained from December 1992 to February 1994 was 0.00290 _+ 0.00194 m s-] (average u, --- 0.47 m s-'). By subtracting the contribution from aerodynamic transport, a surface deposition velocity (see paragraph 2.3) was calculated [4]. The surface deposition velocity Vas(lOc) is similar in
137 magnitude to Vd(loc) and is 0.00320 + 0.00231 m s 1 on average. The measured V~(loc) shows only a poor correlation with u. (Fig. 4). For neutral-unstable conditions the average V~ is 0.00341 + 0.00196 m s ~, for neutral-stable conditions V~(loc) is lower, 0.00230 + 0.00174 m s 4. The largest uncertainty currently concems the extrapolation of Vd(loc) determined at 19 m to a Vo representative for the entire canopy. An attempt to derive this scaling factor was recently made Ruijgrok et al. [9], who estimated a value of 2.5. This factor was determined by calculating the collection efficiency for different layers in the canopy, taking into account the leaf area distribution, the activity distribution of 2~4pb, the wind profile in the canopy and the different deposition processes involved. Furthermore they assumed that the airborne activity is independent of crown depth. If only impaction is considered, a factor of 2.1 was calculated. Another estimate was obtained by examining the local sulphate flux to branches as a function of crown depth, which yielded a scaling factor of 3.7-3.8. Applying these scaling factors of 2.1-3.8 to V~(loc) yields average deposition velocities for 2~4pb of 0.0061 - 0.01 I0 m s -~, with a best estimate of 0.0073 + 0.0049 m s ~. These values are lower than the deposition velocities found for ammonium(bi)sulphate, which is estimated at 0.012-0.015 m s 1 . This is probably related to the difference in size distributions between sulphate aerosol and 2~4pb. Apparently deposition mechanisms are less efficient for the smaller particles with attached 214pb (AMD of 0.4 lam) than for antropogenic sulphate with an MMD of 0.8 lam. A comparison was made between deposition velocities measured for 2~4pb and deposition velocities calculated using the parametrizations by Erisman [10] and by Wesely et al. [4]. The parametrization by Erisman yields deposition velocities which are of approximately the same magnitude as the observed values for 2~4pb, but the correlation between observed and predicted values is poor. The parametrization by Wesely et al. [4] severely underestimates the deposition velocities.
4. C O N C L U S I O N S Deposition velocities for ammonium(bi)sulphate aerosol were derived from vertical concentration gradients measured over a coniferous forest. For different m-ranges the median and average deposition velocities were calculated. Deposition velocities range from _< 0.004 m s -1 for u, < 0.20 m s 1 to 0.034 m s -1 for u, between 0.70 and 1.0 m s 1. The median deposition velocity Vo(50) shows a near-linear increase with u,. Using this relationship the average Vo for the period May - August 1993 is estimated at approximately 0.015 m s 1. This can be considered an upper estimate, due to the possible overestimation of Vo at low turbulence intensity. As a lower estimate a value of 0.012 m sis found. It can be concluded that the deposition velocity for sulphate on coniferous forest in The Netherlands has sofar been underestimated by more than a factor two. The annual flux of sulphate aerosol to this site is estimated at approximately 200-240 tool SO4 ha ~ a 1. Measurement of the deposition of 214pb to branches at the top of the canopy indicates an average Vd(loc) of 0.0029 m s l m -2 leaf area at an average u, of 0.47 m s 4,
138 but only a poor correlation of V~(loc) with u.. Extrapolation of these measurements to the entire canopy results in an average V~ of 0.0061 - 0.0110 m s -1, with a best estimate of 0.0073 m s1. These values for V~ are a factor 1.6-2.1 smaller than deposition velocities derived from gradient measurements of ammonium(bi)sulphate. The difference is attributed to the smaller particle size of aerosol to which 214pb is attached.
Acknowledgements Many colleagues have contributed to this study. We are especially thankfull for the help of Henk Das who formulated the calculational procedures for measurement of low levels of 214pb activity and of Harry ten Brink who derived an estimate of the fraction of unattached airborne radioactivity. This study was carried out as part of the EUROTRAC-BIATEX programme and the Dutch Priority Programme on Acidification. Funding of the BIATEXprogramme by the Ministry of Economic Affairs and the Ministry of Housing, Spatial Planning and Environment is gratefully acknowledged.
REFERENCES Sehmel, G.A. and Hodgson, W.H. (1980) AIChE Symposium Series 76, 218-230. Hicks, B.B., Wesely, M.L., Durham, J.L. and Brown, M.A. (1982) Atmospheric Environment 12, 2899-2903. [3] Hicks, B.B., Matt, D.R., McMillen, R.T., Womack, J.D., Wesely, M.L., Hart, R.L., Cook, D.R., Lindberg, S.E., De Pena, R.G. and Thomson, D.W. (1989) J.Geophys.Res. 94, 13003-13011. [4] Wesely, M.L., Cook, D.R., Hart, R.L. and Speer, R.E. (1985) J.Geophys.Res. 90, 2131 [5] Bosveld, F.C. (1991) Report WR-91-02, KNMI, De Bilt. [6] Weststrate, H. and Duyzer, J.H. (1994) Report MW-R 94/104, TNO, Delft. [7] Fowler, D. and Duyzer, J.H. (1989) In: M.O. Andreae and S.D. Schimel (Eds.) Exchange of trace gases between terrestrial ecosystems and the atmosphere. John Wiley and Sons Ltd., 189-207. [8] Van Aalst, R.M. (1986) In: S.D. Lee, T. Schneider, L.D. Grant and P.J. Verkerk (Eds.) Aerosols. Lewis Publ., Chelsea MI, 933-949. [9] Ruijgrok, W., Tieben, H. and Eisinga, P. (1994) Concept report, KEMA, Arnhem. [10] Erisman, J.W. (1993) Water, Air & Soil Poll., 71(1/2) 81-100. [11] Khlystov, A., Ten Brink, H.M. and Wyers, G.P. (1993) Report ECN-C--93-011, ECN, Petten. [ 12] PorstendOrfer, J. (1994) J.Aerosol Sci. 25, 219-263. [1 ] [2]
[ 13] Bondietti, E.A., Hoffman, F.O.and Larsen, I.L. (1984) J.Environ.Radioactivity 1, 527. [14] Schery, S.D. and Wasiolek, P.T. (1993) J.Geophys.Res. 98, 22915-22923.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
139
Microscopic processes governing the deposition of trace gases and particles to vegetation surfaces
J. Burkhardt Bayreuth Institute of Terrestrial Ecosystem Research (BITOK), Department of Climatology, D-95440 Bayreuth, Germany present address: Institute of Terrestrial Ecology, Edinburgh Research Station, Bush Estate, Penicuik, Midlothian EH26 0QB, UK
Abstract The deposition of fine particles, consisting largely of ammonium and sulphate, as well as the deposition of SO 2 and NH 3, is often dependent on processes happening in the last mm of the atmosphere above the plant surface. Surface wetness plays an important role for trace gas deposition, and at the same time creates a link between particle and trace gas deposition, because water vapour condensation on leaf surfaces is initiated by hygroscopic aerosols. The interdependencies between these processes are described, and the plants role in regulation of deposition by their microroughness and transpiration are mentioned. The importance of acidity and hygroscopicity of deposited particles and salt-forming gases and their possible impact on plants, especially by a kind of 'wick effect', is outlined.
1. Introduction Wind velocity and aerodynamic roughness of canopies are important for the biggest and first step in the transfer of atmospheric substances towards the plant surface. It may be the limiting factor in the deposition process as long as there is high atmospheric stability. In the presence of high atmospheric turbulence, however, the last mm above the plant surface may govern the deposition velocity. The laminar boundary layer, which is a zone of air without turbulence and in direct contact with the plant (Figure 1), represents the major resistance for deposition of fine particles between 0.1 and 1 vrn diameter (Chamberlain and Little, 1981). The water soluble fraction of continental particles of this size consists mostly of ammonium and sulphate. The deposition of trace gases, especially SO 2 and NH 3 is dependent on stomatal opening, and also on the surface wetness (Fowler and Unsworth, 1979, Sutton et al., 1993). The present paper describes important processes during deposition of fine particles and gases, happening directly at the plant surface.
140
(horizontal) ................................... wind
velocity
u
9
.................................... ~ .... 9
9
4
...........................
:::::::l,- :"
. . . . . . . . . . . . . . . . . . . . . . .
stomatal
'
"
~
[ne edle ]
[ needle section'}
(-c)E. tt. Priori
Figure 1. Resistance model for the transport from the free atmosphere to the plant surface
2. The importance of R b and R c for the composition of dew drops Dew drops were collected from small conifers in the Fichtelgebirge, a mountaineous region in NE-Bavaria, over a period of three years (1987, 1990, 1991). The droplets were collected directly from the needles, and were then analysed for inorganic components (Burkhardt and Eiden, 1990). Pairs of samples were taken; one of Scots pine (Pinus sylvestris L.) and one of Norway spruce (Picea abies (L.) Karst.), standing close together. Dew droplets from 30 pairs of trees were examined during the course of the experiment, some of them were sampled repeatedly. Separate samples were taken from different aged needles (recent year, 1-year-old, etc.). Ratios were formed for each sample and each ion, dividing the ion concentration in dew sampled from Scots pine by that from Norway spruce. The geometric means of these ratios are shown in Figure2, with the error bars indicating the 95% confidence range. Only results from needles of the recent year are shown. Concentrations of all ions in dew from Scots pine are significantly enhanced. A
141
difference in the amount of dew between the different tree species could be widely excluded by a comparison with lysimeter balances (Burkhardt, 1994). The exchange of ions over the leaf surface is one possible explanation for the presence of ions in the dew droplets, and another one is atmospheric deposition. While leaching is probably an important source for some of the ions (e.g. Mn, K), this is not likely to be the case for a group of other ions, including Na, CI, H, SO 4, NH 4. The differences observed for this group therefore is concluded to be due to atmospheric deposition. As the sampled pairs of trees stood close together and were of the same height, the air concentration of the trace substances as well as the aerodynamic resistance were the same for both individuals of each pair; as well was their climatological history. The differences in the concentrations are therefore due to different values of R b and/or Rc.
~
.... iiii]iiiiii[iiiiiiiIiiill C
NO 3 SO4
H
NH 4
Ca
Na
Mg
K
Mn
Figure 2: Ratios of the ion content of dew samples taken from pairs of immediately neighboured trees, each pair formed by one Scots pine and one Norway spruce. The geometric mean (i.e. the arithmetic mean of the logarithmic values) of the ratios was formed. The error bars indicate the 95% confidence level The error bars are of different length on both sides due to the retransformation. For absolute concentration values see Burkhardt and Eiden (1990); Burkhardt
(1994).
142
3. A new concept of surface wetting
The influence of humidity without visible dew formation on trace gas deposition first gained scientific attention with the experiments of van Hove et al. (1989, 1990), although it had been recognised several times before (Spedding, 1969; Garland and Branson, 1977; Fowler and Unsworth, 1979). In the years that followed, it became clear that the gap which had been observed between modelled values and measurements of trace gas fluxes to canopies could be explained by gaseous deposition on the plant cuticle at high relative humidity (Draajiers and Erisman, 1993; Padro et al., 1993). Use of newly developed wetness sensors (Burkhardt and Gerchau, 1994) enabled an invisible surface wetness to be detected even on the hydrophobic surfaces of spruce needles (Burkhardt and Eiden, 1994). This was recognised by the very strong correlation between relative humidity and electrical conductance along the needle surface (Figure3)., which can only be explained by a liquid bridge between the two electrodes, i.e. extending over a distance of 5 mm.
- 9- , o0 c: (1)
26
r = 0.90 r=0.59 . . . . . . . . . . . . . . . . .
r '
=
0.73 . .
.
.
.
.
r= .
0.88 '
80
E ._~ 24
tive humidity i---.1
6o "0 ~
40
i-
E r--
"~
2o
"0
20
(D ~ "-~ 16
. 0
conductance control sensor . . . . . . . 12 0 12 0 12
0
~
12
Central European Time 25.6.92
26.6.
27.6.
28.6.
29.6.
Figure3: Electrical conductance of a sensor on spruce needles (14 m height), of a control sensor without contact to a surface (hanging free), and relative humidity. The correlation coefficients between conductance on the needles and relative humidity are indicated on top of the plot. A significant overall decrease of the signal occurred when the relative humidity fell due to a change of air masses. Nevertheless, the correlation was still high. The pointed line and the double arrow indicate the difference of conductance at the same relative humidity, once during the regimen of high humidity, once during a generally lower humidity.
143
Continuous measurements were carried out on needles in the crown of a 40-yearold spruce over a period of 5 months, using a meteorological tower. Apart from the conductance measurements, which took place in 6 heights, relative humidity and temperature were measured in 3 heights. Rainfall was detected by the use of a rain gauge on top of the tower. The results are shown in Table 1. The fraction of time is shown, during which correlation coefficients of 0.7 < r < 1 were found between relative humidity and conductance along the needle (Table 1). This is interpreted as wetting by thin water films. The time of visible wetting by rain or fog which amounted to about 30% of total time, is not included. The remaining 70% without visible wetness are the basis (i.e. 100%) for the numbers indicated.
Table 1: Times of correlation coefficients 0.7 < /r/ < 1, between electrical conductance on the spruce needles and the relative humidity. The basis is the time of visually dry needles, Le. about 70% of total time (for details of data analysis see Burkhardt and Eiden, 1994).
height above ground
time (%) with 0.7 < r < 1
18m 16m 14m 12m 10m 8m
(invisible wetness) 62.4 55.5 59.5 65.6 54.0 20.7
In order to investigate the importance of these water rims on gas deposition, a dew chamber experiment was carried out. By holding the relative humidity stable in the dew chamber, a constant electrical conductance evolved (for detailed description of operating conditions see Burkhardt and Eiden, 1994). After several hours, ammonia was introduced in the dew chamber, which lead to an instantaneous increase of conductance, indicating the dissolution of ammonia in the water film and the formation of ammonium ions (Figure 4). The experiment was repeated at different humidities and applying different ammonia concentrations. The lower values where a significant conductance increase could be detected were 65% relative humidity and 600 ~g m-3. The correlation between relative humidity and conductance disappeared, when needles were washed with deionised water and subsequently allowed to dry before the measurements were taken (Burkhardt and Eiden, 1994).
144
90 '-~ r-
26 80
E O}
r
70
relative humidity
24
= m
"13 = m
E
60 r
_e-
50 r "~
22
m
conductance spruce needles
0
40
L _
30 20
20 0
3
6
time [h]
9
12
-I-15
18
21
N H3 added
Figure 4: Measurement on needles of a spruce seedling, made in a dew chamber. The relative humidity was held constant, which lead to a constant conductance signal on the needle. The sudden conductance increase when ammonia was added to the dew chamber, is due to the dissolution of ammmonia and formation of ammonium ions. It proofs the possibility of gas dissolution in invisible water films.
This supports the hypothesis that particles on the plant surface are responsible for the formation of water films (Eiden et al., 1994). A new concept is thereby confirmed; explaining invisible surface wetness as a consequence of the presence of hygroscopic salts at the needle surface. Water vapour escaping from the stomata creates a humid environment, leading to dissolution of the salts on the leaf surface. As almost all water soluble substances of atmospheric aerosol will be dissolved at a relative humidity of 80% (Winkler, 1988; Pilinis et al., 1989), this is likely to happen to deposited aerosols over extended times. This concept therefore explains the considerable length of observed needle wetness in the field. In addition, the different behaviour on the lowest needles (Table 1) can be explained. This is most probably due to closed stomata since stomata close earlier in the lower part of the tree crown, where photosynthesis is not so effective, to prevent water loss (Larcher, 1984). Partial closure of stomata could as well be the reason for the 'step' in conductance in Figure 3.
145
4. The importance of salts on the leaf surface for gas deposition
This wetting concept has several implications for gas deposition since invisible wetness is likely to control gas deposition on the cuticle. Therefore, the conditions in this wetness are important. As particles form an important prerequisite for the formation of the films, the particle characteristics are of primary importance. Of particular importance are the amount, hygroscopicity, chemical composition, and also the deposition pattern of the particles on the needles. The history of the leaf is therefore important, i.e. for how long did it not rain, which particles were deposited etc. The chemical composition of thin water films is likely to be dominated by deposited particles. The chemical composition of these particles is especially important since SO 2 and NH 3 deposition are pH-dependent. Most continental aerosols contain large amounts of NH 4 and SO 4 which contribute almost 80% of the water soluble part (Ludwig and Klemm, 1990). Particles of about 0.1-1 ~m diameter presumably have pH values between 1 and 3 in the atmosphere (Winkler, 1986). On such deposited particles, therefore, NH 3 would dissolve very easily, whereas in the presence of NaCI salts, with a much higher pH, the deposition of SO 2 will increase and NH 3 deposition will be low. If both gases are in the atmosphere and the pH of the wetting solution is more or less neutral, codeposition of both gases in similar quantities is likely to happen. This will be especially important if SO 2 is oxidised on the leaf surface. In this case, NH4HSO 4 or (NH4)2SO 4 salt will be formed, which will not be volatilized again and will therefore stay on the leaf surface. As new water vapour condenses on the newly formed salt, there is the potential for dilution of more trace gas, producing a positive feedback mechanism. The extent of hygroscopicity of the salt is important because the more water that is absorbed, the more trace gases can be dissolved. Hygroscopicity is characterised by the deliquescence point, i.e. the relative humidity value at which water vapour saturation is reached at the salt surface, or the relative humidity at which the particle will dissolve. The deliquescence points of important atmospheric salts are NaCI 75%, (NH4)2SO 4 80%, and NH4HSO 4 40% (Pilinis et al., 1989). The closer the particles are deposited to transpiring stomata, the longer they will be dissolved. Jagels (1991) found preferential deposition of polydisperse particles at stomatal regions and the same observation was made for monodisperse fine aerosols of about 0.5 ~m and attributed to the presence of structured waxes (Burkhardt, 1994). Particles within this size range are very ineffectively deposited in general, and are therefore transported over long distances (Heintzenberg, 1989).
146
5. The possible Impact of salts on the leaf surface to plants Particles deposited in the stomatal regions of the leaf surface will be nearly permanently dissolved. Some work has been done in the past regarding the acidity on the leaves by evaporation of rain drops or by acidic fog. Most of the continental fine particles are much more acidic (pH 1-3) than these visible drops and so can be NH4HSO 4, which is newly formed on the leaf surface. The pH of a solution is defined operationally, for measurement with the glass electrode. Ludwig and Klemm (1990) proposed a water activity of aw=O.9 as the lower limit of pH definition. Due to aw - rh (e.g. Stumm and Morgan, 1981), this means a relative humidity of 90%. Other concepts of pH for appliance in highly saline environments have also been proposed (Knauss et al., 1990), and could be sensible in the case of particles on the surfaces of leaves. However, pH might not be the most important factor with respect to the damage these particles cause the plants. Moreover, the wick function of aerosols/salts may be more important, a deductive concept which becomes possible as soon as continuous water rims between substomatal cavities and the leaf surface are 'allowed'. Thin water films have been investigated in laboratory measurements by several authors and the water film thickness has been calculated to be far below 100 nm (van Hove, 1989; Benner et al., 1992; Burkhardt and Eiden, 1994; Burkhardt, 1994). It is impossible for water droplets to enter the stomata due to the surface tension and stomatal architecture (SchSnherr and Bukovac, 1972), but water films of a finite thickness notably below the dimensions of the stoma or the openings in the wax mesh could form a liquid connection along the wall with the liquid water film inside the substomatal cavity. Hygroscopic salts on the surface would therefore attract water from inside of the plant (Figure 5). Due to the hydraulic .
.
.
.
.
particle stoma wax mesh
/
_
needle
Figure 5: Model of the wick effect caused by hygroscopic aerosols.
147
junction rapid equilibrium formation will evolve. When the equilibrium is disturbed by the influence of Sun and wind, the evaporating water is replaced by liquid water out of the stomata. Thus, the transpiring area of the tree is increased. Stomatal movements would not affect this mechanism as long as the stomata would not close completely. This mechanism might be more important than the acidity. Needle loss as observed in forest decline is a drought phenomenon (Ulrich, 1990), and winter dessication would also be aggravated by this kind of additional transpiration. 6. Conclusions
Although particles contribute only in part to the overall picture of dry deposition, they may a have an important influence on trace gas deposition by the formation of surface wetness. They could also be of importance by forming an impact to the plants, both by providing very deep pH on the leaf surface for extended times as well as by their hygroscopic action, which leads to an increased transpiration area. This might be of importance during times of drought and times of frozen soil in the winter. Acknowledgements
The help of Dr. A. Gillies in the preparation of the text is gratefully acknowledged. This investigation was supported by the Bayreuth Institute of Terrestrial Ecosystem Research (BITOK) and the German Federal Minister of Science and Technology (BMFT, grant No. OEF 2029).
References
BURKHARDT J. (1994): D~inne Wasserfilme auf Fichtennadeln und ihr Einflul~ auf den Stoffaustausch zwischen Atmosph&re und Pflanze. Bayreuther Forum Okologie, 9, 1-137. ISSN 0944-4122. BURKHARDT J., EIDEN a. (1990): The ion concentration of dew condensed on Norway spruce (Picea abies (L.) Karst.) and Scots pine (Pinus sylvestris L.) needles. Trees, 4, 22-26. BURKHARDT J., EIDEN a. (1994): Thin water films on coniferous needles. Atmospheric Environment, 28A, 2001-2017. BURKHARDT J., GERCHAU J. (1994): A new device for the study of water vapour condensation and gaseous deposition to plant surfaces and particle samples. Atmospheric Environment, 28A, 2012-2017.
148
CHAMBERLAIN A.C., LITTLE P. (1981): Transport and capture of particles by vegetation. In: GRACE J., FORD E.D., JARVlS P.G. (Hrg.): Plants and their Atmospheric Environment, 147-173. DRAAIJERS G.P.J., ERISMANJ.W. (1993): Atmospheric sulphur deposition to forest stands: throughfall estimates compared to estimates from inference. Atmospheric Environment, 27A, 43-55. FOWLER g., UNSWORTH M.H. (1979): Turbulent transfer of sulphur dioxide to a wheat crop. Quarterly Journal of the Royal Meteorological Society, 105, 767-783. GARLAND J.A., BRANSON J.R. (1977): The deposition of sulphur dioxide to pine forest assessed by a radioactive tracer method. Tellus, 29, 445-454. HEINTZENBERGJ. (1989): Fine particles in the global troposphere - A review. Tellus, 41 B, 149-160. JAGELS R. (1991): Biophysical aspects of fog deposition on the needles of three conifers. Journal of Experimental Botany, 42, 757-763. KNAUSS K.G., WOLERY T.J., JACKSON K.J. (1990): A new approach to measuring pH in brines and other concentrated electrolytes. Geochimica et Cosmochimica Acta, 54, 1519-1523. LARCHER W. (1984)" W. Larcher 1984. Okologie der Pflanzen auf physiologischer Grundlage, 4th ed., Ulmer, Stuttgart. LUDWIG J., KLEMM O. (1990): Acidity of size-fractionated aerosol particles. Water, Air and Soil Pollution, 49, 35-50. PADRO J., NEUMANN H.H., DEN HARTOG G. (1993): Dry deposition velocity estimates of SO 2 from models and measurements over a deciduous forest in winter. Water, Air and Sol Pollution, {}8, 325-339. PILINIS C., SEINFELD J.H., GROSJEAN D. (1989): Water content of atmospheric aerosols. Atmospheric Environment, 23, 1601-1606. SCHONHERR J., BUKOVAC M.J. (1972): Penetration of stomata by liquids. Plant Physiology, 49, 813-819. SPEDDING D.J. (1969): Uptake of sulphur dioxide by barley leaves at low sulphur dioxide concentrations. Nature, 224, 1229-1231. SUTTON M.A., PITCAIRN C.E.R., FOWLER D. (1993): The exchange of ammonia between the atmosphere and plant communities. Advances in Ecological Research, 24, 302393. ULRICH B. (1990): Waldsterben: Forest decline in West Germany. Environmental Science and Technology, 24, 436-441. VAN HOVE L.W.A., ADEMA E.H., VREDENBERGW.J., PIETERS G.A. (1989): A study of the adsorption of NH 3 and SO 2 on leaf surfaces. Atmospheric Environment, 23, 1479-1486. VAN HOVE L.W.A., VREDENBERG W.J., ADEMA E.H. (1990): The effect of wind velocity, air temperature and humidity on NH 3 and SO 2 transfer into bean leaves (Phaseolus vulgaris L.). Atmospheric Environment, 24A, 1263-1270. WINKLER P. (1986): Relations between aerosol acidity and ion balance. In: JAESCHKE W. (Hrg.): Chemistry of muitiphase atmospheric systems, Springer, Berlin, 269-298. WINKLER P. (1988): The growth of atmospheric aerosol particles with relative humidity. Physica Scripta, 37, 223-230.
G.J.Heij and J . W Erisman (Editors). Acid Rain Research: Do we have enough answers? 0 1995 Elsevier Science BC! All rights reserved.
149
The atmospheric input of inorganic nitrogen and sulphur by dry deposition of aerosol particles to a spruce stand K. Peters" and G. Bruckner-Schattb "Bayreuth Institute for Terrestrial Ecosystem Research (BITOK), Department of Climatology, University of Bayreuth, D-95440 Bayreuth, Germany bChair of Plant Ecology I, University of Bayreuth, D-95440 Bayreuth, Germany
Abstract The dry deposition of particle-bound NH:, NO, and SO:- to a spruce stand was determined by the inferential method. For determining the deposition velocity the model DEPOSITE was used, based on an analysis of the probability of absorption of particles on plant surfaces due to sedimentation, impaction and molecular diffusion through the laminar boundary layers. The investigation stand was a 45 year old spruce forest with a mean tree height of 20 m and a LAI of 11.4. Airborne particle-bound concentrations of the main inorganic ions were measured as a function of the particle size by Berner cascade impactors mounted at the top of the canopy. From June to November 1992 a total of 23 samples were taken covering 23% of a whole year. In scaling up the input fluxes found for each sampling period to mean values for one year, the results were 2.6 kg NHZ-N ha-' yr-', 2.8 kg NO,-N ha-' yr-' and 3.6 kg SO:--S ha-' yr-' which corresponds to 21% and 10% of the total atmospheric N and S input at this site. 1. INTRODUCTION The input fluxes of particulate and gaseous nitrogen and sulphur compounds from the atmosphere t o vegetation canopies are of major importance in considering the balance of nutrients and pollutants within ecosystems. The dry deposition of these elements is considered to be in the same order of magnitude as the wet deposition [1,2]. Of the dry deposition, the input of the trace gases NH3, NO,, HN03 and SO2 is commonly larger than the flux of the particle-bound ions NH:, NO, and SO:- [3]. Despite this fact the contribution of the particulate deposition cannot be neglected. For determining the dry deposition of aerosol particles a large variety of methods have been proposed [4]. Among these, the sampling of deposited particles by surrogate surfaces [5-71 and the application of the throughfall method [8-101 are most commonly used. The latter suffers from the fact that the portion of ions in the throughfall water originating from leaching or from deposited trace gases is not well known. A shortcoming of surrogate surfaces arises from their aerodynamic properties, which are different from those of real plant leaves. Some studies have dealt with the application of micrometeorological methods, the most important of which are the profile method and the eddy-correlation method [11-
150 14]. The use of these procedures within forested areas is often limited due to the insufficient fetch, especially at mountainous sites. Recently, the inferential method has been established. It is an indirect method, but reduces the fetch requirements of the common micrometeorological methods. Here, a deposition velocity Vd is calculated from the micrometeorological conditions at the site and from parameters characterizing the plant morphology. According to the fundamental equation F(4)
=
(1)
the dry deposition flux F is the product of vd and the airborne particulate concentration c which has to be measured above the canopy. The independent variables dp and z are the particle aerodynamic diameter and a distinctive reference height above the canopy, respectively. If Vd is known from the model calculation, it is sufficient to measure c for determining the dry deposition. This approach is widely used for trace gases, where vd is usually expressed as the inverse of a resistance [15,16]. Studies dealing with Vd of aerosol particles normally do not calculate resistances, but are based on the combination of absorption probabilities [1721]. In this study the model DEPOSITE, a further development of these approaches, was used [22]. It is mainly based on the work of Bache [17-19] who dealt with the deposition of water droplets. In extending the applicability of this model to the whole size range of natural aerosol particles a description of the particle diffusion through laminar sublayers surrounding all plant surfaces was added. The model was used to calculate the dry deposition of the main particulate inorganic ions to a Norway spruce canopy at Wiilfersreuth in the mountainous region Fichtelgebirge in northeastern Bavaria, Germany.
2. M A T E R I A L S
AND
METHODS
2.1. T h e m o d e l D E P O S I T E The model DEPOSITE, which calculates the size specific deposition velocity of aerosol particles to coniferous forests, was explained in detail by Peters and Eiden [22]. Important featured will be repeated here. Neglecting advection and assuming that the mean vertical wind velocity is zero, the mechanisms governing the transport of particles through the canopy are turbulent diffusion, sedimentation and absorption at plant surfaces. In combining these processes, the mass balance within an elemental volume of the canopy is considered. The result is a differential equation describing the vertical gradient of the airborne particle concentration [17]-
d2c
+ f(
z~dc
+
= o
(2)
with
-
and
v
- K Cz) +
K(z)l dK(Z)dz - fl(z)p(z)sin~o(z)
(3)
151
-
-
(4)
In these equations vs is the sedimentation velocity, u is the horizontal wind velocity, and K is the turbulent diffusivity of aerosol particles. ~ois the angle of the particle trajectory to the horizontal and can be expressed by ~o = arctan(v~/u), p is the vegetation density defined as total surface area of the plant elements per volume of the canopy. /3 is the absorption coefficient which combines the absorption efficiencies due to diffusion, impaction and sedimentation. If both height specific functions f(z) and g(z) are replaced by representative constants fa and gR, Equation (2) can be solved analytically [18]. However, in order to make the solution more exact the canopy can be subdivided in several horizontal layers i with equal thickness Az. Now for each of these height intervals individual constants fru and gm can be determined. This procedure, proposed by Bache [19], was used in this study. The calculation of the deposition velocity can then be treated as an initial value problem starting at the lowest height interval. The solution proceedes through all layers to the top of the canopy, finally calculating the overall deposition velocity va. The absorption coefficient/3 is determined by a probabalistic analysis considering diffusion, impaction and sedimentation as well as the mutual interferences between these processes. The result is /3 -- 1 --(1
--
Pz)(1/(l+u2/v~))'/2[(1
--
P~,ai)(1 - P~,im)]0/0+'2"/~'2))~/2,
(5)
where P,,di, Pz,im and Pz are the absorption probability due to diffusion, impaction and sedimentation, respectively. While diffusion and impaction are controlled by the horizontal movement of the particles (denoted by the subscript z), sedimentation occurres during the movement in the z-direction. Other processes like interception, resuspension and bounce-off are less important for coniferous needle surfaces during moderate wind conditions and will be neglected. In this paper diffusion denotes the processes responsible for the transport of particles through the laminar boundary layer that surrounds all surfaces. The diffusion term used in the model is expressed as a function of the friction velocity u, in the vicinity of the plant surfaces. The approach is based on the equation from Friedlander and Johnstone [23] who described the diffusion flux with respect to the molecular diffusivity within the boundary layer and some turbulent portion also contributing to the transfer. Additionally, the microscale roughness of the plant surfaces was introduced since it reduces the thickness of the boundary layer [24]. Moreover, the effect of so-called bursts within boundary layers, first observed by Kline et al. [25], was considered. Downward sweeps of air compensating for these bursts transfer momentum to the particles, pushing them towards the surface. For describing the impaction of particles upon obstacles like needles an empirical approach is used which expresses the absorption probability as a function of the Stokes number. The equation was found to fit well to the experimental results of May and Clifford [26] and Belot and Gauthier [27]. The sedimentation is characterized by the terminal velocity vs of settling particles, when the gravitational force and the drag force are in equilibrium [28].
152 The aerodynamic properties of a canopy needed for the model are the horizontal wind velocity u, the friction velocity u. and the turbulent diffusivity K. Individual values of these parameters are needed for each height interval Az. While u(z) is measured at several heights at the site, the K-theory approach is used to calculate K(z) and u.(z). These expressions include the effect of the thermal stability of the air and are based on the proposals of Goudriaan [29] and Underwood [30]. It is assumed that within canopies the turbulent diffusivity of momentum KM is equal to K for particles. It is not possible to compare the results of the model calculation with deposition velocities really measured at forest sites. Published studies of Vd-Values above forests do not include descriptions of the canopy in the detail required for the model. Therefore, a one-layer version of the model was tested with the experimental data of Chamberlain [31] describing the particle deposition to artificial grass in a wind tunnel. The agreement was fairly good for particles in large and medium size rages. The modelled deposition of small particles was about 50% lower than the experimental data. In summarizing, the input parameters which have to be measured in the field are as follows: - the vertical profile of the horizontal wind velocity u, - the vertical profile of the temperature T, - the vertical profile of the vegetation density p, - the orientation and size of the needles, - the microscale roughness of the needle surface. 2.2. T h e m e a s u r i n g s i t e The model was used to calculate the dry deposition of aerosol particles to a spruce stand (Picea abies (L.) Karst.) near Wiilfersreuth in the subalpine mountainous region Fichtelgebirge, northeastern Bavaria, Germany. The stand, situated at 680 m a.s.l., is an almost rectangular plot 205 m long and 40 m wide. Its longer axis is directed from southwest to northeast. In 1992, when the measurements were made, the age of the trees was about 45 years and their mean height h was 20 m. This stand of younger trees was surrounded on all sides by older-grown spruce trees about 25 m high. The ground has an inclination of 5.9 ° to the northwest [22]. Thus, the site has some horizontal inhomogeneities and is therefore less suitable for micrometeorological investigations. Therefore, the values gained by the model calculation are only valid for the point where the measurements are made, and caution must be exercised in extrapolating them to the whole stand. 2.3. T h e p l a n t m o r p h o l o g y For determining the vegetation density p, the height of the upper and lower boundary of the crown, Zh and Zb, and the cross sectional area A~ of the stem at the height Zb were measured for 29 randomly selected trees. With the regression
A1 - 22.6 + 8500A~.
(6)
the leaf area A1 of each individual tree was determined [32]. For distributing this value over the vertical extend of a crown the chi-square-function
153
z (m) 20
15
<
10
r
S 0
f
-'''' 0
,,,, 1
J,,, 2
I,,, 3
p ( m -1) 4
Figure 1. Distribution of the vegetation density p with height z at the measuring site
A{(z) _ A1
a(za) b exp(czd) -- 0.1 f~z~ a(zd)b exp(cza)dz
(7)
was used [22], where A{ is the needle surface area per meter tree height, za = ( z h z)/(Zh -- Zb), a = 4.76 x 10 4, b = 6.21, and c = - 1 6 . 2 3 . After adding the individual distributions for all 29 trees, the overall function p(z) shown in Figure 1 was achieved. By this procedure a surface area index (SAI) of 28.4 was found, and the leaf area index ( L A I ) was determined to be 11.4 assuming a ratio of total to projected needle area of 2.5. The average needle diameter dl was set to 1.5 mm, and it was assumed t h a t the needles do not have any preferred direction in space. The microscale roughness of the needle surface k, was set to 5 # m which is the elevation of small ridges lying parallel to the needle axis. 2.4. M i c r o m e t e o r o l o g i c a l m e a s u r e m e n t s Using a tower within the forest stand the horizontal wind velocity was continuously measured by a n e m o m e t e r s at heights z of 4.6 m, 11.8 m, 16.6 m, 16.8 m, 18.5 m, 20.2 m, 21.3 m and 28.2 m above the ground. At the uppermost position the wind direction was additionally recorded. At 2.3 m and 20.4 m, i.e. below and above the canopy, the t e m p e r a t u r e and the relative humidity were measured.
154 Table 1 Characteristics of the Berner cascade impactors 5-stage impactor 74.5 1 min -1 Volumetric flow rate Cut diameters (#m) 0.05, 0.14, 0.42, 1.2, 3.5, 10.0 Mean diameters (gin)
2.5. C o n c e n t r a t i o n
0.084, 0.24, 0.71, 2.0, 5.9
l O-stage impactor 25.4 1 min -1 0.015, 0.03, 0.06, 0.125, 0.25, 0.5, 1.0, 2.0, 4.0, 8.0, 16.0 0.021, 0.042, 0.087, 0.18, 0.35, 0.71, 1.4, 2.8, 5.7, 11.3
measurements
Airborne particle-bound concentrations c of the main inorganic ions were measured as a function of the particle diameter dp by Berner cascade impactors mounted at the top of the canopy. Two different types of impactors were used alternately, one 5-stage impactor ranging from dp = 0.05 #m to 10 #m and one 10-stage impactor ranging from 0.015 # m to 16 #m. Some details about the impactors are listed in Table 1. The impaction plates were covered by polyethylene film, which after sampling was extracted with deionized water. The solution was analyzed for the ions NH +, K + , Na+ , SO42-, NO~, C1- by HPLC with conductivity detectors. For Ca 2+ and Mg 2+ ICP-AES analysis was used. By measuring the pH in the solution the concentration of dissociated protons was deduced.
3. R E S U L T S
AND
DISCUSSION
From June to November 1992 a total of 23 samples of aerosol particles were taken. The sampling times were on the average 1 day for the 5-stage impactor and 6 days for the 10-stage impactor in accordance with the different flow rate and size partitioning of both devices. These periods covered 23% of a whole year. 3.1. D e p o s i t i o n velocities For each sampling interval mean values of Vd(dp) w e r e calculated by the model based on 30-rain averages of the meteorological input parameters. To demonstrate the span of the Vd-Values that was occurring, the curves for those periods, when the lowest and highest values were found during the measurements, are depicted in Figure 2. The minimum event is an interval lasting 38 hours from June 29 to July 1, 1992. During this period there were fair weather conditions with almost cloudless skies throughout the days and the nights. The wind velocity at z = 20.2 m never exceeded 2.0 m s -1. The thermal stratification was very stable. On the other hand, the maximum event, lasting about 6 days from September 29 to October 5, was accompanied by varying thermal stability with long periods of unstable and neutral conditions. The sky was cloudy throughout the interval, and during 3 days the wind velocity was large, reaching values of 5.3 m s -1 at z = 20.2 m. The difference in va between these examples is more than an order of magnitude. In Figure 2 the curve representing the sum of the mean values and the standard deviation (thin line) is drawn in addition to the mean function of va vs dp (thick line).
155
a)
(am s -1) 100
/
/
o
/
100
/
=
.01
-vd (cm s-1)
b)
=-
v
.01
= I i lIHH
.001
I I IImi
I IIIIIIIJ
I i illlH
1 dp ( ~ m )
1000
.001
IIIIIIII
I IIIIIII
I IIIIIII
1 dp ( ~ m )
I IIIIIII
IIIIIIII
1000
Figure 2. Calculated deposition velocity Vd at the measuring site vs particle diameter dp. Thick line: mean values; thin line: mean values plus standard deviation, a) measuring period with the smallest Vd; b) measuring period with the largest yd. Table 2 Portions in % of equivalent concentrations of particle-bound inorganic ions averaged for all cascade impactor measurements NH + Na + K + Ca 2+ Mg 2+ H + Cations 84 6 2 3 1 4 100 SO~- NO~ C1Anions 73 22 5 100
This gives an impression of the variability of Vd which is also larger when the mean values are high. 3.2. P a r t i c l e - b o u n d ion c o n c e n t r a t i o n s The atmospheric content of particle-bound inorganic ions at the measuring site is characterized in Table 2 which lists average portions of equivalent concentrations of the individual substances. It can be concluded that NH +, SO~- and NO 3 are the most important ions. The contributions of sea salt and protons are almost negligible. In general the ion balances of individual samples were in equilibrium, so it does not seem that further important ionic species were overseen. In Table 3 the mean concentrations of the prevailing ions and some values characterizing their distribution within the particle size spectrum are listed. As a rule of thumb, the geometric mean diameter dg denotes the particle size where the major portion of the species is occurring. If the size distribution can approximately be described by a log-normal distribution, then dg is equivalent to the maximum of the distribution. The geometric standard deviation % gives an indication of the spread of the distribution. It can be seen that NO 3 is on average found in larger particles than SO42- and NH +.
156 Table 3 Average values of airborne concentrations c-, geometric mean diameters dg and geometric standard deviations o~g,weighted by the measuring time 5-stage samples 10-stage samples all samples (nmol m -a) 152 120 125 NH + d-g (Izm) 0.70 0.77 0.76 cr--g(#m) 2.14 2.26 2.24 74 51 55 SO~- (nmol m -a) d-g (#m) 0.92 0.90 0.90 o~g (~m) 2.82 2.56 2.60 (nmol m -a) 31 33 33 NO~ d-g (#m) 1.59 1.46 1.48 o~g (~m) 2.79 4.13 3.92
The values listed in Table 3 do not show any significant differences between the two cascade impactors. Though the 10-stage impactor covers a broader size range, the mass of sampled material is not larger.
3.3. D e p o s i t i o n fluxes Deposition fluxes F were calculated according to
F - E~=I VdiCi,
(8)
where n is the number of the impactor stages (5 or 10) and dpi is the particle diameter representing the impactor stage i. A problem arises from the fact that towards lager dv the deposition velocity increases steeply. Therefore, the particles sampled in the largest stage of the impactors often contribute the largest portion to the deposition flux, even if their fraction of the total particle concentration is low. Hence the choice of dpi for the uppermost impactor stage very sensitively influences the resulting F. In this study the dpi for every stage was set equal to the arithmetic mean of the upper and lower cutoff (see Table 1). It is assumed that the errors resulting from this choice will be compensated if a large number of measurements is averaged. Another possible procedure is to smooth the impactor data prior to calculation. As is demonstrated in Figure 3, measured size distributions dc/dlndp can sometimes be well fitted by log-normal distributions [33]. Then, upon integration according to F-
rl°°um ~-~ddp J0.001k~m Vg
(9)
the flux F is calculated. However, the majority of the measurements made during this study could not be properly fitted using a log-normal distribution. Therefore, the calculation according to Equation (8) was selected. In Figure 4 the input fluxes F of NH4+-N, NO~-N and SO42--S are shown for all sampling periods. The intensities of the fluxes cover a wide range. Maximum values occurred in periods where the ion concentrations in the larger particle size classes was high. As mentioned earlier, the results further reflect the wind velocity and the thermal
157
150
100
50
0
d, ( P m ) -01
1
100
Figure 3. Airborne particle-bound NH: concentration on June 23 1992 as measured by the 10-stage impactor. The values for the different impactor stages are shown together with the log-normal distribution fitting the measurements.
stability of the air within the canopy. Towards the autumn the fluxes were increased. Note that for the different ions the maxima and minima occurred during different sampling periods. Table 4 lists the mean and maximum fluxes when the results are scaled up to one year.
4. CONCLUSION The flux values obtained in this study are smaller than those found by a rough estimation made at the same site for the time period from Summer 1988 to Summer 1989 and larger than those found in a similar study conducted in the mountainous region Bavarian Forest [34]. The total (wet and dry) atmospheric input estimated by throughfall analysis at various sites in the Fichtelgebirge amounts to 36 kg ha-' yr-' and 25 kg ha-' yr-' of total sulphur and nitrogen, respectively [35]. This means that the aerosol particle deposition found in this study contributes about 10% to the S deposition and 21% to the N deposition. It could be suggested that these values are of minor importance as compared to the total input, but one should keep in mind that the role of deposited particles is not defined solely by their contribution to the total ion input to ecosystems. As could be shown by Burkhardt and Eiden [36], the occurrence of thin water fdms on plant surfaces is enhanced by the presence of deposited particles. In this way the aerosol deposition can also further increase the deposition of gaseous constituents since the gas molecules dissolve in these liquid films.
158
16:
kg
12-
ha -1 y r -1
SO42--8
1992 m
_
m m
m m •
0
1
•
|
i
I
16-
kg NO3--N ha -1 yr -1
12-
1992
t
I
m
m
_
I
I
16-
I
-7
I
kg NH4+-N ha -1 yr -l
12-
1992 m
m _
0
June
-I
July
I
Aug.
I
Sept.
I
Okt.
I
Nov.
Figure 4. Time courses of the input fluxes of major S and N ions to the spruce forest at Wiilfersreuth, Fichtelgebirge
Table 4 Maximum and mean values of yearly input fluxes of the main N and S ions to the spruce stand maximum flux F (kg ha -~ yr -~) mean flux F (kg ha -1 yr -1) NH+-N 14.9 2.6 NO~-N 9.4 2.7 SO24--S 15.8 3.5
159
5. ACKNOWLEDGEMENTS
This study was funded by the Bayreuth Institute for Terrestrial Ecosystem Research and thus by the Bundesminister fur Forschung und Technologie (grant number PT B E 0 51 - 0339476A). Furthermore, thanks are due to Dr. M. McLachlan for final reading of the manuscript.
6. REFERENCES 1 J.W. Erisman, F.A.A.M. de Leeuw and R.M. van Aalst, Atmos. Environ., 23 (1989) 1051. 2 S.E. Lindberg and G.M. Lovett, Atmos. Environ., 26A (1992) 1477. 3 S.E. Lindberg, M. Bredemeier, D.A. Schaefer and L. Qi, Atmos. Environ., 24A (1990) 2207. 4 K.W. Nicholson, Atmos. Environ., 22 (1988) 2653. 5 B.B. Hicks, D.R. Matt, R.T. McMillen, J.D. Womack, M.L. Wesely, R.L. Hart,
D.R. Cook, S.E. Lindberg, R.G. de Pena and D.W. Thomson, J. Geophys. Res., 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
94D (1989) 13003. A. Bytnerowicz, P.J. Dawson, C.L. Morrison and M.P. Poe, Atmos. Environ., 25A (1991) 2203. G.M. Lovett and S.E. Lindberg, Atmos. Environ., 26A (1992) 1469. B. Ulrich, in Eflects of Accumulation of Air Pollutants in Forest Ecosystems, B. Ulrich and J. Pankrath Eds., Reidel, Dordrecht, 1983, pp. 33 - 45. R. Mayer, Staub Reinh. Luft, 45 (1985) 267. G.M. Lovett and S.E. Lindberg, J. Appl. Ecol., 21 (1984) 1013. B. Duan, C.W. Fairall and D.W. Thomson, J. Appl. Meteorol., 27 (1988) 642. I.Y. Lee and M.L. Wesely, J. Appl. Meteorol., 28 (1989) 176. R. Lorenz and C.E. Murphy Jr., Boundary-Layer Meteorol., 46 (1989) 355. A.G. Allen, R.M. Harrison and K.W. Nicholson, Atmos. Environ., 25A (1991) 2671. B.B. Hicks, D.D. Bddocchi, T.P. Meyers, R.P. Hosker Jr. and D.R. Matt, Wat. Air Soil Poll., 36 (1987) 311. J.W. Erisman, A. van Pul and P. Wyers, in Air Pollution Research Report 47, J. Slanina, G. Angeletti and S. Beilke Eds., E. Guyot, Brussels, 1993, pp. 215 - 234. D.H. Bache, Atmos. Environ., 13 (1979) 1257. D.H. Bache, Atmos. Environ., 13 (1979) 1681. D.H. Bache, Atmos. Environ., 18 (1984) 2517. W.G.N. Slinn, Atmos. Environ., 16 (1982) 1785. B.L.B. Wiman and G.I. Agren, Atmos. Environ., 19 (1985) 335. K. Peters and R. Eiden, Atmos. Environ., 26A, (1992) 2555. S.K. Friedlander and H.F. Johnstone, Ind. Eng. Chem., 49 (1957) 1151. N.B. Wood, J. Aerosol Sci., 12 (1981) 275. S.J. Kline, W.C. Reynolds, F.A. Schraub and P.W. Runstadler, J. Fluid Mech., 30 (1967) 741. K.R. May and R. Clifford, Ann. occup. Hyg., 10 (1967) 83. Y. Belot and D. Gauthier, in Heat and Mass Transfer in the Biosphere Part 1, D.A. de Vries and N.H. Afgan Eds., Wiley, New York, 1975, pp. 583 - 591. H.R. Pruppacher and J.D. Klett, Microphysics of Clouds and Precipitation, Reidel, Dordrecht, 1978.
160 29 J. Goudriaan, Crop Micrometeorology: a Simulation Study, Pudoc, Wageningen, 1977. 30 B.Y. Underwood, Atmos. Environ., 21 (1987’) 1573. 31 A.C. Chamberlain, Proc. Roy. SOC.’A, 296 (i967) 45. 32 G. Bauer, University of Bayreuth, Chair of Plant Ecology I, personnel communication. 33 L. Gomes, G. Bergametti, F. Dulac and U. Ezat, J. Aerosol Sci., 21 (1990) 47. 34 K. Peters, J. Ludwig and K. RuoB, J. Aerosol Sci., 22 (1991) S569. 35 B. Manderscheid and E. Matzner, Spatial and Temporal Variability of Soil Solution Chemistry and Seepage Water Ion Fluxes in a Mature Norway Spruce (Picea abies (L.) Karst.) Stand, submitted to Biogeochemistry 1994. 36 J. Burkhardt and R. Eiden, Atmos. Environ., 28 (1994) 2001.
ATMOSPHERIC DEPOSITION
S E S S I O N IV GENERALIZATION; TOTAL ATMOSPHERIC DEPOSITION AND SOIL LOADS; MEASUREMENTS AND MODELS
This Page Intentionally Left Blank
G.J. Heij and J. W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
163
On the Determination of Total Deposition to Remote Areas B. B. Hicks National Oceanic and Atmospheric Administration, Air Resources Laboratory, 1315 East West Highway, Silver Spring, MD 20910, U.S.A.
Abstract Discussion of "critical loads" and "total deposition" has focused attention on the importance of such concepts in the regulatory and policy processes, yet actual quantification of either is exceedingly demanding. Here, alternative measurement and modeling approaches will be explored, and the effects of complicating factors like terrain complexity and vegetation inhomogeneity will be considered. Patchy surfaces tend to receive more deposition than is indicated by a simple area-weighting, as can be easily demonstrated for patches of trees on a grassy countryside. The magnitude of the difference is greatest for substances that are transferred with low surface resistance. In general, for applications that demand consideration of target areas that are currently unstudied, an exploratory measurement program is desirable, coupled with modeling activity that is then benchmarked against the observational data base. Without models, the measurement programs can only address those specific areas that are studied. Without measurements, the models risk failing to allow for site-specific peculiarities that may be dominant. Thus, in general an integrated approach seems optimal, though perhaps difficult to attain in many practical situations.
1. INTRODUCTION In contemporary society, the easy decisions about how to design and impose remedial emission controls have already been made. The remaining difficult decisions require more justification, and decision makers turn to models for this support. Modelers are faced with the need to compress complicated considerations into simple concepts, among which the ideas of "critical loads" and "total deposition" have achieved considerable favor. A "critical load" is a level of pollution input into an ecosystem below which specific damage will not occur, and above which damage is expected. "Total deposition" is the sum of all input derived from the atmosphere. Both concepts are intended to be aids to the decision-making process, but the question arises as to how much the complexity of the subject jeopardizes the conclusions that are reached using such simple concepts. Consider a lake that is potentially susceptible to acid deposition. The concentrations of pollution in the water itself are determined by the flow of pollutant into the lake and the flow out of it m the situation is dynamic, and the pollutants do not merely concentrate in the water
164 as more and more input occurs. But it is not only the deposition to the water surface that defines the "load" of importance, but also the deposition to the surrounding catchment area. There are clearly two limits on what might constitute the quantity of i m p o r t a n c e - the total deposition (wet plus dry plus that due to fog) to the water surface (Fw) on the one extreme, and the total (Fc) to the entire catchment area on the opposite extreme. We must necessarily ask what input property is most r e l e v a n t - Fw, Fc, or some intermediate value. The important measure of deposition is surely that which correlates most directly with the effects to be avoided or alleviated. From the perspective considered above, it is not either of Fc or Fw, but the net effect of these as influences the lake itself. Part of the potential insult delivered to the catchment as Fc will be retained by the catchment and not delivered to the lake. That part which is delivered may well be delayed for years. How the net influence on the lake at any specific time relates to any index of deposition from the atmosphere constitutes a question that is generally unanswered. To obtain the answers would require site-specific research that has usually proved too expensive for replication. What is now available is enough understanding to recognize that SOME deposition to a catchment affects the water bodies in it, but is not yet adequate to enable confident prediction of the retention rate without on-site studies for guidance. There is need to specify the area that must be considered, and to determine (preferably on the basis of external information) the weighting factor to apply to the deposition occurring at different places. All of the discussion above assumes that accurate atmospheric deposition rates can indeed be determined. In fact, the concept of "total deposition" carries different meanings in different communities. Atmospheric sciences are unified on meaning of "total deposition'" it is the sum of deposition from the atmosphere by wet and dry mechanisms, with a (usually small) additional component due to interception of fog droplets and cloud liquid water, all expressed as mass of pollutant per unit time and per unit horizontal area. In materials science, however, "total deposition" refers to the deposition (by all mechanisms) to a portion of an exposed structure; in this regard, it is not the net vertical flux from the atmosphere that is considered, but the rate of transfer to specific elements of the surface, such as the surroundings of a window, the arm of a statue, or the roof of an automobile. This concept extends to many modeling studies, such as in wind tunnels, where even deposition velocities are quantified as being appropriate for a specific leaf, for example, instead of for a layer of trees constituting the lower boundary of the atmosphere. In some modem usage, distinctions between receptor element-specific and atmospheric areal average terminology are made by the use of lower-case symbols for the former and upper case for the latter. Thus, for a given coniferous forested landscape the deposition velocity appropriate to compute the flux of material to a single pine needle may be vd(n), to the trunk of a pine tree Vd(t), and to the underlying forest floor Vd(S). At the same time, the flux from the atmosphere to a unit horizontal area of the same pine forest is computed using a separate quantity Vd. A major challenge is to relate V d to vd(n,t,s). Yet another meaning of "total deposition" arises in some ecological literature, where it refers to deposition arriving at a forest floor, as often measured collecting precipitation samples beneath any vegetation canopy. These samples are affected by (a) wet deposition to
165 the canopy, (b) dry deposition intercepted by the canopy during the period since the previous "wash-off', (c) release of materials following transfer from the root systems to the foliage, and (d) deposition not due to atmospheric transport aloft but to local redistribution from one part of the surface to another. In practice, long-term throughfall measurements also include the leaf-fall component for deciduous vegetation. Measurements of this kind are well suited to quantify deposition as it may affect soils and aquatic ecosystems. The measurements are less well suited to considerations of atmospheric transport or of effects on vegetation. It is clear, however, that the two divergent definitions of total deposition given above both relate to the flux of material to the roughness elements that are exposed to air, and that the third definition given now refers to the deposition to the ground itself, with these surface elements viewed as an interruption. There is no "fight" or "wrong" definition, but there is definite need to recognize that different scientific communities have different working understandings of what is meant by "total deposition." There is no danger of confusion while communication is within specific communities. However, suppose that there is an attempt to refine regulations and policies on the basis of limnology and watershed-scale ecology. Then the critical "total deposition" that is likely to be relevant is that which corresponds to the third case discussed above -- the deposition, by all mechanisms, that arrives at a forest floor. If there are no measurements made at a specific location, then estimates can be produced by models, based on the computation of local air chemistry from assumptions about atmospheric dilution, reactions, and dispersion of emissions perhaps from far upwind. But these models consider "total deposition" to be something different; they contain no local surface-source component, they cannot address well the (sub-grid-scale) peculiarities of particular watersheds, and they do not consider the biological processes that affect the transfer and redistribution of pollutants after they leave the atmosphere and before they enter the soil ecosystem.
2. WET AND DRY DEPOSITION m SOME BASIC DIFFERENCES The processes by which clouds remove pollutants from the atmosphere have been studied with considerable detail over the last forty years, with early attention mainly on the problems of particle removal, especially as they relate to the problem of radioactive fallout. The emphasis of studies over the last ten years has been more on gaseous pollutants. The studies have included consideration of the role of clouds as mechanisms for relocating pollutants throughout the troposphere, as well as on their part in the transformation and removal processes. Because of their recognized efficiency in removing pollutants from the atmosphere, convective storms have been a major focus for these studies. On an event basis, convective storms entrain large amounts of air, remove pollutants with comparative efficiency, transform these pollutants into other chemical species, and deposit the products in intense bursts over relatively small areas. In contrast to convective systems, large scale synoptic precipitation systems scavenge pollutants from elevated layers of the atmosphere, and deposit them far more slowly, over longer time periods, and across larger areas. Different parts of each continent experience different proportions of these two mechanisms. For North America, for example, it is
166 evident that the areas of the continent most affected by convective activity are in the southeastern U.S.A., and that the relative importance of severe storms decreases markedly to the north and the west. The part of the continent that receives the most acidic precipitation is an area where convective activity contributes substantially. Not only does the precipitation process vary with geography, but so does the chemistry that affects wet deposition. Temperature plays an important role in the chemistry of clouds, regardless of whether they rain. As an obvious example, the rate of solution of even soluble gases into solid ice particles is much slower than for liquid droplets. Moreover, the effects on further absorption of dissolved gases such as sulfur dioxide is such that ice crystals quickly become inefficient receptors for transfer from the gas phase. Several repercussions are of practical relevance. First, high altitude cold clouds are likely to be less efficient scavengers of gaseous air pollutants than low altitude warmer clouds (even though solubility of gases increases as temperature decreases). Second, severe thunderstorms that have considerable activity at high altitudes may not be as efficient at removing atmospheric chemicals as the less intense convective cells that are confined to the lower levels of the atmosphere. Third, the temperature effects will be different for different chemical species, thus additionally complicating the patterns of deposition observed at the ground. In general, the deposition "footprint" left by a raining cloud is exceedingly complicated, and depends on the chemical species that is being measured. The geographic and temporal variability of the wet deposition process is sufficiently great that there has been limited success in relating the observations made by any specific wet deposition collector to the chemical and meteorological characteristics of the surrounding atmosphere. Regardless of this difficulty, a major goal of precipitation chemistry research is to derive such a capability, for both event and long-term average situations. Most models represent ensemble average conditions, rather than any single precipitation event. In this regard, it is of special importance to note that deposition by precipitation is essentially an ergodic process in space. Thus, measurements made at any single location will be like those made at any neighboring location (but not, of course, over distances such that orographic effects and climatology become important), provided averages are constructed over a long enough time. The high spatial variability evident in event precipitation chemistry data is greatly reduced as longer time averages are constructed. In contrast to wet deposition, dry deposition is not a spatially ergodic process. In other words, there is no guarantee that values obtained at a particular measurement site will be representative of a neighboring area, even if exposed to much the same air concentrations. Dry deposition is a product of several processes working in combination. First, concentrations of the pollutant in question must be present in the air to which surfaces are exposed. Second, the surfaces in question must be such that the material in the air can be transferred to them. This second factor is what causes dry deposition to be so markedly different from wet, since the nature of the surface itself influences (and sometimes controls) the efficiency of transfer from the air to the ground. In the case of wet deposition, the control is not as related to the surface (excepting, of course, the obvious effects of climatology and topography), but rather to processes within the cloud.
167 On the average, the chemistry of wet deposition does not differ greatly over small spatial scales (i.e., over distances of tens of kilometers). Dry deposition varies from place to place more than wet deposition. The air chemistry affecting wet deposition is mainly that at cloud level, whereas the air chemistry affecting dry deposition is mainly that in air in contact with the surface. Since dry deposition is controlled by the detailed nature of the surface as well as by air concentrations, spatial detail will not be diminished by time averaging. In contrast, for wet deposition any specific event must be expected to have great spatial variability, but this detail will be lost as time averaging proceeds. Thus, the concept of routine monitoring of wet deposition is substantially different from that associated with routine monitoring of dry deposition. For wet deposition, a few locations can be used to represent regional values, which are then representative of the wet deposition of all areas near the stations where measurements are obtained. This is fundamentally not the case for dry deposition.
3. THE DRY TO WET RATIO
A direct repercussion of the lack of spatial ergodicity in dry deposition is that the concept of a "dry to wet ratio" is poorly founded. Even though it is possible to use data obtained at any particular location to define and quantify such a ratio, the value obtained may well be inappropriate even for a neighboring location. It should be noted, however, that spatial averaging serves to smooth much of the variability that characterizes local quantifications of the dry/wet deposition ratio, and hence the ratio approach may well have more validity as the target area increases in size. Data on the relationship between dry and wet deposition can be derived from several US sources. Here, data obtained in the Atmospheric Integrated Research Monitoring Network (AIRMoN) program of NOAA will be used. In AIRMoN, wet deposition is derived from daily measurements at a small array of stations, supporting weekly measurements elsewhere. Dry deposition is derived using the Dry Deposition Inferential Model (see Hicks et al., 1987) at AIRMoN-Dry sites, some of which make considerably more detailed measurements designed to benchmark the inferential methods. Tables of the resulting wet and dry deposition rates are presented elsewhere (e.g. Sisterson, 1990). Table 1 presents a summary of the results for sulfur species (wet deposited as sulfate and dry as the sum of deposits as sulfur dioxide and particulate sulfate) and for a subset of the nitrogen species (nitric acid vapor for the dry component, and nitrate for the wet; hence, ammonia and ammonium are not considered). Figure 1 shows how the ratio of dry to wet deposition varies with the amount of rainfall, across the array of sites of the AIRMoN-Dry network. For purposes of clarity, winter months are excluded from these plots of seasonal data. It is clear that a strong dependence on the amount of precipitation exists. This is anticipated on first principles - - clearly, dry deposition will dominate when there is little rainfall. The overall behavior appears identical for both chemical species; regression yields a slope such that the dry/wet ratio varies according to the inverse square root of the rainfall (exponent = -0.55 __+ 0.04 for sulfur; 0.54 ___ 0.05 for nitrogen).
168 Table 1. Seasonal ratios of dry to wet deposition of sulfur (S) and nitrate (N), for sites of the NOAA AIRMoN network. Contributions associated with ammonium are not included. Mean seasonal values are given. The means are computed geometrically, and the standard errors on these means (a(+)) are quoted as percentages. WM m Whiteface Mountain, NY; WP m West Point, NY; SC - - State College, PA; AR ~ Argonne, IL; BV m Bondville, IL; OR ~ Oak Ridge, TN; P A - Panola, GA; PG ~ Pawnee Grasslands, CO.
WM WP SC AR BV PA OR PG
Winter S N 0.66 0.36 0.61 0.33 0.95 0.73 2.45 1.11 1.49 3.03 1.19 1.46 1.00 0.63 1.08 4.56
Spring S N 0.33 0.54 0.68 1.05 0.37 0.66 1.38 1.00 0.80 1.41 1.28 1.44 1.39 1.14 0.61 2.33
Summer S N 0.19 0.41 0.37 0.42 0.50 0.53 1.28 0.56 0.67 0.62 0.65 0.64 0.89 0.41 0.29 1.32
Autumn S N 0.33 0.52 0.43 0.46 0.59 0.79 1.13 0.83 0.57 0.92 0.71 0.69 1.12 0.67 0.56 2.85
Means a(__+)
1.08 1.02 16% 37%
0.75 1.09 21% 16%
0.52 0.57 23% 13%
0.63 0.81 15% 20%
The information presented here is not new, and is presently being updated with considerably better analyses for a much longer observational period (ten years of data will soon be available, versus the three years used to construct the present tables and figures). However, even the present limited data set demonstrates that the common assumption that dry deposition equals wet is flawed in a consistent fashion that will inject biases. The ratio of wet to dry deposition is a function of location, of rainfall amount, and of the chemical species in question. Using wet deposition as a basis for estimating dry is likely to be misleading unless a site has been calibrated for this purpose. As yet there is no alternative to on-site measurements if accurate dry deposition data are desired.
4. DRY D E P O S I T I O N m EFFECTS OF SURFACE H E T E R O G E N E I T Y There is a further aspect that deserves attention m the consistent underestimation of dry deposition due to omission of edge effects in contemporary models. Consider an area A containing portions A1 of grass and Az of forest. The forested area is composed of a number of n patches, that therefore have a characteristic length scale Ls = (A2/n) ~ As a first approximation, we may consider the exchange of the area as a whole as the sum of exchanges for the two contributing components, so that F = (A1.F 1 + A2.F2)/(AI + A2)
(1)
169
m
1.6
m i _r
1.2K
m m
o~0.8 -
o J
m
[]
m
m
m
>... L_ "(3
m- - -
m
mR
~ H
H m
-. ,,,.,,P"~ . m 0.4
[]
-
_
~
m
[]
m
~
m
m
mm
1.6
m
m
[]
m m
m
1.2
m
m
[]
[]
>,,. "13
m
o .._1
m
i
i i
m
m
m
m
[]
0.4
i
m
m
mm
~
m
m
mi
i
m
mm
m
M
m
,
1
1.1
[]
m m
m
I
I
1
I
1
I
I
I
1.2
1.5
1.4
1.5
1.6
1.7
1.8
1.9
2
Log(pr ecip if at ion)
Figure 1. The relationship between the seasonal ratio of dry to wet deposition, for sulfur (upper) and nitrate-nitrogen (lower) stations of the NOAA AIRMoN array, in the eastern U.S.A.
170 In most numerical models, each of the individual fluxes F~ and F2 is approximated using a resistance model or a surface-specific deposition velocity approach. Land use characteristics are used to determine appropriate values of surface resistances, and of the factors that determine the aerodynamic resistance and/or deposition velocity. Consider, however, the effects of edges. Wind blowing across the grass penetrates into each stand of trees for a distance that scales according to the wind speed, the density of the vegetation, and the height of the trees (h). As this air passes through the canopy, there is interaction between it and the components of the foliage that involve a "filtration efficiency" E. The edge-effect flux can then be written as (SF = E.n.h.Ls.u.(C02- Ca2)/(A~ + A2)
(2)
To a first order, we can consider the exchange by the horizontal edge-effect and filtration mechanism to be additive to the conventional vertically diffusive flux, and hence we can write the total exchange rate for the heterogeneous area as F(new) = [(Al.kl + Az.k2 + E.n.h.Ls.u)/(Al + A2)].(C02- Ca2)
(3)
where k~ and k2 are the appropriate total conductivities (ki = 1/(Rai + Rsi), and where it has been assumed that concentrations in air and at the surface are constant across the area under consideration (a rather questionable assumption). Additional insight can be derived from studies of throughfall under forest canopies, downwind of edges (e.g. see Beier, 1991), which indicate that deposition at the edge can be doubled, but falls off exponentially to the "infinite surface" value with an exponential distance constant that scales with the height of the trees. (Presumably, there will eventually be need to include a factor associated with canopy density and scavenging efficiency here, as indicated in the derivations above, but so far the experimental science does not permit this degree of sophistication. For further guidance, see Slinn, 1982, for the case of particles.) At distance x from a forest edge the flux Fx might be simplified as Fx = F2[1 + C.exp(-x/(c.h))]
(4)
where the edge effect is generalized by introducing an "amplification factor" C. Integration of (4) and comparison with (3) leads to an areal flux amplification that depends on a quantity like l ~ n - - the higher the trees and the patchier the surface the greater the dry deposition.
5. AN EXAMPLE - - THE CHESAPEAKE BAY Several recent studies have emphasized that coastal ecosystems are vulnerable to effects of atmospheric pollution. The focus of much of the concern is nitrogen compounds, deposition of which adds to the input from rivers and groundwater and eventually leads to eutrophication. In general, the most vulnerable ecosystems are biologically productive shallow-water embayments, affected by and downwind of population centers. Most of the Atlantic coast of the U.S.A. is potentially vulnerable (see Fogel and Paerl, 1991, for example), potentially extending as far as Bermuda (see Owens et al., 1992; Michaels et al., 1993). Regulators are currently
171 faced with a difficult problem m how to include air pollution in the complicated equations that underpin regulatory and control decisions. Table 2 summarizes some contemporary assessments for the Chesapeake Bay, the large estuary serving the major population centers in the U.S. mid-Atlantic region. The values tabulated refer to assumptions about the retention of deposited nitrogen by the terrestrial surfaces through which the pollution must migrate on its way to the Bay. The land area is so large and the water body so shallow that a small error in the assumption about retention can have large consequences; as is seen, the assumptions vary considerably.
Table 2. Watershed retention values (in % of nitrogen loading) used in Bay loading studies to date (numbers in parenthesis indicate range tested for sensitivity studies).
Land Use
Tyler (1988)
Hinga et al., (1991)
Fisher and Oppenheimer (1991)
Forest Pasture Cropland Residential
95.2-100.0 93.7-99.96 76.0-99.97 62.0-95.3
80.0 80.0 60.0 25.0
80.0 (51.0-100.0) 70.0 (51.0-90.0) 70.0 35.0 (0.0-70.0)
(25.0-95.0) (25.0-95.0) (45.0-75.0) (10.0-50.0)
In addition, every assessment assumes that dry deposition of nitrogen equals wet. One assessment omits consideration of ammonium. All assessments assume that deposition to the water itself equals that to the surrounding land. All of these assumptions are areas of potentially large error. In fact, it is not until the last year that even one dry and wet deposition monitoring station has been set up to test whether model predictions for the Chesapeake Bay watershed resemble reality. A recent workshop (see STAC, 1994) concluded that the major scientific need was for on-site data to verify model predictions. Although much of the discussion about uncertainty given above relates to the matter of dry deposition, even the wet deposition component remains somewhat unclear. Although wet deposition is highly variable, the spatial distribution of wet deposition ought to have characteristics similar to those of precipitation. Experience has shown that if averaged over a sufficiently long time, the value obtained at a single location is a fairly good approximation of averages at surrounding locations. Average and representative values of wet deposition can be interpolated from isopleth maps derived from independent estimates of precipitation amounts and concentrations. Interpolation requires accounting for the consequences of correlations between the two estimates. Existing estimates appear to overlook the effects of this correlation, and so it is possible that current wet deposition data are slightly biased to overestimate deposition. Although the magnitude of the bias has not been determined, estimates of the bias range from 10% to 25 % depending on location.
172 Estimates of wet deposition to the Chesapeake Bay surface range from 3.45 to 4.2 Gg N O 3N yr -1 (Fisher and Oppenhiemer, 1991; Hinga et al., 1991; Tyler, 1988). These estimates omit consideration of the effects of the correlation mentioned above, and fail to take episodicity into account; in practice it may well eventuate that the key consideration is not the annual delivery of nitrogen to the catchment area but the number of severe inundations that cause atmospheric nitrogen to be wet-delivered in large quantities and which cause nutrients to be scavenged from the landscape as water floods over it.
6. CONCLUSIONS There is accumulating evidence to demonstrate that reliance on a cardinal value for a "dry to wet deposition ratio" will be misleading. In some places, this will cause loadings to be overestimated; in others they will be underestimated. The patchiness of the surface has an influence that can be described in broad terms, but cannot yet be quantified in detail - - the patchier the surface, the greater the underestimation of dry deposition. If the goal is to protect a large area in which endangered ecosystems reside, then it is quite likely that contemporary knowledge will provide adequate guidance. However, if a particularly sensitive ecosystem is singled out, then particular attention should be paid to its specific circumstances. The spatial variability of dry deposition means that it cannot be isoplethed and interpolated like wet deposition, and hence interpolating loading estimates to a specific target area is potentially misleading. The likely biases that will result will not be unidirectional, so that it must be expected that control measures based on use of mapping and isopleths will not be wholly successful in protecting all sensitive regions. The way in which the critical loads concept is applied in a regulatory environment risks asking atmospheric transport models to predict deposition of the kind measured by throughfall studies. This may well be an excellent goal for research of the future, but so far there has been no atmospheric transport and deposition model whose output has been tested in this context. In fact, few models have been tested for their predictions of deposition from the atmosphere. Most tests have focussed on the ability to simulate air chemistry, and some have addressed wet deposition. However, there has been no direct test of predictions of dry deposition. The nearest such test has been in North America, where dry deposition rates computed from model predictions of air chemistry have been shown to agree with estimates of dry deposition derived from field observations of air concentrations of the same chemicals. The sub-models used to derive the dry deposition rates in the two applications are essentially the same. In essence, this is little more than a test of air chemistry predictions. Finally, an obvious question remains about what better can be done. There is no substitute for good on-site data. If the driving goals are related to the fate of pollutants emitted into the atmosphere, then large-scale models and statistically-derived arrays of deposition stations may provide good answers. However, if the intent is to help protect sensitive areas then monitoring should be at those areas, first emphasizing the locations that are best suited for high quality deposition measurements. This conclusion is emphasized by the demonstrations presented here that the smoothed outputs of contemporary models cannot
173 be accurately interpolated to predict the deposition to small (i.e. greatly sub-grid-cell) ecosystems where there are no data. The practical importance of these conclusions will vary from one chemical species to another, in a manner that is not yet understood.
REFERENCES Beier, C. 1991. Separation of gaseous and particulate dry deposition of sulfur at a forest edge in Denmark. J. Environ. Quality, 20:460-466. Fisher, D. C., and M. Oppenheimer. 1991. Atmospheric nitrogen deposition and the Chesapeake Bay estuary. Ambio 23:102-208. Fogel, M. L., and H. M. Paerl. 1991. Nitrogen isotope tracers of atmospheric deposition in coastal waters off North Carolina. Carnegie Institute Geophysical Laboratory. pp. 147-154. Hicks, B. B., D. D. Baldocchi, T. P. Meyers, D. R. Matt, and R. P. Hosker, Jr. 1987. A preliminary multiple resistance routine for deriving dry deposition velocities from measured quantities. Water, Air, and Soil Pollut., 36:311-330. Hinga, K. R., A. A. Keller, and C. A. Oviatt. 1991. Atmospheric deposition and nitrogen inputs to coastal waters. Ambio 20:256-260. Michaels, A. F., D. A. Siegel, R. Johnson, A. H. Knap, and J. N. Galloway. 1993. Episodic inputs of atmospheric nitrogen to the Sargasso Sea: Contributions to new production and phytoplankton blooms. Global Biogeochem. Cycles 7:339-35 I. Owens, N. J. P., J. N. Galloway, and R. A. Duce. 1992. Episodic atmospheric nitrogen input to oligotrophic oceans. Nature 357:397-399. Sisterson, D. L. (ed). 1990. Deposition monitoring: Methods and results. In: Irving, P.M., ed. Acidic Deposition: State of Science and Technology. Vol. 1 Emissions, atmospheric processes, and deposition. Washington D.C.: National Acid Deposition Assessment Program, pp. 320. Slinn, W. G.N. 1982. Predictions for particle deposition to vegetative canopies. Atmos. Environ., 16:1785-1794. STAC. 1994. Atmospheric Loadings to Coastal Areas: Resolving Existing Uncertainties. Workshop Report of the Air Quality Coordination Group and the Scientific and Technical Advisory Committees of the Chesapeake Bay Program, Annapolis, MD. Tyler, M. 1988. Contributions of atmospheric nitrate deposition to nitrate loading in the Chesapeake Bay. Columbia, MD; VERSAR Inc., Maryland Dept. of Natural Resources Report RP 1052.
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
175
Quantifying the scale dependence in estimates of wet and dry deposition and the implications for critical load exceedances R. I. Smith ~ , D. Fowler a and K. R. Bull b alnstitute of Terrestrial Ecology, Edinburgh Research Station, Bush Estate, Penicuik, Midlothian EH26 0QB, United Kingdom bInstitute of Terrestrial Ecology, Monks Wood, Abbots Ripton, Huntingdon PE17 2LS, United Kingdom
Abstract Two sources of uncertainty in the deposition estimates used to calculate critical load exceedances are investigated. An analysis of the uncertainties in the deposition models for the UK suggests predictions of total sulphur deposition on a 20 km scale are within +_40% to +_80% depending on the region, with greater uncertainty over the higher rainfall regions, but catchment studies indicate that predictions within +_30% can be achieved. Critical loads for soils are calculated at a 1 km scale and a simulation study shows that deposition estimates calculated either at a 20 km (UK) or a 150 km (European) scale both underestimate critical load exceedances in complex terrain using the current mapping and modelling procedures. Models for deposition at a 1 km scale are not available for complex terrain but an alternative approach is proposed which gives a probability distribution of critical load exceedances within an area.
INTRODUCTION The procedures developed recently within Europe to regulate emissions of acidifying gases are based on an assessment of the deposition of the various chemical species and their effects on vegetation, soil and freshwaters. In the first instance, emissions of sulphur dioxide (SOz) will be reduced within the Sulphur Protocol agreed within the United Nations Economic Commission for Europe (UNECE, 1994). In developing such protocols the distributions of sources and sensitive receptors throughout Europe are of major importance so that emission reductions can be targeted for maximum benefit. A critical loads approach has been adopted to identify the geographical distribution of sensitive receptors and to assess the magnitude of current effects. In conjunction with long-range transport models, the critical loads approach can also be used to assess the effect of different strategies to control the sources of acidifying gases. The critical load exceedance, or the amount by which the estimated deposition exceeds the
176 estimated critical load for an area, is calculated assuming no uncertainty in either estimate. The aim of this paper is to explore uncertainties in the method associated with the current deposition estimates. Of particular interest is the problem of scale, since critical loads are often derived at different spatial scales from those used in deposition modelling, and these approaches may lead to systematic bias in estimating the magnitude of exceedance and its spatial distribution.
CRITICAL LOADS A critical load is defined by UNECE as 'a quantitative estimate of exposure to one or more pollutants below which significant harmful effects on sensitive elements of the environment do not occur according to present knowledge' (CLAG, 1994). The aim of emission reduction policy is to minimise the area of Europe which has a critical load exceedance. Many of the areas with large critical load exceedances are remote from sources and are often remote from any gas or rainfall monitoring stations, so concentration and deposition fields must be modelled. Maps for critical load exceedances for soils in Britain have been produced with a resolution of 1 km x 1 km. These show the squares where the estimated deposition input, derived from 20 km x 20 km square values (assumed to be constant for all the 400 1 km x 1 km squares within the 20 km x 20 km square), exceeds the calculated critical load value for the 1 km x 1 km square (CLAG, 1994). On a European scale, deposition values are taken from the output of the EMEP models which operate at a 150 km x 150 km scale (Sandnes, 1993). For policy purposes, similar maps are also produced with deposition estimated from various emission reduction scenarios using long-range transport models. The critical load considered in this paper is for acidification of soils with no modification for land use or base cation deposition.
DEPOSITION MODELS The deposition models considered in this paper are those used to provide the deposition of non-marine sulphur to 20 km squares in the UK. Total deposition is calculated as the sum of deposition through 3 deposition pathways: wet, cloud droplet and dry. The importance of the 3 deposition pathways varies considerably across the UK, depending on region and landscape type, and the accuracy of the predicted total deposition will depend both on the accuracy of the inputs and on the particular mix of pathways at a specified location. Wet deposition is modelled as a product of rainfall and 5042-concentration in rain. When the landscape has the rapid altitude variations which are typical of many areas of the UK, both the rainfall amount and rain ion concentrations, enhanced by the presence of polluted orographic cloud at higher elevations, must be adjusted for altitude effects. These effects are included in the model in a simplified form developed for average topography rather than for detailed topographic and meteorological inputs (Dore et al, 1992). Both cloud droplet deposition and dry deposition are modelled using the standard resistance
177 analogy 'big-leaf model (Hicks et al, 1987). Cloud droplet deposition is modelled to high elevation land uses, assumed to be either moorland or forest, at deposition rates close to those for momentum (Fowler et al, 1993). The surface resistances are set to zero and only an aerodynamic resistance, ra, is used to calculate the cloud droplet deposition velocity. For dry deposition the 3 resistances, aerodynamic resistance, r~, a quasi-laminar boundary layer resistance, rb, and a canopy resistance, rc are all required for calculating the gas deposition velocity (Cape et al, 1991). Deposition is a product of deposition velocity and concentration. Assuming all the above models are correct in their formulation, the effect of the accuracy of the input variables on the predicted total deposition at the 20 km scale will now be explored.
Model inputs The wet deposition model is dependent on three inputs. The rainfall amount and rain ion concentrations are taken from interpolated maps derived from a network of monitoring sites (RGAR, 1990). There are approximately 4000 rainfall collectors and there were almost 60 sites recording concentrations in rain of the major ions, although this has now been reduced to about 30 sites. Estimates of rainfall and rain ion concentration for each 20 km square are derived from a kriging interpolation (Webster et al, 1991). The orographic enhancement factor was calculated from 30 year average rainfall records and is assumed to be constant over time. For cloud droplet deposition the aerodynamic resistance, r~, is derived from a land use data base (Bunce et al, 1983) which is used to determine the proportion of each square which is expected to be covered by forest. The roughness length is fixed for the two upland land use categories, forest and moorland, and the wind speeds are 30 year averages (Thompson et al, 1982) for 40 km x 40 km squares interpolated to 20 km values and modified to increase mean wind speeds over high ground. The time for which vegetation is covered by cloud is estimated from long-term cloud cover observations over the UK (Weston, 1992). The cloud ion concentration is derived from the rain ion concentration field. The dry deposition model requires greater detail in land use categorisation and more meteorological information. The land use data base (Bunce et al, 1983) is used to determine the proportions of each square expected to be covered by arable crops, grassland, forest, moorland and urban areas. Climatological data, in the form of 30 year averages for 40 km x 40 km squares interpolated to 20 km values, of wind speed (modified for altitude), temperature, sunshine hours and rainfall (Thompson et al, 1982) are input. The latitude and longitude of the square are used to generate hourly values of incoming solar radiation for one day in each month for clear sky, overcast sky and wet conditions. A stomatal resistance, r s, is generated for each hour of the day from typical vegetation characteristics. Lower optimum temperatures, 25~ are used than appear elsewhere in the literature because there is no explicit constraint on stomatal opening when a plant is water stressed. Soil and leaf surface resistance are combined and set to constant values depending on expected surface wetness, derived from the rainfall and cloud frequency data and estimates of hours of dew. The total canopy resistance with wet surfaces is set to 20 s m ~, a value which reflects average conditions in the UK. The combined canopy resistance, ~, is used with vegetation and wind speed dependent resistances, r~ and rb, to calculate daily deposition velocities for clear sky, overcast sky and wet conditions. The annual sulphur
178
~
deposition is calculated from the deposition velocities using the estimated occurrence of each weather type in each square and the interpolated SO2 concentration field derived from data from about 30 monitoring sites.
Model sensitivity The wet deposition model is linear in both inputs, rainfall and rain ion concentration. The ratio of cloud ion to rain ion concentrations used in determining both the orographic enhancement factor for wet deposition and the cloud ion concentrations for cloud droplet deposition was set to 2. Recent experimental work (Fowler et al, in press) indicates this value is conservative for the high rainfall areas of the UK where ratios of 4 to 8 are common. However, the observational data are biased, since they typically sample the base of the cloud where ion concentrations are highest, and a value of 2 gives an estimate over a wide variety of upland landscapes consistent with more complex modelling approaches (Choularton et al, 1988). The cloud droplet deposition model is linear both in cloud ion concentration and in the time vegetation is covered by cloud and is non-linear in wind speed. In this case, a higher ratio of cloud ion to rain ion concentrations than currently used would be justifiable. The proportion of forest within the square has a substantial effect on modelled cloud droplet deposition at higher altitudes. The dry deposition model is relatively insensitive to several of its inputs, as long as the values are varied within reasonable limits for the UK. Plausible modifications of temperature would alter dry deposition by less than 5%. Wind speed has a non-linear effect on dry deposition and a 40% increase in wind speed would increase dry deposition to forests by 6% and to other land uses by 15%. Doubling the time with wet leaf surfaces would increase deposition to forests by 30% and to other land uses by 10% to 15%. The model is linear with respect to gas concentration. Misspecification of land use has its greatest effect through the presence or absence of forests in remote areas when it can substantially affect deposition. The values for wind speed adjustment, orographic enhancement of rainfall and cloud frequency are averages over large areas and would not necessarily be appropriate for finer scale modelling.
Possible uncertainty in the estimates of total sulphur deposition The deposition models are linear with respect to the input concentrations of SO42 o r S O 2 and errors in these interpolated concentration fields are transmitted directly into the predicted deposition fields. The Review Group on Acid Rain presented data for 1986 to 1988 including error maps for both precipitation-weighted mean concentration of non-marine sulphate and for mean annual wet deposited sulphate (calculated as a product of rainfall and concentration without inclusion of any altitude effect) (RGAR, 1990). Incorporating an orographic enhancement, which was not used for the original maps, the uncertainty, measured as twice the Kriging error, was approximately 2.5 kg S ha ~ y-~ over most of England, 4 kg S ha -~ y~ over south-west England, Wales, and most
179 of Scotland increasing to 10 kg S ha ~ y~ over the high rainfall areas in the west of Scotland. Assuming an error in the rainfall estimate to a 20 km square of • these values indicate a combined uncertainty (i.e. an approximate 95% confidence interval assuming a Normal distribution of residuals) in wet deposition estimates of about • rising to _+80% in high altitude, high rainfall areas. Kriging interpolation has been used to derive the SOz concentration field, but there are a relatively small number of sites for its successful implementation. There is evidence that the assumptions underlying the derivation of a Kriging error map are not satisfied so the spatial error analysis must be considered an approximation. The error map gave values of 1-2 ppbV over most of the UK, ranging from about 25% of the annual mean concentration in central and eastern England to over 100% of the annual mean concentration in the north-west of Scotland (Vincent, K.J. and Campbell, G.W., pers. com.). Assuming an uncertainty in deposition velocity of +_20%, the overall uncertainty in dry deposition estimates would range from • in central England to well over • in many areas of Scotland. The major pathway for deposition varies in different areas of the UK with 70% as dry deposition in central England to 80% as wet deposition in north-west Scotland. Ignoring any uncertainty in cloud droplet deposition, the estimates of total sulphur input to a 20 km square could have an uncertainty of • in central England increasing to • on the west of Scotland and Wales.
C A T C H M E N T STUDIES A number of studies have required more detailed measurements and modelling of specific areas and extra information has become available. These provide valuable support for the interpretation of deposition estimates. In studies of the Plynlimon and Llyn Briane catchments in west Wales deposition inputs were modelled using detailed land use and altitude data and local concentration measurements (Reynolds et al, 1993). Inputs of sulphur to the catchment modelled from the detailed measurements in 1987 and 1988 averaged 26 kg S ha 1 yl while measured averaged output in the stream water for the same period was 27 kg S ha -1 y~. The inputs to the area estimated from the 20 km squares used in the UK map were in the range 20 to 30 kg S ha -~ y~ giving reasonable agreement between the national and the more detailed local model. A similar exercise based on a study in east England at Beacon Hill, an area dominated by dry deposition, between 1985 and 1988 gave an estimated total sulphur input of 37 kg S ha 1 y~ compared against an estimated outflow in the streams of 37 kg S ha ~ y~ (Vitkovic and Black, 1994). If the dry year in 1984 is included, the estimated outflow was 31 kg S ha 1 y~ and the input increased to 40 kg S ha ~ y~. The estimated input from the 20 km national map squares were approximately 30 kg S ha ~ yl, somewhat lower than the more detailed model estimates. In a study in the north of Scotland a specific requirement was to model the current input of sulphur to an area of moorland and to estimate the effect of afforestation. The area lay within
180 2 UK 20 km squares with estimated total sulphur deposition of 5.9 kg S ha t yl. Detailed modelling of 4 areas gave values of 5.7, 6.1, 6.7 and 6.9 kg S ha ~ y~. These 3 studies indicate that the 20 km deposition estimates were within _+30% of more detailed model estimates, with little evidence of gross systematic error in the national maps.
S I M U L A T I N G D E P O S I T I O N DATA AT D I F F E R E N T SPATIAL SCALES Deposition estimates are produced for Europe at the 150 km scale and 50 km scale by EMEP. For the UK, deposition is estimated at a scale of 20 km while in the Netherlands deposition estimates will in time be provided at the 5 km scale. Assuming that the deposition models give reliable estimates at their own scale, do they provide good estimates for critical load exceedances when the critical loads are calculated at a 1 km scale? For this exercise, a method was devised to produce deposition estimates to 1 km squares using data at the 20 km scale. For comparison purposes, deposition was also estimated using exactly the same procedures but implemented at the 20 km scale and the 100 km scale. Each 20 km square was divided into the number of 1 km squares within 7 altitude bands which were assumed to be at 31, 92, 184, 458, 519, 762 and a notional 1066 m above sea level.
m
Figure 1 Total deposition modelled at the 100 km scale.
of
Sulphur
Figure 2 Total deposition of Sulphur modelled at the 20 km scale.
181 The rainfall to the l km squares were estimated from a linear function assuming 600 mm rain at 50 m and 2000 mm rain at 600 m, and then adjusted to match the 20 km square rainfall amounts. The orographic enhancement of wet deposition was estimated by assuming all rainfall amounts in excess of 600 mm had twice the rain ion concentration. Gas and rain ion concentrations were assumed constant over the 20 km square but wind speed was adjusted to the altitude band values. Deposition was calculated for each altitude band and land uses were assigned to the 1 km squares assuming moorland at the highest altitudes, then forest, grass, arable and urban in descending order. This method would not give accurate estimates to any specific 1 km square but does give a distribution of possible deposition values. Deposition maps
The map of total sulphur deposition at the 100 km scale provides an impression of relatively smooth variation (Figure 1). The 20 km map, although not identical to the 'standard' modelled map, gave very similar structure over the UK (Figure 2). At the 1 km scale, the information was in distributional form and could not be accurately located spatially. Therefore, a pair of maps were produced: one giving the value for the whole 20 km square as the minimum deposition from the 400 1 km squares (Figure 3) and the other giving it the maximum deposition from the 400 1 km squares (Figure 4). This indicated the range of possible deposition values at the 1 km scale. The minimum 1 km deposition gave a structure
It :
~ll,~.."
~ ,~
] k g S h a - l y -t In above 30
Im
~
/
Im U
I
20- 3o
12-15 9 - IZ
below 9
: |
L
'
If
...
J
Figure 3 Minimum Sulphur deposition to a 1 km square within the 20 km square.
Figure 4 Maximum Sulphur deposition to a 1 km square within the 20 km square.
182 quite similar to the 20 km map but without showing the areas associated with high wet deposition in west Wales and west Scotland. These areas have rapid changes in altitude and the low level ground did not receive high total deposition according to these models. The maximum 1 km deposition gave values above 20 kg S ha ~ y-~ for large areas of the UK indicating that variations in altitude and land use could give high depositions to local areas practically everywhere.
Distribution of 1 km deposition estimates The categorisation of altitude and land use only allows typically 7 different deposition values for the 400 1 km squares within the 20 km square. This gives little information on the distribution of values. For 5 areas of the UK, ten 20 km squares covering an area of 40 km E-W x 100 km N-S were grouped together to compare distributions in different landscapes.
40-
40"
r
t..
r
~
30.
r r O
30-
r r O
20-
~
10-
0
30-
!0-
tO
~0
30
40
50
Deposition (kg $ ha-t) Figure 5 North-West Scotland 9Distribution of 1 km deposition in a 40 km x 100 km area, dotted line indicates average 20 km deposition.
0
I0
20
30
40
50
Deposition (kg $ ha-t) F i g u r e 6 North-East Scotland : Distribution of 1 km deposition in a 40 km x 100 km area, dotted line indicates average 20 km deposition.
In a high rainfall area in the north-west of Scotland with an average 20 km deposition of 12 kg S ha 1 y~ dominated by wet deposition (over 80% of total deposition), the 1 km deposition values ranged from 2 to 36 kg S ha1 y-~ (Figure 5). In a drier area in the north-east of Scotland wet deposition was 60% of total deposition, the average 20 km deposition was 9 kg S ha ~ y-1 and the range of 1 km deposition values was 2 to 26 kg S ha~ y-~ (Figure 6). 7% and 9% of the two areas respectively received over twice the 20 km average deposition. In more polluted areas further south, in a high rainfall area of Wales with wet deposition at 50% of total deposition the average 20 km value was 21 kg S ha t yl and the range was 8 to 52 kg S ha -~ y-1 (Figure 7). To the east in a drier area in east England where wet deposition is 30% of total deposition the average 20 km value was 25 kg S ha t y~ with a range of 16 to 48 kg S ha -1 y-1 (Figure 8). In Wales about 2% of the area exceeded twice and 14% of the area
183
50-
40-
40-
q.) eo t., t.,
eo Q.) t.., t., eo eo o
30.
eo
eo o
k',e
30.
0 o
~'~
....
II.d2b ,.,.5
lb
i1
am
~
30-
30.
,o.
5b
m . . . . . . . . . .
o
Deposition (kg S ha-t) Figure 7 West Wales 9Distribution of 1 km deposition in a 40 km x 100 km area, dotted line indicates average 20 km deposition.
. . . . . . . .
zb
i
Jill, ~b
3bim,
ii I
~
Mi..,.
5b, ,,
Deposition (kg S ha-s) Figure 8 East England 9Distribution of 1 km deposition in a 40 km x 100 km area, dotted line indicates average 20 km deposition.
exceeded 1.5 times the average 20 km deposition. For east England 4% exceeded 1.5 times the average 20 km deposition.
50-
40q~
ro
In East Anglia, an area more similar to the Netherlands in landscape, the average 20 km deposition was 14 kg S h~f 1 yl (with 50% from wet deposition) with a much shorter range from 10 to 22 kg S ha-~ y~ (Figure 9).
~D
aoo &"e
30.
In all cases the distribution of 1 km values IIIII. was positively skewed. Only in East Anglia, * 10 ~ 30 40 50 with an altitude and land use pattern similar to Deposition (kg S ha-~) the Netherlands, was the range of 1 km Figure 9 East Anglia" Distribution of 1 km estimates relatively short. In the UK context, deposition in a 40 km x 100 km area, dotted where sensitive ecosystems are often on line indicates average 20 km deposition. higher, poorer land, these simulations indicate that in most areas where critical loads are currently exceeded, the 20 km deposition value will be conservative. For substantial areas of land the 1 km deposition would be 1.5 to 2 times greater than the 20 km deposition.
C R I T I C A L L O A D E X C E E D A N C E S AT D I F F E R E N T S P A T I A L SCALES The critical load exceedances were calculated for the different deposition maps. The results (Table 1) indicate that the difference between the total areas exceeded at the 100 km scale and the 20 km scale, using the same model, was small. Use of the minimum 1 km estimate for the
184 whole square reduced the number of exceeded squares by 15% but the maximum 1 km estimates increased the number of exceeded squares by 75%. These figures give some bounds to the likely impact of deposition scale problems on critical load exceedances. Table 1 Critical load exceedances for different spatial scales of deposition Exceedance Number of 1 km x 1 km squares exceeded (keq ha -1 y~) (in thousands) deposition scale 100 km 20 km Min 1 km Max 1 km not exceeded 0.0 - 0.2 0.2 - 0.5 0.5 - 1.0 > 1.0
120 30 35 29 12
124 25 30 29 17
140 24 24 18 20
50 14 25 130
total exceeded
106
101
85
176
6
CONCLUSIONS The analysis of the accuracy of the current UK national 20 km deposition models, although not a full error analysis of the system, indicated that the uncertainty in predicted total sulphur deposition was about • in central England rising to • on the west of Scotland and Wales. The catchment studies with more detailed models and measurements gave improved deposition values which agreed well with measured stream flow and chemistry in the area. The national estimates were within • of the detailed model values, considerably better than the uncertainty analysis suggested. There was an indication of underestimation by using the national estimates but no definite evidence of bias. The simulation study of 1 km deposition values from 20 km data, with extra information on altitude and altitude dependencies, showed that the distribution of deposition to 1 km squares within an area is positively skewed for typical landscapes in the UK. Even for a fiat landscape in East Anglia the skew distribution was apparent. Therefore the 20 km deposition estimates will be biased and will underestimate 1 km deposition in substantial areas. The importance of the bias for critical load exceedance calculations depends on which 1 km squares have low critical load values. In the UK the high deposition, high altitude regions are also often the areas with very sensitive receptors. Therefore the smallest critical loads tend to have depositions towards the upper tail of the distribution of 1 km deposition values. The simulations indicate that 1 km values of 1.5 to 2 times the 20 km values would be appropriate for such sensitive ecosystems. The application of minimum and maximum 1 km estimates to the whole 20 km square indicates bounds to the problems that deposition scale poses for critical load exceedances. The area predicted to exceed its critical load estimated from the 20 km deposition estimate is considerably nearer the minimum 1 km value than the maximum 1 km value. Transferring from
185 a 100 km estimate to a 20 km estimate of deposition makes little difference in the total areas with critical load exceedances. Although using the maximum 1 km deposition value is unrealistic, these results indicate a potential substantial underestimation of areas of critical load exceedance, both at the UK national 20 km scale and at the European 150 km scale (EMEP). The current deposition models are not directly applicable at a scale of 1 km. In the absence of accurate fine scale modelling, the alternative approach is to provide statistical distributional information for deposition to an area. This information would not only incorporate the scale dependency problem but also include the effects of the uncertainties in all the inputs to the models. The critical loads themselves are estimates which have a quantifiable uncertainty. The critical load exceedance for an area would then be provided as a probability distribution of exceedances formed by combining the distributions for deposition and for critical load. This approach would have the further advantage of a direct indication of the uncertainty in the estimates of critical load exceedances. The large scale deposition estimates (150 km) used throughout Europe and the nationally derived estimates at other scales (20 km or 5 km) lead to significant underestimates of the critical load exceedance at the 1 km resolution. The systematic bias results from fine scale variability in deposition not being represented in the coarse scale modelling. In the absence of deposition models for complex terrain at the 1 km scale, a statistical approach is suggested which will provide critical load exceedances in distributional form and give a clearer indication of the uncertainty in the estimates.
ACKNOWLEDGEMENTS The authors gratefully acknowledge the UK Department of the Environment for funding this study.
REFERENCES Bunce, R.G.H., Barr, C.J and Whittaker, H.A. (1983) A stratification system for ecological sampling. In: Ecological mapping from ground, air and space, edited by R.M. Fuller: 39-46. Cambridge: Institute of Terrestrial Ecology. Cape, J.N., Smith, R.I. and Fowler, D. (1991). Modelling dry deposition of SO2 in Britain. In: Computer modelling in the environmental sciences, edited by D.G. Farmer and M.J. Mycroft: 285-298. Oxford: Oxford University Press. Choularton, T.W., Gay, M.J., Jones, A., Fowler, D., Cape, J.N.C. and Leith, I.D. (1988). The influence of altitude on wet deposition. Comparisons between field measurements at Great Dun Fell and predictions of a seeder-feeder model. Atmospheric Environment 22:1363-1371. CLAG (1994). Critical loads of acidity in the United Kingdom. Summary report of the Critical Loads Advisory Group. UK Department of the Environment.
186 Dore, A.J., Choularton, T.W. and Fowler, D. (1992). An improved wet deposition map of the United Kingdom incorporating the seeder-feeder effect over mountainous terrain. Atmospheric Environment 26A: 1375-1381. Fowler, D., Gallagher, M.W. and Lovett, G.M. (1993). Fog, cloudwater and wet deposition. In: Models and Methods for the Quantification of Atmospheric Input to Ecosystems: 51-73. Nordiske Seminar- og Arbejdsrapporter 1993: 573. Copenhagen: Nordic Council of Ministers. Fowler, D., Leith, I.D., Smith, R.I., Choularton, T.W., Inglis, D. and Campbell, G.W. (in press). Atmospheric input of acidity, sulphur and nitrogen in the UK. In: Proceedings of the Critical Loads Workshop, September 1993. London: University College London. Hicks, B.B., Baldocchi, D.D., Meyers, T.P., Hosker, R.D. and Matt, D.R. (1987). A preliminary multiple resistance routine for deriving dry deposition velocities from measured quantities. Water, Air and Soil Pollution 36:311-330. Reynolds, B., Fowler, D. and Smith, R.I. (1993). Modelling atmospheric inputs to catchments. In: European Network of Catchments Organised for Research on Ecosystems (ENCORE). First interim report. April 1993: 59-68. Commission of the European Communities. RGAR (1990). Acidic deposition in the United Kingdom: The Third Report of the Review Group on Acid Rain. UK Department of the Environment. Sandnes, H. (1993). Calculated budgets for airborne acidifying components in Europe, 1985, 1987, 1989, 1990, 1991 and 1992. Det Norske Meteorologiske Institutt. EMEP/MSC-W Report 1/93. Thompson, N., Barrie, I.A. and Ayles, M. (1982) The Meteorological Office rainfall and evaporation calculation system: MORECS (July 1981). Bracknell: The Meteorological Office. UNECE (1994) The UNECE protocol for the 1979 convention for long range trans-boundary air pollution on the further reduction of Sulphur emissions and decision on the structure and function of the implementation committee, as well as procedures for review of compliance. Document number ECE/AB.AI~40. United Nations Economic Commission for Europe. Vitkovic, G. and Black, V.J. (1994). Beacon Hill catchment study. The relationship between the chemical inputs from precipitation and the freshwater chemistry of a small catchment in the East Midlands, UK: dry deposition, modelling and critical loads. Loughborough: Department of Geography, Loughborough University. Webster, R., Campbell, G.W. and Irwin, J. (1991). Spatial analysis and mapping the annual mean concentrations of acidity and major ions in precipitation over the United Kingdom in 1986. Environmental Monitoring and Assessment 16:1-17 Weston, K.J. (1992). Objectivity analysed cloud immersion frequencies for the United Kingdom. Meteorological Magazine 121:108-111.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
187
Uncertainties Associated with the Inferential Modelling of Trace G a s Dry Deposition: A Comparison of Four Models with Observations from Four Surface Types Jeffrey R. Brook and Jacob Padro Atmospheric Environment Ontario, Canada, M3H 5T4
Service,
4905
Dufferin
Street,
Downsview,
Abstract For operational monitoring, it is convenient to use inferential models to estimate dry a n d total deposition. Such models have only been tested for a limited number of surface and atmospheric conditions. To assess the uncertainties associated with the inferential approach 4 different model formulations have been compared with ozone and sulphur dioxide dry deposition velocities derived from flux measurements. These data were collected during 4 separate field studies, each of which involved a different surface type and atmospheric conditions. Averaged over the entire study, the means of the modelled 03 deposition velocities were less than the observed mean by from 16% to 52%, depending upon surface type. There was considerable variation among models (+50%). For SO2 deposition there was greater variability among models and the mean modelled deposition velocity was within -20% of the observed mean. Over shorter time periods there were greater discrepancies between the observations and the model predictions. For example, a majority of the modelled values of mean daily O3 were within 25% to 55% of the observations depending upon model and surface type. However, for each 30 minute measurement period the differences were larger, often exceeding a factor of 2. A diurnal pattern, with larger 03 deposition velocities during daylight hours, was observed over all surfaces, but was most pronounced over a deciduous forest in the summer. The models captured the mean diurnal patterns in the summer reasonably well, but nighttime 03 deposition velocities were generally under-predicted There were larger differences between models and observations over the deciduous forest in early spring (no leaves) and over the cotton crop.
1. I N T R O D U C T I O N The importance of monitoring the total deposition of acidifying chemical species, along with other relevant species such as base cations, is widely
188 recognized. All mechanisms of pollutant deposition, wet, dry or occult, can have an effect on the natural and human environment. It is necessary to develop models for estimating dry and total deposition due to the relative complexity of the techniques for making direct measurements of dry deposition compared to wet deposition. The instrumentation is costly and requires considerable supervision. Hicks et al. (1980) recognized the feasibility of calculating dry deposition fluxes as the product of a modeled deposition velocity and a measured air concentration. Subsequently, much attention has been paid to the development of models for estimating dry deposition velocities. The inferential approach to modelling dry deposition can produce estimates that agree to some extent with available measurements. However, the models have only been tested for a limited number of surface and atmospheric conditions and their uncertainties are not well characterized. The method being developed in Canada to determine dry deposition is based upon the inferential approach. Observations of meteorology, surface conditions and air concentrations are being collected to run site-specific inferential models. In addition, large-scale meteorological data are being used to produce inferential estimates of dry deposition velocity across a larger domain. Four different model formulations have been compared with ozone and sulphur dioxide dry deposition velocities derived from flux measurements. A big leaf model (Big Leaf) (Hicks et al., 1987), a hybrid big leaf/multi-layer model (Multi-Layer) (Baldocchi et al., 1987; Meyers and Baldocchi, 1988), the modified ADOM dry deposition module (Padro et al., 1991) and the RADM module (Wesely, 1989; Byun, 1990) have been run for spring (no leaves) and summer conditions in a deciduous forest and summer conditions over a vineyard and a cotton crop. The results of these model runs are described in this paper. The difference in dry deposition velocity between models and observations provides an indication of the amount of uncertainty associated with inferential modelling of trace gas deposition. The input data and the resistance parameters were examined to identify some of the sources of these differences.
2. DATA AND M E T H O D S Although there are significant differences among the present models, they are all based upon the multiple resistance analogue. Three main resistance terms are considered in the calculation of deposition velocity (Vd). These are the aerodynamic resistance (Ra), the laminar sub-layer resistance (Rb) and the canopy or surface resistance (Re). The largest differences between models are associated with Rr and to a lesser degree Ra. Table I lists the input information required to run the models. The Big Leaf and Multi-Layer
189 models rely on wind speed (u), the standard deviation of wind direction (Ge) and solar radiation to determine Ra (Hicks et al., 1987). RADM and ADOM calculations of Ra are more complex, utilizing the bulk Richardson number and wind speed to determine u* and L. However, ADOM uses the Louis parameterization (Louis, 1979), while RADM uses Byun's method ( Byun, 1990) to compute these parameters. Differences in the Rc calculations are too great to discuss here. One of the main differences is that RADM and ADOM use land-use categories while Big Leaf and Multi-Layer consider specific information on the plant type at the location of interest. This refinement requires more input information, such as plant-specific light response constants (rb,). For a complete description of the models please refer to the literature references given above. Table 1
The main input information required to run the inferential models.
Big Leaf
Atmospheric. Surface
Multi-Layer u, T, Ge, hv, u, T, Ge, hv, wetness, pre. wetness, pre species(%), species (%), LAI, % leaf, profile type, zo, Rstom(min), LAI, % leaf, rb8 Z0, Rstom(min), rb~
ADOM u, AT, pre., wetness season, land-use (8), zo, LAI, Cstom, lat., long.
RADM u, AT, hv, RH, pre. season, land-use (11), z0, Rstom(min)
Information on the observed Vd data that were used in the model comparison is given in Table II. Mean 30 minute (min) 03 fluxes were determined using eddy correlation over 4 different surfaces. The number of 30 min periods with valid observation ranged from 565 over cotton during CODE (California Ozone Deposition Experiment) to 1196 over a summertime maple forest (EMEFS-I, Eulerian Model Evaluation Field Study). There was a limited amount of SO2 flux data which were only collected during EMEFSII also using eddy correlation. These data are not as certain as the 03 data and negative SO2 Vds were excluded from the analysis. However, the magnitude of the Vds and the general diurnal pattern are believed to be true. Table 2
Description of the Field Studies providing Vd Measurements
Study EMEFS I EMEFS II
Surface maple-full leaf maple-no leaves
Time P e r i o d July 15-Aug. 30, 1988 Mar. 17-Apr. 27, 1990
CODE CODE
vineyard cotton crop
July l 1-Aug. 6, 1991 July 15-Aug. 6, 1991
No. 1196 996 247 1183 565
Poll. 03 03 SO 2 O3 O3
190 Model comparisons were only based upon time periods when all models were able to compute Vd values. This was not possible for the entire time periods indicated in Table II because model input data were not available for every 30 min interval. However, for a majority of the time the model values were determined, except for RADM. During the CODE studies no input data were available and for all studies RADM was only capable of determining hourly Vd values.
3. R E S U L T S Observed and modelled deposition velocities were compared for a variety of time scales ranging from 30 min to multiple weeks. In Figures l a and lb the individual 30 min observations are compared to the Multi-Layer and ADOM estimates. Overall, for both models there were more underpredictions of Vd and there tended to be an upper limit to their Vd predictions. There were also a substantial number of observations where the discrepancy between modeled and observed Vd was greater t h a n a factor of 2. For the Multi-Layer and ADOM models, 36% and 35% of the predictions were off by more t h a n a factor of 2. This percentage increased to 44 and 47 for the Big Leaf and RADM models, respectively. As shown on Figures l a and l b the model biases differed between field studies. For example, the Multi-Layer model tended to underestimate the 03 Vd values over the vineyard and cotton crop and overestimate Vd over the leafless maple forest. In contrast, ADOM underestimated Vd over the leafless forest.
3.0 2.5 ~
2.5
2.0
-
~2.0
~ 1.5
~1.5
0.5
0.5
0.0
0.0 0.0
0.5
1.0
1.5
2.0
Observed Vd
2.5
3.0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Observed Vd
Figure la, b Comparison of observed and modelled Vd ,(30 min) for (a) Multi-Layer and (b) ADOM (cm s'l).
191
In the future, Vd values from site-specific inferential models are expected to be used with Canadian Air Precipitation Monitoring Network (CAPMoN) m e a s u r e m e n t s of SO2, SO42" and HNO3 to compute dry deposition fluxes. These air concentration data are collected daily for integrated 24 hour periods. The observed and modelled Vd's were, therefore, averaged for each individual day during the field studies to examine how the models would perform for 24 hour periods. Mean percent RMS errors between modelled and observed 24 hour Vd values, which are listed in Table 3, varied between models and field studies. Errors were smallest over the vineyard (30%) and largest for SO2, for which the errors ranged from 39% for the Multi-Layer model to 130% from ADOM. The day to day variations in m e a n 03 Vd for EMEF I are shown in Figure 2a. A similar plot corresponding to SO2 is included in Figure 2b. Only days with more t h a n 17 of the possible 48, 30 min periods were included in these figures and in Table 3. The models tended to u n d e r e s t i m a t e mean daily 03 Vd (15 of 27 days for E M E F I). There were only 16 days with sufficient SO2 results. The model predictions were above and below the observations with ADOM and the Big Leaf model deviating most from the observed values. Table 3
Mean Percent RMS Errors in Daily Mean Deposition Velocity
............................... Multi-Layer .... Big Leaf ........... ~ 0 M ..... EMEFS I 25% 32% 28% EMEFS II 53% 55% 50% EMEFS II-SO2 39% 85% 130% Vineyard 29% 22% 23% Cotton 50% 40% 37%
a 1.4
(a)
Obs. - - a - . M u l t i - L a y e r - - x -
Big L e a f
=
~ M 41% 37% 81% -
ADOM ~ R A D M
[]
1.2 -~ 1.0 o 0.8
~"
0.6
~09 0.4 op,,q
~. 0.2 0.0
|
209 211 213 215 217 219 221 223 225 227 229 231 233 235 237 239 241 243
Figure 2a
Mean daily Vd for 03 during EMEFS I (cm s'l).
192
~
4.0 3.5
~
3.0
~
2.5
~
2.0
~
1.5
0
"~ 1.0
_{b.) . . . . . . . .
'_ _ _ _' . . . . . . . . . .
-
. . . . . . . .
iiiiiiiiiiiiii ii i iiiiii
o 0.5
0.0
!
-
!
!
!
|
,
,
i
!
i
-
!
;
!
~
!
i
84 86 88 90 92 94 96 98 100 102 104 106 108 110 112 114 116 Julian Day
Figure 2b
Mean daily Vd for SO2 during E M E F S I (cm s-l).
Mean diurnal variations in Vd for each case listed in Table I are shown in Figures 3a-e. According to the observations there tended to be a daytime peak in Vd for all surface types and for 03 and SO2. When there were leaves on the vegetation the peaks tended to occur at midday, when incoming solar radiation was strongest. This behaviour supports the belief t h a t stomatal control of pollutant uptake, which effects Rc, is very i m p o r t a n t for 03 and SO2. The models produced similar diurnal patterns, with the exception of the Big Leaf model for E M E F S II for both O3 and SO2 and RADM for SO2. The Big Leaf model exhibited no diurnal variation for the leafless forest conditions. This was because the leaf area index (LAI) was very small and thus, Rc was set to large constant value corresponding to the soil resistance. There was also very little diurnal variation in the SO2 Rc for RADM over the leafless forest (Wesely, 1989). Given t h a t there were no leaves on the trees, it is somewhat surprising t h a t a diurnal p a t t e r n in 03 Vd was observed and t h a t 3 of the models also predicted such a pattern. RADM increased Vd during the day because Wesely (1989) assumed a Rc dependence on solar radiation even though there are no leaves. In contrast, the ADOM diurnal p a t t e r n was produced because of an assumed t e m p e r a t u r e effect on Rc. Thus, these models both resulted in reasonably good predictions of the m e a n diurnal pattern, but for different reasons. The a m o u n t of agreement between the m e a n hourly observations and the model results varied with surface type, model and time of day. Over cotton (Fig. 3d) all three models (RADM was not used) u n d e r e s t i m a t e d Vd. The largest errors were during the day, when the observations tended to be twice as large as the model results. ADOM and Big Leaf both predicted a sharp decline in Vd in the early afternoon. Big Leaf exhibited this behaviour because of its simplified approach to partition the incoming solar radiation into shaded and sunlit leaves. Apparently, for overhead sun conditions this approach leads to an unrealistic reduction in photosynthetically active
193
radiation to the leaves, particularly at the more southern latitudes. This pattern was also predicted for the s u m m e r t i m e maple forest, but due to the more northern latitude of the E M E F S II site the drop was not as pronounced as during CODE. Possible reasons for the ADOM decline in Vd over cotton were discussed by Padro et al. (1994). During E M E F S I, there was a general tendency for the models to underestimate Vd at night. In particular, in the early morning hours the observations were 2 to 3 times larger than the model values. The low bias at nighttime was also evident in the Multi-Layer and Big Leaf results for both CODE studies. The nighttime biases were not as consistent for A D O M (i.e., ADOM overestimated for the vineyard). { a
Obs.--~-- Multi-Layer--x- Big i~af cm/s
_- ADOM --~--RADM
{ O Obs. - ' ~ ' " Multi-Layer - ~ 1.5 ends (b)
1.5 [ (a)
Big Leaf ~ A D o M
~.
RADM
I
1
_ _
._.,~.,o .
.
.
.
.
. . . .
~'",
.
o.~ . . . . . . .
-
.
.
.
1.0
.
/_,~_ _ . : _ ._---~_-~.o_ ~,_ _
'l~_aO~,,,~ ~
L~/
r"
. . . . . . . . . . . . . . . . . . . . . . . . .
0.5
~ a ~o,-,o ~ 0.0 I ~ X"~~m~V~l~'I ~ t ~
~~
~
'
' ~)~N'~ ~ ) ~ ' ~ ~ ~ I 1
@4C'4,.'Mr Hour
~.~
{'~>
].0 t
.
" .
.
.
.
.
.
.
.
.
.
.
.
~. ~. . , . , . ~ , , N 9
. . . . .
,,
Hour
4.0
,,-,-,
cm]8 (e)
,~ ,,
, ~
~
ADOM I
~ ~176176176176176 ~176
|
o
a
o~
.....
, ~ ,,, , ~ . ~ , ,
--~-- Multi-Layer - ~ - Big Leaf ......
/
. . . . . . . . . . . . . . . . . . . . . . . . .
0.5t 0.0
Hour
i o Obs. 1.5 cm/s
-
o
~
-
-
:,-_'~.~
.....
0.0
Hour
.
.
.
.
3.0 2.0 1.0 Lm. 9 ~ 0.0 Hour
Mean diurnal patterns in Vd (cm s'l). (a) E M E F S I. (b) E M E F S Figure 3 II. (c) V i n e y a r d . (d) C o t t o n (e) E M E F S II SOe Model performance for SO2 was generally not as good as for 03. According to the observations there was an increase in Vd during the day (Fig. 3e), but
194 no obvious midday peak. The Big Leaf model and RADM did not match this pattern. Reasons for this behaviour were discussed above. The Multi-Layer model performed surprising well with no apparent bias during periods without sunlight and only a slight low bias during the day. ADOM predicted some structure in the hourly variations of Vd and there was a tendency for a peak in the early morning hours. This m a y have been a result of surface wetness. There are a n u m b e r of other interesting features in Figures 3a-4e, which are currently being examined in more detail. For example, the Big Leaf model predicts a large peak in 03 Vd in the late afternoon and early evening during E M E F S I. This behaviour has not been fully explained, but some of it was due to the variation in Re. Figure 5 is a plot of the mean diurnal p a t t e r n in Rc+Rb. The Big Leaf model resistances were smallest at approximately 1830 LST and remained below the ADOM, Multi-Layer and RADM values until 2000 LST or later. I
Multi-Layer
o
!
- , * - - -Big Leaf
- N-ADOM
--
RADM I
s/cm ,
"~,/" .
~1
-'~~ 9
_~. . . . . . . . . . . . . . . . . . . . . . .
................ ........ .
0
Figure 4
.
1
.
.
2
~.- : -k- - ' ~
.
.
3
.
.
4
.
5
.............
oo~:9 g
/
ol- i - ~ ~
~j~.~,
.
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour
Diurnal pattern in modelled Rc+Rb (s cm -1) for E M E F S I.
Mean Vds were determined for the entire durations of the field studies. These values are listed in Table 4. For 03, the n u m b e r of 30 min values t h a t were included in the average ranged from about 500 to 1000, depending upon the study. Fewer observations were available for SO2. With the exception of the Multi-Layer model during EMEFS II and ADOM over grape, the overall tendency was for the models to underestimate Vd. For 03, the average amount of bias (difference between the observed Vd and the m e a n of the 3 or 4 models) was -17% for the leafless forest, -22% for the forest with leaves, -52% for cotton and -16% for grape. For SO2 the average bias was -18%. As seen in previous figures the models differed from one another substantially. Compared to the mean Vd across all 3 or 4 models, the range of m e a n Vds for the individual models were: +19% t o - 2 9 % for EMEFS I; +59% t o - 5 2 % for EMEFS II; +14% to -16% for cotton and; +28% t o - 1 8 % for grape. The variability among models was larger for SO2 ranging from + 140% to -84%.
195 Table 4
Comparison of Overall Mean Deposition Velocities (cm s "1)
Site ....... EMEFS I EMEFS II Cotton Grape stable unstable B90-SO2
0bs. ...........Multi=Layer .................B i g L e a f ...............~ 0 M ...........~ M 0.67 0.50 0.62 0.60 0.37 0.18 0.23 0.07 0.15 0.13 0.50 0.20 0.24 0.27 0.31 0.22 0.24 0.34 Stratified by Stability (EMEFS I) 0.44 0.28 0.42 0.44 0.18 0.91 0.73 0.84 0.79 0.59 0.52 0.52 0.07 1.02 0.10
CONCLUSIONS There are a variety of approaches to modelling and/or parameterizing the deposition velocity of acidifying pollutants to surfaces. In this paper, 4 approaches were compared for 4 different sets of conditions. The a m o u n t of discrepancy between models and observations varied among models and from one surface type to the next. In addition, the agreement between models and observations decreased with increasing temporal resolution and reasons for agreement or lack of agreement varied among models. These results clearly suggest t h a t the models are not adequately representing the underlying processes involved in pollutant dry deposition. Results of comparisons, such as those presented here, can be used to improve the models. However, before models can be fully tested, there is a need for more flux m e a s u r e m e n t s for a larger variety of surface and atmospheric conditions and chemical species. More research is needed. Given the range of predictions t h a t arise from different models it is apparent t h a t for monitoring total deposition of air pollutants it will be important for all groups/countries involved to coordinate their modelling efforts.
REFERENCES
Baldocchi D.D., Hicks B.B. and C a m a r a P., 1987: A canopy stomatal resistance model for gaseous deposition to vegetated surfaces. Atmos. Envir., 21, 91-101.
196 Byun D.W., 1990: "Turbulent Transfer in the Atmospheric Surface Layer." in D.A. Haugen, (ed.), Workshop in Micrometeorology, pp. 381-392. American Meteorological Society. Hicks B.B., Wesely, M.L. and Durham J.L., 1980: Critique of methods to measure dry deposition. Workshop Summary, U.S. Environmental Protection Agency EPA-600/9-80-050, NTIS Publication No. PB81-126443, 70pp. Hicks B.B., Baldocchi, D.D., Meyers T.P., Hosker R.P. Jr. and Matt D.R., 1987: A preliminary multiple resistance routine for deriving dry deposition velocities from measured quantities. Wat. Air and Soil Pol., 36, 311-330. Louis J., 1979: A parametric model of vertical eddy fluxes in the atmosphere. Boundary-Layer Meteorol., 17, 187-202. Meyers T.P. and Baldocchi D.D., 1988: A comparison of models for deriving dry deposition fluxes of 0 3 and SO 2 to a forest canopy, Tellus, 40B, No. 4. Padro J., den Hartog G. and Neumann, H.H., 1991: An investigation of the ADOM dry deposition module using summertime 03 measurements obove a deciduous forest. Atmos. Envir., 25A, 1689-1704. Padro J., Massman W.J., Shaw R.H., Delany A. and Oncley S.P., 1994: A comparison of some aerodynamic resistance methods using measurements over cotton and grass from the 1991 California Ozone Deposition Experiment. Boundary-Layer Meteorol., (in press). Wesely M.L., 1989: Parameterization of surgace resistance to gaseous dry deposition in regional-scale numerical models. Atmos. Envir., 23, 12931304.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
197
EDACS: European Deposition maps of Acidifying Components on a Small scale
W.A.J. van Pul, C.J.M. Potma, E.P. van Leeuwen, G.P.J. Draaijers, J.W. Erisman Laboratory for Air Research, LLO National Institute of Public Health and Environmental Protection, RIVM P.O. Box 1, 3720 BA Bilthoven, the Netherlands
Abstract In this paper a description is given of the EDACS model (European Deposition of Acidifying Components on Small scale), with which the deposition of acidifying components on a small scale over Europe is calculated for 1989. The acidifying components considered in EDACS are sulphur and reduced and oxidized nitrogen compounds. Dry deposition is estimated with the inference method i.e the deposition at the surface is inferred from the concentration and the deposition velocity at the same height. The deposition velocity is calculated using a resistance model in which the transport to and absorption or uptake of a component by the surface are described. Dry deposition velocity fields over Europe are constructed from a detailed land-use map (1/6~ ~ lat/long grid, made by RIVM) and meteorological information using a detailed parameterization of the dry deposition process. These small-scale dry deposition velocity fields are combined with concentration fields from the EMEP Lagrangian long range transport model to yield dry deposition amounts on a small scale. Wet deposition is also estimated, based on measurements, to obtain a total acidifying deposition map at a European scale. These deposition fields clearly reflect the spatial detailed land-use information and the large-scale concentration pattern over Europe. The maps of the acidifying components over Europe on a small scale are made in cooperation with the EMEPkMSC-W, Oslo, Norway.
I Introduction In Europe, sulphur and reduced and oxidized nitrogen compounds are found to acidify soils and surface waters. Furthermore, nitrogen deposition causes eutrophication. The effects of acidification and eutrophication have been described extensively in the literature (e.g. Heij and Schneider, 1991, Grennfelt and ThOrnel6f, 1992). Especially the effects on ecosystems have obtained great attention due to their vulnerability. In order to protect these ecosystems critical loads have been defined above which there is an increased risk of damage to the ecosystems. The main atmospheric pathways via which emitted acidifying components reach the earth's surface are dry and wet deposition. The acidifying components can be transported over a long range up to 1000 km or more dependent on the component properties and the dry and wet deposition processes. Several models exist for estimating long-range transport of acidifying components on a European scale (e.g. EMEP: Sandnes, 1993, TREND: van Jaarsveld and Onderdelinden, 1994). The purpose of these models is to describe the relation between source
198 and receptor. The model results are used in quantifying country to country budgets, as a basis for the sulphur and nitrogen protocols and in assessments of the effects of acidification. The horizontal spatial scale on which these models operate is typically 50x50 km. However, describing the effects of acidification on the level of ecosystems, the acid load should be available at the scale of ecosystems or at scales which allow for comparison with critical loads (i.e. typically in the order of a lxl km resolution, Hettelingh et al., 1991). So the above model resolutions are not appropriate to describe the acidifying load at the level of ecosystems. A method was presented by van Pul et al., 1992a, with which the deposition of acidifying components on a small scale over Europe can be mapped. Their method was discussed at the ECE-EMEP/BIATEX Workshops on deposition at GOteborg, Sweden (LOvblad et al., 1993) and Aveiro, Portugal (Slanina et al., 1993) and accepted as currently the best available method to describe local acid deposition fluxes. In this paper the method is described and the fast, preliminary, maps of small-scale fluxes of acidifying components over Europe are presented. The calculations are made with the EDACS model (European Deposition of Acidifying Components on Small scale) in which the method is adopted. Recommendations made during the Workshops are incorporated in the model. The emphasis in the method is on modelling local scale dry deposition fluxes. A detailed parameterization of the dry deposition process for each acidifying component is based on available experimental results (EUROTRAC~IATEX project) and literature (Erisman et al. 1994a). Wet deposition is also estimated to obtain a total acidifying deposition map over Europe (van Leeuwen et al., 1994). The dry deposition maps were presented at the EUROTRAC symposium at Garmisch Partenkirchen, April 1994 by Erisman et al.(1994b) and van Pul et a/.(1994a). The maps of the acidifying components over Europe on a small scale are made in cooperation with the EMEtXaMSC-W, Oslo, Norway.
2 General descriptionof E D A C S An overview of the input for and calculation scheme of EDACS is presented in Figure 1. In EDACS the dry deposition is estimated with the inference method (Hicks, 1986). The deposition at the surface is inferred from the concentration and the deposition velocity at the same height. The deposition velocity is calculated using a resistance model in which the transport to and absorption or uptake by the surface of a component are described (see Section 3). The parameterizations are dependent on surface characteristics and other environmental and meteorological conditions. Dry deposition velocity fields, on a 6-hourly basis, are constructed from a detailed land-use map using these parameterizations along with the meteorological information (Potma, 1993). Here a RIVM data base is used which contains land-use data on a 1/6~ ~ lat/long grid over Europe (van de Velde et al. 1994). Finally the dry deposition amounts are calculated by multiplying these dry deposition velocity fields with concentration fields. In principle the concentration data can originate from measurements, model calculations or a combination of both. However, the spatial resolution of the operational European and national networks (ECE-EMEP, EUROTRAC) is too coarse to provide the necessary data and not every component is measured. This will lead to large uncertainties in the interpolated concentration fields. On a local scale, national or local networks may provide the concentration data. When the local maps are aggregated into one European map, problems
199 may arise about the inconsistency between the networks and it is foreseen that still not a full coverage of Europe can be obtained. However, for parts of Europe with small horizontal concentration gradients this can be done e.g. for UK:UK Review group on acid rain, 1990; Sweden: l_2)vblad et al. 1993, The Netherlands; Erisman 1992). In this paper the concentration data of the EMEP Lagrangian long-range transport model (hence EMEP-LRT) on a 150x150 km scale were used (as described in Sandnes, 1993) to obtain a consistent concentration field over Europe. Using calculated concentration fields, the relation between emissions and deposition is maintained and assessments or scenario studies can be made at different scales. In the inference method it is assumed that a constant flux layer is present between the reference height and the surface i.e. the atmospheric surface layer. This assumption implies that there is no significant advection in the layer and the air flow is well-adapted to the surface properties of the depositing surface and chemical reactions are not present. In that
Figure 1 Overview of the input for and calculation scheme of EDACS. For explanation see texL
200 case the deposition flux at the reference height equals the deposition flux at the surface. The adaptation of the air flow to the surface is strongly dependent on the surface roughness and the stability of the air. The choice of the reference height is a compromise between the height where the concentration is not severely affected by local deposition and is below the surface layer height. In EDACS the concentration at 50m is taken which is the lowest LRT model level above the surface. This concentration then is assumed to be representative for a certain area, here an EMEP LRT gridcell of 150x150 km, and consequently can be used in estimating the deposition to surfaces within this area. The wet deposition of acidifying components is based on measurements of the concentrations in precipitation and rain amounts (van Leeuwen et al., 1994). A concentration map over Europe was constructed from these data by kriging. A data set with interpolated values of long-term yearly precipitation amounts was used to calculate the wet deposition. The components considered here are SO 2 and SO42--aerosol (SOx), NO, NO 2 (NOx), HNO 3 and NO3--aerosol and NH 3 and NHa+-aerosol (NHx). NH x is considered to be acidifying because of the nitrification processes in the soil in which H ÷ is produced (van Breemen et alo 1982). If all above components which are deposited produce one equivalent of acid this leads to a total potential acid load which is estimated from: potential acid = 2 SO x + NOy + NH x. The actual acid load differs from the potential load because of an incomplete nitrification of NH 3 and by neutralization of the acidity by base cations. HONO, PAN and HNO 2 are not taken into account. However, the contribution of these components to the total acidifying deposition is very small (e.g. L6vblad et al., p 19).
3 Parameterization of the dry deposition velocity The dry deposition flux of gases and particles from the atmosphere to a receptor surface is governed by the concentration in air and turbulent transport processes in the boundary layer, by the chemical and physical nature of the depositing species and by the efficiency of the surface to capture or absorb gases and particles. The flux of a trace gas is given as: F : Va(z) c(z)
(1)
where c(z) is the concentration at height z and V,t is the dry deposition velocity (Chamberlain, 1966). z is the reference height above the surface: here taken as 50m. If the surface is covered with vegetation, a zero-plane displacement, d, is included: z=z-d. The absorbing surface is often assumed to have zero surface concentration. This holds only for depositing gases and not for gases that might also be emitted, such as NH 3 and NO. For these gases a non-zero surface concentration, a compensation point cp, might exist, which can be higher than the ambient concentration, in which case the gas is emitted. Here the concentration at the various surfaces, c s, is assumed to be zero for all components because of insufficient knowledge of the compensation point. The parameterization of the dry deposition velocity is based on a description of this process via a resistance analogy or Big Leaf Model (see e.g. Thom, 1975, Hicks et al., 1987, Fowler, 1978). In this resistance model the most important deposition pathways via which the component is transported and subsequently destroyed at, or taken up by the surface, are parameterized. The resistance model used here is shown in Figure 2.
201
C(z) ra
rb
rstom
rinc rext rs~
~
C~(soil) C~(ext)
rm
Cs(m)
m
Figure 2 Resistance or Big Leaf model used in EDACS. c,(soil), c,(ext) and cs(m) denote the concentration at the soil external or mesophyl surface respectively, for explanation of the other symbols see text). V d is the inverse of three resistances:
(2)
V d - ( r a + r b + r s ) -1
These three resistances indicate the three stages of transport: the aerodynamic resistance, r a, represents the resistance against turbulent transport of the component close to the surface; the quasi-laminar sublayer resistance, r b, accounts for the transport of the component through a laminar layer adjacent to the surface by molecular diffusion and the surface resistance r s for the uptake or destruction at the surface. This surface resistance is composed of the resistances of the various destruction or uptake processes at the surface. For a surface covered with vegetation this is: -the stomatal resistance, rstom , the resistance to the transport through the stomata of leaves and needles; -the mesophyl resistance, r m, the resistance of the internal plant tissues against the uptake or destruction (in a chemical way); -the cuticle resistance, reut, or external surface resistance, rext, the resistance of the exterior plant parts against the uptake or destruction of the component; -the ri~ the in-canopy aerodynamic resistance to account for the transport of air above the vegetation towards the soil and lower plant parts; -rsoa, the soil resistance, the resistance against destruction or absorption at the soil surface; These resistances which act in parallel or series are summed up to yield a (total) surface resistance, rs: rs =
-1
(rin c + rsoil )-1 + rex t + (r m + rstom)
-1] -1
(3)
202 For a water surface: r s = rwat, where rw~t is the resistance against the solution of gases in water. For bare soil: r s = rsoa and for urban areas: r s = rurban. When the surface is covered with snow r~ = rsnow.
In turn, these resistances are affected by meteorology, leaf area, stomatal physiology, soil and external leaf surface pH, and presence and chemistry of water drops and films. Especially the state of the leaf and soil surface i.e. the presence of water films and snow, is an important variable in the deposition of soluble gases such as SO 2 and NH 3. The process of dry deposition of particles of acidifying components is not very well known compared to the gaseous counterparts (Ruijgrok et a/.,1993). As a best estimate the dry deposition of particles is described using a parameterization by Wesely et al.(1985) and Erisman (1992). Recent information on the deposition of particles to forests has come available (Ruijgrok et al. 1994) and will be used in a future version of the model. The scheme used here to derive the surface resistances for SO 2, N O 2, NO, HNO 3, NH 3 is described in Erisman et al.(1994a). This scheme is based on previous publications among others Wesely (1989), l./Svblad et a1.(1993) and recent dry deposition measurements (among others in the BIATEX project of EUROTRAC). More details on the actual parameterizations used in EDACS can be found in van Pul et a1.(1995).
4 Results
The 6-hourly deposition velocity fields were averaged to daily values and multiplied with the daily EMEP/LRT concentrations. These daily dry deposition maps were summed to annual totals. In Figure 3 the dry deposition of total potential acid over Europe estimated with EDACS is shown. The dry deposition values for most components vary greatly over Europe. This is partly explained by variations in the deposition velocity caused by variations in land use and meteorological conditions over Europe. The concentration pattern of the components over Europe, which are associated with the distribution of emissions, introduce variations on a larger scale (150xl50km i.e. the EMEP grid). Large emission areas can be detected in the maps e.g. for SO 2 and NO~ this is the so-called black triangle (Eastern Germany - PolandCzech Republic), for ammonia e.g. north western Europe (The Netherlands, Denmark). The total potential acidification map which is the sum of the dry deposition and wet deposition is presented in Figure 4. This figure reflects the above mentioned variations. The relative contribution of dry and wet deposition can be observed. For instance in the Scandinavian countries the surface inhomogeneities are not represented due to the large contribution of the wet deposition which has a smooth distribution over this area. In Figure 5 the standard deviation of the total deposition of EDACS cells in an EMEP-grid cell (about 100 EDACS cells in one EMEP-grid) is given as absolute values and values relative to the average per EMEP-grid. It can be seen that the largest absolute values of the standard deviation can be found in the above described emission areas. Whereas the relative standard deviation is largest in areas with a small deposition and so variations in land use, meteorology etcetera are reflected.
V
~
~
~
~
A
I DDlnnnn
..
...
~
_
aqi,
,
_
-=.
"-~
,,,
203
g~
r~
g~
t~
Figure 4 Potential acidification map over Europe on 1/6°xl/6 ° lat/long grid for 1989 (tool ha "1 year1).
205
Figure 5 Standard deviation of the total deposition of EDACS grid cells per EMEP grid in absolute values in mol ha -1 year -~ (above) and values relative to the total deposition per EMEP-grid cell in % (below).
206
5 Uncertainties The maps shown in Section 4 have a limited accuracy and are therefore preliminary. The aim of these maps is to show the variations of the deposition of acidifying components on a small scale. Several uncertainties and shortcomings are present which need some discussion. We will adress these items here and will suggest a quantification of the uncertainties. Also some recommendations for improvements of these maps are given. One of the main uncertainties in the maps is in the simple resistance model and especially the surface resistance parameterization for estimating the dry deposition of different gases and particles. The resistance model is a simple approach for a highly variable process. It assumes a constant flux layer, i.e. there are no surface inhomogeneities, edge effects or chemical reactions. How much these simplifications contribute to the total uncertainty in the annual average deposition fluxes has not been investigated. The uncertainty in the surface resistance parameterization is the largest uncertainty in this simplified scheme. Therefore more, and more accurate, parameterizations are needed for various vegetations and surfaces. Moreover there is a lack of measurements on which these parameterizations can be based especially for southern and eastern European climates and surfaces. This is needed to obtain parameterizations for use in LRT models which are valid for the whole of Europe. Part of the uncertainty in the surface resistance parameterization is due to the mismatch between the available parameterizations for a limited number of landcover types and the landuse classifications used in the RIVM land-use data base. Surface wetness is found to be one of the major factors influencing the deposition process. In the present version of EDACS only rain and an indication of dew is used. In the next version the dew amount will be modelled in more detail using appropriate surface properties. The evaporation of rain and dew will also be parameterized. This means that an administration of the available energy and moisture flux during the day has to be made. An indicator on the presence and condition of a snow layer will also be taken into account. The overall uncertainty in the surface resistance due to the above factors is different for each component and surface type. This uncertainty, on a annual basis, is a few tens percent points but can easily exceed 100%. In the current version of EDACS, the EMEP-LRT concentration maps on a 150x150 grid are used. The uncertainty in the concentrations are estimated at 40-70% by a statistical analysis with the EMEP measurements (Krtiger, 1993). These concentrations represent the background situation in Europe. It is assumed that the concentration distribution within a grid is homogeneous. This is not the case in a grid which contains industrialised areas or many scattered sources such as of NH 3 and NO x. For such conditions, subgrid concentration variations are present and will lead to underestimates of the deposition in that grid. To obtain an indication of the errors, a small-scale, short-range model can be useful here to resolve subgrid concentration gradients for dense source areas. The uncertainty in the deposition in an EDACS grid cell due to these gradients is estimated at 25% (Berg and Schaug, 1994). The deposition in EDACS is based on the EMEP-LRT concentrations which in turn are dependent on EMEP deposition estimates. The deposition in the EMEP-LRT model and in EDACS are calculated in different ways. By using other dry deposition velocities in EDACS a mass inconsistency, between the EMEP calculated deposition and the small scale maps by EDACS, is introduced. However, if the differences in the used deposition descriptions between the two models are not very large and non-systematic over a larger region, this will
207 not lead to large mass inconsistencies. In Figure 6 a comparison between the sulphur dry deposition per country estimated by EMEP and EDACS is shown. It can be seen that on average there is a good agreement indicating that for the model area the mass consistency is not violated to a large extent. However, for some countries the deviations can be as large as 50%. To avoid this mass inconsistency it is planned to implement the deposition module in the EMEP-LRT model. In this way the calculated concentration fields are consistent with the EDACS deposition description.
2000 0 0
E 1500
.S
0 >
o
1000
0 0
w
500
QO
I
2 0 0
500
1000
1500
2000
EMEP country averages (mol/ha/y)
Figure 6 Comparison between the country averaged dry deposition of sulphur calculated with EMEP and EDACS (deposition in tool ha s year'1). In inferential modelling the concept is used that the surfaces, at which the deposition is calculated, should have a certain horizontal length. As an approximation this is about 100 times the reference height (Pasquill, 1972). The typical horizontal length scale of the surface, using 50m as the reference height, is 5 km. This means that only surfaces with larger length scales are modelled correctly. Variations in land cover on this scale are regularly present. At each surface transition the deposition is altered. Especially forest edges give rise to large enhancements of the deposition. This enhanced deposition can be dealt with in a very simplified way using correcting factors defined by e.g. van Pul et al. 1992b, Draaijers et al., 1994. Since in this land-use data base only percentages of the land-use type per grid are given and not the geographical position, only a statistical approach of the uncertainty can be carded out. For instance, the enhanced deposition at forest edges as a whole for all forest stands in the Netherlands, is estimated at 10-30% (van Pul et al., 1992b, Draaijers et al. 1994). The accuracy of the presented results depends on the availability and quality of the input data such as the land-use map and the meteorological observations. In the gridded version of the RIVM land-use data base, forest is not subdivided into deciduous and coniferous. All forest is classified as coniferous forest. This will probably lead to overestimates of the deposition
208 velocity to deciduous forests for all components during winter. However, the stomatal resistance in winter will be large due to low temperatures. So this overestimate will be somewhat leveled out. A version in which the forest data are subdivided in the above categories will be available in 1995. In this new version the quality of the land-use data for some areas in Eastern Europe will also be improved. The dry deposition is calculated on a daily basis. However, due to the daily averaging of the concentration and the deposition velocity, a loss in temporal correlation is introduced between the concentration and the deposition velocity. This error is component specific and is estimated to be smaller than 20% (van Pul et al.,1993). In the future the deposition will be calculated on a 6-hourly basis and the above error avoided. The uncertainty in the wet deposition estimates is relatively small compared to the uncertainty in the dry deposition estimates. A comparison between the derived wet deposition maps and EMEP long-range transport model results was carded out by means of calculation of differences and ratios between the grids derived by the two methods (van Leeuwen et al., 1994). In most parts of Europe deviations were found smaller than 200 mol/ha/year in absolute terms for individual components and smaller than 50% in relative terms. The uncertainty in the wet deposition is most important in areas where wet deposition is equal to or higher than dry deposition. In such areas, however, wet deposition usually shows a smooth pattern. This is not true for mountainous regions where additional deposition pathways such as fog and cloud deposition are present. Corrections can be applied to the wet deposition if local data on fog and cloud composition, occurrence and liquid water content are available (Fowler, 1991). The overall uncertainty in the depostion maps consists of the above-mentioned uncertainties. However, the uncertainty in the surface resistance and the occurence of sub-grid concentration gradients will act as the largest uncertainty sources. Given these uncertainty estimates, the uncertainty in the deposition of a component of a 1/6~ ~ lat/long gridcell is typically 100200%. However, the uncertainty in the total potential acidification map is smaller because the total acidification consists of components such as wet deposition which have a smaller uncertainty. 6 Conclusions In this paper a description is given of the EDACS model, with which the deposition of acidifying components on a small scale over Europe is calculated. Dry deposition velocity fields are constructed from a detailed land-use map (1/6~ ~ lat/long, made by RIVM) and meteorological information using a detailed parameterization of the dry deposition process. These small-scale dry deposition velocity fields are combined with air-concentration fields (taken from the EMEP Long range transport model) to yield dry deposition amounts on a small scale. Wet deposition is also estimated to obtain a total acidifying deposition map over Europe (van Leeuwen et al., 1994). These deposition fields clearly reflect the spatially detailed land-use information and the large-scale concentration pattern over Europe. With these fields a better match is obtained between the critical and actual loads when ecosystems are concerned. The presented deposition fields are preliminary because of several shortcomings present in the method and data bases. An update of the deposition fields and calculations for more recent years will be available in 1995. A more thorough uncertainty analysis of the deposition maps,
209 a validation with (throughfall and micro-meteorological) measurements and corrections on the wet deposition caused by cloud and fog deposition will also be carded out.
Acknowledgements A.Eliassen, E.Berge and H.Styve of EMEP MSC-W are thanked for their cooperation in providing the model concentration data. The ECMWF and KNMI are acknowledged for the WMO synops data. F.de Leeuw is acknowledged for his comments on the draft and J.Burn for editing the manuscript.
References Breemen,N.van, P.A.Burrough, E.J.Velthorst,H.F.van Dobben, T.de Wit,T.B.Ridder and H.F.R.Reinders, 1982. Nature,299,548-550. Chamberlain A.C., 1966. Proc. R. Soc. Lond. A290, 236-265. Draaijers,G.P.J.,R.van Ek, and W.Bleuten, 1994. Boundary-Layer Meteorology 69, 343-366. Erisman J.W., 1992. PhD-thesis. University of Utrecht, the Netherlands. Erisman, J.W.,W.A.J. van Pul and G.P.Wyers, 1994a. Atmospheric Environment Vol. 28, No. 16: 2595-2607. Erisman, J.W.,C.J.M.Potma, G.P.J.Draaijers, E. van Leeuwen and W.A.J. van Pul., 1994b. In Proceedings of EUROTRAC symposium'94, P.Borrell (Ed.), SPB Academic Publishing, The Hague, the Netherlands. Erisman J.W., 1994c. Atmospheric Environment Vol.28, No.16: 2583-2594. Erisman J.W. and Baldocchi D.D., 1994d. Tellus 46B, 159-171. Fowler D., 1978. Atmospheric Environment 12,369-373. Fowler D., J.H.Duyzer, D.D.Baldocchi, 1991. Proc. R. Soc. Edinburgh 97B,35-59. Grennfelt,P. and E.Th6rnel6f, 1992. Report from a workshop at L6keberg, Sweden, April 610, 1992. Report No. Nord 41, Nordic Council of Ministers, Copenhagen. Heij, G.J. and T.Schneider (Ed.), 1991. Studies in Environmental Science 46. Elsevier,Amsterdam. Hettelingh,J.P.,R.J.Downing and P.A.M.de Smet, 1991. CEC technical report no.1 RIVM report 259101001. Hicks, B.B.,1986. Water,Air and Soil Pollution 30: 75-90. Hicks B.B., D.D.Baldocchi, T.P.Meyers, R.P.Hosker Jr. and D.R.Matt, 1987. Water Air Soil Pollut. 36, 311-330. Hicks B.B., R.R.Draxler, D.L.Albritton, F.C.Fehsenfeld, J.M.Hales, T.P.Meyers, R.L.Vong, M.Dodge, S.E.Schwartz, R.L.Tanner, C.I.Davidson, S.E.Lindberg and M.L.Wesely, 1989. State of Science/Technology, Report no. 2. National Acid Precipitation Assessment Program. Jaarsveld, J.A. van and D.Onderdelinden, 1995. RIVM report in preparation. Kriiger,O., 1993. In: ProceeAings CEC/BIATEX Workshop 4-7 May 1993 Aveiro, Portugal. Ed. J.Slanina. pp 31-38. L6vblad, G.,J.W.Erisman and D.Fowler, 1993. Proceedings Nordic Council/EMEP/BIATEX workshop in G6tenborg 3-6 November 1992. Leeuwen, E.P. van, J.W.Erisman, G.P.J.Draaijers, C.J.M.Potma and W.A.J.van Pul, 1994. Report no. 722108008. RIVM, Bilthoven, the Netherlands. Pasquill,F., 1972. Quarterly Journal of the Royal Meteorological Society 98:469-494. Potma, C.J., 1993. RIVM report 722401001 Bilthoven, the Nethedands.
210 Pul W.A.J. van, J.W.Erisman, J.A. van Jaarsveld and F.A.A.M. de Leeuw, 1992a. In Acidification research: evaluation and policy application. (edited by T. Schneider), Studies in Environmental Science. Elsevier, Amsterdam. Pul W.A.J. van, R.M.van Aalst and J.W.Erisman, 1992b. In: J.Slanina,ed., EUROTRACK/BIATEX annual report,Garmisch-Partenkirchen, FRG 248-254. Pul W.A.J. van, J.W.Erisman, J.A. van Jaarsveld and F.A.A.M. de Leeuw, 1993. In: Proceedings CEC/BIATEX Workshop 4-7 May 1993 Aveiro, Portugal. Ed. J.Slanina. pp 95-115. Pul W.A.J. van and A.F.G. Jacobs, 1994. Boundary-Layer Meteorology 69: 83-99. Pul W.A.J. van, C.J.M.Potma, G.P.J.Draaijers, E. van Leeuwen and J.W.Erisman, 1994. In: Proceedings of EUROTRAC symposium'94, P.Borrell (Ed.), SPB Academic Publishing, The Hague, the Netherlands. Pul W.A.J. van, C.J.M.Potma, E.P. van Leeuwen, G.P.J.Draaijers and J.W.Erisman, 1995. RIVM report 722401005 Bilthoven, the Netherlands. Ruijgrok W. and C.I.Davidson C.I., 1993. In: Proceedings Nordic Council/EMEP/BIATEX workshop, GStenborg, Sweden, 3-6 November 1992. Ed. l.~vblad, G., J.W.Erisman and D.Fowler. Ruijgrok W., H.Tieben and P.Eisinga, 1994. Report 20159-KES/MLU, KEMA, Arnhem, the Netherlands. Sandnes H., 1993. EMEP report 1/93. MSC-West, Oslo, Norway. Slanina, J., G.Angeletti and S.Beilke, 1993. Proceedings CEC~IATEX Workshop 4-7 May 1993 Aveiro, Portugal. Thorn A.S., 1975. In: Vegetation and Atmosphere, pp. 58-109 (Ed. Monteith J.L.), Academic Press, London. UK Review Group on Acid Rain, 1990. Warren Spring Laboratory, Stevenage, UK. Velde van de, R.J., W.Faber, V.Katwijk, H.J.Scholten, T.J.M.Thewessen, M.Verspuy and M.Zevenbergen, 1994. report 712401001, RIVM, Bilthoven, the Netherlands. Voldner E.C., L.A.Barrie and A.Sirois, 1986. Atmospheric Environment 20,2101- 2123. Walcek, C.J., R.A.Brost, J.S. Chang and M.Wesely, 1986. Atmospheric Environment 20, 949964. Wesely M.L., D.R.Cook and R.L.Hart, 1985. J. geophys. Res. 90,2131-2143. Wesely M.L., 1989. Atmospheric Environment 23,1293-1304.
E F F E C T S OF A C I D D E P O S I T I O N O N F O R E S T E C O S Y S T E M S I N T H E NETHERLANDS SESSION V
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
213
A S S E S S M E N T AND E V A L U A T I O N OF C R I T I C A L LEVELS FOR O3 AND NH3
Eveliene Steingrfver', Tom Dueck b & Ludger van der Eerden b
9 IBN-DLO Institute for Forestry and Nature Research, P.O. Box 23, 6700 AA Wageningen, The Netherlands b AB-DLO Research Institute for Agrobiology and Soil Fertility, P.O. Box 14, 6700 AA Wageningen, The Netherlands
Abstract The effect of 03 on three different tree species was similar. A threshold of 40 ppb was found for Pinus and for Pseudotsuga. Over two growing seasons growth was inhibited in Pinus and Fagus saplings and total assimilation per tree was inhibited over one growing season in Pseudotsuga trees. Pinus was most sensitive to 03, but a critical level of 10 ppm.hrs is sufficient to protect the three species from O3 damage. The effect of NH3 on growth was ambiguous. Growth was unchanged in Fagus, stimulated in Pseudotsuga and inhibited in Pinus. The effect of NH3 on tree architecture and stress sensitivity was similar. Tree architecture was changed in both Fagus and Pseudotsuga and drought and frost sensitivity were increased in Pinus and Pseudotsuga. At the moment both the critical level for 03 and for N H~ are exceeded in The Netherlands in general and in selected areas in particular.
1. INTRODUCTION
The current estimation of the mean nitrogen load in the Netherlands is 35 kg ha -I y r I of which the largest proportion (16 kg ha -~ yr -~) is gaseous ammonia. The regional mean NH3 concentrations, however, may be very different from the average national concentration, resulting in a much higher exposure level for forest trees in areas with intensive live stock farming [ 1]. Ozone, another relevant air pollutant, is more evenly distributed over the Netherlands (Tab. l). The national mean concentration exceeds the phytotoxic level for crops [2], and the economic impact on Dutch crops is large 13]. The effect of 03 on mature forest trees, on forest ecosystems and on natural ecosystems in general is largely unknown. The available information is mainly acquired from experiments with seedlings and saplings 141. Two types of air quality standards are currently in use: critical levels and critical loads. Critical levels are based on exposure values, while critical loads are based on deposition values. The use of two different air quality standards is .justified, because critical levels focus on individual air pollution components and on nondepositing oxidants such as 03, whereas critical loads lump either nitrogen or
214 acidifying components together. Moreover critical levels are usually used for shortterm exposures (hours - months) and critical loads for long-term exposures (years). The aim of the present investigation was to assess the impact of NH3 and 03 on forest ecosystems at concentrations presently occurring in the Netherlands in relation to the critical level. This was done by comparing the effects on tree saplings exposed to both pollutants in OTC's (Open Top Chambers) with the effects on mature forest trees under ambient field conditions.
2. MATERIAL AND METHODS
2.1 Fumigation experiments in OTC's Two experiments were performed in OTC's, in which 3-yr-old Beech (Fagus sylvatica L.) and Scots pine (I~'nus sylvestris L.) saplings were fumigated for 15 months, largely covering two growing seasons (from June l to September 1). The OTC's have been described earlier [5]. Ammonia and ozone were injected into the air stream prior to the blower via thermal mass-flow controllers (Brooks 5850 TR). Air pollutant concentrations were sequentially monitored with an ozone analyzer (8810, Monitor Labs) and a NH3 monitor (thermoconverter model 8750 followed by a chemiluminescent NOx analyzer model 8840, both Monitor Labs) and were computercontrolled. The data were recorded with a HP data acquisition system. A duplicated range of 03 concentrations (0, 30, 60, 90, 120 and 150 #g m -3) was used in the first experiment and these concentrations were both higher and lower than current ambient concentrations in the Netherlands (Tab. l).
Table 1 Average concentrations of N H 3 and 03 (#g m 3) in 1992.
The Netherlands Wageningen, OTC's Veluwe, field
NH3
03
3.4 16 2.5
60 72 78
Trees were exposed to 03 during a 9 h day and to a third of the daytime concentrations during the remaining 15 h. To one of the 03 ranges, a concentration of 40/~g m 3 NH3 was added (24 h dayS), which is somewhat higher than the highest mean concentrations experienced in the Netherlands. In the control treatments, the NH3 concentration was also higher (15 #g m -3) than the national mean due to the tact that Wageningen is located in a region with high NH3 concentrations (Tab. 1), and the filters used to clean the air have less that 50% capacity for NH3. In the second experiment, 03 and NH3 were applied in factorial design, resulting in a triplicated set of 03 concentrations (setpoints 0, 90, and 135/~g m3), supplemented with ambient air, or with NH3 to 40 ~g m -3 and 80 /~g m 3 NH3. The water potential on 1-year-old
215 needles of P~'nus sylvestris was measured with a pressure bomb 161. The measurements were performed in six OTC's only, in all three NH3 treatments combined with filtered air and the highest 03 concentration and were performed between 08.00 and 13.00 hours in fully watered pots and again after five days without water. The soil water potential was measured daily to ensure that the drought treatment was not prolonged to the point where excessive drought injury to the trees occurred in order to relate the soil water potential to the needle water status. This paper discusses the main results on biomass production and drought sensitivity.
2.2 Correlation studies in the field All field measurements were performed in a stand of 34-year-old Douglas fir (Pseudotsuga menziesii (Mirb.) Franco L., provenance Arlington), located at the Veluwe, in the central part of the Netherlands. The Speuld site has a stand density of approximately 800. In 1993, the average tree height was 22.2 m and the average DBH was 25.4 cm 17]. The mean 03 and NH3 concentration at Speuld are shown in Tab. 1. A computer controlled field gas exchange system was installed at the site. With this system up to 16 different branch assimilation chambers can be measured continuously. A 22 m high scaffolding was built in the middle of the stand, from which the branch assimilation chambers could be mounted in 8 different trees. The chambers were ventilated with ambient air and contained one year class of needles. Temperature, relative humidity and CO2 concentration of the air entering the chambers closely resembled that of ambient outside air. A PAR sensor was mounted at the outside of each chamber. Each chamber was sampled twice every hour during 24 hours per day from March 1992 up to December 1993. The light response curve was modelled using the equation of Goudriaan I81 with the measured photosynthetic rate and photoactive radiation levels. The unexplained variance in the data could be reduced by 40% by taking into account changes in vapour pressure deficit (VPD) and the 03 concentration. In this way, the direct effects of 03, NH3 and NOx and VPD on the photosynthetic rate were statistically estimated. This paper discusses the results on CO2 assimilation, per month and per year, and on the needle nutrient status.
2.3
Exposure-response relationship for 03
The levels of exposure to 03 used here are not expressed as concentrations, but as the accumulated exposure over a threshold concentration, abbreviated with AOT [41. The AOT adds all exposures above a certain threshold concentration over the period of interest, i.e. growing season, but also over several years in which mean concentrations can strongly fluctuate over the seasons. The threshold value is the concentration above which 03 toxicity becomes evident. The AOT approach proved to be very useful for evaluation of crop loss in which relative yield reductions could be linearly fitted to AOT values. Recently, this approach has also been applied to forest trees, although reliable field data are scarce [4].
216
3. RESULTS 3.1. Fumigation experiments in OTC's Fig. 1 shows the effects of O3 and NH3 on the total biomass of P. sylvestris. The biomass of trees exposed to the lowest 3 levels of 03 was significantly higher than that in the 3 highest levels. An effect threshold for 03 is visible between 5 and 15 ppm-hrs. The 3 lowest 03 concentrations below 40 ppb were logically similar in terms of AOT40, and this was also reflected in the biomass production in those treatments, indicating the validity of the threshold for P. sylvestris. The absence of a linear exposure-response relationship is likely due to the fact that high concentrations of 03 inhibit stomatal conductance 19] and thus the uptake of 03. The UNECE and the WHO recently recommended an AOT40 of 10 ppm-hrs as a critical level for trees, at which a 10% reduction in biomass production should not be exceeded. The results shown for P. sylvestris in Fig. 1 appear to support this critical level. Current concentrations of O3 in the Netherlands however, exceed this critical level by more than a factor 2, indicating that O3 is significantly reducing the growth of P. sylvestris. Fig. 1 also indicates that at low 03 concentration, NH3 tends to inhibit growth, but growth is not further reduced at higher 03 concentrations. 1000
A OI
9OO
-NHa
0
0
Jr ..,.. O
3=
aE"
9
+NH 3
800
700
. 0
.
. 20
.
. 40
.
. 60
.
. 80
100
AOT40 (ppm-h)
Figure 1. Effect of 03 and NH3 on the total biomass production Pinus ,~ylvestris. o O3 alone; 9 - O3 + NH3. In the same fumigation experiment, F. sylvatica was found to be less sensitive to 03; a 10% reduction in biomass production was found at an AOT40 of 30 ppm-hrs, while NH3 had no effect on the total biomass. The growth of F. ,~ylvatica was differentially affected by NH3 and 03 in tree architecture rather than in biomass production (Fig. 2). Although 03 inhibited tree height, it increased stem diameter, which resulted in relatively sturdier trees. NH3 did not influence tree height and reduced stem diameter, resulting in relatively smaller trees at higher levels of 03. Lateral branch growth was also reduced by NH3 with increasing concentrations of 03, thus reducing the potential for light interception. In the second fumigation experiment, special attention was paid to the effect of NH3 and 03 on the drought sensitivity of P. sylvestris. The data in Fig. 3 show that needles of fully watered P. ,~ylvestris saplings have a significantly higher (more negative) water potential when exposed to NH3 alone than when exposed to NH3 +03.
217
100
Length increment (cm)
lO
Diameter incr.ement (cm) o
gO
9
80 70 60
o l
0
,
i
20
,
l
,
40
l
,
60
80
,,
.
~
9
9
0
1 O0
20
AOT40 (ppm-h)
160
"
40
60
80
..i
100
AOT40 (ppm-h)
Branch dry wt (g)
150 140
130
o
o
-NH3
120 9 +NH 3
110
100 90 80
"
0
20
40
60
80
'
100
AOT40 (ppm-h)
Figure 2. Effect of O3 and NH3 on mean length increment, stem diameter increment and branch dry weight of Fagus sylvatica, o = O3 alone 9 = O3 + NH,.
No difference was observed in the needle water potentials between the three treatments of only NHs. It appears that, irrespective ~)I the NH, ct)ncentrati~)n. O, inhibits stomatal conductance and thus transpiration in the absence c)f water stress. allowing the trees to maintain turgor at a lower (200-400 k Pa) needle water potential than when exposed to NH3 alone. When the trees were droughted for 5 days, the water potential increased linearly with increasing concentrations of NHs in the absence of 03. This indicates that under conditions of drought stress, N H~ disrupts the stomatal control, resulting in increased transpiration and reduced water use efficiency I! 0-i 11. When O, was added to the NH3 treatments, the effect of increasing N H, c()ncentrati~)ns was masked, and the water potential in all three Os + NH~ treatments remained lower than in treatments with NHs alone. This confirms earlier experiments, in which the reduced transpiration of various species exposed to 03 is attributed to inhibition of stomatal al~rture [91. Thus, even at the concentrations of N H3 used in this experiment, O3 appears able to reduce water loss through transpiration and is thus able to reduce drought stress.
218 1400 w a t e r e d [ - - ' ] drought A a
1160
Jr
"~ ...
920
C 0 0
=.
680
=_
=
~:
440
i
200
FA
40N
80N
03
40N+03 80N+03
Figure 3. Effect of O3 and NH3 on drought stress of Pinus sylvestris. Dark bars indicate water potential (means + SE) of watered trees and white bars that of droughted trees (n= 10).
3.2. Correlation studies in the field The vitality of the Speuld stand was characterized in 1986 according to international EC standards of defoliation and foliar discoloration. The vitality was higher than the nationwide average for P. menziesii, which was classified as less healthy for 50% of all trees 1121. The vitality was re-assessed in 1994, and was similar to the nationwide average, indicating that the vitality of the stand had decreased to the level of most other P. menziesii stands in The Netherlands, which was classified as less healthy for 80% of all trees 171. Notwithstanding the poor vitality, biomass production was high. According to yield tables for Dutch P. menziesii stands I131, based on data from the pre-intensive livestock age, annual wood volume increment was higher than expected. Speuld can be qualified as an average soil quality site (class 111) with an expected mean annual volume increment of around 13.4 m 3 ha 1. The average annual increment over the period 1987-1993 however, was 24.7 m 3 ha -1, which is even higher than the expected volume increment of 17.3 m 3 ha 1 on the best quality site class.
Table 2. Biomass partitioning in Pseudotsuga menziesii using foliage I71 and root !141 data from Speuld and from i15-161. Speuld
low productive, slow grow rate
high productive, fast grow rate
5.5
0.95-1.1
2.7-3.6
0.67
0.21-0.34
0.19-0.31
Foliage/ fine root
Foliage/ coarse root
219 The dry weight distribution in Speuld in which the N deposition is lower than the Dutch mean, was compared to that of other P. menziesii stands. Speuld can be best compared to high productive, fast growing stands of the same age. The foliage/root ratio in Speuld is exceptionally high compared to the other stands (Tab. 2). The high foliage/root ratio results from both lower root biomass and higher needle biomass. The high amount of needle biomass is also reflected in the LAi of the stand which was 10.7 in 1992. The foliage/coarse root ratio is also high (Tab. 2). The branch dry weight however, was lower compared to the other stands. The nutrient status of the needles changed significantly during the experimental period [8]. The average N concentration increased from 1.7 % in 1987 to 2 % in 1993. The optimal concentration for biomass production in P. menziesii is 1.8%. The K concentration decreased from 0.7% to 0.5 %, which is below the deficiency level of 0.6 %. The P concentration remained constant, but was also below the deficiency level of 0.14%. The N/P ratio was constantly below/above the deficiency level, while the N/K ratio increased from 2.7 to 3.7. Extrapolating the nutrient trends found over the past 7 years, the ratio of N/K can be expected to reach the deficiency level within 1-2 years. High N/K ratio's, as found in the Speuld stand, are often considered as an indicator for increased stress sensitivity, e.g. frost sensitivity !171. There seems to be little doubt that the high nitrogen status, the high productivity, the high foliage/root ratio, the high LAI and the increased stress sensitivity are caused by the high N deposition into the stand.
Table 3. Proportional reduction of the monthly CO2 assimilation due to 03 and VPD (mean over 2 needle age classes and three crown levels). 03
VPD
03
1992 April
3.3
May
VPD 1993
2.4
28.0
14.0
12.2
22.9
13.2
8.1
June
4.4
20.1
4.4
6.1
July
5.3
6.3
7.6
2.9
August
4.1
7.0
5.4
NS
Net CO2 assimilation was measured more or less constantly throughout 1992 and 1993 and was related to changes in meterological parameters and air pollution concentration. No direct effects of NH3 and/or NOx on net CO2 assimilation of P. menziesii were found. NH3 tended to stimulate assimilation, but the differences were not significant. The only obvious relationship of CO2 assimilation with air pollution was that with 03 (Tab. 3). The reduction of net CO2 assimilation by 03 on a monthly basis can be quite considerable, and was in the same range as the reduction by VPD.
220 The annual reduction of total CO2 assimilation per tree was estimated to be in the range of 3-10% in 1992. The annual reduction over 1993 will probably be higher, as the monthly reductions are higher. In order to assess the critical level for 03 for adult P. menziesii trees under ambient conditions, the reduction in biomass production by 03 has to be known. We assumed that the annual reduction in total CO2 assimilation is in the same range as the annual reduction in biomass production. Matyssek et al. however, (pers. commun.) found that a 40% reduction in total CO~ assimilation resulted in a 60% reduction in biomass.
4. EVALUATION A N D CONCLUSIONS
Despite the fact that different tree species were used, the reaction of saplings and adult trees under field conditions to high NH3 concentrations showed both similarities and discrepancies. Most of the work done on effects of NH3 on plants to date indicate that NH3 stimulates growth, but the results presented in this paper show that NH3 may also inhibit growth in the presence of low concentrations of 03 in Pinus. In Fagus however, NH3 had no effect on growth, but changed tree architecture by decreasing tree height and branch biomass. Under field conditions the high N input resulted in both increased growth and in a changed architecture. Dry weight distribution in Pseudotsuga was affected, as foliar biomass was higher and fine root and branch biomass was lower. Another similarity was the increased sensitivity. N H 3 was found to increase drought sensitivity in Pinus saplings, and both drought and frost sensitivity in adult Pseudotsuga trees under ambient conditions. NH~ is the maior part of the total N deposition in the Netherlands. The aim to reduce emissions until critical loads are not exceeded, will surely reduce NH3 levels below the critical level for NH3. The critical level proposed for NH3 is 8 and 270/,g m 3 respectively, for an annual and a 24 hrs mean !181. The annual mean is currently being exceeded in half of the Dutch areal I1|. On sites removed from the direct influence of point sources, the 24 h critical level is probably not being exceeded, not even in the Netherlands. Exceedances are frequent only on a local scale, in the first 300 meters from a point source. If environmental policy will be based on damage estimates in relation to emission and dispersion, knowledge of effect thresholds of N H3 is insufficient. Although NOx is only a small part of the total N deposition, its abatement would be very profitable because it would reduce the 03 concentration as well. High 03 concentrations resulted in decreased growth in all tree species. However, a 10% reduction in biomass production was reached at different AOT40 values (Fig. 4). It seems that the critical level for forest trees of 10 ppm.hrs is able to protect our trees from damage by 03. From the range of A OT40 values between '87-'93 it is evident that these values are exceeded in general and in the forested area of the Veluwe in particular. Timber producers may not bother about a 10% reduction in biomass by 03 if the fertilizing effect of atmospheric N deposition compensates for the losses by 03. However, this does not hold for the Dutch situation. First of all growth stimulation by N is of a temporary nature. An initial growth stimulation by N is accompanied by increasing ratio's of N to cations and following luxurious N consumption will eventually lead to a decreased growth due to severe cation deficiencies [19|. The progressing nutrient status of Speuld suggests that it will not
221 take decades to reach this situation. Furthermore, a number of forested areas in The Netherlands appears to have reached the stage of growth reduction already 1191. A O T 4 0 for 10% r e d u c t i o n of Speuld
biomass production
75
50 9-"
Douglas
9 9-,~
Beech
25
l
NL '87-'93
Scots pine
9
10
Critical Level for forest trees
0
,
I
I
I
I
,
Figure 4. AOT40 values for forest trees. The columns on the left indicate the range of 03 concentrations measured during 1987-1993 in Speuld (25m height) and as a national average (3m). The arrows indicate the AOT40 values for Pl'nus ,sylvestris, Fagus sylvatica saplings and for mature Pseudotsuga menziesii, presented in this paper in relation to the critical level proposed for forest trees.
Secondly, forests are more than trees alone. The impact of air pollution on trees (and crops) should not be seen as indicative of the unknown effects on other parts of the forest ecosystem. Effects on herbs and grasses in the undergrowth and on biodiversity of plants and animals are likely to be more pronounced 1201. The results presented in this paper suggest that all three tree species used in the experiments are adversely affected by current 03 concentrations in The Netherlands. The possibilities to quantify this damage has increased strongly in recent years, but much has yet to be done.
5. REFERENCES G.J. Heij and T. Schneider (eds.), Acidification research in the Netherlands, Elsevier, Amsterdam, 1991. T. Schneider, S.D. Lee, G.J.R. Wolters and L.D. Grant (eds.), Atmospheric Ozone Research and its Policy Implications, Elsevier, Amsterdam, (1989).
222
9 10
11 12 13 14 15 16 17 18 19 20
L.J. Van der Eerden, A.E.G. Tonneijck and J.H.M. Wijnands, Environmental Pollution, 53 (1988) 365. J. Fuhrer and B. Achermann (eds.), Critical Levels for Ozone, FAC Report no. 16, Liebefeld-Bern, 1994. Th. A. Dueck, Functional Ecology 4 (1990) 109. P.F. Scholander, H.T. Hammel, E.D. Bradstreet and E.A. Hemmingsen, Science 48 (1965) 339. E.G. Steingr/Sver and W.W.P. Jans, Dutch Priority Programme on Acidification, Report No. 793315-03, Bilthoven, 1994. F.W.T. Penning de Vries and H.H. van Laar (eds.), Simulation of plant growth and crop production, PUDOC, Wageningen, 1992 M. Pearson and T.A. Mansfield, New Phytologist 123 (1993) 351. M. Tesche and S. Feiler (eds.), Air Pollution and Interactions between Organisms in Forest Ecosystems, Proc. IUFRO-Centennial Congress, Tharandt/Dresden, 1992. L.J. Van der Eerden and M.G.F.J. P~rez-Soba, Trees-Magazine 6 (1992) 48. T.F.C. Smits and G. van Tol (eds.), IKC-NBLF Report, 1987. J.G.A. LaBastide and P.J. Faber, Stichting Bosbouwproefstation "De Dorschkamp" 11(1), 1972. A.F.M. Olsthoorn and A. Tiktak, Neth. J. Agric. Sci. 39 (1991) 61. M.A. Espinosa Bancalari and D.A. Perry, Can. J. For. Res. 17 (1987) 722. M.R. Keyes and C.C. Grier, Can. J. For. Res. 11 (1981) 599. O. Skre, Commun. of the Norwegian Forest Research Station, 40(9), 1988. L.J. Van der Eerden, Th. A. Dueck, A.C. Posthumus and A.E.G. Tonneijck, UNECE Workshop on Critical Levels, Egham, UK, 1992. J.G.M. Roelofs et al., this proceedings, 1995. R. Bobbink et al., this proceedings, 1995.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers? © 1995 Elsevier Science BV. All rights reserved.
223
EXPERIMENTAL MANIPULATIONS: FOREST ECOSYSTEM RESPONSES TO CHANGES IN WATER, NUTRIENTS AND ATMOSPHERIC LOADS
Andries W. Boxman 1, Pieter H.B. de Visser2 and Jan G.M. Roelofs ~ 1 Department of Ecology, Research Group of Environmental Biology, University of Nijmegen, P.O. Box 9010, 6500 GL Nijmegen, The Netherlands 2 Department of Soil Science and Geology, Agricultural University of Wageningen, P.O. Box 37, 6700 AA Wageningen, The Netherlands
ABSTRACT
In four Dutch coniferous forest ecosystems water and nutrient supply, as well as atmospheric loads, were manipulated for three or more years. Four approaches were used: (1) optimal supply of water and nutrients, (2) decrease of nitrogen and sulphur loads to pre-industrial levels, (3) increase of nitrogen and sulphur loads to excess levels (120 kg N ha -1 yrl). Nutrient supply was optimized according to tree demand in optimal proportions relative to ambient nitrogen supply. Tree growth was strongly enhanced (ca. 30%) by optimal water supply but not further enhanced by nutrient additions. Water additions tended to lower nitrogen concentrations in the needles by 5-10%, probably by growth dilution. Nutrient applications improved the nutritional balance in trees of phosphorus and potassium relative to nitrogen. Exchange of applied base cations with protons and aluminium from the soil, temporarily increased the acidity of the soil solution. Large applications oi nitrogen in a Scots pine stand increased the nitrogen concentration in the needles. Excess nitrogen stimulated tree growth during the first two years, and depressed growth in the fifth treatment year. Phosphorus deficiency was induced but no visible tree damage occurred. When atmospheric deposition of nitrogen and sulphur was reduced to preindustrial levels in a nitrogen and sulphur saturated Scots pine and Douglas fir stand, a few months after reduction of the input, output to the groundwater was also strongly reduced. This implies a tight input-output coupling. As a result, leaching of aluminium and base cations (to counteract nitrate and sulphate leaching) decreased and the mineral balance in the soil solution improved. Consequently, tree health improved as shown by increased root and shoot growth and by reduced totaI-N and arginine-N concentrations as well as an improved mineral balance of the needles
224
INTRODUCTION
High atmospheric input of ammonium results in loss of base cations from the soil. due to either direct exchange at the soil absorption complex or indirect via acidification, as a result of nitrification. This may result in nutrient deficiencies in trees (Roelofs eta/., 1985; Schulze, 1989). The nutritional balance of trees can be affected even further by preferential ammonium uptake to supraoptimal levels and can be disturbed by the ion competition of ammonium, aluminum and protons on the potassium, magnesium and calcium uptake. Furthermore acid soils can have Ca/AI ratios that are unfavourable for roots. The high inputs of nitrogen can result in lower root/shoot ratios and may increase the drought sensitivity of trees. This study deals with the growth and nutrition of two tree species, Scots pine
(Pinus sy/vestris L.) and Douglas fir (Pseudotsuga menziesii (Mirb.) Franco), under influence of acidification and eutrophication. The aim was to determine whether
the
nutrient supply is growth-limiting,
under the constraint of
aluminum stress and competition with ammonium and how nutrition and root functioning are related to the water economy of the tree. Therefore, the effects of a change in input of water, nutrients and acidifying substances to the soil, on trees were quantified. Atmospheric loads were increased or decreased to study their impact on nutrient cycling and forest decline. Secondly, it was hypothesed that the effects of increased soil acidification and eutrophication on trees could be mitigated by optimization of the water and nutrient supply. Experimental variation of the water supply may elucidate the impact of water stress on tree functioning. Reducing the nitrogen input to pre-industrial levels may assess the reversibility of nitrogen saturation on the different compartments of the ecosystem. These ecosystem manipulation experiments were conducted within the EXMAN and NITREX framework (Beier and Rasmussen, 1992; Dise and Wright, 1992). MATERIALS AND METHODS
Manipulations were carried out at four forest stands in the Netherlands. In a Scots pine stand near Harclerwijk and in a Douglas fir stand near Kootwijk either water and nutrients were applied, or rates of soil acidification
were
changed by exclusion of atmospheric loads or by increasing acid loads. Irrigation of demineralized water (I plot) amounted 3 to 4 mm day 1 maximally on days without rain. Fertigation
(IF plot) consisted of a complete set of
dissolved nutrients, given very frequently and in addition to irrigation during four growing seasons. The total annual application rate was equal to the
225
estimated gross uptake in trees. Phosphorus and potassium additions were quantitatively the most important and ranged from 13 (P) and 65 (K) for Scots pine to 36 and 60 kg ha 1 yr4 for Douglas fir respectively. A roof construction above the forest floor prevented the infiltration of throughfall water, being polluted with atmospheric substances, and clean rain was irrigated below in combination with a fertigation treatment (IF+R plot). Fertilization with dissolved (NH4)2SO4 amounted to 120 kg N ha 4 yr 4 (treatment N+S). In the Harderwijk stand the treatments C (control), I, IF and N+S, in the Kootwijk stand C, I, IF and IF+R were carried out. In all treatments soil water contents and composition, tree growth, needle chemistry and needle fall were monitored. (see De Visser (1994) for details on treatments and measurements). In 1989 two research sites were established in a Scots pine stand near Ysselsteyn and in a Douglas fir stand near Speuld in which ambient throughfall water was intercepted by means of a roof, and replaced by demineralized water to which all nutrients were added in the same amount as present in the throughfall, except for nitrogen and sulphur concentrations. Underneath the roof, two plots (10xl0m) were designed to receive either clean water (roof-clean plot) or ambient throughfall (roof-control plot). Outside the roof a second control plot was established, receiving ambient throughfall (control plot). A detailed description of the sites and of the methods has been given elsewhere (Dise and Wright, 1992; Van Dijk et aL, 1992a,b; Boxman et aL, 1994; Boxman et aL,1995). For statistical analyses the software packages Systat 5.0 and Statgraphics 6.0 were used. RESULTS AND DISCUSSION
Ecosystem responses to changes in water and nutrients Tree growth and nutrition in relation to water and nutrient supply Three out of four irrigated forest stands showed a water-limited growth in the examined period. An increase of 40% for Douglas fir to 50% for Scots pine in basal area growth was observed upon irrigation of 3 to 4 mm day 1 (Figure 1; p_<0.05, analysis of variance, followed by LSD test; Statgraphics 8.0). Figure 1 These data are in agreement with those found in a Norway spruce stand at Klosterhede (Denmark). In a Norway spruce stand at H6glwald (Germany) no response was observed, since the soil had a high water storage capacity (De Visser, 1994).
226
The concentrations of potassium and phosphorus in the needles were raised in all treatment years by fertigation. Needle nitrogen remained stable and consequently K/N (Figure 2) and P/N ratios increased, resulting in an improved nutritional balance in trees, with supra-optimal values (K/N > 50) in plot IF of Douglas fir. Figure 2 Fertigation did not increase total Douglas fir growth over the four-year period in addition to the growth effect of irrigation alone, and in Scots pine in one out of four treatment years only (Figure 1: 1992; p_<0.05, analysis of variance, followed by LSD test; Statgraphics 6.0). The lack of response did confirm that nutrient shortages were hardly present in the soil at the sites, although ratios of the base cations to nitrogen seemed unprofitable. Input-output budgets suggested that applied nutrients were mostly retained in the soil, although more potassium was returned with needle fall and some potassium losses were observed in the Scots pine stand. Litter fall in the Douglas fir stand decreased during the experimental period by irrigation, averaging 2.75 Mg ha 1 yr 1 (1989 to 1992), relative to 3.24 in the control. Fertigation increased aluminium concentrations in soil solution in both stands due to exchange of applied base cations with aluminium from soil. The same effect was found due to nutrient applications in a pot trial (De Visser and Keltjens, 1993) and so the treatment increased soil solution acidity, instead of ameliorating the root environment. Tree growth at different rates of soil acidification
The application of (NH4)2SO4 during five years to a Scots pine stand in Harderwijk (N+S plot) strongly increased soil acidification rates and resulted in aluminium concentrations up to 2.22 mM in soil solution from the third year onwards. Yet, diameter growth was first enhanced, probably by N-induced aboveground growth. In Figure 1 the cumulative basal area growth indicates whether a growth effect sustains during a number of years. It can be seen that after a slight initial stimulation, in the fifth treatment year growth was depressed at N+S relative to C. This coincided with phosphorus limitation, that might have resulted from AI-phosphates precipitates in soil and roots. Non of the mentioned acidifying treatments had changed the tree vitality (De Visser, 1994). At the Douglas fir site near Kootwijk, growth increased drastically at decreased rates of soil acidification (Figure 1:plot IF+R). This growth effect added up to the positive effect of water+nutrient applications (De Visser, 1994). This
227
growth increase was probably related to the decrease of NH 4 supply and uptake relative to NO 3, that resulted in slightly decreased soil solution concentrations of aluminium and protons, hence soil conditions were probably more favourable for roots. Soil solution concentrations of nitrate decreased under the roof in winter, in the summers upto 1992 nitrate concentrations rose due to small additions of fertilizer-N and due to mineralization (Figure 3).
Figure 3 Needle N concentrations decreased in irrigated and roofed plot (Table 1), whereas N uptake in biomass was estimated similar to (I) or higher than (IF+R) the control. Probably a growth dilution took place and growth was not hampered by the lower nitrogen contents.
Table 1. Nitrogen concentrations in current year, light adapted needles (% of D.W.) of Douglas fir at Kootwijk. Different letters between different treabnent indicate means which are significantly different at p~_O.05. Treatment
1989
1990
1991
1992
Control
1.86
1.99
1.74bc
1.90a
Irrigation
1.70
1.85
1.59c
1.62b
Fertigation
2.04
1.85
1.79ab
1.94a
Fe rt+Roof
1.95
1.87
1.57bc
1.70b
In a number of foreign coniferous stands, the rate of soil acidification was being increased or decreased as well. Acid irrigation (4 kmol H÷ ha -1 yr 1) on a fertile soil in HSglwald (Germany) only slightly depressed element contents in needles of Norway spruce and no change in growth was observed (De Visser, 1994).
Nitrogen saturation Fertigation, in combination with exclusion of atmospheric nitrogen and sulphur loads by a roof, decreased soil acidification slightly and increased tree growth considerably relative to fertigation at ambient nitrogen and sulphur loads. The reduction in nitrogen loads had reduced leaching losses of NO 3 to the same extent, whereas all other fertigation treatments showed that increased nitrogen inputs of approx. 30 to 40 kg N ha ~ yr ~ resulted in equally increased outputs (De Visser, 1994), suggesting a nitrogen-saturation of the ecosystems. The high growth rates in the stand at low nitrogen loads may have resulted from a higher supply ratio of NO 3 to NH4÷, that can result in less acid soil conditions
228
and can increase stem growth, as was shown in a pot trial (De Visser and Keltjens, 1993).
Ecosystem responses to reduced nitrogen and sulphur inputs Throughfa// fluxes to the forest floor Interception of throughfall in the roof-clean plot reduced atmospheric input of nitrogen and sulphur to a few kg's ha -~ yr -1 (Figure 4). During the experimental period the deposition to the ambient control plots was high (55-70 kg totaI-N ha -1 yr 1 and 35-40 kg S ha 4 yr 1 at Ysselsteyn and Speuld, respectively) and remained high, particularly in comparison with a critical load value of 15-20 kg N ha I yr4 (Bobbink eta/., 1992).
Figure 4 Atmospheric deposition to the roof-control plot was somewhat lower than to the control plot but this was attributed to the way of watering and storage of the water, and some water losses during heavy showers. Input of the other nutrients was approximately equal in all plots (Boxman eta/., 1995).
Soft solution chemistry At both locations the inorganic nitrogen and sulphur concentrations in the soil solution of the roof-clean plot responded within six months to a reduced input of nitrogen and sulphur (Figures 5 to 7). Unfortunately, no pretreatment soil solution data are available, but soil solution data (including pretreatment data) from the Soiling NITREX site (Germany) show the same rapid response (Bredemeier et al., 1995).
Figures 5,6,7,8 However, soil extracts made before the start of the treatment revealed no significant differences between the plots in availability of ammonium, nitrate or sulphur. During the first two years the way of watering determined differences between the control plots, but automation of the watering regime (early 1992) converged the data of these plots (Figure 5 to 10). NH4+ availability in the upper soil layer of control plots might have been determined by meteorological conditions. The ammonium concentration in the soil solution decreased rapidly in the very wet autumns of 1992 and 1993, which may be due to stimulated nitrification and/or leaching to deeper soil layers. The same seasonality was observed for nitrate. After an initial decrease of NO 3- concentrations in the soil solution of the roof-clean plot a tendency was observed to increasing levels (Figure 6 and also Figure 4), which is related to increased dry-deposition to the plots (data not shown). In throughfall a dominance of ammonium over
229
nitrate was found, in the soil solution the reverse situation was found probably due to nitrification, preferential uptake of ammonium by the vegetation or immobilization. As a result, the NH4/K ratio in the soil solution of the roof-clean plot decreased to levels below 5, Figure 8) considered to be favourable for a balanced nutrient uptake (Roelofs et al., 1985).
Figure 9 and 10 Nitrogen and sulphur leaving the ecosystem at 90cm have decreased significantly as a result of the treatment. Because of this, the leaching of the accompanying cations (AI, Mg, Ca and K) was significantly reduced (Boxman et al., 1995). This means that the mineral balance in the roof-clean plots improved. Nutrients in the needles Before the start of the manipulation experiment no significant differences were
found in nutrient concentrations of the needles at both locations. At Ysselsteyn the needles had high nitrogen concentrations, whereas the other nutrient concentrations were very low. During the years of treatment the concentrations of nutrients can vary considerably, which may be related to meteorological conditions (Boxman et al., 1995 and references therein). As yet, the needles in the roof-clean plot have responded to the treatment after a lag-time of approximately four years (Figure 11). The nitrogen concentration is still above 2%, which is considered as very high (optimal level is approximately 1.4-1.8%, Anonymous, 1990). The potassium and magnesium concentrations have increased significantly in the needles of the roof-clean plot as compared to both control plots (data not shown) and may be related to 1) a favourable NH4+/NO3- ratio in the soil solution, 2) a decreased leaching of base cations and 3) an increased root biomass (Boxman et aL, 1995). Consequently, the nutritional balance in the needles of the roof-clean plot has improved for potassium (Figure 12) and magnesium (data not shown) relative to nitrogen, the former even to a level above that is considered deficient (K:N>25, Anonymous, 1990).
Figures 11, 12, 13 and 14 Although the ecosystem at Speuld is also nitrogen-saturated the trees have normal nitrogen concentrations in their needles. In the control plot, however, nitrogen tends to increase (Figure 13). Since the trees are growing reasonably well, dilution effects in the needles may prevent nitrogen to become toxic. The nitrogen saturation effect is most pronounced in the older needles, which have higher nitrogen concentrations than the current ones (data not shown).
230
Potassium, magnesium and calcium are sufficient in the needles, while phosphorus is somewhat low. No significant differences were found between the plots. As a result the nutritional balance of potassium (data not shown) and magnesium (Figure 11) relative to nitrogen have improved, although ratios are above the levels that are considered deficient (25 and 5, respectively; Anonymous, 1990). Nitrogen, taken up by the trees is incorporated into amino acids, and subsequently into proteins. The assimilation of ammonium is absolutely necessary as free ammonium is toxic because of its interference with many processes in the cell (see e.g. Puritch and Barker, 1967; Wakiuchi et aL, 1970; Van der Eerden, 1982). If the rates of nitrogen uptake and subsequent amino acid synthesis exceed that of protein synthesis, free amino acids accumulate. Upon changes in nitrogen supply these amino acids show more sensitive changes in concentrations than the total nitrogen content. In coniferous trees arginine is most important in this respect, because of the low C/N ratio. When the nitrogen concentrations in the needles increased, arginine seemed to accumulate at nitrogen concentrations above 1.5 to 1.6% (Figure 15). Figure 15 Irrespective whether totaI-N in the needles was still high, the treatment significantly decreased the arginine concentration in the needles at Ysselsteyn, while the same trend was observed at Speuld. The response of arginine-N was fast: within one year changes could be observed. In this respect arginine can be regarded as a good indicator of detrimentally high ammonium deposition (Van Dijk and Roelofs, 1988; Ferm et a/. (1990); N&sholm and Ericsson, 1990; Pietil& eta/., 1991). Figure 16 and 17 Diameter growth of the dominant trees in the roof-clean plot improved (p=0.06 for both roofed plots) and was inversely related to the arginine-N concentration in the needles (Figure 18). This is in agreement with the observation of Krau~ et aL (1986) who found a clear reduction in growth in relation to increasing arginine concentrations. Figure 18 With respect to concentrations of arginine-N reported in the needles from pristine areas in northern Scandinavia (<0.06%) concentrations in the needles of Speuld were high (0.1-0.3%), but lower than at Ysselsteyn (0.65-0.9%). At the latter site this implies that almost 30% of totaI-N was stored in the form of arginine-N.
231
Fine root growth A survey in 1992 revealed a significantly increased fine root biomass and number of fine root tips in the roof-clean plot at Ysselsteyn (Boxman et aL,
1995). Data from litter-bag experiments of the Vrije Universiteit of Amsterdam only confirmed these data. After one year more fine roots were grown into the litter-bags, contained more root tips and had a higher degree of mycorrhizal infection (data not shown). These data suggest an enhanced uptake capacity of the roots, which is in accordance with the improved nutritional balance in the needles. Final conclusions Decreasing the input of nitrogen and sulphur strongly reduced the output of nitrate, sulphate, aluminium and base cations from both ecosystems, implying a tight input-output coupling. The N-cycle changed from an open to a more closed one, indicating reversibility of nitrogen-saturation. However, the soil still contains a large amount of immobilized nitrogen and at this moment it is uncertain what will happen with this amount in the future. Both ecosystems are recovering from excess nitrogen availability, indicated by 1) a decreased leaching of base cations, 2) a more favourable NH4÷/NO3 ratio in the soil solution, 3) increased root and 4) shoot growth, 5) a decrease in totaI-N and arginine-N in the needles and 6) an enhanced nutrient uptake and an improved nutritional balance in the needles. REFERENCES Aber, J.D., Nadelhoffer, K.J., Steudler, P. and Melillo, J.M., 1989. Nitrogen saturation in northern forest ecosystems. Bioscience 39, 378-386. Anonymous, 1990. Final Report Commission Advice Forest fertilization. Report 1990-11. Ministry of Agriculture, Nature Conservation and Fishery. pp. 63. (in Dutch). Beier, C., Rasmussen, L., De Visser, P., Kreutzer, K., Schierl, R., Zuleger, M., Steinberg, N., Bredemeier, M., Farrell, E.P., Collins, J., and Cummins, T. 1993. Effects of changing the atmospheric input to forest ecosystems Results of the "EXMAN" project. In: Experimental manipulations of biota and biogeochemical cycling in ecosystems - Approach - Methodologies Findings. (Eds. L. Rasmussen, T. Brydges and P. Mathy). Commision of the European Communities, Ecosystem Reseach Report 4. pp.138-154.
232
Bredemeier, M., Blanck, K., Lamersdorf, N. and Wiedey, G.A., 1995. Response of soil water chemistry to experimental "clean rain" in the NITREX roof experiment at Soiling, Germany. For. Ecol. Manage., in press. Bobbink, R., Boxman, D., Fremstad, E., Heil, G., Houdijk, A. and Roelofs, J. 1992. Critical loads for nitrogen eutrofication of terrestrial and wetland ecosystems based upon changes in vegetation and fauna. In: Critical loads for nitrogen. Eds.: Grennfelt, P. and Th6rnel6f, E. pp. 111-159. The Nordic Council, Nord 1992: 41. ISBN 92 9120 121 9. Boxman, A.W., Van Dam, D., Van Dijk, H.F.G., Hogervorst, R.F. and Koopmans, C.J., 1995. Ecosystem responses to reduced nitrogen and sulphur inputs into two coniferous forest stands in the Netherlands. Forest Ecol. Manage., in press. De Visser, P.H.B. and Keltjens, W.G., 1993. Growth and nutrient uptake of Douglas fir seedlings at different rates of ammonium supply, with or without additional nitrate and other nutrients. Neth. J. Agric. Sci., 41,327-341. De Visser, P.H.B., 1994. Growth and nutrition of Douglas fir, Scots pine and pedunculate oak in relation to soil acidification. Ph.D. thesis, Wageningen Agricultural University, The Netherlands. Dise, N.B and Wright, R.F., 1992. The NITREX project (Nitrogen saturation experiments). Commision of the European Communities, Ecosystem Reseach Report 2. ISBN 2-87263-077-5. Encke, B.G. 1986. Stikstoff und Waldsterben. AIIg. Forstzeitschr. 37, 922-923. Ferm, A., Hyt6nen, J., L~.hdesm&ki, P., Pietil~.inen, P. and P&til&, A., 1990. Effects of high nitrogen deposition on forests: case studies close to fur animal farms. In: Kauppi et aL (eds.), Acidification in Finland. SpringerVerlag, Berlin. pp. 635-668. Kraul3, H.H., Heinsdorf, D., Hippeli, P. and T611e, H., 1986. Untersuchungen zu Ern~hrung und Wachstum wirtschaftlich wichtiger Nadelbaumarten im Tiefland der DDR. Beitr. Forstwirtsch. 20: 156-164. N&sholm, T. and Ericsson, A., 1990. Seasonal changes in amino acids, protein and total nitrogen in needles of fertilized Scots pine trees. Tree Physiol. 6, 267-281. Pietil&, M., L&hdesm&ki, P., Pietil&inen, P., Ferm, A., Hyt6nen, J. and P&til&, A., 1991. High nitrogen deposition causes changes in amino acid concentrations and protein spectra in needles of the Scots pine (Pinus sylvestris). Environ. Pollut. 72, 103-115. Puritch, G.S. and Barker, A.V., 1967. Structure and function of tomato leaf chloroplasts during ammonium toxicity. Plant Physiol. 42, 1229-1238.
233
Roelofs, J.G.M., Kempers, A.J., Houdijk, A.L.F.M. and Jansen, J., 1985. The effect of air-borne ammonium sulphate on Pinus nigra var. maritima in the Netherlands. Plant Soil, 84, 45-56. Schulze, E.-D., 1989. Air pollution and forest decline in a spruce (Picea abies) forest. Science 244, 776-783. Smits, T.F.C., 1992. The vitality of the Dutch forests 10. Report of the national survey 1992. Report IKC-NBLF 1992-8. Ministry of Agriculture, Nature Conservation and Fishery. pp. 39. (in Dutch). Wakiuchi, N., Matsumoto, H. and Takahashi, E., 1970. Changes of some enzyme activities of cucumber during ammonium toxicity. Physiol. Plant. 24, 248-253. Van der Eerden, L., 1982. Toxicity of ammonia to plants. Agric. Environ. 7, 223-235. Van Dijk, H.F.G. and Roelofs, J.G.M., 1988. Effects of excessive ammonium deposition on the nutritional status and condition of pine needles. Physiol. Plant. 73, 494-501. Van Dijk, H.F.G., Boxman, A.W. and Roelofs, J.G.M., 1992a. Effects of a decrease in atmospheric deposition of nitrogen and sulphur on the mineral balance and vitality of a Scots pine and a Douglas stand in the Netherlands, Interim Project report: 1988-1991. Department of Ecology, Section Environmental Biology, University of Nijmegen, The Netherlands. pp. 1-43. Van Dijk, H.F.G., Boxman, A.W. and Roelofs, J.G.M., 1992b. Effects of a decrease in atmospheric deposition of nitrogen and sulphur on the mineral balance and vitality of a Scots pine stand in the Netherlands, For. Ecol. Manage., 51,207-215.
234
Cumulative basal area growth
Scots pine 10
s
N+S
1
6
4
2
0
1987
1988
1989
1990
1991
1992
1993
1994
18
Douglas fir
16
IF+R
14
IF
12 10
C 8 6 4 2 0
1987
1988
1989
1990
1991
1992
1993
1994
Figure 1. Cumulative basal area growth (m 2 ha 1 yr 1) in the Scots pine stand at Harderwijk and the Douglas fir stand at Kootwijk. The treatments started in 1989.
235
50 :
[ Scots pine IF
- 45
'~40
~
I C
35
o
°v..~ 4.a
30
25
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
-"-1 N + S
deficient 20
1987
1988
1989
1990
1991
1992
1993
55
Douglas fir 50
...................
opt_im__a!............~
IF .... IF+R
. ~ 45
5 .--q 40 .e.a
35
30
1987
1988
1989
1990
1991
1992
1993
Figure 2. Elemental ratio of K to N (in %) in the Scots pine stand at Harderwijk and the Douglas fir stand at Kootwijk. For both tree species a ratio above 50% is judged optimal and a ratio below 25% insufficient for tree growth (Anonymous,
236
NO3 at 20 cm depth
5
"6 E E ¢.-
4
tO
3 tO
2
Dec-88
Jul-89 •
Jan-90 Aug-90 Mar-91 Sep-91 Apr-92 C
'
,~
IF
I
~
Oct-92
IF+R
Figure 3. Nitrate soil solution concentrations at 20 cm depth in four treatment plots at Kootwijk.
70 60
"7 ~ 4o O9
~ 30
g ~20 10 89
91 90
93
89
92
91 90
93 92
89
93
91 90
92
year
roo~clean
roo~control []
TotaI-N [ ]
control
TotaI-S I
Figure 4. Deposition fluxes of total-N and total-S to the forest floor of the plots at Ysselsteyn.
237
8OO 700
.
, .......... ............ ...............
600
~,500
i", '
--
,
/ ,
400
Z 300
.
-" . . . . .
/' ~
-~, ,\
/
,
/
, \ I
/\ i
. I
~li
/
200
,
i
,,
/'
,
' ,
L
~
,
\-~,'
100
'~
o
. ,
01101190
01101191
I--
01/01192 Date
roof-clean
Figure 5. A m m o n i u m
01101193
01101194
. . . . roof-control ...... control
l
c o n c e n t r a t i o n s at l O c m depth in t h e soil solution at Y s s e l s t e y n .
1800 1600 1400 1200
.
~
~" lOOO 0 z
J
'
~ /' '~ ~
,.
r
80O 600
f
~
'
-"\ \ i
~ ,f"
400 200 01101190
01101191
I --
roof-clean
01101192 Date
01/01/93
. . . . roof-control . . . . control
01/01/94
I
Figure 6. Nitrate c o n c e n t r a t i o n s at l O c m depth in the soil solution at Y s s e l s t e y n .
238
700
j'~'\
600 ", ,L/// //:"
500 /I
~'400
~\\
¢
-x ~.,,° ..
\\
. ,
I
,-i
'_
¢n 300 20O 100 0 01/01/90
01/01/91
I--
roof-clean
01/01/92 Date
01/01/93
. . . . roof-control
01101194
...... control
I
Figure 7. TotaI-S concentrations at 10cm depth in the soil solution at Ysselsteyn.
20 18 16 ~ 1 4
0
Is
•
\
E 12 O
~1o ~
8
Z
6
T
4 2 0
01/01/90
i
h
01101191
I--
01101192 Date
L
01/01/93
roof-clean - - - roof-control ...... control
I
01101194
I
Figure 8. NHJK ratio at 10cm depth in the soil solution at Ysselsteyn.
239
3000 2500 2000 1500 O z 1000 5OO 0 01101190
q
i
01101191
--
~
01/01/92 Date
roof-clean
01/01/93
t
01101194
. . . . roof-control ........ control
I
Figure 9. Nitrate concentrations at 90cm depth in the soil solution at Ysselsteyn.
800 ,k,
-
700 -
,
k.
6O0 r'\
~,5oo
] i'
/ ,\ \ - - - - / /
o~ 400 300 200 100 01101190
01/01/91
I--
roof-clean
01/01/92 Date . . . . roof-control
01/01/93
. . . . . . . . .
control
01101194
J
Figure 10. TotaI-S concentrations at 90cm depth in the soil solution at Ysselsteyn.
240
3.1
2.9 2.7
/
/
//
/
\\
\
\ ./4
~'2.5 o~ z2.3 2.1 1.9
1.7
I
r
1989
1990
I
I
1991
1992
1993
1994
Year Im
roof-clean
-=~ roof-control ~ - control
I
Figure 11. Nitrogen concentration in the V2-year-old needles at Ysselsteyn. *: roofclean plot is significantly different from both control plots at p_<0.05. Systat 5.0: ANOVA correlation.
29 ~RB-~j
\\
27
25
iS
\/
Z
o~ 23
~21 Z
~d
19 17 15
\ ~ 1990
i 1991
J-m- roof-clean
i 1992 Year
t 1993
--~ roof-control -e= control
O I
1994
I
Figure 12. K:N balance in the V2-year-old needles at Ysselsteyn. *: roof-clean plot is significantly different from both control plots at p<0.05. Systat 5.0: ANOVA correlation.
241 2.2-
~'1.8 C]
z 1.6 1.4 1.2
' 1989
~ 1990
~ 1991
"t 1992
', 1993
I 1994
Year
~-=- roof-control-=- roof-clean
-~-.,Control
Figure 13. Nitrogen concentrations in the
89
l needles at Speuld. Plots are not
significally different. Systat 5.0: ANOVA correlation.
11 10 o Z
9 8
O)
Z
6
4
4
.
1991
.
.
.
t
-'
I
1992
-
1993
'
,
1994
Year
~-m- roof-control ~
roof-~'ean -~- .c,ontiol' -
J
Figure 14. Mg:N balance in the 89 needles significally different. Systat 5.0: ANOVA correlation.
at Speuld.
Plots
are
not
242
1.4
Jp
[
,
.
I
0.8
8 o.6 Z O.4 "2
~, < 0.2 m
0
I
1.6
t
i
1.8
I
I
2
t
I
t
i
I
t
2.2 2.4 2.6 N concentration [%DW]
I
i
I
2.8
3
3.2
Figure 15. Correlation between the totaI-N and arginine-N concentration in the V2year-old needles (collected January 1994) at Ysselsteyn. Systat 5.0: Spearman correlation.
n.s.
.
.
**
**
1990
1991
1992 Year
1993
1994
0.9
~,0.8 c~0. 7 o~ Z0.6 (D C
i~_ 0.5 < 0.4 0.3 0.2 i 1989
I -m- roof-clean
-e- roof-control - * - control
'1
,,,
Figure 16. Free arginine-N concentrations in the V2-year-old needles at Ysselsteyn. Roof-clean plot is significantly different from both control plots at *' p<0.05, **: p<0.01, n.s" not significant. Systat 5.0: ANOVA correlation.
243
0.35 0.3 ~' 0.25 D 0.2
z $ ,'0.15 . J
C-
L-
<
0.10.05
m m
="=
|
"
1989
I
'
!
'"
1990
i
,
1991
i
r
1992 Year
i -~-roof-clean
I
r
I
1993
1994
-~- roof-control-A- control
!
Figure 17. Free arginine-N concentrations in the V2-year-old needles at Speuld. Plots are not significally different. Systat 5.0: ANOVA correlation.
14 r=-0.85 p<0.001
E 12 '7 o~ 10 t.-
o= 8
A
o}
E 6 t~ a ,,=..
O ,
'T
0.2
,
i
0.4
I
i
0.6
~
i
0.8
:
i
1
,
i
1.2
w
1.4
Arginine-.N [% DW] Figure 18. Correlation between arginine-N concentration in the Y2-year-old-needles (collected in January 1994) and diameter growth of five dominant trees at Ysselsteyn. m: roof-clean; z~: roof-control" 0 : control. Systat5.0: Spearman correlation.
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
245
Effects of acid deposition on forest e c o s y s t e m s in T h e N e t h e r l a n d s : A n a l y s i s o f t h e S p e u l d D o u g l a s fir s i t e Hans van Grinsven a, Bert-Jan Groenenberg ~, Kees van Heerden a, Hans Kros c, Frits Mohren b, Carolien van der Salm c, Evelien SteingrSver b, Aaldrik Tiktak ~ and Jan-Renger van de Veen b aNational Institute of Public Health and Environmental Protection, P.O. Box 1, 3720 BA Bilthoven, The Netherlands bInstitute for Forestry and Nature Research, P.O. Box 23, 6700 AA Wageningen, The Netherlands ~rhe Winand Staring Centre for Integrated land, Soil and Water Research, P.O. Box 125, 6700 AC Wageningen, The Netherlands Abstract Large efforts have been dedicated to investigate effects of acid atmospheric deposition on trees and soil at the forest stand level. For this purpose intensive monitoring programs and integrated simulation models were developed. This paper describes the application of the models NuCSAM, SoilVeg and ForGro to the Speuld site, a Douglas fir stand on a podzolic soil. The site was monitored between 1987 and 1992. Atmospheric inputs and element outputs at Speuld are fairly representative for Dutch forests. The models were parameterized and calibrated for this site. Simulated soil water contents, soil solution chemistry, foliage biomass and nutrient status and stem growth between 1987 and 1992 were comparable with observations. The generality of the SoilVeg and ForGro model was further tested by an approximate simulation of a site irrigation and fertigation experiment at a nearby Douglas fir site between 1987 and 1991. The direction and magnitude of simulated effects of irrigation and fertigation on stem growth, litter fall and needle nutrient status were generally right, but the observed enhanced N-mineralization could not be simulated. Simulations of site response to the Dutch target deposition scenario between 1994 and 2050 showed large differences between the three models, particularly for nitrogen cycling and foliage nutrient status. Nonetheless all models indicate a fast response of soil solution chemistry to changing deposition. Both SoilVeg and ForGro indicate that direct effects of air pollution and effects of soil pH and A1 are subsidiary to effects of drought and nitrogen. Our understanding of effects of acid atmospheric deposition on forests, based on lab trials, relatively short site monitoring studies and integrated simulation models, is still inadequate to quantitatively predict the long-term impact of acid deposition on forests on a nationwide scale.
246 1. F O R E S T SITE R E S E A R C H IN A P E R S P E C T I V E Small scale occurrence of forest dieback in central Europe is evident (Kandler, 1992). Proof for ongoing large scale forest dieback due to acid atmospheric deposition is still not found. Concern for adverse effects of acid atmospheric deposition in the Netherlands was triggered by the observation in the early eighties that tree vitality, expressed by leaf occupancy and discolouring, was poor throughout the Netherlands, and the observation that ambient concentrations of S02, 03, NO,, and NH 3 exceeded no-effect levels. At the some time evidence was found in the field that forests in the Netherlands received high atmospheric inputs of sulphur and nitrogen compounds leading to elevated concentrations and accelerated cycling of SO,, N and A1 in underlying soils (Van Breemen et al., 1982, De Vries et al., 1994). These simultaneous observations led to several hypotheses linking forest growth and forest vitality to air pollution, atmospheric deposition, soil acidification and disturbed nutrient cycling. Examples are the Al-toxicity hypothesis (IYlrich, 1989) and the nitrogen saturation hypothesis (Gundersen, 1992). A major problem when proving these hypotheses was the large difference between scales of time and space at which various effect observations were done, ranging from pot trials to nation wide surveys and from hours to decades. Another specific problem for field trials was the separation of air pollution and acid deposition related response from natural variation of tree growth and nutrient status. This was one of the motivations to start intensive integrated field studies in forest stands, in which weather, air quality, atmospheric deposition, soil hydrology, soil chemistry, nutrient cycling and forest growth were monitored over a number of years. These integrated monitoring program8 were started in the US (Hubbard Brook), Norway (Birkeness, Sogndal), Sweden (Gardsjon, Skogaby), Germany (Soiling, H~glwald), Denmark (Klosterhede) and the Netherlands (Speuld, Kootwijk). An important tool to analyse and integrate the observations at the forest stand level was the budgeting approach (De Vries et al., 1994; Posma et al., 1994), in which the fate of carbon, nutrients or protons was determined by quantifying all relevant inputs, outputs and stores. However, the budgeting approach was not suitable to test hypothetical effect mechanisms and to predict the response of the forest stand to changes of atmospheric inputs. This was one of the motivations to develop mechanistic and comprehensive simulation models of a forest stand. A second motivation was the need to generalize and quantify the effect relationships for air pollution and atmospheric deposition, to be able to evaluate the effectiveness of national environmental policies for the Dutch forests. The integrated models describe the water, carbon and nutrient status of the forest and forest soil as a function of weather, atmospheric deposition and air quality, given the inherent properties of the soil and the tree species, and taking into account the relevant transformation and transport processes in tree and soil. These models have been fairly successful to explain the water and chemistry status of the forest ecosystem (Landsberg 1991; Van Grinsven et al., 1995). Efforts to develop fully integrated models, that include forest hydrology and soil chemistry and forest growth, and apply these models to comprehensive
247 site datasets, are r at her scarce. It was soon clear t h a t at the integrated monitoring sites boundary conditions changed slowly, and the variation in the response of tree and soil were sm~ll or non-existent, hempering a m e a n i n ~ test of any integrated stand effect model. This knowledge was a major motivation to start experimental field manipulation, very oiten at the location of the intensive monitoring sites (De Visser, 1994). By artificially and drastically changing the inputs of water and nutrients to the stand, the soil water, nutrient and soft chemical status were forced to change, thus providing information about the rate and magnitude of response of the ecosystem. Data from these experiments are presently coming available and appear to be very valuable for testing models and effect hypotheses. However, as the duration of the experiments seldom exceeds three years, there is a danger t hat transient artifacts initially dominate the response of the ecosystem. This paper describes the application of the soil acidification model NuCSAM and two integrated models SoilVeg and ForGro to a Douglas fir stand at Speuld in the Netherlands. First the Speuld monitoring progr~_m is described, followed by a discussion of the modelling philosophy and model principles. Next, model calibration to Speuld, results of application to a nearby manipulation experiment and analysis of one deposition scenario are described. Finally, conclusions are presented with respect to monitoring and modelling at the forest stand level.
2. THE S P E U L D E X P E R I M E N T A L F O R E S T Since 1980 a total number of 17 forest sites has been monitored intensively (Van Breemen and Verstraten, 1991) and 150 sites have been studied extensively (De Vries et al., 1994). In intensive monitoring studies, boundary fluxes and states for hydrology, soil chemistry and sometimes tree structure and carbon and nutrient status have been measured at various times within a year, in general over a period of more than three years. The extensive survey is characterized by a one time assessment of a smAUer selection of soil chemical and tree nutrient states. The most complete intensive monitoring studies, with a relative emphasis on pollution climate and on tree carbon and nutrient status, were carried out on two Douglas fir stands at Speuld and Kootwijk (Evers et al., 1987). The Speuld Douglas fir site is located at 52.1 ON, 5.4 ~ 50 m above sea level. Annual average rainfall is 808 ram, potential evapotranspiration is 534 ram, and temperature 9.3 ~ Average air temperature in J a n u a r y is 2 ~ in July 17 ~ The age of the Douglas fir in 1990 was 32 years, stand density 812 ha ~, mean tree height 20.4 m and basal area 35.2 m ~ ha ~. The underlying soil is a Cambic podzol developed on sandy losm. The saturated volumetric water fraction ranges from 0.21 to 0.33, the saturated conductivity ranges between 1 to 8 m/d. Soil pH-H20 ranges from 3.7 to 4.3, CEC from 28 to 105 mmolJkg, Base Saturation is below detection limit. The organic matter content ranges
248 from 0.2 to 2.4% (w/w), the free &l-oxide content from 160-490 mmolJkg. The total mineral CaO content is about 30, the MgO content ranges from 30-60, the K20 content is about 200 mmolJkg. The air pollution climate can be characterized by an ambient SO~ concentration of 10 pg m "3 (1990), ambient O3 concentration of 40 ~ g m "3, an annual N-deposition of 3.4 k m o l h a "~ ( 2 . 7 N H 4, 0.7 NO3) and S-deposition of 1.1 kmol ha "1. Monitoring at Speuld was carried out between 1986 and 1993, however, hydrological and soil chemical monitoring was restricted to the period between 1987 and 1990 (Table 1). Table 1 Characterization of major monitoring activities at Speuld. 1.
hourly
2.
hourly
3.
hourly
4. 5. 6. 7.
bi-weekly bi-weekly bi-weekly daily
8.
daily
9. 10 11. 12. 13.
bi-weekly monthly monthly annually variable
Meteorology (rainfall, temperature, radiation, vapour pressure, wind speed, sunshine duration) Ambient concentrations of SO2, NH3, 03 and NO2 (average and/or peak values) Dry atmospheric deposition fluxes (SO 2, total S, NH 3, NOffi, total N, base cation aerosols, total C1) Wet deposition of all major compounds Bulk deposition of all major compounds Throughfall of all major compounds Transpiration, litter evaporation and net CO2 flux (only in 1989) Soil water potential, water content and temperature at various depths Soil solution concentration at various depths Weight, length and nutrient status of roots DBH and stem basal area Volumetric growth, ratio of sap and heart wood Tree height, height of lowest living whorl, biomass of stem, branches and foliage (per needle year class), nutrient content (C, N, P, S and base cations) of wood and foliage, stem and branch density, LAI, SLA, sugar, starch, cellulose and lignin content of foliage, litterfall and nutrient status of litter, vitality classification
Speuld has a high foliage mass (18-23 t o n g a between 1987 and 1992) and a very high ratio of foliage to fine root mass (6-7). Further Speuld is deficient for phosphorus in foliage (0.08-0.12% w/w) and high for nitrogen in foliage (1.5-2.0 % w/w). Speuld is classified as a fairly vital and well growing stand producing over 5 tons of stem wood per year. Element inputs and outputs at Speuld are very comparable to those of other intensively monitored stands (Table 2; Van Breemen and Verstraten, 1991), but S-inputs and N-outputs, and by result also S and Al-outputs, are high compared
249 to fluxes inferred from a nation wide extensive forest survey (n=147) in 1990 (De Vries et al., 1994). The discrepancy between the S-budget of the intensive and extensive sites can be attributed to a substantial decrease of S-deposition in the past decade (Erisman et al, 1993). The discrepancy for N can be explained in p a r t by the relative absence of wet oak stands with a high potential for Nimmobilization in the set of intensively monitored sites. Table 2 Comparison of element budgets for Speuld with intensively and extensively monitored forest sites
those
for other
Dutch
Element flux or ratio kmol c ha x a x
Speuld (n=l) 1987-1990
Intensive (n=17) 1980-1990
Extensive (n=147) 1990
S-input N-input S-output NO3-output NH4-output Al-output S-out/S-in N-out/N-in Al-ouU(N+S)-out
2.3 3.6 3.1 2.2 0.0 4.6 1.3 0.6 0.9
2.8 4.0 3.0 3.0 0.3 4.6 1.1 0.8 0.9
1.8 4.2 1.9 0.8 0.0 1.0 1.1 0.2 0.4
A major problem for integrated model analysis of the Speuld data set was the large spatial variability of throughfall - m o u n t and chemistry, soil hydrology (Bouten et al., 1992) and soil solution chemistry. An example is given for SO 4 concentration at 60 cm depth (Figure 1), were both m e a n values and trends are incomparable for three plots at short distance. In view of the 1987 1988 1989 number of replicates as Figure 1. Concentration of SO4 at 60 small cm depth for three adjacent plots under compared to this large variability it had to be decided to use either a a Douglas fir stand at Speulderbos. unreliable site mean value (n=6) of observations for model validation or to use unrepresentative mean values of observations (n=2) for hydrological and soil chemical variables taken in or near one soil pit. We chose for the latter to m~intain reliable temporal dynamics in the soil solution data. SO42" (tool c m "3) (6o on. ck~h)
250 3.
THE MODELS
In the Dutch Priority Progrsm on Acidification (Heij and Schneider, 1993) considerable emphasis was put on the development and application of the dynamic process-oriented simulation models N u C S A M , a site version of the model R e S A M (de Vries et al., 1994), SoilVeg (Berdowski et al., 1991; V a n Heerden et al., 1995) and ForGro (Mohren et al., 1993). All three models aim at simulating the behaviour of a forest stand over one rotation as a function of climate, air pollution and atmospheric deposition at a near to daily resolution. R e S A M and SoilVeg were also applied at a nationwide scale and an annual resolution (Heij and Schneider, 1991). The major characteristics of the three applied models are summarized in Table 3. The three models consider all major chemical species and plant nutrients, H, AI, SO4, Na, C1, N H o NOs, K, Ca, M g and PO4 (not in SoiIVeg). ForGro uses N u C S A M as soil chemistry submodel. The models are particularly useful for testing the hypotheses that episodes with high Al, low p H and low water content, typically occurring in summer, or episodes with high ,mbient concentrations of 08, SO~ or N H 3 can cause a longterm decrease of foliage and root mass. The models N u C S A M , SoiIVeg and ForGro were already applied to the Soiling site (Van Grinsven et al., 1995) together with 13 other models of various complexity and completeness with respect to the description of the water, nutrient and carbon cycles. Some important conclusions from this workshop were: differences between the models were largest for the calculation of nutrient demand and uptake all models reproduced the average observed water, nutrient and carbon status of the Soiling spruce stand fairly well, but seasonal dynamics and year to year variation poorly none of the models could explain the observed NOs mobilization and leaching the input demand and complexity of the models varied widely, but the perform-nce of models did not systematically become better with increasing complexity, at a c o m m o n level of output aggregation The application to the Speuld Douglas fir stand is described in more detail by Tiktak et al. (1995). Both for the Soiling and the Speuld application the three models followed similar guidelines for evaluation of model performance, based on Janssen and Heuberger (1995). -
-
-
4.
M O D E L P A R A M E T E R I Z A T I O N AND C A L I B R A T I O N
Although calibration procedures for the three models were different and rather unstructured, a logical sequence for calibration was adapted which was guided by the strength of the interactions between hydrology, soil chemistry and tree processes. Hydrology strongly affects soil solution chemistry but not vice versa. The effects of soil hydrology and soil solution chemistry on nutrient uptake and tree growth are stronger than vice versa.
251 Table 3 Major principles of the forest stand level models NuCSAM, SoilVeg and ForGro. ,,,
,,
,,
NuCSAM
SoilVeg
ForGro
hydrology
Solution of water potential in Richard's and D a r e r s equation. Potential transpiration from Makkink and an empirical crop factor
Solution of water content from empirical equations with soft water flux and water uptake. Potential transpiration as in NuCSAM
Solution of water content from empirical relation with water flux. Transpiration from Penman-Monteith
soil chemistry
Gaines-Thomas cation exchange. 1~ order nitrification and silicate weathering. Elovich Al~xide weathering
Equations and parameterization as in NuCSAM. PO, is not considered
ForGro uses NuCSAM
Forced by stem growth, biomass turnover and fixed nutrient contents
Driven by water uptake, nutrient concentration in soil solution and selectivity coefficients
Driven by demand and limited by radial diffusion from the bulk soil to the root
forest growth
Logistic stem growth and fixed biomass ratios
Grmm photosynthesis is forcing function. Adapted, multiple nutrient productivity concept
Photosynthesis is
tree effect relations
no effect model. Empirical relation between N-deposition and N-content of foliage
empirical reduction of photosynthesis with ambient SO: and 0,. Al and pH effecta on root uptake and root growth. Direct canopy uptake of ambient NH:. Increased respiration plant N status
nutrient uptake
as submodel
driven by light interception. Gross carbon assimilation is summed per leaf layer Nutrient shortage, and stomatal uptake of SOs and O, in foliage will reduce photceynthesis. AI effect on root growth and nutrient uptake. Direct canopy uptake of ambient NH r Increased respiration plant N status
Parometerization of conductivity and percolation functions, and water uptake distribution with depth is fairly straightforward for all three models. The hydrology modules were validated against a common data set of water contents in 1989 and against C1 and SO4 data. Due to the low sorption capacities of the soils at Speuld, C1 and SO, are practically conservative tracers. The derivation of exchange coefficients, silicate and oxide weathering rates, mineralization and nitrification rate constants and sulphate and phosphate sorption isotherms could be harmonized, in view of the similarity of the soil chemistry submodels. Parameterization of the nutrient uptake and growth was model specific and posed most problems. Calibration of these submodels for NuCSAM was simple as growth is a forcing function. Calibration needs for SoilVeg probably were largest, mainly because this model uses empirical multi-parameter S-shaped age dependent forcing functions for gross nutrient uptake and allocation of
252 carbon and nutrients. Adjustment of the default parameterization of the growth module in ForGro was kept to a minimum. However, allocation of assimilates to foliage had to be increased to simulate foliage masses in accordance with observations at Speuld. An example of the complexity of the calibration was the parameterization of the carbon turnover and allocation process using litterfall data. Initial estimates of annual litterfall from litter traps (1 m2), litter nets (45 m 2) and inferred from the standing needle mass ranged from 2.2 tot 6.8 Mg ha "1 a 1. Taking into account spatial variability of biomass in the stand and aider matching litterfall to standing foliage mass a most probable annual litterfall of about 5 Mg ha 1 a 1 was used. This example illustrates t h a t biomass and nutrient data for Speuld in fact did not allow any fine-tuning or validation of the carbon turnover and allocation parameters. The parameters for effect submodels were inferred from fumigation and hydroculture experiments. Model performance was judged by comparing model output with observations at a rather aggregated level: viz. water contents, soil solution concentrations, stem and foliage mass and the nutrient status of foliage. For soil hydrology and soil solution chemistry objective performance criteria were used. Differences between hydrology submodels are largest in the growing season. In this period observations show a response of water content to individual rain events at depths up to one meter, while none of the models showed such a response. This discrepancy could be an indication of short circuit flow. Simulated annual hydrological fluxes showed considerable differences; eg. in 1988 transpiration ranged from 301 to 323 mm, interception from 303 to 456 mm, soil evaporation from 19 to 79 mm and drainage from 17 to 255 mm. Validation of the hydrological calibration with C1 and SO4 was unsuccessful. As compared to experiences from model application to Solling (Grinsven et al., 1995), model performance for C1 and SO4 is rather poor (Figure 2). Correct simulation of pH and Ca generally failed. Simulated concentration increases for A1 and NOs in the growing season are larger t han the observed increase (Figure 2). This discrepancy can only partly be attributed to the impossibility to use suction cups for soil solution sampling when the soil water suction exceeds the air entry values of the cup lysimeters (-- 800 mbar). The moderate success of the soil chemistry calibration is also illustrated by values (Table 4) of Normalized mean absolute errors (NMAE) and Normalized Mean Errors (NME) (Janssen and Heuberger, 1995). The NMAE and NME are zero when model results and observations are identical. Actual values lie around 50% for NMAE and strongly vary for NME. Results for 90 cm depth are better than for 20 cm depth. Overall model performance for Solling was better t han for Speuld, which can be explained by smaller spatial variability at Solling and a longer data set (17 years for Solling as compared to 3 years for Speuld).
253 [. o
o=..,=,,
~,~
~.,
......... ~o
I
5.0 4.5
-'
4.0
-
3.5
-
H
.~-
-...-. ,~-----.,..-" -%,. ..... "'r176
.._.../
,. ..,,o ,,= . o o
---t
oooo,.o.~176
o....oO ....
~
9
__~
o
3.0 8"0~
(AI=') (m~ m~)
I
4.0
,--'"~,
2.0
. ~ _
0.0 k - " ~ ~ ! - = ' " ' ~ " ~ " 1.o |
.,,,.,,
" ~" " = = ' , , " " - ~ " -
~
~
1
o
7
6
1
7 ;'' 6
1
7
6
(ca") (mot, rr,~)
0.8 I
o
0.6
i
0.4
ooo o Oo 0
/
o
.j
"~.A
.--
..
/'"x. %
o
0000
.
.
s
~176
o
o 00
0
.
o
,t~ ,...t...: oo 000
~ ,Mltt,.r'r
~ ..
9
00
~ ~"
="'[~-*t.
I"M.:--l"
.
%'
-'.~"~.
0.0
5.0
|
(NO,~(mot,m3)
R
4.0 3.0 2.0 r"
1.0
r
,
o -
o
o
9 o ~
Ooooo
~a,, ~to.,...o
*"
t ~o'1
0.0
1 ,.0
~ O _ _ o" o o O. . .o. . .
-r
O,O
~o~. ',~-o-~.."~:~._..o.o.~o__/~....-'--L,_-'',.,l "'--"~ "" ' ...... "" 9
~,.
0.0
~~ '~176
J ~"'~::"! I
il
3.0
~0
.-J
t
1.0 -4 0.0
~ _ o
,~"
~o...__ 1987
,.
o
oooOo
-o
2 #'~--'~. . . . ~'t,=,;.
~o~_~~..~~~. 1988
"p
OOo , .; ,,":'~-oo i OoooO
"~"~"--
~
o
""1~
1989
Figure 2. Observed and simulated soil solutior, concentrations at 20 cm depth for a Douglas fir site at Speulderbos between 1987 and 1991.
254 Tabel 4. Performance for NuCSAM, SoilVeg and ForGro for application to the Speuld Douglas fir dataset and for SoilVeg to the Solling spruce dataset, expressed by the Normalized Mean Absolute Error (NMAE) Normalized Mean Error (NME) Depth
NuCSAM
SoilVeg ForGro Speuld NMAE NME NMAE NME NMAE NME
SoilVeg Solling NMAE NME
[%]
H A1 Mg NO3 SO4 C1
20 20 20 20 20 20
81 49 86 41 44 47
-81 10 -86 -24 -33 -40
21 104 48 35 73 81
7 54 -48 -14 5 27
91 52 38 95 28 110
-91 12 -27 -95 -15 81
45 28 62 47 48
-24 -14 -41 4 -5
H A1 Mg NO3 SOt C1
90 90 90 90 90 90
32 57 54 53 40 52
20 28 -54 2 2 4
40 55 25 62 41 50
34 -40 -10 34 -18 24
25 30 22 73 68 44
-20 22 18 -53 68 3
34 37 41 30 22
-32 33 -11 -20 -11
SoilVeg and ForGro could reproduce the high observed needle mass at Speuld (Figure 3), but tended to overestimate stem mass which is a consequence of the high gross assimilate production predicted by the models given the high needle mass. Average contents of N, Ca, Mg and K (and P for ForGro) in foliage could also be reproduced by SoilVeg and ForGro. The time series of nutrient contents and biomass were too short to evaluate if SoilVeg and ForGro could explain the observed year to year and seasonal variations. The uncertainty and validity of the model calibrations for the Speuld site was not systematically analysed, which in fact is a prerequisite before m e a n i n g ~ l scenario analysis. Questions arose like "How unique is the present calibration? "What is the predictive power of the calibrated models?" and "What is the added value of the model to the original observations.. " ~" The uncertainty of the model calibration can be illustrated by the variation of simulated mean total annual plant uptake of NO3, and NH 4 between 1988 and 1991, respectively 98 and 5 for SoilVeg, 39 and 41 for ForGro and 54 and 49 kg/ha for NuCSAM. Nitrogen uptake is a key process, but can not be checked against observations. However, the validity of the model calibrations was qualitatively touched by evaluating whether SoilVeg and ForGro could reproduce the observed effects of experimental manipulations as conducted for a nearby Douglas fir site (De Visser, 1994). For this purpose these results were compared in a relative way to a simulation of his irrigation and fertigation experiments by the models calibrated for Speuld (Table 5). It should be stressed that the simulation was
255
25
Needle mass (Mg ha"1) --o- observed -o-. SOILVEG -.o-- FORGRO
20 o........ ~ ........ o ........ ~ ' " ~
15
10 220
. Stem mass (Mg ha"t) - o - observed - o - . SOILVEG ..o.. FORGRO
200
180 160 oooO~ ~:~
140
8:'~ " 120
19'88
J 19'89
1990
19"91
Figure 3. Observed and simulated needle and stem mass for a Douglas fir stand at Speuld.
only approximate, as site conditions at Speuld and the manipulation plot are different and model implementations of elimination of water and n u t r i e n t stress are not identical to experimental procedures. Both in SoilVeg and ForGro irrigation a m o u n t s were equal to the difference between daily potential transpiration and precipitation if the water content was below field capacity. Elimination of n u t r i e n t stress in SoilVeg was accomplished by simulating the m a x i m u m fertilizer additions of the field experiment, while ForGro switched off all effects of n u t r i e n t shortage. Observations and simulation were compared after four years of treatment. From this single comparison (Table 5) it can be concluded t h a t directions and magnitudes of predicted changes in the tree and soil are about right b u t t h a t differences can be considerable. In particular prediction of effects on Nmineralization and the effect of fertigation on N-content in foliage are not very good.
Table 5 Comparison of observed effects of irrigation (I) and fertigation (IF) on a Douglas fir site at Kootwijk (De Visser, 1994) with simulated effects by SoilVeg and ForGro for a Douglas fir site at Speuld, relative (%) to an u n t r e a t e d control case.
[C1] 20 cm depth [NO3] 20 cm depth [Mg] 20 cm depth [A1] 20 cm depth stem increment litter fall N content foliage Mg content foliage
Observed
SoilVeg
Forgro
I
F
I
F
I
F
-52 +46 -37 +16 +19 -7 - 10 -2
-53 +82 -10 +28 +25 +9 +1 +8
-17 -6 -16 -17 +6 +1 -2 -4
-11 +46 +60 +34 +62 +18 +8 +11
+19 +9 -8 -2
+24 +5 -7 0
256 5. S C E N A R I O ANALYSIS All models were applied to evaluate a deposition scenario between 1992 and ....... NOx 2050 representing the present targets of . . . . NHx Dutch environmental policy (Figure 4). ~H This scenario is obvious an optimistic one with respect to the reduction of deposition. Simulations were started in /-i ,,.,---"- ~,.. ~ 1950. For this purpose model ~176 S~ ~ "~. parameterizion for Speuld was adapted ...................... ::.:.=.:.:.-.=.--..=~. .:.-.-.--.~.E-.:-.~. to match a so-called generic Douglas fir 1950 1970 1990 2010 2030 205C site a Cambic podzol and a generic Scots Figure 4. Deposition scenario for pine site on a Albic arenosol. Generic SOx, NOx, NHx and potential means that standard databases are used acidity for Douglas fir in the central for soil and tree parameters. The part of the Netherlands evaluated generic soil-forest combinations are assumed to be representative and suitable for regional presentation. Weather data were randomly selected by a statistical model of historically observed weather data. The results of this scenario analysis are only meant as an example of model use for predictive purposes as only one deposition scenario and one realization of weather data was evaluated. For some key stand states, eg. [A1], AUCa ratio, stem mass (and growth), the three models give comparable output differences between tree species and between results for 2010 and 2050 (Table 6). However, for pH, nitrate leaching, nitrogen uptake by trees (not shown), needle mass and N-content in foliage differences are dominated by differences between models. Some apparent effects of the scenario, are a fast improvement of [A1] and A1/Ca ratio in soil solution, a slow improvement of soil solution pH and N content in foliage, small effects on needle mass and considerable depletion of Al-oxides, ranging from 13-25% for Douglas fir. 10
Deposition (kmol= ha "1 a "1)
-----
SOx
257 Table 6 Mean predicted soil and tree s t a t u s s i m u l a t e d by NuCSAM, SoilVeg a n d ForGro between 1980 and 1990, and between 2040 and 2050 for Speuld (SD), a generic Douglas s t a n d on a Cambic podzol (DF) and a generic Scots pine s t a n d on a Albic arenosol (SP). NuCSAM
SoilVeg
ForGro
DF
DF
SD
SP
SP
DF
SP
4.4 0.9 95 5 15.7 143 2.0
4.5 0.3 711 7 8.1 64 1.9
4.7 0.2 85 5 17.7 288 2.0
4.7 0.2 621 3 5.0 179 2.0
m e a n 1980-1990 pH 20 cm d e p t h [All 20 cm depth ( m o l J m 3) Al-oxide 0-20 cm (mmolJkg) NO 3 leaching (kg h a "1 a ] ) A1/Ca ratio 0-20 cm Needle m a s s (ton/ha) S t e m mass ( t o n g a ) N content foliage (% w/w)
3.8 1.1 89 47 3.4 (10.9) 2.6
4.0 0.6 59 33 1.5 (7.5) 2 2.4
3.2 1.4 92 23 1.9 16.2 150 1.9
3.8 0.3 58 7 9.5 8.6 70 1.8
4.3 1.6 159 6 16.5 150 2.1
m e a n 2040-2050 pH 20 cm d e p t h [A1] 20 cm d e p t h ( m o l J m 3) Al-oxide 0-20 cm (mmolJkg) NO 3 leaching (kg ha "1 a "~) A1/Ca ratio 0-20 cm Needle m a s s (ton/ha) S t e m mass ( t o , h a ) N content foliage (% w/w)
4.0 0.1 66 9 0.8 (10.9) 1.0
4.8 0.0 54 3 0.0 (7.5) 2 1.0
3.3 0.9 79 18 0.4 12.9 228 1.6
4.0 0.3 50 9 0.3 8.6 177 1.8
4.6 0.4 133 18.4 303 1.9
ForGro data apply to the soil layer between 10-20 cm depth Needle mass and stem growth are boundary conditions for NuCSAM
6. C O N C L U S I O N S A N D D I S C U S S I O N Conclusions t h a t can be d r a w n from model simulations of the Speuld site, of the m a n i p u l a t i o n e x p e r i m e n t and of the single scenario are provisional because models could not be validated, and are specific because only one s t a n d a n d two generic d a t a s e t s were analysed. In this context the following conclusions are drawn: ~ it is not possible to find unique p a r a m e t e r sets for site models and to validate i n t e g r a t e d site models a g a i n s t observation data, 9 average observed water, n u t r i e n t a n d carbon s t a t u s of the Solling spruce s t a n d are fairly reproduced, b u t seasonal dynamics as well as y e a r to
258
year variation are poorly reproduced, directions and magnitudes of the response of forest soft to naturally occurring variations of weather and deposition and to artificial manipulations agree with observations, for the tree component both observation data and model behaviour are too uncertain to make statistical comparisons or quantitative predictions, different components of the ecosystem respond at different rates to reduced deposition: SO, and A1 in soil solution respond within 1-2 years, NO 3 in soil solution after 5-10 years, N in foliage takes more than 50 years. models predict a quick decrease of A1 and AI/Ca ratios in soil solution to below critical values after a reduction of deposition, models indicate that nitrogen related effects dominate the effects of air pollution and atmosphere deposition on forest in the Netherlands site observations in combination with model application indicate that, for the Dutch pollution climate, direct effects of air pollution with SO2, O3 and NH3 on tree growth, and indirect effect of pH and AI on root functioning, are relevant at stand scale; direct effects appear to be less important than soil mediated effects, SoilVeg and ForGro predict high foliage and low but sufficient fine root (not shown) biomass throughout the scenario, and that stand growth is sustained; the models are inconclusive about the long-term forest production; it should be noted that both stand age and the applied thinning scheme are not realistic for a production stand; similar provisional results are found for Scots pine on a more sensitive Albic arenosol, models do not indicate that short-lived unfavourable soil conditions will cause dieback of the presently productive and vital Speuld stand. models predict considerable depletion of Al-oxides and accumulation of organic nitrogen. These conclusions do not allow a general statement about success or failure of the approach, let alone a statement about the long-term fate of Dutch forests in relation to future air pollution climate and atmospheric deposition. A major shortcoming of the stand level approach is the inability, as yet, to include known catastrophic effects on stand growth and vitality like (i) bacterial, fungal and insect infestations and (ii) bud d~mage due to frost and (iii) severe drought and (iv) wind throw. Events like these will very likely occur several times within one stand rotation and the susceptibility of the stand to such effects is related to nutrient (nitrogen) status and biomass of foliage and roots. For now, these effects can only be considered in terms of increased risk for occurrence. Given these limitations, and acknowledging that we still cannot bridge the gap in time and spatial scales between experimental effect assessment and regional effect observations, any prediction of long-term effect of air pollution on forests is bound to be very uncertain. Assessment of long-term effects and response of the soil forest is less uncertain. The long-term depletion of Al-oxide pool, which
259 is the major source of acid buffering in poor sandy soils, is beyond discussion. Depletion will further lower pH and may destabilize organic m a t t e r and enhance DOC and heavy metal mobility (Westerhof et al., 1994). Also the strong accumulation of organic N is evident, but its long-term stability is u n c e ~ . Although the reliability of the submodels for soil hydrology and soil chemistry allows m e a n i n ~ regional application of the models, the uncertainty about the growth and effect submodels is still too large too allow m e a n i n ~ long-term and regional prediction of the carbon and nutrient status of forests. Forest stand models do not confirm the validity of present critical levels for A1, Ca/A1 ratio and loads for acidity and nitrogen (De Vries, 1993) for prevention of meaningful harmfid effects to the forest component. The models do not provide better alternatives, but models could be helpful to redefine critical loads for the tree component (De Vries et al., 1995), if harmfid effect are defined more clearly. Critical levels and loads are best supported by effects on other, more sensitive component of the forest ecosystem. Perhaps expectations with respect to the answers from site monitoring and modelling have been too high, although enormous progress was made in the past five years. There cert~nly is a need to continue monitoring at the stand level, but perhaps lower sample frequencies (eg. annual) and a smaller selection of parameters are adequate to detect magnitudes and trends of effects of weather, air pollution and atmospheric deposition on soil N and proton buffer status, tree nutrient status, and growth. There is not a good reason to rebuild the site models. At present, process-oriented forest stand models are, above all, suitable and probably the only available tools to integrate and synthesize mechanistic formulations of hypothetical or proven relationships between the various states of a forest stand. After calibration, these models can be used to indicate directions and magnitudes of short-term and long-term responses of the stand to changes of the pollution climate and weather. Present model versions should be used to further explore available observation sets (EXMAN, NITREX) and present site calibration could be used to assess the uncertainty of qualitative predictions for the calibration, deposition scenarios and weather. Perhaps uncertainty analysis of the integrated forest stand models could be carried out in a risk perspective, after empirical incorporation of the indirect "catastrophic" effects. Acknowledgements The data used for this paper were originally collected and compiled by M. van der Maas, P. de Visser, T. Pape and N. van Breemen d, W. Bouten and J. Verstraten (University of Amsterdam), W. dans b, A. Olsthoom b, d.W. Erism~n a, for some cases in cooperation with the authors. W. de Vries c was strongly involved in the development of NuCSAM. We thank Jan Mulder (NISK), and Geert Draaijers for their contributions.
260 7. R E F E R E N C E S
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
20 21 22 23 24
O. Kandler. Evironm. Toxic. Chemistry. Vol. 11 (1992) 1077. B. Ulrich, in D.C. Adriano and A.H. Johnson (eds), Acidic Precipitation, Volume 2" Biological and ecological effects, Springer Verlag, New York, 1989. N. van Breemen, P.A. Burrough, E.J. Velthorst, H.F. van Dobben, T. de Wit, T.B. de Ridder and H.F.R. Reijnders. Nature, 299 (1982) 548. W. de Vries, J.J.M. van Grinsven, N. van Breemen, E.E.J.M. Leeters and P.C Jansen. in press for Geoderma (1994). P. Gundersen. Nord 1992:41 (1992) 55. M. Posma, W.W.P. Jans and E.G. SteingrSver, IBN Research Report 94/2, ISSN 0928-6896 (1994). J.J. Landsberg, M.R. Kaufmann, D. Binkley, J. Isebrand, and P.G. Jarvis. Tree Physiology 9 (1991) 1. J.J.M. van Grinsven, C.T. Driscoll and A. Tiktak. accepted for Ecol. Model. (1995). P.W. Evers, C.J.M. Konsten and A.W.M. Vermetten. Proc. Syrup. Effects of Air Pollut. On terrest. Ecosyst, Grenoble (1987) 887. N. van Breemen and J.M. Verstraten, in G.J. Heij and T. Schneider (eds), Acidification Research in the Netherlands, Elsevier, Amsterdam, 1991. J.W. Erisman, Water, Air Soil Pollut. 71 (1993) 51. W. Bouten, T.J. Heimovaara and A. Tiktak. Water Res. Res. 28 (1992) 3227. T. Schneider, and G.J. Heij. Dutch Prior. Progr. on Acidif. rep. 300-1 (1993). W. de Vries, J. Kros and C. van der Salm. in press for Ecol. Model. (1995). Berdowski, J.J.M., C. van Heerden, J.J.M. van Grinsven, J.G. van Minnen and W. de Vries. Dutch Prior. Progr. on Acidif. rep. 114.1-02 (1991). Heij and Schneider (eds), Acidification Research in the Netherlands, Stud. Envir. Sci. 46, Elsevier, Amsterdam, The Netherlands, 1991. C. van Heerden, J.J.M. van Grinsven and A. Tiktak. accepted for Ecol. Model. (1995) G.M.J. Mohren, H.H. Bartelink, I.T.M. Jorritsma and K. Kramer, in M.E.A. Broekmeijer, W. Vos and H. Koop (eds), PUDOC, Wageningen, 1993. A. Tiktak, J.J.M. van Grinsven, J.E. Groenenberg, C. van Heerden, P. H.M. Janssen, J. Kros, G.M.J. Mohren, C. van der Salm, J.R. van der Veen, W. de Vries, in prep. for Heij and Schneider (eds), Elsevier, Amsterdam, 1995. P.H.M. Janssen and P.S.C Heuberger. accepted for Ecol. Modelling (1995). P.H.B. de Visser. PhD thesis, ISBN 90-5485-290-9, 1994. R. Westerhof, J. Mulder and D. Bergren. submitted to European J. Soil Sci. (1995). W. de Vries, Water, Air, Soil Pollut. 68 (1993) 399. W. de Vries, M. Posch, T. Oja, H. van Oene, J. Kros, P. Warfvinge and P.A. Arp. accepted for Ecol. Modelling (1995).
G.J. Heij and J. W. Erisman (Editors). Acid Rain Research: Do we have enough answers? 9 1995 Elsevier Science BV. All rights reserved.
261
LARGE SCALE IMPACTS OF ACID DEPOSITION ON FORESTS AND FOREST SOILS IN THE NETHERLANDS
W. de V r i e s ~, E . E . J . M . L e e t e r s ~, C . M . A . H e n d r i k s ~, H. v a n D o b b e n b, J. v a n den Burg b and L.J.M. Boumans c
a DLO Winand Staring Centre for integrated Land Soil and Water Research, PO Box 125, 6700 AC Wageningen, The Netherlands b
DLO Institute for Forestry and N a t u r e Research, PO Box 23, 6700 AA Wageningen, The Netherlands National Institute of Public Health and E n v i r o n m e n t a l Protection, PO Box 1, 3720 BA Bilthoven, The Netherlands
Abstract Since the early eighties, effects of acid atmospheric deposion have received much attention in the Netherlands. Effects of elevated S and N deposition on soil solution chemistry is mainly manifested by increased concentrations of A1 associated with increased concentrations of SO 4 and NO 3. P r e s u m e d critical A1 concentrations (0.2 molc m 1) and A1/Ca ratios (1.0 mol mol 1) are generally exceeded below 20 cm soil depth. There is also ample circumstantial evidence t h a t elevated N deposion during the last decades affected the forest n u t r i e n t status and caused large changes in forest vegetation. About half of the Dutch forests have foliar N contents exceeding a critical limit (1.8%). Field evidence for a relationship between soil acidification and n u t r i e n t imbalances in the soil and the foliage on one h a n d and the vitality of forests (mainly expressed by defoliation class) on the other h a n d is, however, lacking. This result implies t h a t an exceedance of critical acid loads, based on critical A1 concentrations and A1/Ca ratios observed experiments in relation to effects on root (uptake), do not imply visible effects or even the dieback of forests.
1. I N T R O D U C T I O N Since the early eighties it is recognized t h a t Dutch forests receive large inputs of NH4 and SO4 from the atmosphere (Van Breemen et al., 1982). At the same time the possible role of atmospheric sulphur (SOx) and nitrogen (NO,,, NH~) deposition in forest dieback in central Europe became a subject of wide public and political discussion. Since then, effects of atmospheric deposition of SOx, NO~ and NHx on forests have received much attention in the Netherlands. Research efforts were focused on the impacts of acid deposition on soil (solution) chemistry (e.g. Van Breemen and Verstraten,
262 1991; De Vries and Leeters, 1994), forest vegetation (e.g. Hommel et al., 1990; Van Dobben et al, 1994), forest n u t r i e n t status and forest vitality (e.g. Van den Burg et al., 1988; De Visser, 1994) and the relationship between soil solution chemistry and forest vitality (e.g. Roelofs et al., 1985; Hendriks et al., 1994). I m p o r t a n t effects of elevated SOx, NOx and NH~ deposition on the soil are: (i) a decrease on base saturation and pH and (ii) an increase in the concentrations of SO4, NO3, NH4 and A1. Controlled laboratory experiments on seedlings have shown t h a t increased concentrations of NH4 and A1 in relation to Ca, Mg and K affect the root system of trees and the uptake of these base cations. Because of the importance of soil mediated effects, much experimental research on the effects of acid atmospheric deposition on Dutch forests dealt with the impacts on the soil solution chemistry. Several local monitoring studies were performed in a total of eighteen forest stands. F u r t h e r m o r e , a one-time national survey was carried out in early spring in 150 representative forest stands on non-calcareous sandy soils across the country. The chemical composition of the soil solution in both the monitoring and survey sites was mainly measured to gain insight in (i) the fate of SO4, NH4 and NO3 in forest soils and (ii) the mobilization of A1 and base cations (Ca, Mg, K and Na), which mainly neutralize the acid input, based on i n p u t - o u t p u t budgets (Van Breemen and Verstraten, 1991. De Vries and J a n s e n , 1994). The one-time soil solution survey in 150 forest stands was also carried out to study the relationship between the soil and soil solution chemistry and forest vitality characteristics, such as needle loss and needle colour, (Hendriks et al., 1994). Consequenly, the 150 survey stands were selected from about 3000 stands belonging to the Dutch forest vitality inventory. According to the frequency distribution of tree species in the Netherlands, 45 stands of Pinus sylvestris (Scots pine; 30%), 30 stands of Quercus robur (oak; 20%) and 15 stands (each 10%) of Pinus nigra subsp maritima (Corsican pine), Pseudotsuga menziesii (Douglas fir), Picea abies (Norway spruce), Larix kaempferii (Japanese larch) and Fagus sylvatica (beech) were selected. More information on the selection procedure is given in De Vries and Leeters (1994). Other i m p o r t a n t aims of the one-time national survey were the (i) determination of the chemical composition of the foliage to assess the n u t r i e n t supply (Hendriks et al., 1994) and (ii) assessment of the chemical composition of the soil solution on a national scale from relationships with deposition level, stand characteristics (such as tree species) and site characteristics (such as soil type; Leeters et al., 1994). This paper summarizes the impacts of acid deposition on forests in the N e t h e r l a n d s with respect to soil solution and g r o u n d w a t e r chemistry, forest n u t r i e n t status, forest vegetation and forest vitality. Critical loads for nitrogen and acidity t h a t have been derived in relation to several effects will be reviewed critically in the light of these results.
263
2 EFFECTS ON FOREST SOILS 2.1 S o i l s o l u t i o n c h e m i s t r y Exceedances of critical chemical values The most pronounced reflections of atmospheric deposition of S and N on the solution chemistry of Dutch acid sandy soils are the high SO4 and NO3 concentrations in the soil solution and concomitantly high concentrations of acidity (H+A1). Results of linear regression analyses for the various monitoring and survey sites, showed t h a t the concentration of H+A1 gets closer to the concentration of SO4+NO3 with increasing depth for nearly all sites (de Vries et al., 1994b). Moreover, the linearity of the relationship increased with increasing depth as shown by an increasing value for the adjusted coefficient of determination (R 2adj. ) A 1"1 relationship (on an equivalent basis) between [H+A1] and [SO4+NO3] in the soil layers below the rootzone, indicates t h a t external inputs of N and S to the soil (corrected for N and S retention in the soil) will cause mobilization and leaching of equivalent a m o u n ts of H and A1. In all monitoring sites with a non-calcareous subsoil (16 out of 18 sites), the calculated a n n u a l average A1 concentration in the subsoil largely exceeded the Dutch drinking water standard (0.2 mg 11 = 0.02 mole m -a) in the subsoil. The a n n u a l average NO3 concentration exceeded the EC drinking water s t a n d a r d for NO3 (50 mg 11 = 0.8 mole m a) in 14 sites and the Dutch t arget value (25 mg 1 -1 = 0.4 mol~ m 3) in 16 sites. Molar A1/Ca and NH4/K ratios, which are regarded as indicators of potential reduction of n u t r i e n t u p t a k e by roots, showed strong gradients with depth in the intensively monitored sites. The A1/Ca ratios, based on annual average A1 and Ca concentrations, for these sites generally exceeded an assumed critical value of 1.0 at a depth of 20 cm, which is within the predominant zone of root activity. In contrast to A1/Ca ratios, NH4/K ratios tended to decrease, and so become more favourable for biota, with increasing depth. In several intensively monitored sites molar NH4/K ratios were above an assumed critical value of 5 in the upper 20 cm of the soil. In the non-calcareous survey sites (147 since three sites appeared to be calcareous) the EC drinking water standards in the subsoil (60 - 100 cm below soil surface) were exceeded every where for A1 (Fig 1A) and in 31% of the cases for NO3, whereas 55% of the sites exceeded the Dutch target value for NO 3 (cf Fig. 1B). An A1 concentration of 0.2 mol c m -3, which is considered critical with respect to effects of roots, was exceeded in 80% of the sites (Fig 1A). However, A1 and NO3 concentrations measured at the survey sites m ay deviate from a n n u a l average values. Values for the molar A1/Ca and NH4/K ratio in the survey sites were generally lower t h a n those at the monitoring sites. In the forest topsoil (0-30 cm depth), the critical A1/Ca ratio of 1.0 was exceeded in 57% of the 147 survey sites (Fig 1C), whereas only 4% of the
264 100
i
A Topsoil
,
.......... Subsoil
Critical limit "";
~ v >, u
,-
0
o-
";"~
I
I '"
........
I
|
1
2
. ................
""',.
. ....
I
I
I
1
2
I
NO 3 concentration (mol c m -3)
AI concentration (mol c m -3)
.>_ E
100
I
L.2
\,,
",%
2
3
AI/Ca ratio (mol mol 1)
Figure 1.
4
5
o
I
I
I " ..............
~ ..............
1
2
3
4
....
-J
5
NH4/K ratio (mol mo1-1)
Inverse cumulative frequency distributions of Al concentrations (A), NO3 concentrations (B) Al/CA ratios (C) and NH4/K ratios (D) in the topsoil (O30cm) and subsoil (60-100cm) of 147 survey sites
survey sites exceeded the critical NH4/K ratio of 5 (Fig. 1D). The relatively low A1/Ca ratios were mainly a result of a high Ca concentration. This may partly be the result of a high Ca input from the atmosphere due to strong filtering of base cations by the forest canopy, especially near forest edges. The most important explanation might, however, be the different methodology to obtain soil solution (cf De Vries et al., 1994b). The various soil solution parameters in the survey sites were largely influenced by tree species. Lowest pH values and highest concentrations in A1, SO4 and NO3 occurred below Douglas fir. The reverse was true for oak and beech, whereas Scots pine en black pine occupied an intermediate position.
265 The increase in solute concentrations between tree species, going from deciduous forest to pine forests to spruce forests, is most probably caused by increased dry deposition and evapotranspiration. Concentrations of A1, SO4 and NO3 also increased with increasing tree height and canopy coverage. This is most likely due to an increase in atmospheric S and N deposition (cf De Vries et al., 1994 b). Relatively good regression relationships (R2~di > 0.5) were found between the SO4 and A1 concentration and the tree species, tree height, percentage of forest in the surrounding area and acid atmospheric deposition. Even tree species and tree height alone already explained n e a r l y 40% of the variation in SO4 and A1 concentration in the forest topsoil. For NO3, the relationships were slightly worse (Re~dj = 0.4; cf Leeters et al., 1994). The regression relationships t h a t were found were used to produce SO4, NO3 and A1 concentration maps at a grid resolution of 0.5 km x 0.5 km. Relatively low concentrations of SO4, NO3 and A1 were generally predicted in the large forest complexes in the central part (Veluwe area) and the n o r t h e r n part (Drenthe) of the Netherlands, where atmospheric deposition is comparatively low. High concentrations were generally predicted in the small forest complexes in the eastern and southern p a r t of the Netherlands. Small scale variation was caused by variation in tree species and tree height (cf Leeters et al., 1994)
input - output budgets Soil solution chemistry data can be used to calculate drainage outputs from the soil. Comparison of such outputs with atmospheric inputs gives quantitative information on the retention or mobilisation of elements. Regarding S and N, it also gives insight whether the deposition of potential acidity (SOx, NOx and NH x) is actually realised in the soil. This is only the case when all deposited S and N leaves the soil in the form of SO4 and NO3. I n p u t - output budgets for ten intensively monitored sites (Van Breemen and Verstraten, 1991) and the 147 survey sites (De Vries and J a n s e n , 1994) showed that. Dutch forest soils are SO4 saturated, but N is still largely retained (Table 1) Table 1. Average atmospheric inputs and drainage outputs of S04, NH4, NOa and total N in ten intensively monitored soils between 1981 and 1990 and 147 soils in which a single measurement took place in 1990
Type of research
Flux (kmolc ha 1 yr1) SO 4
in Monitoring Survey
NH4
out
2.77 2.69 1.74 1.77
in
NO3
out
2.98 0.12 3.19 0.13
in
N
out
0.87 1.78 0.97 0.70
in
out
3.85 4.16
1.90 0.83
Both the i n p u t and output of SO4 was lower in the 147 survey sites t h a n in the ten monitoring sites, which is most probably due to the decrease in SO4 deposition during the eighties. However, in both types of research, average
266 SO4 inputs equalled average S04 outputs, implying that S04 deposition contributes for 100% to the actual acidification of the soil, Atmospheric deposition of NH 4 and NO3 was comparable for the monitoring sites and survey sites. In the survey sites retention of N in the forest (by uptake) and the soil (by immobilization) was, however, larger t h a n in the monitoring sites, since NO3 leaching was considerably lower; even below the NO3 input (cf Table 2). Theoretically, NH4 deposition may not contribute to the acidification of these sites since it is either taken up a immobilised. In the intensively monitored sites, however, the difference between NO3 leaching and NO3 deposition of ca 0.9 kmol c ha 1 yr 1 is due to nitrification, and ammonia at least contributes ca 20% to the actual acidification of these soils (ca 4.5 kmolc ha 1 yrl). Soil acidification is mainly manifested by A1 mobilization from secondary A1 compounds (De Vries et al., 1994b). Model calculations indicated t h a t forest soils may become depleted in these compound within several decades at ongoing present atmospheric inputs (De Vries and Kros, 1989). This may cause a strong decline in pH, which may be an important stress to the forest ecosystem. However, the risk of A1 depletion will strongly decrease at expected emission reductions. De Vries et al. (1993) calculated t h a t the time period to reach complete A1 depletion will increase from approximately 100 years at present acid loads to 2000 years at expected emission reductions in 50% of all forest soils. In the near future (2050), a relative A1 depletion above 50% was only predicted in small areas in the southern and eastern parts of the Netherlands in soils with low present amounts of secondary A1 compounds (De Vries et al., 1993). Model calculations also showed t h a t emission reductions will lead to a fast improvement of the soil solution quality (decreased concentrations in SO4, NO3 and A1 and increased pH; De Vries et al 1994a).
2.2 G r o u n d w a t e r c h e m i s t r y As with the soil solution, elevated atmospheric deposition has increased the concentrations of SO4, NO3 and A1 in groundwater. This can be derived from a study on the chemical composition of phreatic groundwater in 156 gridcells of 0.5 km x 0.5 km , containing at least 0.1 ha of forest or heathland, t h a t was carried out between 1989 and 1990 (Boumans and Beltman, 1991). P a r t of the m e a s u r e m e n t s coincided with the 150 forest stands where the chemical composition of foliage, soil and soil solution was measured (a total of 71 stands) Median SO4 concentrations in phreatic groundwater at these sites were comparable to those in the soil solution, but concentrations of A1 and NO 3 were lower (Table 2). Still, A1 concentrations exceeded the EC drinking water standard of 0.2 mg 11 in ca 80% of the sites (forest and healthlands), whereas NO 3 concentrations exceeded the EC drinking water s t a n d a r d (50 mg 11) and the Dutch target value (25 mg 11) in 20% and 37% of the sites
267 respectively (after Boumans and Beltman, 1991). The quality of drinking water, pumped up at much greater depths, will, however, be less affected because of A1 retention and denitrification in the groundwater aquifer and because of mixing with various other watertypes. Unlike soil solution, the N O 3 and A1 concentration in groundwater was best 'explained' by a regression model including soil type and to a lesser extent by tree species (coniferous deciduous forest), tree height, surrounding land use and atmospheric deposition (Leeters et al., 1994; Boumans, 1994). NO3 concentrations, for example, increase according to peaty soils < moderately drained, poor sandy soils < well drained rich sandy soils. (Boumans, 1994). As with the soil solution, however, lowest concentrations are generally found in the large forest complexes in the central part of the Netherlands and in the low deposition areas in the northern part of the Netherlands (Leeters et al., 1994). Table 2. Median concentrations of S04, NO3, NH4 and Al in the soil solution and in phreatic groundwater of 71 Dutch forest stands, sampled between 1989 and 1990. (After De Vries and Jansen, 1994)
Element SO4 NO 3
NH4 A1
3.
3.1
Median concentration (molc m "3) 0 - 30 cm
60 - 100 cm
groundwater
0.97 0.53 0.20 0.69
1.08 0.48 0.09 0.67
1.04 0.24 0.00 0.54
E F F E C T S ON T H E F O R E S T E C O S Y S T E M
Foliar c o m p o s i t i o n
Elevated atmospheric deposition of N and S compounds in the Netherlands during the last decades has led to an increase in the N content and a decrease in the P, K and Ca content in foliage. This can be derived from a study by van den Burg and Kiewit (1989), who compared the foliar composition of stands of Scots pine, black pine and Douglas fir in 1956 and 1988 in the 'Peel' area with intensive animal husbandry. Surprisingly, the Mg content did not decrease during t h a t period However, even in 1956, the Mg content was already low. Furthermore, as with P, K and Ca, the Mg supply relative to N decreased (Table 3).
268 Table 3. Average N contents and ratios of P, K and Mg to N in half years old foliage of stands of Scots pine, black pine and Douglas fir in 1956 and 1988 (After Van den Burg and Kiewit, 1988). Tree species
Scots pine Black pine Douglas fir
Nutrient content (%)
Nutrient ratio x 100
N
P
1956 1988
1956
1988
1956
1.5 1.2 1.4
0.15 0.16 0.25
0.14 0.11 0.12
9.9 12.2 18.1
2.3 1.7 2.2
P/N
(gg-1)
K/N
Mg/N
1988
1956 1988
1956 1988
6.1 6.7 5.1
34 58 68
3.0 4.0 6.1
27 35 24
2.7 3.8 5.0
T h e a v e r a g e i n c r e a s e in N c o n t e n t b e t w e e n 1956 a n d 1988 v a r i e d f r o m 0.5% ( b l a c k p i n e ) to 0.8% (Scots p i n e a n d D o u g l a s fir). A c c o r d i n g to c r i t e r i a t h a t h a v e b e e n g i v e n to j u d g e t h e n u t r i e n t s u p p l y of t r e e s in r e l a t i o n to g r o w t h , t h e c o n t e n t c h a n g e d f r o m s h o r t a g e to e x c e s s (cf T a b l e 5). T h e h i g h N cont e n t s m a y also i n c r e a s e t h e s u s c e p t i b i l i t y to f r o s t ( A r o n s s o n , 1980) a n d f u n g a l d i s e a s e s s u c h as G r e m m e n i e l l e a b i t i e n a a n d S p h a e r o p s i s s a p i n e a (Roelofs et al., 1985; B o x m a n a n d V a n Dijk, 1988). T h e l a r g e s t c h a n g e s in t h e f o l i a r c o m p o s i t i o n o c c u r r e d for D o u g l a s fir. N u t r i e n t r a t i o s in t h i s t r e e w e r e g e n e r a l l y o p t i m a l for g r o w t h in 1956, w h e r e a s t h e y w e r e n e a r t h e level of d e f i c i e n c y i n 1988. A s i n g l e m e a s u r e m e n t of t h e foliar c o m p o s i t i o n of t h e f o r e s t s t a n d s in t h e s u r v e y of 1990 also i n d i c a t e d n u t r i e n t deficiencies in s e v e r a l t r e e s p e c i e s ( T a b l e 4). In e i g h t s t a n d s w i t h too h i g h t r e e s no m e a s u r e m e n t s w e r e m a d e . I n f o r m a t i o n on t h e c r i t e r i a t h a t w e r e u s e d to j u d g e t h e n u t r i e n t c o n t e n t s a n d n u t r i e n t r a t i o s in foliage is g i v e n in H e n d r i k s et al (1994). Table 4 Exceedances of the lower limit of critical foliar nutrient contents and ratios of seven tree species in forest stands (After Hendriks et al., 1994)
Tree species
Nr
Exceedance (%) N 1)
Scots pine Corsican pine Douglas fir Norway spruce Japanese larch Oak Beech All
43 14 16 15 13 27 14 142
91 14 69 67 36 55 13 49
P
K
Ca
Mg
P/N
K/N
Mg/N
37 79 75 33 92 26 100 63
2 14 50 67 23 4 43 29
28 36 0 33 31 63 21 30
23 7 0 27 0 59 86 29
9 7 50 7 23 30 64 27
23 7 38 33 23 4 29 22
95 57 19 60 23 44 71 53
1) For N it is the upper limit T h e m o s t s t r i k i n g c o n c l u s i o n s f r o m T a b l e 4 a r e (i) t h e a b s o l u t e e x c e s s of N, (ii) t h e a b s o l u t e s h o r t a g e of P a n d , (iii) t h e r e l a t i v e s h o r t a g e of M g ( a n d to a
269 lesser extent also K) compared to N. Foliar nutrient contents were most significantly related to tree species, and to a lesser extent to n u t r i e n t contents in mineral soil, soil solution and humus layer. A high heavy metal content (Pb, Zn, Cd) of the humus layer negatively influenced the foliar nutrient content of P and Mg. Part of the 150 stands, i.e. four stands of Scots pine, Douglas fir and oak, have been monitored since 1992. However, the time period is too short to derive reliable trends. Simulations with the integrated dynamic forest/soil model SOILVEG indicate that expected N emission reductions in the coming decade may lead to a decrease in the foliar N content of ca 0.2 - 0.3% (Van Grinsven et al., 1991).
3.2 Forest v e g e t a t i o n Circumstantial evidence is available for large changes in forest understory in the Netherlands over the period 1950-1990. These changes entail: (i) a decline of terrestrial lichens ('reindeer lichens') and of ectomycorrhiza mushrooms; (ii) an increase of grasses, notably Deschampsia flexuosa and (iii) a general increase of mosses and vascular plants that typically occur on nitrogen-rich soils. However, detailed information at site level is scarce. A pilot study was carried out by De Vries (1983) at 'Boswachterij Kootwijk'. This is an area of pine forests on dry, sandy soil, where vegetation maps from. 1957 were available. This study showed a complete changeover in understorey vegetation, from a moss- and lichen-dominated type to a grassdominated type. Scattered information on other sites shows that the changes observed at this site are probably typical for most Dutch pine forests on poor soils. A comparable study was carried out by Hommel et al. (1990) in a neighbouring area ('Speulderbos'), with both deciduous and coniferous forest. Comparable results were obtained in this study; in general, the nitrogen indicator value of the vegetation has strongly increased. A statistical evaluation of the geographical distribution of the vegetation changes showed t h a t these changes were stronger at shorter distances to agricultural areas. Thus, this study yielded a direct indication for agriculturally derived ammonia as a cause for the observed changes. Changes in the Dutch mushroom flora have been studied by comparing old (circa 1950-1970) excursion reports with recent inventories. Extensive studies of this type, carried out by Arnolds (1991) and others, showed a strong decline of fruitbodies of ectomycorrhizal fungi, and an increase of fruitbodies of wood-inhabiting saprotrophic and parasitic fungi. Among the soil-inhabiting saprotrophic species those of nutrient-rich soils had increased, while those of nutrient-poor soils had decreased. Changes in the understory of Dutch pine forests in a more recent period were studied by comparing vegetation descriptions made in 177 p e r m a n e n t plots in 1984 and in 1993 (Van Dobben et al. 1994). This study showed a
270 significant decrease in the cover of Erica tetralix and Calluna vulgaris and a strong increase of m any nitrophilous species. As a consequence, a highly significant increase in Ellenberg N-indicator value was observed. Interestingly, the Ellenberg pH-indicator value had also significantly increased, indicating soil alkalization. This might be due to the decrease in S deposition during this period. Although direct information on the cause of these changes is lacking, most authors agree on n a t u r a l succession and atmospheric N deposition as the most i m p o r t a n t factors. However, at present it is hardly possible to estimate the relative importance of these two factors. Studies comparing vegetation development in u n t r e a t e d and fertilized pine forests in areas with a low background deposition of nitrogen can be used to estimate the effects of nitrogen on forest vegetation. Such studies have shown t h a t the changes in Dutch pine forests can be r a t h e r well simulated by the addition of nitrogen fertilizer at rates comparable to the rate of atmospheric nitrogen deposition (Van Dobben 1993). On the other hand, vegetation changes provoked by experimental soil acidification are generally unrelated to those presently observed in the Netherlands. The observed increase in Ellenberg pH-indi cator value also implies t h a t soil acidification is not the main cause for the vegetation changes, In t h a t case a decrease in pH-indicator value would be expected. However, note t h a t the increase in pH-indicator value may also be an artifact due to succession. Studies on the significance of the changes in the undergrowth for the tree layer are virtually lacking. There are no indications t hat the dense grass cover h a m p e r s tree juvenation or succession in the tree layer. The decline of ectomycorrhizal fungi may be a factor contributing to a general decline of tree vitality, but a hard proof for this hypothesis is lacking.
3.3 F o r e s t V i t a l i t y The h ealth of forests in the Netherlands has been monitored since 1983. To describe the state of health of a forest the term "vitality" is used. The vitality, is influenced by "traditional" factors such as fertility of the soil, provenance of tree species, rainfall, frost and drought, and pests and plagues, as well as by "new" anthropogenic factors such as air pollution and acid deposition. The annual vitality survey mainly has as an indicator function. It only gives an indication of the possible occurrence of combined stress. The commonly used indicators for the vitality of trees are defoliation and foliar discolouration. The defoliation class is judged to be the most i m p o r t a n t aspect of vitality. Since the beginning of the forest vitality inventories in 1984, the vitality of Dutch forests decreased until 1989. In 1990, the year in which the field inventory in 150 stands was carried out, the vitality was stabilised at the level of 1989. The vitality in 1991 was about the same. In 1992 the vitality decreased strongly, followed by a steady
271 increase in 1993 and 1994. Forest vitality characteristics, such as defoliation and foliar discolouration, are not only a function of the nutrient status of the tree and of factors influencing nutrient availability (e.g. the heavy metal contents in the humus layer and the A1 concentration in the soil solution, which affect n u t r i e n t mineralisation and nutrient uptake respectively), but also of site factors such as rainfall, soil type, groundwater level (all affect soil water supply) and stand age. Results from a recent forest vitality inventory in Europe shows that especially stand age can be very important (UN-ECE/CEC, 1992). The defoliation class measured in 150 forest stands was related to these, so called, explanatory variables with multiple linear regression. Apart from tree species, soil type, groundwater level and stand age, stand characteristics which influence the input of elements by atmospheric deposition, i.e. tree height, canopy coverage and distance of the trees to the forest edge, were also included. Chemical variables that were used to predict the defoliation class were limited to the N content in foliage, humus layer and soil solution, the ratios of P, K, Ca and Mg to N in foliage, heavy metal contents in the humus layer and the pH, A1/Ca ratio and N H J K ratio in soil solution. This limitation was based on the assumption that the chemical composition of the foliage is the best reflection of the nutrient status of the forest, whereas elevated A1/Ca and NH4/K ratios in solution and heavy metal contents in the humus layer may limit nutrient uptake due to root damage. Tree species and stand age appeared to be very important explanatory variables. Together they explained about 44% of the variation in defoliation class (cf Table 5). The defoliation class becomes higher (worse vitality) with an increasing stand age. Using Mallows Cp as criterium to select the best subset of explanatory variables, the models 2, 3 and 4 gave the best description of the relation with the defoliation class. The percentage variance accounted for was 46% for all of the three equations. Using the percentage accounted for as criterium, the best description of the defoliation class was obtained by model 5. The value of R2~aj was 53%. According to model 5 the defoliation class increases (worse vitality) with an increasing foliar N content and/or a decreasing pH of the soil solution. This is a well acceptable explanation in view of existing theories on forest vitality (e.g. Ulrich and Matzner, 1983; Boxman en Van Dijk, 1988, Van den Burg and Olsthoorn, 1994). The nutrient status of the foliage and soil was, however, relatively unimportant in explaining forest vitality compared to tree species and stand age. A decrease in vitality could only partly be related to a relative P deficiency, due to N excess in the foliage, and to a decreasing pH of the soil solution. Results of this study do not confirm the theory t h a t an increased A1 concentration causes a decreasing vitality (Ulrich and Matzner, 1983), which is mainly based on laboratory experiments, showing root damage and limiting nutrient uptake at increased A1 levels.
272 Table 5 Percentage of variance accounted for (R 2 adj) in several regression models between defoliation class, site characteristics and chemical variables
Explaining variable
Explaining models 1
2
3
4
5
Tree species Stand age Canopy coverage Soil type Foliar content N Soil pH
* *
* *
* *
* * * *
* * * * * *
R 2 adj
0.44
0.46
0.53
*
0.46
0.46
4 DISCUSSION AND CONCLUSIONS Critical loads
In the N e t h e r l a n d s , critical loads for n i t r o g e n a n d acidity h a v e been derived using (i) empirical d a t a t h a t directly r e l a t e loads to effects a n d (ii) steadys t a t e soil models t h a t calculate critical loads from critical chemical v a l u e s for ion c o n c e n t r a t i o n s or ratios in foliage, soil solution a n d g r o u n d w a t e r . Critical loads t h a t t h u s h a v e been derived are shown in Table 6 (cf De Vries, 1993). B a s e d on the d a t a in Table 6, a t a r g e t acid load of 1400 mole h a 1 yr -1 h a s been set for the y e a r 2010 with an N i n p u t below 1000 molr ha 1 yr 1. The u n c e r t a i n l y in critical loads given in Table 6 is s t r o n g l y i n f l u e n c e d by the r e l i a b i l i t y of the critical values for the chemical p a r a m e t e r s . A critical review on the various criteria for A1 ( S v e r d r u p a n d Warfvinge, 1993) shows t h a t the A1 c o n c e n t r a t i o n criterion of 0.2 molc m 3 is v e r y u n r e l i a b l e , w h e r e a s the m o l a r A1/Ca ratio (or b e t t e r A1/(Ca + Mg + K) ratio) should be defined as a function of tree species. S v e r d r u p a n d W a r f v i n g e (1993) derived critical A1/(Ca + Mg + K) ratios of 0.2 for Willow, 0.5 for larch, a s h and black alder, 0.8 for Scots pine a n d N o r w a y Spruce, 1.2 for Birch, 1.7 for O a k a n d Beech a n d 3.3 for Douglas fir, which are all common tree species in t h e N e t h e r lands. T h e y took a 20% reduction in e i t h e r b i o m a s s growth, root l e n g t h or root g r o w t h as the c r i t e r i u m to derive such ratios while u s i n g a large compilation of l i t e r a t u r e d a t a on the effects of A1 on trees. A p a r t from A1, a critical NO3 c o n c e n t r a t i o n r e l a t e d to v e g e t a t i o n c h a n g e s is also v e r y u n r e l i a b l e . A n i t r o g e n m a s s b a l a n c e for a calcareous g r a s s l a n d in the N e t h e r l a n d s indicates t h a t v e g e t a t i o n s changes m a y t a k e place in a s i t u a t i o n w h e r e N l e a c h i n g h a r d l y i n c r e a s e s above n a t u r a l b a c k g r o u n d v a l u e s (Van D a m , 1990). Similarly, N l e a c h i n g is n e a r l y negligible in D u t c h h e a t h l a n d s c h a n g i n g into
273
Table 6
Average critical loads for acidity and nitrogen for forest ecosystems in the Netherlands (After De Vries, 1993)
Compound
Effects
Criteria ~)
Critical loads (molr ha ~ yr% Coniferous forests
Acidity
Nitrogen
Root damage Inhibition of uptake A1 depletion A1 pollution Inhibition of uptake Increased susceptibility Vegetation changes Nitrate pollution
A1 < 0.2 molr m 3 Al/Ca < 1.0 mol mol 1 aAl(OH)s=0 mmolr kg 1 A1 < 0.02 molr m 3 N H / K < 5 mol mol1 N < 1.8 % NO3 < 0.1 molr m 3 NO~ < 0.4-0.8 mole m "3
Deciduous forests
14002) 11002) 11002) 14002) 13002) 12002) 3002) 5002) 1250-50003) 1500-30004) 500-14005) 800-14005) 1700-2900 e) 900-15008)
1) Background information on the various criteria is given in De Vries (1993). Critical A1 and NO8 concentrations and critical A1/Ca and NH4/K ratios related to root damage, inhibition of nutrient uptake and vegetation changes refer to the soil solution. Critical A1 and NOa concentrations related to pollution refer to phreatic groundwater. Critical N contents related to an increased risk for frost damage and diseases refer to the foliage. 2) Derived by a steady-state model. A1 pollution refers to phreatic groundwater. For ground water used for the preparation of drinking water, a critical acid load of 1600 tool c ha 1 yr 1 was derived (cf De Vries 1993). 3) Derived by a steady-state model assuming to nitrification (first value; worst case) and 50% nitrification in the mineral topsoil (second value). 4) Empirical data on the relation between N deposition and foliar N contents. 5) The first value is derived by a steady-state model ('worst case') and the second value is based on empirical data. 8) Derived by a steady-state model using a critical NO3 concentration of 0.4 and 0.8 mole m 3 respectively. NO3 pollution refers to phreatric groundwater. For deep groundwater, the critical load will be higher because of denitrification. g r a s s l a n d s . It is t h e i n c r e a s e in N a v a i l a b i l i t y t h r o u g h e n h a n c e d N c y c l i n g t h a t t r i g g e r s t h e v e g e t a t i o n c h a n g e s ( B e r e n d s e e t al., 1987). T h e s e c o n d v a l u e in T a b l e 6 b a s e d on e m p i r i c a l d a t a is t h i s l i k e l y to b e m o r e r e l i a b l e .
Observed effects T h e p a s t d e c a d e of a c i d i f i c a t i o n r e s e a r c h in D u t c h f o r e s t s h a s u n e q u i v o c a l l y s h o w n t h a t acid a t m o s p h e r i c d e p o s i t i o n c a u s e s l a r g e c h a n g e s in t h e c h e m i c a l c o m p o s i t i o n of foliage, (soil) s o l u t i o n a n d g r o u n d w a t e r and in the u n d e r s t o r e y of f o r e s t s . C r i t i c a l l o a d s r e l a t e d to t h e s e effects (cf T a b l e 6) a r e s t r o n g l y e x c e e d e d ( t h e a v e r a g e acid l o a d on D u t c h f o r e s t s e x c e e d s 4 0 0 0 molo h a 1 y r 1) a n d t h i s d o e s h a v e c l e a r i m p a c t s on t h e f o r e s t e c o s y s t e m (cf T a b l e 7). E f f e c t s of e l e v a t e d S a n d N d e p o s i t i o n on soil ( s o l u t i o n ) a n d g r o u n d w a t e r c h e m i s t r y a r e m o s t e v i d e n t . F i e l d s t u d i e s s h o w e d t h a t SO4 b e h a v e s c o n s e r v a tive in D u t c h f o r e s t soils, w h e r e a s N is l a r g e l y r e t a i n e d . D e s p i t e t h e h i g h N
274 Table 7 Possible effects when critical loads are exceeded and observed effects in Dutch forests.
Possible effects
Average Observed effects in Critical load the field (mole ha" yr 1)
Rootdamage
1100-1400 ')
Inhibition of uptake
1100-1400 ') 1250-50002) 1200-1300 ') 300-500 ')
A1 depletion Groundwater pollution
900-29002) Increased susceptibility Vegetation changes Defoliation
1500-30002) 500-14002)
Large exceedances of critical A1 concentrations Large exceedances of critical Al/Ca ratios Small exceedances of critical NH4/K ratios Depletion of secondary A1 compounds Large exceedances of critical A1 concentrations Substantial exceedances of critical NO3 concentrations Substantial exceedances of critical N contents; Nutrient imbalances; Increased shoot/root ratios Strong increase in nitrophilous species Decline in past ten years, but no clear relationship with abiotic effects
') Refers to acid loads 2) Refers to nitrogen loads deposition, actual soil acidification, which is m a i n l y m a n i f e s t e d by l e a c h i n g of A1 associated with SO4 and NO3 leaching, is d o m i n a n t l y c a u s e d by S deposition. SO4, NOs and A1 concentrations increase in the order: deciduous forests < pine forests < spruce forests. P r e s u m e d critical A1 c o n c e n t r a t i o n s (0.2 mole m 3) a n d A1/Ca ratios (1.0 mol mol 1) r e l a t e d to effects on roots are mostly (ca. 60 - 80%) exceeded below 20 cm soil depth. C o n c e n t r a t i o n s of A1 a n d NO3 in g r o u n d w a t e r often exceed EC d r i n k i n g w a t e r s t a n d a r d s (80% for A1 and 20% for NO3). The p r e s e n t a t m o s p h e r i c i n p u t has also c a u s e d a decline in the content of readily available secondary A1 compounds, t h a t m a i n l y buffer the acid i n p u t (Wesselink et a1.,1994). This m a y cause a f u r t h e r pH decline of the soil, which in t u r n m a y strongly affect the v i t a l i t y of the forest(Houdijk,1993). The relative small contribution of nitrogen to the acidification of D u t c h forest soils as compared to s u l p h u r does not imply t h a t s u l p h u r h a s a l a r g e r i m p a c t on the v i t a l i t y of Dutch forests since the relation b e t w e e n soil acidification and forest vitality in the field is not very evident. This can be derived from both correlative field r e s e a r c h (cf section 3.3) a n d from l i m i n g e x p e r i m e n t s in s t a n d s of Douglas fir and J a p a n e s e l a r c h (Van den B u r g a n d Olsthoorn, 1994). Most probably, at this m o m e n t the e u t r o p h y i n g i m p a c t of n i t r o g e n is more i m p o r t a n t t h a n the acidifying impact. Firstly, elevated N deposition strongly affects the forest n u t r i e n t status. Field studies showed an increase in N content of more t h a n 50% in the l a s t t h r e e decades. At
275 present, about half of the Dutch forests have foliar N contents exceeding a critical limit related to an increased risk for frost and fungal diseases. Compared to N there is a considerable relative Mg deficiency (observed in 53% of the forests) and a large absolute P deficiency (observed in 63% of the forests). Recent results from fertilization experiments have shown t h a t Douglas fir and J a p a n e s e larch reacted positively on P fertilization. Secondly, a large N input initially stimulates above ground growth but it leads to a stronger sensitivity to drought due to an increase in shoot/root ratio (De Visser,1994). Finally, there is ample circumstantial evidence t h a t elevated N deposition during the last decades has caused large changes in forest undergrowth. Comparison of recent and old vegetation descriptions at h u n d r e d s of sites indicates an increase in nitrophilous species (both mosses and vascular plants) and grasses and a decrease of lichens and ectomycorrhiza mushrooms. Unlike the effects on forest undergrowth, circumstantial field evidence for a relationship between soil acidification and n u t r i e n t imbalances in the soil and the foliage on one hand and the vitality of forests (mainly expressed by defoliation class) on the other hand is lacking. Tree species and st and age explains most of the variation in the defoliation class. S t a n d age m ay indirectly be related to air pollution since the period of exposition increases with stand age. Defoliation increased with an increase in N contents (relative P deficiency) and a decrease in pH, but the explanation of the defoliation class increased only very slightly when these variables were included. Aluminium in the soil solution appeared to have no significant effect on the defoliation class, even though there is ample evidence for its toxic effect in laboratory experiments. This result implies t h a t an exceedance of critical acid loads, based on critical A1/Ca or A1/(Ca + Mg + K) ratios observed in laboratory experiments in relation to effects on root (uptake), do not imply visible effects or even the dieback of forests. However, an exceedance of critical loads does affect the long-term sustainability of forests due to depletion of secondary A1 compounds. This risk increases when the rate at which present loads exceed critical loads is higher and the d u r a t i o n is longer.
References Arnolds, E. 1991. Decline of ectomycorrhizal fungi in The Netherlands. Agriculture Ecosystems Environment 35:209-244. Aronsson, A., 1980. Frost hardiness in Scot s pine. II Hardiness during winter and spring in young trees of different mineral status. Studia Forestalia Suecica 155: 1-27. Berendse, F., Beltman, B., Bobbink, R., Kwant, R. and Schmitz, M.B. 1987. Primary production and nutrient availability in wet heathland ecosystems. Acta Oec./Oecol. Plant.: 265-276.
276 Boumans, L.J.M. en W. Beltman, 1991. Kwaliteit van het bovenste freatische grondwater in de zandgebieden van Nederland onder bos- en heidevelden. Bilthoven, Rijksinstituut voor Volksgezondheid en Milieuhygi~ne, Rapport 724901001, 65 pp. Boumans, L.J.M., 1994. Nitraat in het bovenste grondwater onder natuurgebieden op zandgrond in Nederland door atmosferische stikstof depositie. Bilthoven, Rijksinstituut voor Volksgezondheid en Milieuhygi~ne. Rapport 712300002, 52 pp Boxman, A.W. en H.F.G. Van Dijk, 1988. Het effect van landbouw ammonium deposities op bos- en heidevegetaties. Katholieke Universiteit Nijmegen, 96 pp. CAD-BLB, 1990. Eindrapport commissie advies bosbemesting. Utrecht, CADBLB, Report 1990-11.63 pp. De Visser, P.H.B., 1994. Growth and nutrition of Douglas fir, Scots pine and pedunculate oak in relation to soil acidification. Wageningen, Agricultural University, Ph.D. Thesis, 185 pp. De Vries, I M. 1982. De invloed van luchtverontreiniging/zure neerslag op hogere planten. Rapport RU UtrechffRIN Leersum, 180 p + bijl. De Vries, W., 1993. Average critical loads for nitrogen and sulfur and its use in acidification abatement policy in the Netherlands. Water Air and Soil Poll. 68: 399-434. De Vries, W. and J. Kros, 1989. The long-term impact of acid deposition on the aluminium chemistry of an acid forest soil. In: J. K~im~iri, D.F. Brakke, A. Jenkins, S.A. Norton and R.F. Wright (Eds.), Regional Acidification Models. Geographic Extent and Time Development: 113-128. De Vries, W. and P.C Jansen, 1994. Effects of acid deposition on 150 forest stands in the Netherlands. 3. Input output budgets for sulphur, nitrogen, base cations and aluminium. Wageningen, the Netherlands, DLO Winand Staring Centre, Report 69.3, 58 pp. De Vries, W. and E.E.J.M. Leeters, 1994. Effects of acid deposition on 150 forest stands in the Netherlands. 1. Chemical composition of the humus layer, mineral soil and soil solution. Wageningen, the Netherlands, DLO Winand Staring Centre, Report 69.1. De Vries, W., J. Kros and C. Van der Salm, 1994a. The long-term impact of three emission-deposition scenarios on Dutch forest soils. Water Air and Soil Poll. 75: 1-35. De Vries, W., J.J.M. Van Grinsven, N. Van Breemen, E.E.J.M. Leeters and P.C. Jansen, 1994b. Impacts of acid atmospheric deposition on concentrations and fluxes of solutes in acid sandy forest soils in the Netherlands. Geoderma (accepted). Hendriks, C.M.A., W. De Vries and J. Van den Burg, 1994. Effects of acid deposition on 150 forest stands in the Netherlands. 2. Relationship between forest vitality and the chemical composition of the foliage, humus layer and the soil solution. Wageningen, the Netherlands, DLO Winand Staring Centre, Report 69.2, 55 pp. Hilgen, P (Ed.), 1994. De vitaliteit van het Nederlandse bos 11. Verslag van de landelijke inventarisatie 1994. Utrecht, The Netherlands, Information and knowledge centre for Nature, Forest, Landscape and Wildlife, Report 2, 39 pp. Hommel, P.W.F.M., E.E.J.M. Leeters, P. Mekkink and J.G. Vrielink, 1990. Vegetation changes in the Speulderbos (the Netherlands) during the period 1958-1988. Wageningen, the Netherlands, DLO Winand Staring Centre, Report 23, 9 pp. Houdijk, A.L.F.M., 1993. De invloed van verhoogde aluminium-calcium verhoudingen in aanwezigheid van humuszuur en van de uitputting van de aluminium voorraad in de bodem op de vitaliteit van de Corsicaanse den. Katholieke Universiteit Nijmegen, 51 pp. Leeters, E.E.J.M., Hartholt., W. de Vries and L.J.M. Boumans, 1994. Effects of acid deposition on 150 forest stands in the Netherlands, 4. Assessment of the chemical composition of foliage, mineral soil, soil solution and ground
277
water on a national scale. Wageningen, the Netherlands, DLO Winand Staring Centre, Report 69.4, 163 pp. Roelofs, J.G.M., A.J. Kempers, A.L.F.M. Houdijk and J. Jansen, 1985. The effect of airborne ammonium sulphate on Pinus nigra var. maritima in the Netherlands. Plant and Soil 84: 45-56. Sverdrup, H. and P. Warfvinge, 1993. The effect of soil acidification on the growth of trees, grass and herbs as expressed by the (Ca+Mg+K)/A1 ratio. Reports in Ecology and Environmental Engineering 1993: 2, lund University, Department of Chemical Engineering II, 108 pp. Ulrich, B. und E. Matzner, 1983. Abiotische Folgewirkungen der weitrafimigen Ausbreit ung von Luftverunreinigung. Umweltforschungsplan der Bundesminister des Inneren. Forschungsbericht 10402615, BRD, 221 pp. UN-ECE/CEC, 1992. Forest condition in Europe. CEE-UN/ECE, Brussels, Geneva. 159 pp. Van Breemen, N. and J.M. Verstraten, 1991. Soil acidification and N cycling. In: T. Schneider and G.J. Heij (Eds.), Acidification research in the Netherlands. Final report of the Dutch Priority Programme on Acidification. Studies in Environmental Science 46, Elsevier Science Publishers, Amsterdam, the Netherlands: 289-352. Van Breemen, N., P.A. Burrough, E.J. Velthorst, H.F. Van Dobben, T. De Wit, T.B. De Ridder and H.F.R. Reynders, 1982. Soil acidification from atmospheric ammonium sulfate in forest canopy throughfall. Nature 299: 548-550. Van den Burg, J. en H.P. Kiewiet, 1989. Veebezetting en de naaldsamenstelling van grove den, Douglas en Corsicaanse den in het Peelgebied in de periode 1956 t/m 1988. Een onderzoek naar de betekenis van de veebezetting voor het optreden van bosschade. Wageningen, Instituut voor Bosbouw en Groenbeheer, "De Dorschkamp", Rapport nr. 559, 76 pp. Van den Burg, J. and A.F.M. Olsthoorn, 1994. Het landelijke bemestingsonderzoek in bossen 1986 t/m 1991. Deelrapport 6: Overzicht en bespreking van de resultaten. Wageningen, DLO Instituut voor Bos en Natuurbeheer (IBN-DLO), Rapport 106, 126 pp. Van den Burg, J., P.W. Evers, G.F.P. Martakis, J.P.M. Relou en D.C. Van der Werf, 1988. De conditie en de minerale-voedingstoestand van opstanden van grove den (Pinus silvestris) en Corsicaanse den (Pinus nigra var. Maritima) in de Peel en op de zuidoostelijke Veluwe najaar 1986. Wageningen, Instituut voor Bosbouw en Groenbeheer, "De Dorschkamp", Rapport nr. 519, 66 pp. Van Dam, D. 1990. Atmospheric deposition and nutrient cycling in chalk grassland. PhD thesis, University of Utrecht, the Netherlands 119 pp. Van Dobben, H F. 1993. Vegetation as a monitor for deposition of nitrogen and acidity. University of Utrecht, Ph.D. Thesis, 214 p. Van Dobben, H F., M.J.M.R. Vocks., E, Jansen, en G.M. Dirkse. 1994. Veranderingen in de ondergroei van het Nederlandse dennenbos over de periode 1985-1993. IBN Rapport 085.,37 pp. Van Grinsven, J.J.M., J. Van Minnen and C. Van Heerden, 1991. Effects on growth of Douglas fir. In: T. Schneider and G.J. Heij (Eds.), Acidification research in the Netherlands. Final report of the Dutch Priority Programme on Acidification. Studies in Environmental Science 46, Elsevier Science Publishers, Amsterdam, the Netherlands: 180-190. Wesselink, L.G., 1994. Time trends and mechanisms of soil acidification. Wageningen, agricultural University, Ph.D. Thesis, 129 pp.
This Page Intentionally Left Blank
G.J. Heij and J. W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
Ecological effects of atmospheric ecosystems in Western Europe
deposition
on
279
non-forest
Roland Bobbink and Jan G.M. Roelofs
Department of Ecology, Research group Environmental Biology, University of Nijmegen, Toernooiveld 1, 6525 ED Nijmegen, The Netherlands
A b s t r a c t Mans activities pose a number of threats to natural vegetation. One of the major threats is the increase in atmospheric deposition. In this paper information on the ecological effects of increased atmospheric deposition upon non-forest ecosystems of high nature conservation importance (shallow soft-water lakes, heathlands, matgrass swards, and calcareous grasslands) have been presented. Empirical nitrogen critical loads are discussed and finally summarized.
1. I N T R O D U C T I O N
Most of earths biodiversity is found in (semi-)natural ecosystems, both in aquatic and terrestrial habitats. Mans activities pose a number of threats to the structure and the functioning of these ecosystems, and thus to the natural variety of plant and animal species. One of the major antropogenic threats is the increase in sulphur (SOy) and nitrogen (NH x and NOy) air pollution. S and N pollutants have been shown to acidify ecosystems. Furthermore, N deposition may cause eutrophication, because nitrogen is limiting for plant growth in many of these (semi-)natural ecosystems. The impacts of increased nitrogen deposition upon biological systems are diverse, but the most important effects are: (i) direct toxicity of nitrogen gases and aerosols to individual species; (ii) soil-mediated effects upon vegetation; (iii) increased susceptibility to stress factors and, (iiii) changes in (competitive) relationships between species, resulting in loss of biodiversity. The aim of this paper is to review information of ecological effects of an increase in atmospheric deposition upon West-European non-forest ecosystems of high nature conservation importance (shallow soft-water lakes, heathlands, matgrass swards, and calcareous grasslands).
2. SHALLOW SOFTWATER LAKES In the lowlands of Western Europe many soft waters are found on sandy soils which are poor in calcium carbonate or almost devoid of it. The waters are poorly buffered and the concentrations of calcium in the water layer are very low; they are shallow and fully mixed water bodies, with periodically fluctuating water levels, and are mainly fed by rain water, and thus oligotrophic. These softwater ecosystems are characterized by plant communities from the phytosociological
280 alliance LITTORELLION (Arts 1990; Schamin~e et al. 1992). The stands of these communities are characterized by the presence of rare and endangered (Red-list) isoetids, such as Littorella uniflora, Lobelia dortmanna, Isoetes lacustris, I. echinospora, Echinodorus species, Luronium natans and many other softwater macrophytes. These soft waters are nowadays almost all within nature reserves and have become very rare in western Europe. The effects of air pollutants on these soft waters have been intensively studied in the Netherlands both in field surveys and experimental studies. Field observations in ca. 70 soft waters (with well-developed isoetid vegetation in the 1950s) showed that the waters in which these macrophytes were still a b u n d a n t in the early 1980s, were poorly buffered (alkalinity 50-500 lueq 11), circumneutral (pH=5-6) and very poor in nitrogen (Roelofs 1983; Arts et al. 1990). The softwater sites where these plant species had disappeared, could be divided into two groups. In 12 of the 53 softwater sites eutrophication, resulting from enriched water, seemed to be the cause of the decline. In the second group of lakes and pools (41 out of 53) another development had taken place: the isoetid species were replaced by dense stands of Juncus bulbosus or aquatic mosses such as Sphagnum cuspidatum or Drepanocladus fluitans. This clearly indicates acidification of these soft waters in recent decades, probably caused by enhanced atmospheric deposition. In the same field study it has been shown that the nitrogen levels of the water layer were higher in ecosystems where the natural vegetation had disappeared, compared with ecosystems where the LITTORELLION stands were still present (Roelofs 1983). This strongly suggests the detrimental effects of atmospheric nitrogen deposition in these softwater lakes. A number of ecophysiological studies has revealed the importance of (i) inorganic carbon status of the water as a result of intermediate levels of alkalinity, and, (ii) low nitrogen concentrations, for the growth of the endangered isoetid macrophytes. Furthermore, almost all of the typical softwater plants had a relatively low potential growth rate. Increased acidity and higher concentrations of ammonium in the water layer clearly stimulated the development of Juncus bulbosus and submerged mosses such as S p h a g n u m and Drepanocladus species (Roelofs et al. 1984; Den Hartog 1986). It has also been shown in cultivation experiments that the nitrogen species involved (ammonium or nitrate) differentially influenced the growth of the studied species of water plants. Almost all of the characteristic softwater isoetids developed better with nitrate instead of ammonium addition, whereas Juncus bulbosus and aquatic mosses (Sphagnum & Drepanocladus), were clearly stimulated by ammonium nutrition (Schuurkes et al. 1986). The effects of atmospheric deposition have been studied in small-scale softwater systems during a 2-year treatment with different artificial rainwaters. Acidification, without air-borne nitrogen input (sulphuric acid), has not resulted in a mass growth of Juncus bulbosus and a diverse isoetid vegetation remains present. However, after increasing the nitrogen concentration in the precipitation (as ammonium sulphate), similar changes in floristic composition as under field conditions have been observed: a dramatic increase in dominance of Juncus
281
bulbosus, of submerged aquatic mosses and of Agrostis canina (Schuurkes et al. 1987). In these small-scale ecosystems, the recovery of the small-scale systems have been followed during ca. 8 years. Reduction of the acid load leads to a quick recovery of the acidified systems, whereas the situation in the eutrophied did not improved (personal observ. J.G.M. Roelofs). It became obvious that the observed changes occurred because of the effects of ammonium sulphate deposition, leading to both eutrophication and acidification. The increased levels of ammonium in the system stimulated directly the growth of plants such as Juncus bulbosus, whereas the surplus of the extra ammonium will be nitrified in these waters (pH>4.0). During this nitrification process H § ions are produced, which increase the acidity of the system. The results of this study clearly demonstrated that the changes in composition of the vegetation already occurred after 2-year treatment with >19 kg N ha 1 yr 1. A reliable critical load for nitrogen deposition in these shallow softwater lakes is thus most likely below 19 kg N ha ~ yr ~ and probably between 5 and 10 kg N ha 1 yr 1. This value is supported by the observation that the strongest decline in the species composition of the Dutch LITTORELLION communities has coincided with nitrogen loads of ca. 10-13 kg N ha -1 yr 1 (Arts 1990).
3. H E A T H L A N D S Heathlands have for a long time been part in the West European landscape. Heaths are plant communities where the dominant life-form is that of the smallleaved dwarf-shrubs forming a canopy of 1 m or less above soil surface. Dwarfshrub heathlands are wide spread in the atlantic and sub-atlantic parts of Europe. In these parts of the European continent natural heathland is limited to a narrow coastal zone. Inland lowland heathlands are certainly man-made (seminatural), although they have existed for several centuries. Lowland heaths are widely dominated by some Ericaceae, especially Calluna vulgaris in the dry- and Erica tetralix in the wet-heathlands (Gimingham et al. 1979). Until the beginning of this century, the balance of nutrient input and output was in equilibrium in the lowland heathlands of Western Europe. The original land use implied a regular, periodic removal of nutrients out of the ecosystem via grazing and sod removal (Heil & Aerts 1993). The original land use of the lowland heathland ceased in the early 1900s and the area occupied by this community decreased markedly all over its distribution area (e.g. Ellenberg 1988). Because of their conservational importance, many lowland heathlands have become nature reserves in recent years. In W Europe many lowland dry-heathlands have become dominated by grass species. An evaluation, using aerial photographs, has shown that more t han 35% of Dutch heathland has been altered into grassland (Van Kootwijk & Van der Voet 1989). It has been suggested that nitrogen eutrophication might be a significant factor in this transition to grasslands. Field and laboratory experiments affirm the importance of nutrients, especially in the early phase of heathland development (Heil & Diemont 1983; Roelofs 1986; Heil & Bruggink
282 1987; Aerts et al. 1990). However, Calluna can compete successfully with the grasses even at high nitrogen loading, if its canopy remains closed (Aerts et al. 1990). Apart from the changes in competitive interactions between Calluna and the grasses, heather beetle plagues, and nitrogen accumulation in the soil are important factors in the changing lowland heaths. Furthermore, evidence is growing that frost and drought sensitivity of the dominant dwarf-shrubs may also be affected by increasing nitrogen inputs. In Calluna heathland outbreaks of the chrysomelid heather beetle (Lochmaea suturalis) occur frequently. The beetles feed exclusively on the green parts of Calluna and the closed Calluna canopy is opened over large areas and the interception of light by Calluna decreases strongly (Berdowski 1987; 1993). Thus the growth of the understorey grasses (Deschampsia or Molinia) will be enhanced significantly. In general (insect) grazing is affected by the nutritive value of the plant material. The nitrogen content is especially important in this respect. Experimental applications of nitrogen to heathland vegetation have demonstrated that the concentrations of this element in the (green) parts of Calluna increased (Heil & Bruggink 1987; Bobbink & Heil 1993). It is, therefore, very likely, that the frequency and intensity of the heather beetle outbreaks are stimulated by the increased atmospheric nitrogen input in Dutch heathland. This hypothesis is supported by the observations of Blankwaardt (1977); he reported that from 1915 onwards heather beetle outbreaks have been observed in the Netherlands with an interval of ca. 20 years, whereas in the last 15 years the outbreaks occur within periods of less than 8 years. Furthermore, it has also been observed that during a heather beetle outbreak Calluna plants were more severely damaged in nitrogen-fertilized vegetation (Hell & Diemont 1983). Brunsting and Heil (1985) have done a rearing experiment with larvae of the heather beetle and demonstrated that the growth of the larvae of the heather beetle increased by higher leave nitrogen concentrations of Calluna. Van der Eerden et al. (1990) found no significant effect of ammonium sulphate treatments on total number and on biomass of the 1st stage larvae after an beetle outbreak. However, the development of subsequent larval stages has been accelerated by the application of ammonium sulphate in the artificial rain. Furthermore, heather beetle larvae had been put on Calluna shoots taken from plants which has been fumigated with ammonia in open top chambers (12 months; 4 to 105 l~g m 3) (Van der Eerden et al. 1991). Both the mass and the development rate of the larvae clearly increased with increasing concentrations of ammonia prior applied to Calluna. It can be concluded that nitrogen inputs influence the outbreaks of heather beetle, although the exact relationship between both processes needs further research. Nowadays most Dutch heathlands are managed by mechanical 'sod removal'. It is likely that changes in the rate of nitrogen accumulation during secondary heathland succession will occur due to the increased nitrogen deposition. Berendse (1990) found a large increase in total nitrogen storage in the first 20/30 years after sod removal. Furthermore, he demonstrated that nitrogen mineralization was low in the first 10 years (ca. 10 kg N ha 1 yr 1) after sod removal, but strongly increased in the next 20 years to 50-110 kg N ha ~ yr a.
283 Thus, the total organic matter and nitrogen amounts increased, as usual, during secondary succession after sod removal. However, this process is accelerated by the enhanced dry matter production and litter production of the dwarf-shrubs caused by the extra nitrogen inputs. Hardly any nitrogen disappeared from the system because nitrate leaching to deeper layers is only of minor importance in Dutch heathlands (De Boer 1989; Van Der Maas 1990). Nitrogen availability from atmospheric inputs, in addition to mineralization, is within a relatively short period of 10 years high enough to stimulate the transition of heathland to grassland, especially after the opening of the heather canopy by secondary causes. It has been demonstrated that frost sensitivity in some tree species increased with increasing concentrations of air pollutants. This increase in frost sensitivity is sometimes correlated with the enhanced nitrogen concentrations in the foliage of the trees. Long-term effects of air pollutants on the frost sensitivity of CaUuna and Erica may be expected, because of (i) the evergreen growth form of these species and, (ii) the increasing contents of nitrogen in the leaves of Calluna, associated with increased nitrogen deposition in the Netherlands. It is suggested that damage of the Calluna shoots in the successive severe winters of the mid1980s is at least partly caused by the increased frost sensitivity. After fumigation with sulphur dioxide (90 ~g m3; 3 months) increased frost injury in Calluna was only found at temperatures which hardly occur in the Netherlands (< -20 ~ (Van der Eerden et al. 1990). Fumigation with ammonia of Calluna plants in Open Top Chambers during 4-7 month periods (100 ~g m 3) revealed that frost sensitivity was not affected in autumn (September or November), whereas in February, just before growth started, frost injury increased significantly at -12 ~ (Van der Eerden et al. 1991). They also studied the frost sensitivity in Calluna vegetation which was artificially sprayed with 6 different levels of ammonium sulphate (3-91 kg N ha 1 yrl). The frost sensitivity of Calluna increased slightly, although significantly, after 5 months in vegetation treated with the highest level of ammonium sulphate (400 ~mol 11; 91 kg N ha -1 yrl). However, frost sensitivity of Calluna decreased again two months later and no significant effects have been measured at that time. Thus, high levels of ammonia or ammonium sulphate seem to increase the frost sensitivity of Calluna plants, although the significance of this phenomenom is still uncertain at ambient nitrogen inputs. It is shown that atmospheric nitrogen is the trigger for the changes of lowland dry-heathlands into grass swards in the Netherlands. A dynamic ecosystem simulation model has been used which integrated processes, such as atmospheric nitrogen input, heather beetle outbreaks, soil nitrogen accumulation, sod removal and competition between species, to establish nitrogen critical load in lowland dry-heathlands (Heil & Bobbink 1993). The model has been calibrated with data from field and laboratory experiments in the Netherlands. As an indicator of the effects of atmospheric nitrogen the proportion and increase of grasses in the heathland system are used. Atmospheric nitrogen deposition has been varied between 5 and 75 kg N ha 1 yr 1 in steps of 5-10 kg N during different
284
simulations. From these simulations it became obvious that the nitrogen critical load for the changes from dwarf-shrubs to grasses is 15-20 kg N ha 1 yr 1 (Fig. 1). 20 kg N ha "1 yr "1
1 0 k g N h a "1 ),r "1
100 80 70 60 50 40 C:D ffi 5:1 30 20 100 -9 0
C:D
=
1:1
2
4
6
8
10
12
14
16
18
20
22
24
26
2
100 90 80 70 60 5O 4O
100
3O 20 10 0
30 2O 10 0
4
6
,-
8O 70 6O 5O 4O
4
6
8
10
12
14
16
18
20
22
24
26
0
10
/
''
9O
2
8
2
4
6
8
12
14
16
18
20
22
24
26
,
10
12
14
16
18
. . . . . . . . 20 22 24 26
Year
Fig. 1. Model results of interaction between Calluna vulgaris and Deschampsia flexuosa at two levels of atmospheric nitrogen deposition and two initial ratios of both species (C:D). The sudden reduction of Calluna cover is due to heather beetle attacks (adapted with permission from Heil & Bobbink 1993a). Closed circles: Calluna; open circles: Deschampsia.
4. M A T G R A S S S W A R D S & S P E C I E S - R I C H W E T H E A T H L A N D S In recent decades, besides the transition from dwarf-shrub dominated to grass dominated heathlands, a reduced species diversity in these ecosystems has been observed. Species of the acidic NARDETALIA grasslands and the related dryand w e t - h e a t h l a n d s (CALLUNO-GENISTION and ERICION TETRALICES) seem to be especially sensitive. Many of these herbaceous species (e.g. Arnica
montana, Antennaria dioica, Dactylorhiza maculata, Gentiana pneumonanthe, Genista pilosa, Genista tinctoria, Lycopodium inundatum, Narthecium ossifragum, Pedicularis sylvatica, Polygala serpyllifolia and Thymus serpyllum) are declining or have even become locally extinct in the Netherlands. The distribution of these species is related to small-scale, spatial variability of the heathland soils. It is suggested that atmospheric deposition has caused such drastic abiotic changes of these species that they can not survive (Van D a m et al. 1986). Dwarf-shrubs as well as grass species are nowadays dominant in former habitats of these endangered species.
285 Enhanced nitrogen fluxes onto the nutrient-poor heathland soils lead to an increased nitrogen availability in the soil. However, most of the deposited nitrogen in W Europe originates from ammonia/ammonium deposition and may also cause acidification as a result of nitrification. Whether eutrophication or acidification or a combination of both processes is important, depends on pH, buffer capacity and nitrification rates of the soil. Roelofs et al. (1985) found t h a t in dwarf-shrub dominated heathland soils nitrification has been inhibited at pH 4.0-4.2, and th at ammonium accumulated while nitrate decreased to almost zero at these or lower pH values. Furthermore, nitrification has been observed in the soils from the habitats of the endangered species, due to its somewhat higher pH and higher buffer capacity. In soils within the pH range of 4.1-5.9, the produced acidity is buffered by cation exchange processes (Ulrich 1983). The pH will drop when calcium is depleted and this may cause the decline of those species t h a t are generally found on soils with somewhat higher pH. To study the pH effects on root growth and survival rate, hydroculture experiments have been done over 4-week periods with several of the endangered species (Arnica, Antennaria, Viola, Hieracium pilosella and Gentiana) and with the dominant species (Molinia and Deschampsia) (Van Dobben 1991). The dominant species indeed have a lower pH optimum (3.5 and 4.0, respectively) than the endangered species (4.2-6.0). However, the endangered species could survive very low pH without visible injuries during this short experimental period. The pH decrease may indirectly result in an increased leaching of base cations, increased aluminium mobilization and thus enhanced A1/Ca ratios of the soil (Van Breemen et al. 1982). Furthermore, the reduction of the soil pH may inhibit nitrification and result in ammonium accumulation and consequently increased NHt/NO 3 ratios. In a recent field study the characteristics of the soil of several of these threatened heathland species have been compared with the soil characteristics of the dominant species. Generally the endangered species grow on soils with higher pH, lower nitrogen content, and lower A1/Ca ratios t h a n the dominant species. The NHt/NO 3 ratios were higher in the dwarf-shrub dominated soils compared with the ratios in the soil of the endangered species. Fennema (1990; 1992) has demonstrated that soils from locations where Arnica is still present, had higher pH and lower A1/Ca ratios than soils of former Arnica stands. However, he found no differences in total soil nitrogen and NHt/NO 3 ratios. Both these studies indicate that high A1/Ca ratios or even increased NH4/NO 3 ratios are associated with the decline of these species. However, no significant effects of A1 and A1/Ca on growth rates have been observed in hydroculture experiments at high nutrient levels in which the effects of A1 and A1/Ca ratios on root growth and survival rate were studied (Pegtel 1987; Kroeze et al. 1990; Van Dobben 1991). However, results of a hydroculture experiment with Arnica showed that this species is very sensitive to enhanced A1/Ca ratios at intermediate or low nutrient levels, whereas another Red-list species (Cirsium dissectum) is also very sensitive for high ammonium concentrations and high NH4/NO 3 ratios (Fig. 2) (De Graaf et al. 1994). Pot experiments with acidic heathland soil have indicated that increased NH4/NO3 ratios, because of ammonium accumulation, have caused a decreased
286 vitality of T h y m u s . Only in artificially buffered soils, nitrification rates were high enough to balance ammonium and nitrate. T h y m u s plants on these soils were vital despite high nitrogen applications (Houdijk et al. 1993). Hydroculture experiments with this plant species confirmed t h a t increased NH4/NO3 ratios affected the cation uptake (Houdijk 1993). At present, however, there is too little information available on these rare h e a t h l a n d and acidic grassland species to formulate a critical load for nitrogen. The observation t h a t these species mostly disappear before dwarf-shrubs are replaced by grasses, leads to the assumption t h a t their critical load is lower t h a n the critical load for the transition to grasses (thus < 15-20 kg N ha 1 yrl).
Arnica
Cirsium
250
125
200
1O0
150
75
L--
"1o o} E
L
"t3 100
-7
E
50
50 25
0 0/100
100/1 oo 500/1 oo 100/1000500/5000
AI/Ca ratio
!
0
_
01100
100/1 oo
1000/100
N H4/NO3
100/0
ratio
Fig. 2. Plant dry weight of Arnica montana (A) aider cultivation on water cultures with different A1/Ca ratios and of Cirsium dissectum (B) on cultures with different ammonium/nitrate ratios. The ratios are given in ~mol 11 (adapted with permission from De Graaf et al. 1994).
5. C A L C A R E O U S G R A S S L A N D S Calcareous grasslands are communities on limestone, which is wide spread in the hilly and mountainous regions of Western and Central Europe. Subsoils consist of different kinds of limestone with high contents of calcium carbonate (> 90%), covered by shallow well-buffered rendzina soils (A/C-profiles; pH of the top soil: 7-8 with calcium carbonate content of ca. 10%). The depth of the soil varies between 10-50 cm and the availability of nitrogen and phosphorus is low. A large part of the E u r o p e a n calcareous grasslands are MESOBROMION communities; a grassland type found in areas with precipitation quantities between 500-900 m m per year (Willems 1982). Plant productivity is low and peak standing crop is in general between 150-400 g m 2. Calcareous grasslands are among the most species-rich plant communities in Europe and contain a large n u m b e r of rare and endangered species. These semi-natural grasslands decreased strongly in a r e a during the second half of this century (e.g. Wolkinger & P l a n k 1981). Some
287 remnants became nature reserve in several European countries. To maintain the characteristic calcareous vegetation a specific management is needed to prevent their natural succession towards woodland (e.g. Ellenberg 1988). The effects of nitrogen enrichment in Dutch calcareous grasslands on vegetation composition have been investigated in field experiments (Bobbink et al. 1988, Bobbink 1991). Either potassium (100 kg K ha 1 yrl), phosphorus (30 kg P ha ~ yr ~) or nitrogen (100 kg N ha a yr ~) as well as a complete fertilization (N+P+K) have been applied during 3 years to study 'long-term' effects on vegetation composition. Total above-ground biomass increased considerably, as expected, after three years of N+P+K fertilization. In the separate application of nutrients, a moderate increase in above-ground dry weight was only seen with nitrogen addition: ca. 330 g m 2 compared with ca. 210 g m 2 in the untreated plots. Dry weight distribution of the species was dramactically affected by nutrient treatments. In the N-treated vegetation the dry weight of the grass species Brachypodium pinnatum was ca. 3 times higher than in the control plots (Fig. 3). Nitrogen application resulted furthermore in a drastic reduction of the biomass of forb species (including several Dutch Red List species) and of the total number of species (Fig. 3). The observed decrease in species diversity in the nitrogen treated vegetation is certainly not caused by nitrogen toxicity, but by the change in vertical structure of the grassland vegetation. The growth of Brachypodium is strongly stimulated and its overtopping leaves reduce the light quantity and quality in the vegetation. It overshadowed the other characteristic species and growth of these species declined rapidly (Bobbink et al. 1988; Bobbink 1991). The effects of excess nitrogen supply on the massive expansion of Brachypodium and a drastic reduction in species diversity have also been observed in a long-term permanent plot study using a factorial design (Willems et al. 1993.)
DRY
WEIGHT
BRACHYI::::)ODIUM 99
350"
I
300 250
SPECIES
NUMBER
30 25
I
20
200 15 150 10
100
84
5-"!
50 0 L
0 0
K
P N Nutrient treatment
NPK
O
K
P N Nutrient treatment
NPK
Fig. 3. Above-ground biomass of Brachypodium (g m2) and phanerogamic species number (per 50x50 cm) after 3 years of nutrient additions (adapted with permission from Bobbink 1991). (*): p<0.10; **: p<0.01.
288 With the repeated harvest approach it has been shown that B r a c h y p o d i u m had a very efficient nitrogen acquisition and a very efficient withdrawal from its senescent shoots into its well-developed rhizome system (Bobbink et al. 1989). It thus benefits from the extra nitrogen redistributed to the below-ground rhizomes by enhanced growth in the next spring. After 3-year of nitrogen addition B r a c h y p o d i u m had strongly monopolized (>75%) the nitrogen storage in both the above-ground and below-ground compartments of the vegetation with increasing nitrogen availability (Bobbink et al. 1988; Bobbink 1991). Nitrogen cycling and accumulation in calcareous grassland can be significantly influenced by two major outputs from the system: (i) leaching from the soil, and (ii) removal with management regimes. Nitrogen losses by denitrification in dry calcareous grasslands are low (<1 kg N ha ~ yr ~) (Mosier et al. 1981). Ammonium and nitrate leaching has been studied in Dutch calcareous grasslands by Van Dam et al. (1992). Both the water fluxes and the solute fluxes at different soil depths have been measured over two years in untreated vegetation and in calcareous grassland vegetation sprayed with ammonium sulphate (50 kg N ha 1 yr-~). The observed nitrogen leaching from the untreated vegetation is very low (0.7 kg N ha ~ yrl), and only 2% of the total atmospheric deposition of N. After two-weekly spraying of ammonium sulphate, nitrogen leaching has significantly increased to 3.5 kg N ha ~ yr ~, although this figure is also a very small proportion (4%) of the nitrogen input (Van Dam et al. 1992). It is thus evident that calcareous grassland ecosystems almost completely retain nitrogen in the system. This is caused by a combination of enhanced plant uptake (Bobbink et al. 1988; Bobbink 1991) and increased immobilization in the soil organic matter (Van Dam et al. 1992). The most important output of nitrogen from the calcareous grassland is by exploitation or management. From the 1950s onwards almost all of the calcareous grasslands in the Netherlands were mown in autumn with removal of the hay. The annual nitrogen removal in the hay varies slightly between years and sites, but in general between 17-22 kg N ha -1 is removed from the system with the usual management (Bobbink 1991; Bobbink & Willems 1991). The ambient nitrogen deposition in Dutch calcareous grasslands, as determined by Van Dam (1990), is high (35-40 kg N ha 1 yr -1) and is nowadays the major nitrogen input to the system. Legume species ( L e g u m i n o s a e ) also occur in calcareous vegetation, and form an additional nitrogen input with the nitrogenfixing microorganisms in their root nodules (ca. 5 kg N ha ~ yr-~). The nitrogen mass balance of Dutch calcareous grasslands is summarized in Table 1. It is obvious that calcareous grasslands nowadays significantly accumulate nitrogen (16-26 kg N ha ~ yr 1) in the Netherlands. A critical nitrogen load can be determined with a steady-state mass balance model (e.g. De Vries 1994). Assuming a critical long-term immobilization rate for N of 0-6 kg N ha 1 yr ~, the critical nitrogen load can be derived by adding the N fluxes due to net uptake, denitrification and leaching, corrected for the N input by fixation. Thus, 14-25 kg N ha 1 yr ~ is considered as nitrogen critical load for this system.
289 T a b l e 1. Nitrogen mass balance (kg N ha -1 yr 1) for dry calcareous grassland in the Netherlands.
INPUT Atmospheric deposition Nitrogen fixation
OUTPUT 35-40 5
Harvest Denitrification Soil leaching
17-22 1 1
6. CONCLUSION: E M P I R I C A L N I T R O G E N CRITICAL LOADS In this paper the effects of nitrogen deposition on (semi-)natural non-forest ecosystems are evaluated with published evidence. Empirical critical loads for excess nitrogen deposition are summarized in Table 2.
T a b l e 2. Summary of nitrogen critical loads
(kg N ha 1 yr 1) to non-forest ecosystems. ## reliable; # quite reliable and (#) best guess (after Bobbink et al. 1992; Bobbink & Roelofs 1995). Critical load
Indication
Shallow soft-water lakes 5-10 ## Mesotrophic fens 20-35 # Ombrotrophic bogs 5-10 #
Decline isoetid species Increase tall graminoids, decl. diversity Decrease Sphagnum and subordinate species, increase tall graminoids
Calcareous grassland Neutral-acid grassland
14-25 ## 20-30 #
Increase tall grass, decline diversity Increase tall grass, decline diversity
Lowland dry-heathland Lowland wet-heathland
15-20 ## 17-22 ##
Transition heather to grass Transition heather to grass
7. R E F E R E N C E S
Aerts, R. Berendse, F. De Caluwe, H. & Schmitz, M. 1990. Competiton in heathland along an experimental gradiet of nutrient availability. Oikos 57: 310318. Arts, G.H.P. 1990. Deterioration of atlantic soft-water systems and their flora, a historical account. PhD thesis, University of Nijmegen. Arts, G.H.P., Van Der Velde, G., Roelofs, J.G.M. & Vanv Swaay, C.A.M. 1990. Successional changes in the soft-water macrophyte vegetation of (sub)atlantic, sandy, lowland regions during this century. Freshwater Biol. 24: 287-294.
290 Berdowski, J.J.M. 1987. The catastrophic death of Calluna vulgaris in Dutch heathlands. PhD thesis, University of Utrecht. Berdowski, J.J.M. 1993. The effect of external stress and disturbance factors on Calluna-dominated heathland vegetation. In: Aerts, R. & Heil, G.W. (Eds.), Heathland: Patterns and Processes in a changing environment. Geobotany 20, Kluwer, Dordrecht, pp. 85-124. Berendse, F. 1990. Organic matter accumulation and nitrogen mineralization during secondary succession in heathland ecosystems. J. Ecol. 78: 413-427. Blankwaardt, H.F.H. 1977. Het optreden van de heidekever (Lochmaea suturalis Thomson) in Nederland sedert 1915. Entomologische Berichten. 37" 34-40. (In Dutch). Bobbink, R. 1991. Effects of nutrient enrichment in Dutch chalk grassland. J. Appl. Ecol. 28: 28-41. Bobbink, R., Bik, L. & Willems, J.H. 1988. Effects of nitrogen fertilization on vegetation structure and dominance of Brachypodium pinnatum (L.) Beauv. in chalk grassland. Acta Bot. Neerl. 37: 231-242. Bobbink, R., Den Dubbelden, K.C. & Willems, J.H. 1989. Seasonal dynamics of phytomass and nutrients in chalk grassland. Oikos 55: 216-224. Bobbink, R. & Willems, J.H. 1991. Impact of different cutting regimes on the performance of Brachypodium pinnatum in Dutch chalk grassland. Biol. Conserv. 56: 1-21. Bobbink, R., Boxman, D., Fremstad, E., Heil, G., Houdijk, A. & Roelofs, J. (1992). Critical loads for nitrogen eutrophication of terrestrial and wetland ecosystems based upon changes in vegetation and fauna. In: Grennfelt, P. & Th6rnel6f, E. (Eds.), Critical loads for nitrogen. Nord (MiljSrapport) 41, 111-159. Nordic Council of Ministers, Copenhagen. Bobbink, R. & Heil, G.W. 1993. Atmospheric deposition of sulphur and nitrogen in heathland ecosystems. In: Aerts, R. & Heil, G.W., Heathland: Patterns and Processes in a changing environment. Geobotany 20, Kluwer, Dordrecht, pp.2550. Bobbink, R. & Roelofs, J.G.M. (1955). Empirical nitrogen critical loads: update since LSkeberg (1992). In: Sutton, M.A. (ed.). Critical loads for nitrogen. A workshop report of the UN/ECE meeting in Grange-over-Sands, xx-xx. UK Ministery of the Environment. Brunsting, A.M.H. & Heil G.W. 1985. The role of nutrients in the interaction between a herbivorous beetle and some competing plant species in heathlands. Oikos 44: 23-26. De Boer, W. 1989. Nitrification in Dutch heathland soils. PhD thesis, Agricultural University of Wageningen. De Graaf, M.C.C., Verbeek P.J.M., Cals, M.J.R. & Roelofs, J.G.M. 1994. Effectgerichte Maatregelen in matig mineraalrijke heide en schraallanden. Eindrapport monitoringsrapport eerste fase. Vakgroep Oecologie, Katholieke Universiteit Nijmegen. Den Hartog, C. 1986. The effects of acid and ammonium deposition on aquatic vegetations in the Netherlands. In: Proceedings 1st. Internat. Symposium on water milfoil (Myriophyllum spicatum) and related Haloragaceae species. Vancouver, Canada, pp. 51-58.
291 De Vries, W. 1994. Soil response to acid deposition at different regional scales. PhD Thesis, Agricultural University of Wageningen. Ellenberg, H. 1988. Vegetation ecology of Central Europe. Cambridge Univ. Press, Cambridge. Fennema, F. 1990. Effects of exposure to atmospheric SO2, NH 3 and (NH4)2SO4 on survival and extinction of Arnica montana L. and Viola canina L.. RIN, Arnhem, The Netherlands. Report no. 90/14, 1-61. Fennema, F. 1992. SO 2 and NH 3 deposition as possible causes for the extinction of Arnica montana L. Water Air Soil Poll. 62: 325-336. Gimingham, C.H., Chapman, S.B. & Webb, N.R. 1979. European heathlands. In: Specht, R.L. (ed.) Ecosystems of the world, 9A. Elsevier, Amsterdam, pp.365-386. Heil, G.W. & Diemont, W.H. 1983. Raised nutrient levels change heathland into grassland. Vegetatio 53: 113-120. Heil, G.W. & Bruggink, M. 1987. Competition for nutrients between Calluna vulgaris (L.) Hull and Molinia caerulea (L.) Moench. Oecologia 73" 105-108. Heil, G.W. & Aerts, R. 1993b. General introduction. In: Aerts, R. & Hell, G.W. , Heathland: Patterns and Processes in a changing environment. Geobotany 20, Kluwer, Dordrecht, pp. 1-24. Heil, G.W. & Bobbink, R. 1993. Impact of atmospheric nitrogen deposition on dry heathlands: a stochastic model simulating competition between Calluna vulgaris and two grass species. In: Aerts, R. & Hell, G.W., Heathland: Patterns and Processes in a changing environment. Geobotany 20, Kluwer, Dordrecht, pp. 181200. Houdijk, A.L.F.M. 1993. Atmospheric ammonium deposition and the nutritional balance of terrestrial ecosystems. PhD Thesis, University of Nijmegen. Houdijk, A.L.F.M., Verbeek, P.J.M., Van Dijk, H.F.G. & Roelofs, J.G.M. 1993. Distribution and decline of endangered herbaceous heathland species in relation to the chemical composition of the soil. Plant Soil 148: 137-143. Kroeze, C., Pegtel, D.M. & Blom, C.J.C. 1989. An experimental comparison of aluminium and manganese susceptibility in Antennaria dioica, Viola canina, Filago minima and Deschampsia flexuosa. Acta Bot. Neerl. 38: 165-172. Mosier, A.R. Stillwel, M. Paton, W.J. & Woodmansee, R.G. 1981. Nitrous oxide emissions from a native shortgrass prairie. Soil Sci. Soc. Am. J. 45: 617-619. Moss, B. 1988. Ecology of fresh waters - Man and Medium. Blackwell, Oxford. Pegtel, D.M. 1987. Effects of ionic A1 in culture solutions on the growth of Arnica montana L. and Deschampsia flexuosa (L.) Trin. Plant Soil 102:85-92. Roelofs, J.G.M. 1983. Impact of acidification and eutrophication on macrophyte coomunities in soft waters in the Netherlands. I. Field observations. Aquat. Bot. 17: 139-155. Roelofs, J.G.M. 1986. The effect of airborne sulphur and nitrogen deposition on aquatic and terrestrial heathland vegetation. Experientia 42: 372-377. Roelofs, J.G.M., Schuurkens, J.A.A.R. & Smits, A.J.M. 1984. Impact of acidification and eutrophication on macrophyte coomunities in soft waters in the Netherlands. II. Experimental studies Aquat. Bot. 18: 389-411. Roelofs, J.G.M., Kempers, A.J., Houdijk, A.L.F.M. & Jansen, J. 1985. The effect of air-borne ammonium on Pinus nigra var. maritima in the Netherlands. Plant Soil 84: 45-56.
292 Schamin~e, J.H.J., Westhoff, V. & A r t s , G.H.P. 1992. Die Strandlinggesellschaften (Littorelletea Br.-B1. et Tx. 43) der Niederlande, in europ~iische Rahmen gefasst. Phytocoenologia 20: 529-558. Schuurkens, J.A.A.R., Kok, C.J. & Den Hartog, C. 1986. Ammonium and nitrate uptake by aquatic plants from poorly buffered and acidified waters. Aquat. Bot. 24: 131-146. Schuurkes, J.A.A.R. Elbers, M.A., Gudden, J.J.F. & Roelofs, J.G.M. 1987. Effects of simulated ammonium sulphate and sulphuric acid rain on acidification, water quality and flora of small-scale soft water systems. Aquat. Bot. 28: 199-225. Ulrich, B. 1983. Interaction of forest canopies with atmospheric constituents: SO2, alkali and earth alkali cations and chloride. In: Ulrich, B. & Pankrath, J. (eds.) Effects of accumulation of air pollutants in forest ecosystems. D. Reidel Publ., Dordrecht, pp.33-45. Van Breemen, N., Burrough, P.A., Velthorst, E.J., Dobben, H.F. van, Wit, T. de, Ridder, T.B. & Reijnders H.F.R. 1982. Soil acidification from atmospheric ammonium sulphate in forest canopy throughfall. Nature 299: 548-550. Van Dam, D. 1990. Atmospheric deposition and nutrient cycling in chalk grassland. PhD thesis, University of Utrecht. Van Dam, D. Van Dobben, H.F. Ter Braak, C.F.J. & De Wit, T. 1986. Air pollution as a possible cause for the decline of some phanerogamic species in The Netherlands. Vegetatio. 65: 47-52. Van Dam, D. Heil, G.W. Bobbink, R. & Heijne, B. 1992. Impact of atmospheric deposition on nutrient cycling in chalk grassland: throughfall, canopy exchange, nitrogen turnover and input/output budgets. Oecologia. Van Der Eerden L.J. Dueck, Th.A. Elderson, J. Van Dobben, H.F. Berdowski, J.J.M. & Latuhihin M. 1990. Effects of NH 3 and (NH4)2SO4 deposition on terrestrial semi-natural vegetation on nutrient-poor soils. Report IPO/RIN. Van Der Eerden, L.J. Dueck, Th. A. Berdowski, J.J.M. & Greven, H. & Van Dobben, H.F. 1991. Influence of NH 3 and (NH4)2SO4 on heathland vegetation. Acta Bot. Neerl. 40: 281-297. Van Der Maas, M.P. 1990. Hydrochemistry of two douglas fir ecosystems and a heather ecosystem in the Veluwe, The Netherlands. Report Agricultural University of Wageningen. Van Dobben, H.F. 1991. Integrated effects (Low vegeatation). In: Heij, G.J. & Schneider, T., Acidification research in the Netherlands Final report of the Dutch Priority Programme on Acidification. Elsevier, Amsterdam, pp 464-524. Van Kootwijk, E.J. & Van der Voet, H. 1989. De kartering van heidevergrassing in Nederland met de Landsat Thematic Mapper sattelietbeelden. Report RIN 89/2, Arnhem (In Dutch). Willems, J.H. 1982. Phytosociological and geographical survey of Mesobromion communities in Western Europe. Vegetatio 48: 227-240. Willems, J.H., Peet, R.K. & Bik, L. 1993. Changes in chalk-grassland structure and species richness resulting from selective nutrient additions. J. Veg. Science 4: 203-212. Wolkinger, F. & Plank, S. 1981. Dry grasslands of Europe. Council of Europe, Strasbourg.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
293
Evaluation; integration H.F. van Dobben DLO Institute for Forest and Nature Research, P.O.B. 23, 6700 AA Wageningen, Netherlands
Introduction Since the early 1980s the ecological effects of atmospheric deposition have become the subject of intense research activities. In this research the main focus of interest was forest, with emphasis on soil chemistry and tree physiology. The choice for this subject was primarily inspired by the work of Ulrich and co-workers (Ulrich et al. 1979, Ulrich 1983). Their publications resulted in a broad public concern over forest decline, in analogy with the fish decline in lakes and streams, which was a known result of acid deposition. Van Breemen et al. (1982) observed a nearly complete nitrification of atmospheric ammonia in acid forest soil, resulting in an internal generation of acidity. This publication strongly stimulated research in the field of forest soil chemistry. The past decade of research has unequivocally shown a strong relationship between atmospheric deposition and a large number of biological and chemical changes in forest ecosystems. These deposition-related changes are described in detail in the preceding papers. Some of the highlights are summarized below for the two main topics: the soil and the trees. Effects of atmospheric deposition on forest soil The most important results can be summarized in three points: In spite of the high deposition, nitrogen is still largely retained in most forest soils; for sulphur, however, leaching equals deposition (De Vries 1994). In Dutch forests, soil solution is dominated by SO4, NO3 and AI. Absolute concentrations of these ions strongly depend upon the tree species, and decrease in the order spruce forest (Norway spruce, Douglas fir) > pine forest (Scots pine, black pine) > deciduous forest (oak, beech). Leaching of NH4
294
hardly occurs; if nitrogen is leached this is in the form of NO3, which may be partly derived from atmospheric ammonia through nitrification. If net nitrification occurs (i.e., if the leaching of NO3 exceeds the deposition of NOx), acidity is generated internally.
The deposited or generated acidity is largely buffered by the release of AI from secondary AI-compounds (Van Breemen et al. 1982; De Vries 1994). This buffer mechanism has two important implications. Firstly, a strong drop in pH may be expected if the store of secondary AI-compounds becomes exhausted. For some Dutch forest soils this may be the case within a few decades (see De Vries et al. 1995 for further discussion). Secondly, phytotoxic effects of AI may be expected. Critical levels of AI in soil solution for the occurrence of such effects have often been stated as AI/Ca = 1 (relative) or AI = 0.2 mmolc.m-3 (absolute) (De Vries & Heij 1991). Atmospheric deposition of nitrogen compounds leads to a strong increase in the amount of plant-available N. A good quantification is lacking, but changes in both the physiology of trees (Steingr5ver & Jans 1995), and in the species composition of understorey vegetation (De Vries et al. 1995) indirectly show the importance of this effect. There are a few general comments to be made. (1) deposition of ammonia is often stated as an important source of acidity in forest soil. It should however be stressed that in the present situation in The Netherlands, the contribution of SO2 to soil acidification exceeds the contribution of NH3, due to the large retention of N in the soil/vegetation system (De Vries 1994). The relative importance of atmospheric compounds may however change in the future, depending upon the success of various abatement measures. (2) These is still a large uncertainty about the phytotoxic effect of AI. The critical levels for AI in soil solution (and, consequently, the critical loads for acid) are based on model calculations and laboratory observations (De Vries & Kros 1991; Sverdrup & Warfvinge 1993). However, in a recent field survey no direct relation was found between AI concentration and tree vitality (Hendriks et al. 1994), and the kinetic model approach has also been criticized (HSgberg & Jens~n 1994). In laboratory experiments with herbaceous species the critical
295
levels for AI (both absolute and relative) far exceeded the above-stated values of N/Ca = 1 or AI = 0.2 mmolc.m-3 (Pegtel 1987, Kroeze et al. 1989, Van Dobben 1991). Effects of atmospheric deposition on tree physiology Good knowledge exists concerning deposition-related changes in leaf and needle nutrient concentration in Dutch forests (Van den Burg & Kiewit 1989). However, there is considerable uncertainty about other effects of deposition on tree physiology. The most important known effects are summarized below. Leaf and needle N concentrations are generally high (often exceeding the concentrations considered optimal for biomass increment), and base cation and P concentrations are low (often approaching or exceeding deficiency levels) (Hendriks et al. 1994). The low base cation levels are often ascribed to interference of AI or NH4 with uptake (Roelofs 1991). Again, the relation with tree vitality is unclear. In the field survey mentioned earlier (Hendriks et al. 1994) vitality appeared to be only weakly correlated with leaf and needle N and
P concentration. During the past few decades, forest growth has strongly increased over large parts of Europe. A high production was noted at the Speuld monitoring site (SteingrSver & Jans 1995). Survey studies based on tree-ring analyses indicate
a 20-50% increase of forest production since c. 1950 (Kenk & Fisher 1988, Becker et al. 1994). This increase is usually ascribed to a higher availability of N through atmospheric deposition, although some authors state increased atmospheric CO2 as a possible cause. At the Speuld site the increase in growth was accompanied by a strong increase in needle to fine root ratio (SteingrSver & Jans 1995). This is probably a general phenomenon, which might increase the drought sensitivity of trees. A strong increase in productivity and concomitant increase in shoot/root ratio as a result of NH3 deposition was also shown for some herbaceous species (Van Dobben 1991). Very high levels of NH3 deposition or concentration (> c. 50 kg N ha-l.y -1 or > c. 10 #g.m-3 NH3) probably result in a reduction of tree growth, at least for Pinus sylvestris. This was shown in both laboratory (fumigation) experiments
296
(Steingr6ver et al. 1995) and in field (fertilization) experiments (Tamm & Popovic 1995). Negative effects of NH3 on tree growth may be enhanced in the presence of 03 (Steingr6ver et al. 1995). For most tree species there is no significant trend in vitality in The Nether~ands since 1984 (Hilgen 1994). As stated above, the hypothesized relations between
loss of vitality (defined as defoliation or discolouration) and soil chemistry (acidification or high AI or NH4 concentrations) could not be confirmed in the field. Critical levels and critical loads
The relation of some of the above-described phenomena with atmospheric deposition (of acidity, nitrogen, or both) is now well understood and in some cases quantitatively well described. The known quantitative relations have in turn led to the formulation of critical loads and critical levels, which have become important tools for policy makers (De Vries & Kros 1991). However, critical values are based on chemical or physiological criteria, and not on observed visible effects in the field. In general, it seems that up to now researchers have failed to confirm the hypothesized relationship between atmospheric deposition and forest decline (or, in particular, tree vitality) in the field. Therefore critical levels and critical loads must be considered as best guesses according to the present state of knowledge, of concentration or deposition values where the risk of ecological effects is reduced to an acceptable level. For a discussion of the differences in philosophy underlying critical levels and critical loads the reader is referred to Van der Eerden (1995). The apparent absence of a simple and direct link between abiotic parameters and vitality has gradually caused a shift in research focus towards combinations of atmospheric deposition and other abiotic or biotic factors (see Van der Eerden 1995). However, the study of such combination stresses is a difficult one, be it only because the number of possible combinations tends to multiply explosively.
297
Table 1: Overview of changes in non-forest ecosystems in The Netherlands (including non-tree components of forests) that are partly or completely ascribed to atmospheric deposition.
effect
cause
source
extinction of many diatom spp., dominance of a few spp.
acidification
Van Dam et al. (1981), Van Dobben et al. (1992)
decline of isoetid spp., increase of submerged mosses
acidification, eutrophication
Roelofs (1983), Arts (1990)
decline of amphibians, hampered reproduction
acidification
Leuven et al. (1986), Van Dam & Buskens (1993)
replacement of Calluna and Erica by grasses
eutrophication
Aerts et al. (1990) Berdowski (1987),
decline of rare spp.
acidification (eutrophication)
Van Dobben (1991), Houdijk et al. (1993)
chalk grassland
decrease in species richness, dominance of Brachypodium
eutrophication
Bobbink & Willems (1987), Bobbink (1989)
pine forest
extinction of terrestrial lichens, increase of grasses
eutrophication
De Vries (1982), Van Dobben et al. (1994)
decline of mycorrhizaforming fungi
eutrophication, (acidification?)
Jansen & Van Dobben (1987), Arnolds & Jansen (1992)
hampered reproduction of great tit
acidification
Graveland et al. (1994)
epiphytic lichens
extinction of many spp.
SO2 toxicity
De Wit (1976), Van Dobben (1993)
change in dominance of spp.
acidification
wet oligotrophic grassland
strong decline
eutrophication, (acidification?), desiccation, reclamation
Horsthuis & Schaminee (1993), Bink et al. (1994)
alder swamp
strong decline of characteristic spp. in undergrowth
eutrophication, desiccation
Bink et al. (1994)
dune grassland
decline of lichens, increase of grasses
eutrophication
Veer et al. (1993)
floating rich-fen
change in dominance of moss spp.
acidification
Kooyman (1993a,b), Van Wirdum (1993)
various pioneer communities
strong decline
eutrophication
Bink et al. (1994)
ecosystem soft water
heathland
298
Non-forest ecosystems Research into the ecological effects of atmospheric deposition was not completely restricted to forests; other (semi-)natural ecosystems have also been the subject of study. Surprisingly, clear effects of atmospheric deposition on presence or vitality of organisms in non-forest ecosystems are claimed by a great number of authors, and for a highly diverse group of ecosystems, despite a significantly lower research effort. These ecosystems include freshwater (both plankton and macrophytes), cryptogamic epiphytes, heathland, coastal and inland dunes, but also forest undergrowth. Effects on animals have been reported as well, e.g. on amphibians and birds, besides the well-known fish decline. These effects are treated in detail by e.g. Cals et al. (1993), Bink et al. (1994) or Bobbink & Roelofs (1995), and are summarized in Table 1. In some of these cases, field observations are supported by experimental results, both from field (manipulation) and laboratory studies. In the terrestrial environment the latter is especially true for the heathland ecosystem (Aerts & Berendse 1989, Van Dobben 1991, Houdijk 1993). Comparison of forest and non-forest ecosystems We might wonder to what extent the insight gained in the study of non-forest ecosystems could also be applicable to forests. In general, two mechanisms seem to underlie the field effects of atmospheric deposition: (a) intolerance to acidification, and (b) differential growth response of species to nitrogen enrichment, resulting in competitive displacement. Now the most intensively studied forest trees generally belong to acid-resistent species (most conifers, oak, beech etc.); and inter-species competition is either absent (in monocultures) or only effective on a very long time-scale (possibly longer than one rotation period). Still, some of the processes shown in lower vegetation might also be effective in forests. This can be made clear with an example. Common heather, Calluna vulgaris, is a woody and long-living species (like forest trees), but by the simple fact of its lower stature, lends itself better for an experimental approach (Hell 1984, Berdowski 1987, Van Dobben 1991). The results of field and experimental work clone with this species are summarized below.
299
An example: Calluna Calluna appeared to be highly resistent to both high SO2 concentrations and low soil pH. In monoculture, a high availability of nitrogen (either as dissolved NH4 or as gaseous NH3) initially leads to a strongly enhanced growth. However, under high levels of atmospheric deposition, several mechanisms cause trouble for this species in the long run: (1) an increased shoot/root ratio decreases its drought tolerance; (2) an increased amino acid content increases its palatability for plague insects; (3) a decreased sugar content decreases its frost hardiness. In general, the species becomes more susceptible to various stress factors, which will ultimately lead to its replacement by grasses which are more competitive in a nitrogen-rich environment. It is tempting to transfer this mechanism directly to forest trees. Both the enhanced growth, the change in shoot/root ratio and the increase in amino acid content have also been shown there. There is little reason to suppose that the enhanced susceptibility for plagues or extreme weather conditions found for Calluna is absent in trees. Increased sensitivity to frost and fungal infections has in fact been shown in some cases (Aronsson 1980, Van Dijk et al. 1992). The analogy with lower vegetation may therefore be considered as a warning signal for the forest (De Visser 1994). If this analogy is true, strong effects have not occurred hitherto in forest trees because of (a) the absence of inter-species competition, and (b) a long response period.
Conclusion The above warning signal was derived from the physiological response of the trees to atmospheric deposition. However, some of the effects on other forest components than the trees may be considered as warning signals indicating a tendency towards destabilization of the forest ecosystem as a whole. The most apparent of these signals are (a) the depletion of soil buffer substances, in particular AI, (b) the increase of ruderal species in undergrowth, and (c) the decline of mycorrhiza-forming fungi.
300
The relation between the above changes and tree vitality is still unclear. However, each has a possible link with tree vitality. Aluminium is phytotoxic, but also serves as a buffer that protects forest soils from a dramatic drop in pH. Forest undergrowth may be an important sink for nutrients (Mugasha & Pluth 1994, Melin et al. 1983) but it may also hamper rejuvenation of trees (Wagner 1994). Mycorrhiza fungi are important for water and nutrient uptake by trees. Changes in these factors show us that the forest is at risk, and at any given moment large changes might be triggered by some secondary factor. Such changes could be comparable to the changes observed in the ecosystems summed up in Table 1, and might therefore include a decline of the dominant species (i.e., forest dieback). It is clear that this risk will become larger as the magnitude and duration of critical load exceedence increases. On the other hand manipulation experiments (Boxman et al. 1995) have shown that at least some of the parameters that are now at 'danger' levels may rapidly return to 'safe' levels if deposition drops below the critical load.
Acknowledgements: Thanks are due to E. Steingr~ver and W. de Vries for helpful comments on an earlier version of the manuscript.
301 References Aerts, R, Berendse, F. 1989. Above-ground nutrient turnover and net primary production of an evergreen and a deciduous species in a heathland ecosystem. Journal of Ecology 77:343-356. Aerts, R, Berendse, F, de Caluwe, H, Schmitz, M. 1990. Competition in heathland along an experimental gradient of nutrient availability. Oikos 57:310-318. Arnolds, E J M, Jansen, A E. 1992. New evidence for changes in the macromycete flora of The Netherlands. Nova Hedwigia 55:325-351. Aronsson, A. 1980. Frost hardiness in Scots pine (Pinus sylvestris L.) I1: hardiness during winter and spring in young trees of different mineral nutrient status. Studia Forestalia Suecica 155:1-27. Arts, G. 1990. Deterioration of atlantic soft water systems and their flora, a historical account. Diss., Nijmegen, 197 p. Becker, M, Bert, G D, Bouchon, J, Picard, J-F, Ulrich, E. 1994. Tendences & long terme observ(~es dans las croissance de divers feuillus et r~sineux du nord-est de la France depuis le milieu du 19e si~cle. Revue Foresti~.re Fran(;aise 56:355-341. Berdowski, J J M. 1987. The catastrophic death of Calluna vulgaris in Dutch heathland. Diss., Utrecht, 132 p. Bink, R J, Bal, D, van dan Berk, V M, Draaijer, L J. 1994. Toestand van de natuur 2. Rapport IKC-NBLF 4:246 p. Wageningen. Bobbink, R. 1989. Brachypodium pinnatum and the species diversity in chalk grassland. Diss., Utrecht, 141 p. Bobbink, R, Willems, J H. 1987. Increasing dominance of Brachypodium pinnatum (L) Beau in chalk grasslands: a threat to species-rich ecosystems. Biological Conservation 40:301-314. Bobbink, R, Roelofs, J G M. 1995. Ecological effects of atmospheric deposition on non-forest ecosystems in Western Europe. In: J-W Erisman, G J Heij & T Schneider (eds.): Proceedings Speciality Conference Acid Rain Research: Do we have enough answers? October 10-12, 1994, 's-Hertogenbosch, Netherlands. Boxman, A W, De Visser, P H B, Roelofs, J G M. 1995. Experimental manipulations: forest ecosystem responses to changes in water, nutrients and atmospheric loads. In: J-W Erisman, G J Heij & T Schneider (eds.)" Proceedings Speciality Conference Acid Rain Research: Do we have enough answers? October 10-12, 1994, 's-Hertogenbosch, Netherlands. Cals, M, de Graaf, M, Roelofs, J. 1993. Effectgerichte maatregelen tegen verzuring en eutrofi~ring in natuurterreinen, 188 p. Vakgroep Oecologie KUN / Directie NBLF Ministerie van L, N en V. De Visser, P H B. 1994. Growth and nutrition of Douglas fir, Scots pine and pedunculate oak in relation to soil acidification. Diss., Wageningen, 185 p + app. De Vries, I M. 1982. De invIoed van luchtverontreiniging / zure neerslag op hogere planten. Rapport RU Utrecht / RIN Leersum :180 p + bijl. De Vries, W. 1994. Soil response to acid deposition at different regional scales: field and laboratory data, critical loads and model predictions. Diss., Wageningen, 487 p. De Vries, W, Heij, G J. 1991. Critical loads and critical levels for the environmental effects of air pollutants. In: G J Heij & T Schneider (eds.): Acidification Research in The Netherlands: final report on the Dutch Priority Programme on acidification, 205-214. Elsevier, Amsterdam. De Vries, W, Kros, J. 1991. Assessment of critical loads and the impact of deposition scenarios by steady state and dynamic soil acidification models. In: G J Heij & T Schneider (eds.): Acidification Research in The Netherlands, final report of the Dutch Priority Programme on acidification, 569-624. Elsevier, Amsterdam. De Vries, W, Leeters, E E J M, Hendriks, C M A, Van Dobben, H F, Van den Burg, J, Boumans, L J M. 1995. Large scale impacts of acid deposition on forests and forest soils in The Netherlands. In: J-W Erisman, G J Heij & T Schneider (eds.): Proceedings Speciality Conference Acid Rain Research: Do we have enough answers? October 10-12, 1994, 's-Hertogenbosch, Netherlands.
302 De Wit, A. 1976. Epiphytic lichens and air pollution in The Netherlands. Bibliotheca LichenoIogica 5:115 p. + ann. Cramer, Vaduz. Graveland, J, Van der Wal, R, Van Balen, J H, Van Noordwijk, A J. 1994. Poor reproduction in forest passerines from decline of snail abundance on acidified soils. Nature 368:446-448. Heil, G. 1984. Nutrients and the species composition of heathland. Diss., Utrecht, 139 p. Hendriks, C M A, de Vries, W, Van den Burg, J. 1994. Effects of acid deposition on 150 forest stands in The Netherlands. Report DLO Winand Staring Centre 69.2:55 p. Wageningen. Hilgen, P R. 1994. De vitaliteit van het Nededandse bos 12: Verslag van de landelijke inventarisatie in 1994. Rapport IKC-N 10:41 p. Wageningen. Horsthuis, M A P, Schaminee, J H J. 1993. Verspreiding en ecologische spectra van 24 plantengemeenschappen in Nederland. IBN Rapport 021:170 p. Wageningen. Houdijk, A L F M. 1993. Atmospheric ammonium deposition and the nutritional balance of terrestrial ecosystems. Diss., Nijmegen, 127 p. H6gberg, P, Jens~n, P. 1994. Aluminium and uptake of base cations by tree roots: a critique of the model proposed by Sverdrup et al. Water, Air and Soil Pollution 75:121-126. Jansen, E, Van Dobben, H F. 1987. Is decline of Cantharellus cibarius in The Netherlands due to air pollution? Ambio 16:211-213. Kenk, G, Fischer, H. 1988. Evidence from nitrogen fertilization in the forests of Germany. Environmental Pollution 54:199-218. Kooijman, A M. 1993a. Changes in the bryophyte layer of rich fens as controlled by acidification and eutrophication. Diss., Utrecht, 159 p. Kooijman, A M. 1993b. The decrease of rich fen bryophytes in The Netherlands. Biological Conservation 59:139-143. Kroeze, C, Pegtel, D M, Blom, C J C. 1989. An experimental comparison of aluminium and manganese susceptibility in Antennaria dioica, Arnica montana, Viola canina, Filago minima and Deschampsia flexuosa. Acta Botanica Neerlandica 38:165-172. Leuven, S E W, den Hartog, C, Christiaans, M M C, Heijligers, W H C. 1986. Effects of water acidification on the distribution pattern and the reproductive success of amphibians. Experientia 42:495-503. Melin, J, NSmmik, H, Lohm, U, Flower-Ellis, J. 1983. Fertilizer nitrogen budget in a Scots pine ecosystem attained by using root-isolated plots and 15-N tracter technique. Plant and Soil 74:249-263. Mugasha, A G, Pluth, D J. 1994. Distribution and recovery of 15-N urea in a tamarack/black spruce mixed stand on a drained minerotrophic peatland. Forest Ecololy and Management 68:353-363. Pegtel, D. 1987. Effect of ionic AI in culture solutions on the growth of Arnica montana L and Deschampsia flexuosa (L) Trin. Plant and Soil 102:85-92. Roelofs, J G M. 1983. Impact of acidification and eutrophication on macrophyte communities in soft waters in The Netherlands I: Field observations. Aquatic Botany 17:139-155. Roelofs, J. 1991. Vegetation under chemical stress: effects of acidification, eutrophication and alkalinization. Diss., Nijmegen, 165 p. SteingrSver, E, Dueck, T, Van der Eeerden, L. 1995. Assessment and evaluation of critical levels for 03 and NH3. In: J-W Erisman, G J Heij & T Schneider (eds.): Proceedings Speciality Conference Acid Rain Research: Do we have enough answers? October 10-12, 1994, 's-Hertogenbosch, Netherlands. Steingr5ver, E, Jans, W W P. 1995. Physiology of forest-grown Douglas-fir trees: Effects of pollution and drought. Report RIVM 793315. Biithoven. Sverdrup, H, Warfvinge, P. 1993. The effect of soil acidification on the growth of trees, grass and herbs as expressed by the (Ca+Mg+ K)/AI ratio. Reports in ecology and environmental engineering 2:108 p. University of Lund. Tamm, C-O, Popovic, B. 1995. Long-term field experiments simulating increased deposition of sulphur and nitrogen to forest plots, manuscript. Ulrich, B. 1983. A concept of forest ecosystem stability and of acid deposition as driving for
303 destabilization. In: Ulrich B, Pankrath J (eds.): Effects of accumulation of air pollutants in forest ecosystems; Proceedings of a workshop held at G5ttingen, May 16-18, 1982, 1-29. Reidel, Dordrecht. Ulrich, B, Mayer, R, Khanna, P K. 1979. Deposition von Luftveruntreinigungen und ihre Auswirkung in WaldSkosystemen in Soiling, 291 p. Saued~inder, Frankfurt a/M. Van Breemen, N, Burrough, P A, Velthorst, E J, Van Dobben, H F, De Wit, T, Reijnders, H F R. 1982. Soil acidification from atmospheric ammonium sulphate in forest canopy throughfall. Nature 299:548-550. Van Dam, H, Suurmond, G, Ter Braak, C J F. 1981. Impact of acidification on diatoms and chemistry of Dutch moorland pools. Hydrobiologia 83:425-459. Van Dam, H, Buskens, R F M. 1993. Ecology and Management of Moorland Pools - Balancing Acidification and Eutrophication. Hydrobiologia 265:225-263. Van Dijk, H F G, Van Der Gaag, M, Perik, P J M, Roelofs, J G M. 1992. Nutrient availability in Corsican pine stands in the netherlands and the occurrence of Sphaeropsis sapinea - A field study. Canadian Journal of Botany 70:870-875. Van Dobben, H F. 1991. Integrated effects (low vegetation). In: G J Heij & T Schneider (eds.): Acidification Research in The Netherlands, final report of the Dutch Priority Programme on acidification, 465-523. Elsevier, Amsterdam. Van Dobben, H F. 1993. Vegetation as a monitor for deposition of nitrogen and acidity. Diss., Utrecht, 214 p. Van Dobben, H F, Mulder, J, Van Dam, H, Houweling, H. 1992. Impact of acid atmospheric deposition on the biogeochemistry of moorland pools and surrounding terrestrial environment. Agricultural Research Reports 931:232 p. Pudoc, Wageningen. Van Dobben, H F, Vocks, M J M R, Jansen, E, Dirkse, G M. 1994. Veranderingen in de ondergroei van het Nederlandse dennenbos over de periode 1985-1993. IBN Rapport 085:37 p. Wageningen. Van Wirdum, G. 1993. Basenverzadiging in soortenrijke trilvenen. In: M Cals, M de Graaf en J Roelofs (eds.): Effectgerichte maatregelen tegen verzuring en eutrofi&ing in natuurterreinen; proceedings van een symposium op 30 oktober 1992, 97-126. Katholieke Universiteit, Nijmegen. Van den Burg, J, Kiewiet, H P. 1989. Veebezetting en de naaldsamenstelling van groveden, douglas en Corsicaanse den in het Peelgebied in de periode 1956 t/m 1988: een onderzoek naar de betekenis van de veebezetting voor het optreden van bosschade. Rapport De Dorschkamp 559:77 p. Wageningen. Van der Eerden, L J M. 1995. Thematic report on effects (with special emphasis on combination stress). In: G J Heij & T Schneider (eds.): Final report of the Dutch Priority Programme on Acidification, third phase. Elsevier, Amsterdam, in prep. Veer, M, van der Meulen, F, Hell, G W, Kooijman, A M. 1993. Ecotoopbeheer in droge duinen: sturende processen en limiterende factoren. In: M Cals, M de Graaf en J Roelofs (eds.): Effectgerichte maatregelen tegen verzuring en eutrofi&ing in natuurterreinen, proceedings van een symposium op 30 oktober 1992, 147-170. Katholieke Universiteit, Nijmegen. Wagner, R G. 1994. Toward integrated forest management. Journal of Forestry 92:26-30.
This Page Intentionally Left Blank
F U T U R E OF A C I D I F I C A T I O N R E S E A R C H
S E S S I O N VI TODAY'S KNOWLEDGE; IS I T SUFFICIENT FOR TOMORROW'S DECISION MAKING P U R P O S E S
This Page Intentionally Left Blank
G.J. Heij and J. W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
307
Lessons learned in acidification research: Implications for future environmental research and a s s e s s m e n t s Ellis B. Cowling College of Forest Resources, North Carolina State University, Raleigh, North Carolina 27695 USA Abstract The purpose of this paper is to summarize lessons that are available to be learned from the American experience in acidification research. During the years between 1975 and the early 1990s, various acidification research and policy-focused assessment programs were initiated in the United States and Canada. The objectives of these programs were to better understand the phenomenon of acid precipitation and its effects on lakes, streams, fish, crops, forests, visibility, human health, engineering materials, and cultural resources in North America. The most comprehensive of these programs were the Memorandum of Intent (MOI) Work Group Processes initiated in 1978, the Office of Technology Assessment (OTA) Report completed in 1984, and the National Acid Precipitation Assessment Program (NAPAP) completed in 1990. These were the first efforts by governments in Canada and the United States to develop scientific and policy-focused assessments of alternative strategies for management of environmental risks on a continental scale. As such, they provide a useful basis for evaluation of: 1) scientific learning about atmosphere/biosphere interactions, and 2) social learning about the interface between science and public policy.
1. INTRODUCTION Many insights developed in this paper are outgrowths of an international research project at Harvard University's Kennedy School Of Government. This project, entitled "Social Learning in the Management of Global Environmental Risks" is led by Dr. William Clark. The purpose of the project is to compare and contrast the processes used in various countries as they make decisions about what each nation should do, if anything, about certain contemporary environmental risks. Dr. Clark asked that I prepared the '~Evaluation" chapter for the "Acidi~cation" case study for the United States (Cowling, 1994). This chapter was developed using the following protocol which was designed by an international committee of experts within the Social Learning Project: Three environmental risks were selected as case studies: 1) Global climate change,
308 2) Depletion of stratospheric ozone, and 3) Acidification and long-range transport of air pollutants. Nine nations on three continents were selected for inclusion in the Social Learning Project. Risk assessment and risk management efforts in each country were studied by teams of researchers working under the guidance of the following national team leaders: 1) The Netherlands --Josee van Eijnhoven, University of Utrecht, Utrecht, The Netherlands 2) Germany -- Jill Yeager, International Institute for Applied Systems Analysis, Laxenburg, Austria 3) United Kingdom -- Brian Wynne, University of Lancaster, United Kingdom 4) Hungary -- Ferenc Toth, Potsdam Institut fiir K]imafolgenforschung, Potsdam, Germany 5) Russia --Vassily Sokolov, Russian Academy of Sciences, Moscow, Russia 6) Japan -- iranda Schreurs, University of Maryland, College Park, MD 7) Canada--Rod Dobell, University of Victoria, British Columbia, Canada 8) Mexico --Diana Liverman, Pennsylvania State University, University Park, PA 9) United States -- William Clark and Nancy Dickson, Kennedy School of Government, Harvard University, Cambridge, MA The roles played by seven actor groups were examined within each nation: 1) Expert communities -- mainly scientists and engineers with special knowledge about each environmental risk; 2) Business and industry groups -- mainly the electric utility, coal, automobile, and other industries; 3) Other non-governmental groups -- mainly environmental groups, trade associations, and private foundations; 4) Executive branches of government -- main|y federal government agencies, but including some state and provincial organizations as well; 5) Legislative branches of government such as Congress in the USA or Parliament in Canada; 6) Judicial branches of government; and 7) The media -- including both print and electronic media. Studies were then made of the risk-assessment and risk-management functions performed by each actor group within each nation with respect to each environmental risk. The definitions listed below are those used for the case study of acidification and long-range transport of air pollutants: 1) Issue Framing -- framing the issue of acidification and associated long-range transport of air pollutants (LRTAP); 2) Risk Assessment -- assessing the nature and magnitude of the risks imposed by acidification and LRTAP; 3) Response Assessment -- assessing alternative strategies by which the risk of acidification and LRTAP could be managed; 4) Goal and Strategy Formulation -- formulating specific goals and strategies by which the risks of acidification and LRTAP could be managed;
309
5) Implementation -- actions taken by one or more actor groups in an effort to manage the risk of acidification and LRTAP; 6) Monitoring -- measuring the amount and intensity of risk and the nature and magnitude of damage or injury caused by acidification and LRTAP; and 7) Evaluation -- self-conscious efforts by different actor groups to determine the effectiveness of their own efforts, or the efforts of other actor groups, in fulfilling one or more of the preceding six risk-assessment and/or riskmanagement functions.
2. EVALUATION O F AMERICAN A C I D I F I C A T I O N R E S E A R C H From a broad array of evaluation documents produced by the seven actor groups listed above, 16 were selected for detailed analysis: --Eight documents from four different e x p e r t c o m m u n i t i e s -- one by a group of four scientists which led to establishment of the National Acid Precipitation Assessment Program (NAPAP) (Galloway et al, 1978), one by a professional society dealing with the NAPAP Interim Assessment (LeFohn and Krupa, 1988a, 1988b), five documents by scientists selected by the Ecological Society of America (Levine, 1992; Russell, 1992; Cowling, 1992; Loucks, 1992; Schindler, 1992), and another noted scientist (Likens, 1992); --Two documents from b u s i n e s s a n d i n d u s t r y -- both by a former Director of Environmental Research at the Electric Power Research Institute (Perhac, 199 la, 199 lb); --Four documents from e x e c u t i v e b r a n c h o r g a n i z a t i o n s -- two by Ad Hoc evaluation committees (EPA, 1983; OSTP, 1984) (the latter by Presidential appointment), the NAPAP Oversight Review Board report (NAPAP, 1991a), and one by a former director of NAPAP (Mahoney, 1990); --Two documents from C o n g r e s s -- a report by the Congress' Office of Technology Assessment (OTA, 1984), and an oversight hearing on plans for NAPAP's 1990 Integrated Assessment (House of Representatives, 1988). The principal foci of these documents were successes and shortcomings of two North American research and assessment programs or activities: 1) M e m o r a n d u m of Intent Work Group Reports (MOI, 1982-83) which were developed in response to a Congressional resolution calling for negotiation of an air-quality treaty between Canada and the United States, and 2) The N a t i o n a l Acid P r e c i p i t a t i o n A s s e s s m e n t P r o g r a m (NAPAP) -- the 10-year-federal research and assessment program in the United States (NAPAP, 199 lb, 199 lc).
3. INSIGHTS FROM THESE SIXTEEN EVALUATION DOCUMENTS Preparation of the Evaluation Chapter for the United States Case Study on Acidification (Cowling, 1994) revealed five major types of insights:
310 1) Many incremental changes in understanding of the phenomena and effects of air pollution and acidification have occurred in the United States since 1960. 2) Despite many gaps in scientific understanding, adequate information about the phenomenon and effects of acidification were available in 1984 for the United States and Canada to take initial steps to decrease emissions of sulfur and nitrogen oxides on both sides of the border and to reconvene negotiations to set a preliminary target load for acid deposition in aquatic ecosystems of the northeastern United States and southeastern Canada. This is the judgment of the Nierenberg Committee (OSTP, 1984). On the basis of an entirely separate evaluation process, substantially the same conclusion was reached by the Office of Technology Assessment in its Report to Congress -- Final Report on Acid Rain and Transported Air Pollutants: Implications for Public Policy (OTA, 1984). 3) Significant innovations took place during the 1980's and 1990's in institutional arrangements by which the United States sought to assess and manage the risks of acidification and long range transport of air pollutants. 4) A total of 8 general lessons and 19 more specific lessons are available to be learned from the United States' experience with acid deposition research. Some of these general and specific lessons may be of value in improving the policy performance of the United States and other countries as they seek solutions to the acidification and other global, national, and regional environmental problems. 5) The NAPAP Oversight Review Board provided a series of guidelines for formulation of research program findings to be used for policy purposes. Several of these incremental changes, institutional and process innovations, lessons available to be learned, and guidelines for formulation of statements of findings are summarized below.
3.1 Incremental Acidification
Changes
in
Understanding
Air
Pollution
and
1) During the 1960s, air pollution was regarded as a local, mostly urban problem m~inly affecting h u m a n health. In contrast, by the early 1990s, air pollution has come to be regarded as both a local and a regional problem in urban, rural, and even some remote areas of the United States, where it has been having significant influences on the stability of ecosystems, engineering materials, historical monuments, visibility, etc. 2) During the 1960s, "dilution" was confidently believed to be "a solution to pollution". In recent years, we have come to realize that "what goes up comes down somewhere and probably has some kind of effect when it gets there".
311 3) During the 1960s, we thought of pollutant exposure in terms of air concentrations and/or annual _~_mounts of deposition. Now we realize that cumulative (multi-decade) exposures also are important and that we must understand that softs, have a finite "sulfate absorption capacity" or "assimilative capacity" for acidifying substances and other airborne pollutant chemicals. 4) During the 1960s, we thought that most pollutants of concern were primary pollutants, that is, the substances that caused a pollution problem were emitted directly by a pollution source. In recent years we have come to recognize that secondary pollutants also are important. Secondary pollutants are injurious substances formed in the atmosphere from primary pollutants. Acid deposition and ozone near the ground are examples of secondary pollutants. 5) During the 1970s, we thought acid deposition was largely a problem of long distance transport. Now we recogni_ze that acid deposition is both a long-distance and a short-distance problem and that areas of high chemical loading in the United States are substantially coincident with areas of high emission. This is not true in some (mostly remote) parts of eastern Canada, however. Here, distant sources of pollutants are more important than local sources. 6) During the early 1970's, we thought that sulfur dioxide was the principal cause of airborne acids and acidifying substances. Now we recognize that a wide variety of both sulfur and nitrogen compounds are important causes of acidifying deposition from the atmosphere in North America and Europe. 7) During the early 1970's, we thought of acid deposition mostly as a wetdeposition phenomenon, hence the terms "acid rain" or "acid precipitation". Now we think of acid deposition as both a wet, and a dry, and a cloud-water deposition phenomenon involving acid rain, acid snow, acid fog, acid dew, acid aerosols, acid particles, acidic gases, and other acidifying substances. 8) During the 1980's, we used the terms "acidic rain" or "acidic deposition". Increasingly now we use the terms "acidifying deposition" and "acidification" because not all acidifying substances are acidic, and not all substances that cause acidification are acidic either. 9) During the 1980's, a major convergence of extreme views about acidification took place both in scientific and in public opinion. In 1980, some scientists, and many in the public, believed that air pollution in general and acid deposition in particular posed serious and immediate threats to the health and productivity of aquatic and terrestrial ecosystems. By contrast, another group of scientists, and some in the public, believed that air pollution was mainly an urban problem, the acids present in precipitation were too dilute to have significant effects on any but the most sensitive lakes and streams, and acid deposition was unlikely to affect crops or forests. By the end of this decade, these extreme views converged into a more moderate middle ground:
312 --There are important effects of acidification on aquatic ecosystems in some regions of North America where sulfate absorption capacity of softs have been exhausted; --Nitrogen saturation of some forest softs is occurring in the Untied States; --High-mountain red spruce forests exposed to acid cloud water in the northeastern United States are predisposed to damage by winter frost; --Agricultural crops are not seriously threatened by acid deposition, but are being impacted by ozone near the ground in large parts of eastern North America; --Continuing deposition of large amounts of sulfur and nitrogen oxides over decades of time leads to continuous alteration of terrestrial and aquatic ecosystems; --Decreased loading of the atmosphere with sulfur and nitrogen oxides will ~meliorate these ecological effects, improve visibility, and decrease h a r m to engineering materials and historical monuments. 10) In the Clear Air Act Amendments of 1990, the United States initiated a system of marketable pollution-trading permits, continuous emissions monitoring, requirements for decreases in emission of VOC, etc. 11) During the 1980's, the United States found it diffctdt to consider specific critical or even target loads of acidifying substances. Recently, as the concept of critical loads and associated target loads has gained wider acceptance among our own citizens and in many European countries, some rethinking of our earlier national reluctance to consider such approaches appears to be developing. 3.2 I n n o v a t i o n s in S c i e n t i f i c R e s e a r c h a n d A s s e s s m e n t P r o c e s s e s
The most important institutional and process innovations deriving from the NAPAP experience were the following: 1) Development of a unique mechanism for federal interagency coordination. NAPAP was governed by a federal interagency coordinating committee with six "leading agencies", each with different specific roles in research and assessment, and six other "contributing agencies". This organizational construct h a d no true "lead agency" with both administrative authority and budgetary control over most of the NAPAP progr_~m. Such an approach had not been used before in the United States. NAPAP's federal interagency coordinating committee was called the Acid Precipitation Task Force. It served from the time of NAPAP's establishment in 1981 through 1990 when NAPAP produced its 1990 Integrated Assessment. This mode of organization h a d the constructive effect of bringing together in a single coordinated program a total of 12 different federal agencies. Many of these agencies had widely different agency missions. After 10 years of working together, many agency representatives felt that they were in a better position to help the country make policy decisions that transcend different agency jurisdictions, regions, industrial sectors, and value systems within both Canadian and United States society.
313 2) Apparently in deference to this organizational innovation, however, NAPAP developed a tradition that all six "leading agencies"in the progr_~m would have to approve, prior to publication: all state of science reports, state of technology reports, annual reports, interim and final assessment documents, scientific findings, testimony for congressional hearings, and other public documents. This requirement for unanimous approval by six different agencies of government in light of advice received from both an Interagency Science Committee and an Interagency Policy Committee, frequently led to compromise, "watering down", and, almost always, to substantial delay in release of research and assessment findings. 3) The most significant process innovation deriving from the NAPAP experience was development of a new paradigm for "integrated assessment" (House of Representatives, 1988; Mahoney, 1990; NAPAP, 1989)which should: a) Be based on credible scientific information which is well documented and available to all users; b) Be developed according to a structured plan which includes i) definition of all effects studied, and ii) specific policy and scientific questions to be addressed; c) Be developed with its plan and specific questions open for review by all potential users in the United States and Canada and modified to include consideration of user comments; d) Document all data, quality assurance methods, computational analysis methods, and computer software used, to facilitate subsequent analyses by user groups; e) Define uncertainties, to allow users to evaluate confidence in reported findings; f) Be structured to facilitate incorporation of new information on trends, doseresponse models, emissions estimates, etc.; and g) Include illustrative management scenarios that consider the benefits, costs, and timing of alternative control measures and technologies. 3.3 G e n e r a l a n d S p e c i f i c L e s s o n s T h a t are A v a i l a b l e to be L e a r n e d f r o m t h e A c i d D e p o s i t i o n E x p e r i e n c e in t h e U n i t e d S t a t e s
1) "Integrated Assessment" is very different from "scientific research and reporting" (see items 3a-3g in Section 3.2 above). So also are "Guidelines for
Formulation of Research Program Findings To Be Used for Policy Purposes" different from those for preparation of scientific findings not planned for use for policy purposes. The guidelines developed by the NAPAP Oversight Review Board (NAPAP, 1991a) are of sufficient value for other policy-focused purposes that they are reproduced in their entirety in Section 3.4 of this paper. 2) The key actors in developing a science-based integrated assessment are not two in n u m b e r - - scientists and government decision makers -- as many had assumed before the acidification research and assessment experience, but rather at least four in number-- scientists; policy analysts; communicators; and decision
314 makers -- including decision makers in industry, governments, and public interest groups (NAPAP, 199 la; Cowling, 1992, Russell, 1992). 3) It is desirable to include very highly skilled and experienced representatives from each of these actor groups in developing peer review panels, arranging oversight hearings, and establishing Ad Hoc evaluation committees. In NAPAP, scientists were selected disproportionately for service on such evaluation panels and committees. Policy analysts, representatives of states, other user groups, and experienced decision makers were conspicuous by their ~lmost complete absence from such panels and committees. Partly as a result of this bias, NAPAP science was generally regarded as more satisfactory than NAPAP assessments. 4) The United States is just beginning to learn how to develop, maintain, and police the necessary "series of highly selective semi-permeable barriers that allow or block influence of different types from one set of players to another [see list of key actor groups in item 2 above]. For ex_~mple, scientist doing and reporting their research must be isolated from influence over what they find and report, but be subject to direction over the questions that are of importance to the decision. Policy-makers must be protected from policy-analysts or scientists telling them what they should decide, but open to information about what the consequences of alternative decisions are likely to be" (For further information on this topic see Russell, 1992). 5) In a policy-focused scientific research and assessment process, success will be enhanced ff the program will consider and apply the following 19 specific lessons and/or recommendations: --Give assessment primacy. --Understand the role of science and how to use it effectively to answer critical policy questions and associated scientific questions identified within an integrated assessment framework (Mahoney, 1990; NAPAP, 1989, 1991a). --Develop and maintain liaison with relevant research and assessment groups in other countries, especially those in which scientific and policy understanding may be more advanced than in the United States, and also in those countries whose interests also may be affected by the policy decisions contemplated in this country. --Provide adequate financial and h u m a n resources to get the research and assessment jobs done well and on time. --Maintain continuous and open communication between the assessment preparing and user/decision maker communities in this country and abroad. --Provide an appropriate fraction of the total funds for the program_ (10 to 20% has been suggested) to explore unexpected research leads, novel research approaches, and, especially, for individual-investigator-initiated fundomental research on promising new hypotheses. --Configure organization and authority to match responsibility. --Obtain and maintain political commitment. --Develop effective methodologies for estimating benefits as well as costs. --Report scientific uncertainties explicitly.
315 --Maintain scientific credibility through extensive and repeated interdisciplinary peer reviews with peer reviewers selected from user communities, both in this country and abroad. --Provide for independent external scientific oversight. --Take special care with the timeliness, quality, and completeness of communications. --Refuse to publish in the name of the program, any assessment report that does not meet appropriately high scientific and technical standards. --Identify authors of all assessment documents by name, institutional _Af~liation, and other potential sources of bias. --Describe accurately the peer review and other quality assurance processes actually used in preparing all parts of all assessment documents. --Beware of large data sets and large models. Also beware of advocates for use of such resources who, in the interest of protecting their favorite data set or model, inhibit their early use to give approximate findings that could guide further research and data gathering activities. --In large-scale field programs of research and assessment, be sure that the time, financial, and h u m a n resources needed for advance planning, scientific quality assurance, data quality assurance, data archiving, and data analysis and interpretation are adequate. --Take steps to ensure continuity. Especially, beware of the (conscious or unconscious) tendency to cut monitoring programs as soon as policy decisions have been made. Monitoring programs are essential to verify if expected benefits are achieved at something close to expected costs of implementation. 6) The electric utility and coal industries in the United States should learn several important lessons including the following (Perhac, 199 la, 199 lb): --"Most environmental issues end up as legislation or regulation. Therefore, the industry should prepare early to meet issues and to work with legislators (and their staffs) and environmental groups to arrive at a solution." --"Industry should coordinate its efforts better and strive for a consensus approach to environmental issues." --Industry must "overcome the common perception that it holds a rigid three-fold approach to environmental issues: a) The problem is not serious, b) If it is serious, industry isn't part of it, and c) If industry is part of it, nothing can be done about it." --"Presenting scientific or technological uncertainty as a rationale for inaction simply won't stand up. The environmental community won't be swayed and Congress won't be impressed". --"More emphasis should be placed on policy-related research". In this connection, integrated assessment and benefit/cost analyses are especially important". --"Industry should put more effort into providing cost/benefit information, especially that related to technological options."
316 7) There appeared to be no significant correlation between the amount of financial and h u m a n resources devoted to an evaluation document, or the scientific prestige of a review panel or oversight committee, and its impact on either social learning or policy improvement. Among the 16 documents selected for analysis, the least costly were those prepared by Director Mahoney ($500$2,000) in 1990, by four NADP scientists in 1978 ($10,000), and by the Ecological Society of America in 1992 ($50,000). The most costly evaluation documents were the Deutch Committee ($100,000-200,000) and Nierenberg Committee Reports ($150,000-200,000), the APCA Technical Amplification of the NAPAP Interim Assessment ($200,000-300,000), and the OTA Report ($400,000-500,000). In fact, if there is a correlation between the value and the cost of the documents, the correlation is probably inverse. The criteria for this evaluation of relative value are entirely subjective, however (Cowling, 1994)! 8) The nearly 10-year-long delay that took place in the United States about the timing of passage of the Clean Air Act Amendments of 1990 -- in 1990 instead of 1981, 1982, 1983, 1984 .... 1989 (as they were originally scheduled or attempted by other persons than the President) bears out Ralph Perhac's conclusion that the power of the President of the United States in dealing with contentious environmental issues is extraordinary (Perhac, 199 la, 199 lb)!
3.4 G u i d e l i n e s for F o r m u l a t i o n of R e s e a r c h P r o g r a m F i n d i n g s To Be U s e d for P o l i c y Purposes: The NAPAP Oversight Review Board developed the following set of guidelines for formulation of statements of scientific findings to be used for policy purposes. These guidelines are reproduced here in the form of checklist questions to illustrate the care that needs to be taken by scientists and policy analysts in research and assessment programs at the interface between science and public policy (NAPAP, 199 la). 1) Is the statement sound? Have the central issues been dearly identified? Does each statement contain the distilled essence of present scientific and technical understanding of the phenomenon or process to which it applies? Is the statement consistent with all relevant evidence -- evidence developed either through this research program or through analysis of research conducted outside of this program? Is the statement contradicted by any important evidence developed through research inside or outside of the program? Have apparent contradictions or interpretations of available evidence been considered in formulating the statement of principal findings? 2) Is the statement directional and, where appropriate, quantitative? Does the statement correctly quantify both the direction and magnitude of trends and relationships in the phenomenon or process to which the statement is relevant? When possible, is a range of uncerto_inty given for each quantitative result? Have various sources of uncertainty been identified and quantified, for example, does the statement include or acknowledge errors in actual measurements, standard
317 errors of estimate, possible biases in the availability of data, extrapolation of results beyond the mathematical, geographical, or temporal relevancy of available information, etc.? In short, are there numbers in the statement? Are the numbers correct? Are the numbers relevant to the general meaning of the statement? 3) Is th.e degree of certainty or uncertainty of the statement indicated clearly? Have appropriate statistical tests been applied to the data used in drawing the conclusion set forth in the statement? If the statement is based on a mathematical or novel conceptual model, has the model or concept been validated? Does the statement describe the model or concept on which it is based and the degree of validity of that model or concept? 4) Is the stateme.nt correct without quMification? Are there limitations of time, space, or other special circumstances in which the statement is true? If the statement is true only in some circumstances, are these limitations described adequately and briefly? 5) Is the statement clear and unambiguous? Are the words and phrases used in the statement understandable by the decision makers of our society? Is the statement free of specialized jargon? Will too many people misunderstand its meaning? 6) Is the statement as concise as it can be made without risk of misunderstanding? Are there any excess words, phrases, or ideas in the statement which are not necessary to communicate the meaning of the statement? Are there so many caveats in the statement that the statement itself is trivial, confusing, or ambiguous? 7) Is the stateme.nt free of scientific or other biases or implications of societal value judgments? Is the statement free of influence by specific schools of scientific thought? Is the statement also free of words, phrases, or concepts which have political, economic, ideological, religious, moral, or other personal-, agency-, or organization-specific values, overtones, or implications? Does the choice of how the statement is expressed rather than its specific words suggest underlying biases or value judgments? Is the tone impartial and free of special pleading? If societal value judgments have been discussed, have these judgments been identified as such and described both clearly and objectively? 8) Have societal implications been described objectively? Consideration of alternative courses of action and their consequences inherently involves judgments of their feasibility and the importance of effects. For this reason, it is important to ask if a reasonable range of alternative policies or courses of action have been evaluated? Have societal implications of alternative courses of action been stated in the following general form?: "If this [particular option] were adopted then that [particular outcome] would be expected."
318 9) Have the professional biases of authors and reviewers been described openly? Acknowledgment of potential sources of bias is important so that readers can judge for themselves the credibility of reports and assessments. 4. CONTRASTS BETWEEN E U R O P E A N AND AMERICAN ATTITUDES ABOUT ACIDIFICATION AND LONG-RANGE T R A N S P O R T OF AIR POLLUTANTS As discussed further by Nilsson and Cowling (1992), there are a number of important differencesbetween European and American attitudes about research and management of acidification and long-range transport of air pollutants. A m o n g the more intriguing of these differencesare the following: --Greater acceptance of the concepts of criticalloads and target loads in Europe, --Greater concern about visibilitydegradation in North America, --Greater concern about pollution-induced changes in water quality and aquatic ecosystems in Canada and northern Europe than in the U S A or central Europe, --Greater trust of Europeans generally in their governments, --Greater acceptance of the concept of tradeable pollution permits in the USA, --Greater acceptance of the idea of carbon taxes in Europe, and --Differences in the roles that scientists and policy analysts play in public decision making in large and small democratic nations. 5. R E F E R E N C E S
Cowling, E . B . 1992. The Performance and Legacy of NAPAP. Ecological Applications 2(2): 111-116. EPA. 1983. Report of the Ad Hoc Committee to Review the National Acid Precipitation Assessment Program (NAPAP). So-called Deutch Committee Report. U. S. Environmental Protection Agency, Science Advisory Board, Washington D.C. 26 pp. Galloway, J. N., E. B. Cowling, E. Gotham, and W. W. McFee. 1978. A National Progr_~m for Assessing the Problem of Atmospheric Deposition (Acid Rain). A Report to the President's Council on Environmental Quality. National Atmospheric Deposition Program, Colorado State University, Fort Collins, CO. 155 pp. House of Representatives. 1988. Hearings on the National Acid Precipitation Assessment Program held on April 27, 1988 before Congressman James Scheuer's Subcommittee on Natural Resources, Agricultural Research and Environment of the Committee on Science, Space, and Technology. U . S . House of Representatives, Washington D.C. 246 pp. Lefohn, A. S., and S. V. Kr-upa. 1988a. Acidic Precipitation: A Technical Amplification of NAPAP's Findings, Proceedings of an Air Pollution Control Association International Conference held on January 26-28, 1988. Air and Waste Management Association, Pittsburgh, PA. 239 pp.
319 Lefohn, A. S., and S. V. Krupa. 1988b. Acidic precipitation: A technical amplification of NAPAP's findings. (A summary of the Proceedings of an APCA International Conference). Air Pollution Control Association Journal 38(6):766-776. Levin, S . A . 1992. Orchestrating environmental research and assessment. Ecological Applications 2(2):103-106. Likens, G.E. 1992. The ecosystem concept: Its use and abuse. Ecology Institute, Odendorf-Luhe, Germany. 166 pp. Loucks, O.I. 1992. Forest response research in NAPAP: Potentially successful linkage of policy and science. Ecological Applications 2(2): 117-123. Perhac, R . M . 1991a. Making credible science usable: lessons from CAA, NAPAP. Power Engineering 95(9):38-40. Perhac, R.M. 199 lb. Usable science: lessons from acid r~in legislation, NAPAP. Power Engineering 95(10):26-29. Mahoney, J . R . 1990. Lessons Learned from NAPAP: for Consideration in Planning and Implementing Other Environmental Assessments. National Acid Precipitation Assessment Program, Washington, D.C. 5 pp. MOI, 1982-83. United States-Canada Memorandum of Intent on Transboundary Air Pollution: Final Reports by the Impact Assessment Work Group I, Atmospheric Modeling Work Group 2; Strategies Development and Implementation Work Group 3A. U.S. Department of State and Embassy of Canada, Washington, D. C. NAPAP. 1989. Assessment Plan Update. National Acid Precipitation Assessment Program, Washington, D.C. 180 pp. NAPAP. 1991a. The Experience and Legacy of NAPAP. Report of the NAPAP Oversight Review Board to the Joint Chairs Council of the Interagency Task Force on Acidic Deposition. 40 pp. NAPAP. 1991b. Acid Deposition: State of Science and Technology: Vol. I Emissions, Atmospheric Processes and Deposition; Vol. II - Aquatic Processes and Effects; Vol. III- Terrestrial, Materials, Health, and Visibility Effects; and Vol. IV - Control Technologies, Future Emissions, and Effects Valuation. National Acid Precipitation Assessment Program, Washington, D.C. NAPAP. 199 lc. 1990 Integrated Assessment Report. National Acid Precipitation Assessment Progrom, Washington, D.C. OSTP. 1984. Report of the Acid Rain Peer Review Panel. William A. Nierenberg, Chairman. Office of Science and Technology Policy, Washington D.C. 80 pp. OTA. 1984. Final report on Acid Rain and Transported Air Pollutants: Implications for Public Policy. Congressional Office of Technology Assessment, Washington D. C. 186 pp. Nilsson, J., and E. B. Cowling. 1992. A comparison of some national assessments, pp. 463-517. I__nnT. Schneider, ed. Acidi~cation Research: Evaluation and Policy Applications. Elsevier, Amsterdam. Russell, M. 1992. Lessons from NAPAP. Ecological Applications 2(2)" 107-110. Schindler, D.W. 1992. A view of NAPAP from north of the border. Ecological Applications 2(2):124-130.
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
321
Exceedence, Damage and Area Minimisation Approaches to Integrated Acidic Deposition Modelling Clair Gough a, Johan Kuylenstiernaa, Peter Baileya and Michael J. Chadwick b aStockholm Environment Institute at York, University of York, York YO 1 5DD, UK bStockholm Environment Institute, Box 2142, S-103 14 Stockholm, Sweden
Abstract Optimisation procedures used in abatement strategy models have made use of the relaxation of targets if feasible solutions cannot be found. However, exceedence, damage and area minimisation approaches provide alternative options. This paper explains and demonstrates how these approaches can be used to develop acidic deposition abatement strategies in Europe. The Coordinated Abatement Strategy Model (CASM) structure and input data are described and results for the alternative optimisation procedures are presented. The paper concludes with a discussion of the advantages of the different options available for developing cost-effective abatement strategies.
1. INTRODUCTION Acidification of terrestrial and aquatic ecosystems is recognised as one of Europe's most significant international environmental problems. It is accepted that a major cause of this ecological damage is the transboundary atmospheric transport of certain gaseous pollutants - oxides of sulphur and nitrogen (SO x and NO x) and reduced nitrogen (NHx). The main sources of SO x and NO x are processes involving the combustion of fossil fuels, whilst NH x is released predominantly from agricultural sources. The United Nations Economic Commission for Europe (UN-ECE) Convention on Long-range Transboundary Air Pollution (CLRTAP) has recognised this problem in the Helsinki and Sofia Protocols which specify flat rate reduction levels for emissions of SO 2 and NO x respectively, for each Party to the Protocol. In June 1994, the Oslo Protocol on further reductions of sulphur emissions (UN-ECE, 1994) was signed and this marked a significant change in approach. This Protocol differentiates national emission reduction obligations according to the relative harm emissions cause and the relative cost of controlling emissions. Integrated assessment models have played a key role in this process by providing optimised abatement strategies from which the final policy has been negotiated. Three models have provided analyses of abatement strategies to the UN-ECE Task Force on Integrated Assessment Modelling (UN-ECE, 1993). These are: ASAM (ApSimon and Warren, 1992), RAINS (Alcamo et al., 1990) and CASM (Gough et al., 1994) Each utilises input data specifying emissions, abatement costs, atmospheric transport and deposition targets based on critical loads but the models differ in their mode of operation. The CASM model, detailed here,
322 was developed by the Stockholm Environment Institute and uses linear programming optimisation techniques to generate cost-effective, environmentally targeted strategies; its main strength lies in its flexibility and the choice of optimisation approaches that it offers.
2. CO-ORDINATED ABATEMENT STRATEGY MODEL
CASM is shown diagrammatically in Figure 1. The input data requirements include base case quantifies of emissions, abatement cost curves for emissions sources, air transfer coefficients, receptor sensitivity data, and problem definition specifications. Calculations have been restricted to analysis of the abatement of sulphur dioxide emissions for the work for the Oslo Protocol; future developments w121extend the model to include other long-range transboundary air pollutants.
Emissions
Transfer Coefficients
Fi~ ~taining linear equations
Optimal Solution
Mapsshowing
effects in Europe
Figure 1. CASM flow diagram.
Critical loads
323 Estimates of emissions for the target year (in this case, 2000) are made from projected energy balances for each of 35 source regions. Marginal abatement cost curves are produced using a detailed database of point sources in conjunction with the energy forecasts and national cost data. Th~se specify estimates of annual abatement costs of different options of sulphur removal. They consist of cost steps in order of increasing marginal cost and only describe costs for combinations of "bolt-on" technologies. The assumption behind the use of these cost curves is that abatement measures will be applied in leastcost order and without structural changes in a region's energy consumption. One reason behind this last assumption is that official energy forecasts are used; any provision for fuel switching or energy efficiency measures should be taken into account as part of these energy balances. It is also likely that such changes would be motivated by reasons other than sulphur abatement for acidification (for example, control of greenhouse gas emissions) and to allocate costs becomes complex. In order to consider such measures the use of specific energy scenarios are generally involved; these incorporate these issues and may still be used in conjunction with abatement cost curves. Atmospheric transfer coefficients have been taken from the EMEP model 0versen et al., 1989, 1991) which provides estimates of annual deposition from each source region across a European grid of 150 km by 150 km (the EMEP grid). Critical loads applied to the SEI map of relative sensitivity of ecosystems to acid deposition have been used (Chadwick and Kuylenstiema, 1990). Atmospheric deposition of base cations has a neutralising effect on acidic deposition and a methodology for estimating this contribution for each EMEP grid square has been developed. With the exception of the EMEP data, all data used to generate the results presented in this paper in the model have been prepared by SEI but data from other sources (for example "official" data) may be used in CASM. There are two broad categories of approaches for deriving abatement strategies; they may be source-based (non-targeted approaches) or receptor-based (targeted approaches). Non-targeted approaches take emission standards, reduction levels or other emission source constraints and find the least-cost solution; these approaches do not incorporate environmental constraints. Targeted approaches, however, aim for a least-cost solution but include deposition constraints which may be based on the sensitivity of the environment to acidic deposition. Thus the destination and relative effects of a source's emissions are taken into account in the optimisation procedure. The optimisation problems discussed here are developed as sets of equations which can be solved using linear or integer programming techniques. In each case, there is one "objective function" (e.g., to minimise the total cost of abatement, or to minimise the deposition in excess of target loads) that is optimised subject to various "constraints" (e.g., that the deposition in a particular receptor cannot exceed a specified value). Linear Programming (LP) is a methodology for expressing problems in a series of linear equations and then finding a solution that optimises (maximises or minimises) the specified objective. The restriction of the problem formulation to linear equations ensures that one can always find the best solution - if any solution does exist. The algorithm for finding the best solution is known as the Simplex Method (see Walsh, 1971; Williams, 1985) and is implemented in the CASM model with a commercial package called LINDO (Schrage, 1989). 3. APPROACHES TO DEVELOPING ABATEMENT STRATEGIES The Oslo Protocol is different from the Helsinki and Sofia Protocols since it has aimed to incorporate the relative environmental implications of European abatement strategies and has been developed using information provided by optimisation models. This marks a change from the earlier, source-based, approaches which adopted uniform emission cutbacks in all countries.
324 Strategies in which emission levels are specified initially may be investigated using the CASM model, which provides information concerning the cost and environmental performance of the strategy. However, the main advantage of this model is that it provides a decision maker with a choice of policy goals (objectives) upon which to optimise. The types of objective that may be considered and how these are implemented in the model are discussed here. The difference between the approaches lies in the choice of objective function - the attribute of the strategy to be minimised. The way the problems are formulated using the abatement cost curves ensures that the most cost-effective solution will always be found, such that, even if cost does not form the objective function, there will not be an alternative solution that would achieve the same level of the objective for a lower cost than the final solution. The approaches available are described and these cover the range between source-based approaches and receptor-based approaches.
(i) Uniform Percentage Reductions (UPR) This is the type of approach adopted by the Helsinki Protocol in which each source region reduces its emissions by a fixed percentage relative to a base year. It is a source-based approach and does not involve optimisation techniques. CASM may be used to estimate the costs of implementing this type of strategy and to simulate the resulting deposition pattern in Europe.
(ii) Emissions Minimisation This is a source-based approach and incorporates emissions or the environmental status of receptors. of abatement is achieved in Europe for a given cost. achieved most economically does not necessarily reduced.
no information concerning the destination of It simply ensures that the maximum amount Directing abatement to areas where it can be ensure that the most harmful emissions are
(iii) Targeted Cost Minimisation In this case the objective is to minimise the total cost of abatement, whilst meeting deposition targets at the receptors. This approach is attractive in that it provides for sustainable levels of deposition at all receptor locations, but some drawbacks must be recognised. Most importantly, for actual data sets, where deposition targets are set at scientifically determined critical loads and emissions reductions cannot be entirely eliminated, one may find that it is not technically feasible to meet all receptor targets, i.e. even after all of the "available" abatement is implemented one or more receptors continue to receive deposition in excess of the critical load. If this is the case, then the analyst may decide to loosen some, or all, of the receptor targets, or to provide for higher levels of abatement through changes in the energy system. An example of a targeted cost minimisation approach where the receptor targets have been relaxed from the critical load value is the Gap Closure approach (UN-ECE, 1993). In this approach the receptor targets are relaxed to a fixed reduction point between present deposition levels and the critical load (for example, 60 per cent). The Oslo Protocol was based on results from a 60 per cent Gap Closure between 1990 deposition levels and 5 percentile critical loads (the 5 percentile critical load is the value that ensures the protection of at least 95 per cent of the area of an EMEP grid square)~ if current deposition is below the critical load no modification is required.
325 This approach constitutes a stage between source-based strategies and optimised receptorbased strategies. Although the solution is driven by receptor targets based on critical loads, these targets are governed by additional factors to the environmental sensitivity of a receptor, for example, the Gap Closure targets are influenced by the level of abatement already in place in some countries. In any receptor, exceedence will be reduced beyond the 60 per cent target only as a by-product of achieving neighbouring targets; there is no pressure on the model to go beyond the specified targets. Gap Closure provides a useful means of defining intermediate targets externally to the optimisation process.
(iv)Exceedence Minimisation Exceedence minimisation is a receptor-based optimisation approach. There is no requirement for environmental targets to be modified if they cannot be achieved - critical loads may be used. The model reduces the total exceedence in Europe to the smallest amount achievable within a given total cost. Receptor sensitivity may be described by single square values (e.g. the 5 percentile) or deposition may be compared to an area weighted distribution of classes of critical loads (the "subsquare approach"). In LP terms this is a goal programming approach (Nijkamp, 1980; Williams, 1985), which is a technique used for solving problems allowing the relaxation of certain constraints (in this case the achievement of critical loads). The amount by which these constraints are violated is minimised, for example, critical load exceedence. Goal programming techniques have been used in the development of abatement strategies for control of acid deposition in the past by Ellis (1988). Ellis's aim was to determine opportunities for cost savings through the relaxation of deposition targets given a level of uncertainty associated with the atmospheric transfer models. Exceedence Minimisation adopts the assumption that ecosystem damage is proportional to critical load exceedence. The definition of critical loads does not specify the shape of the response of a receptor in excess, but there is empirical evidence that damage to ecosystems is related to exceedence of critical loads. Several studies have documented instances of acidification in Europe (GEMS, 1988-1991; Kuylenstiema, 1993; Rosseland and Henriksen, 1990) and comparisons of data from these studies with modelled exceedence for these locations have revealed correlations between damage and exceedence (Kuylenstiema, 1993). Minimising the total exceedence in Europe subject to a total cost constraint generates a strategy that brings deposition as close as possible to the goal of reaching critical loads throughout Europe. Statements concerning the shape of such a dose-response function should be treated with caution at this stage, but existing research does not rule out the exceedence model of damage, illustrated in Figure 2.
Damage
Sensitivity
/
/'~
/
Sensitivity
Deposition Figure 2. Damage function used in Exceedence Minimisation.
326
(v) Damage Minimisation Exceedence Minimisation considers all receptors to be equal, in terms of the potential damage expected from a unit of exceedence. It is possible to differentiate between receptors by applying weights to different classes of sensitivity and hence prioritise reductions in exceedence over, for example, more sensitive receptors. Another way of viewing this is to consider that these receptorbased models incorporate different types of damage function. Exceedence Minimisation assumes a linear damage function of equal slope across all receptors. Damage Minimisation assumes a linear response with different slopes for different critical loads or receptors, shown in Figure 3. The objective of this approach is to minimise the total weighted exceedence in Europe, where the weights describe the slopes of the damage function. Although at present there is insufficient evidence to suggest that this is an accurate model of ecosystem response, the method may still be useful for exploring the implications of prioritising the protection of certain areas and to accommodate developments in dose response analysis. These receptor-based approaches are expressed in terms of assumed damage responses; in this sense the Gap Closure optimisation does not conveniently fit into this type of analogy.
Damage Sensitivity f
Deposition Figure 3. Damage function used in Damage Minimisation.
(v) Area Minimisation Assuming deposition is not expected to be reduced to below critical loads in every receptor, this approach could be seen as a literal interpretation of the definition of critical loads. The objective is to minimise the area over which critical loads are exceeded; abatement is priodtised at sources through which deposition can be reduced to below critical loads. The assumption behind this approach is that once the critical load has been exceeded there will be no benefit to an ecosystem of limiting any excess deposition. This model of dose response is illustrated in Figure 4. Area Minimisation introduces a new type of formulation, requiring a different solution technique. In standard linear programming, the variables can take on any value in the continuous range from zero to their specified upper limits. With Integer Programming (IP), some of the variables are permitted to have only integer values in the solution. This is especially desirable if a variable represents a discrete countable number of some entity, such as number of receptors exceeded. This additional requirement greatly complicates the solution process, since the Simplex Method can no longer be relied upon to find the optimal solution. The number of potential solutions for an
327 IP can be very large and difficult to solve. A number of algorithms have been developed to solve these problems in a reasonable time. The most common method used for Integer Problems is called the "Branch and Bound" (B&B) technique (Williams, 1985); this method is used in CASM (Schrage, 1989). Although the B&B technique is much more efficient than simply enumerating the possible solutions, it is still much more computationally expensive than the solution of an LP using the Simplex Method. On further consideration this may not appear to be such an appropriate method. It may be very costly to implement small reductions necessary to reach critical loads in some receptors when, for the same cost, large reductions in exceedence could be made elsewhere. This places a great emphasis on the certainty of both the critical load values and deposition estimates; the model solution would be highly sensitive to changes in these values. The concentration on achieving critical loads in some areas may also result in areas of extremely high exceedence, i.e. in those locations where critical loads are infeasible targets, no effort will be placed on reducing deposition.
Damage
,
sensitivity
sensitivity
class 5
class w,-
4 w-
Deposition Figure 4. Damage function used in Area Minimisation.
4. COMPARING THE PERFORMANCE OF OPTIMISED STRATEGIES Results from the CASM model are presented for the approaches listed above to illustrate the implications to the abatement strategies they derive. These have all been implemented using SEI data wherever possible, including the critical loads data. These results are presented to provide an indication of the consequences of adopting these approaches; since they have not been calculated using official data, they differ from those results presented to the UN-ECE for work within the Convention and should not be compared directly with similar results in other documents. Each type of run has been analysed across a range of abatement levels. The initial point of comparison has been a series of uniform reduction levels [approach type (i)]; the total cost estimated by the model for implementing these strategies have subsequently been used as budget constraints where one is required [approaches (iv), (v) and (vi)]. Two targeted cost minimisation [approach (fii)] results are presented for targets set to 50 and 60 per cent Gap Closure levels. The two most commonly used criteria for assessing the environmental performance of abatement strategies are total excess of critical loads and area of critical load exceedence. Plots of these indicators against total annual cost of implementation are presented in Figures 5 and 6, which illustrate European totals for the entire range of runs carried out. The presentation of the results over the two different evaluation criteria (amount and area of critical load exceedence) demonstrates how the choice of strategy will change according to the policy goal used in the model. The benefits, in these terms, of adopting an optirnised targeted approach over a source-based approach
328 are immediately apparent. In Figure 5, as the strategy has been defined on this basis, Exceedence Minimisation clearly offers the most efficient solution if reducing exceedence is taken as a policy goal; similarly, in Figure 6, Area Minimisation provides the solution leading to the smallest area of exceedence. The position of the third approach, targeted cost minimisation (Gap Closure), should also be noted. In both Figures these strategies lie on the "upper" edge of the targeted strategies, between the receptor-based and source based approaches. The graphs illustrate that, for a given expenditure, there are great benefits in terms of environmental protection to be gained through taking an optimised targeted approach. As the total cost of these strategies increases, i.e. as more sources approach their maximum feasible reductions, these lines begin to converge. This happens as there become fewer alternatives for locating emission reductions. The implications for critical load exceedence under these strategies may be seen more clearly in the maps shown in Figures 7 to 9. These also give an indication of the differences in the amount of exceedence in the different regions. The improvement, in terms of reduced exceedence, gained by adopting a targeted approach is obvious. The advantage of Exceedence Minimisation over Gap Closure shows up on these maps in the Kola Peninsular, Bulgaria, the Alps, Spain and the UK; these benefits are reflected in the level of effort required in these regions.
Cost ($miUion)
14000 UPR 12000
;
\
- - Emissions Min
,
\'-,
- - X ~ - - - Gap Closure ......
10000
:~
\
Exeeedence Min - - Damage Min
",
Area Min
\,,~'- . . .
8000 kk)' " 6000
~
~
'-,
" . . . ~
4000
~
",, ~
2000
0
0
t
i
t
I
I
t
I
100
200
300
400
500
600
700
Figure 5. Summary of model results.
I
sooe x c e s s
(
I
I
tqo
1000
329
c o s t t , rnuuon ) 14000 UPR - - Emissi(ms Min
12000
10000
~g~
Gap Closure
......
Excee~ence Min
-- -- -- Damage Min
\~,
~-a
\
-..
.... Area Min.
8000
6000
~
-..<
_ "-<
:...:
4000
2000
0
#
I
I
#
I
1
2
3
4
5
I
#
6 % a r e a exceeded7
Figure 6. Summary of model results.
g of S m2 per yr 0
iiiiiiiiiio - o.s ~ ! i 0 . 5 - 1.0 ..... 1.0 - 2 . 0 >2.0
Figure 7. Critical load exceedence after 60% Uniform Percentage Reductions.
330
g of S mY, per y r 0
iiii!iiiiiiioo.s ~0.51 . 0
-
>2.0 q
9
Figure 8. Critical load exceedence after 50% Gap Closure.
.._.,
g of S m2 per yr ::::::::::::i:::::: 0 - 0.5
:~
~o.s-1.o
~
1 . 0
-
2.0
m>2.o 9
" ~
Figure 9. Critical load exceedence after Exceedence Minimization.
1.0 2.0
331 5. DISCUSSION The examples of model runs presented in this paper have shown the potential for developing cost effective abatement strategies using optimisation models. These can achieve significant improvements in the environmental effectiveness of a strategy for equivalent levels of expenditure in Europe. The most efficient of these approaches are those that take a receptor-based approach and derive cost effective solutions according to different environmental policy goals. However, the notion of efficiency here applies to considering Europe as a whole. The intermediate to such approaches, Gap Closure, has the advantage that the proportional reductions in exceedence are equal in every country. This does not necessarily guarantee that the environmental protection in every country will be equal or that there are not other strategies that benefit more regions. These issues arise because there is no straightforward solution to the acid rain problem in Europe with different relative sensitivity of environments, and non-uniform mixing and transport of pollutants across national boundaries. It has been shown that integrated models can provide the decision maker with valuable information to enable the best to be gained from international policy. Before a single approach is adopted the decision maker must decide exactly what the aim of the strategy is, what it is to achieve and how the benefits are to be assessed. 6. REFERENCES
10 11
UN-ECE, Protocol to the 1979 Convention on Long-range Transboundary Air Pollution on Further Reduction of Sulphur Emissions. United Nations Economic Commission for Europe. ECE/EB.AIR/40 (1994). UN-ECE, State of Transboundary Air Pollution: 1992 Update. Air Pollution Studies Number 9. ECE/EB.AIR/34. United Nations Economic Commission for Europe, Geneva (1993). H.M. ApSimon and R.E Warren, Abatement strategies for sulphur dioxide, and analysis of the role of emissions from Central and Eastern Europe. Idtij~r~ 96 No. 4. 207-21 (1992). J. Alcamo, R. Shaw and L. Hordijk (eds.), The RAINS Model of Acidification, Science and Strategies in Europe. Kluwer Academic Publishers, Dordrecht (1990). C.A. Gough, ED. Bailey, B. Biewald, J.C.I. Kuylenstierna and M.J. Chadwick, Environmentally Targeted Objectives for Reducing Acidification in Europe. Energy Policy (in press) (1994). T. Iversen, J. Saltbones, H. Sandnes, A. Eliassen and O. Hov, Airborne Transport of Sulphur and Nitrogen over Europe - Model Descriptions and Calculations. EMEP/MSC-W Report 2/89. Oslo (1989). T. Iversen, N.E. Halvorsen, S. Mylona and H. Sandnes, Calculated Budgets for Airborne Acidifying Components in Europe, 1985, 1987, 1989 and 1990. EMEP/MSC-W. Technical Report No. 91, Oslo (1991). M.J. Chadwick and J.C.I. Kuylenstiema, The Relative Sensitivity of Ecosystems in Europe to Acidic Depositions. The Stockholm Environment Institute, Stockholm (1990). G.R. Walsh, An Introduction to Linear Programming. Holt, Rinehart and Winston Ltd. London (1971). H.P. Williams, Model Building in Mathematical Programming. Wiley Interscience, Chichester, UK. 20-30 (1985). L. Schrage, User's Manual for Linear, Integer and Quadratic Programming with LINDO. The Scientific Press, Redwood City, California (1989).
332 12 13 14
15 16
E Nijkamp, Environmental Policy Analysis, Operational Methods and Models. John Wiley and Sons, New York (1980). J.H. Ellis, Acid Rain Control Strategies, Options exist despite uncertainties. Environmental Science and Technology. Vol. 22, No. 11. P 1248 (1988). GEMS, Forest Damage and Air Pollution. Reports of the 1987-1990 forest damage surveys in Europe. Convention on Long-range Transboundary Air Pollution -Intemational Cooperative Programme on Assessment and Monitoring of Air Pollution Effects on Forests. UNEP/ECE, Geneva (1988 - 1991 ). J.C.I. Kuylenstiema, Assessment of Ecosystem Sensitivity to Acidic Deposition: Critical Load Estimates, Detriment and Damage. PhD Thesis (Unpublished) (1993). B.O. Rosseland and A. Henriksen, Acidification in Norway - loss of fish populations and the 1000-lake survey 1986. The Science of the Total Environment 96,45-56 (1990).
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BE All rights reserved.
333
Reliability of environmental information obtained by modelling and monitoring J.A. Hoekstra, J.C.H. van Eijkeren, A.L.M. Dekkers, B.J. de Haan, P.S.C. Heuberger, P.H.M. Janssen, A.U.C.J. van Beurden, A.A.M. Kusse, M.J.H. Pastoors National Institute of Public Health and Environmental Protection, P.O. Box 1, 3720 BA Bilthoven, The Netherlands
Abstract Environmental information on cause-effect chains is obtained by combining results of measurements and survey data with statistical and process oriented models. For proper use of the information it is important to have a good understanding of its reliability. This paper addresses the problem of assessing the reliability of information obtained from different models and measurements. A reliability factor is introduced for standardized quantification of reliability. The approach is illustrated by results of a socalled "nitrogen chain", starting with estimates of nitrogen emission and ending with predicted future concentrations in untreated water at drinking water stations.
1. INTRODUCTION Environmental information is often a composition of results from different sources. This certainly holds for the information provided by the National Institute of Public Health and Environmental Protection (e.g. RIVM, 1994). The information includes amounts of pollutants emitted, concentrations in air, water and soil, and related effects on human and ecosystem health. It also involves predictions of future emissions, concentrations and effects. The data are collected using monitoring networks, surveys and process oriented models, and constitute an important ingredient in the support of the policy-making process of the Dutch government. Insight in the reliability of the information would be an extremely valuable addition. This insight will also help in optimizing the monitoring networks and modelling activities. This paper discusses the problem of assessing the reliability of information obtained from different types of models and measurements. The problem will be illustrated by a practical case, a so-called "nitrogen chain", starting with estimates of nitrogen emission and ending with predicted future concentrations in untreated water at drinking water stations. Section 2 gives an overview of the studied system. In section 3 a reliability factor is defined for quantifying the reliability of the data in an uniform way. In section 4 it is indicated how the reliability factor can be calculated. Section 5 presents some results for the nitrogen chain, which are discussed in section 6.
334 2. THE NITROGEN CHAIN Nitrogen is emitted into the atmosphere by traffic, industry and intensive animal husbandry. After transport, these pollutants are deposited to natural soils, farm lands and urban areas. Added to the directly applied fertilization on farm lands, the excess nitrates percolate to the groundwater. High nitrate concentrations in groundwater threaten its exploitation for drinking water production. Figure 1 shows the data production process related to this causal chain in a simplified form. Grey boxes refer to monitoring networks and surveys, open boxes to mathematical models. Arrows denote physical processes like transport and deposition. Bold arrows indicate that the reliability of the models simulating such processes is investigated in this study. Broken arrows mark processes that have not been considered here. Figure 1 shows that reduced nitrogen concentration data result from a model with emission survey data on animal husbandry as a main input (Asman and Van Jaarsveld, 1990), while oxidized nitrogen concentration data stem from a monitoring network (Erisman, 1992). In the first column emission registration data are given. E.g. for animal husbandry, the data include number and type of animals per farm and stable type. Based on these data estimates of NH3-emission are obtained for each municipality in the Netherlands (Van der Hock, 1994). These emission data are reallocated to a 5x5 km 2 grid. The reallocation scheme draws on a high resolution land use database. Ambient air concentrations, second column, are processed into dry deposition data using a model (Erisman, 1992). Wet deposition is obtained from a monitoring network. The nitrate load to the groundwater below urbanized regions and natural vegetation soils, third column, is also obtained by measurements. The thin arrow pointing to the box 'natural soils' symbolises the actual use of the 1986 deposition data in a regression model with nitrate concentration in freatic groundwater as response variable (Boumans, 1994). Finally, the loads to freatic groundwater, fourth column, feed a model describing the transport to the inlet of the drinking water production plants (Lieste, 1989). In summary, this study involves the following types of data and models: emission registration data per municipality a model to reallocate these data to a grid at the appropriate scale monitoring networks in combination with interpolation schemes - regression models obtained from case studies - process oriented models describing transport, transformation and deposition - land use data, meteorological data, parameter values for process models c.q. interpolation schemes, etc. -
-
-
This nitrogen chain consists of several chains of models and monitoring systems starting with the actual polluting activities. Most important to the quality of untreated water is the short chain starting with agricultural fertilization loads.
335
emission registration
ambient air concentration
I .......... ~'~
deposition on soil surface
NO3' load to groundwater
NO~' concentration in groundwater
NO~ ~et NOx ary
NHx dry
.
.
.
.
j orb., .
.
,
_[
.
.
.
.
soils
natural
Unw ette
NHx ~4~f
NO3'
T
..J agricultural "-] soils
1 The data production process. Grey boxes: obtained from monitoring networks and surveys. Open boxes: data from process oriented models. Bold arrows indicate for which models reliability analyses have been carried out.
Figure
3. T H E R E L I A B I L I T Y F A C T O R A standardized measure of reliability is helpful to obtain a complete picture of the quality of information rendered by data and models.
Estimated and "true" value Suppose yO is the value or set of values for which the reliability is to be assessed. Examples of yO are: - estimated average N-deposition (mol/ha) on 5x5 km 2 in a given year estimated total N-input into soil by agriculture (million kg) in a given year predicted nitrate concentration (mol/1) at a given drinking water station in 2100. The symbol y* represents the corresponding unknown "true" value. Intuitively, reliability should measure the nearness between yO and y*. The concept of reliability is
336 only meaningful if the true value is clearly defined. This implies that one should be specific about the intended temporal and spatial representativity of the information. When a value in the distant future is presented, it is difficult to define precisely what is meant by the "true" value. Scientists generally do not have tools to predict e.g. the actual N-concentrations in 2100. Therefore, the "true" value meant in the prediction procedure is not what the N-concentration will really be in 2100. Instead, the prediction is calculated under specific and limited model assumptions, assuming "all other things being equal" (which we know not to be the case). A reliability analysis therefore has equally limited scope; it should specify clearly which parameters and input are varied and which are kept constant. Such limited predictions can still be very useful. The calculated predictions will render important information by comparing the influence of different environmental strategies on the future state of the environment.
Systematic and random error The value yO is considered to be the outcome of a stochastic process Y, which could generate a number of outcomes each with a given probability. For instance, Y represents a measurement process with error or a computer model with imprecise knowledge of the parameter values and/or stochastic input. Figure 2 shows two examples of distributions of Y, together with the true value y*. In Figure 2a the process generates outcomes with a small variance and a large systematic error or bias, as y* is not in the centre of the distribution. In Figure 2b, the systematic error is much smaller, but the random error is larger.
b: large random error
a: large systematic error
0 c
Off) k..
b_
y*
Figure 2
Y Frequency distribution of Y, with "true" value y*
y*
337 The reliability factor
Mostly, the distribution of Y is skew to the right, as in Figure 2, and takes on only positive values. In this case a relative error is an appropriate measure of variation. Moreover, relative errors are dimensionless and more easy to compare for different variables than absolute errors. The reliability factor is therefore based on the geometric standard error:
S geometric
_ _
ll/,og~
exp n
(1) I1 n 0 - e x p n;__~lOogyi-(logv~
+ ((logy~
,)2
= expCvariance(logy~ +bias(logy 0)2
where Ylo,Y2,...,Yn o o represent realisations of Y, e.g. measured or computed values, and 10gy~ = 1 ~ n
logyO.
The reliability factor is defined as: B(yO) : 1/s g. . . .
tric
(2)
B(y ~ takes on values between 0 (extremely tmreliable, value completely unknown) and 1 (exactly known). It is closely related to the reliability index proposed by Leggett and Williams (1990) for comparison of model predictions with measurements, and to Kirkwood's geometric measure of dispersion for comparing variables with different dimensions (Kirkwood, 1988). If Y is approximately lognormal and contains no systematic error, the factor B can be used to construct a confidence interval for y*:
68% Confidence Interval: By ~ - y~ 95% Confidence Interval: B2y ~ -
y~
4. METHOD OF CALCULATION Usually y* is unknown. In that case, only the variance component of the error can be assessed. If yO is obtained by application of a statistical method to measurement data, the variance is given by the statistical procedure. E.g., a field of nitrogen oxide concentrations is obtained by applying Kriging to results of a monitoring network, and the Kriging error model provides standard errors for each grid cell (Cressie, 1991). The model
338 assumptions can to some extent be checked using residual plots and cross-validation. If the measurements are unbiased and have the correct spatial and temporal representativity, it can be assumed that the systematic error is small in comparison with the random error. Especially the second condition: spatial and temporal representativity is difficult to fulfil. To calculate the reliability factor for outcomes of process models, two different approaches can be followed:
.
"External" examination by comparison with real measurement data. Sometimes good quality measurements are available to validate the model. The measurements can then be used as a substitute for the true value y* in the above formula. The reliability factor will contain both variance and bias. "Internal" examination using Monte Carlo methods as implemented in the simulation package UNCSAM (Janssen et al., 1992). A set of outcomes is generated using a probability distribution for uncertain model parameters, and input (error propagation). These data are used to calculate the variance of the outcome of the process model. The nominal value, y0.... is usually not in the centre of the distribution of outcomes. The bias part in the formula above is replaced by (Iogyn0om- 1ogy0)2. Evidently, expert judgement plays an important role in specifying the probability distributions of the parameters (including correlations). Therefore, the internal reliability factor contains an expert judgement element as well. The two approaches produce complementary information as is exemplified later.
5. PRELIMINARY RESULTS Table 1 gives reliability factors for some elements of the nitrogen chain described in section 2. The analysis is in a preliminary stage. Therefore, the results are rough estimates, presented only to illustrate the discussion. More detailed results will be given in Hoekstra and Heuberger (in prep). Reliability factors were calculated for each grid cell separately. The presented values in the table are the medians of the reliability per cell; large fluctuations around this median reliability do exist.
6. DISCUSSION The reliability factors presented in Table 1 should be interpreted with care, since their values heavily rely on subjective choices concerning parameter uncertainties, neglecting potential systematic errors due to erroneous model specifications etc. Improvements are being made, by considering more parameter uncertainties and carrying out more model checks. However, the intrinsic problem: that frequently the "true" values are not known and therefore formula 1 is applied with substitutes of y*, cannot be solved. Therefore, it should be realized and accepted that reliability factors can have a considerable uncertainty themselves too.
339 Table 1
Reliability factors for elements of a nitrogen chain (annual averages). Preliminary results. Type of Data Emissions N-load on arable soil (5002 m 2) NH3-emissions to air (52 km 2) Air NH3-concentrations (52 km 2)
Uncertainty Analysis
process model process model
internal internal
process model
B
0.8 0.8"
internal external process model internal monitoring network internal process model internal
0.8 0.6 0.6 0.9 0.8
internal internal external
0.4 0.5 0.7
internal
0.6
NHx-depositions (52 km 2) NO2-concentrations (52 km 2) NOy-depositions (52 km 2) Soil/groundwater Nitrate in freatic groundwater (5002 m 2) nature field measurements agriculture process model Predicted value at a drinking water station in 2100 Nitrate in untreated water process model
" Partially based on expert judgement. Other estimate: 0.6 (Erisman, 1992) This being stated, what can be concluded from the data in Table 1? First of all, the provided environmental information can be highly uncertain. Some variables have a reliability factor of 0.5, which means that the variable is not known within a factor 2. A 95% confidence interval would even run from y~ to y~ Considering the small differences between target and limit values for environmental quality parameters, this is a large uncertainty. One should not be surprised by the large uncertainty: in several cases the local predictions in Table 1 are obtained by models with parameters that are held constant over large regions out of need rather than truth. This raises the question whether the models are appropriate for this type of prediction, and vice versa, whether information on such a fine scale is really necessary. The NO2-concentrations in air have higher reliability, because this variable exhibits comparatively modest local variation. Comparison of internal and external reliability factors is interesting from a validation point of view. The discrepancy between the two reliability factors for NH3-concentrations in air, indicates that uncertainty on input and parameters as presently considered feasible, is not sufficient to explain the limited performance of the process model in matching the NH3-measurements. Further analyses will be carried out to study spatial variation in reliability and to include external reliability factors and new environmental variables.
340 7.
REFERENCES
Asman, W.A.H. and Jaarsveld, J.A. van (1990). A variable-resolution statistical transport model applied for ammonia and ammonium. RIVM-report no. 228471007, Bilthoven, The Netherlands Boumans, L.J.M. (1994). Nitrate in freatic groundwater under natural area's on sandy soils in the Netherlands. RIVM-report no. 712300002, Bilthoven, The Netherlands. (In Dutch). Cressie, N.A.C. (1991). Statistics for spatial data. John Wiley & Sons, New York. Drecht, G. van, (1993). Modelling of regional scale nitrate leaching from agricultural soils in the Netherlands. Applied Geochemistry, Suppl. Issue, no. 2, 175-178. Erisman, J.W. (1992). Atmospheric deposition of acidifying compounds in The Netherlands. Doctoral Thesis, University of Utrecht, The Netherlands. Hoek, K.W. van der, (1994). Calculation method for ammonisa emission in the Netherlands in the years 1990, 1991 and 1992. RIVM-report no. 773004003, Bilthoven, The Netherlands. (In Dutch). Hoekstra, J.A. and Heuberger, P.S.C. (Eds) (in prep). The reliability of environmental information: an analysis of the nitrogen-chain. RIVM-report in preparation, Bilthoven, The Netherlands. (In Dutch). Janssen, P.H.M., Heuberger, P.S.C. and Sanders, R. (1992). UNCSAM 1.1: A software package for sensitivity and uncertainty analysis: Manual. RIVM-report no. 959101004, Bilthoven, The Netherlands. Leggett, R. and Williams, L.R. (1990). A reliability index for models. Ecological
Modelling 13, 303-312. Lieste, R. (1989). Computer Program FLOPZ1. Path lines in quasi-three-dimensional groundwater flow in a layered homogeneous aquifer. RIVM-report no. 728520004, Bilthoven, The Netherlands. Kirkwood, B.R. (1988). Essentials of medical statistics. Blackwell Scientific Publications, Oxford. RIVM (1994). National Environmental Outlook 3 1993-2015. Samson H.D., Tjeenk Willink, Alphen aan de Rijn.
F U T U R E OF A C I D I F I C A T I O N R E S E A R C H
S E S S I O N VII F U T U R E RESEARCH; COMBINATION WITH OTHER E N V I R O N M E N T A L TOPICS
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
343
Future of acidification research T. Paces Czech Geological Survey, Kl~irov 3, CZ-118 21 Prague, Czech Republic Abstract Acidification is not a simple process operating separately from other geochemical, biochemical and biological processes. It is a part of an extremely complex ecological development of our nature influenced by energy- hungry ever growing h u m a n population. Acidification as a general biogeochemical process will be implicitly studied in most of the future ecological research projects dealing with global climate changes and land use changes There are many scientific, technological and even political questions to be solved to fight acidic pollution effectively. It is just a question of the significance of acidification with respect to other environmental processes and issues. The most important questions to be answered by the future research fall into following topics: (1) Questions related to the effects of reduction of emissions and mitigation practices on ecosystem development and present status of improvement of water and soil quality, (2) acidification impact on climate change, (3) acidification during land-use changes and organic matter transformations, (4) acidification and its influence on trace metal toxicity, (5) acidification during biogeochemical cycles of nitrogen, carbon and sulfur (6) socioeconomic factors with respect to environmental acidification. Other miscellaneous topics should be subject to future investigations before the process can be fully controlled and its harmful effects eliminated.
1. I N T R O D U C T I O N Future of the research into environmental acidification will be determined by the past development of the subject. The occurrence of acid rain has a long history. It has been discussed since 1661 (Cowling, 1982). Smith (1852) wrote about air and rain of Manchester. An systematic monitoring and research of the process and its consequences to ecosystems started in seventies of this century following a news release and a scientific paper by Oddn (1967, 1968). The history of the development of the environmental topic is well documented by Gorham (1981, 1991). Internationally, acidification was recognized as a serious environmental problem in Europe through the United Nations Convention on Long Range Transboundary Air Pollution signed in 1979. Within this Convention, research was lunched to learn more about causes and consequences of acidification in Europe. The National Acid Precipitation Assessment Program (NAPAP) in the U.S.A. was created by Congress in 1980 (Irving ,1992). Almost $400 million have been invested in sponsored research by NAPAP till 1989 (NAPAP, 1989). Acidification of water and forest soils has been an important
344 research topic during the last 25 years by individual researchers, many government sponsored research programs, especially in Sweden (Anonymous, 1992), as well as by joint efforts of environmental programs lunched by the Commission of European Communities. The investigation of acidification has began later in Asia (Ogura N., 1992, Environmental Agency, 1993). Acid rain was reported even in an Amazon rain forest (Haines et al., 1983). During the last 25 years, detail knowledge has been gained about relationships between acidifying emissions of pollutants and physico - chemical and biological state of surface water, ground water and soil. It has been recognized that the intensified environmental acidification due to human activity is detrimental to aquatic and terrestrial ecosystems (e.g. Schindler, 1988, Schneider, 1992, Ulrich and Pankrath, 1983, Steinberg and Wright, 1994, Teller et al., 1992). The most important information for future development of the topic is that acidification is not a simple process operating separately from other geochemical, biochemical and biological processes. To the contrary, it is a part of an extremely complex ecological development of our nature influenced by energy- hungry ever growing human population. Acidification is not a global problem, however, it is a regional phenomenon occurring frequently in industrial districts with a high intensity. The process has a character of a time chemical bomb. Often, it proceeds unnoticed, until a buffering capacity of nature is exhausted, and acids become detrimental to various natural and man made materials as well as to organisms. It is phenomenon which will stay with us as far as people will burn fossil fuels. During the last two decades, scientists have gained a solid understanding of the process (Teller et al., 1992, Steinberg and Wright, 1994, Anonymous, 1993). In spite of that, there are still many scientific, technological and even political questions to be solved to fight acidic pollution effectively. 2. A C I D I F I C A T I O N I N A F R A M E W O R K ENVIRONMENTAL ISSUES
OF OTHER
Acidification is a physico-chemical and biological process which will be always active part of our dynamic environment. An excellent example of the position and significance of acidification within a broad range of other environmental issues was presented in a Dutch publication "Concern for Tomorrow" (Langeweg, 1989). Acidification is just a part of the complex interaction between biota, bedrock, soils, sediments, water and air and the interaction is influenced by social and economic factors. These relationships are illustrated in Fig. 1. Most of the past effort focused on the acidification as a result of burning of fossil fuels and the interaction of S02 and NOx with water and soils. Acidification in Scandinavia has been attributed not only to emission sources but also to n a t u r a l processes connected with changes in forestry practices (Rosenquist et al., 1980, Rosenquist I.Th., (1985). Attention was paid to the physiological influence of acid water on fish (Hultberg, 1983) and acid atmospheric deposition on forests (Ulrich and Pankrath, 1983). The importance of depletion of base cations in soils and toxicity of aluminum ions has been stressed by Ulrich (1981) and others. Much less is known about the response of organic matter to acidification, namely about the rates of release and fixation of
345
CO2 due to accumulation and decomposition of humic substances in soils.To w h a t extent acidification reduces biodeversity in aquatic and t e r r e s t r i a l ecosystems is known only in few cases.It is often assumed, that the changes in land use and climate can influence the rates of environmental acidification. Unknown is, whether regional acidification can influence the global climatic changes and future land use changes significantly. There is an u n c e r t a i n t y whether acidification is still a social and economic issue. Do people and decision m a k e r s care about acidification any more? What are the differences in the opinion within rich countries where substantial reduction of acid emissions has t a k e n place and poor countries where such measures cannot be implemented because of insufficient funding? Who is going to finance future research into acidification, critical loads of acidifying compounds and regional exceedences of the loads?
tl
I
GLOBAL CLIMATIC I CHANGES
LAND - USE CHANGES
\
? t l YES illll I
MINING BURNING
?tl,,,
ACIDIFICATION. WEATHERING. CHEMICAL EROSION
//
liI
BIOTA
%
tl r WASTE DISPOSAL
SEDIMENTS i
,,
I'i ,~_
RAW MATERIALS
WATER $
,,i
Fig. 1 Acidification within the broad scope of e n v i r o n m e n t a l a n d socioeconomic relationships. YES indicates where the relationship is assumed to be significant. ? indicates an uncertainty whether the feedback is significant or not. 3. A C I D I F I C A T I O N PROGRAMS
WITHIN
FUTURE
RESEARCH
Terrestrial ecosystem research initiative of European scientists 1994) introduces 6 major topics for future environmental research: 1. Land-use change 2. C02 atmospheric concentration and climatic drivers
(Manaut,
346 3. Soil organic dynamics 4. Biogeochemical cycles 5. Biodiversity and ecosystem functioning 6. Modeling and other integration activities. These topics should be developed across geographical gradients from boreal to M e d i t e r r a n e a n European environments and from maritime to continental climate gradients. The research effort should be interdisciplinary b e t w e e n biochemical, biophysical and socioeconomic fields. Another i m p o r t a n t feature will be the transition between various scales of resolution and global perspective of changing environment. Manipulation field experiments and long-term monitoring schemes will be necessary to reach meaningful results. Research focus will shift from sulfur cycle to carbon and nitrogen cycle and to the role of trace elements and trace organics. In the United States, the relations between the acid emissions and their effects have been presented to U.S. Congress in 1990 (NAPAP, 1993). The summary indicates that the understanding of acidification and its influence on forests, aquatic ecosystems, human health and materials and cultural resources is considered to be sufficient to reach meaningful policy-oriented a s s e s s m e n t s (NAPAP, 1991). The research and monitoring strategy will be probably directed towards answering questions contributing to the evaluation of a particular acid deposition control approach. The research should be directed towards better understanding of the impacts of pollutants on forest and ecosystem n u t r i e n t cycling along climate gradients, and towards r e g e n e r a t i o n of d a m a g e d ecosystems and forest resources. Ozone-acid rain interaction should be also an important research topic as well as the plant responses to complex impact of atmospheric ozone, NOx,NH3, SO2, CO2 and trace metal toxicities a n d deficiencies. It has been recognized, that we have an insufficient knowledge on the biochemical role of antioxidants and free radicals scavengers in oxidative metabolism. There is a general opinion among scientists that there is a need for a continual long-term monitoring program on sensitive ecosystem parameters. As stated by Rask and Hendershot (1992), "the researchers should a d a p t to changing priorities for research while granting agencies should recognize the benefit of uninterrupted time series". Small catchment studies are the tool to better u n d e r s t a n d i n g of the causes of the observed long-term changes in well defined ecosystems (Paces, 1992, Molclan and Cerny, 1994, Cerny et al., in print). The catchment and plot experimental manipulations are a new tool which has been well developed in Europe (Rasmussen et al., 1992) and has been considered for operation in the United States too (NAPAP 1991). Acidification as a general biogeochemical process will be implicitly studied in most of the future ecological research projects. It is just a question of the significance of acidification with respect to other environmental processes and issues. Future research should concentrate on quantification of nitrogen and carbon cycles at different scales from soil-water micro systems via small and large catchment studies toward continental and global scale synthesis of field data. More information is desirable on proton generation and consumption during the water pathways in order to interpret dynamics of biogeochemical reactions. The use of environmental isotopes D, 180, 13C, 15N, 348, 87/868r can be helpful in this respect (Cerny et al., in print). Biogeochemical studies will most probably relay more on m a n i p u l a t i o n experiments on more diverse v e g e t a t i o n types, particularly on deciduous woodlands, grasslands and wetlands. Acidification
347 should be considered together with the impact of temperature and variable levels of ozone and nitrogen. Reversibility of acidification should be studied in combination with the impact of increased summer draughts and winter rapid decreases in temperature. Future research should focus more towards combined effects of air pollutants and climate change (Beier and Cummins, 1992). Soil sinks for sulfur should be better understood to explain commonly found imbalance of sulfur in catchment budget studies and possible irreversibility in lowering the buffer capacity of acid soils. Acidification of water and soils is a recognized problem and it is sometimes considered as scientifically well understood process. Still many important aspects need further development. A recent Dahlem Conference addressed explicitly this topic (Steinberg and Wright, 1994). A summary of the most important topics for future research which appeared at the Conference is given in the Appendix. 4. C O N C L U S I O N S Future acidification research will not be an isolated effort. It is expected that it will become an integral part of research into aquatic and terrestrial ecosystems and that it will be more integrated with economic and political assessments. In spite of that acidification is a regional problem, it will be considered as one of the factors contributing to global climatic changes because of its impact on behavior of soils and forests. Acidification will be implicitly considered in the research dealing with biogeochemical cycling of potentially toxic trace elements, within nitrogen and carbon biogeochemical cycles and in studies of biodiversity. Ecological research along transects across Europe will necessary include acidification gradients. It will be especially important in transects including the Black Triangle of central Europe with acidification maxima at the German - Czech - Polish borders. Goal oriented research organized by individual governments and intragovernmental bodies, should be supported by an independent individual basic research. The new and often most important discoveries in the acidification and other ecological problems were reached by individual efforts by dedicated scientists (Odin, 1968) or small scientific teams (Likens and Bormann, 1974, Douglass and Hoover, 1988) rather than through large costly projects. On the other hand only the continuous, interdisciplinary and well organized projects are able to describe the environmental problems on continental and global scale. Such projects, when analyzing the behavior of the global systems, should apply the knowledge gained through basic research on local scale. 5. R E F E R E N C E S Anonymous (1982) Acidification today and tomorrow. Swedish Ministry of Agriculture, Environment'82 Committee. Anonymous (1993) The expert meeting on acid precipitation monitoring network in East Asia, Proceedings, October 26-28, 198"93, Toyama City, Environmental Agency, Government of Japan. Beier C., Cummins T. (1992) The future, and current limitations of manipulation and monitoring of terrestrial ecosystems. In: Experimental manipulations of biota and biogeochemical cycling in ecosystems. (L.
348 R a s m u s s e n et al., eds.), Ecosystem Research Report 4, 338-340, Commission of European Communities. Cerny J., Novak M., Paces T. and Wieder K. (eds.), (in print): Water, Air, and Soil Pollution, Special Issue. Cowling E. B. (1982) Acid precipitation in historical perspective. Environ. Sci. Technol., 16, No. 2. Douglass J.E. and Hoover M.D. (1988) History of Coweeta. in Forest Hydrology and Ecology at Coweeta (Swank W.T. and Crossley Jr. D.A., eds.), 17-31, Springer/Verlag. Environmental Agency (1993) Proceedings, the expert meeting on acid precipitation monitoring network in east Asia. Oct. 16-28, 1993, Toyama, Japan. Environmental Agency, Government of Japan, Toyama. Gorham E. (1981) Scientific understanding of atmosphere - biosphere interactions: A historical overview. In: Atmosphere-biosphere interactions: toward a better assessment of the ecological consequences of fossil fuel combustion. Capt. 2, 9-21, National Academy Press, Washington D.C. Gorham E. (1991) Atmospheric deposition to lakes and its ecological effects: /k retrospective and prospective view of research. Proceedings of Int. Syrup. Impacts of salinization and acidification on terrestrial ecosystem and its rehabilitation. (N. Ogura, ed.), 25-80, Tokyo University of Agriculture and Technology, Fuchu, Tokyo. Haines B., Jordan C., Clark H., Clark K.E. (1983) Acid rain in an Amazon rain forest. Tellus, 35B, 77-80. Hultberg H. (1983) Effects of acid deposition on aquatic ecosystems. Proceedings, preliminary edition, 167-185, Symposium, Acid Deposition a Challenge for Europe (Ott H. and Stangl H, eds.), Karlsruhe, CEC DG-XII, Brussels. Irving P.M. (1992) The United States national acid precipitation assessment program. In Acidification Research, Evaluation and Policy Applications (T, Schneider, ed.), 365-374, Elsevier Science Publishers. Langeweg Ir.F. (ed.) (1989) Concern for tomorrow. Rijksinstituut voor volksgezondheid en milieuhygiene, Bilthoven. Likens G.E. and Bormann F.H. (1994) Acid rain: a serious environmental problem. Science, 184, 1176-1179. Manaut J.-C. (1994) Unpublished letters. Moldan B., Cerny J.(eds.) (1994) Biogeochemistry of small catchments. John Wiley & Sons. NAPAP, (1991) Mission, goals, and program plan post 1990. NAPAP, Office of the Director, 722 Jackson Place, NW, Washington DC 20503. NAPAP (1993) NAPAP 1992 Report to Congress. NAPAP, Office of the Director, 722 Jackson Place, NW, Washington DC 20503. Oddn (1967) Dagens Nyheter, Stockholm 24. Oct., 1967. Oddn (1968) The acidification of air and precipitation and its consequences in n a t u r a l environment. Ecology Committee Bulletin No. 1, National Research Council, Stockholm. Ogura N. (ed.) (1992) Proceedings of Int. Symp. Impacts of salinization and acidification on terrestrial ecosystem and its rehabilitation. Tokyo University of Agriculture and Technology, Fuchu, Tokyo. Paces T. (1992) Monitoring for the future: integrated biogeochemical cycles in representative catchments. In: Acidification Research, Evaluation and
349 Policy Applications. (Schneider T., ed.), 145-159, Elsevier Science Publishers. Rask M., Hendershot W., (1992) Theme III: Aquatic ecosystem studies. In: Experimental manipulations of biota and biogeochemical cycling in ecosystems. (L. Rasmussen et al., eds.), Ecosystem Research Report 4, 341-343, Commission of European Communities. R a s m u s s e n L. (1992) Experimental manipulations of biota and biogeochemical cycling in ecosystems. Ecosystem Research Report 4, Commission of European Communities. Rosenquist I.Th. (1985) Acid rain, acid precipitation and acid soil in fresh water chemistry. Land use Policy, 70-73. Rosenquist I.Th, Jorgensen P. and Ruesl~itten H. (1980) The importance of natural H§ production for acidity in soil and water. Ecological impact of acid precipitation proceedings. SNSF Report. Int. Conference 1980. Schneider T. (ed.) (1992) Acidification Research, Evaluation and Policy Applications. Elsevier Science Publishers. Schindler D.W. (1988) Effects of acid rain on freshwater ecosystems. Science 239, 149-239. Smith (1852) on the air and rain of Manchester. Mem. Lit. Phil. Soc. Manchester, series 2, 10, 207-217. Steinberg C.E.W., Wright R.F. (eds.) (1994) Acidification of freshwater ecosystems, Implication for the future. Dahlem Workshop Reports, Environmental Sciences Research Report 14, John Wiley & Sons. Teller A., Mathy P., Jeffers J.N.R. (eds.) (1992) Responses of forest ecosystems to environmental changes. Elsevier Applied Science,, London. Ulrich B. (1981) Destabilisierung von Wald(ikosystem durch Akkumulation yon Luftverunreinigungen. Der Forst und Holzwirt, 36, 525-5322. Ulrich B., Pankrath J. (eds.) (1983) Effect of accumulation of air pollutants in forest ecosystems. D. Reidel Publ. Comp., Dordrecht
5. A P P E N D I X Topics for future research into environmental acidification (selected from Steinberg and Wright ,1994 and supplemented by the author). 5.1 Q u e s t i o n s r e l a t e d to the effects of r e d u c t i o n o f e m i s s i o n s a n d m i t i g a t i o n p r a c t i c e s on e c o s y s t e m d e v e l o p m e n t a n d the s t a t u s of i m p r o v e m e n t o f w a t e r a n d soil quality:
Sites responses to reduction of SO2 and increase in NOx and 03 Effect of liming on pool of exchangeable base cations and acidity in soils Ecological effects of substantial decrease in SO2 deposition in central Europe. 5.2 A c i d i f i c a t i o n and c l i m a t e c h a n g e s : Relationships between climate change and changes in hydrological pathways to environmental acidification.
350 Strong acid anion pulses due fluctuation in climatic conditions (e.g. changes in wetting and drying periods) as an analogue of some expected global climate changes. Release of N 2 0 during denitrification and nitrification influenced by acidification and global warming. N 2 0 is more radiatively active t h a n C 0 2 and CH4. Scaling-up of local or small catchment acidification studies to regional or even global scale ecological changes. 5.3 L a n d - u s e c h a n g e s a n d o r g a n i c matter: Changes in quantity and quality of organic acids with changes in land use and changes in anthropogenic acid deposition Vegetation changes or management other than afforestation as a cause of chronic (more t h a n 5 years) surface water acidification in the absence of acidic deposition? Organic acidification due to production, transport and behavior of organic acids in ecosystems Impact of acidic deposition on production and release of dissolved organic carbon (DOS) Does acidification by inorganic acids influences metal completing with organic carbon significantly? Incorporation of DOS to critical load calculations. 5.4 A c i d i f i c a t i o n and t r a c e m e t a l s and o t h e r e l e m e n t toxicity: The response of trace, potentially toxic metals to acidic deposition and its relationship to general physico - chemical properties of ions (ion ratios, Pauling's negativity). Toxicity of trace metals enhanced by acidification in water and soil with respect to fish, aquatic invertebrates, soil and sediment microorganisms and to sensitive plants. Methylation of mercury and other elements in environment due to changes in pH and DOC 5.5 A c i d i f i c a t i o n and cycling of nitrogen: Acidification effect of increased nitrogen input on N cycling and physiology of soil - plant system before and after the nitrogen saturation of forests. Historical status of nitrogen in acidified and non-acidified surface waters derived from paleoecological observations. 5.6 A c i d i f i c a t i o n and cycling of sulfur: Forms of sulfate which accumulates in soils. Reversibility of fixation of sulfate in connection with release of toxic aluminum. Reduction of the buffering capacity of geological materials in hydrologically important source areas due to deposition of SO2 and products of its oxidation.. 5.7 S o c i o e c o n o m i c factors and e n v i r o n m e n t a l acidification: Do people care about acidification any more? What are the differences in opinions in rich "non-polluting" countries and poor polluting countries? Who is going to finance future research of acidification, critical loads and the state of exceedences of the load? Past Communistic governments were notorious for
351 hiding fats about environmental deteriorations in their countries. However, there are strong indications that even the new democratic governments in central Europe do no support sufficient research of environmental acidification and the exceedences of critical loads in their countries. There is a general agreement among environmental scientists about the need of long-term data on behavior of aquatic and terrestrial ecosystems, especially on changes in biodiversity due to acidification, changes in temperature and land use. How such long term monitoring projects can be financed? Comparative data are needed along climatic and pollution transects. Important issue is the reversibility of ecosystem damages due to acidification under conditions of reducing emissions. Will the present emission control measures be cost effective and ecologically sufficient? 5.8 M i s c e l l a n e o u s topics: Differentiation between biological responses to natural and to anthropogenic acidification. Importance of acidic episodes due to sea salt in combination with acidic deposition. Is it important or unimportant phenomenon? Episodic vs. chronical effects of acidification on ecosystems. Full explanation of why species are lost from acidified environment. Multiple stress impact is not fully understood. We refer to it usually in cases when we do not know the exact cause of an ecological damage. Reduction of emissions is probably the single most important factor to lead to ecosystem recovery. Yet, are existing predictive models reliable? do they reflect real processes of acidification and its recovery? Are not they sets of equations which after proper calibration just mathematically simulate observed trends without any real predictive power?
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BE All rights reserved.
353
Potential ecological risk due to acidification of heavy industrialized areas the Upper Silesia case A. Worsztynowicz, W. Mill Institute for Ecology of Industrial Areas, ul. Kossutha 6, 40-833 Katowice, Poland
Abstract Currently the critical loads concept applies to natural aquatic and terrestrial ecosystems. Critical loads calculation and mapping carried out to support the second Sulfur Protocol negotiations have shown that there are some areas in Poland where the commonly used ecosystems sensitivity criteria for acidification do not reflect the real level of risk. In the Upper Silesia region, the most heavily industrialized in Poland, a number of specific phenomena occur. The energy production is mainly based on hard coal burning what results in intensive emission of acidic gases and solid particles. These particles, alkaline in nature, neutralize substantially the acidification effect of the emitted gases. Excessive reduction of solid particles can lead to excessive acidification of precipitation. On the other hand this area is heavily polluted with heavy metals of natural and anthropogenic origin. Leaving this system out of control can result in spontaneous release of heavy metals and their migration into soil, surface and ground waters. This may lead to unforeseen consequences the more dangerous that addressed to one of the most populated areas in Europe. Preliminary study done in the Institute for Ecology of Industrial Areas in Katowice revealed some evident symptoms of the above mentioned hazards. From this it should be concluded that a comprehensive research program should be undertaken aimed at a thorough identification of the problem and development of an effective environmental policy for heavy industrialized areas.
1. I N T R O D U C T I O N Progressing acidification of atmospheric precipitation belongs to the phenomena more and more often observed in the world. The effects of acidification have impact on a broader and broader territory and their forms are constantly becoming more and more drastic like destruction of forests in Central Europe or biological degradation of Scandinavian lakes. The problem of acid precipitation impact on the areas loaded with heavy metals accumulated as a result of industrial activity is a phenomenon remaining still not fully recognized. Generally, at present heavy metals exist in soil in the form of insoluble compounds, however excessive acidification of natural environment in which they are deposited may lead to their mobility in the forms that may be absorbed by living organisms. Plants turn out to be quite resistant to heavy metals contamination but animals, and humans are prone to the toxic impact of those metals. The situation is particularly serious in the case of industrial agglomeration inhabitants in that for children first of all. Insufficient recognition of the problem on one hand and wide range of possible hazardous effects on the other, impose a prompt necessity to undertake adequate investigations. As study area the Upper Silesian Industrial Region has been selected, in which the problem of impact caused by heavy metals seems to be of particular importance and the risk of quick progress in the acidification of natural environment most real. It should be stressed here that the obtained research results can find application in risk assessment in other industrial
354 agglomerations.
2. G E N E R A L
DESCRIPTION
OF THE STUDY AREA.
The Katowice province is situated in south-western Poland. The province is a typical industrial area, with high level of urbanization and serious environmental problems, which place it on the list of most polluted regions in Europe. The province covers the area of 6650 km 2, which makes 2.1% of the total area of Poland, and it is inhabited by about 4 million people (10.5 % of total population in the country). The average population density amounts to 602 people per one km2. About 87 % of the population lives in towns where the average density amounts to 2051 people per one km 2. High level of urbanization is the result of intensive industrialization based on local mineral resources. The main branches of industry developed in the Katowice District include: coal mining, coal-fired electricity and heat generation, non-ferrous metal processing and mining (including zinc and lead ores), chemical industry, production of building materials. The extraction and raw materials processing, machinery and chemical industries developed in an uncontrolled manner. They have not been modernized, causing many ecological problems, proper living and working conditions cannot be secured for Silesia inhabitants. According to the reports of the Regional Sanitary-Epidemiological Station in Katowice the air pollution in the Katowice province belongs to the highest in Poland. In 1991 24% of gaseous pollutants and 20.2% dust pollutants in Poland came from the emission sources located within the Katowice province. The annual permissible concentrations of most of the monitored pollutants are exceeded many times, particularly within industrial-urban areas (Chorz6w, Swi~toch/owice, Bytom, Ruda Sl~ska, Zabrze), and also in peripheral areas of the province. Particularly high concentrations are found for such pollutants as carbon monoxide, nitrogen dioxide and of suspended dust - within which high concentrations of lead and benzo-a-pyrene occur (Table 1). Atmospheric precipitation causes contamination of surface waters and soil. In the vicinity of plants processing non-ferrous metals (zinc smelters), the lead fallout in dust has resulted in contamination of the topsoil, reaching drastically high levels of several grams of lead per kilogram of soil (Miasteczko Sl~skie, Bytom, Bukowno). The Upper Silesia region suffers a severe water resources shortage. The deficit of water is estimated at about 250000 m3/day for municipal use and about 300000 m 3 for industrial use. In 1991 the total use of water was of about 1081.4 hm3. Of this, about 753.5 hm 3 was discharged as effluent, and 517.6 hm 3 was treated, but only 26-27 % of total effluent were treated properly. The remaining 235.9 hm 3 was dumped into the surface water system without any treatment. Industrial effluent discharged into rivers contain excessive amounts of heavy metals, phenols, cyanide, ammonia nitrogen and salt. Discharge of municipal waste water and saline mine waters has resulted in catastrophic degree of river pollution. Water quality monitoring indicates that 62.8% of the overall length of the rivers of Katowice province carry excessively polluted water, unsuitable for any use (in Poland - 35 %).
355 Table 1 Pollution levels in Katowice province in 1991. Substance
Annual concentrations Units
(observed) from to
Permissible mean annual conc. Da
ammonia
#g/m 3
18
79
51
phenol
#g/m 3
5.2
27
2.5
fluorine
/~g/m3
0.8
1.9
1.6
formaldehyde
/zg/m3
2.6
44
3.8
carbon dioxide
g/m 3
0.76
0.94
-
carbon monoxide
mg/m 3
4.4
7
0.12
sulphur dioxide
/~g/m3
21
87
32
nitrogen dioxide
#g/m 3
25
134
50
volatile hydrocarbons
mg/m 3
3.4
4.8
-
suspended dust
/zg/m3
66
253
50
tar substances
#g/m 3
7
32
-
benzo-a-pyrene
ng/m 3
15
125
1
perylene
ng/m 3
3
37
-
lead
#g/m 3
0.11
1.51
0.2
zinc
#g/m 3
0.19
7.6
-
manganese
/zg/m3
0.03
0.21
1
iron(Fe202)
/.t/m3
2
24
-
copper
/z/m 3
0.3
4.4
0.6
cadmium
ng/m 3
2.1
72
10
chromium
ng/m 3
2.1
44
400
nickel
ng/m 3
2.3
59
25
dust deposition
g/m2y
35
466
200
lead deposition
mg/m2y
2
2822
100
cadmium deposition
mg/m2y
0
87
10
Reprinted from: Cimander, B. and J.Szeliga (1992) Air Pollution in Katowice Province. "AURA" No 11/92. Data collected by Regional Sanitary-Epidemiological Station in Katowice
356 Annually, the Katowice province industry generates 60-70 mln tons of wastes (1991), from which: coal mining 53.0 mln tons - power generation industry 5.0 iron and steel metallurgy 4.5 - non-ferrous metallurgy 3.5 chemical industry 1.0 It makes about 50% of wastes generated in Poland (128 mln tons). 42.0 mln tons of wastes generated annually are utilized (for production purposes, filling excavation, reclamation, land levelling, soil fertilizing) - 0.005 mln tons are neutralized - 23.0 mln tons are dumped Annually 30 mln tons of wastes (mainly from coal mining) are used for non-industry purposes, such as: filling excavations, levelling and building roads. Only 10 mln tons of wastes are used as row material for building materials production. The degradation of the natural environment is closely related to lower values of many health indices for children and young people, and also with the all-revealing infant mortality rate. With the average value for the province at 16.2 in 1991, it is much higher in most of the cites within Upper Silesia ( in 1990 in Katowice province - 17.8, in Zabrze and Sosnowiec - 22.7, in Piekary Sl~skie - 24.8, in Tychy - 26.7, in Czeladz - 27.7 and in Swi~tochtowice- 31.3). In Poland: 1990 year- 15.9, 1991 year- 15.0. Among the diseases resulting from environmental pollution the predominant ones are pulmonary diseases (around the coke plant at D~tbrowa G6rnicza), heart conditions, allergies, deficient cellular immunity and chromosomal disorders (in areas affected by lead, cadmium and zinc pollution). -
-
-
-
3. E M I S S I O N
SOURCES
OF HEAVY
METALS
It has been assessed that since the beginning of industrial activities in the Upper Silesian Region up to present times about 2 mln mln tons of industrial wastes have been generated. It makes approx. 300 thousand ton of wastes per km a i.e. the concentration highest in the world in such a densely populated areas as the Upper Silesian Region. About 78 % of wastes is accumulated on the earth surface which in a significant way enhance their impact on natural environment causing harm to soil, water and air quality in the area of the Katowice Province. The dispersion of landfill sites makes the situation even more difficult; only about 680 landfill sites are inventoried and the number of 'wild' landfill sites is assessed to be 280 300. Those landfill sites are not adopted to hazardous wastes disposal, neither are there suitable technologies, equipment and selected disposal areas [1]. From the point of view of heavy metals impact on environment the wastes generated in energy production as well as those from non-ferrous metals industrial plants, especially in the processes of cadmium, lead and zinc production create special risk to environment. These metals along with Hg, Cu, T1, Sn, Cr, Sb, Ag and Au are classified into the group of elements of especially high probability of potential risk for living organisms [2]. Annual increase of solid wastes mass connected with energy production (i.e. slag and ashes) is assessed as approx. 6 mln tons whereas the total mass of the already disposed wastes, according to the same assessments, is assumed as 68 mln tons. Emission of volatile ashes
357 into the atmosphere is currently on the level of 200 thousand tons. It includes about several tens of tons of zinc in the form of oxygen compounds, several tens of tons of lead and 1-2 tons of cadmium. The impact of non-ferrous metals industry (Zn, Cu, Cd) located in the Upper Silesia Region is observable already in the phase of ores exploitation, their enrichment and during processing. The solid wastes disposed on landfill sites are assessed as approx. 80 mln tons and the annual increase is assessed on the level of 4 mln tons [3,4]. In 1992 the discharges into the atmosphere in the form of dusts included 90 tons of ZN and 20 tons of Pb [5]. One should also not forget about the role that the 800 year history of zinc and lead ore mining not fully registered in documents played in the development of the present state of natural environment in the Upper Silesian Region. Transport is another source of lead in the discussed region. Annual emission from that source is assessed on the level of 150 tons. Summarizing it should be said that the major sources of heavy metals pollution in the Katowice Province are: energy production industry and non-ferrous metals like zinc, lead and cadmium industry and in the case of lead also transport. The disposal sites of slag and ashes as well as flotation wastes store at present approx. 150 mln tons of wastes. Their impact on environment is of regional character. Emission of dusts into the atmosphere has impact on a much more extensive area. Annually, about 100 tons of Zn, 160 tons of Pb and 1-2 tons of Cd are transported mainly in the form of oxygen compounds. Power plants and their disposal sites are situated uniformly in the central and at the same time most densely populated part of the province, whereas the non-ferrous metals industry is located in the northern and eastern parts of the province i.e. the area of Bytom, Tarnowskie G6ry, Chrzan6w and Bukowno.
4. T R E N D S
IN ACIDIFICATION
OF ATMOSPHERIC
DEPOSITION
Generally hazardous situation concerning acid deposition in the Katowice Province is well known. It can be illustrated by the results of the investigations carried out by IEIA within the international research program for Convention on Long-Range Transboundary Air Pollution Transport [6]. That research proved considerable exceedances of the critical loads of acidity, especially for sulphur compounds in the South-Western part of Poland, in that in the Katowice Province (Fig. 1 and 2) [7]. The research on atmospheric pollution is carried out among others in the Institute for Ecology of Industrial Areas in Katowice - Zat~e. In order to obtain information on the time variability of rains pH the data collected from monitoring were processed. The results of the calculations are presented in Fig. 3 and 4. On their basis it is possible to state that the pH of rains measured at IEIA station has been showing a significant tendency to decrease in the recent four years (Fig.3). Also the presented in Fig.4 correlation between the concentration of particulates in the air and precipitation pH is statistically significant. The decrease of precipitation acidity together with the increase of dusts concentration proves its basic character. An independent confirmation of the observed in IEIA tendency of precipitation acidity increase is the research performed in recent years at the Silesian University in Katowice.
358
~-~
Z~
~0
o
z
,,, ,~
2o"
F~
F~ ~,m~
F~
"T,L_
~'(II (.-
.~.
o _o
o o o
o o o o4
r-
oo°
o~~ O~~ ~o
mNmmmm
o
o
A
~
I-w
&
o,-~
¢0
(If
G)
£ (-
"o .i L_
c}. G}
fY
z~~ ~0~ ~0
"~
0
z
~~ °
~.~
o-
o~ v o~
o
o
o o
~ ~ ~_
~))mmm
° o ^
&-i
& o~
359
Or) (17)
°~
0 $...
e"
360
Y = 6.1386
correlation
-
8
-
-
~
.
.
.
.
"
.................................................................... o
.
.
.
.
.
- .0350
* X
coefficient: r = -.6866
.
.
" ...............................
.
~
: ................................
-
-
-
i
i ...............................
7 o
0
o
~
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
o
o
.
.
o-
o
! °o
o
i
.
.
.
.....................................................
i
. . . . . . . . . . .
I.... ° ...................... ~...,................................ .~,.....
0
15
5
.
:::::::::---S~II ..... i ~ .............~ ...........~................................
...................................................................................................................................
o~
.
ii
. . . . . . . . . . . . . . . . . . . . . . . . . . .
................................ i"................................ i ................................ i.................
03 O)
.
ii
...................
6
.
-! .................................................................
:.: o
"IE:L
.
25
o
35
!;
]
45
55
..... .....
......
65
Time [months]
C)
Fig.3. Changes
of precipitation
p H in t i m e
Y = 3.9487 + .01106 * X Correlation
coefficient: r = .53495
7
i
i ....................................................................................................................
6.2
:............................................i.............................................i ............................................i............-::~ .....................i .......................................
5.8 5.4 7-
......i..............
6.6
5
o
:
. 0 ............................................................ ::~,;~.......................................
i
i
i
i
o
!
............~
io
o
o
i
................. !
....... i ....
~ . i ~
~
o
...........
! ............................................ i ...... 5 .......... ..~:.:--::..:.c, ..... i ................ i~;i:::::::;:::"-:~: .......................................... i ............................................ i i : 6 ................... ':.................... ...................... -::::::---~ ......... ~: ................ ---::: .............::f ............................................. i..................................o .......i............................................ .................
...0 ..... 0
!
i
4.6 4.2
,..:-:-::-::..........::.:...................... o .................. o..0--.-.-,b-............................ .o.........................................................................................................................
3.8
................................................................... .5o.i6. ...........................................................................................................................................................................................................................
3.4 ,0
80
120 Particulates
Fig 4. C o r r e l a t i o n
between
parcitulates
160 concetration
concentration
200
240
[ p g / m 3]
in a i r a n d p r e c i p i t a t i o n
pH
361
5. POTENTIAL RISK TO ECOSYSTEMS DUE TO INCREASING ACIDIFICATION Practically speaking, all components of environment in the Upper Silesian Region are exposed to the effects of the increasing acidity of precipitation and in consequence to the enhanced migration of heavy metals. The presence of heavy metals especially of Pb, Zn, and Cd in soil water and air creates a real risk of penetration of those toxic elements into human organism. Acid rains even rise that phenomenon in power. 5.1. Soils A considerable amount of heavy metals migrating into soil is immobilized as a result of complexing process taking place in humus or unmeasurable sorption by clay materials or finally in sparingly soluble compounds etc. The main factor influencing the concentration of heavy metals in soil solution (and at the same time the possibility of toxic impact on plants) is soil pH. As a rule the concentration of heavy metals in soil solution increases parallel with the decrease of soil pH. When the area under the emission of heavy metals covers with the area of acid rains appearance the situation turns out to be especially dangerous [8]. R.Kucharski, M.Marchiwfiska et al [9,10] carried out research on the concentration of heavy metals in soils used for agricultural purposes and plants cultivated on them. Their developed an agricultural land assessment method, which allows to classify the area, considering environmental pollution [11]. The classification is performed using the method of simple indices, easy to operate, and showing which species of plants could be safely grown regarding consumer's exposure. With the application of this method and on the basis of environmental quality data, more than half of total arable land area has been evaluated in the district. Up till now, about 140000 hectares of arable soils have been assessed, out of which 55% were classified as A category (safe zone), 38% - as B category (limited agricultural activity) and 7 % - as C category (unsafe zone). It seems also important to pay attention to the fact that the majority of research was carried out in the time when the problem of acid rains was not so glaring as today. The fate of heavy metals accumulation in arable soils is different than in forest areas. Due to ploughing a process of "dilution" of the immited elements in big mass of soil is observable. In the case of forest soils very high concentrations appear in the surface layers, mainly in the humus, whereas in the deeper levels they are usually low with the values often comparable to those observable in unpolluted soils [8][12]. In the case of forest a more serious problem is the mobility of the active A1 ÷3 ions resulting from the excessive acidification. Those ions have toxic impact on the root systems of trees. It is assumed that the presence of heavy metals in soil solution strengthens that effect. Research on the identification and inventory of forest damages in the area of the Upper Silesia is carried out by the Research Institute of Forestry. The documentation of those damages is published in annual statistic reports and for the year 1990 forest area endangered by air pollution in Katowice district amounts to 99.2 %, within the third class of trees injury relates to nearly 15 % of the total forest area in district [13].
362 5.2 Surface waters Heavy metals exist in surface waters in the form of soluble compounds (ionic and complex), colloids and suspended matters. Soluble forms easily accessible to organisms are only a negligible percentage of the metals total mass. The predominant amounts of metals are deposited in the form of sediments. Sediments are that part of water ecosystem in which the accumulation of metals takes place and they may at the same time be the terminal stage of metals migration in environment. However, in some conditions, heavy metals accumulated in sediments can be released and migrate into water causing its degradation. That is the reason why the heavy metals accumulated in sediments are often called "chemical time bomb". The processes of metals release from sediments are the effect of the activity of various mechanisms conditioned both by interior and outerior factors of a given ecosystem. The mechanisms of those processes are still not well recognized, and it especially concerns water ecosystems loaded with heavy metals. Preliminary research consisted in the determination of contamination with heavy metals and their distribution in river between the water phase, suspended matter and sediments as well as the determination of the form in which those metals are bounded in sediments on the example of the Biata Przemsza River [14]. The Biala Przemsza River and its tributaries remaining under strong impact of zinc-lead ore exploitation and processing industry, showed very high level of contamination with zinc, lead and cadmium (Table 2). Table 2 Concentrations of heavy metals in the Biata Przemsza River Basin Range of concentration in the cross sections: Below the point sources
Above the point sources
1. Water phase Zn, mg/1
0 . 0 3 - 3.7
u.d.* - 1.2
Pb, mg/1
0.02 - 1.3
u.d.* - 0.2
Cd, mg/1
u.d.* - 0.01
u.d.* - 0.002
4,513- 68,000
8 4 0 - 9,900
Pb, mg/kg
588 - 42,700
866 - 12,000
Cd, mg/kg
28 - 1,200
3 9 - 840
Zn, mg/kg d.w.**
166 - 16,545
31 - 164
Pb, mg/kg
167 - 4,100
13 - 183
Cd, mg/kg
0 . 9 - 73
0.1 - 2.1
2. Suspended matter Zn, mg/kg .d.w.**
3. Sediment
*u.d.- under determination level
**d.w.- dry weight
363 Metals in soluble form were only an insignificant quota in comparison to their contents in suspended matter and sediments. Maximal concentrations of metals in soluble form reached in the cross sections loaded by emission from point sources the following values (mg/dm3): Zn - 3.7; Pb - 1.3; Cd - 0.01; whereas in suspended matter they reached the values respectively (mg/kg of dry weight): Zn - 68,000; Pb - 42,700 and Cd - 1,200. Sediments showed also high level of contamination with heavy metals yet still lower than in suspended matter. The highest measured values were (mg/kg of dry weight): Zn - 16,500; Pb - 4,100 and Cd - 73. The natural contents of those metals in sandstones padding the Bia~a Przemsza Valley is rather low and amounts (mg/kg): Z n - 16; Pb - 7 and Cd - 0.02. In that context the contents of metals in sediments in the control cross section - above the point sources of emission was also many times lower and proves the impact of non-point emission sources. The presented results of sediment contamination with heavy metals, limited due to insufficient measurements number only to the Biata Przemsza River Basin, show that the risk that metals accumulated in sediments has already been real.
5.3. Groundwater The groundwater resources of the Katowice Province are 5 main underground water reservoirs of the total surface 3350 krn~. Those reservoirs create a slit - karst water bearing supply complex of the carbonate series of Trias. They are the main source of groundwater supply in the urbanized and industrialized agglomeration of Upper Silesia. Observations show that groundwater is especially exposed to pollution from earth surface in the catchment areas [15]. In water in the vicinity of zinc and lead ore mines decreased pH as well as considerable concentration of Zn up to 21 mg/dm 3 and Pb up to 7 mg/dm 3, exceeding the permissible values (respectively 5 and 0.5 mg/dm 3) were stated. Also exceedance of Cd concentration permissible value is noted.
6. S U M M A R Y
AND CONCLUSIONS
The above presented overview enables to draw the following conclusions: 1. The results of the precipitation pH measurements in the area of the Katowice Province prove the progressing process of its acidification. 2. In the territory of the Katowice Province there are areas considerably loaded with heavy metals as a result of industrial activity connected with exploitation and enrichment of nonferrous metals ores and production of those metals, mainly cadmium, zinc and lead. That situation concerns first of all such cities and communes like Bukowno, Bytom, Chorz6w, Chrzan6w, Katowice, Swi~tochtowice and Tarnowskie G6ry. They cover approx. 10% of the province area and are inhabited by about 20% of the province population. 3. In the Katowice Province the effects of heavy metals mobilization have already been tangible in the form of excessive contamination of agricultural products as well as surface and underground water. Progressing acidification may only enhance the processes of heavy metals mobilization.
364 4. The assessment of heavy metals mobilization processes in time is rather difficult due to nearly complete lack of data on the macrokinetics of those processes. Generally, it can be stated that the negative effects of the excessive pollution with heavy metals in the area of the Katowice Province are tangible. The impact of the increased precipitation acidity may progress only in one direction i.e. towards the increase of heavy metals mobility making at the same time the situation very dangerous. The time scale of the discussed phenomena is difficult to determine. While the data on the steady-state thermodynamics in soils are relatively rich, the questions addressing the rate of soil processes connected with heavy metals mobilization have not found satisfactory answers so far. The reason for this are considerable experimental difficulties connected with the collection of kinetic data for soil systems. Currently the critical loads concept applies to natural aquatic and terrestrial ecosystems. Critical loads calculation and mapping have shown that there are some areas in Poland where the commonly used ecosystems sensitivity criteria for acidification do not reflect the real level of risk. Especially in the Upper Silesia region a number of specific phenomena occur. In this connection a comprehensive research program should be undertaken aimed at a thorough identification of the problem and development of an effective environmental policy for Upper Silesia region.
7. R E F E R E N C E S 1 L.Sieja. J.Borkiewicz, A.Goszcz: Minimalization of industrial wastes and improvements in handling of communal wastes in Upper Silesia as a result of regional activities. In: Proecological activities aimed at upgrading of living standard in Upper Silesia Regionmaterial of plenary session of Ecological Council at the President of Poland. Katowice, 22-23 June 1993. (in Polish) 2 A.Kabata-Pendias, H.Pendias, Trace elements in biological systems, Wydawnictwo Geologiczne, p. 62, 1974. (in Polish) 3 M.Szuwarzyfiski, A.Kryza, Zinc and lead ores flotation wastes in Silesian-Cracow excavation region, Przegl~d Geologiczny (Geological Survey), 841, 9, pp. 629-633, 1993. (in Polish) 4 W.Janusz, E.Popiotek, Environment in Olkusz region of zinc and lead ores excavation, Rudy Metali (Metal ores) 38, 1, pp. 6-9, 1993. (in Polish) 5 S.Wolff: Politechnika Slaska (Silesian University of Technology), Katowice - personal communication, 1993. 6 W.Mill et al., Critical loads mapping for Poland, In: Calculation and mapping of critical loads in Europe - Status report 1993, UN/ECE Coordination Center for Effects, RIVM, Bilthoven, pp. 97-101, 1994. 7 K.Abert, K.Budzifiski, K.Juda-Rezler, Regional scale air pollution models for Poland, Paper accepted for publication in Ecological Engineering, 1993. 8 Z.Prusinkiewicz, W.Pokojska, Influence of industrial imission on soils, In: Trees in polluted environment, PWN Warszawa, pp. 239, 1986. (in Polish) 9 R.Kucharski, E.Marchwifiska et al., Agriculture in Katowice Province and industrially polluted environment, Aura 11, pp. 22-23, 1992. (in Polish) 10 R.Kucharski, E.Marchwifiska et al., Heavy metals contamination of cultivated soils and
365 plants and means for diminishing consumers risk, Instytut Ochrony Srodowiska Oddziat (Institute of Environmental Protection), Katowice, 1990. ( manuscript in Polish) 11 R. Kucharski, E. Marchwifiska, J. Gzyl, Agricultural policy in polluted areas~ Ecological Engineering, 3, 1994. 12 J.Greszta, E.Panek, Influence of heavy metals on trees, In: Trees in polluted environment, PWN Warszawa, p. 202, 1986. (in Polish) 13 Environment Protection 1991, GUS (Main Statistics Office), Warszawa, 1991. 14 S.Ryborz et al., Mechanisms of heavy metals accumulation in sediments and living organisms in surface waters, Instytut Ochrony Srodowiska (Institute of Environmental Protection), Katowice, 1991. (manuscript in Polish) 15 R6zkowski A., R6zkowski J., Pacholewski A, Water quality from the slit-karst formations in Jura Krakowsko-Cz~stochowska region and sources of its deterioration, Kras i Speleologia (Karst and Speleology), Katowice, 1993. (in Polish)
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
Acidification and metal changes on Cd mobility
mobilization:
effects
of land
367
use
P.F. RSmkens a and W. de Vries b aAB-DLO Institute for Agrobiology and Soil Fertility Research, PO Box 129, 9750 AC Haren, the Netherlands bSC-DLO W i n a n d Staring Center for Integrated Land, Soil and Research, PO Box 125, 6700 AC Wageningen, the Netherlands
Water
Abstract Due to the actual EC agricultural policy, arable land is transformed to forest. Termination of liming will lead to a decrease in soil pH and an increase in soil organic m a t t e r and Dissolved Organic Carbon concentration. Both field m e a s u r e m e n t s and model predictions indicate t h a t the decrease in soil pH from near n e u t r a l values to ca. 3 to 4 takes place within 40 to 60 years after forestation. As a result of this, Cd solution concentrations increase with a factor 2 to 6 and leaching losses and/or plant uptake are expected to increase accordingly.
1. I N T R O D U C T I O N In several countries of the European Community, arable land is abandoned in an a t t e m p t to reduce arable crop production. Up to 30% of the actual arable land (approx. 8.5 million hectares) is estimated to be converted to nonagricultural purposes due to technical improvements and large scale agriculture. One of the options available is the creation of forests. These forests could both act as a sink for atmospheric CO 2 and be used for recreational purposes in densely populated areas like the Netherlands. However, the creation of forests on a large scale will also affect the chemical conditions in the soil. Since the majority of the areas t h a t are to be converted to forest have been used for agricultural purposes for almost a century, changes in the trace metal status of these soils have occurred. As a result of m a n u r e application and fertilizer t r e a t m e n t , the total Cd and Cu content has increased compared to forest soils (van Driel and Smilde, 1990). The mobility of these (potentially) harmful metals was limited, however, due to the regular application of lime. In general the soil pH has been maintained at near neutral or slightly acid values
368 (5 to 6). In this pH range the solubility of most metals is limited due to strong sorption processes onto several solid soil components like clay particles, Fe- and Al-(hydr)oxides and organic m a t t e r (McBride, 1989). In case of t e r m i n a t i o n of liming, however, the soil pH will decrease with time. Apart from n a t u r a l acidification, the elevated input of acidifying components from the atmosphere accelerates the acidification, especially in soils with a low buffering capacity. In general the decrease of soil pH will cause an increase in the mobility of most metals due to desorption from the soil surface (Harter, 1982). It would therefore be preferable to convert only those areas which contain soils with a low degree of contamination and a high buffering capacity. Unfortunately, these well buffered, non-contaminated soils (e.g. the calcareous clay soils in the n o r t h e r n parts of the Netherlands) are highly productive and in general only the marginally productive soils will be transformed to forest. In the N e t h e r l a n d s these areas consist of poorly buffered sandy soils with a low clay content and sometimes elevated concentrations of metals like Cu and Cd due to the application of pig m a n u r e and/or phosphate fertilizers. Apart from soil pH also the organic m a t t e r chemistry will change. Due to the increased input of organic m a t t e r (e.g. leaf litter, tree needles, branches, roots etc.) an organic horizon will develop. This will affect metal mobility in two ways. First, the increase in soil organic m a t t e r could lead to an increase in the CEC of the soil and metals can be bound to the organic matter. Secondly, the increase in soil organic m a t t e r also leads to an increase in the production of soluble organic colloids (Dissolved Organic C a r b o n - DOC). Especially the mobility of metals like Cu and Pb, which form stable aqueous complexes with organic ligands, could increase as a result of the elevated DOC levels. In column studies the mobility of Cu was enhanced in the presence of DOC, compared to a system without DOC although not all types of DOC studied showed similar results. (Oden et al., 1993). In order to estimate the mobility of trace metals in case of a conversion of arable land to forest it is therefore important to know the rate of soil acidification and the effect of an increase in the soil organic m a t t e r content. It is the purpose of this paper to show how soil acidification combined with a change in the organic matter chemistry affects the mobility of Cd after conversion of arable land to forest. Field data combined with model calculations are presented which show the consequences of land use changes for trace metal mobility.
2. M A T E R I A L S AND M E T H O D S The field study and model calculations presented here, have been carried out independently. The field study is an integral part of a national study concerning the mobility of trace metals in both arable and forest soils and the effect of soil parameters like pH and DOC on metal solution concentrations. The model calculations form part of a study referring to the response of soils to acidification.
369 Field data - trace metals To study the role of total metal content, soil pH and DOC on Cd, Zn and Cu solubility in Dutch soils, 30 sites have been investigated which r e p r e s e n t the majority of Dutch soils and land use types. Both arable soils (pasture, crop production and fruit tree stands) and forested sites were included in the study. Soil samples have been collected at 10-cm depth intervals from 0 tot 80 cm. Soil solution samples were obtained by centrifugation of 100 cm 3 field-moist samples in a centrifuge applying a rotational force of 20.103 N.kg ~1. In the centrifuge tube the solution samples were expelled through a non-acid MN680M filter with a nominal pore diameter of 6~m. In this way more 'natural' soil solution samples were obtained compared to extraction with specific extractants (e.g. 0.01M CaC12). Soil solution analyses included pH (ORION combination electrode model 91-02), total dissolved carbon and inorganic carbon (Shimadzu TOC-500 Total Organic Carbon Analyzer) and trace metal analysis. Cd was m e a s u r e d on a Graphite Furnace Atomic Absorption Spectrophotometer (Perkin E l m e r model 5100) using s t a n d a r d addition methods and Z e e m a n background correction. Total metal content was determined using a hot-acid extraction as described by Del Castilho et al. (1993). Total organic m a t t e r content, CEC and soil pH were determined according to s t a n d a r d methods (Page et al., 1987). Field data - p H / D O C change with time To study the effect of forestation of arable land on soil pH and Dissolved Organic Carbon dynamics with time, 8 forests have been investigated. All forests have been planted on limed arable soils and vary in age from 1 to 70 years. On the terrain managed by the Dutch Forest Service (Staatsbosbeheer) in Jipsing Boertange (province of Groningen, the Netherlands) soil solution samples have been collected from the 0 tot 30 cm soil layer. Extraction of the solution was obtained by centrifugation as described previously. Modelling approach In order to describe the effect of acidification, as a result of land use changes, on the Cd solubility a Cd sorption model (Bril, 1993) was linked to a soil acidification model (de Vries, 1994). The calculation of the Cd solution concentration is based on the assumption that, with a decrease in soil pH the adsorption of Cd decreases. Therefore, we used the calculated change in soil pH with time to calculate the Cd adsorption coefficient (described below). With this coefficient the solution concentration of Cd as a function of time was calculated and compared with the measured Cd solution concentration in existing field sites throughout the Netherlands. Cd adsorption model In the model applied, the change in the dissolved Cd concentration was described based on the Freundlich adsorption isotherm:
370
Cdac = KCdad. [Cd]n
(1)
Cdac : sorbed Cd content (mg.kg-1) KCdad : adsorption constant for Cd (m3-kg-1) [Cd] n
: dissolved Cd concentration : empirical constant, depending on soil type
The Cd adsorption constant was related to several soil p a r a m e t e r s according to: log (KCdad) = -6.15 + 1.00. Iog(CEC) - 0.24. log(lut) + 0.50. pH- 0.41 9log[Ca]
lut: [Ca]:
(2)
clay content (%) dissolved Ca concentration (mol.L -1)
The values in equation 2 are based on data from Chardon (1984) and Christensen (1989); from these data also the value for n was estimated at 0.82. Based on the modelled change in soil pH and Ca solution concentration, the change in the adsorption constant for Cd can be calculated. Assuming t h a t the sorbed Cd content does not change with time:
Cdac(t p) = Cdac(t O) aD:
to:
(3)
time period p after start of acidification time zero, s t a r t of acidification
Equation I can be rewritten as:
KCdad(tp). [Cd]n(tp) = KCdad(to). [Cdln(to )
(4)
From this the change in the dissolved Cd concentration at a given time p after forestation, [Cd](tp), was calculated according to: [Cd](tp) _ [Cd](t o )
KCdad(to ) KCd ad (t p )
)l/n
(5)
Soil acidification model The two m a i n p a r a m e t e r s needed to calculate the dissolved Cd concentration are soil pH and dissolved Ca concentration. The soil acidification model applied in the pH calculation is based on the "critical load" concept (see de Vries, 1994 for more detail). In those cases where the critical load is exceeded (in this approach the m a x i m u m input of acidifying components t h a t can be buffered by a specified soil by weathering and N retention), the base s a t u r a t i o n of the soil decreases which is related to a decrease in soil pH.
371 The critical deposition level (critical load) has been defined as:
Std+ Ntd(crit) = BCtd+ BCwe- BCgu + Ngu + Nim + Nde(crit) + Ac/e(crit)
S, N: BC: Ac: gu: we:
Sulfur and Nitrogen Base Cation Acidity growth uptake weathering
le: td: ira: crit:
(6)
leached total deposition net immobilization critical level
All pools are expressed in molc.ha-l.yr -1 The assumptions in this approach and their justifications (e.g. total nitrification, no nitrogen- and sulfur fixation) are given by de Vries (1994). The critical a m o u n t of acidity leached from the soil (Ac/e(crit)) is based on the critical concentration of acid components in the soil solution multiplied by the net precipitation surplus. Assuming that the acidity has to stay constant at values encountered at near neutral pH, a pH of 5 was used to calculate the critical leaching of acidity. This has been defined as: (7)
[Ac] = [H] + [Ali] - [HCO 3] where [Ali] is dissolved inorganic Aluminum (mol c. m -3)
The change in the soil base saturation has been calculated as the difference between the actual deposition and the critical deposition according to:
t=tp frBac(t p )= frBac(tO )- Z[(Std § Ntd)(act)-(Std § Ntd)(crit)]/ (p.D.CEC-10) t=t 0 frBac: base saturation p: soil bulk density
(8)
D: thickness of soil layer CEC: Cation exchange capacity
In the calculations the initial values are based on 1990 (t o) and calculations are made for 2000, 2010 and 2050. If the critical load is not exceeded, the base saturation will remain constant and the acidic input will be compensated completely by net cation release and net N retention (cf Eq. 6) resulting from weathering and denitrification processes. If the acid input exceeds the critical load, the base saturation and pH will drop according to an empirical relationship determined by Bloom and Grigal (1985): pH = 4.96 + 0.80 9log ( frBac / (1 - frBac))
(9)
372 The change in the dissolved Ca concentration was calculated from the the charge balance equation: [Ca] = [SO 4] + [NO 3] + [HCO 3] + [C1]- [ H ] - [A1]- Z{Mg, K, Na}
(10)
Concentrations of SO4, NO3, C1 and base cations were based on their deposition, u p t a k e (for N and base cations) and w e a t h e r i n g (base cations). Dissolved concentrations of both A1 and HCO 3 were derived from equilibrium calculations (cf Eq. 11 and 12) [HCO 3] = g c o 2 9pCO 2 ] [HI
(11)
The dissolved A1 solution concentration was calculated from the equilibrium with amorphous Al-hydroxide according to: [Ali] = KAl_ox. [H] 3
(12)
Data requirements for the model Based on the actual deposition levels for S and N, estimates are m a d e for the total deposition until 2050. In the scenario applied here, a constant i n p u t of acidifying components has been assumed using the 1990 input levels. For the Netherlands, data for the actual deposition were available for 1985 and 1994 on a 10 x 10 km grid basis. The values for 1990 have been calculated using a linear interpolation program. The actual base saturation has been calculated for the upper 40 cm of the soil. Both CEC and bulk density have been e s t i m a t e d using transfer functions based on the organic m a t t e r and clay content ( B r e e u w s m a et al., 1986; Hoekstra and Poelman, 1982): CEC= 15.OM +5.1ut OM: lut:
(13)
Organic M a t t e r content (%) clay content (%) 1000
P = b0 +b 1 . 0 M + b 2.1ut bi:
(14)
empirical constants
Organic m a t t e r and clay content were also used to calculate the adsorption constant for Cd (cf Eq. 2).
373 3. R E S U L T S 3.1 Field
data
In figure 1, the changes in soil pH and DOC concentration with time are shown. After termination of liming, the soil pH decreased to 3.5 - 4 within 3 to 4 decades, indicating a rapid acidification of the topsoil due to a lack of buffering capacity in these sandy soils. Despite the lower soil pH, the DOC concentration increased with time. It has been shown by Bergkvist, 1987) that under field conditions, metals like Cu and Pb are mobilized in the presence of DOC. The increase from an initial DOC concentration of 5 - 10 mg.L -1 to 50 mg-L -1 within 40 years could therefore be of significance for the mobility of those metals which form stable aqueous complexes with DOC. Results for Cu (not presented here) indeed showed t h a t solution concentrations increase with increasing DOC concentrations (RSmkens and Salomons, 1993). However, the increase of the total organic m a t t e r content also increases the capacity of the soil to retain metals from solution and the net effect of the change in the organic m a t t e r chemistry (net release or retention of metals) t h u s remains unclear until now. However, the solubility of metals like Cd and Zn is mainly controlled by pH and the acidification due to the land use changes will lead to a significant increase in the dissolved Cd concentrations. In figure 2 the results from the screening of 30 sites in the Netherlands are shown. It appeared t h a t the solubility of Cd was controlled mostly by soil pH and, to a much lesser extent, by DOC. Despite the large range in the dissolved Cd concentration at a given pH, there is a significant increase with a decrease in soil pH. Especially below pH 5, dissolved Cd concentrations increase rapidly due to pH dependent desorption from the soil surface. Similar results were obtained or Zn (not shown here). 6
soil pH
DOC (mg/L)
. . . . . . . . . .
- 50
,r ', :
-...........................
,
-
1992
'~'
I
.
40
.(, ': ....
....................
.
I
: .
. -
.
:: - - i . j .
!:7,:.
~':
~,:,
.
-
60
30 ~:
~!~
1985 1977 1965 1960 1934 year of forestation
20
~ 1920
. I0 0
Figure 1. Change in soil pH and DOC concentration due to forestation of arable land (0 - 30 cm soil layer).
374
100
Cd (ug/L)
30-
3
J! I
--
. 9
== v.._
~----_.~i.
--
C
.t
9
9
9
.
-
B A
0.3
0.1 0.03 0.01 3
Soil type: sand silt cl~y peat I
4
I
5
I
6 soil pH
I
7
I
8
9
Figure 2. Relation between pH and dissolved Cd concentrations in arable soils and forest soils (0 - 80 cm; A, B, and C refer to Dutch Reference Values, see fig. 3 and text) The high dissolved Cd concentrations at pH 8 were found in a tree stand treated repeatedly with a pesticide containing Cd. The range of dissolved Cd concentrations at a given pH is due to the fact t h a t in this plot all samples are shown; both forest soils with low Cd content (Cdto t < 0.1 mg.kg-1) and arable soils containing elevated amounts of Cd (Cdto t _> 0.5 mg.kg-1). However, the highest dissolved Cd concentrations were found in forest soils (Cdaqranged between 6 and 24 ~tg.L-1) despite their very low Cd content (sometimes less t h a n 0.05 mg.kg-1). The total metal content in arable soils was higher compared to forest soils due to application of fertilizers, m a n u r e and lime. Dissolved Cd concentrations however, were usually very low due to the near neutral soil pH (5.5 - 7.5). In figure 3, the dissolved Cd concentrations are compared to the Dutch soil quality standards for the soil solution and the upper ground water (Stoop and Rennen, 1990). These standards are divided in three classes: A-, B-, and C-value. Concentrations below the A-value are considered 'multifunctional': the site can be used without any restriction. If the dissolved Cd concentration exceeds the B-value, further research is necessary and values above the C-level require t r e a t m e n t (sanitation). For Cd (and similar results were obtained for Zn), more t h a n 70% of all samples in forest soils exceeded the B-value and 15% even exceeded the C-value. For arable soils only 20% of all samples exceeded the B-value (Figure 3). In arable soils more t h a n 90% of these values were found in the topsoil and are directly related to the application of Cd to the soil. In forest soils, however, a large p a r t of the high dissolved Cd concentrations were encountered in the deeper soil
375 horizons (between 50 and 80 cm) which reflects the poor binding capacity of the soil due to a low soil pH combined with a low organic m a t t e r and clay content.
A r a b l e soils
F o r e s t soils I
0
I
20
D
I
I
40 60 80 percentage (%)
D
D
m
100
standard A B C
(ug/L) 1,5 2,5 10
A-B B-C >C
Figure 3. Exceedances of soil quality standards for dissolved Cd concentrations in Dutch arable soils and forest soils.
3.2 M o d e l r e s u l t s a n d c o m p a r i s o n w i t h f i e l d d a t a
Soil acidification The model results indicate t h a t termination of liming leads to a significant acidification of the topsoil within 40 years as is shown in figure 4. This g r a p h shows the cumulative frequency distribution of the predicted pH values in clayey and sandy soils for the 4 time periods (1990, 2000, 2010, 2050). In all simulations, the deposition level exceeded the critical level, which resulted in a significant reduction of the base saturation and soil pH. The predicted decrease in the s a n d y soil from 5 tot 3 - 3.5 was in reasonable agreement with the m e a s u r e d pH in t h e old forest stands. The difference between the m e a s u r e d and calculated pH is partly due to the fact t h a t the model takes into account the 0 40 cm layer, whereas the measured values stands for the 0 - 30cm layer. The predicted decrease in soil pH in clay soils is larger compared to forest soils during the first few decades. This is partly due to the stronger relationship between soil pH and base saturation in the 'high' pH range (5 - 7). For clay soils, the initial base saturation was set at 100% compared to 50% in the s a n d y soils. Consequently the base saturation in forest soil diminishes w h e r e a s the base s a t u r a t i o n in clay soil remains higher t h a n 50%.
376
C u m u l a t i v e f r e q u e n c y (%)
C u m u l a t i v e f r e q u e n c y (%)
100
,,~
2
!
100
!
,.
.o
t
9 f"
I
9
r
80 . . . . . . . . . . . . . . . . . . . . . . . . . . . . " ..... +. . . . . . . . . .
80 . . . . . . r.- --~. . . . . . . . . . . . . . . . . . . . . . . i. . . . . . . . . ! !
9
~-'-"-.,"'. f
9
!
60 . . . . . . I. . . . . . . . . . . . . . . . . . . . . . . . :"'~ . . . . . . . . . . . :9 I' A
] . , 60 . . . . . . . . . . . . . . ] . . . . . . . . . . ::...... ] . . . . . . . . . . . . . B " tf "9
40 ...... [. . . . . . . . . . . . . . . . . . . . . . . :---4. . . . . . . . . . . . .
40
..........
20
........
I 20
. . . . . . 1"
,"
I..I
! !
.........
3.5
I. . . . . .
4
" ""
I
4.5
2000:
-'. .... l .................
.
! ...........
,/
! . S
.I. . . . . . . . . . . . . . . . . . .
/9 .....
I
.
!
I
5
5.5
o;
-
9
..
:
, i
,
.', 6
6.5
pH
pH 1990:
l ..........
,
-'! . . . . . . . . . . . . .
9
I 03
9 9
2010: ........
2050: - ~. -
F i g u r e 4. S i m u l a t e d f r e q u e n c y d i s t r i b u t i o n s o f soil p H i n s a n d y s o i l s (A) a n d c l a y soils (B)
Cd release D u e to t h e d e c r e a s e i n soil p H , a l s o t h e a d s o r p t i o n - c o n s t a n t for C d d e c r e a s e s ( e q u a t i o n 2) a s is i l l u s t r a t e d i n t a b l e 1. A c c o r d i n g to e q u a t i o n 1 t h i s l e a d s to higher dissolved Cd concentrations. In figure 5 the cumulative distribution of t h e r e l a t i v e i n c r e a s e i n t h e d i s s o l v e d C d c o n c e n t r a t i o n i n s h o w n . I n c l a y soils, t h e r e l a t i v e i n c r e a s e is e v e n l a r g e r c o m p a r e d to s a n d y soils, b u t t h e a b s o l u t e c o n c e n t r a t i o n w i l l b e l o w e r d u e to t h e h i g h e r a d s o r p t i o n - c o n s t a n t s i n c l a y s o i l s a s is i l l u s t r a t e d i n t a b l e 1. Table 1 A d s o r p t i o n c o n s t a n t s for C d i n soil ( m 3 . k g -1) clay sand 1990 2.101 0.19 2000 1.30 0.17 2010 0.95 0.13 2050 0.60 0.08 1: t h e s e v a l u e s r e p r e s e n t t h e m e d i a n v a l u e for b o t h soil t y p e s I n c a s e o f a p H d r o p f r o m 5.5 to 3.5, d i s s o l v e d C d c o n c e n t r a t i o n s a r e e x p e c t e d to i n c r e a s e w i t h a f a c t o r 2 to 6. T h i s w a s a l s o o b s e r v e d for s a n d y s o i l s u n d e r f i e l d c o n d i t i o n s w h e r e C d s o l u t i o n c o n c e n t r a t i o n s w e r e f o u n d to i n c r e a s e f r o m 2 - 5~t g . L -1 to 10 - 15~tg.L -1 w i t h a d e c r e a s e i n soil p H f r o m 5.5 to 4.
377
100
Cumulative frequency (%) ,,," g . ~ . - - " - - ~ . . - , ~ |
'
!
80
.........
9
/
100
I9
80
| " " " t . . . . . . . . . . . . . . .. . . . . . ... ... . . . . . . . . . I 9
:l
/
:!
........ i-I ...... -/--- ........ /
20 ........ i-i ..... t-----
~
1
2
........
3
...........
/
:
./
4
/---
[
;, .... ! .............
A
/ ........
I I 60 ........... : i
/
'i
:,
~ g "
60 ........ :..j ....... ./ .................. 40
Cumulative frequency (%)
: /
t ooo
40
t
20
/
............ j ...........
/
,' '/ ........
: !
B
~
........
.............1.
20001
~,
......
12056 ........ !...i ....... /...
:~o
.j,/ 5
6
~
1
2
I
I
I
I
3
4
5
6
I, 7
Relative Cd-concentration (1990= 1)
Figure 5. Simulated frequency distributions of the relative Cd concentration in sandy soils (A) and clay (B) soils.
4. D I S C U S S I O N A N D C O N C L U S I O N S
Soil acidification a n d Cd release The results from both field measurements and model predictions indicate t h a t the conversion of arable land to forest leads to major changes in the chemistry of the soil solution: i. After termination of liming, the pH of the soil decreases from n e a r - n e u t r a l (5.5 - 6.5) to acid (4 - 4.5) within 30 to 40 years. Despite some assumptions made in the model (see below) the model predictions are in good a g r e e m e n t with the m e a s u r e d pH decrease in the field. ii. The pH decrease will lead to a significant increase in the dissolved Cd concentration. In both field and model results, dissolved Cd concentrations increased with a factor 2 to 6 with a decrease in soil pH. In m a t u r e n a t u r a l forest soils, dissolved Cd concentrations are high compared to actual arable soils despite the much lower total metal content. This indicates t h a t if soils with an elevated Cd total content (0.5 to 2 mg-kg -1) are converted to forest, dissolved Cd concentrations will even exceed the actual concentrations encountered in forests.
378
Uncertainties and model improvements Despite the good agreement between field data and model predictions several assumptions have been made t h a t deserve further attention: i. The acid input was based on external deposition only. After forestation a strong internal acidification will occur due to a net mineralization and nitrification which is not compensated for by NO 3- uptake. On a longer term, a net immobilization will occur due to the formation of a an organic horizon. ii. Interactions between the amounts of H and A1 leached and the base saturation on one h a n d and the distribution between dissolved Cd and the adsorbed fraction on the other h a n d have not been included in the model. The cumulative amount of Cd mobilized in sandy soils during the simulation run, amounted up to 75% of the total Cd content in the soil. This decrease will probably affect the equilibrium between the solid phase and the dissolved Cd concentrations as well. iii. The interaction between DOC and Cd has not been t a k e n into account. Although the interaction between Cd and DOC is less strong compared to Cu or Pb, the increase in the DOC solution concentration (figure 1) could still affect the dissolved Cd concentration as well (see point 4) as plant uptake. iv. One of the major uncertainties at the moment is the role of the vegetation. Due to acidification, Cd is released to the soil solution. However as a result of the forest development, a net increase in the plant uptake from the soil solution seems possible. This could mitigate the effect of the Cd mobilization and prevent any leaching to the ground water. However, due to the increase in the DOC concentration, the degree of Cd-complexation (in solution) also increases. Several studies showed that the uptake of organically complexed metals is strongly reduced compared to the uptake of the free metal (Kuiters and Mulder; 1993). Recent work done by E r n s t showed t h a t the uptake of metals by plants on a contaminated site decreased after a few years due to the increase in the organic m a t t e r content and the complexation of the metals in solution (Ernst, 1994). The role of organic matter is therefore two-fold: it increases the capacity of the soil to retain metals from solution (to the solid organic phase) but it simultaneously decreases the uptake by plants as a result of the formation of aqueous complexes.
Future needs and integration of Acidification research with Trace Metal studies This study clearly shows t h a t the combination of soil acidification research combined with trace metal monitoring studies can be used as a tool to predict the consequences of land use changes. At present several large-scale activities affecting both land use and m a n a g e m e n t are being carried out. In the Netherlands this includes both forestation of arable land as well as the creation of wetlands along the main rivers. In order to assess the chemical consequences of these new m a n a g e m e n t strategies it will be necessary to combine studies concerning both the macro-chemistry (e.g. Ca-A1 chemistry in combination with soil pH as well as redox chemistry) and the micro-chemistry (behaviour of trace
379 elements and organic chemicals). In this context, the large experience gained in acidification and trace metal research (both in field/laboratory experiments and modelling studies) will be useful to couple various models developed in the different fields. In case of forestation, especially the processes of acidification combined with the organic m a t t e r dynamics (development of an O h layer, increase in DOC content, A1 - organic m a t t e r interactions) need further attention in order to assess future consequences of present activities in relation to soil quality and protection. With respect to land use changes in general, it can be concluded that there is still a lack of information concerning the dynamics of soil chemical processes on a time-frame of 10 to 100 years.
5. R E F E R E N C E S
Bergkvist, B. 1987. Soil solution chemistry and metal budgets of spruce forest ecosystems in S. Sweden. Water, Air, and Soil Pol., 33:130-154. Bloom, P.R., and D.F. Grigal. 1985. Modelling soil response to acidic deposition in non-sulfate adsorbing soils. J. Environ. Qual., 14:489-495. Breeuwsma, A., J.H.M. WSsten, J.J. Vleeshouwer, A.M. Van Slobbe and J. Bouma. 1986. Derivation of land qualities to assess environmental problems from soil surveys. Soil Sci. Soc. Am. J., 50:186-190. Bril, J. 1993. Transfer functions between adsorption constants for heavy metals and soil characteristics, AB-DLO Inst. for Agrobiology and Soil Fertility Res., internal document. Castilho, del, P., W.J. Chardon, and W. Salomons. 1993. Influence of cattlemanure slurry application on the solubility of Cadmium, Copper and Zinc in a manured acidic, loamy-sand soil. J. Environ. Qual., 22:689-697. Chardon, W.J. 1984. Mobility of Cd in soils (In Dutch). Ph.D. thesis Wageningen Agricultural University. Soil Protection Series no. 36, Staatsuitgeverij, the Hague, 200pp. Christensen, T.H. 1989. Cadmium soil sorption at low concentrations. Department of Environmental Engineering, Technical University of Denmark, Polyteknisk Forlag, 314pp. Driel, van, W., and K.W. Smilde. 1990. Micro-nutrients and heavy metals in Dutch agriculture. Fertilizer Res., 25:115-126. Ernst, W.H.O. 1994. Bioavailability of heavy metals and decontamination of soils by plants, p. 110. In E. Helios Rybicka and W.S. Sikora. Proc. from the 3 rd Int. Symp. on Environmental Geochemistry, Krakow, Poland, 12-15 September, Primar Druk, Krakow. Harter, R.D. 1982. Effect of soil pH on adsorption of lead, copper, zinc, and nickel. Soil Sci. Soc. Am. J., 47:47-51. Hoekstra, C. and J.N.B. Poelman. 1982. Relation between soil bulk density and general soil characteristics in the Netherlands (In Dutch). STIBOKA report no. 1582, Wageningen, 47pp.
380 Kuiters, A.T., and W. Mulder. 1993. Water-soluble organic matter in forest soils II: Interference with plant cation uptake. Plant and Soil, 152:225-235. McBride, M.B. 1989. Reactions controlling heavy metal solubility in soils, p.155. In B.A. Stewart (Ed.). Advances in Soil Science, vol. 10, Springer-Verlag, New York. Oden, W.I., G.L. Amy, and M. Conklin. 1993. Subsurface interactions of humic substances with Cu(II) in saturated media. Environ. Sci. Technol., 27:10451051. Page, A.L., R.H. Miller, and D.R. Keeney (eds.). 1987. Methods of Soil Analysis~ part 2: Chemical and microbiological properties. Agronomy Monograph 9, 2na ed. ASAJSSSA Inc. Madison, WI. 1184pp. RSmkens, P.F., and W. Salomons. 1993. The non-applicability of the simple K dapproach in modelling trace metal behaviour: a field study, p. 496-499. In R.J. Allen and J. Nriagu (eds.) Proc. from the Int. Conf. Heavy Metals in the Environment, Toronto, Canada, 12-16 September, CEP consultants Ltd, Edinburgh. Stoop, J.M., and A.J.M. Rennen. 1990. Harmful chemicals in agriculture and horticulture, volume 1 (in Dutch). Center for Agriculture and Environment, Utrecht, the Netherlands, 122pp. Vries, de, W. 1994. Soil response to acid deposition at different regional scales; Field and laboratory data, critical loads and model predictions. PhD thesis Wageningen Agricultural University, 487pp. Vries, de, W., and P. RSmkens. 1994. Mobilization of Cd as a result of land use changes (in Dutch). Bodem, 4:76-79.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
381
Acidification interacting with global changes" research to manage drii~ing systems R.S.A.R. van Rompaey P r o g r a m m e Office of the Dutch National Research Programme on Global Air Pollution and Climate Change NRP, c/o RIVM (pb 59), P.O. Box 1, NL-3720 BA Bilthoven, The Netherlands, tel: +31-30-743781, fax: +31-30-251932, e-mail: [email protected]
Abstract A comparison is made between acidification and the climate change problems. The common cause and the variety of effects with an impact on m a n y different ecosystems and aspects of society are discussed. A number of specific aspects of environmental pollution research are highlighted. For the Dutch Climate Change Program 1995-2000, a more system-oriented approach in stead of an effectoriented approach is chosen. Research should determine the degrees of freedom the society has to manage these different systems.
1. Acidification and global change Air pollution has many negative effects, both on our natural environment as on our health and well-being. Since the industrial revolution, fossil fuel has provided the energy for a spectacular improvement of hu m an welfare. In the mean time, all combustion products were emitted to the atmosphere where they were supposed to infinitely dilute. Certain emissions like dust and heavy metals, do not travel a long distance and have to be sanitated in an early stage. A number of compounds, like the acidifying ones, travel across the national border and international negotiations are needed to abate these emissions. Last but not least, the lightest gases like CO2 and CH4 travel around the world and still proved to have an impact on the emitting societies by influencing the climate system in different ways. The major and common cause to both acidification and climate change is the
emission of fossil fuel combustion products to the atmosphere. Besides, NH3 emission from manure also causes acidification, and land cover change, e.g. in the tropics, causes climate change. The direct relation t h a t exists between h u m a n welfare, energy consumption and air pollution, implies t h a t all three increase exponentially and t h a t a clean environment can only be achieved by adopting a less materialistic life style.
382 2. E m i s s i o n s ,
effects and impact
By burning fossil fuel a variety of compounds are emitted: H20, CO 2, CO, SO 2, NOx, VOC, etc. Each compound has a specific atmospheric chemistry, transport rate, residence time and thus dispersal over the globe and, finally, deposition rate. A first category of effects of their deposition are of (bio-)chemical nature: - acidification - oxidation - intoxication - fertilisation The second category has to do with the physics of the atmosphere: - radiative forcing by reflection of the earth's emitting infra-red radiation back to earth, resulting in global warming (Houghton et al. 1992), - regional cooling because of reflection of solar radiation by aerosols (Taylor & Penner 1994) - absorption/transmittance of UV radiation (De Backer 1994) These effects act on the earth and may induce secondary effects: - sea level rise - more frequent extreme weather (drought, heavy rain, storm, ...) - large scale biotic feedbacks to the climate system because of increased nutrient levels (CO2, N) and improved water efficiency (Goudriaan 1992). All these effects have an impact on a wide variety of ecosystems t h a t are of importance to man, and also directly on the h u m a n populations and societies. Ecosystems touched by these environmental changes are both terrestrial (forests, lakes, crop lands and non-woody systems like prairies, savannah, tundra, etc.) and marine (coastal shelf systems, open ocean, (ant-)arctic systems. H u m a n society is also directly touched by these effects. Health is threatened: respiratory problems by smog, loss of immunity and skin cancer by UV radiation, etc. Man's security against floods and extreme weather is endangered. On a global scale, also food security may be in danger. This may lead to migration and subsequent political instability. A much greater variety of systems is subject to global change research than to acidification research. This also means that the available research funds are to be shared between much more disciplines. 3. A s p e c t s
of environmental
pollution
research
A number of aspects makes environmental pollution research particulary difficult, and should be considered when defining future research programmes on these problems. As indicated above, the geographic scale of each pollutant and its effects may differ considerably. Consequently, there is a need for georeferenced modelling that can handle different scale levels or resolutions simultaneously (Alcamo et al. 1994, Leemans et al. 1994).
383 From a more conceptual point of view, there is a need to attain a higher level of aggregation. The environment is permanently changing and adapted research methods are needed to study this "moving target". In most cases there is a lack of a reference state or control which underwent no treatment. Also modelling exercises are embarrassed by the permanent drift of the object of study, making validation difficult. Ecological sciences need to adapt their concepts to this situation. The policy makers ask environmental scientists for predictions of future evolutions and possible damage or risk. This aspect of scanning the future often forces the scientist to pronounce his expert guess beyond the scientifically certain conclusions of his research work. A last but major aspect of this kind of research is the complexity of the systems under study and the chaotic behaviour of these systems. Biosphere, atmosphere and h y d r o s p h e r e have plenty of feedback mechanisms and multi-factor interactions which make single effect of single factors difficult to separate out. Non-linear relationships result in multi-equilibrium systems, difficult to model or predict.
4. A p p r o a c h of N R P 1995-2000 Based on the experience gained in the first phase of NRP (1990-1994), the Dutch National Research Programme on Global Air Pollution and Climate Change, NRP, has chosen the following approach: - from effect-oriented to system-oriented research
In the first phase many projects focussed on dose-effect relationships. Many climate variables may be considered and almost any system is sensitive to change of one of those. To concentrate research efforts on the problematic cases of climate change, we invited research proposals using a system-oriented approach by selecting a system under study, e.g. the Rhine river basin or the Wadden Sea, and then situating possible impact of climate change against other antropogenous changes affecting the system. - assess the system's vulnerability and adaption capacity
Given the uncertainty and the lack of detail in the prediction of future climate, it is not desirable to consider the impact of only one possible scenario. By assessing the system's vulnerability to climate change, many possible future climate states are considered. At the same time the severity of the damage to the system or its eventual collapse is also assessed. Both natural systems and society are able to adapt to climate change and research should be focussed on adaption capacity and adaptation rates versus climate change rates. - determine degrees of freedom to manage the drifting system As the NRP is a policy oriented research programme, research is also invited in the domain of h u m a n induced adaptation. All the knowledge on foregoing aspects should be used for a wiser management of the environment for present and future generations.
384
5. Message for the acidification research c o m m u n i t y I conclude t h a t the e n v i r o n m e n t a l problem of acidification shows m a n y interactions with global environmental changes such as climate change. To pick up the shift in public and policy maker interest towards global change, I suggest the research community the following: - for atmospheric chemists and physicists: Switch compounds and include all elements present in the atmosphere in the modelling work - for stress ecologists: Include the climate factors when studying plant and ecosystem response. Try to shift from plant to ecosystem level and even to biome level. - for modellers: Introduce a georeferenced framework in the modelling work. Solve the time and space scale problems. Apply a whole system approach and provide handles for management. This conference addressed the question whether we have enough answers in the domain of acidification research?. After this overview of global change problems, one might ask: Don't we have enough problems left ?
References Alcamo J. (ed., 1994). Image 2.0: Integrated modeling of global climate change. Kluwer Acad. Publ., Dordrecht, 318 p. De Backer H. (1994). Analysis and interpretation of ozone observations at Uccle (1969-1993). PhD thesis, Vrije Univ., Brussel, 160 p. Goudriaan J. (1992). Biosphere structure, carbon sequestering potential and the atmospheric 14C carbon record. Journal of Experimental Botany 43 (253): 1111-1119. Houghton J.T., Jenkins G.J. and Ephraums J.J. (editors, 1992). Climate change: the IPCC scientific assessment. Cambridge Univ. Press, 365 p. Leemans R. and van den Born G.J. (1994). Determining the potential distribution of vegetation, crops and agricultural productivity. Water, Air, and Soil Pollution 76: 133-161. Taylor K.E. & Penner J.E. (1994). Response of the climate system to atmospheric aerosols and greenhouse gases. Nature 369:734-737
POSTERS S E S S I O N VIII CRITICAL LOADS / EXCEEDANCES
This Page Intentionally Left Blank
G.J. Heij and J. W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
Critical loads of h e a v y m e t a l s for E u r o p e a n
387
f o r e s t soils.
G.J. Reinds ~, J.Bril b, W.de Vries a, J.E. Groenenberg ~, A Breeuwsma% DLO Winand Staring Centre for Integrated Land, Soil and Water Research. Postbox 125, 6700 AC Wageningen, The Netherlands b DLO Research Institute for Agrobiology and Soil Fertility Postbox 129, 9750 AC Haren (Gn.), The Netherlands
Abstract Recently, concern has arisen about the impact of the dispersion of heavy metals and persistent organic pollutants in Europe. Therefore, a study (ESQUAD) was initiated to assess critical loads and steady state concentrations of cadmium, copper and lead and lindane and benzo(a)pyrene for European forest soils and the North sea. This study was carried out by five Dutch research institutes. This poster will present critical loads and steady state concentrations, in both adsorbed and dissolved phase, of Cd, Cu and Pb for European forest soils. Heavy metal adsorption in the soil was computed using transfer functions between adsorption constants and soil properties such as CEC and pH. The (steady state) pH was calculated using models and databases developed for acidification research. Excess loads were computed using results from the emission inventory and subsequent deposition calculations for Cd, Cu and Pb. Results show that the computed critical loads and associated excess loads strongly depend on the threshold values chosen and on the soil phase (adsorbed/dissolved) considered. When the most stringent threshold values are used, excess loads are found all over Europe, whereas the less stringent threshold values lead to critical loads that are hardly exceeded. Further research is needed to improve both input data and the modelling of heavy metal adsorption.
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV All rights reserved.
389
Setting critical loads of acidity for dystrophic peat- a new approach E J Wilson ~, R A Skeffington ~, C J Downer ~, E Maltby 2, P Immirzi 2 & C Swanson 2
National Power Research and Engineering, Windmill Hill Business Park, Whitehill Way, Swindon, Wiltshire SN5 6PB, UK.
Wetland Ecosystems Research Group, Department of Geography, University of Exeter, Amory Building, Rennes Drive, Exeter EX4 4RJ, UK.
Abstract Current methods of determining critical loads for mineral soils cannot be applied to dystrophic peat. The experiment described here aims to a) investigate how peat responds to increases and decreases in acid deposition and b) calculate a critical load.
1. I N T R O D U C T I O N
The critical loads approach is now widely accepted as a tool for pollutant abatement strategy. The aim is to maximise environmental benefits by targeting emission reductions in ecologically sensitive areas, rather than uniformly. Critical load maps for acidity identify soil as the sensitive element, and aim to protect it from further acidification. For mineral soils, the critical load of H + is related to the rate of production of base cations (or weathering rate) and is a function of soil type and underlying mineralogy. This approach cannot be used for dystrophic peat soils where the underlying bedrock has no influence on the availability of base cations in the surface layers. Most countries in Europe have arbitrarily allocated these soils to the lowest (most sensitive) weathering class. National Power Research & Engineering and the Wetland Ecosystems Research Group at Exeter University are using a combination of field and experimental investigations to determine the effect of acid deposition on dystrophic peat with a view to estimating a critical load.
2. M E T H O D S
Intact cores (including the vegetation) of the hydrologically active portion of the peat profile (acrotelm) were taken at 8 sites, using a 10 cm diameter drainpipe. The sites were selected to encompass the range of H + deposition and peat type in the UK. In the laboratory they
390 were treated with simulated rain which mimicked the ion composition at the site. The six treatments are equivalent to a reduction in acid loading (0.2x and 0.5x H § deposition), an increase in acid loading (2x, 4x and 6x H § deposition) and current inputs (lx H § deposition). Two years rainfall was applied over 3 months and the leachate analysed weekly, pH and exchangeable cations were determined on the peat itself at the end of the experiment.
3. R E S U L T S
Figure 1 shows the change in pH of the six different rain solutions after passing through Calluna peat from one of the more polluted sites in the South Pennines. Where the change is zero, the pH of leachate leaving the core is the same as that of the rain applied, indicating that the rain and peat are in approximate equilibrium. At this site, peat appears to be in equilibrium with rain between 0.5x and Ix current acid loading. It is thus unlikely to be acidified by current deposition. Increasing acid deposition above current levels did acidify the peat (ie the pH of leachate was increased relative to the rain applied). Data from other sites show that peat can also be in equilibrium with rain that is both more and less acid than it is currently receiving.
0
4
o3 r~ E .r-!
I I O.
o
2
0
0
-0
1
-0
2
-0
3
-0
5 6
QI
.~ c"
u
E
-o
$s u
-o
7
I r~
-o
(3 ~
2
'
~
3
'
'
4
~
5
'
~
'
6
~
~
7
8
'
~
9
~
10
'
'
r
~1
~
Months
Treatment
-~
~
~
0 .2x
"
"
"
2x
H+ H+
'
12
of
v^ ~
v^
~
13
~
~
14
f
15
~
~
~
16
~
17
~
18
~
19
~
20
21
22
Rainfa12
v^
0 .5x 4x
H+ H+
~" ~"
~~
~x
H+
o
o
6x
H+
o
Figure 1. The change in pH of rain passing through Calluna peat. The concentration of H § in the Ix treatment is equivalent to that in ambient rainfall at the site.
391
4. D I S C U S S I O N
Preliminary results suggest that drainage waters from peat are more sensitive to changes in acid deposition than the soil itself: changes in leachate pH during the experiment were not reflected in peat pH at the end of the experiment. Since the existing pool of exchangeable H + is so large in these acid peat soils, it may take many years for changes in rain acidity to make an impact on the pH of peat. Peats from different sites responded differently to increases and decreases in acid deposition and cannot be treated as a homogenous material. Data from the 8 sampling sites should provide information on how factors such as peat type and deposition history influence this response. When fully analysed, these results ought to enable us to predict the chemical response of peats to variations in acid load. The data will be used to define a damage function and determine a critical load.
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BE All rights reserved.
393
A comparison of models for the assessment of critical loads on different scales of observation R.J.M. Lenz,, S. Mendler,, R. Staryo 9G S F - F o r s c h u n g s z e n t r u m ftir U m w e l t und G e s u n d h e i t , Projektgruppe Umweltgef~ihrdungspotentiale von Chemikalien, Neuherberg, Postfach 1129, 85758 Oberschleigheim, F.R.G. bIngenieur-Btiro f'tir Landschaflsinformatik, Freising oLehrstuhl f'tir Landschafts6kologie, TU Mtinchen-Weihenstephan ABSTRACT
Three simple models for the assessment of exceedances of critical loads for acidity, and eutrophication due to nitrogen, were applied in northeastern Bavaria on a mapscale of about 1:25.000. In order to get some rough ideas about their validity, the results were compared with three degrees of severe stand-level diebacks in spruce stands (Picea abies L.). The areal percentages show a similar pattern, and the most simple model seems to fit best. Hence it is concluded, that the accuracy and availability of data e.g. deposition and base saturation, soil depth, and lateral transports, are often more crucial than the quality of the models itself. 1. I N T R O D U C T I O N In 1986 Nilsson & Grennfelt (see in 1) described the concept of critical loads for the first time; the definition itself: "The highest load that will not cause chemical changes leading to long-term harmful effects on the most sensitive ecological systems" was outlining the long term effects and was so far precise in its temporal scale. Ecological systems are somehow dimensionless and can be chosen according to the specific situation. Two years later the definition of critical loads was changed a little bit: "A quantitative estimate of an exposure to one or more pollutants below which significant harmful effects on specified elements of the environment do not occur according to present knowledge". Now it is more open to all kind of temporal and spatial scales, which on the one hand has the advantage of a wider use of this evaluation concept; but on the other hand the applicants and modifier of this concept must define their own and relevant scales, also depending on the resolution of the data (2). 2. M A T E R I A L AND M E T H O D S Up to now a distinction was made among three types of critical loads calculations (see in 1): level O (- use of existing data to assign critical load classes to ecosystems based on ecosystem sensitivity); level 1 (= steady state modelling, divided into a water chemistry method and a mass balance method); and level 2 (- dynamic modelling). This classification only shows the degree of resolution in time and space. Their adequacy for specific loads, receptors, data bases and other purposes (e.g. specific locations, time horizons of prognosis) can not be recognised itself. With the following examples three model conceptions (two level 1, one level 0) are applied and compared with severe stand level diebacks in a mountainous area in northeastern Bavaria. Unfortunately the maps cannot be printed in this paper, but could be recognized on the poster.
394 3. R E S U L T S The simulation results are shown in Table 1. Table 1 Frequencies of exceedance classes of critical loads in comparison to different degrees of severe stand-level dieback (%) Exceedance class
1
2
3
4
Model 1 Severe dieback Medium dieback No dieback
1
11 47
92 72 39
8 17 13
Model 2 Severe dieback Medium dieback No dieback
40 62 92
44 30 7
16 8 1
Model 3 Severe dieback Medium dieback No dieback
8
5 18 44
79 73 40
16 9 8
Model 1: The most simple modelling concept for the exceedance of critical loads substracts the weathering rate from the deposition rate. In this application the weathering rate was derived from geology and soil depth. The deposition rate was calculated by some open air measurements, multiplied with factors derived from surface roughness a.o. (level 1). Model 2: In this example the N-deposition rate was compared with the empirical threshold (critical load) of 13 kg N ha-1 a-1 for vegetation changes in coniferous forests (level 0). Model 3: Due to the fact that the pool of base cations in the soil buffers acidity there should be a delay of harmful effects proportionally to the size of the exchangeable pool of base cations. Hence reaction types were derived from soil estimates in combination with acid depositions (level 1). 3. F I N A L R E M A R K S In already longtermed heavily polluted areas the stand level dieback correlates with the exceedances of critical loads of acidity, nitrogen surplus causing eutrophication, and ecosystem reactions due to losses of base cations. In general, for the calculation of exceedances with different resolution and parametrization the availability of data seems to be more limiting than the quality of models. Lateral transports of material in the soils seem to be keyfactors at the landscape level in this area and should be implemented in the models in a next step.
4. R E F E R E N C E S CCE (Coordination Centre for Effects), RIVM-report (1993) 163 R. Lenz and W. Haber, Vegetatio 89 (1990) 121-135
POSTERS SESSION IX W E T D E P O S I T I O N / THROUGHFALL
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
397
DEPOSITION AND LEACHING AT FOREST STANDS IN SUMAVA Mts. Jana Kubiznfikovgt * Jirf Kubiznfik * and Ota Rauch** *Institute of Landscape Ecology, Czech Academy of Sciences, Na sadkach 7, 370 05 Cesk6 Budejovice, Czech Republic **Botanical Institute, Czech Academy of Sciences, Dukelska 145, 379 82 Trebon, Czech Republic
Abstract Bulk deposition, throughfall, and soil water samples were collected in Sumava Mts. Czech Republic from 1988 to 1992. Canopy exchange budget and soil water output for the major ions at mountain spruce forest stands were estimated. The deposition and output from soil of determinated ions differ greatly among forest stands. The positive ecosystem budgets were calculated for most elements.
1. INTRODUCTION Knowledge of the transfer and budgets of ions in forest ecosystems is important for the estimation of cycling rates and long-term effects on biological systems. The prevalent part of Bohemia is known for heavy acid deposition. Sumava Mts. are a mountain range situated in South Bohemia, relatively the most protected area. Nevertheless, in the Sumava National Park more than one third of the forest is damaged.
2. STUDY SITES The forest stands were situated at elevation range from 800 to 1300 m. The vegetation is dominated by Norway spruce. The soils are brown forest soils and podzols developed on the slope deposits derived from gneiss. Mean annual precipitation sum is about 950 mm, mean annual temperature reaches about 4.8 ~
3. METHODS The samples were collected by using bulk vessels with separating of coarse particulates and lysimeters for mineral soil leachates approximately monthly. All water samples were analyzed for pH by combined glass electrode and specific conductance with conductivity cell. Anions were determined by isotachophoretic method and cations by ICP method.
398 4. RESULTS A N D DISCUSSION Sulphates, nitrates and potassium dominated in bulk deposition on the all stands (see Figure). With elevation is enhanced precipitation and pH- values decreased in bulk deposition, but reach the higher value in throughfall. Flux of major cations in the throughfall was higher than in bulk deposit on the lower stands. Oposite effect was observed at elevation above 1000m. Dry deposition was calculated about 10% of total deposition. Some results confirmed prevailing wind direction and flux of major ions from bulk, throughfall (wash-effect) and leaching from needles. The Sindlov throughfall of potassium was the highest. Chronical exposure of canopy by sulphates and nitrates are reflected by higher throughfall values. Tremendous variability in percolated volume were observed both temporally and spatially. Percolation of water at stands Nove Hute and Vodarnaoccurs mainly after snowmelting or after heavy rain. At Sindlov leaching was prevented due to higher water retention capacity of humus horizon and low snow accumulation. In most cases the concentrations of ions in lysimetric waters were independent on the water volume, making the water flow desicive for the budget. The leachability in forest soils differs greatly among ions but positive budgets were calculated for the most of the ions. There is net release of some metals from the B horizon at stands Nove Hute and Vodarna. The only metals with a net outflow throughout the A and the B horizons of these stands was AI.
Annual flux [kg/km2]
Annual flux [kg/km2]
Vodarna
Nove Hute
/
/ / A 8000" r / J A 7ooo-r. /
eooo-/
/
/, 450O
V~ /// / / 2soo. / /// 2oo0/// Isoo. Iooo./ / /
A
30o0
4ooo./ / A aooo./ A /
2000./ A / ooo-/ L4 o.A~/~
ughf~ ,po.~ CA
K
MG
NA
H
NO3
S04
Sindlov
/1 /
ALX k/1 / /
A A
/
/I
/
/I A
/, ughfall ~posit AL
u~ ,posit
soo. O"
Annual flux [kg/km2]
/I
/ /
4oo0 V / / / asoo
sooo./ / A
AL
/
,i AL
CA
K
MG
NA
H
NO3
S04
G.J. Heij and J. W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.
399
P r e c i p i t a t i o n input of inorganic c h e m i c a l s in the S. Vitale p i n e s t a n d of Ravenna (Italy) T. Georgiadisa, F. Fortezzab, L. Albertib, P. Rossinib and V. Strocchib aC.N.R.- FISBAT, Via Gobetti 101, 1-40129 Bologna, Italy bPMP-U.S.L.35, Via Alberoni 17, 1-48100 Ravenna, Italy
Abstract This study reports the differences in the local precipitation input of inorganic chemicals at urban, industrial and woodland sites of a coastal area in n o r t h e r n Italy. The role of the local wind circulation was investigated to determine the contribution of the industrialised area to pollution above the woodland in terms of chemical (SO4) and particle (Ca++) depositions. Data collected via the monitoring network of greater Ravenna were compared with those collected at an Apennine station (Brisighella, 80 km SW from Ravenna) taken as a "clean reference". Findings show the influence of wind direction in contributing to the local dispersal of chemicals at the different sites. 1. Introduction The solution for environmental problems related to anthropogenic emissions is often connected to the quantification of the airborne chemical input to the terrestrial ecosystem, as well to the definition of the atmospheric patterns which act as "carriers" in returning the pollutants to the earth's surface. The main pathway of this pollutant-return to the ground must be attributed to pollutant deposition in both wet and dry processes (1). Wet deposition can be roughly defined as the process by which hydrometeors deliver to the ground the primary pollutants or the chemicals produced via several reactions in the presence of water (2-3). This takes place by means of wash-out (impaction and interception) and rain-out (in-cloud nucleation, molecular diffusion, phoretic forces and condensation) mechanisms. It is clear that the ground concentrations produced by wash-out can be considered indicative of the local atmospheric pollutant contents. On the other hand, the depositions produced by the rain-out mechanism are indicators of the chemical composition of the clouds, which can travel long distances after their formation, picking up varying chemical "contributions". While wash-out gives specific information on the local atmospheric load of pollutants, wash-out and rain-out are often difficult to separate. This study was designed to elucidate the importance of wet deposition in determining the environmental profile of sites belonging to the same geographical area but characterised by varying topographic features. The possible role played by the local wind circulation in modulating the deposition of chemicals produced locally over the ancient woodland area of the S. Vitale pine stand was also investigated.
400
2. Site d e s c r i p t i o n and e x p e r i m e n t a l Located in northern Italy in the part of the Po valley facing the sea, great er Ravenna is marked by a flat terrain bordered on the north by m arshl and and on the east by the sea. The Apennine m o u n t a i n range forms the boundaries, stretching across approximately 100 km westward and 60 km southward. A strip of sea pines along the shoreline reaches across the whole area. The city proper is located 8 km inland and an extensive industrial belt (petrochemical plants, an electric power station and other industrial sites) extends from the city to the coast. Pollutants were monitored at three sites (urban, industrial and a pine stand) to determine the differences in chemical composition, concentration (SO4-- and Ca++) and pH value for each meteorological condition. An additional station was set up at Brisighella (40 km SW of Ravenna in the Apennine range) to act as an indicator of clean atmosphere. The monitoring stations were equipped with wet and dry collectors (MTX, Italy); all samples were examined using ionic chromatography (Dionex 4000I, USA) and atomic absorption spectrophotometry (Perkin-Elmer 2280, USA); pH was measured with an Orion EA 940 (USA) analyser.
3. R e s u l t s and d i s c u s s i o n Figure 1 shows the average precipitation intensity for the prevailing wind directions (NNE, WNW, ESE and SSW) at the four sites considered (1: city, 2: industrial zone, 3: pine stand and 4: Brisighella) over the period 1992-1993. Measurements were taken for each wind direction classification. For each class, the first three stations of the Ravenna area showed no significant differences in the precipitation amount, while station 4 recorded a larger amount of precipitation in each class due to its mountain position. If the different wind direction classes are compared, those having a westerly orientation (normally associated with the lows originating over the Tirrenian Sea) representing a prevail, common climatic pattern of Italy. Figure 2 indicates the pH maxima and minima along with the upper and lower quartiles, and the median value. Despite the m axi m um pH value of 8.2, the woodland area exhibits more pronounced acidification in 25 % and 75 % of the data. By contrast, the lowest acidification is recorded at the industrial sites where chemicals are emitted. The analysis of SO4-- and Ca++ concentrations (Figures 3 and 4) indicates a possible explanation of the pH patterns. The industrial area appears much more vulnerable to the presence of SO4-, reaching twice the concentration of the Apennine station, although the pH median value is about 5.8, the least acid. Particles may be posited as having a role (where Ca++ is their indicator) in neutralising the acidity of the atmospheric wash-out. The more marked concentration of Ca++ in the city and industrial areas can lead to a neutralisation of the acid compound emitted by the industrial plants through a "buffer effect". The mean deposition levels of SO4-- and Ca++ in the pine stand appear influenced by the wind direction (Table 1). The SO4-- concentration values increase for westerly winds when the pine stand is downwind from the city and the industrial belt. Analogous to this is the pattern in the Ca++ deposition, along with an increase for ESE winds ascribable to the transport of sand from the nearby shoreline.
401 Table 1 Mean deposition (rag/m2) SO4-site site site site
1 2 3 4
NNE 64.9 81.9 56.5 51.4
WNW 95.5 104.3 67.0 66.2
ESE 54.3 82.8 51.4 51.7
Ca++ SSW 88.1 101.9 63.8 50.5
NNE 13.9 14.9 11.8 21.4
WNW 20.1 19.0 14.8 18.1
ESE 17.8 18.9 18.4 41.7
SSW 28.9 28.2 22.9 54.5
4. C o n c l u s i o n s
The findings support the hypothesis that the local acidification of the woodland area should be ascribed to the urban and industrial emissions of greater Ravenna. The deposition levels recorded for the two chemicals considered are related to the prevailing wind direction during the precipitation events. 5. R e f e r e n c e s
1. B.B. Hicks, Water Air Soil Pollut., 30 (1986) 75. 2. J.M. Hales, Atmos. Environ., 12 (1978) 389. 3. L.A. Barrie, J. Geophys. Res., 90 (1985) 5789.
402
50
82
40
6,9
6,9
4,3
4,3
site 1
site 2
6,9
30
~6 20 10 4,1
4,2
site 3
site 4
,/
E-SE N-NE W-NW S-SW I~site 1 i s i t e 2 I~lsite 3 r~site 4
Figure 1. Average precipitation.
Figure 2. pH values recorded at the four m e a s u r e m e n t sites.
8
2,5
7
6
I
5 0) 4 E 3
m
1,5
J
2
0,5 1 0
,
i
i
i
site 1
site 2
site 3
site 4
Figure 3. 8 0 4 concentration values.
site 1
site 2
site 3
site 4
Figure 4. Ca ++concentration values.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
403
Trends of some Components of Wet Deposition in East Germany after the Unification E.BrLiggemann and W.Rolle Institute of Tropospheric Research, Permoserstr. 15, 04303 Leipzig, Germany
Abstract A precipitation network was established in East Germany consisting of 15 sites in 1991. Rain water samples were analysed and characterized regarding acidity, conductivity, main water soluble components, volume, and meteorological parameters. By means of long-term data records can be shown a significant decrease of sulfate-S and calcium in the precipitation. The acidity and the nitrogen-containing ions have been remained unchanged up to now.
1. INTRODUCTION In 1991 an integrated research project (SANA) was established (supported by the Bundesministerium for Forschung und Technologie) to observe the changing air pollution situation and the effects on the biosphere from 1992 to 1995 in East Germany. The air situation was characterized of enormous SO 2- and alkaline dusts pollution before 1990. After the unification of Germany in 1990 the industrial structure was reformed, the stock farmings were reduced, and the individual and goods traffic has increased drastically, causing great changes in the emission pattern. A part of this project is the determination of the changing of air pollution following wet deposition.
404
2. EXPERIMENTALS In 1 991 a precipitation network, consisting of 1 5 sites, was established in East Germany. Three older stations (S,G,W) with long-term data records since 1 983, 1 984, and 1 988, resp., could be integrated. Wet-only collectors (automatically opening at start of precipitation and closing after event ) are used to sample rain water (1). Usually, collection intervals of 4h or 24h are chosen. The main components in the precipitation (CI-,NO3-,SO42-,Na + ,NH 4 + ,K +, Ca 2 + , M g 2 + ) are analysed by ion chromatography (before 1991 with spectralphotometrical methods and AAS). Additionally, acidity, conductivity, and meteorological parameters are determined (2,3). A QA/QC plan was adopted from the U.S. Air Quality Monitoring Network (4).
3. RESULTS AND CONCLUSIONS The changes of air pollution begun in 1990 and 1991. The measurement program of the project SANA started in 1 992. Figure 1 shows the SANA precipitation network (description in 2). We have stations in industrial polluted areas (W,L,H,M,R,GZ,CH),in rural areas (G,S,N,A,AR), and at three mountainous sites (O,UE,C) influenced of emissions by the northern Bohemian industrial area.
Greifswald (G)
Neuglobsow • / (N) Angeml{mde
Iiesenburg "* (W)
Brocken
Artern (AR}
~-( _
SchmOcke
Lindenberg (LI /)
Radebeul
j.~
Figure 1 PRECIRTATION NETWORK SANA COLLECTION TIME : o - 24-hour • - 4-hour
)berb~renburg (0)
405
In industrial regions the wet deposition rates of sulfate-S and calcium are about twice as high as in rural areas. WET DEPOSITION 1992/1993 SULFATE-
S
kg/ha
mm
16 I
1600
14 -t
1400
1
~""
~"
1200
1
lOOO
'00 'oo'O0 0O
WET DEPOSITION 1992/1993 CALCIUM kg/ha
,
l
l
,
,
,
,
,
I
,
G
S
N
A
W
L
AR
H
M
R
[~
1992
~
1993
t
i
G Z OH
- .t,-- R R - 9 2
T
,
,
O
UE
C
0
mm
12['
600 f =e,os
lO-I
400
,~,
200
"'~ - R R - 9 3
000
Figure 2
-.-.: 2 "
O0
2 } "
.
O0
O0 O0 o
i
,
~
f
,
,
t
,
j
,
G
S
N
A
W
L
AR
H
M
R
i
1992
~
1993
,
,
G Z CH
- ~ - RR-92
,
i
,
O
UE
C
- ~ - RR-93
Figure 3 The wet deposition rate of calcium decreased from 1992 to 1993 (except G) and of the other alkaline components also, while the wet deposition of sulfateS in industrial areas has decreased too (Fig. 2 and 3). No significant change of components nitrate-N, ammonium-N, and the acidity could be observed. Wet deposition trends of air pollutants can be seen by means of long-term data. The main reason for the decrease of sulfate-S in the precipitation is the reduction of SO 2 emission in consequence of the collapse of industry in 1 9 9 0 and 1991 (Fig. 4). The calcium deposition was also reduced since 1991 because of the removal of dust emissions by fly ash filter in power plants (Fig. 5). WET DEPOSITION 1983-1993 SULFATE- S
WET DEPOSITION 1983-1993 CALCIUM
kg/ha
18t 16
kg/ha
iiii, s
14 7
12~
I
,ira! 83
84
85
86
87
88
89
YEAR L~S
~ G
~ 90
, 91
, 92
9'3 o
1 83
i 84
w 86
~ 86
~ W
~ 87
, 88
, 89
YEAR
Figure 4
i S
Figure 5
~ G
CZW
I 90
i 91
~ 92
i 93
406 A change of nitrate-N rate in precipitation could not be observed. The decline of industrial NO x emissions and the increase of road traffic emissions are probably balancing (Fig. 6). The deposition rates of ammonium-N and acidity (Fig. 7) have been remained unchanged since 1988. An expected decrease of anthropogenic ammonium-N impact may be compensated by long-range transport. WET DEPOSITION 1983-1993 NITRATE - N kg/ha e~
83
84
85
86
m~s
87
88 89 YEAR I~G
90
91
92
93
~---lw
Figure 6
WET DEPOSITION 1983-1993 ACIDITY kg/ha o,6
1I
O,4 J
0,2 0,1
o
i
83
84
85
86
87
88
89
90
91
92
93
YEAR Es
~G
'--]w
Figure 7 Based on the decreasing concentration of alkaline components in the precipitation, the stagnation of sulfate-S since 1991, and the assumed increasing of nitrate-N in 1994 an increase of acidity in the precipitation will expected in 1994.
4. REFERENCES Marquardt,W. , Ihle,P. , and Kappe,W. , Chemische Technik, 38 (1986), 262 Br0ggemann,E. , Gnauk,T. , and Renner, E. , Umweltwissenschaften und Schadstoff-Forschung (1993 accepted) Marquardt,W. , Ihle,P. , and Br0ggemann,E. , Atm. Env. a (1993 submitted TopoI,L.E. and Ozdemir,S., U.S. Env.Prot.Agency, Contract-No. 68-0241 25 (1986)
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
407
Eight years studing bulk and wet deposition in the Spanish Basque Country D. Encinas and H. Casado Univ. del Pals Vasco. Dpto. Ffsica Aplicada II. Ftad. Farmacia. 01006 VITORIA (SPAIN)
Abstract Chemical characteristics of bulk and wet deposition collected on six sites in the Basque Country (Spain) from 1986 to 1993 are discussed. Wet deposition (kg.ha-l.y 1) is included between: CI = 118.5 and 5.0, NO3- = 27.7 and 7.4, SO42- = 60.6 and 14.0, NIL + = 18.0 and 3.7 and Ca2+ = 23.3 and 5.1. Mean rainwater pH fluctuates strongly from a site to another between 5.3 and 4.4. Maximum CI/Na ÷ ratio has a value of 3 denoting the influence of a waste incinerator. Wet deposition variation throughout precipitation process presents different trends for H ÷ and for the remaining ions. Bulk deposition is clearly dominated by SO42- with 51.3 kg.ha-l.y -1.
1. INTRODUCTION As a part of the EPOCA programme (Estudios en el Pirineo Occidental de la Contaminaci6n Acida) [ 1-3], the results between 1986 and 1993 concerning the wet deposition of six sites in the Basque Country are presented in this paper. Moreover, scavenging process by means a sequential collector is studied in four sites. Finally, the bulk deposition of a forested area in the Basque Country is presented.
2. EXPERIMENTAL
The Basque Country is located in the north of the Iberian Peninsula, next to the Cantabric Sea, between the Western Pyrenees and the Cantabrian region. It covers an area of 7261 km 2 and is an highly industrialized region but with half of its area covered with woods (Figure 1). Mundaka
North
T Cantabric Sea
, ~-
)~"
I
-Fra-""
!r, a s q u e ~ o u m r y _ . . - -
,,I,,,~,: .~, ,.,, ..,LT,/.I.,.." ....,
.t, - ~ /
,.7-
rJ~~i Spain /
t.----_X~
~,e
i
~...'
" :2o km ~
\
L~ Hoy,,, ......* Bulk arid wet \v/-~, ~
~_...__
Wet-only
Figure 1. Map of the Basque Country showing the location of the precipitation collectors.
408
3. RESULTS AND DISCUSSION 3.1. Average chemical characteristics of wet deposition Table 1 shows the spatial distribution of the ionic wet deposition on each site. Table 1 Average wet deposition recorded in the rainwater collecting sites. Units: kg.ha-l.y-1. Site
C1-
N O 3-
SO2- Na ÷ NH4÷ K ÷ Ca2÷ Mg2+
Vitoria 11.7 7.9 15.2 5.2 Igueldo 118.5 27.7 60.6 55.0 Mundaka 103.2 16.7 38.9 57.1 Salvatierra 24.8 12.8 23.9 10.0 La Hoya 5.0 7.4 14.0 2.3 Olaeta 45.4 13.9 25.9 9.9
3.8 18.0 7.1 8.87 3.7 13.2
0.8 5.8 6.4 2.5 2.2 2.4
6.4 23.1 11.9 13.2 5.1 13.8
1.0 7.6 8.2 1.5 0.6 1.2
Molar CI-/Na + ratio varies between 1.1 and 1.5 at both coastal sites and at the southern sites. However, this ratio takes higher values at the central sites: 1.8 at Salvatierra and 3.0 at Olaeta. These results show the influence of a waste incinerator. The incinerator is located 11 km north-easterly from Olaeta, which affects the ionic composition and the pH of the precipitation at the nearest sites, especially at Olaeta.
3.2. Rain acidity Table 2 gives the volume weighted mean concentration of H ÷ for each site. The corresponding average pH is also included in it. Table 2 Average H ÷ concentration and pH in the precipitation collected at each site. Units: #eq.1-1. Site
Vitoria
Igueldo
Mundaka
Salvatierra
La Hoya
Olaeta
H ÷
10.9 5.0
12.8 4.9
18.3 4.7
5.4 5.3
8.5 5.1
41.0 4.4
pH
In general, these pH values are higher than those found in different areas of central Europe [4,5] which have similar concentrations of acid anions, SO42- and NO3-. This is due to the importance of the CO3Ca, which acts as a neutralizing element at the study area [3]. The precipitation of Olaeta has very acid characteristics which are typical of very industrialized areas. This can be explained by the influence of the mentioned waste incinerator.
3.3. Variation of wet deposition with meteorological class Four types of back-trajectories have been distinguished, as shown in Figure 2: - Local trajectories: They do not exceed 200 kin.- Iberian trajectories" They come from the Peninsula inland. - Marine trajectories: They come from the Cantabric Sea. - Continental trajectories: They come from the Europe. Figure 3 shows the mean wet deposition of the different species at each site, according to back-trajectories.
409
MARINE ~
CONTINENTAL
Figure 2. Examples of the four sectors types.
l Mg'* Na' CI
"E] 4OO
ii
¢D "lO
m
V~a ~
M u ~ ~erra ~ ~ya ~a~
~
~d0rmIguel~M~ka ~
mmzd~
~ ~a ~
Figure 3. Average wet deposition collected at each site for the different trajectories types. Anthropogenic ions, NO3-, 5042- and NH4 + show maximum deposition values in Local and/or Continental trajectories depending on the site. This is probably due to the accumulation of contaminants in the case of Local trajectories and to a long-range transport phenomenon in the case of Continental ones. H ÷ shows maximum deposition values in Continental trajectories at all study sites. It also shows an high deposition value in Local trajectories at Olaeta. This can be easily explainable if we take into account that the acid anion (NO 3- and SO42- at all sites and CI- at Olaeta) have the maximum deposition level in this trajectories type. Ca2÷ has higher wet deposition levels in Local and/or Iberian trajectories, to which would indicate the peninsular origin of this species. This would also be the reason for the lower acidity levels of these trajectories type.
3.4. Scavenging process: Evolution of wet deposition during a continuous precipitation event Table 3 gives the wet deposition percentage of each ion in each precipitation fraction. All ionic species experience an important washing process, as is proved in the reduction of wet deposition percentage along the precipitation event. This process is more important at Igueldo and Olaeta. The anthropogenic-crustal ions experience a greater washing process than the marine ones. The wet H ÷ deposition percentage increases along the rainy event at Vitoria and Salvatierra. This is probably due to the decrease of Ca 2+ and consecuently the neutralizing power of precipitation. However, H ÷ behaviour at Igueldo and Olaeta is similar to the remaining ions, probably due to these sites having a continuous source near sampling points.
410
Table 3 W e t deposition percentage in the 12 first m m of precipitation. Units" %kg.ha -1. CI" Vitoria 2 4 6 8 10 12 Igueldo 2 4 6 8 10 12 Salvatierra 2 4 6 8 10 12 Olaeta4 6 8 10 12
NO3- SO42"
23.9 15.8 15.8 15.5 14.7 9.9 28.4 12.5 11.3 15.8 9.5 8.2 24.1 16.1 12.0 11.3 12.3 12.0 42.4 13.8 14.6 16.5 11.3
27.8 22.7 13.2 13.8 13.7 8.9 37.2 13.7 10.5 12.5 8.3 9.0 23.1 15.1 12.4 12.4 12.6 13.6 36.6 19.7 14.5 15.5 12.6
27.9 21.2 13.3 13.3 13.0 10.2 32.4 14.0 12.5 14.4 9.6 7.0 23.5 14.3 13.2 11.0 12.4 14.0 31.6 20.2 15.9 16.8 14.2
H+
Na +
NH4+
K+
Ca2÷
Mg2+
9.6 16.0 15.9 17.3 21.4 20.6 23.2 14.5 14.3 15.9 10.7 9.1 7.9 11.8 11.1 13.9 15.3 21.8 31.0 14.1 17.7 16.1 19.9
27.2 18.2 11.4 12.9 15.9 7.3 28.6 12.5 12.4 15.8 8.5 8.6 15.8 14.4 12.3 13.5 13.4 13.6 34.1 18.8 14.1 18.5 12.6
31.3 22.3 11.8 13.6 14.1 7.8 39.3 11.4 9.6 12.3 9.2 8.2 29.8 16.0 12.1 9.7 12.5 12.2 37.3 17.4 16.0 15.7 12.0
24.7 20.5 12.0 16.9 13.1 11.2 30.8 9.5 13.7 14.4 8.0 10.2 17.1 15.8 13.6 13.2 13.6 13.8 36.4 18.9 18.5 15.4 10.2
34.1 17.7 13.0 11.9 11.9 9.9 29.6 15.1 11.7 13.4 9.8 9.3 26.9 16.7 12.3 11.1 11.4 11.7 42.1 17.0 15.3 14.1 10.4
28.1 19.2 14.7 13.5 12.6 9.6 26.0 13.2 12.1 16.5 9.7 8.9 32.8 14.0 12.2 8.0 11.3 8.4 37.3 18.5 13.0 17.3 13.8
3 . 5 . B u l k deposition Bulk deposition has been collected at one-monthly intervals at Olaeta, between Jun 88 and May 91. Average chemical characteristics of the 36 samples can be seen in Table 4. Dry deposition, calculated as the difference between bulk and wet deposition, are also shown in Table 4. Table 4 Average characteristics of bulk and dry deposition collected at Olaeta. Units" kg.ha-l.y -1. Ion
CI
Bulk Dry
31.4 negative
NO3- SO42-
Na +
NIL +
K+
C a 2+
Mg2+
24.6 10.9
21.7 11.0
34.7 21.6
11.1 8.4
36.3 22.5
2.9 1.6
51.3 25.2
CI- undergoes reactions in bulk deposition and later evaporations, probably in the form of HCI or NH4CI, or both at the same time.
4. REFERENCES 1 H. Casado and D. Encinas, Geoffsica, 48 (1992) 99. 2 D. Encinas, H. Casado, J.P. Lacaux and Pham Van Dinh, Environ. Sci. Health, A29 1 (1994) 99. 3 H. Casado, D. Encinas and J.P. Lacaux, Atm. Env., 26A 6 (1992) 1175. 4 R.D. Saylor, K . M . Butt and L.K. Peters, Atm. Env., 26A 6 (1992) 1147. 5 A . F . M . Ahmed, R.P. Singh and A.H. Elmubarak, Atm. Env., 24A 12 (1990) 2927.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
411
SCAVENGING OF GASES DURING GROWTH OF ICE CRYSTALS G.Santachiara, F.Prodi and F.VivareUi Istituto FISBAT-CNR, Bologna, Italy
ABSTRACT Scavenging of HC1 and SO2 during snow-crystal growth by water vapour from supercooled crystals tested
liquid
was
droplets
independent
is
studied.
of the
1.2-7.4 ppm range.
The
mean
SO2 gaseous
sulfur
phase
(as
sulfate
concentration
) in
in the
The results with HC1 indicate that for a C1 -
concentration in the liquid phase of less than 20 ppm, all the HC1 goes from
the
droplets
into
ice
and
enhance
knowledge
of
the
acidification
precipitation.
1.EXPERIMENTAL METHODS AND RESULTS The
chemical
interactions particles
composition
of numerous
and
gases
of
precipitation
processes
within
and
, e.g.
below
is
determined
scavenging processes
clouds,
chemical
by of
the
aerosol
transformation
in
liquid or gas phase, mechanism of precipitation formation and so forth. As in many climates precipitation forms by processes that include the ice-phase, it is important to study scavenging processes of gases during snow crystal growth by sublimation of water vapour or by accretion of cloud droplets (riming) .These processes influence precipitation pH. As
to
riming,
previous
laboratory
experiments
values of the SO 2 retention coefficient configuration
(Lamb
water-soluble
gases
et
al.,
(HCL,
1987; HNO 3,
(F)
Iribarne H202)
have
measured
different
, depending on the operating et
,
al.,
1990
Iribarne
and
).
With
highly
Pyshnov
(1990)
measured F as equal to about unity. Experiments
on
scavenging
of
gases
during
crystal
growth
from
water
vapour have mainly concerned SO2 and HNO 3. Hydrogen chloride also plays an important role in the acidification of precipitation. In laboratory experiments
we
studied
the
sorption
of
SO2 and
HC1 during
growth
of
crystals by water vapour. Droplets produced by spraying Super-Q water were
412 introduced at the bottom of a cylindrical chamber housed in a cold room (T=-13
o C). Crystals grow via vapour diffusion from supercooled droplets in
the presence
of SO2 or HC1. Collected ice crystal were
melting
ionic
for
species
(SO 4
analyzed
after
by ion chromatography. SO2 concentration in the gas phase was in the 1.2-7.4 ppmv range. The mean crystal
concentration
of sulfur
'
( as
C1-)
sulfate)
was independent
of the
SO
2
gaseous concentration. HC1 test data indicate that for a C1- concentration in the liquid phase of less than 20 ppm, HC1 goes entirely into the ice phase. This finding can be explained
by assuming the
existence
of a liquid-like
surface
layer
during
crystal growth in which HC1 and SO2 dissolve and ionize. The
existence
of a liquid-layer
at
T<
0°C
during
crystal
growth
confirmed by further experiment. Finnegan and Pitter (1988) and et
al.
(1991)
demonstrated
that
differential
ion
incorporation
is
Finnegan during
the
growth of ice crystals from supercooled water containing halide or ammonium salts
leads
to
charge
separation
in
the
ice
that
affects
the
ice
crystal
shape. In addition, both oxidation and reduction reactions occur in growing ice crystals at T=-16 o C. Both these phenomenon require the existence at the growing
interface
of
a
liquid
layer
which
is
sufficient
to
dissolve
and
ionize the included ionic salt. Given our data and the fact that the HC1 retention coefficient during riming is about one, there is an almost complete transfer of HC1 from liquid to solid phase, whether crystal growth occurs via vapour diffusion from supercooled droplets containing dissolved HC1 or by riming. This result has important consequences in the acidification of precipitation. 2. REFERENCES 1.D.Lamb and R. Blumenstein, Atmos.Env. 21 (1987) 1765 2 J.V.Iribarne, T.Pyshnov and B.Naik, Atmos.Env. 24A (1990) 389 3 J.V.Iribarne and T.Pyshnov, Atmos.Env. 24A (1990) 383 4 W.G.Finnegan and R.L. Pitter, Atmos.Res., 22 (1988) 235 5 W.G.Finnegan, R.L.Pitter and L.G.Young, Atmos.Env., 25A (1991) 2531
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
Contribution of Root-derived Sulphur Throughfall in a Douglas Fir Forest
to
Sulphate
413
in
A.C. Veltkamp, M. Geusebroek, G.P. Wyers Netherlands Energy Research Foundation (ECN) P.O. Box 1, 1755 ZG, Petten, The Netherlands
Abstract
The relative contribution of root-derived sulphur to sulphate in throughfall is determined for a Douglas Fir forest using 35S-radiotracer. The experiment took place at Speuld, The Netherlands, from July 1993 until April 1994. It is shown that the relative contribution of root-derived sulphur contributes to throughfall of 3% on average, with slightly higher values (8-12%) during autumn. Leaching is higher for an experimental plot receiving a double annual input of ammoniumnitrate and ammoniumsulphate on the forest floor compared to a reference plot. The calculated contribution is probably overestimated by a factor of about 2. This is subject to further study.
1. INTRODUCTION Analysis of throughfall and stemflow can be used effectively for the estimation of atmospheric deposition to a forest if certain conditions are met. Wash-off of components deposited on the canopy via dry- and occult deposition should be quantitative. Furthermore, the leaching of root-derived plant components should be small relative to wash-off of deposition. In this report, results are given from a 35S radiotracer experiment in a Douglas Fir forest. The experiment took place at Speuld, The Netherlands, from July 1993 until April 1994. The aim of the study is to quantify foliar leaching of root-derived sulphur to throughfall. The study was initiated in order to find a possible explanation for the observed difference for several locations in The Netherlands between sulphate deposition estimated via inference and throughfall measurements, respectively. Additionally, the influence of acid deposition on foliar leaching of sulphate was studied. The relative contribution of root-derived sulphur to throughfall is given by the ratio of the 35S specific radioactivity (i.e., Bq/mg sulphate) of throughfall sulphate and the leachable sulphur pool in the canopy. One of the major uncertainties in this isotopic study therefore arises from the radioactivity-determination of the leachable sulphur pool if isotopic equilibrium is not obtained for all sulphur pools. It is further assumed that leaching occurs only via foliage, neglecting leaching from twigs.
414 2. EXPERIMENTAL 35S-labeled sulphate is applied to the forest floor of two selected plots at Speuld, The Netherlands, at unfrequent intervals as aqueous solutions using a high-density drainage tubing system. The first period of radiotracer application started on June 29th (1993) and ended on November 16th (1993, 8 events). The second period started on March 10th (1994) and ended in October (1994). In this report, only results obtained between June 1993 and April 1994 are discussed. For plot 1, the 35S-sulphate radiotracer is dissolved in artificial rain water. For plot 2, the tracer is added to a concentrated solution of ammoniumsulphate and ammoniumnitrate. In this way, the annual input of ammonium, nitrate and sulphate to the soil of plot 2 is doubled as compared to the estimated atmospheric input in 1989. Throughfall and rainwater are collected from the site at 1-4 week intervals, depending on season and rain events. Foliage is collected from the canopy at approximately monthly intervals starting on July 20th (1993). Throughfall composition is determined using ion chromatographic methods. The 35S specific radioactivity is determined in needles (total sulphur), throughfall (as sulphate) and water-extractable sulphate from the foliage using liquid scintillation counting techniques. Detailed description of all methods will be published elsewhere.
RESULTS AND DISCUSSION As explained in the Introduction, the relative contribution of foliar leaching of sulphate to sulphate in total throughfall is given by the ratio of the 35S specific activity of sulphate in throughfall and the water-leachable sulphur from the canopy at any time. The results have been plotted in Fig. 1. The contribution is approximately 3% on average for both plot 1 and 2. Only during one occasion (November 1993) the relative contribution was larger than 10%. The contribution of root-derived sulphur to sulphate in throughfall is therefore relatively small. The result are in accordance with results obtained at other sites for different types of vegetation (Garten 1988, Cape 1993). Despite the large uncertainties in the specific radioactivity of the sulphur pools, results for plot 1 and 2 are highly correlated. The differences between plot 1 and 2 observed for certain periods are therefore realistic. Foliar sulphate leaching from the canopy was relatively large during the autumn of 1993. During this period, sulphur leaching is larger for plot 2. The question remains if the underlying assumptions for calculating the contribution of foliar leaching to throughall are valid. It is assumed that by soaking freshly collected needles for 24 h in water, the 'leachable' portion of root-derived sulphur from the needles is obtained. Most likely, the water extract will also contain sulphate from dry deposited sulphate aerosol and sulphur dioxide. This means that the root-derived 35S labeled 'leachable' sulphate is diluted with unlabeled, dry deposited sulphur. Under these circumstances, the specific activity of sulphate measured in the extract is not representative for the rootderived leachable sulphate. This effect will increase if sampling of needles is preceeded by long dry periods. The relative contribution of foliar leaching to sulphate in bulk-throughfall given in Fig. 1 should therefore be regarded as upper limit at any time for leaching of root-derived sulphur.
415 REFERENCES J.N. Cape, 1993. The Use of 35S to Study Sulphur Cycling in Forests, Envir. Geochem and Health, 15 (2/3), 113-118. C.T. Garten, 1988. Fate and Distribution of Sulphur-35 in yellow poplar and red maple trees. Oecologia, 76, 43-50.
plot 1
plot 2
15
12 tO
°.,..,
91 .m L_
tO
o
6
/
"
L
3
\
50
100
150
200
250
300
days after start 35 S labeling (29-6-93)
Fig. 1 Relative contribution of Douglas Fir foliar leaching to sulphate in throughfall calculated from 35S specific activity of water-leachable sulphate and throughfall-sulphate measured at two selected plots at Speuld, The Netherlands.
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BE All rights reserved.
417
Fate of nitrogen in Spruce and Pine Ecosystems T. Staszewski, S. Godzik and J. Szdzuj
Institute for Ecology of Industrial Areas, Kossutha 6, 40-833 Katowice, Poland
Abstract Investigation at two permanent forest research plots - spruce - at Brenna (The Beskidy Mountains) and - pine - in Katowice (Upper Silesian Industrial Region) was carried out. The plots were equipped with collectors for bulk deposition, throughfall, and vacuum cup lysimeters (25 and 50 cm depth). The mean year concentration and annual load of NO 3 and NH 4 was calculated. The data obtained by measurements were compared with amounts suggested as critical loads for forest soil.
1. METHODS APPLIED Collectors for bulk deposition and throughfall, and vacuum cup lysimeters were the same as those used in studies of Integrated Monitoring in Northern Countries. Samples were taken at monthly intervals, and analyzed for nitrates by Ion Chromatography. The concentration of NH 4 was determined using the Nessler method. Loads were calculated as kg/ha/year, based on concentration and volume of deposited rain or snow collected as bulk deposition and as throughfall.
2. RESULTS AND CONCLUSION
Data characterizing both deposition and soil solution are given in Table 1. At both locations, bulk deposition, throughfall and soil solution from both depths are acidic. As compared to bulk deposition, pH decrease by 0.7 units in Brenna, but a slight increase is observed in Katowice. An increase of pH, on both plots has been found in soil solutions. These differences in pH may be explained by soil properties (Godzik et al 1994). Annual mean concentrations of NO 3 and NH 4 in bulk deposition between Brenna and Katowice differ markedly. The increase of both NO 3 and NH 4 in throughfall in spruce stand has been found. The concentration of N O 3 in soil solution (25 cm) increases for both locations, but concentration of NH 4 is 1.6 times lower in soil solution for the Brenna site and 1.14 times higher for the Katowice site, when compared to concentration in throughfall (Table 1). The loads of nitrogen for both sites exceed the value suggested as critical values (Convention
418 on Long Range Transboundary Air Pollution). Annual loads calculated from bulk deposition were (kg/ha): 25.0 and 16.5 for Brenna, and 30.0 and 13.0 for Katowice, of NO 3 and NH4, respectively. Similar values of loads, in spite of significant differences in concentration, are due to amounts of precipitations: Brenna (1300 mm) and Katowice (700 mm). Based on the throughfall data, the following annual loads were found (kg/ha): 51.5 and 22.0 for Brenna, and 24.0 and 13.0 for Katowice, of NO 3 and NH 4. A significant decrease in NO3, and a smaller decrease in NH 4 concentrations in soil solution from 50 cm depth in Katowice were found. Soil at the Brenna site has higher concentration of nitrogen than the soil of the Katowice site amounting to 1.8 % and 1.0 %, respectively [1]. The data indicate that the nitrogen supply from air pollutants is higher than the stands can use. The Katowice pine stand seems to be less saturated with nitrogen as compared to the spruce stand in Brenna. These data are from a one year long investigation from which no firm conclusion can be made - at this time. However: 1. Loads of nitrogen to both ecosystems exceed the suggested critical values, and 2. The amount of nitrogen from air pollution, deposited to the forest ecosystems studied is higher than required by ecosystems. Table 1 pH and concentration [ppm] of nitrogen compounds in collected waters. Katowice
Bulk deposition Troughfall Soil solution 25 cm Soil solution 50 cm
Brenna
pH
Concentration NO 3 NH4
pH
Concentration NO3 NH4
3.94 3.99 4.09 4.03
6.72 5.60 6.97 2.95
4.4 3.67 4.22 4.39
2.38 6.89 8.49 8.53
2.95 2.99 3.4 2.72
1.68 2.59 1.6 1.52
3. REFERENCES S. Godzik, W. Lukasik, P. Poborski, T. Staszewski, J. Szdzuj, B. Andrzejaczek, Oddziatywanie i obieg zwiazk6w siarki i azotu zawartych w powietrzu i opadach w ekosystemach le~nych- badania w gradiencie st~zefi i klimatu. Annual Report of Institute for Ecology of Industrail Ar
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
419
Acid deposition" D a t a from the S w i s s Alps Sabine Braun and Walter Fliickiger Institute for Applied Plant Biology, Sandgrubenstr. 25, CH-4124 Sch5nenbuch, Switzerland
Abstract Deposition of nitrogen and sulfur was measured in spruce stands at different altitudes in Switzerland. Both NH4+-N and NO.~-N in throughfall showed a clear altitude dependence whereas SOa2 was not correlated with altitude. Critical loads for nitrogen and acidity were calculated for six sites in Central Switzerland.
1. INTRODUCTION The topography and the geology of Switzerland are very heterogeneous. Therefore, the application of the critical loads concept asks for knowledge about the influence of altitude and geology. Critical loads for acidity and nitrogen were calculated in a case study on sites in an area with sensitive geology and compared with actual loads measured by means of throughfall deposition. Some factors of these sites are listed in Table 1. Table 1: Sites used for the critical load study (canton of Uri, main valley) Site
Abbreviation
Silenen Frentschenberg Waldiberg Riggstfifeli RStiboden Eggberge
SIL FRB WAB RIG ROE EGG
Altitude
Geology
Age (years)
Stock Stems (m s ha -1) ha -1
630 910
Riss-Wtirm Moraine Riss-Wtirm Moraine
150 230
578 460
628 558
1220 1740 1580 1540
Riss-Wiirm Moraine Riss-Wiirm Moraine Granite Flysch
200 210 300 180
274 794 327 837
467 633 233 633
(m)
420 2. R E S U L T S
Deposition Deposition was measured by collecting throughfall in spruce stands and rainwater in adjacent clearings or in open field. No attempt was made to correct the N loads for direct uptake by the trees although this was noticeable especially for NHA÷-N in the more elevated sites where NH4÷-concentrations in throughfall were often lower than those in rainwater. Also, there is no information about stomatal uptake of NH3 or NO2. Thus, the values reported here are minimum estimates. Throughfall loads of N were mainly correlated with altitude (Fig. 1). This holds true for NH4÷-N as well as for NO3--N. The altitude dependence was much more important than regional differences. In contrast, the loads in rainwater showed increased values at an altitude of 1000-1200 m, but no other altitude dependence. This is the altitude with the highest cloud frequency. The deposition of SO4 ~-- in throughfall was not correlated with altitude. In the open field a similar increase of SO42 at 1000-1200 in was observed as in the case of NH4 ÷ and NO~. 15
15
i
,
, NOs-N
NH,-N
-
10
10
-
•
•
e
g .1=
m J=
\
\
Z
Z
•
00
.x
-
5
5
•
• 0
•
-
••
o
o
o
o 5O0
1000
1500
Altitude (m)
2000
500
1000
1500
2000
A l t i t u d e (m)
Fig. 1: Throughfall loads of NH4+-N (left) and NOa--N (right) during the summer (6 months)
Actual loads for acidity The potential acidity was calculated as the sum of NH4 +, NOa-, Sea 2 and CI minus the base cations Ca ++, Mg ++, K + and Na +. The dry deposition of Ca ++, Mg ++, K + which enters this calculation was derived from dry deposition on inert surfaces and throughfall loads according to Lindberg et al. (1988) for the sites in the case study or taken from model calculations by Rihm (1994) for the other sites. All other ions were extrapolated from throughfall summer loads to annual loads basing on whole year measurements of open field bulk deposition.
421 The actual loads for potential acidity are dominated by NH4 ÷ and NO:~ which causes a similar altitude dependence as for the two nitrogen compounds. Of the N and S compounds, NO3- contributed 37%, NH4 ÷ 28% and SO4 ~-- 35% to the acid load on an average.
Critical loads for acidity Weathering rates of single cations were calculated with the soil model PROFILE Vers. 3.2 (Warhringe and Sverdrup (1992). These weathering rates were entered into the steady-state mass balance (SSMB) equations for critical loads in the form developed for alpine areas (UBA 1993). Critical loads based on both vegetation sensitivity (Tab. 2, CL) and soil stability (Tab. 2, CL*) were calculated. Because of a rather high calculated Al-weathering, CL* is lower in 4 out of 6 sites than CL and causes quite low critical loads. Table 2: Weathering rate in the rooting zone (ANCw), critical load for acidity (CL and CL*; see text) and deposition of potential acidity (DepAc) at different sites (abbreviations of site names see Tab. 1). Numbers are in keq ha -~ a -1. Site
SIL FRB WAB RIG ROE EGG
ANCw CL
0.63 0.35 0.98 1.49 0.09 0.13
actual harvest 2.12 1.58 2.85 3.85 0.78 0.77
CL
CL*
CL*
max. harvest 1.44 1.17 2.57 3.48 0.63 0.77
rooting zone 1.90 1.05 2.94 4.48 0.28 0.38
whole profile 14.7 1.05 3.47 4.91 1.14 0.38
DepAc Exceedance
2.25 1.73 1.50 0.88 1.35 1.08
0.35 0.68 0 0 1.07 0.70
Critical loads for nitrogen Critical loads for nitrogen based on empirical data are given by Bobbink et al. (1992). For acidic coniferous forests, they report values of 15-20 kg N*ha-l*a -1 although these values only hold true for managed forests. In the alps, forest m a n a g e m e n t is quite extensive because of slow growth and bad site accessability. The harvest rates in the sites were mainly between 0.6 and 1 m 3ha -~ a -1 (Tab. 3). Thus, the amount of N taken up is much less important than the question where the nitrogen is going in the ecosystem and how much can be tolerated. With the present data, this question cannot be answered. At least in one site, a high C/N ratio and the h u m u s form (raw humus) suggests that no nitrification takes place, which will lead to an accumulation of NH4+-N. The critical loads for nitrogen were also considered by nutrient balance calculations according to Gundersen (1992). First, the N loads in deposition can be compared with short and long term uptake. Nitrogen deposition exceeded the current immobilization rate in woody parts by a factor of 2-5. Second, N uptake must be
422 balanced by the supply of other nutrients in weathering or deposition in order to maintain optimal nutrition. Thus, the critical load of N can be seen in view of optimal nutrition (nutrient balance approach). This method suggested that P and /or K was limiting at the sites under examination. It is interesting to note that needle analysis indeed revealed a quite short supply of P, at one site also of K. Table 3: Comparison of N-deposition, N uptake (immobilization in woody parts) and results of the nutrient balance calculations (maximum amount of N which can be balanced by weathering/deposition input of other elements). Site Harvest rate (m 3 ha "1 a l) Stem increment (m 3 ha 1 a 1) Deposition of N (kg N ha -1 a -~ ) Immobilization (kg N ha 1 a -1) N in harvest (kg N ha -1 a l) Nutrient balance (kg N ha ~ a -i) Limiting element
SIL 0.7 6.0 32 7.3 0.25 1.6 P
FRB 0.6 5.7 26 5.8 0.18 2.3 P
WAB 0.6 3.9 21 4.1 0.19 4.2 P
RIG 1.0 6.2 11 5.7 0.32 14.4 P,K
ROE 0.8 3.1 16 3.4 0.23 0.3 P
EGG 5.1 5.3 17 4.8 1.45 5.1 P
3. C O N C L U S I O N S The case study suggests that acid loads are of ecological relevance in some parts of Switzerland. It has to be noticed that the examinated sites are forests which protect settlements and main traffic lines. The effect of N deposition on alpine forests deserves further attention.
4. A C K N O W L E D G E M E N T The project was supported by the Federal Office for Environment, Forest and Landscape. We would like to t h a n k the supporters, Dr. R. Volz and B. Achermann, as well as the local farmers and foresters who carried out the sampling. 5. R E F E R E N C E S • Bobbink, R., D. Boxman, E. Fremstad, G. Heft, A. Houdijk and J. Roelofs (1992). Nord 1992:41, pp. 111-160 • Gundersen P. 1992. Nord 1992:41, pp. 55-110 • Lindberg, S.E., G.M. Lovett, D.A. Schaefer, M. Bredemeier (1988). J. Aerosol Sci. 19 (7), 1187-1190 • Rihm, B. (1994). Status report on mapping of critical loads of acidity for Swiss forest softs and alpine l a k e s - steady state mass balance method. Bern, 83 pp. • UBA (Umweltbundesamt) (1993). UBA-Report 93-083, Wien • Warfvinge, P. und H. Sverdrup (1992). Water, Air, Soft Poll. 63, 119-143
G.J. Heij and J. W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
423
Dry deposition to bulk samplers underneath a roof In a spruce (picea abies KarsL) forest Michael Bredemeier and Martin Meyer Forest Ecosystems Res. Ctr., Univ. of Goettingen, Habichtsweg 55, D-37075 Goettingen, FRG
Abstract Standard bulk precipitation samplers of the funnel+bottle type were exposed underneath a roof facility in a spruce forest at Soiling, Germany. They contained 200 ml of distilled water adjusted to a series of pH steps and were placed for 1, 2, and 4 weeks, respectively. Considerable enrichment rates of chemical elements, particularly nitrate-N, ammonium-N and sulfate-S were observed, which must be attributed to dry deposition.
1. The experiment In the Soiling mountain area of central Germany a roof manipulation study is conducted in 60 y old norway spruce (picea abies [L.] KARST.) stand. One of the three roofs of 300 m '~ area each is currently used for drought experiments. During an extensive experimental drought phase in the summer of 1992 bulk samplers with polyethylene funnels and bottles, of the same type as regularly used for throughfaU sampling, were exposed underneath that roof. Time periods of exposure were 1, 2, and 4 weeks, respectively. The bottles contained 200 ml of standard solution with pH 3, 4, 5, (H2Odest.), 8 and 9. Each pH step was in 3 replicates. During the drought phase the sprinkler systems underneath the roof do not operate. The roof cover prevents rain or canopy drip from entering the bulk samplers. Therefore, any enrichment of chemical constituents in the samplers after exposure can be attributed to either dry deposition or organic contamination (by insects, bird droppings, etc.). The pH steps were employed in the experiment to test the dependence of dry deposition on the pH of the exposed sample solution.
2. Results The results show appreciable enrichment rates of chemical constituents, particularly of nitrate-N and ammonium-N, in the exposed bulk samplers (figs. 1, 2). There is no indication that this input is due to an organic contamination by insects or birds. The enrichment is independent of the pH of the exposed sample. In the case of NH4-N it seems to be related to the time of exposure (fig. 2), but not in the case of NO3-N, where the highest concentrations gained were found in the second exposure period lastin-g only two weeks (fig. 1). The enrichment is most probably caused by high deposition rates of particles and gaseous compounds.
3. Discussion It is quite surprising how high the final element concentrations in the samplers exposed underneath the roof were. They were actually in the same range as those normally observed in monthly samples from the same type of bulk samplers exposed outside the roofs (figs. 1, 2, lower parts). The enrichment of elements seems to be independent from the pH of the exposed solution. This provides evidence that the nature of the dry deposited material is particulate. The depo-
424
sition velocities calculated on the basis of the observed enrichment (table 1) are much higher, however, than particle deposition velocities for these constituents reported in the literature. 4. Conclusions - dry deposition can make up a considerable part of the element input to bulk samplers - no pH-dependence of the enrichment of NO3-N, NH4-N and SO4-S in the exposed samples was observed - calculation of deposition velocities for NO3-N and NH4-N from their observed enrichment in this experiment leads to very (unreahstically?) high values.
Fig. 1" Enrichment of nitrate-N in solutions exposed in bulk samplers underneath a roof
exposed in bulk samplers underneath a roof
nitrate-N [mg/I]
i
Fig. 2: Enrichment of ammonlum-N in solutions ammonium-N [mg/I] 5
-
-
2
3
4 5 d.w. iniUal pH of solution
8
2
9
3
B
B I~ I:~r~d~2nd ~d(xt~31d p~cx! 1 week ~ 2
weeks ~
weeks
Nitrate-N concentrations measured in throughfall
4 5 d.w k~t~ pH of sdutk~n
8
lst periodl~2nd p e r i o d ~ r d period 1 week i ~ 2 w e e k s ~ w e e k s
Ammonlum-N cone. measured in throughfall outside roof (monthly values for the year 1991)
outside roof (monthly values for the year 1991)
ammonium-N [m~l] 12
nitrate-N [rag/I]
10
. . . . .
8
2
3
4
5
6
7
8
9
10
11
12
. . . . .
6
.
4
.
2
.
0 1
9
1
.
.
.
.
.
.
2
3
.
.
.
. .
. 5
B
bulk TF
Table 1: Calculated deposition velocities [cm/s] for nitrate and ammoniumat varying pH nitrate-N 2rid sarnplir~lF~fiod
ammonium-N 3rd samplin~lperiod
pH2
10.1
1.7
pH3
20.6
7.6
pH 4
19.0
5.3
pH5
19.5
7.6
d.w.
3.8
3.9
pH8
14.6
1.1
pH9
10.1
6.1
.
.
. 4
month
B monlhly~nc.
.
.
.
.
.
.
.
.
.
. 6 7 month
8
monthly corm. buk TF
9
10
11
12
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BE All rights reserved.
425
Deposition of Air Pollutants to Forest Ecosystems Along Pollution and Climatic Gradients in Poland S. Godzik, T. Staszewski and J. Szdzuj
Institute for Ecology of Industrial Areas, Kossutha 6, 40-833 Katowice, Poland
Abstract Measurements of bulk deposition, throughfall and soil solution were taken at six permanent plots located in spruce and/or pine stands along pollution and climatic gradients. The investigation was accompanied by continuous measurements of SO2 and NO2 concentration in the air. Concentrations of sulphur and nitrogen in pine and spruce needles were determined. Health quality of forests stands was assessed. Annual loads of sulphur and nitrogen compounds reaching the forest ecosystems were calculated.
1. AIM OF THE STUDY AND METHODS APPLIED
The aim of this study is to determine the direct and indirect effects of air pollutants to forest ecosystems. Concentration of sulphur dioxide and nitrogen dioxide are measured. Bulk deposition, troughfall and soil solution are collected at six permanent forest plots. Both annual mean concentration and load of pollutants: sulphur and nitrogen were calculated. Permanent plots were established in spruce (Brenna - the Beskidy Mountains and SmoMzino the Baltic Sea shore) and pine stands (Katowice, WoZniki - Upper Silesia, Puszczykowo Central-West Poland, and Smotdzino). Pollution and climatic gradients are represented by the South- North locations. Methods recommended for Integrated Monitoring for Northern Countries and ICP Forests (under the Convention on Long Range Transboundary Air Pollution) have been used.
2. RESULTS AND CONCLUSIONS
The highest concentration of sulphur dioxide and nitrogen dioxide was found at the Katowice site, followed by Puszczykowo and Brenna. The lowest concentration of air pollutants was found at the Smoldzino site (Tab.l). The Polish standard for SO2 (annual mean - 32/~g m3), has been violated in Katowice only. Bulk deposition, throughfall and soil solution from all locations are acidic, ranging from 3.32 in soil solution in Smoidzino (spruce stand) to 4.79 in soil solution in Smotdzino (pine stand). In spruce stands, a decrease in the throughfall pH when compared to bulk precipitation was
426 Table 1 Concentration of SO2, N O 2 in the air and sulphates, nitrates and ammonium in rainfall and soil solution from different locations along a pollution and climatic gradients in Poland. Air [/zg/m3]
Brenna Katowice Wo2niki Puszczykowo Smoldzino
Bulk [ppm]
SO2
NO2
SO 4
NO3
NH4
25.59 55.73 34.34 4.60
7.23 4.45 2.38 23.10 8.25 6.72 6.17 3.20 9.00 13.70 5.00 4.60 4.85 5.88
1.68 2.95 1.75 2.80 2.68
Throughfall [ppm]
Soil solution [ppm]
504 14.80 20.05 14.88 25.20 7.77
NO3
NH4
SO4
6.89 5.60 5.26 6.50 4.67
2.59 2.93 1.53 3.70 1.27
17.90 58.00 142.00 90.00 27.70
NO3 8.50 3.20 0.50 1.00 1.80
NH4 1.50 2.50 1.70 1.80 2.50
found (Brenna, Smoldzino). The pH of throughfall from pine stands is increasing slightly (Puszczykowo, Katowice). A small decrease is observed for the Smotdzino site. No differences in pH of throughfall in Smotdzino spruce and pine stands have been found. It can be concluded, that differences between locations with spruce and pine are caused by the chemical character of deposited air pollutants. The concentrations of sulphates and nitrogen compounds in Brenna and Smotdzino do not differ markedly. They differ for soil solution (Tab. 1). No differences in load of sulphates, but large differences for nitrogen compounds have been found for these two location (Tab. 1). One possible explanation are differences in the amount of rainfall. The data for the pine stands seem to favour this explanation. The concentration of sulphur in spruce needles at Brenna is 0.18%, and in the Smotdzino needles 0.11% [1]. The highest concentration of sulphur in pine needles has been found in Katowice (0.16 %) followed by Puszczykowo (0.14 %), and Smo~dzino (0.11%) These concentrations are higher than accepted as "normal". The concentration of nitrogen are in the range described as normal ( 1.3 - 1.7% for spruce and 1.4- 1.8% for pine)[2-3]. Trees injury (crown transparency) for spruce stands at Brenna and the Smoldzino does not differ. For pine stands the most severe injury has been found for the Katowice stand, followed by Puszczykowo and Smotdzino [1]. This reflects the gradient in air pollution. In the authors opinion, lack or poor correlations between some parameters investigated may be caused, at least in part, by the short measurement time.
3. REFERENCES S. Godzik, W. Lukasik, P. Poborski, T. Staszewski, J. Szdzuj, B. Andrzejaczek, Oddziatywanie i obieg zwi~zk6w siarki i azotu zawartych w powietrzu i opadach w ekosystemach le~nych- badania w gradiencie st~zefi i klimatu. Annual Report of Institute for Ecology of Industrial Areas.(1994) Anonymus, Commission Advice Forest Fertilization (1990) J.N. Cape, P.H. Freer-Smith, I.S. Paterson, J.A. Parkinson, J. Wolfenden. Trees, 4 (1990) 211 - 224.
POSTERS SESSIONX DRY DEPOSITION / CONCENTRATIONS
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers? © 1995 Elsevier Science BV. All rights reserved.
429
Immission and Dry Deposition of SO2 and Ozone lee side of the Conurbation of Leipzig in Eastern Germany G. Spindler, W. Rolle
Institut Mr Tropospharenforschung, Permoserstr. 15, D-04303 Leipzig, Germany
ABSTRACT In 1991 the integrated research project SANA has been established to study the air pollution situation and its changing after the German unification in East Germany. In this project a measurement station was built at leeside of the strong air polluted city area of Leipzig. In the poster we presented results of the continously measurement of the immission of SO 2 and 0 3 by the gradient technique. The ratio of dry to wet deposition l) of total S is between 0,5 to 2,3 and depends on the quantity of precipitation.
EXPERIMENTAL AND RESULTS Figure 1 shows the regions with much SO2-Emission and the location of the measurement place Melpitz in Saxonia.
m.
~1~1¥1_/-~3
[[ ~
,~
1 ~,..,,.~ j / ~
~_..~.r
\N
Measurementplace Melpitz
| helghestS02 hnnlission
/ k. so ,pzil toelpi 4,
Figure 1:
Location of the measurement place near Melpitz
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1)
Detenninalion ofsulphaleSbyE.Brfiggemann(personalcommunication1994)
430 The measurement station is placed in an old large meadow which is well suited for micrometeorological experiments. The gradient method is based on measurements of meteorological parameters such as wind velocity, temperature and relative humidity in different heights. The experimental realization of the measurement of the profiles for the concentration of trace gases in Melpitz in different heights is shown in figure 2. 11,69
-~ main wind direction WSW
luarz
Fri~e
7,89 ~ "
height [m]
5,32
multiplexer
0,5011~13/3,59/7,89/zero air 0,89/2,42/5,32/11,6Slzero air 3,59
every hour 6 profils per hour tubes SO2
NOx
[
03
rotameter
air pump data
acquisition gas multiplexer
Figure 2:
Gradient system for chemical species
The fluxes are calculated from hourly mean values by logarithmical fits of the profiles from concentration and wind speed. Results are presented in Table 1 Table 1
Fluxes calculated by gradient technique
o
I10-6,19- 24(nigth) ii7< ay)
-0,02 -0,30 -0,07 -0,39
1 [
0,58
2)
-0,03 -0,08 -0,07 -0,!5
2)
-0,0! -0,04
-0,!9 -0,29
monthly impact lkgha-lmonth -11 as S or 0 3
8,94
0,66
2,97
depends on the quality of the measured profiles; they are influenced by instationarities
0,34
6,33
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
431
Dry deposition of acidifying and alkaline particles to forests" model and experimental results compared W. Ruijgrok KEMA, P.O. Box 9035, 6800 ET Arnhem, The Netherlands
Abstract This paper provides a comparison of modelling and experimental results on the dry deposition of acidifying and alkaline particles to forests. On average, a fair to good agreement is found between model and experimental results. Results show that the mean v d of acidifying aerosol (such as SO 4) is around 1 cm s 1, while for base cations (e.g. Ca) v d is approx. 5 cm s 1.
1. I N T R O D U C T I O N In 1991, the second phase of the Dutch Priority Programme on Acidification ended [1]. A major uncertainty recognized was the atmospheric input of acidifying and alkaline aerosol to forests. The contribution of acidifying aerosol deposition to total input is unclear. Most commonly, the deposition velocity of such particles is believed to be small. However, over very rough surfaces it is poorly quantified and may be larger than assumed upto now [2-3]. In the 3 rd phase of the Acidification Programme, a joint study was initiated to address the contribution of acidifying aerosol and base cations to atmospheric input and soil loads. This study consisted of a large experimental set-up together with a modelling study [4]. This paper compares estimates of particle deposition based on model and experimental results. The principal aim is to validate a model developed by Slinn [5] in such a way that deposition estimates derived from this model can be used for an assessment of the atmospheric input by particle deposition within the Netherlands. To this end a number of experiments were carried out in the course of 1992 and 1993 at the acidification research location Speulderbos in the Netherlands. The experiments differ in their set up, methods used, accuracy and each has its own advantages and disadvantages. Although results of all these experiments may not be directly comparable with each other, the whole array of results should bring about a clearer picture and understanding of particle deposition processes and of representative values for dry deposition velocities.
432 The model of Slinn was chosen following a review of current dry deposition models for forest conditions [6]. An extensive sensitivity analysis showed t h a t the uncertainty in v d may amount to 65% for acidifying compounds (60% for base cations). Some modifications were made: the influence of relative humidity on particle size and corrections for atmospheric stability were included. To compare modelled v d with measured values, an integration of v d was carried out over particle size distributions. Estimates for these distributions were made from concentration measurements carried Speuld or elsewhere in the Netherlands [6]. To model v d a number of parameters characterizing the canopy are required; for the Speulder forest these were estimated from data measured or from literature values. To describe particle collection efficiencies, equations given in [7] were adopted instead of the original equations of Slinn, as they provided a better agreement with observational data. We will present here a selection of the comparison with measuring data only. A full description can be found in [6].
2. M O D E L E V A L U A T I O N WITH M E A S U R E M E N T S
Fog deposition In December 1992 and F e b r u a r y 1993 a number of fog events occurred during which eddy correlation measurements were taken of the turbulent deposition flux of fog droplets [8]. Model estimates of fog water deposition were made by integrating modelled v d over the fog droplet spectrum which was measured. In December fog droplets showed a considerably larger mean diameter t h a n in February. This will influence the contribution of sedimentation to the total flux of fog w a t e r droplets. Sedimentation was not measured directly but calculated from fog droplet spectra. Results show a large contribution of sedimentation to the total fog deposition (57% of the total flux in December versus 13% in February). Fig. 1 shows a comparison of modelled and measured time series. On the whole, there is a reasonable agreement between modelled and measured values with a notable exception on F e b r u a r y 9. Model results show a clear relation of the t u r b u l e n t deposition velocity with u, 2 (Vd = 0.20 u, 2) which is close to the relation derived from measurements (vt = 0.195 u,2). A response of v d to u, 2 reflects the dominating influence of impaction.
214Pb accumulation Over a period of more t h a n a year 214pb accumulation was measured [8]. The values obtained can provide an analogue for the deposition of acidifying particulate m a t t e r in the submicron size range (such as sulphate), because of a similar size distribution. We selected 26 measurements for comparison with model results; the other data may have been biased by interference of u n a t t a c h e d 214pb. The experimental data relate to a local in-canopy deposition velocity at the measuring level, while the model provides a value referring to the entire canopy. To make the experimental data compatible with model estimates they have been scaled by a factor of 2.5. The uncertainty for this scaling factor ranges between about 1.5 and 4.
433
- = model o = measured 8-
4-"
6-
E E4X
LL 2-
0I 5
4
1 6
I 7
I 8
I 9
I 10
I 11
I 12
I 13
I 14
I 15
16
Day of month
F i g u r e 1 T i m e series of t h e deposition of liquid w a t e r over t h e S p e u l d e r forest in F e b r u a r y 1993. The line indicates model r e s u l t s a n d circles m e a s u r e m e n t s 2.0
j/" /"
2.0
/- /
//" 1.5-
/"
E
./
_9.o 1.0m
1"51 ---~ 1.0-~
,,,,,"
(D
0 E
It/"
//"
0.5-
//" // •
/ • •
y/
/o
oO
•
•
:
• °
• •
o ; o °o
:
•
°
// 0.0
0.0
o.o
•
o.o V. moasurod [cm/s]
L
oi,
,.o
U* [m/s]
F i g u r e 2 (a) C o m p a r i s o n of v d m e a s u r e d (incorporating a scaling factor of 2.5) a n d v d modelled. (b) The relation of v d modelled w i t h m e a s u r e d v a l u e s of u, The model p e r f o r m s relatively poor in r e p r e s e n t i n g the deposition velocity of 214Pb as m e a s u r e d in the canopy. This s e e m s largely caused by two groups: one w i t h a m e a s u r e d v d b e t w e e n 1.3 a n d 2 cm s 1 for which the v d is d r a m a t i c a l l y u n d e r p r e d i c t e d ; t h e other concerns a group w i t h a calculated v d of a r o u n d 1.3 cm s -1 while m e a s u r e m e n t s indicate a v d of a r o u n d 0.5 cm s 1. This r e s u l t s in h a r d l y a n y correlation b e t w e e n m e a s u r e m e n t s a n d model results. We e x a m i n e d if two possible factors (relative h u m i d i t y a n d friction velocity) m a y h a v e c a u s e d
434 the highest deposition velocities measured, but no clear relations were found. Despite the small correlation, the average v d calculated agrees quite well with the mean of the observations (although slightly underpredicted).
Concentration gradient measurements A number of concentration gradient measurements (NO 3, SO4 and particles) was carried out at the Speulder forest. All results were used for comparison with modelled values; however, we will present here results for SO4 fluxes from thermodenuder gradient measurements. A relatively large dataset was available after selection, which shows a reasonable comparison with modelled fluxes below 0.3 llg m 3 (Fig. 3). Above this value, measured fluxes are systematically underestimated. Large positive and negative measured values seem to be the result of random measuring errors in individual concentration measurements. Model and measuring results show by and large a similar response of v d to u,, a driving force of deposition (vd - u,12). Averaged model results of vd fall within 95% confidence limits of the averaged measured data. 0.3
,/ o/
i
/
7
o ~/
,._~0.2-" 0
0
O0
o
o
OD
0
0 0
//
0
o
o
0
~ o
1~?6go o I o oo o noO¢~.,o !
o o
o(b
0
o
oQ o
0
0
/
0 "lO
0
/
0
0
X D
o
~o
o,,/ooo
o
o
o °
0
o°
° o o
:S.¢~o
o~
0
0
ooo 0
0
o
o 0.0
o
o
o°O
.,~
-0.1
o
o
0.0
0.1
0
0
O0
o I
I
-, 1.2
o oo
°o~~° i °°,
0.2
'
I
0.3
'
I
0.4
'
I
0.5
).6
Measured SO4 flux [,u,g m 2 s 2]
Figure 3 results)
Comparison of modelled and measured SO 4 fluxes (thermodenuder
Throughfall Draaijers et al. [9] made an analysis of throughfall fluxes and canopy exchange processes at Speuld. As part of this analysis atmospheric deposition estimates were made for the various acidifying and base cation compounds. A summary of results is presented in Table 1. To assess the performance of the particle deposition model, the components in throughfall most suitable for comparison are Na and C1. The dry deposition of these two elements is via particle deposition only (as opposed to the sulfur and nitrogen compounds),
435 while canopy exchange processes do not influence their throughfall fluxes significantly. To establish the atmospheric input by particle dry deposition we inferred values of v d for acidifying and alkaline particles using the modified model of Slinn and continuous meteorological data measured at Speulder forest. Size distributions as derived from measurements at Speuld [6] were applied to obtain an integrated value of v d. Fluxes were calculated from these integrated Vd values and the series of aerosol concentration measurements [10]. Results are shown in Table 1 for the period November 1992 to May 1993. The estimates are based on a time coverage of around 65%. Estimates of the input by dry gaseous deposition and by fog have been derived from gradient m e a s u r e m e n t s and inferential modelling [4,8,11]. Table 1 Estimates of total atmospheric input (consisting of gaseous dry deposition, aerosol dry deposition and fog deposition) at Speulder forest. For comparison, net throughfall fluxes as well as net throughfall corrected for the canopy exchange processes are indicated. Fluxes are in tool ha 1 yr 1. Also shown are mean and standard deviation of aerosol v d used in these calculations (cm s ~) SO x
NOy
• gas • aerosol • fog total atmospheric input uncertainty
663 216 34 913 40%
net throughfall net TF - canopy exchange uncertainty
924 924 30%
average v d s t a n d a r d deviation
1.3 1.5
NH x
Na
C1
Ca
K
Mg
356 1443 414 645 23 96 793 2184 40% 50%
0 599 2 599 50%
0 855 4 889 50%
0 101 1 102 50%
0 34 1 35 50%
0 118 0 118 50%
394 1728 394 1983 40% 40%
692 692 30%
802 802 30%
159 86 40%
305 35 50%
138 98 40%
1.9 2.4
5.0 4.1
5.0 4.1
5.7 4.4
2.8 2.9
5.8 4.5
1.0 1.1
The long-term atmospheric deposition estimate of SO x nearly equals the net throughfall flux measured. For Na and C1, there is only a small difference between both ways of estimating input. In the case of NH x and NOy and most base cations, total atmospheric deposition is lower t h a n net throughfall. However, if a correction is applied for the influence of canopy exchange processes atmospheric deposition and net throughfall fluxes compare reasonably well in the case of NH x and base cations. Corrected net throughfall fluxes of Ca and Mg are slightly smaller than the atmospheric deposition estimate. However, this is probably caused by an overestimation of the canopy leaching of Ca and Mg by the canopy exchange model [9]. On the whole, long-
436 a n d SO 4 m e a s u r e m e n t s for w h i c h t h e l a r g e s t d a t a s e t s w e r e a v a i l a b l e . F o r t h e s e two c o m p o u d s , also a r e a s o n a b l e s i m u l a t i o n of t h e t i m e s e r i e s could be m a d e . M o r e o v e r , a c o m p a r i s o n of m o d e l a n d m e a s u r i n g e r r o r s gives no s t r o n g s u p p o r t for a s i g n i f i c a n t b i a s e s of t h e d e p o s i t i o n velocity by t h e m o d e l used. Table 2 Model p e r f o r m a n c e to s i m u l a t e v d m e a s u r e d w i t h v a r i o u s t e c h n i q u e s a t S p e u l d e r forest. F o u r d i m e n s i o n l e s s i n d i c a t o r s of m o d e l p e r f o r m a n c e a r e p r e s e n t e d : c o r r e l a t i o n coefficient (R), f r a c t i o n a l b i a s e s of m e a n (FBM) a n d v a r i a n c e s (FBV) a n d n o r m a l i z e d m e a n s q u a r e e r r o r (NMSE). F o r c o m p a r i s o n , m e a n a n d s t a n d a r d d e v i a t i o n (in cm/s) a r e s h o w n as well for m o d e l a n d measuring results Intercomparison Fog 214pb NO 3 (filterpack) SO4 (filterpack) SO4 (thermodenuder) Ca (DFM, unscaled) Ca (DFM, scaled) Na (throughfall) SO 4 (throughfall)
0.10
R
FBM
FBV NMSE
0.57 0.15 0.55 0.42 0.33 0.78 0.78 0.52 0.77
0.18 -0.37 0.02 0.55 -0.08 1.44 0.62 -0.14 0.01
0.71 -0.66 -1.06 -0.60 -1.54 1.86 0.93 -0.10 0.52
0.06-
e 0.04-
2.9 0.5 1.2 1.1 2.1 4.1 4.1 5.0 1.3
0.06--
/
2.3 0.5 1.9 1.4 3.4 0.5 0.9 -
116 26 23 23 169 14 14 8 8
Fog
NO,(TNO) 1.~SO'(ECN)
/ ~ ~ 0 , ,
(TNO)
i) o.oo-
' 0.0
,=
/
0.00-0.02
2.8 0.7 1.2 0.7 2.3 0.8 2.4 -
(B)
0.02>~ 0.04-
_.
1.8 0.3 1.1 1.0 1.2 2.5 2.5 1.5 4.1
n
0.08-
/ :°o:oo,
0.08-
>~ o.02-
1.12 1.01 1.75 2.66 1.94 5.16 0.59 0.26 0.04
0.10
(A)
vd ~ vd ~ model model meas. meas.
I 0.4
'
U. m/s
I 0.8
'
........
-0.02 1.2
' 0.0
I 0.4
'
I 0.8
1.2
U. m/s
F i g u r e 4 A s u m m a r y of t h e a v e r a g e d m e a s u r e d (A) a n d m o d e l l e d (B) v d v a l u e s as a f u n c t i o n of a v e r a g e d v a l u e s of u. (in i n t e r v a l s of 0.1 m/s) F r o m o u r c o m p a r i s o n one can o b t a i n little g r o u n d to reject m o d e l r e s u l t s , a l t h o u g h m o d e l p e r f o r m a n c e is not v e r y satisfying. H o w e v e r , p r e s e n t o u t c o m e s c o n f i r m e a r l i e r o b s e r v a t i o n s of h i g h e r d e p o s i t i o n velocities of acidifying p a r t i c l e s onto f o r e s t s a n d o t h e r v e r y r o u g h surfaces. This does not hold only for t h e d e p o s i t i o n m e a s u r e m e n t s c a r r i e d out in Speuld, b u t also for t h e m o d e l
437 term averaged particle deposition fluxes seem to be in line with input estimates obtained from (corrected) net throughfall. This provides confidence in the longterm averaged values of v d calculated for particle dry deposition as well as in the performance of the canopy exchange model. In order to make an evaluation on shorter time scales atmospheric deposition estimates were compared with net throughfall fluxes of 16 individual sampling periods (varying in duration from 57 to 493 hours). For these 16 individual events significant relationships between atmospheric deposition estimates and net throughfall fluxes could not be detected. This can largely be explained by incomplete wash-off of dry deposition from the canopy by precipitation [9]. This incomplete wash-off created a time dependency of throughfall samples on previously collected samples. A second reason for a lack of correspondence is the limited coverage in time of measurements on which some of the atmospheric deposition estimates are based. The limited time coverage can potentially introduce large errors in deposition estimates. Enlarging the averaging period over which deposition is calculated (by taking two consecutive periods together), modelled fluxes and net throughfall agree reasonably well (see results for Na and SO4 in Table 2).
3. S Y N T H E S I S
We presented a comparison between deposition flux measurements of various compounds and tracers with estimates based on calculations with Slinn's dry deposition model. A summary of results can be found in Table 2. The model version used for these calculations included alternative descriptions of particle collection efficiencies [7], instead of Slinn's original equations. These alternatives were adopted as they yielded significantly higher deposition velocities. Such differences reflect the uncertainty in current understanding of the particle deposition process. The option of adopting alternative efficiency descriptions was chosen as it seemed at first hand to match more closely with the data measured. To assess the performance of the model we used four different indicators to measure correspondence between measured and modelled fluxes. None of these indicators point at a good agreement; at best, their values illustrate a reasonable performance due to the large scatter. The cause for this are the relatively large inaccuracies of measuring methods used and the sensitivity of the model which is of the same order (60 - 130%). Depending on the compound, the averaged modelled fluxes differ with 2% to 144% from the averaged measured fluxes. The largest difference applies to a comparison with dry deposition fluxes from a filter method, a technique which may be characterized as unreliable. With respect to reproducing the variability in measured fluxes, the model seems to underestimate this variability systematically. On the other hand, the model adequately reproduces the response of v d to u, (an important driving force of particle deposition) by showing similar features as the measurements (see Fig. 4). This holds in particular for the case studies with fog
438 simulations. They also support the adjustment by Erisman [11] of the original parametrisation of v d by Wesely et al. [12] in order to obtain higher values for acidifying particles. With respect to Slinn's deposition model, such higher values can only be calculated if alternative descriptions of particle collection efficiencies are used. The present, adjusted version of Slinn's deposition model can in principal be useful to generalize dry deposition of acidifying particles to forests and other rough surfaces. A more convenient way (because of computational reasons) is to rely on a parametrisation of v d with u.. Secondly, as most input parameters of the model are not available for making deposition estimates for the entire Netherlands and also detailed particle size data are lacking, the model response will be primarily depending on u.. Such a relation has been successfully derived from this model [4,6]. A parametrisation will, no doubt, cause inaccuracies and extrapolations have to done with caution, but, at the same time, these inaccuracies will be mostly due to the uncertainties of the model. Using the model itself will hardly reduce uncertainty in v d for such purposes. It may only make the sources of such uncertainties more explicit.
4. R E F E R E N C E S 1 Heij G.J. and Schneider T. (1991). Studies in Environmental Sciences 46, Elsevier, Amsterdam. 2 Ruijgrok W., Davidson C.I., and Nicholson K.W. (1994). Dry deposition of particles: implications and recommendations for mapping of deposition over Europe. Accepted for publication in Tellus. 3 Nicholson K.W. (1988). Atmos. Environ. 22, 2653-2666. 4 Erisman, J.W., Draaijers G., Duijzer J., Hofschreuder P., Van Leeuwen N., RSmer F. Ruijgrok W. and Wyers P. (1994). Particle deposition to forest. Elsewhere in this volume. 5 Slinn W.G.N. (1982). Predictions for particle deposition to vegetative surfaces. Atmos. Environ. 16, 1785-1794. 6 Ruijgrok W., Tieben H. and Eisinga P. (1994). KEMA, Arnhem, report 20159-KES 94-xxxx. 7 Wiman B.L.B., and/kgren G.I. (1985). Atmos. Environ. 19, 335-362. 8 Wyers G.P., Veltkamp A.C., Vermeulen A.T., Geusebroek M., Wayers A., MSls J.J. (1994). ECN, Petten, report ECN-C-94-0xx. 9 Draaijers G.P.J., Erisman J.W., Van Leeuwen N.F.M., RSmer F.G., Te Winkel B.H., Wyers G.P. (1994). RIVM, Bilthoven, report 712108004. 10 RSmer F.G. and Te Winkel B.H. (1994). KEMA, Arnhem, report 63591KES/MLU 93-3243. 11 Erisman J.W. (1993). Water, Air and Soil Poll., 71, 51-80. 12 Wesely M.L., Cook D.R., Hart R.L., and Speer R.E. (1985). J. Geophys. Res. 90, 2131-2143.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
439
The measurement of ammonia in the National Air Quality Monitoring Network (LML)" (1) instrumentation and network set-up B.G. van Elzakker a, E. Buijsman a, G.P. Wyers b and R.P. Otjes b a National Institute of Public Health and Environmental Protection (RIVM), P.O. Box 1, 3720 BA Bilthoven, The Netherlands b Netherlands Energy Research Foundation (ECN), P.O. Box 1, 1755 ZG Petten, The Netherlands
1. introduction Ammonia (NH3) plays an important role in the total acid deposition in the Netherlands. On the average, its relative contribution to this deposition (SOx+NOx+NHx) is estimated to be about 50% for the whole country. Immission and deposition levels are calculated with a transport model which takes into account emissions on a 5 x 5-km scale, meteorologics and NH3-specific deposition parameters. To calibrate and validate such a model, measurements are, however, indispensable. The instrumentation and LML configuration used will be described.
2. instrumentation The instrument used is the continuous-flow wet denuder manufactured by the Netherlands Energy Research Foundation (1). It is not primarily designed for use in an automatic network; the set-up is more-or-less experimental. Therefore it is only used in an interim network. An operational instrument has been developed in the meantime for use in the final network (see §4). The principal of the continuous-flow wet denuder is as follows (see Figure 1). Air is sampled at a flowrate of about 301/min through a rotating annular denuder which contains an absorption solution (0.1M NaHSO4). NH3 in the air sampled is efficiently absorbed by the solution, which is continuously withdrawn from the denuder to the detector at a flowrate of ~1.5 ml/min. After adding a 0.5M NaOH solution, by which dissolved NH4 + is converted into gaseous ammonia, the solution passes through a semi-permeable membrane. The dissolved gaseous ammonia diffuses through the membrane and dissolves as NH4 + in demineralised water (with a certain NH4 + offset) on the other side.
440
to air pump
air inlet
rotating annular denuder
NH4+ in N a O H ~
liquid film
t
_detector demineral~ed w
waste
I
Teflon membrane
-/
I-- -- -- ~ a t a l o g g e r
ter~e~at~-re ~nd' conductivity signals
Figure 1. Schematic view of the continuous-flow wet denuder. This NH4 + -containing solution then proceeds to a conductivity cell. The temperaturecorrected conductivity is a direct measure for the NH4 + concentration. The NH3 concentration in the sampled air is calculated from this value, gas and liquid flow, and calibration data. All data is stored in the datalogger of the instrument, which can be read out by a computer. Time resolution is 1 min. During an extensive test - programme at RIVM (2,3) the instrument performed as follows: detection limit (3s) precision (ls) linear range drift max. deviation
0.02 <2 >500 -0.3 < 0.4 <4 < 1.2
~tg/m3 % ~tg/m3 % per day ~tg/m3 % %
(at zero air)
(at zero air) (at 10 ~tg/m3 NH3) (a 100 ~tg/m3 NH3)
This performance meets the required specifications. It should be noted that the above stated maximum deviation assumes an environmental temperature of the instrument in the range 15 - 25 oc. Therefore the measurement containers have to be temperature conditioned for use in the field.
441
3. network configuration LML stations
The interim network consists of 8 stations located in areas with different levels of ammonia emission densities. For model validation purposes this is necessary. The emission levels vary from < 1 ton NH3/km2 per year (background areas) to ~ 30 ton NH3/km2 per year (high emission areas). The network was set up in August 1992. Since then, the selected locations have changed slightly. From May 1994 the configuration has been as follows (see also Figure 2): stations in areas with average emission density: LML 633 Zegveld LML 928 Witteveen
stations in background- areas: LML 235 Huijbergen LML AA.A. De Zilk LML 538 Wieringerwerf stations in LML 131 LML 722 LML 738
areas with high emisssion density: Vredepeel Eibergen Wekerom f
f
0 Wieringe
De 7ilk2
eft
~ f"/" t •" ~ ~' \ / e ~ , , , , " r M-~E lberge Zegveld Wekerom ,,-,>.,1
Figure 2. Ammonia measurement locations in the Netherlands (August 1994).
442 Because ammonia sticks to almost every surface, special attention is given to the inlet configuration, which consists of a polyethene funnel (with the opening faced downwards) connected to the instrument by polyethene tubing. Also, the instrument is installed at the top of the measurement container to keep the inlet length as short as possible. Sampling height is 3 m above ground level. van - locations f o r representativeness measurements
For the distance to local emission sources, these locations are selected in the same way as the LML stations. This means at least 300 m from local ammonia sources in emission-and average areas, and about 1000 m in background areas. Because the expected gradient of ammonia levels in the emission areas is higher than in the background areas, the necessary number of van - locations is 8, 6 and 4 for emission, average and background areas, respectively. logistics
Each LML - station is visited once a week. The measurement period is finished with a final calibration of detector-cel and flows. Data is stored in a portable computer for validation and storage in Bilthoven. Solutions, tubing etc. are replaced and after calibration, a new measurement period is started. These field - activities take about 3 to 4 hours, depending on the chance of malfunctioning of the instrument. Representativeness measurements are carried out frequently.
4. future developments As previously stated, the instrument is not configured for use in an automatic network like LML. So, from the beginning of 1993 large efforts have been put in the development and extensive testing of an operational continuous-flow wet denuder in the cooperation of RIVM with ECN. The most important improvements with this instrument are the continuous control of important measurement parameters (flows, temperatures etc.) and essential functions, like rotation of the annular denuder. For this, an RS232 communication with local station processors (SPS) will become possible. These processors (which are part of the existing network) are connected with the central station at RIVM by a telephone line. Maintenance interval time for the instruments will be increased from 1 week to 1 month. These new ammonia monitors are expected to be completely implemented at the beginning of 1995.
4. references G.P. Wyers, R.P. Otjes and J. Slanina, A continuous-flow denuder for the measurement of ambient concentrations and surface-exchange fluxes of ammonia, Atmospheric Environment 27A (1993) 2085-2090 E.M. van Putten, M.G. Mennen, T. Regts and J.W. Uiterwijk, Performance study of four automatic ammonia monitors under controlled conditions, RIVM report 723101004, 1994 B.G. van Elzakker, J. Stuiver and G.J.B.M. van Uden, Onderzoek naar 11 ammoniak monitoren voor het interim meetnet ammoniak, RIVM rapport 723101007 (in prep.)
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
443
The measurement of ammonia in the National Air Quality Monitoring Network (LML)" (2) results and performance B.G. van Elzakker, J.T. Schippers, J. Stuiver and G.J.B.M. van Uden National Institute of Public Health and Environmental Protection (RIVM), P.O. Box 1, 3720 BA Bilthoven, The Netherlands
1. i n t r o d u c t i o n
Since August 1992 ammonia (NH3) concentrations have been measured at 8 locations in the National Air Quality Monitoring Network (LML) using instrumentation and network configurations as described in references 1 and 2. This paper gives a general overview of: a. the measurement results of the LML - stations (network configuration 1992-1993) b. the representativeness measurements carried out with a mobile van in the surroundings of the LML stations c. quality assurance and control d. operational performance.
2. m e a s u r e m e n t results network
Table 1 summarises the results (average, maximum and several percentile values) for each measurement location for the period August 1992 to July 1993 to give an impression of NH3 concentration levels in the Netherlands. Since May 1994, the stations at Lunteren and Leiduin have been replaced by Wekerom and De Zilk respectively, and so for these last 2 locations, measurement results over a whole year are not yet available. It can be noted that some stations show different concentration levels, as could be expected from the initial classification (emission, average or background). For ammonia, local disturbances can cause these differences fairly easy. Witteveen, for example, is a station located in a forest, while Zegveld is probably influenced by cows grazing nearby. Evaluation of this, with the help of the representativeness measurements in the vicinity of the locations (discussed below), is necessary for correct interpretations.
444 Table 1 Measurement results based on 1-hour average concentrations for the period August 1992 to July 1993 average station
max
(gg/m3) (gg/m 3)
percentile values (gg/m 3) 50
70
90
95
98
99.9
emission areas:
Vredepeel
18.5
256
12
21
47
63
90
162
Lunteren
23.5
428
17
26
47
62
85
155
Eibergen
11.8
236
9
13
22
31
47
129
Zegveld
11.2
252
7
12
24
35
55
167
Witteveen
2.7
34
2
3
6
9
13
26
Huijbergen
3.3
37
2
4
8
10
14
25
Wieringerwerf
5.6
346
3
6
12
18
28
188
Leiduin
2.4
41
1
2
7
10
14
34
average areas:
background:
Figure 1 shows the variation in the monthly average concentrations during one year in a high emission-density area (Eibergen station). September and March show the highest monthly concentrations, reflecting the (beginning of the) manure application to the fields.
m 25 20 t~
Z
15
o
10
=
5
o
0
.~,.~
aug sep oct nov dec jan feb mar apr may jun
jul
month Figure 1. Monthly variation of NH3 at Eibergen station (period 8-92 to 7-93). The diurnal variations can be of special importance for modelling purposes. In Figure 2 this variation is plotted for Vredepeel and Witteveen stations.
445
30 ~25 20
"= 10 o
5
o
0
Witteveen ,
;
2
•
;
4
'.
- - * - - ,
6
!
8
.'
•
;
;
10
;
12
,
,.
14
;
;
;
16
18
'.
,.
;
;
20
'.
22
,.
24
hour of the day Figure 2.
Average diurnal variation (period 9-92 to 7-93).
Figure 2 shows that this diurnal variation can be complete inversed due to several phenomena (for instance, meteorologics), which are not always easy to interpret.
representativeness measurements As stated before, these measurements, which are carried out with a mobile van, are necessary to interpret the continuous measurements at the LML stations and especially the relation with concentration levels in the surroundings (5 x 5 km2). Modelling is carried out on this scale. The representativeness measurements are scheduled to be carried out under different meteorological conditions. In total, 16 classes with different temperatures, wind speed and wind direction are defined (3). Figure 3 presents the results of just one representativeness measurement, for the Eibergen station (November 3 1993).
~" 20 18 ~ 16
g f
h
•= 12
i~ 10
~Z
/I
IX a t/m x" mobile measurements
2 0
,
t'-i
I
e6
,
I
~
,
I
~
,
I
,6
,
I
t'-:
,
I
,
06
time (hours) Figure 3. Representativeness measurement at Eibergen (Nov. 3 1993).
,
446 The concentrations measured at each van location are plotted against the concentration of the LML - station at the same time. Measurement x is the mobile measurement taken as close as possible to the LML station. Note: 1) the good agreement of this measurement with the LML station and 2) the sometimes neglected stability time for the mobile measurements. Typical measuring time at a mobile-van location is in effect only 15 min for several reasons (2). Interpretation of these measurements turns out to be very difficult because, for instance: measurements can't be carried out frequently during the whole year because of the substantial personnel capacity needed measurements are carded out only during daytime (for the same reason). -
-
For this, alternative measuring methods, preferably continuous in time, simple and low cost have to be investigated.
3.
quality
- assurance
and control
quality assurance Quality assurance is obtained by: * using only standard operating procedures in the field as well as in the laboratory. * validation procedures carried out for each single weekly data File. This validation includes the comparison of calibration data at the beginning and end of the measurement period, increase or decrease of air and liquid flows, condition of pumps, valves etc. A special software package was developed for this. Because of the previously mentioned experimental set-up of the equipment, this validation has to be carded out manually (however, computer-supported), thus requiring another person to check this validation.
quality control Quality control is organised as follows: checking calibration liquids at an independent laboratory. checking response in the field with an NH3 calibration device (2). comparison with another instrument. For this, a continuous-flow wet denuder is installed in a measuring van positioned as close as possible to the LML - station. This is done frequently for all stations of the network. Results are shown in Figure 4.
447
15%
50
1:1
40 -15% 30 ,~ 20 o
10
0
10
20
30
40
50
NH3 concentration at LML station (ug/m3) Figure 4. Quality control measurements (all stations). Intercomparabilities are within 15% (C > 10 gg/m3) or < 1.5 l.tg/m3 (C < 10 gg/m3).
4. operational performance Figure 5 shows the monthly availability of the (1-min) measurement data averaged over the 8 network locations.
100.0 90.0 ~" 80.0 7oo 60.0 .--z 50.0 40.0 30.0 20.0 10.0 0.0 . '
~,~
~ ''
: :'
: ~''
: ''
•
, :',
,',
, : ,
•
date Figure 5. Monthly availability of NH3 measurement data (all stations).
It can be seen that during the first 6 months of the interim network the availability increases from 50 to 80%, mainly due to improvements which could be made to the instruments.
448 Although the equipment used is experimental, the operational set-up with frequent visits to the LML - stations, use of the standard operating procedures and, last but not least, the employment of qualified personnel for maintenance etc. yields fairly satisfactory results. From February 1993 the average availability has been about 90% with a variation of about 10%, most often caused by just one station which performs less adequately during a certain period. One has to take into account that there is no continuous on-line control of these instruments in contradistinction to the automated equipment for other air pollutants in LML. Therefore the average availibility of 90% is high compared to the performance of the fully automated equipment (95-98%), taking into account the experimental set-up and complexity of the continuous-flow wet denuder. The availability will even increase with the anticipated implementation of less experimental equipment at the beginning of 1995 (1).
5. references B.G. van Elzakker, E. Buijsman, G.P. Wyers and R. Otjes, The measurement of ammonia in the National Air Quality Monitoring Network (LML): (1) instrumentation and network set up (this volume) B.G. van Elzakker, J. Aben, J.W. Erisman and M.G. Mennen, Het Interim Meetnet Ammoniak, RIVM rapport 723101008 (in prep.) G.M.F. Boermans and J.W. Erisman, Meetstrategieontwikkeling voor het representativiteitsonderzoek als onderdeel van het additioneel meetprogramma anamoniak; fenomenologie van NH3 en meetritsimulaties, RIVM report 222105001, September 1990
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
449
Fine resolution m o d e l l i n g of a m m o n i a dry d e p o s i t i o n in Great Britain R.Singles: M A S u t t o n + and K J Weston* * Department of Meteorology, University of Edinburgh, Scotland. + Institute of Terrestrial Ecology, Edinburgh, Scotland. Abstract
Work is presented on the initial development of a detailed treatment of NH3 dry deposition in an atmospheric transport model for Great Britain. There is considerable local variability in agricultural emissions of NH3. Because of the ground level nature of the emissions and variable deposition velocities (Vd) to different land
surfaces, a model is required that incorporates the land-use dependence of NH3 dry deposition, together with a detailed (multi-laver) treatment of atmospheric
diffusion and transport. The initial results of applying a variable Vd are presented together with a budget of net NH3 dry deposition over the country. Differences in concentration due to the use of variable Vd, compared with a spatially uniform value, are indicated. The Model
The model described, referred to as UKTERN, has been developed from the TERN model, of which a more detailed description can be found in ApSimon at al. (1994). It is a Lagrangian model and follows columns of air, of defined cross section and which extends from the earth's surface to the top of the mixing layer. The model consists of 33 vertical layers of variable depth, with the resolution becoming especially detailed close to the earths surface. The grid resolution used for UKTERN is 20 kin, with spatiallydisaggregated NH3 emission estimates derived from agricultural census information (Eager, 1992; Sutton et a1.,1993), as shown in Figure 1. A straight line trajectory approach is employed, with trajectories originating from 8 wind sectors. The results are combined statistically, suitably weighted by the frequency of the winds in each sector. An assumed mean wind speed of 8ms -1 is used to move the air column. There are two main aspects of this model, which distinguishes it from previous work (Metcalfe et al., 1989; Sandnes et al., 1992; Asman and van Jaarsveld, 1992). The first feature is the enhanced treatment of vertical diffusion, utilising a multi-layer approach and an assumed profile of the turbulent diffusivity. Vertical mixing is determined by the equation
On__ O (ic Ox=_) Ot
Oz
oz "
where X: is the concentration of the species under consideration. The vertical diffusivity (K,) is defined as a function of height, dependent on the diurnal cycle and prevailing stability conditions. In the model, K, has been assumed to increase in proportion to height over a layer of depth Hz to a maximum value Kmax and remain constant over the rest of the mixing layer. The second part of this investigation is on the treatment of the land-use dependence of NH3 dry deposition within an atmospheric transport model.
450 Gaseous NHa is removed from the atmosphere by dry deposition, wet deposition and by chemical conversion. Dry deposition is by way of deposition velocities (Vd) which remove material from the lowest layer of the air column. In the case of NHa, a landuse database is used which contains information on the % land cover of the main types of landuse for each 20 km square. For this initial study, a separate Vd is assigned to each landuse to reflect the overall effectiveness of the surface to be a recipient for NHa dry deposition. These landuse categories, and the initial values of Vd assigned to them for NHa, are listed in Table 1. In future work, it is expected that resistances will be used to more accurately describe the temporal and spatial variations of Vd. Other chemical species in the model are listed in Table 2, together with the applied Vd for each species. These values are taken to be independent of landuse. A detailed description of the chemical processes in the model can be found in ApSimon at al. (1994). Wet deposition is calculated using a washout coefficient dependent on a specified rainfall rate. A background concentration of 0.1 #gm -3 is used for NHa and 0.2 #gm -3 for NH +. All other species have initially been given zero background values. R e s u l t s a n d Discussion The motivation behind this work was to see whether the detailed treatment of the diffusion process and dry deposition, described in the previous section, would produce significantly different results from previous models which assumed instantaneous mixing (Metcalfe at al., 1989, Sandnes at al., 1992) and dry deposition rates independent of landuse type (Asman and van Jaarsveld, 1992, Metcalfe at al., 1989, Sandnes at al., 1992). Figure 2 shows the results fl'om the model with the detailed treatment of diffusion included, together with the landuse dependent values of Va for NH3. There is a strong correlation between areas of high emission and large ground level concentrations. This is more evident in this model because of the use of multi-layer diffusion and limited rates of dry deposition to the main emission areas. Figure 3 is a comparison plot of two runs of the model. The first run uses the landuse-dependent Vd (as shown in Figure 2), whereas the in the second run a constant value of 0.01 ms -1 was assigned to the NH3 Vd, independent of landuse. The difference map produced can be split into two areas of interest. In the east of England, there is a greater concentration due to smaller landuse-dependent Vd, compared with the assumed constant value of Vd. In Wales, northern England and Scotland, the reverse in true with the more detailed treatment of Va resulting in more deposition and smaller air concentration. These differences can be attributed to the 'type' of land prevalent in these areas. In the east, the land is relatively fiat and has a large amount of arable crop cover. On the rest of the map however, there are regions where more semi-natural ecosystems dominate, such as moorland and forests. Thus the eastern area will on average be assigned a value of Va smaller than 0.01 ms -1, resulting in less NHa being removed from the atmosphere, whereas the rest of the land will have in general a greater value than this being used. This will affect air concentrations. Figure 4 shows the NHa dry deposition budget for Great Britain produced by the model using the enhanced diffusion process and the inclusion of landuse-dependent deposition velocities. These results show the importance for NHa allowing Vd to vary with landuse. The multi-layer approach is also important since this provides a more realistic treatment of
451 atmospheric mixing, allowing the land-use effect on deposition to be seem more clearly. Future work will focus on a more detailed resistance analysis and compensation point treatment of the dry deposition process.
LandClass
Vd/ms -1
Specie
Vd/ms -1
Arable Grassland Moorland Forest Urban Sea
0.003 0.01 0.02 0.03 0.004 0.004
SO2 SO42-, NO3, NH + aerosol PAN NO NO2 HNO3
0.008 0.001 0.002 0.000 0.001 0.04
Table 1. Initial NH3 values of Vo assigned to the landuse database.
Table 2. Dry deposition velocities for other chemical species used by the model.
Acknowledgements. Financial support for this work and its presentation is gratefully acknowledged from the UK Natural Environmental Research Council (RJS) and the UK Department of the Environment (MAS).
References
ApSimon, H.M., Barker, B.M., and Kayin, S. (1994). Modelling studies of tile atmospheric release and transport of ammonia in anticyclonic episodes. Atmospheric Environment 28, 665-678. Asman W.A.H. and van Jaarsveld J.A. (1992) A variable-resolution transport model applied for NHx deposition for Europe. Atmospheric Environment 26A, 445-464. Eager M. (1992). The development of an ammonia emissions inventory for Great Britain using GIS techniques. M.Sc. Thesis, University of Edinburgh, Edinburgh, U.K. Metcalfe S.E., Atkins D.H.F. and Derwent R.G. (1989). Acid deposition modelling and the interpretation of the United Kingdom secondary precipitation network data. Atmospheric Environment 23, 2033-2052. Sandnes H. and Styve H (1992). Calculated Budgets for airborne a cidiL'ing components in Europe, 1985, 1987, 1988, 1989, 1990 and 1991. EMEP Technical Report no.97. Sutton M.A.. Fowler D., Smith R.I., Eager M.. Place C.J. and Asman W.A.H. (1993). Modelling the net exchange of reduced nitrogen. In General assessment of biogenic emissions and deposition of nitrogen compounds, sulphur compounds and oxidants in Europe. Proceedings of the joint CEC/BIATEX workshop. Aveiro (~lay 1993). Commission for the European Communities, Brussels. Published in Air Pollution Research Report 47, 117131.
452
+ 6.00 4.50
3.00 t .50 0.00
Figure I.N-NH, emission mop of Greot Britain.
Figure 2. NH, concentration map of Great Britain using land-dependent V, and detoiled vertical diffusion.
+
t
0.80
12.00 0.40
0.00
8.00
-0.40 4.00
-0.80
-
Fi ure 3. Difference in concentration %etween o run of the model with lond-dependent V, and that with constant V., Dotted line is zero contour.
0.00
Figure 4. N-NH, dr deposition map of Great B d a i n using 1a nd depe nde nt V.,
-
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
453
Fog deposition on Douglas Fir Forest A.T. Vermeulen a, G.P. Wyers a, F.G. R6mer b, G.P.J. Draaijers c'd, N.F.M. van Leeuwen ¢, J.W. Erisman d aFossil Fuels Unit, Netherlands Energy Research Foundation, P.O. Box 1, 1755 ZG Petten, The Netherlands bKEMA, P.O. Box 9035, 6800 ET Arnhem, The Netherlands CDepartment of Physical Geography, State University of Utrecht, P.O. Box 80115, 2508 TC Utrecht, The Netherlands dAir Research Laboratory, National Institute of Public Health and Environmental Protection, P.O. Box 1, 3720 BA Bilthoven, The Netherlands
Abstract In December 1992 and February 1993 two periods in which fog occurred were monitored at the location Speulderbos in The Netherlands. The occult deposition during these periods was 3.4 and 2.0 mg/(m2.s) respectively. The contribution of occult deposition to the total acid deposition to forests in The Netherlands is estimated to be about 5%. High correlation between the friction velocity u. and turbulent deposition velocity vt of fog droplets could be derived.
1. INTRODUCTION Occult deposition is the downward transport of atmospheric pollutants by direct transfer of fog or cloud droplets to a receptor surface [1]. The turbulent deposition of fog droplets or other large particles can be measured directly or indirectly with micrometeorological methods. In order to infer deposition rates from simple routine measurements it is necessary to relate these measurements to more direct but expensive measurement techniques.
2. MEASUREMENTS The measurements were carried out at Speulderbos, one of the research sites of the Dutch Priority Programme on Acidification in the centre of the Netherlands [2]. Direct measurements of the turbulent deposition of the fog droplets were made with the eddy correlation technique. Gravitational settling was calculated from the measured droplet size distribution. The eddy-correlation
454 measurements were carried out using a Solent 3D symmetric head sonic anemometer. The Liquid Water Content (LWC) was measured with a Gerber PVM-100 sensor. Samples of the fog water were drawn with a CWP string collector operated by KEMA. The size of the fog droplets was measured with a Forward Scattering Spectrometer Probe (FSSP) of Particle Measurements Systems.
3. R E S U L T S
The measured throughfall flux was a factor two higher than the measured occult deposition flux. The concentrations of acidifying ions in fog water were a factor 2 to 3 higher than in throughfall water. The occult deposition velocity vt showed a high correlation with the squared friction velocity (u2., r=0.83) and with the deposition velocity for momentum (vm, r=0.75). The measured turbulent deposition velocity for fog is half the deposition velocity for impulse. Table 1. Calculated turbulent and gravitational fog water fluxes in mg/(m2.s) for data selected from two periods (n=537). Data within brackets concerns all data. Period Fturb Fgrav Ftotal n Fthroughfall December '92 1.47 (1.81) 1.97 (1.74) 3.44 (2.55) 77 .5.1 February. '93 1.76 (1.52) 0.26 (0.30) 2.02 (1.82) 460 5.3 Total 1.72 (1.57) 0.51 (0.38) 2.23 (1.95) 537 5.2 Table 2. Estimated annual contribution of acidic components of deposited fog water compared with the total deposition (wet+dry) for Speulderbos in 1989 [3] usizng measured concentrations in mist in 1989 [4]. Component Fog dep. (Mol/ha) Total Dep. (Mol/ha) Relative (%)
NOy
47
940
5
SOx
46
910
5
NHx
134
2880
5
4. R E F E R E N C E S
1 2 3 4
J.F. Nagel, Q.J.R. Meteorol. Soc, 82 (1956), 452-460 A.T. Vermeulen, G.P. Wyers, F.G. R0mer, G.P.J. Draaijers, N.F.M. van Leeuwen and J.W.Erisman, in preparation, 1994. T. Schneider and G.J. Heij, Thematic reports Dutch Priority program on Acidification, Elsevier, Amsterdam, 1990. J. Slanina, B.G. Arends, M.P. Keuken, A.C. Veltkamp, G.P. Wyers, ECN, Petten, 1990.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
455
The contribution of canopy exchange to differences observed between atmospheric deposition and throughfall fluxes G.P.J. Draaijers a), J.W. Erisman a), N.F.M.van Leeuwen b), F.G. ROmere), B.H. Te Winkel e), A.T. Vermeulen, G.P. Wyers d), K. Hansen e)
National Institute of Public Health and Environmental Protection (RIVM), Air Research Laboratory (LLO), P.O. box 1, 3720 BA Bilthoven, the Netherlands University of Utrecht (UU), Department of Physical Geography, P.O. box 80.115, 3508 TC Utrecht, the Netherlands Electric Power Research Institute (KEMA), Environmental Research Department, P.O. box 9035, 6800 ET Arnhem, the Netherlands Energy Research Foundation (ECN), P.O. box 1, 1755 ZG Petten, the Netherlands Danish Forest and Landscape Research Institute, Department of Forest Health and Forest Ecosystems, Skovbrynet 16, 2800 Lyngby, Denmark.
Introduction For the evaluation of emission abatement measures it is essential that a relation can be established between emission of air pollutants and adverse effects as a result of exposure and deposition. Most adverse effects of air pollutants in forest ecosystems are found to occur due to changes in the soil system (Hey & Schneider, 1991). Critical loads are therefore directly referring to soil loads. The soil load (usually determined by measuring throughfaU and stemflow) may differ from the deposition flux (determined by a combination of air concentration and meteorological measurements and inferential modelling) as a result of canopy exchange processes. The mechanisms of these processes are still not well known and debate continues on their contribution to the gap between soil loads and deposition fluxes (Draaijers et al., 1994). Methods ThroughfaU and precipitation fluxes were measured at the Speulder forest research site in the Netherlands with different time resolutions allowing the application of two canopy exchange models. Sequential sampling during throughfall events allowed a detailed study of the dynamic mechanisms of canopy exchange. Information on canopy exchange was also provided by comparing throughfaU deposition estimates with estimates from micrometeorological measurements and inferential modelling, and by comparing deposition estimates from surface wash experiments using real and artificial twigs. Specific information on canopy exchange of root-derived sulphur at the Speulder forest was provided by a S35 nutrition experiment. Canopy exchange rates at the Speulder forest Sulphur was found to behave conservative within the canopy, with SO 2 uptake (35 mol/ha.year) more or less balancing leaching of soil-derived SO42" (80 mol/ha.year). Stomatal uptake of NO 2 and HNO 2 amounted 130 mol/ha.year. Experiments did not indicate significant uptake of NO 3" from water layers coveting the tree surface, leaving an inexplicable gap of
456 approximately 270 mol/ha.year between the NO 3" soil load and the NO. deposition estimate. Stomatal uptake of NH 3 amounted 140 mol/ha.year, whereas uptake Yof NH4+ in solution equalled 115 mol/ha.year. About 180-200 mol H+/ha.year was retained within the canopy. Canopy uptake of H+ and NH4÷ was encountered by leaching of K ÷ (270 mol/ha.year), Ca z÷ (50-75 tool/ha.year) and Mg z÷ (0-40 mol/ha.year). Part of the leaching of K ÷, Ca z÷ and Mg 2÷ (15%) took place along with weak organic acids. No significant canopy exchange was found for Na + and C1-. Differences observed between atmospheric deposition and throughfall fluxes could almost completely be explained by canopy exchange. For closing the gap between the soil load of NO"3 and the deposition flux of N Oy, additional research is necessary More knowledge regarding canopy exchange of nitrogen compounds can be obtained by using radioactive tracers (I5N) in ecosystem studies. At the same time, NO 2, HNO z, HNO 3 and NO 3" dry deposition estimates from micrometeorological measurements and inferential modelling need to be improved. Generalisation of measurement results
Field experiments at the Speulder forest were mainly performed in the winter period (November until May) when the vegetation is physiologically less active. As measurement results were scaled to one year, stomatal uptake as well as uptake and leaching in solution is most probably underestimated. During the measurement period no episodes with winter smog, frost, drought or an insect plague occurred. Such stress factors may intensify canopy exchange processes considerably. Canopy exchange rates derived for the Speulder forest may not be considered representative for other forests in the Netherlands as canopy exchange is found to depend strongly on tree species and ecological setting. A canopy exchange model developed by Van der Maas & Pape (1991) has proven to be a useful tool for determining the impact of canopy exchange on throughfaU fluxes. The combination of throughfaU measurements and this model results in deposition estimates which are similar to deposition estimates derived from micrometeorological measurements and inferential modelling. Unfortunately, several basic assumptions in the canopy exchange model are not properly evaluated under different environmental conditions (ecological setting, pollution climate), which limits the models' utility up to now to forest stands growing on dry and sandy, nutrient poor podzolic soils under current air pollution levels. The model can be improved by taken into account the different mass median diameters of Mg 2+, Ca 2+ and K÷ containing particles compared to Na+ containing particles in the calculation of the dry deposition factors. Moreover, stomatal uptake of NO 2 and HNO 2 has to be included in the model (Draaijers et al., 1994). References
Draaijers, G.P.J., J.W. Erisman, N.F.M. van Leeuwen, F.G. ROmer, B.H. te Winkel, A.T. Vermeulen, G.P. Wyers & K. Hansen (1994), A comparison of methods to estimate canopy exchange at the Speulder forest. RIVM-report 722108004. Hey, G.J. & T. Schneider (1991), Final report on the second phase of the Dutch Priority Programme on Acidification, RIVM report no. 220-09. Van der Maas, M.P. & Th. Pape (1991), Hydrochemistry of two Douglas fir ecosystems and a heather ecosystem in the Veluwe, the Netherlands. Dutch Priority Programme on Acidification, report no. 102.1.01.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BE All rights reserved.
457
Dry deposition monitoring of SO2, NH3 and NO2 over a coniferous forest
Jos Hogenkamp§, Jan Willem Erisman§, Marcel Mennen§, Erik Kemkers§, Addo van Pul§, Geert Draaijers§, Jan Duyzer~ (TNO) and Paul Wyers (ECN)~ §RIVM, P.O.Box 1, 3720 BA Bilthoven, the Netherlands, ~TNO, P.O.Box 6011, 2600 JA Delft, the Netherlands, ~ECN, P.O.Box 1, 1755 ZG Petten, the Netherlands
1. INTRODUCTION In 1990 a large project was initiated to develop a monitoring method for measurement of deposition of acidifying components onto forests. This project is the continuation of the successful development of monitoring methods for SO2, NO2 and NH3 deposition to low vegetation [1:3]. From January to August 1992 tests were done with the equipment over low vegetation at the heathland Elspeetsche Veld [4]. These tests included experiments determining effects of obstacles (monitor housing) to be installed in the mast above the forest on the momentum and heat flux measurements, optimisation of the gradient system for NO2 and SO2, and tests of eddy correlation measurements of the NO2 flux. At the end of 1992 the optimised monitoring systems for SO2, NO2 (RIVM) and NH3 (ECN) were installed at the forest site. Since then continuous vertical concentration gradient measurements of these components are available. In June 1992 also eddy correlation measurements of NO 2 were started. These data, however, have not yet been validated and evaluated. In this paper, the first results of the measurements made between November 1992 and September 1993 are presented.
2. SITE DESCRIPTION AND FLUX PROFILE RELATIONS The measuring site consists of a homogeneous 2.5 ha monoculture of Douglas fir, 35 years old with a stem density of 785 ha-1. The mean tree height is 20 m. The canopy is well closed with the maximum leaf area density, one-sided LAI = 9, at a height of 10-14 m. The measuring system is extensively described in [4]. Flux profile relations for ozone and heat were evaluated from simultaneous eddy correlation and gradient measurements by TNO in 1993 [5]. The ¢xh factors found by TNO for 0 3 and heat are similar to those found by [6]. It is concluded that the conventional flux-profile relations can be used for the estimation fluxes of heat and ozone using gradient measurements. It is assumed that these relations can be applied for other trace gases (SO 2, NO 2 and NH3). The eddy correlation measurements of u . and H at 30 m height by TNO and those at 36 m height by RIVM were compared to evaluate the constant flux layer assumption (Fig. 1 and 2). The agreement between the 30 m and 36 m u . values was reasonable, whereas the difference between the heat fluxes measured at the two levels was much. From the two figures it was concluded that during these measurements the constant flux layer assumption for momentum and sensible heat fluxes is valid
458
I
1.5 E
6-
=
I
I
-230 .___~~
"~ o.5
0
0.5
I
I
1.5
2
u" 36 m (m/s)
FIGURE 1. Eddy correlation u. measurements at 30 m compared to those measured at 36 m height.
lid
310
T-' I H0 36 m 0N/m2)
FIGURE 2. Eddy correlationH measurementsat 30 m compared to those measured at 36 m height).
3. A V E R A G E DEPOSITION PARAMETERS FOR SELECTED PERIODS For each hour three vertical SO2 and NO2 gradients were averaged and c., F, Vd, Ra, Rb and Rc were calculated. The ten NH 3 gradients measured within one hour were also averaged and the deposition values were calculated accordingly. The dataset thus obtained has to be 'cleaned' by selection of hours during which the theoretical demands for the gradient technique were fulfilled [1], during which the concentrations were well above the detection limit, and during which there was no loss of necessary measurements due to technical problems. Rc values show strong variations with time; when the surface becomes wet, Rc values drop to zero, whereas at very dry conditions Rc values can easily increase up to values of 1000 s m-1 at night. An Rc parameterization derived from analogous measurements over a heathland during three years [2] was tested. This parameterization can be applied for routinely measured components; it is based on literature values for the stomatal resistance, on empirical values for wet surfaces (due to rain or at high relative humidity) and for snow covered surfaces. Parameterized and 'measured' Rc values for SO 2 show reasonable agreement (40% of variance accounted for), with high values during dry periods at night, low values during daytime and values approaching zero during wet surface conditions. Rc values for NH 3 show similar variations. Emission of NH 3 was observed on several occasions and seemed to be strongly related to drying of the canopy. Emission was observed mainly during the day at a relative humidity decreasing below 80%. When the canopy is dry and the flux is towards the surface, this flux is higher than the inferred stomatal flux, indicating that the external leaf surface is also an important receptor for NH 3. For the whole dataset no influence of NH3 on deposition parameters for SO2 and vice versa was observed. This in contrast to the observations of [3] who demonstrated influence of both gases on the deposition of each other under extreme conditions.
4. References [1] Erisman J.W., Versluis A.H., Verplanke T.A.J.W., Haan D. de, Anink D., Elzakker B.G. van, Mennen M.G. and Aalst R.M. van (1993) AtmosphericEnvironment 27A,1153-1161. [2] Erisman J.W., Elzakker B. van, Mennen M., Hogenkamp J., Zwart E., Beld L. van den, R6mer F.G., Bobbink R., Heil G.,Raessen M., Duyzer J., Verhage H., Wyers G., Otjes R., M61s J. (1993) AtmosphericEnvironment,28,487496. [3] Erisman J.W. and Wyers G.P.(1993) AtmosphericEnvironment,27A,1937-1949. [4] Zwart H.J.M.A., Hogenkamp J.E.M., Mennen M.G. (1994) Report no. 722108001, National Institute of Public Health and Environmental Protection, Bilthoven, the Netherlands. [5] Duyzer J.H., this volume [6] Bosveld F.C. (1991) Report WR-91-02, KNMI, de Bilt, the Netherlands.
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
459
Gas depositionof sulphurdioxide on the territoryof the Czech Republic in 1991 Milo~
Zapletal
Environmental 746 01 Opava,
M o n i t o r i n g Centre Czech Republic
- Ekotoxa
Opava,
Horn/
n~m. 2,
Abstract In this c o m m u n i c a t i o n detailed map of d e p o s i t i o n of s u l p h u r dioxide gas is presented on the territory of the Czech R e p u b l i c in the year 1991. The c o n c e n t r a t i o n and deposition values have been c a l c u l a t e d for a rectangular g r i d - s i z e of 10xl0 km. In the areas where c o n c e n t r a t i o n data were not available, the values have been computed by spatial interpolation based on linear statistical techniques of Kriging. The deposition velocities were estimated in a c c o r d a n c e w i t h the dominant land-use types a s s i g n e d to 10xl0 km grid squares, using a geographical i n f o r m a t i o n s y s t e m for the grid analysis.
i.
INTRODUCTION
The s i g n i f i c a n t a c i d i f y i n g pollutants in the Czech R e p u b l i c are s u l p h u r dioxide, nitrogen oxides and ammonia, and their reaction products. There is no doubt about the n e g a t i v e influence of these acidifying components on soils, w a t e r and v e g e t a t i o n [i]. The C z e c h R e p u b l i c decided in the year 1992 to p a r t i c i p a t e in UN ECE program of mapping critical loads and critical levels. One of aims of the National p r o g r a m of critical loads is the c a l c u l a t i o n of d e p o s i t i o n and its g e o g r a p h i c v a r i a t i o n by means of thematic maps. There can then be c o m p a r e d w i t h critical load maps to deduce exceedance. The M i n i s t r y of E n v i r o n m e n t of the Czech Republic has d e l e g a t e d a c t i v i t y in this d o m a i n to the Environmental M o n i t o r i n g Centre situated in Opava. For e s t i m a t i o n of gas deposition SO 2 on the t e r r i t o r y of the Czech Republic distinguishing by the proportions of a g r i c u l t u r a l and forestry ecosystems, a d e p o s i t i o n model has
460 been used deriving deposition fluxes from annual mean c o n c e n t r a t i o n s of SO 2 in air above the dominant land-use types (ecosystems) and from appropriate d e p o s i t i o n velocities.
2. C A L C U L A T I O N
OF G A S D E P O S I T I O N
OF S U L P H U R D I O X I D E
To define air concentrations over the grid spatial i n t e r p o l a t i o n of air SO2 concentration values was used with a set of 333 points irregularly dislocated on t e r r i t o r y of the C z e c h Republic. Monitoring stations had been placed at these points. On the basis of daily measurements at these stations in the year 1991 the annual concentration values were c a l c u l a t e d [2]. For the transformation from the i r r e g u l a r grid to the regular (square) one the spatial i n t e r p o l a t i o n m e t h o d used t e c h n i q u e s of Kriging [3]. The f o l l o w i n g deposition model for the c a l c u l a t i o n of gas d e p o s i t i o n of SO 2 was applied [4]: DEP(SO2 )GAS = C(SO2 )'vd (SO2)'Fu
(i)
where C(SO 2) represents the averag~ annual c o n c e n t r a t i o n of SO 2 in the grqund air layer (ug.m-°), V d d e p o s i t i o n v e l o c i t y of SO~ (mm.s -x) related to mean height value z = 1 m above grouna and F~_~ 9,855 corrsspon~s to t r a n s f o r m a t i o n factor from u g . m m . m -3 to m o l c . h a - ~ . y r --. The values of deposition velocity, Vd, have been d e f i n e d to d e p e n d on the d~minant land-use types: v d = 8 mm.s -~ for forests = 5 mm.s -I for forest-arable land = 4 mm.s -I for crops and meadows = 2 mm.s -I for water areas, residential and p r o d u c t i o n a l built-up areas, mining and d e v a s t a t e d areas. These d e p o s i t i o n velocity values V d are average data s e l e c t e d from estimates published in the literature [5,6]. The c a r t o g r a p h i c model of land-use types was d e r i v e d from the map "Land use" published in "Atlas of the e n v i r o n m e n t and health of the population of the ~SFR" in scale i:i000000 [7]. A single land use type was defined for each 10xl0 km grid square a c c o r d i n g to the dominant (i.e. most prevalent) land use in that square.
3. R E S U L T S
AND CONCLUSIONS
The d i s t r i b u t i o n of gaseous SO d e p o s i t i o n values is m a p p e d on the 10xl0 km grid for all 2the territory of the Czech Republic (Figure i.). T~e t~ 5% of values of gas SO 2 d e p o s i t i o n (>3335 molc.ha .yr is again in n o r t h - w e s t part
]as Deposition
of SOe - 1991 Units : rnOJc.~ I.yr •
EGEND
"
> lOOO SO00
-
3000-
q[--"~
. . . . . . ~ ~ - ~ ~ "--.~EE~L-.___.
Fig.
i. Gas d e p o s i t i o n
nil ...... :
~':"...
~ ~ ~ m
of SO 2 in the Czech R e p u b l i c
in 1991 (molc.ha-l.yr-l).
m
ZOO()
.5000
2&O0-
3000
20~-
2500
ISO() -
2000
lOOO-
ISO0
~OO-
IOO0
O-
SO0
~
462
of Bohemia and in certain localities of middle and east 5% share of lowest values (<540 m o l c . h a - l . y r -I) Bohemia. The are located in the region of south Bohemia and in a small area of n o r t h - e a s t Bohemia. The annual mean value of g a s e o u s SO 2 d e p o s i t i o n in the year 1991 was 1741 m o l c . h a - l . [ r -I. The range of gas SO, d e p o s i t i o n values is 7446 m o l _ . h a - ~ . y r -I (min=138 molc.ha-l.yr-l,, median=1532 m o l c . h a ~ l . y r -I, max=7584 mol~-X.ha-±.yr-±). An areal d i s t r i b u t i o n of gas SO 2 d e p o s i t i o n on the t e r r i t o r y of the C z e c h Republic is characterized by an uneven t r a n s i t i o n from the h i g h e s t values in north-west region of Bohemia to the lowest values in south Bohemia and in south Moravia. Considerable v a r i a b i l i t y of high deposition values is shown over all the territory. The n u m e r i c results of the calculations have been c l a s s i f i e d into 9 categories for the purpose of the cartographic displaying. For better orientation the map is supplemented with the frontier of the Czech Republic and the EMEP grid (size 150x150 km), which is the spatial reference for the E u r o p e a n maps used within UN ECE [8].
4.
REFERENCES
1
B. Moldan, Acid rains and acidifying of e n v i r o n m e n t (in Czech). Vesm~r (The Universe), 61(5) (1982) 142-144. CzIH, Numerical year - book on air purity in the territory of the Czech Republic in 1991 (in Czech), The Czech Institute of Hydrometeorology, Praha, 1992. M. Zapletal, Model of immission SO 2 c o n c e n t r a t i o n in the territory of the north Moravia and Silesia and its interpretation by means of computer graphical output (in Czech). In: Modelling of the biological systems. Proceedings of CSAZ The 10-th summer school of biometrics, ZS VTS, Nitra, 1992. B. Rihm, S. Kunz, J. Engel, Mapping critical loads for Switzerland, METEOTEST, Bern, 1992. D. Fowler, Removal of sulphur and nitrogen c o m p o s i t i o n from the a t m o s p h e r e in rain and by dry deposition. In: D. Drablos and A. Tollan (eds.), Ecological Impact of Acid Precipitation. Proceedings of an International Converence, Sandefjord, Norway, 1980. J.A. Garland, D.H.F. Atkins, C.J. Readings, S.J. Caughey, Technical Note; Deposition of gaseous sulphur dioxide to the ground. Atmospheric Environment, 8 (1974) 75-79. IGCAS, Atlas of the Environment and Health of the Population of the CSFR, Institut of G e o g r a p h y CAS, Brno, 1992. CCE, M a p p i n g Critical Loads for Europe, C o o r d i n a t i o n Center for Effects, Technical Report No. I, National Institute for Public Health and Environmental Protection, Bilthoven, 1991.
2
3
4 5
6
7
8
POSTERS
SESSION X l MISCELLANEOUS
This Page Intentionally Left Blank
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
465
Forest condition in Europe and North America: What have we learnt over the past ten years? H. Visser KEMA Environmental Services, P.O. Box 9035, 6800 ET Arnhem, The Netherlands
Abstract Due to alarming signals of forest decline in Europe and the U.S.A., a number of research activities started around 1984 to find the causes and to quantify the role of air pollution. Now, in the year 1994, the topic of "acid rain" has faded into the background. Studying the literature of the past ten years, this action seems justified. Little proof can be found for causal relations between forest condition and any form of pollution. Exceptions are the declines occurring when gaseous concentrations are far beyond critical levels. Here, harmful effects are directly observable. However, appearances may be deceptive when concentrations are low. Soil acidification and its subsequent effect on the functioning of roots appears to be the dominant mechanism of injury in this case. Acidification is a cumulating phenomenon and it will not stop when deposition levels decrease. Therefore, the quality of soils has to be carefully monitored in the future. Further development of the concept of critical loads will be an important topic in the forthcoming years.
1. I N T R O D U C T I O N During the 1980s, considerable concern about the condition of trees developed in many parts of the world. Visitors to forests saw dead and dying trees and foresters described the increasing necessity to fell trees prematurely in order to prevent the spread of insects. The story was picked up by the increasingly environmentally aware media and by politicians. As a result, the issue of forest decline rapidly became one of the most widely discussed environmental topics of the 1980s. Since then, public interest in the subject has waned and it no longer draws the same level of political interest. This was partly because of an improvement in the health of forests in the late 1980s ([1]). The central issue in the debate about forest decline has been the detrimental effect of air pollution. The seeming absence of any long-term, large-scale decline in the overall condition of forests does not preclude any adverse effects of air pollution. Whereas acute injury is usually easy to diagnose, chronic effects are
466 much more difficult to identify. This has not prevented some from making such statements about air pollution as e.g. "in recent years forest damage has increased in country X because of air pollution, possibly in combination with climatic effects". As with other such claims, the evidence supporting this assertion is limited in many cases. In the following sections, three questions will be answered. First, what body of evidence is sufficient to prove causal relations? Second, what scientific knowledge has been gained over the past ten years? Third, what are the implications for future research?
2. CAUSALITY 2.1
Criteria
It is important to clarify which body of evidence is sufficient to prove causal relations. Innes in his textbook on forest decline ([1]), recalls criteria from [2]: (i) the infecting agent must be present in all patients showing symptoms of the disease, (ii) the infecting agent must be isolated from the patient and (iii) the infecting agent must produce the disease under controlled laboratory conditions. These criteria were expanded by the Committee on Biologic Markers of Air Pollution Damage in Trees [3]. Their five primary criteria for establishing cause-effect relationships were: - strong correlation; - plausibility of mechanism; - responsiveness or experimental replication; - temporality and - weight of evidence. These criteria speak for themselves. However, many claims in literature do not satisfy all of these requirements. E.g., the presence of a correlation alone is a poor indication of a cause-effect relationship. A correlation does not necessarily imply causation; see the example in §2.2. Thus it is necessary to include other evidence. As for experimental replication, a major criticism of most experimental studies of 'forest decline' is that there has been a tendency to generate symptoms under laboratory conditions and then to look for these symptoms in the field (the reverse process should be applied). Temporality means that the timing of decline should coincide with changes in air pollution levels. However, the evidence of such temporal associations is very weak in general ([1]). 2.2 Examples
An example of the problems associated with a lack of temporal association is described by [1]. Rehfuess re-examined data presented by Ulrich et al. in 1980, who proposed that the fine root biomass of trees in the Solling area of Germany had decreased in parallel with an increase in the aluminium concentration of the soil solution. Rehfuess, however, showed that the decline of the fine root biomass had occurred before the rise in aluminium concentrations and that
467 moisture stress was the primary cause of the decline. A second example is given by [4,6]. There is vast literature on the dying of silver fir in Europe, known as "Tannensterben". Visser and Molenaar analyzed ring-width data of declining silver firs in the Bavarian forest (south of Germany). They conclude from correlative statistical inferences that the drastic drop in wood production since the 1960s and the subsequent recovery in the 1980s cannot be explained by meteorological conditions. See Figure 1. They conclude t h a t the significant correlation between filtered ring widths and an index for SOz emissions points to a relation between tree growth and pollution. However, Kandler ([5]) questions their conclusions. He states that these results are due to the use of an inadequate measure of the degree of air pollution and a
"2~;g~ '~-'o°'%';°o,..-~_ [:'~;~ ~' t ~ "~ ©
" ~ " ~ % ° 4 ~ " .o, o,
5.0
tl~
t
J, t --"
t
4.0 0
\
i
3.0 0 2.0
\
1.0 I
0.0
I
~
I
~
1
J
1
L
1
~
1
1
1
l
l
1
l
1
1890
1900
1910
1920 1930 1940 1950 Year of emission
1960
I970
1980
1990
1890
1900
1910
1920 1930 1940 Year of growth
1960
1970
1980
1990
4.00
E E ~- 3.00 "0
o, 2.00 .c
rr"
1.00
0.00 1950
Figure 1. SO2 emissions in the former F.R.G. are shown in the graph at the top The lower graph shows a ring-width curve of four silver fir stands in the Bavarian forest (solid line). The curve has been filtered for weather influences. An increase or decrease during specific periods is indicated by arrows. Periods of economic events are marked by vertical lines (from [8]).
468 biased set of tree-ring data. On the other hand, Elling ([7)] analyzed silver fir in the same region and found high values of Sulphur in needles as well as a large n u m b e r of missing rings at m a n y sites. Finally, silver fir is known to be very sensitive to SO 2. There is no definite proof in the light of the criteria stated in §2.1. There is a (strong) correlation, temporality is more or less fullfilled and the mechanism is plausible. However, there is no experimental replication and the weight of evidence is not complete satisfactory (no consensus between researchers).
3. F O R E S T C O N D I T I O N AND A C I D I F I C A T I O N Summarizing the scientific knowledge on forest decline and air pollution, the criteria listed in §2.1 have been taken into account. The conclusions are listed concisely and point by point. Main references are [1], [9], [10], [11] and [12].
3.1. F o r e s t d e c l i n e • The term 'forest decline' is misleading. First, m a n y types of decline can exist within individual species. Second, most signs of decline are limited to one or two species within a specific forest. A better term would be: tree decline. • There are m a n y documented cases of n a t u r a l forest decline. Sometimes the dieback can be attributed to a single pathogen, i.e. severe insect defoliation or drought. Often the reasons for decline are unclear. • If concentrations of pollutants are high (in the order of rag/m3), direct aboveground effects on trees are likely to occur. There are numerous examples of direct injury around emission sources in the past. A well-known example is the triangle Bohemia, southern Poland and the south of the former G e r m a n Democratic Republic. Here, the atmospheric concentrations of SO 2 are far beyond critical levels. Another example is the high Ammonia emission from bioindustry in the Netherlands. The detrimental effects are clearly observable in the vicinity of farmlands. An example in Central America is the metropolitan district of Mexico City with nearby forests and other nature reserve areas. It is an area of 800 km 2 within a basin surrounded by mountains, where photochemical oxidants reach high values. An example in North America is the west coast of the U.S.A., where in some parts extreme ozone concentrations occur (maximum hourly concentrations 200-300 ppb). • Apart from examples such as those given above, no other instances of widespread, large-scale forest decline in Europe and North America have been demonstrated at the present time. There are few instances of tree decline in which air pollution plays a dominant role, as in the decline of pines in the San Bernadino Mountains (U.S.A.). 3.2. S u r v e y s of forest c o n d i t i o n • The concept of vitality classes (loss and discoloration of foliage) has generally been chosen as an indicator for forest condition. Since 1984, nation-wide surveys have been performed annually. Since 1988 temporal and spatial
469 patterns of damage classes in Europe have been analysed by the International Cooperative Programme on the Assessment and M o n i t o r i n g of Air Pollution Effects on Forests (ICP Forests) for UN/ECE. The 1993 report [12] is a cooperation with EEC. There are in total 34 participating countries. • Temporal patterns of the distribution of trees in specific damage classes show statistically significant increases over the years 1988-1992 for m a n y tree species. Data of four species, averaged for the whole of Europe, are shown in Figure 2. The upward trend in foliage loss for Scots pine, beech and Norway spruce is significant (a=0.05). There are no trends present in the discoloration data.
A
B
year
year 1993 1991 199{ 1989 1988 - - - -40
1992 1991 1990 1989 1988 40- - ~
-30
.~ 20- J ]20~
-io
Oak
---¢._ 199a 1991 1990 Ye~ 1
1988
N°~eJ"spruce ~eech SCOtspi~e
Figure 2. Percentage of trees in Europe with needle/leaf loss >25% (A) and with a discoloration >10% (B). Results of four species are shown. Data from [12].
• Figure 3A shows the result of a spatial comparison of the percentage of conifers with needle loss >25% and acid deposition. There appears to be a positive but statistically non-significant correlation between both variables (R= 0.36, tested with a=0.05). The same result is found for broadleaf trees (Figure 3B). • ICP Forests concludes, among others, that vitality alone is not suitable to prove causal relations between tree condition and levels of acidification. Also, the usefullness of vitality as a representative indicator for forest condition is questioned by many (not however by ICP Forests): a tree cannot be classed as unhealthy simply because it has lost any of its foliage. Foliage loss is often the
470 end-result of a series of changes in the tree's condition. Therefore, it is unclear what to conclude from Figures such as the ones shown here.
B
A 6O.
60-
Sso.
~50-
gb
~40nl
by
30-
.~.-
dk no
10o
o
It it
gr ua fr
~
x~
.~o-
i- -'''''f'I"
.~~ ' 1 ' ' ~ b ' ~ ' ' h u
so.~""
8
..cs de
~o.
ro lu at
~o
~
1
10-
~o
~o
lu gr
ch
acid deposition (mol I-r/ha/year)
pt
~o
~o
o
o
ltit
.
~o
.
~
dc-"~:s
ro
se" by fieSfr .
nl
..~..~"~""""'" hu ~be at
.
~o
.
.
~
.
~
~o
~o
~o
acid deposition (mol I-I+/ha/year)
Figure 3. Relation between trees with needle loss (A) or leaf loss (B) over 25% and acid deposition. The data are average values for the year 1992 and cover 29 European countries. Deposition data are calculated from the deposition of Sand N-compounds, listed in [13]; vitality data are from [12]. Abbreviations: Albania=al, Austria=at, Belgium=be, Bulgaria=bg, Byelorussia=by, Croatia=hr, Czechia/Slovakia=cs, Denmark=dk, Estonia=ee, Finland=fi, France=fr, Germany=de, Greece=gr, Hungary=hu, Ireland=ie, Italy=it, Latvia=Iv, Lithuania=It, Luxembourg=lu, Netherlands=nl, Norway=no, Poland=pl, Portugal=pt, Romania=ro, Russia=ru, Spain=es, Sweden=se, Switzerland=ch, United Kingdom=gb, Ukraina=ua.
3.3 M u l t i p l e s t r e s s • The weight of evidence for climate as having a major effect on trees and as being a cause of decline is overwhelming. • Consequently, the idea that trees are affected by combinations of stresses with air pollution being one of them, has been generally recognized and accepted (multiple-stress hypothesis). Other factors are weather conditions, as mentioned above, insect attacks and silvicultural treatment. As for the Netherlands, the systematic lowering of groundwater levels should be mentioned. • The global increase in carbon dioxide concentrations that has occurred over the past century, may have offset growth reductions attributable to pollution. The same holds for the deposition of N-compounds, which stimulates tree growth initially. Such enhanced growth behaviour has been detected by Briffa ([14]), who found increasing productivity in many western European conifers over the last century.
471 4. D O WE H A V E A L L T H E A N S W E R S ?
The public fear of dying forests has proven to be groundless. However, from a scientific point of view, many questions remain unanswered, despite an enormous quantity of publications on the topic of "acid rain". The majority of researchers believe that acidification will harm trees through acidification of soils and the subsequent damage to roots and mycorrhiza. Because soil acidification is an accumulating process, the present decrease of deposition levels in many European countries will not stop the deterioration of soils. Therefore, forest condition and soil chemistry have to be monitored in the forthcoming years. From these observations the following conclusions can be drawn: • The concept of critical loads combines soil chemistry and its effects on the root system. Although calculations of critical loads and critical levels have drawbacks at present, it is a promising approach for the forthcoming years. See e.g. [15]. • There should be international agreement on what parameters are needed to quantify forest condition. Research shows that vitality should be monitored in conjunction with other indicators of forest condition: measurements of ring widths, nutritional status of soils, needle/leaf analyses and the quality of groundwater. Also, the selection criteria of forest stands should be formulated with care (see e.g. [16]). • Experimental studies (i.e. open and closed top chambers) fail to reproduce symptoms and damage observed in the field. Such reproduction is essential for the assessment of causal relations. F u r t h e r research is needed here. • Improvement of silvicultural treatment and the right choices of species and provenances will better the condition of forests in the long term. However, to achieve this, economic gain should not be the main goal of forestry any longer. As a final r e m a r k it should be noticed that of the entire forest ecosystem forest trees seem to be the least sensitive part to air pollution. They are less sensitive than forest ground vegetation, surface waters or crops [17].
472 5. R E F E R E N C E S
1 2 3
4 5 6 7 8 9
10 11 12 13 14 15 16 17
J.L. Innes, Forest Health: Its Assessment and Status, CAB International, Wallingford, 1993. R. Koch, Beitr~ige zur Biologie der Pflanzen, 2 (1876) 277. Committee on Biologic Markers of Air Pollution Damage in Trees, Biologic Markers of Air Pollution Stress and Damage in Forests, National Academy Press, Washington DC, 1989. H. Visser and J. Molenaar, Forest Science, 38(2) (1992) 221. O. Kandler, Forest Science, 38(4) (1992) 866. H. Visser and J. Molenaar, Forest Science, 38(4) (1992) 870. W. Elling, Allgemeine Forst Zeitschrift, 48(2) (1993) 87. H. Visser, Responses of Trees to Weather Variations and Air Pollution: a Tree-ring based Approach. KEMA report 50385-MOF 90-3394/NPZR 73-7. M.R. Ashmore, J.N.B. Bell and I.J. Brown, Air Pollution and Forest Ecosystems in the European Community, CEC Environmental Research Programme, report 29, Brussels, 1990. O. Kandler, Unasylva, 44 (1993) 39. R. Schlaepfer, in: Acidification Research, Evaluation and Policy Implications (ed. T. Schneider), Elsevier, Amsterdam (1992) 27. Anonymous, Forest Condition in Europe, Results of the 1992 Survey, CECUN/ECE report, Brussels, 1993. Anonymous, Calculated budgets for airborne acidifying components in Europe, EMEP report 1/93, 1993. K.R. Briffa, in: Tree Rings and the Environment (eds. T.S. Bartholin, B.E. Berglund, D. Eckstein and F.H. Schweingruber), Lundqua report 34, Lund, (1992) 64. R.J. Downing, J. Hettelingh and P.A.M. de Smet, Calculation and Mapping of Critical Loads in Europe: Status Report 1993, RIVM report 259101003, Bilthoven, 1993. J.A.M. van den Ancker, et al., Nederlands Bosbouwtijdschrift 12 (1987) 405 (in Dutch). G. Landmann, in: Acidification Research, Evaluation and Policy Implications (ed. T. Schneider), Elsevier, Amsterdam, (1992) 383.
G.J. Heij and J. W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
STOMATAL
W.W.P.
REGULATION
Jans
and E.G.
IN F I E L D - G R O W N
473
DOUGLAS-FIR.
SteingrSver
IBN-DLO, I n s t i t u t e for F o r e s t r y P . O . B o x 23, 6700 AA W a g e n i n g e n ,
and N a t u r e R e s e a r c h The N e t h e r l a n d s .
1. I N T R O D U C T I O N
S t o m a t a l a p e r t u r e is i m p o r t a n t in r e g u l a t i n g CO 2 and H20 exchange between leaf and environment. Knowledge of s t o m a t a l regulation under various ambient conditions is i m p o r t a n t for m o d e l s in w h i c h u p t a k e of g a s e o u s air p o l l u t a n t s is estimated. N e e d l e s h i g h and low in the tree crown are s u b j e c t to d i f f e r e n t a m b i e n t c o n d i t i o n s and may also d i f f e r in s t o m a t a l r e g u l a t i o n (I). T h e r e f o r e , s t o m a t a l b e h a v i o u r was s t u d i e d on c u r r e n t - y e a r and o n e - y e a r - o l d needles high (sun adapted) and low (shade adapted) in the tree crown. B e t w e e n May and S e p t e m b e r in 1992 and 1993 2 trees w e r e c l i m b e d every 2 w e e k s to m e a s u r e gas exchange with a portable gas exchange unit (ADC, LCA-2). T r a n s p i r a t i o n rate (E), s t o m a t a l c o n d u c t a n c e (gs) and m e t e o r o logical data w e r e measured. A f t e r w a r d the shoot was cut and the shoot w a t e r p o t e n t i a l (P) was m e a s u r e d d i r e c t l y at the forest floor, u s i n g a p r e s s u r e bomb.
2 .RESULTS
In g e n e r a l the gs and E of n e e d l e s high in the c r o w n was h i g h e r c o m p a r e d w i t h n e e d l e s low in the crown. No d i f f e r e n c e was found in gs and E b e t w e e n c u r r e n t - y e a r and o n e - y e a r old needles. H i g h in the crown s t o m a t a l c l o s u r e was only found u n d e r low light conditions. No r e l a t i o n was e v i d e n t b e t w e e n gs and PAR. A clear relation was found b e t w e e n E and gs: E increased with increasing gs, but the slope was highly d e p e n d e n t on VPD (fig I). At high VPD, t r a n s p i r a t i o n was more s e n s i t i v e to c h a n g e s in gs. This was also found for o n e - y e a r old n e e d l e s and n e e d l e s low in the crown. E was also more s e n s i t i v e to c h a n g e s in gs at low P v a l u e s (more negative) (fig 2). P decreased (became more negative) with increasing t r a n s p i r a t i o n to r e a c h a m i n i m u m at ca -17 Bar,
474
w h i l e E still increased (fig 3). Evidently low P values did not induce stomatal closure, as sometimes reported in literature (2) Needles low in the crown showed much the same relation b e t w e e n E and P, although E did not exceed i000 ~Mol H20/m2.s.
SO00
"S ~.
80O0
{__
f."
mo
1"o
o •
•
~
Q
3:
"5 • ,..,% I000
~,
o
,~
o
I| I '5
oo
oee
•
. ~4..%~#~ 0 0
,
i |
1
$
0
1
1
o |
Figure
A 3
1
•
,
•
8
4
4
~_0
•
o~'o~_ o o
.lm
+ 6
o
-15-0
Figure 2
•
<-lS
oo
o~
omooo
oo
0 o
| 0
a¢ (m~)
G8 (ramie) •
o,
o ~o
o
, 4
|
'
bo o "10 i 0
lOOO
dlOOO
Jooo
E ( p M . I H.O,~flt.8)
Figure 3
Figure 1: The relation between gs and E under d i f f e r e n t VPDranges of c u r r e n t - y e a r needles high in the tree crown. (VPD (Pa): 1<500; 2=500-1000; 3=1000-2000; 4=2000-3000; 5>3000). Figure 2: The relation between gs and E under d i f f e r e n t P ranges of c u r r e n t - y e a r needles high in the tree crown. (P (Bar): >-15 and <-15). Figure 3: The relation between E and P of c u r r e n t - y e a r needles high in the tree crown.
3 • CONCLUSIONS
1 T r a n s p i r a t i o n rates were more sensitive to gs at high VPD ranges, and low shoot water potentials. 2 Under these conditions, shoot water potentials did not induce stomatal closure, high or low in the crown. 3 No d i f f e r e n c e s were found in stomatal r e g u l a t i o n between needles high and low in the crown and between c u r r e n t - y e a r and o n e - y e a r - o l d needles.
4 • REFERENCES
1 W. Jans and E. Steingr~ver. IBN, Wageningen, The Netherlands, D o r s c h k a m p Report 695, 1992. 2 C. Tan and T. Black. B o u n d a r y - L a y e r M e t e o r o l o g y i0 (1976).
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
CARBON PARTITIONING
E. S t e i n g r ~ v e r ,
475
IN D O U G L A S - F I R
M. P o s m a and W. Jans
IBN-DLO, I n s t i t u t e for F o r e s t r y and N a t u r e R e s e a r c h P.O. B o x 23, 6700 AA, W a g e n i n g e n , The N e t h e r l a n d s .
I.
INTRODUCTION
A f i e l d gas e x c h a n g e laboratory, with specially developed b r a n c h a s s i m i l a t i o n c h a m b e r s o p e r a b l e t h r o u g h o u t the year, was i n s t a l l e d in a 3 2 - y e a r old D o u g l a s fir s t a n d in the c e n t r a l p a r t of the N e t h e r l a n d s . D u r i n g 1992 p h o t o s y n t h e s i s of c u r r e n t y e a r and 1 - y e a r - o l d n e e d l e s from 5 c r o w n levels w a s m o n i t o r e d .
2.
RESULTS
Data on t o t a l n u m b e r of combined with data on c o n t r i b u t i o n of d i f f e r e n t t o t a l net CO 2 u p t a k e over
n e e d l e s and n e e d l e s u r f a c e area w e r e net assimilation to determine the m o n t h s and d i f f e r e n t c r o w n levels to 1992 (Figure i).
14
T o t a l net CO 2 u p t a k e w a s not evenly distributed over the crown. Net assimilation rate 0 4 increased with tree ° . h e i g h t and c u r r e n t y e a r ~l_J. needles had higher -| assimilation rates j,. Fs..,. , P , . A Y J U . J U , A U . . s P OCT . O r . S O compared to 1-year old needles. Needles around 2*1 the c a n o p y c l o s u r e p o i n t (1 eve 1 3) were responsible for more Figure l:Total net CO 2 u p t a k e per t h a n 40% of the total m o n t h and per c r o w n level in 1992. uptake in 1992, due to the large number of needles and a s s o c i a t e d needle surface area at this level. A l t h o u g h 33% of t o t a l n e e d l e s u r f a c e area was a s s o c i a t e d w i t h 1£
,.
1[
476 c r o w n l e v e l s 1 + 2, CO 2 u p t a k e in 1992.
these
levels
contributed
only
3%
to
total
A c a r b o n b u d g e t for 1992 is p r e s e n t e d in T a b l e 3. A total 31.7 M g C per ha was t a k e n up by the stand. As no m a j o r f l o w e r i n g o c c u r r e d d u r i n g the g r o w i n g season, the a l l o c a t i o n of c a r b o n to g r o w t h and p r o d u c t i o n of cones was n e g l e c t e d . Of the t o t a l c a r b o n t a k e n up by the stand, 47% was u s e d for b i o m a s s increment: 37% a b o v e g r o u n d and 10% b e l o w ground. As n e e d l e r e s p i r a t i o n is a l r e a d y i n c l u d e d in the total c a r b o n uptake, 53% of t o t a l c a r b o n u p t a k e by the s t a n d was a s s o c i a t e d w i t h g r o w t h and m a i n t e n a n c e r e s p i r a t i o n of fine r o o t s and t h a t of w o o d y tissues in branches, stems and c o a r s e roots. The a m o u n t of c a r b o n u s e d for m y c o r r h i z a and in e x u d a t i o n m i g h t be important, but was not d e t e r m i n e d .
Table
3:
Carbon budget
Net u p t a k e Needles Branches Stem Roots Total
for 1992
in Mg C p e r ha.
Biomass
increment
31.7,
3.6 0.8 7.4 3.2 ** 15.0
31.7
Respiration
16.7
* Needle respiration included ** F r o m O l s t h o o r n (1991)
3. C O N C L U S I O N S
-
M a j o r p a r t of total CO2 was t a k e n up in June and J u l y and in the w i n t e r m o n t h s no net CO 2 u p t a k e was found 40% of total CO2 u p t a k e was a s s o c i a t e d w i t h the n e e d l e s a r o u n d the c a n o p y c l o s u r e p o i n t 47% of total CO 2 u p t a k e was used for b i o m a s s i n c r e m e n t 53% of total CO2 u p t a k e was a s s o c i a t e d w i t h r e s p i r a t i o n of fine roots and w o o d y t i s s u e s
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BE All rights reserved.
477
Decreasing concentration of air pollutants and the rate of dry and wet acidic deposition at three forestry monitoring stations in Hungary L. Horv~ith and Gy. Baranka Institute for Atmospheric Physics, H-1675 Budapest, P.O.Box 39, Hungary and
E. Gy. F/ihrer Forest Research Institute, H-I023 Budapest, Frankel Le6 u. 44, Hungary
Abstract Concentration and dry + wet deposition of acidic atmospheric pollutants were determined in three Hungarian forestry monitoring stations between 1988 and 1992. Concentration of most important acidic compounds (sulfur and nitrogen dioxide) does not reach the critical level therefore forest decline as consequence of direct effects (foliar uptake through stomata) may be excluded. Total (dry + wet) deposition of acidic pollutants are 187 mg H ÷ m-2 yr-~ as an average of three stations. The acidic load at Hungarian forestry stations frequently exceeds the critical value recommended by international organizations. Though concentration and deposition of atmospheric acidic compounds has considerably decreased in the recent years, futher monitoring of acid deposition in forests is necessarry because the accumulation effects of acid load to forest soil.
1 Introduction Air pollution affect directly and/or indirectly the forest health. Air pollutants through stomata initiate biochemical processes which can affect the leaves directly. Indirect effect of acidic substances appears in the long term acidification of forest soil. Acidification of forest soils, forest decline, furthermore the fact that Hungary belongs to moderately polluted countries in Europe support the need for continuous monitoring of the rate and trend of atmospheric acidic deposition in Hungarian forests. Forest Research Institute and Institute for Atmospheric Physics of Hungarian
478 Meteorological Service have started a monitoring program in 1988 at the three forestry monitoring stations. The aim of this paper is to summarize the result of the first five years of the monitoring program (1988-1992), to calculate the concentration and deposition of acidic compounds for forests and compare them with standards and recommendations.
2 Measuring programs The three forestry monitoring stations are situated in pine forests at the west (Farkasfa:~,=45°55 ', ~,=16°18 ', H=312 m masl) and the middle of Hungary (K-puszta: ~=46°58 ', ~,=19°33 ', H=126 m masl) as well as in the M~itra-Mountains, north-east of Hungary (Nyirjes: ~=47°57 ', ~,= 19058 , H=560 m masl). Forestry stations were located on openings in the forest. The sampling period is 24 hours for the most important acidic gaseous pollutants (sulfur dioxide, nitrogen dioxide, nitric acid and ammonia) and for the aerosol particles (their ammonium, sulfate and nitrate contents are determined). Chemical composition of precipitation is also measured on the basis of monthly wet-only sampling. Detailed description of sampling and analytical methods including the estimated average dry deposition velocities can be found in Horv~ith [1] or Pais and Horv~ith [3]. Calculation of acid deposition rate is described by Horv~ith [1]. It can be respected as the upper limit of acidity. Therefore, instead of uncertain net (acidic and alkalinic) deposition we prefer to use the term of possible highest rate of acid deposition.
3 Concentration of air pollutants Direct effects of acidic air pollutants are related to their high atmospheric concentrations. There are different approaches to estimate the air quality standards for forest ecological systems. According to WHO [6] the harmful level of sulfur dioxide and nitrogen dioxide for forests (for the two most important air pollutants from the point of view of acidification) is 30 /~g m3 as a yearly average. These figures correspond to Hungarian standards determined for "particularly protected" areas (MSZ [2]). Proposed critical loads for forests are 20 and 30/~g m-3 as a yearly average for sulfur dioxide and nitrogen dioxide, respectively (V~allyay et al. [4]). Yearly averages for sulfur dioxide and nitrogen dioxide as well as the variation of mean monthly values can be seen in Fig. 1. The yearly mean values of pollutants do not reach the harmful level. From west to east concentrations of sulfur and nitrogen dioxide increase.
479
I~gm
8o~I
SO
i
2
,
K-puszta
ii ...... ] Nyfrjes ..J
i
5o I
:i,!
II
40~!ii
~
Farkasfa
:i
3o !I!
It if!
,
il
i,,' ," ti,, LIL
201'I
I[
ii
~oliI~,,
ii
o ,~,~ !l!h!
,~ii;:;
II.
ol i
";,
ilii
,
i
1988
~gm -3
25-,
1989
1991
1990
1992
NO2
K-puszta
i
, .
Nyirjes
201
5ii i
i~.
ii ~
ii!i''
i l!!illi
5!
ol
d
~
1988
1989
h
!~1! 1990
1991
i
lli!,i,! 1992
Figure 1. Variation of the mean monthly concentration of sulfur dioxide and nitrogen dioxide
Further important information is that higher sulfur dioxide and nitrogen dioxide levels occur during the winter half-year, out of the vegetation period. For deciduous forests the foliar uptake of gases is possible only during the vegetation period (practically during summer half-year, from April to November). In the case of coniferous forest the uptake of sulfur and nitrogen dioxide is minimized in the winter half-year because of closed stage of stomata.
480
4 Deposition of air pollutants In contrast with high atmospheric level of pollutants dry and wet deposition of acidic substances (sulfur and nitrogen compounds) may affect the forest health by indirect processes through accumulation in the forest soil. The average rate of total deposition (for the period of 1988-92 as an average of three stations) is 187 mg H ÷ m 2 yr ~. The share of dry and wet deposition is approximately the same. However, in the summer half-year the wet, in winter half-year the dry deposition dominates (Fig. 2). Dry deposition has a winter, wet deposition has a summer peak. Consequently, there is no expressed annual variation in the total deposition. mg H+rn2 mo1 20-
deposition total
/
I
/ / '~'"""--.~
.
10
..-/"
• • 1
,I
//
/
,"
,
\
//
\
X
/ \
...-"'
\
/
,
"
oT;'"
"
\
/
X
.~-"
.~
/
)
wet
/ f\
\
/
/
.
/
/
~ "~
...
•
" :~. " , , , ,
"
..
dry -.
...-
m
..+ . . . . . . . . .
/
---=-I--
1
F
2
F - - ~ I
3
i
4
5
T ..........
I ........
fi 7 month
F ......
8
~......
9
I ..........
10
1. . . . . . . . .
11
F. . . . .
12
Figure 2. Annual variation of dry, wet and total acidic deposition as an average of the three forestry stations.
The ratio of sulfur to nitrogen compounds in the total deposition is 42 and 58 %. The relative importance of sulfur compounds is continuously decreased from 46 % to 36 % during the 5 years. It is in agreement with our earlier results (Horv~ith, [ 1]) suggesting that N/S tatio is increasing in the acidic deposition. The average yearly acidic depositions of Farkasfa and K-puszta stations are similar, 164 and 172 mg H ÷ m2 yr~, which can be regarded as a "background" acid deposition for
481 Hungary. For Nyfrjes station the acidic deposition is higher, 224 mg H + m 2 yr -~ indicating local pollution effect. The yearly deposition figures for the tree stations generally do not exce~ the proposed critical load for Hungary. The critical load of total acid deposition is 280 mg H ÷ m"2yr ~ (VLrallyay et al. [4]). According to the recommendation of UN and ECE (RIVM [5]), Hungary is divided into different grids according to sensibility of the given area. The critical loads for the 4 different areas for Hungary are 20-50; 50-100; 100-200 and > 200 mg H ÷ m 2 yr 1. Lowest figures are generally determined for western and northern part of Hungary. Taking into consideration these recommendations we can see that in the majority of the cases the critical load is exceeded.
5 Trends in concentration and deposition Concentration of most important acidic compounds (sulfur and nitrogen dioxide) as well as total (dry+wet) deposition of pollutants have decreased during the 5 years of investigation. This result is summarized in Fig 3. The atmospheric level of pollutants and the rate of acidic deposition have reached to the half of the value measured in 1988. It is probably due to reduced sulfur and nitrogen emission in Hungary and in the neighbouring countries as a consequence of change in economical structure of East European region. -3 I~g m 15
+
H
SO 2 , NO 2
-2 -1 mg m yr -300
200
10
"'"-.
N\\ ".
".. •~...
-N N N NNN
j 1 / 1 tit ~
~
.., ,, .0,0'~.,,\ N ~ \ " \ N \ ~
'".....
NO 2
\\
1OO
". " . "'.. "'..
0
19'88
19'89
19'90
19'91
19'92
Figure 3. Variation of the concentration of air pollutants and the rate of total acidic deposition
482 This fortunate tendency suggests that we can not probably face the danger of direct forest decline in the near future caused by increased level of pollutants. In spite of the decreasing deposition rate the critical loads recomended by international organizations are frequently exceeded. For this reason the continuous monitoring of acid deposition is needed for the effects of pollution may accumulate in the forest soil (e.g. accumulation of nitrogen loading, mobilization of heavy metals, decrease of buffer capacity against acid stress) causing long term indirect damages in forest.
References Horv~ith, L., 1989: Measurement of atmospheric acid deposition in Hungary (in Hungarian), Periodicals of the Hungarian Meteorological Service No. 65, Budapest.
.
0
.
11
Q
MSZ, 1990: Requirements of cleanness of ambient air (Hungarian Standard), MSZ 21854. Pais, I. and Horv~ith, L., 1990: Atmospheric acidic deposition and its environmental effect in Hungary. Advances in Environmental Science (ed.: D.C. Adriano), University of Georgia, USA V~irallyay, Gy., Fekete, K., Fiihrer, E., Gopcsa, E., Haszpra, L. and V~irkonyi, T., 1992: System for Rural Air Quality Standards. Institute for Environmental Management, Budapest, Hungary, (manuscript). RIVM, 1991: Mapping critical loads for Europe. CCE Technical Report No. 1. National Institute of Public Health and Environmental Protection, Bilthoven, The Netherlands. WHO, 1985: Air Quality Guidelines-Ecological Effects of Air Pollutants. World Health Organization ICP/CEH 902/m 71(S).
G.J. HeO and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BE All rights reserved.
483
THE CHARACTERISTICS OF ACID PRECIPITATION IN SOUTHERN CHINA Yuhua Bai a and Xiaoyan Tang 2 aDepartment of Technical Physics, 2The Centre of Environmental Sciences Peking University, Beijing, 100871, P.R. China
EXTENDED ABSTRACT Research on acid rain started in 1979 in several cities of China. This paper presents the results of precipitation and gaseous pollutants monitored in Guangdong and Guangxi provinces during spring 1988 and spring 1989. Four mountain sites (380-2140 m altitude) and five city monitoring sites have been installed. An aircraft campaign at 500-4000 m altitude was also performed. The precipitation in these areas was seriously acidified. The percentage of samples with pH less than 5.6 was 95, 87, 88 and 99.5% for upper air cloud, mountain rain, mountain cloud and rainwater collected at the city stations, respectively. The percentages of samples with pH less than 5.0 was 88, 84, 80 and 96% for the type of samples mentioned above, respectively [1]. The acidity was closely related to specific meterological conditions. With the northern cold air front, the percentage of samples with pH less than 5.0 was almost quantitative. The average of total cation or total anion concentration is much higher as compared to Lijiang background site (Table 1). Sulphate and nitrate were the major contributors to acidity, with lower sulphate-to-nitrate ratio as compared to South-western area (Chongqing city, see also Table 1). The major acidic compound is sulphuric acid originating from sulphur dioxide. The sulphur dioxide average concentration in the cities ranged from 0.07 to 0.13 mg/m3 (1988). The concentrations of secondary pollutants such as 03, H202, HNO3, aldehydes and fine particles were rather high, resulting in high photochemical activity and strong oxidising capacity of the atmosphere. The average 03 and HNO 3 concentrations in the cities varied between 29-114 and 0.05-0.21 ppbv, respectively. The H202 concentration in precipitation collected at the city sites varied between 0.4-60 Iamol/l, depending on the SO2 concentration in air and S(IV) concentration in precipitation. High H202 concentrations are only found under low SO2 atmospheric condition. This indicates that H20 z is the principal oxidant for precipitation during spring. For the precipitation samples, high NH4 concentrations were observed which were most likely caused by high NH 3 concentrations in air (ranging from 6-11 [ttg/m3). Below-cloud scavenging modelling showed that ammonia was the most important compound impacting pH and total sulphur concentration of precipitation. With respect to the sources of acid precipitation, coal with relatively high sulphur content, ash content and low caloric value was the major energy resource used in this area. In total about 0.87 X 109 kg SO z and 0.65 x 109 kg smoke dust was emitted in this area in 1988. These emissions were mainly responsible for the acid precipitation. The precipitation
484 collected on the mountain sites was acidified mostly during northern cold front passages. Therefore, long-range transportation of acidifying pollutants accumulated in the northern cold front also contributed to acid precipitation in the Southern provinces. 1. Wang Meirong, ACTA Scientiae Circumstantiae, 12 (1) (1992) 37. 2. Sheng Peixuan, Mao Jietai et al. ib, 12 (1) (1992) 16. Table 1. Cat- and anion concentration (l.teq/1) and corresponding ratios site
SO42-/ani -
5042-
on
/NO3-
Avg. H÷
sum (anion)
sum (cation)
NH4*/ca -
upper air Mt. rain Mt. cloud ground rain
135 81.0 98.4 138
356 177 531 347
495 250 595 418
30.0 29.5 50.9 33.8
34.6 26.0 19.5 14.7
40.3 56.5 70.4 65.3
1.2 2.5 3.9 5.1
Chongqing Lijiang
129 10.0
216 13.7
270 23.9
30.0 23.8
1.1 13.9
78.1 59.1
71.0 4.2
tion
NO3-/anion
(~)
(%)
(~)
G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? © 1995 Elsevier Science BV. All rights reserved.
485
THE RESPONSE OF PEAT WETLAND METHANE EMISSIONS TO TEMPERATURE, WATER TABLE AND SULPHATE DEPOSITION D. Fowler, J. MacDonald, I.D. Leith, K.J. Hargreaves & R. Martynoga Institute of Terrestrial Ecology, Bush Estate, Penicuik, Midlothian, Scotland. Sixty peat monoliths, each 30 cm in diameter and 40 cm deep, collected in the blanket bogs of Sutherland were installed in open-top chambers at ITE. A dynamic CH 4 flux measuring system capable of measuring emission or deposition over 15 to 30 minute periods using a flame ionisation detector was developed and installed at the site. The initial fluxes showed a clear positive response to temperatures of the peat in the surface 10 cm (temperatures were continuously monitored at 6 depths in the peat down to 40 cm at which the surface diurnal cycle in temperature was not detectable). Following the initial measurements a systematic study of the temperature response of 24 monoliths, using the natural changes in temperature of the peat with changing season and weather took place during the spring and summer of 1993. The monoliths were divided between pools (water table maintained at the surface), lawns (water table maintained 2 cm below the surface) and hummocks (water table maintained 15 cm below the surface). The temperature responses of the pool monoliths range from 3.6 to 14.9/~mol m "2 h 1 °C'1 with fluxes (normalised to 15°C) ranging from 42.7 to 183.5 /~mol m "2 h "~ with the largest fluxes from monoliths with the largest leaf area of higher plants. An example of the quasilinear temperature response for one of the pool monoliths is presented as Figure 1 and is typical of the pool monoliths studied.
300 250
Activation energy = 95.78 KJ mo1-1
..o ..o" o
"-"
.*•
200 .•'°*
-~
O
150 • -*•°~"
loo
0
• •.,0 °°'°
°. .° . . . .
so ° .o
-50
'1
i
!
!
i
!
!
!
!
8
10
12
14
16
18
20
22
24
26
Temperature (012) Figure 1: Effect of temperature on m e t h a n e e m i s s i o n f r o m a p o o l core
Emissions of methane from the hummock monoliths were an order of magnitude smaller than those of the pool and lawn monoliths (Table 1). This was attributed to
486 water table depth, hummock monoliths having a water table depth of approximately 15 cm, providing an oxidising layer through which the methane has to pass before reaching the atmosphere. The effects of water table and temperature fluctuations have important consequences for the response of methane flux to climate change and any feedback mechanisms that may occur.
n
Temperature response/zmol
SD
Mean Flux (15°C)
SD
Q1o between 5°C and 15°C
SD
m-2h-1 Pool
8
8.91
3.7
111.21
32.3
3.90
1.68
Lawn
6
11.52
3.8
103.18
38.1
6.90
5.12
Hummock
5
6.48
5.4
8.48
9.3
1.82
1.62
Table 1: Effect of t e m p e r a t u r e and water table on methane emission
More detailed and precise studies of the temperature response of CH4 emissions are provided by work in controlled environment chambers (CONVIRONS). The same monoliths as those used above were used to obtain temperature responses of CH4 emission with constant light, water table and relative humidity over long and short periods. Mean fluxes (normalised to 15°C) compare well with the fluxes from the peat monoliths in the OTC's. The results above show a clear exponential response with Qlo values in the range 2.10 - 2.82 and activation energies in the range 51.8- 72.8 in good agreement with values in the literature.
400
QI0 = 2.82 350
Activation energy ffi 72.77 kJ mol"1
...
-
300 .e~
250-
~
200-
,,,.
~
150-
4
100 50-
i
i
i
!
i
I
I
0
$
10
15
20
25
30
Temperature (°C)
I
35
J
I
i
I
#
I
'
3.95 4.00 4.05 4.10 4.15 4.20 4.25 4.30 4.35 1/RT (tool j-l) ,10-4
Figure 2: Temperature response of methane emissions from monoliths maintained in CONVIRONS
487 THE RESPONSE OF METHANE EMISSIONS TO INPUTS OF SULPHUR Large areas of European and North American peat wetlands receive inputs of sulphur as a consequence of the long range transport of the products of fossil fuel combustion in industrialised countries. These inputs may influence the emission of methane by providing alternative electron acceptors for the microorganisms and an increase in the redox potential in the surface layers of the peat. The effect of atmospheric inputs were simulated by applying a 'wet deposition' equivalent to 100 kg ha "1 sulphur in an aqueous "-.. • • solution of (NH4)~SO4 and • "'-.o Na2SO4. An equivalent amount of t -... -4) m ,, • "'-.,Q "-.. deionised water was added to the "'"6--. ""'--. 8ocontrols and the methane flux I ?o",... "'..... • was monitored using static and | ""-. o'"-.. 6Odynamic methods. The inputs used in this study were large and i $0° ! at the upper range of the I~ 40atmospheric input of sulphur in 30 u | ! ! v | | 2 4 6 I i0 12 14 ,, the U.K. However it is an O 33he (d~a) appropriate amount for an initial Figure 3: Effect of sulphate application on methane emission in an s t u d y since a substantial fraction open top chamber of the annual deposit may occur on a single day. Following the simulated input of sulphur the emission of methane declined by approximately 50% within two weeks (Figure 3). It ~0 is intriguing that the reduction in ~0 "IROI. flux should be 50% as it has been ,j shown that > 80% of the methane ! (NH4)2SO 4 flux is produced by the layers within 20 cm of the surface. One explanation could be the ISO 1 prescence of methanogens which utilise non-competitive so w I , , | w , w o s lO is 2o 2s 3o ~ 4o 45 substrates and are therefore not 33m c (d~s) affected by the increased activity Figure 4: Effect of sulphate application on methane emission in a of the sulphur reducing bacteria.
j
CONVIRON
This Page Intentionally Left Blank
ANNMI LIST OF PARTICIPANTS
This Page Intentionally Left Blank
491
J.M.M. Aben National Instiute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742938 Fax: 31-30-287591
A.W.F. van Alphen Ministry of Housing, Spacial Planning and the Environment (ipc 64) P.O.Box 30945 2500 GX DEN HAAG Tel: 31-70-3394677 Fax: 31-70-3391313
W.A.H. Asman National Environmental Research Fredriksborgvej 399 4000 ROSKILDE Denmark
Y. Bai Netherlands Energy Research Foundation P.O.Box 1 1755 ZG PETTEN The Netherlands
Tel: 45-42371137 Fax: 45-46301214
Tel: 31-2246-4155 Fax: 31-2246-3488
Gy. Baranka Institute for Atmospheric Physics, Department for Analysis of Atmospheric Environment Gilice Ter 39 1181 BUDAPEST Hungary Tel: 36-1-2900163 Fax: 36-1-2904174
C.G.B. Barquero CIEMAT, Environment Institute Avda Complutense 22 28040 MADRID Spain
U. Bertills Swedish Environmental Protection Agency 171 85 SOLNA Sweden
K.H. Bishop Department of Forest Ecology Swedish University of Agricultural Sciences 90183 UMEA Sweden
Tel: 46-8-7991502 Fax: 46-8-283008
Tel: 46-90-166625 Fax: 46-90-167750
A. Bleeker National Institute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742938 Fax: 31-30-287531
Dr. R. Bobbink University of Nijmegen, Department of Ecology, Research Group Environmental Biology Toernooiveld 1 6525 ED NIJMEGEN The Netherlands Tel: 31-80-652477 Fax: 31-80-553450
F.C. Bosveld Royal Netherlands Meteorological Institute P.O.Box 201 3730 AE DE BILT The Netherlands
A.W. Boxman University of Nijmegen, Department of Ecology, Research Group Environmental Biology P.O.Box 9010 6500 GL NIJMEGEN The Netherlands Tel: 31-80-652860 Fax: 31-80-553450
Tel: 31-30-206911 (386) Fax: 31-30-210407 A. Branding Projektzentrum Okosystemforschung, Christian-Albrechts-Universit~it Schauenburgerstrasse 112 D-24118 KIEL Germany Tel: 49-431-8804030 Fax: 49-431-8804083
Tel: 34-1-3466062 Fax: 34-1-3466121
S. Braun Institute for Applied Plant Biology Sandgrubenstrasse 25 CH-4124 SCHONENBUCH Switzerland Tel: 41-61-4813224 Fax: 41-61-4813436
492
Dr. M. Bredemeier Forschungzentrum Waldtikosysteme der Universittit G6ttingen Habichtsweg 55 37075 GOTTINGEN Germany Tel: 43-551-393512 Fax: 43-551-393762
A.H.M. Bresser National Institute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-743756 Fax: 31-30-252066
J.R. Brook Atmospheric Environment Service of Canada, Air Quality Processes Research Division 4905 Dufferin Street DOWNSVIEW, Ontario M3H 5T4 Canada Tel: 1-416-7394916 Fax: 1-416-7395708
Mrs.Dr. E. Brtiggeman Institut ftir Troposphtirenforschung e.v. Permoserstrasse 15 04303 LEIPZIG Germany
J. Burkhardt Bayreuth Institute of Terrestrial Ecosystem Research, Department of Climatology, Bayreuth, Germany Institute of Terrestrial Ecol9gy Edinburgh PENICUIK, Midlothian EO26 0QB United Kingdom Tel: 44-31-4454343 Fax: 44-31-4453943
H. Casado Universidad Pais Vasco Facultad de Farmacia Ci Marques de Urovijo s/n VITORIAA 01006 Spain Tel: 34-945-140005 Fax: 34-945-130756
E.B. Cowling North Carolina State University Forestry Department Box 8002 RALEIGH, NC 27695-8002 USA Tel: 1-919-5157564 Fax: 1-919-5151700
R. Dlugi Meteorological Institute University Theresienstrasse 37 80333 MUNICH Germany Tel: 49-89-188185 Fax: 49-89-182260
H.F. van Dobben IBN-DLO P.O.Box 23 6700 AA WAGENINGEN The Netherlands Tel: 31-8370-77936 Fax: 31-8370-24988
G.P.J. Draaijers National Institute of Public Health and Environmental Protection, Air Research Laboratory P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742824 Fax: 31-30-287531
Th.A. Dueck AB-DLO P.O.Box 14 6700 AA WAGENINGEN The Netherlands
J.H. Duyzer TNO Institute of Environmental Sciences P.O. Box 6011 2600 JA DELFT The Netherlands
Tel: 31-8370-75918 Fax: 31-8370-75787
Tel: 31-15-696263 Fax: 31-15-616812
L.J. van der Eerden AB-DLO P.O.Box 14 6700 AA WAGENINGEN The Netherlands
N.D. van Egmond National Institute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-743100 Fax: 31-30-292677
Tel: 31-8370-75919 Fax: 31-8370-75787
Tel: 49-341-2352535 Fax: 49-341-2352361
493
A.J. Elshout KF_aMA P.O. Box 9035 6800 ET ARNHEM The Netherlands
Tel: 31-85-562381 Fax: 31-85-515022
B.G. van Elzakker National Institute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742820 Fax: 31-30-287531
D. Encinas Universidad Pais Vasco Facultad de Farmacia Ci Marques de Urovijo s/n VITOR/AA 01006 Spain Tel: 34-945-140005 Fax: 34-945-130756
J.W. Erisman National Institute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742824 Fax: 31-30-287531
R.C. Ferrier Macaulay Land Use Research Institute Craigiebuckler ABERDEEN AB9 2QJ United Kingdom
S. Fleischer National Swedish Research Program on Acid Waters, County Administration Board S-30186 HALMSTAD Sweden
Tel: 44-224-318611 Fax: 44-224-311556
Tel: 46-35-132206 Fax: 46-35-158640
D. Fowler Institute of Terrestrial Ecology Bush Estate PENICUICK, Middlothian United Kingdom
P.H. Freer-Smith Forestry Authority, Research Division Alice Holte Lode WRECCLESHAM, FARNHAM, Surrey G410 4LH United Kingdom
Tel: 44-31-4454343 Fax: 44-31-4453943
Tel: 44-420-22255 Fax: 44-420-23653
A.P. Gaasbeek Shell Nederland B.V. Hofplein 20 3032 AC ROTI'ERDAM The Netherlands Tel: 31-10-4696594 Fax: 31-10-4696605
L.N. Ganzeveld Agricultural University Wageningen Biotechnion P.O.Box 8129 6700 EV WAGENINGEN The Netherlands Tel: 31-8370-85096 Fax: 31-8370-84457
T. Georgidas CCNR-FISBAT Via Gobetti 104 BOLOGNA 40129 Italy
S. Godzik Institute for Ecology of Industrial Areas Kossutha 40-833 KATOWICE Poland
Tel: 39-51-6399586 Fax: 39-51-6399652
Tel: 48-3-1546031 Fax: 48-3-1541717
C.A. Gough Stockholm Environment Institute University of York Heslington YORK Y01 500 United Kingdom Tel: 44-904-432895 Fax: 44-904-432898
J.J.M. van Grinsven National Institute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-743397 Fax: 31-30-292897
494
J.E. Groenenberg SC-DLO P.O.Box 125 6700 AC WAGENINGEN The Netherlands Tel: 31-8370-21336 Fax: 31-8370-24812
H.J. Haanstra Ministry of Housing, Spacial Planning and Environment P.O.Box 30945 (ipc 630) 2500 GX DEN HAAG The Netherlands Tel: 31-70-3394837 Fax: 31-70-3391289
J. van Ham TNO-MW P.O.Box 6011 2600 JA DELFT The Netherlands
H. Hannessen Regionale Inspectie MilieuhygiSne Noord-Brabant P.O.Box 6195 5600 HD EINDHOVEN The Netherlands
Tel: 31-15-696885 Fax:
Tel: 31-40-652911 Fax: 31-40-460902
B.J. Heij National Institute of Public Health and Environmental Protection (postbak 59) P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742970 Fax: 31-30-251932
B.B. Hicks National Oceanic and Atmospheric Administration, Air Resources Laboratory 1315 East West Highway, R/E/AR SILVER SPRING MD 20910 USA Tel: 1-301-7130684 Fax: 1-301-7130119
P.R. Hilgen IKC Coordination of Dutch Forest Health Survey P.O.Box 30 6700 AA WAGENINGEN The Netherlands Tel: 31-8370-74874 Fax: 31-8370-74930
J.A. Hoekstra National Institute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742914 Fax: 31-30-252973
P. Hofschreuder Agricultural University Wageningen Biotechnion P.O.Box 8129 6700 EV WAGENINGEN The Netherlands Tel: 31-8370-85096 Fax: 31-8370-84457
J. Hogenkamp National Institute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742379 Fax: 31-30-287531
L. Horvath Institute for Atmospheric Physics, Department for Analysis of Atmospheric Environment Gilice Ter 39 1181 BUDAPEST Hungary Tel: 36-1-2900163 Fax: 36-1-2904174
L.W.A. van Hove Wageningen Agricultural University, Department of Air Quality P.O.Box 8129 6700 EV WAGENINGEN The Netherlands Tel: 31-8370-82684/8208 Fax: 31-8370-84457
W.W.P. Jans Institute for Forestry and Nature Research (IBN-DLO) P.O.Box 23 6700 AA WAGENINGEN The Netherlands
A. Jenkins Institute of Hydrology Crowmarsh Gifford WALLINGFORD, Oxon OX10 8BB United Kingdom
Tel: 31-8370-7789 Fax: 31-8370-24988
Tel: 44-491-838800 Fax: 44-491-832256
495
J. Jones Climatic Research Unit NR4 7TJ NORWICH United Kingdom
D. KaUweit Umweltbundesamt Bismarckplatz 1 14193 BERLIN Germany
Tel: 44-603-593161 Fax: 44-603-507784
Tel: 49-30-89032839 Fax: 49-30-89032282
N. Kanchana Kanti University of Bayreuth Bit/Sk, Abt. Klimatologie 95440 BAYREUTH Germany
P.E. Kauppi Finnish Forest Research Institute Unioninkatv 40a SF-00170 HELSINKI Finland
Tel: 49-921-7857172 Fax: 49-921-7857299
Tel: 358-0-857051 Fax: 358-0-625308
V.G. Keizer Ministry of Housing, Spacial Planning and the Environment (ipc 64) P.O.Box 30945 2500 GX DEN HAAG The Netherlands Tel: 31-70-3394363 Fax: 31-70-3391313
Dr. J. van der Kooij N.V. SEP Utrechtseweg 310 6800 ET ARNHEM The Netherlands
L. Koster Shell Nederland B.V. Hofplein 20 3032 AC ROTTERDAM The Netherlands
Dr. G.H.M. Krause Landesumweltamt NRW Wallneyerstrasse 6 45133 ESSEN Germany
Tel: 31-10-4696035 Fax: 31-10-4696605
Tel: 49-201-7995215 Fax: 49-201-7995446
J. Kros DLO Winand Staring Centre P.O.Box 125 6700 AC WAGENINGEN The Netherlands
J. Kubuznfikov~i Insitute of Landscape Ecology, Czech Academy of Sciences Na S~idk~ich7 CESKE BUDEJOVICE Czech Republic Tel: 42-38-817516 Fax: 42-38-45719
Tel: 31-8370-74366 Fax: 31-8370-24812 R. Lenz GSF/PVC P.O. Box 1129 85758 OBERSCHLEISSHEM Germany Tel: 49-89-31872954 Fax: 49-89-31873369 P. Mayerhofer IER Hessbruhlstrasse 49 70565 STUTTGART GermanY Tel: 49-711-7806149 Fax: 49-711-7803353
Tel: 31-85-721473 Fax: 31-85-430858
H. Marseille Ministry of Housing, Spacial Planning and the Environment (ipc 64) P.O.Box 30945 2500 GX DEN HAAG Tel: 31-70-3394363 Fax: 31-70-3391313 M.G. Mennen National Institute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742389 Fax: 31-30-287531
496
W. Mill Institute for Ecology of Industrial Areas UI. Kossuthag 40-833 KATOWICE Poland
G.M.J. Mohren Institute for Forestry and Nature Research P.O.Box 23 6700 AA WAGENINGEN The Netherlands
Tel: 00-48-3-1546031 Fax: 00-48-3-1541717
Tel: 31-8370-95322 Fax: 31-8370-24988
G.J. Monteny Institute for Agricultural and Envionmental Engineering (IMAC-DLO) P.O.Box 43 6700 AA WAGENINGEN The Netherlands Tel: 31-8370-76580 Fax: 31-8370-25670
T. Paces CGU Kl&ov 3 118 21 PRAHA Czech Republic
K. Peters University of Bayreuth Bit6k, Abt. Klimatologie 95440 BAYREUTH Germany
K. Pilegaard RISO National Laboratory P.O.Box 49 DK 4000 ROSKILDE Denmark
Tel: 49-921-7857172 Fax: 49-921-7857299
Tel: 45-46774677 Fax: 45-42370025
W.A.J. van Pul National Institute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742818 Fax: 31-30-287531
G.J. Reinds SC-DLO P.O.Box 125 6700 AC WAGENINGEN The Netherlands
Dr. J.G.M. Roelofs University of Nijmegen, Department of Ecology, Research Group Environmental Biology Toemooiveld 1 6525 ED NIJMEGEN The Netherlands Tel: 31-80-652340 Fax: 31-80-553450
Prof.Dr. W. Rolle Institut fur Troposph~enforschung ev. Permoserstrasse 15 04303 LEIPZIG Germany
F.G. R6mer KEMA P.O.Box 9035 6800 ET ARNHEM The Netherlands
P.F.A.M. R6mkens AB-DLO P.O.Box 129 9750 AC HAREN (Gr.) The Netherlands
Tel: 31-85-562599 Fax: 31-85-515022
Tel: 31-50-337368 Fax: 31-50-337291
R.S.A.R. van Rompaey Programme Bureau NRP (postbak 59) P.O. Box 1 3720 BA BILTHOVEN The Netherlands
W. Ruijgrok KEMA P.O.Box 9035 6800 ET ARNHEM The Netherlands
Tel: 31-30-743781 Fax: 31-30-251932
Tel: 31-85-562374 Fax: 31-85-515022
Tel: 42-2-590282 Fax: 42-2-5816748
Tel: 31-8370-74697 Fax: 31-8370-24812
Tel: 49-341-2352446 Fax: 49-341-2352361
497
F. Santachiara CCNR-FISBAT Via Gobetti 104 BOLOGNA 40129 Tel: 39-51-6399562 Fax: 39-51-6399568
Dr. T. Schneider National Institute of Public Health and Environmental Protection (postbak 59) P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742970 Fax: 31-30-251932
B. Schneider GSF Forschungszentrum ftir Umwelt und Gesundheit Kiahbachstrasse 11 D-81543 MUNCHEN Germany
K.G. Schnitzler Institut far Bioklimatologie Btisgenweg 1 37077 GOTTINGEN Germany
Tel: 49-89-65108862 Fax: 49-89-65108854
Tel: 00-49-551393685 Fax:
A.J. Schouten National Institute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-743134 Fax: 31-30-292897
R.J. Singles University of Edinburgh Department of Meteorology JCMB, King's Buildings Mayfield Road EDINBURGH EH93JZ Tel: 44-31-6508745 Fax: 44-31-6624269
F.P. Sival University of Groningen Laboratory of Plant Ecology P.O.Box 14 9750 AA HAREN The Netherlands Tel: 31-50-632231 Fax: 31-50-632273
R.I. Smith Institute of Terrestrial Ecology Edinburgh Research Station Bush Estate PENICUIK, Midlothian EH26 0QB United Kingdom Tel: 44-31-4454343 Fax: 44-31-4453943
Dr. G. Spindler Institut far Troposph~enforschung e.v. Permoserstrasse 15 04303 LEIPZIG Germany
T. Spranger Umweltbundesamt Mayerstrasse 52 10117 BERLIN Germany
Tel: 49-341-2352865 Fax: 49-341-2352361
Tel: 49-30-23145842 Fax:
H. Staaf Swedish Environmental Protection Agency S- 171 85 SOLNA Sweden
T. Staszweski Institute for Ecology of Industrial Areas Kossutha 40-833 KATOWICE Poland
Tel: 46-8-7991442 Fax: 46-8-283008
Tel: 48-3-1546031 Fax: 48-3-1541717
Ottelien van Steenis National Institute of Public Health and Environmental Protection (postbak 59) P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742970 Fax: 31-30-251932
E.G. Steingr6ver Institute for Forestry and Nature Research (IBN-DLO) P.O.Box 23 6700 AA WAGENINGEN The Netherlands
Italy
Tel: 31-8370-77895 Fax: 31-8370-24988
498
A.P. Stolk National Institute of Public Health and Environmental Protection P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-742412 Fax:
Dr. M.A. Sutton Department of Environment Air Quality, Room B354, Ronmey House 43 Marsham Street LONDON SW 1 3PY United Kingdom Tel: 44-71-2768155 Fax: 44-71-2768299
J. Szdzuj Institute for Ecology of Industrial Areas Kossutha 40-833 KATOWICE Poland
J.R. van de Veen Institute for Forestry and Nature Research P.O.Box 23 6700 AA WAGENINGEN The Netherlands
Tel: 48-3-1546031 Fax: 48-3-1541717
Tel: 31-8370-77894 Fax: 31-8370-24988
C. Veithen Institut ftir Bioklimatologie Btisgenweg 1 37077 GOTTINGEN Germany
A.C. Veltkamp ECN, Netherlands Energy Research Foundation P.O.Box 1 1755 ZG PETTEN The Netherlands
Tel: 00-49-551393685 Fax:
Tel: 31-2246-4251 Fax: 31-2246-3488
C.H. Venderbos Province Noord-Brabant Postbus 90151 5200 MC DEN BOSCH The Netherlands
A.T. Vermeulen Netherlands Energy Research Foundation P.O.Box 1 1755 ZG PETTEN The Netherlands
Tel: 00-73-812733 Fax:
Tel: 31-2246-4 194 Fax: 31-2246-3488
J.M. Verstraten University of Amsterdam Landscape and Environmental Research Group Nieuwe Prinsengracht 130 1018 VZ AMSTERDAM The Netherlands Tel: 31-20-5257415 Fax: 31-20-5257431
K.J. Vincent AEA Technology, National Environment Technology Centre E5 Culham ABINGDON OX14 3DB United Kingdom Tel: 44-235-463184 Fax: 44-235-463005
H. Visser KEMA Environmental Services P.O.Box 9035 6800 ET ARNHEM The Netherlands
P. de Visser Zevenwoude 173 3524 CN UTRECHT The Netherlands
Tel: 31-85-563012 Fax: 31-85-515022
Tel: 31-30-887432 Fax:
Marianne Vonk National Institute of Public Health and Environmental Protection (postbak 59) P.O.Box 1 3720 BA BILTHOVEN The Netherlands Tel: 31-30-743211 Fax: 31-30-251932
E.H.M. de Voogd Agricultural Research Department DLO P.O.Box 59 6700 AB WAGENINGEN The Netherlands Tel: 31-8370-74034 Fax: 31-8370-24060
499
W. de Vries DLO Winand Staring Centre P.O.Box 125 6700 AC WAGENINGEN The Netherlands
J. Vrubel Environmental Monitoring Centre, Ekotoxa Opava Homi n~n. 2 OPAVA 74601 Czech Republic
Tel: 31-8370-74353 Fax: 31-8370-24812
Tel: 42-653-215180 Fax: 42-653-218049
P. Warfvinge Department Chemical Engineering II Chemical Center P.O. Box 124 S-22100 LUND Sweden Tel: 46-46-103626 Fax: 46-46-108274
Dr. K.C. Weathers Institute of Ecosystem Studies Box AB MILLBROOK NY 12545 USA
J. Webb ADAS, Rosemaund Research Centre Woodthome Weres Road WOLVERHAMPTON WVE 8TQ United Kingdom Tel: 44-902-754190 Fax:
E. Weber Bundesministerium ftir Umwelt, Naturschutz und Reaktorsicherheit Bernkastelerstrasse 8 53175 BONN Germany Tel: 49-228-214137 Fax: 49-228-214838
Dr. E.J. Wilson National Power Research and Engineering National Power PLC, Windmill Hill, Business Park, Whitehill Way SWINDON Wiltshire SN5 6PB United Kingdom Tel: 44-793-896242 Fax: 44-793-896251
B. Wilson Wilson Associates Talworth, Carperby Leybum Nth Yorks United Kingdom Tel: 44-969-663133 Fax: 44-969-663133
G.P. Wyers Netherlands Energy Research Foundation P.O.Box 1 1755 ZG PETTEN The Netherlands
M. Zapletal Environmental Monitoring Centre, Ekotoxa Opava Homf nLrn. 2 OPAVA 74601 Czech Republic
Tel: 31-2246-4155 Fax: 31-2246-3488
Tel: 42-653-215180 Fax: 42-653-218049
U. Zelger Meteorological Institute University Theresienstrasse 37 80333 MUNICH Germany Tel: 49-89-188185 Fax: 49-89-182260
Y. Zhang Netherlands Energy Research Foundation P.O.Box 1 1755 ZG PETTEN The Netherlands
F.C. Zuidema Ministry of Agriculture, Nature Management and Fisheries P.O.Box 20401 2500 EK THE HAGUE The Netherlands Tel: 31-70-3793459 Fax: 31-70-3478167
Tel: 1-914-6775343 Fax: 1-914-6775976
Tel: 31-2246-4155 Fax: 31-2246-3488
This Page Intentionally Left Blank
501
Studies in Environmental Science Other volumes in this series
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
Atmospheric Pollution 1978 edited by M.M. Benarie Air Pollution Reference Measurement Methods and Systems edited by T. Schneider, H.W. de Koning and L.J. Brasser Biogeochemical Cycling of Mineral-Forming Elements edited by P.A. Trudinger and D.J. Swaine Potential Industrial Carcinogens and Mutagens by L. Fishbein Industrial Waste Management by S.E. Jorgensen Trade and Environment: A Theoretical Enquiry by H. Siebert, J. Eichberger, R. Gronych and R. Pethig Field Worker Exposure during Pesticide Application edited by W.F. Tordoir and E.A.H. van Heemstra-Lequin Atmospheric Pollution 1980 edited by M.M. Benarie Energetics and Technology of Biological Elimination of Wastes edited by G. Milazzo Bioengineering, Thermal Physiology and Comfort edited by K. Cena and J.A. Clark Atmospheric Chemistry. Fundamental Aspects by E. M~sz~ros Water Supply and Health edited by H. van Lelyveld and B.C.J. Zoeteman Man under Vibration, Suffering and Protection edited by G. Binachi, K.V. Frolov and A. Oledzki Principles of Environmental Science and Technology by S.E. Jergensen and I. Johnsen Disposal of Radioactive Wastes by Z. Dlouh~ Mankind and Energy edited by A. Blanc-Lapierre Quality of Groundwater edited by W. van Duijvenbooden, P. Glasbergen and H. van Lelyveld Education and Safe Handling in Pesticide Application edited by E.A.H. van Heemstra-Lequin and W.F. Tordoir Physicochemical Methods for Waste and Wastewater Treatment edited by L. Pawlowski Atmospheric Pollution 1982 edited by M.M. Benarie Air Pollution by Nitrogen Oxides edited by T. Schneider and L. Grant Environmental Radioanalysis by H.A. Das, A. Faanhof and H.A. van der Sloot Chemistry for Protection of the Environment edited by L. Pawlowski, A.J. Verdier and W.J. Lacy Determination and Assessment of Pesticide Exposure edited by M. Siewierski The Biosphere: Problems and Solutions edited by T.N. Veziroglu Chemical Events in the Atmosphere and their Impact on the Environment edited by G.B. Marini-Bett61o Fluoride Research 1985 edited by H. Tsunoda and M.-H. Yu Algal Biofouling edited by L.V. Evans and K.D. Hoagland Chemistry for Protection of the Environment 1985 edited by L. Pawlowski, G. Alaerts and W.J. Lacy Acidification and its Policy Implications edited by T. Schneider Teratogens: Chemicals which cause birth defects edited by V.M. Kolb Meyers Pesticide Chemistry by G. Matolcsy, M. N~dasy and V. Andriska Principles of Environmental Science and Technology (second revised edition) by S.E. Jorgensen and I. Johnson
502
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
61 62 63
Chemistry for the Protection of the Environment 1987 edited by L. Pawlowski, E. mentasti, C. Sarzanini and W.J. Lacy Atmospheric Ozone Research and its Policy Implications edited by T. Schneider, S.D. Lee, G.J.R. Wolters and L.D. Grant Valuation Methods and Policy Making in Environmental Economics edited by H. Folmer and E. van lerland Asbestos in the Natural Environment by H. Schreier How to Conquer Air Pollution. A Japanese Experience edited by H. Nishimura Aquatic Bioenvironmental Studies by C.D. Becket Radon in the Environment by M.M. Wilkening Evaluation of Environmental Data for Regulatory and Impact Assessment by S.Ramamoorthy and E. Baddaloo Environmental Biotechnology edited by A. Blazej and V. Privarovd Applied Isotope Hydrogeology by F.J. Pearson Jr., W. Balderer, H.H. Loosli, B.E. Lehmann, A. Matter, Tj. Peters, H. Schmassmann and A. Gautschi Highway Pollution edited by R.S. Hamilton and R.M. Harrison Freight Transport and the Environment edited by M. Kroon, R. Smit and J. van Ham Acidification Research in the Netherlands edited by G.J. Hey and T. Schneider Handbook of Radioactive Contamination and Decontamination by J. Severa and J. Bdr Waste Materials in Construction edited by J.J.J.M. Goumans, H.A. van der Sloot and Th. Aalbers Statistical Methods in Water Research by D.R. Helsel and R.M. Hirsch Acidification Research: Evaluation and Policy Applications edited by T. Schneider Biotechniques for Air Pollution Abatement and Odour Control Policies edited by A.J. Dragt and J. van Ham Environmental Science Theory by W.T. de Groot Chemistry and Biology of Water, Air and Soil edited by J. T~lgyessy The Removal of Nitrogen Compounds from Wastewater edited by B. Hailing Serensen and S.E. Jergensen Environmental Contamination edited by J.-P. Vernet Reclamation of Former Coal Mines and Steel Works by I.G. Richards, J.P. Palmer and P.A. Barratt Natural Analogue Studies in the Geological Disposal of Radioactive Wastes by W.M. Miller Water and Peace in the Middle East edited by J. Isaac and H. Shuval Environmental Oriented Electrochemistry edited by C.A.C. Sequeira Environmental Aspects of Construction with Waste Materials edited by J.J.J.M. Goumans, H.A. van der Sloot and Th.G. Aalbers Characterization and Control of Odours and VOC in the Process Industry edited by S. Vigneron, J. Hermia and J. Chaouki Nordic Radio Ecology, edited by Dahlgaard Atmospheric Deposition in Relation to Acidification and Eutrophication by J.W. Erisman and G.P.J. Draaijers