Coastal Tectonics
Geological Society Special Publications Series Editors
A. J. FLEET R. E. HOLDSWORTH A. C. MORTON M. S. STOKER
It is recommended that reference to all or part of this book should be made in one of the following ways. STEWART, I. S. & VITA-FINZI, C. (eds) 1998. Coastal Tectonics. Geological Society, London, Special Publications, 146. CHAPPELL, J., OTA, Y. & CAMPBELL, C. 1998. Decoupling post-glacial tectonism and eustasy at Huon Peninsula, Papua New Guinea. In: STEWART, I. S. & V~TA-FINZI, C. (eds) 1998. Coastal Tectonics. Geological Society, London, Special Publications, 146, 31-40.
GEOLOGICAL SOCIETY SPECIAL PUBLICATION NO. 146
Coastal Tectonics
EDITED BY
I A I N S. S T E W A R T Brunel University, UK AND
CLAUDIO VITA-FINZI University College London, UK
1998 Published by The Geological Society London
THE GEOLOGICAL SOCIETY The Society was founded in 1807 as The Geological Society of London and is the oldest geological society in the world. It received its Royal Charter in 1825 for the purpose of 'investigating the mineral structure of the Earth'. The Society is Britain's national society for geology with a membership of around 8500. It has countrywide coverage and approximately 1500 members reside overseas. The Society is responsible for all aspects of the geological sciences including professional matters. The Society has its own publishing house, which produces the Society's international journals, books and maps, and which acts as the European distributor for publications of the American Association of Petroleum Geologists, SEPM and the Geological Society of America. Fellowship is open to those holding a recognized honours degree in geology or cognate subject and who have at least two years' relevant postgraduate experience, or who have not less than six years' relevant experience in geology or a cognate subject. A Fellow who has not less than five years' relevant postgraduate experience in the practice of geology may apply for validation and, subject to approval, may be able to use the designatory letters C Geol (Chartered Geologist). Further information about the Society is available from the Membership Manager, The Geological Society, Burlington House, Piccadilly, London WlV 0JU, UK. The Society is a Registered Charity, No. 210161. Published by The Geological Society from: The Geological Society Publishing House Unit 7, BrassmiU Enterprise Centre Brassmill Lane Bath BA1 3JN UK (Orders: Tel. 01225 445046 Fax 01225 442836)
Distributors
USA AAPG Bookstore PO Box 979 Tulsa OK 74101-0979 USA (Orders: Tel. (918) 584-2555 Fax (918) 560-2652)
First published 1998
Australia The publishers make no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any legal responsibility for any errors or omissions that may be made. 9 The Geological Society 1998. All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with the provisions of the Copyright Licensing Agency, 90 Tottenham Court Road, London WlP 9HE. Users registered with the Copyright Clearance Center, 27 Congress Street, Salem, MA 01970, USA: the item-fee code for this publication is 0305-8719/98/$10.00.
Australian Mineral Foundation 63 Conyngham Street Glenside South Australia 5065 Australia (Orders: Tel. (08) 379-0444 Fax (08) 379-4634)
India Affiliated East-West Press PVT Ltd G- 1/ 16 Ansari Road New Delhi 110 002 India (Orders: Tel. (11) 327-9113 Fax (11) 326-0538)
Japan British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. ISBN 1-86239-024-X
Typeset by Aarontype Ltd, Unit 47, Easton Business Centre, Felix Road, Bristol BS5 0HE, UK Printed by The Alden Press, Osney Mead, Oxford, UK
Kanda Book Trading Co. Cityhouse Tama 204 Tsurumaki 1-3-10 Tama-Shi Tokyo 0206-0034 Japan (Orders: Tel. (0423) 57-7650 Fax (0423) 57-7651)
Contents
Preface PELTIER, W. R. Global glacial adjustment and coastal tectonics CHAPPELL,J., OTA, Y. • CAMPBELL,C. Decoupling post-glacial tectonism and eustasy at Huon Peninsula, Papua New Guinea SOTER, S. Holocene uplift and subsidence of the Helike Delta, Gulf of Corinth, Greece TRECKER, M. A., GURROLA, L. D. & KELLER, E. A. Oxygen isotope correlation of marine terraces and uplift of the Mesa Hills, Santa Barbara, California, USA BORDONI, P. & VALENSISE, G. Deformation of the 125 ka marine terrace in Italy: tectonic implications CORNET, Y. 8z DEMOULIN, A. Neotectonic implications of a lineament-coplanarity analysis in Southern Calabria, Italy FLEMMING, N. C. Archaeological evidence for vertical tectonic movement on the continental shelf during the Palaeolithic, Neolithic and Bronze Age periods GALILI, E. & SHARVIT,J. Ancient coastal installations and the tectonic stability of the Israeli coast in historical times FOULGER, G. R. & HOFTON, M. A. Regional vertical motion in Iceland 1987-1992, determined using GPS surveying ORME, A. R. Late Quaternary tectonism along the Pacific coast of the Californias: a contrast in style THACKRAY, G. D. Convergent-margin deformation of Pleistocene strata on the Olympic Coast of Washington, USA MERRITTS, D., EBY, R., HARRIS, R., EDWARDS, R. L. & CHENG, H. Variable rates of Late Quaternary surface uplift along the Banda Arc-Australian plate collision zone, eastern Indonesia REYSS, J. L., PIRAZZOLI, P. A., HAGHIPOUR, A., HATTIe, C. & FONTUGNE, M. Quaternary marine terraces and tectonic uplift rates on the south coast of Iran MEGHRAOUI, M.; OUTTANI, F., CHOUKRI, A. & FRIZONDE LAMOTTE, O. Coastal tectonics across the South Atlas Thrust Front and the Agadir Active Zone, Morocco MURRAY-WALLACE, C. V., BELPERIO, A. P. & CANN, J. H. Quaternary neotectonism and intra-plate volcanism: the Coorong to Mount Gambier Coastal Plain, southeastern Australia: a review NUNN, P. Late Cenozoic emergence of the islands of the northern Lau-Colville Ridge, southwest Pacific BEZERRA, F. H. R., LIMA-FILHO, F. P., AMARAL, R. F., CALDAS, L. H. O. & COSTA-NETO, L. X. Holocene coastal tectonics in NE Brazil BILHAM, R. Slip parameters for the Rann of Kachchh, India, I6 June 1819, earthquake, quantified from contemporary accounts MCNEILL, L. C., GOLDFINGER, C., YEATS, R. S. & KULM, L. D. The effects of upper plate deformation on records of prehistoric Cascadia subduction zone earthquakes DOMINEY-HOWES, D., DAWSON, A. & SMITH, D. Late Holocene coastal tectonics at Falasama, western Crete, (Greece): a sedimentary study GOFF, J. R., CROZIER, M., SUTHERLAND, V., COCHRAN, U. & SHANE, P. Possible tsunami deposits from the 1855 earthquake, North Island, New Zealand Index
vii 1 31 41 57 71 111 129 147 165 179 199 213
225 239 255
269 279 295 319 341 351 373
Preface At first glance, coastal tectonics is as redundant a category as inland tectonics, for the shoreline does not necessarily coincide with a distinctive geodynamic environment. What prompted the international conference on the subject that led to this book was the Editors' conviction that coasts favour the study of active tectonics (a) by providing a reference d a t u m - namely sea l e v e l - against which deformation can be measured and (b) by supplying datable material and environmental clues with which the progress of deformation can be traced. As a bonus we have coasts which temporarily coincide with a tectonic boundary or major structure and wash it clean for our inspection. Consider plate boundaries such as those of the western Americas where subduction and transform displacement are now operating, or the extensional coasts of the Gulf of Corinth where normal faulting will perpetuate tectonic conditions on the coast for some time to come. There are also countless locations, notably oceanic islands, which are tectonic at one remove, as their uplift or subsidence reflects the dynamic behaviour of the lithosphere elsewhere. There was a further question to be resolved. The original plan had been to focus on Late Quaternary coastal tectonics, but this soon emerged as unnecessarily restricting: why 'late', and why Quaternary, when many active coasts began to deform in the Tertiary or even earlier, and when much illuminating work depends on the evidence of seismology and geodesy? We have an excellent precedent for our title: that of the survey by Ken Lajoie (1986) that did much to define the scope and procedures of tectonic investigations on coasts. Lajoie opened his discussion by observing that between one third and one half of the Earth's marine coastlines lie along or near tectonically active plate boundaries. By implication he was emphasizing mechanism rather than narrative, and that was our intention when we organized a conference around the application of high-resolution coastal chronologies to the testing and refining of crustal models at local, regional and global scales. The papers that follow (which include seven that were solicited after the meeting) have accordingly been grouped into sections which deal in turn with the extraction of tectonic data from the many kinds of noise in the coastal record and with their bearing on the analysis of interplate and intraplate tectonics and the construction of earthquake sequences. Lack of space meant that some themes, such as salt tectonics, receive little mention; conversely, areas which have attracted investigation from various viewpoints, notably coastal California, are discussed in more than one paper. Of the many possible dating methods, the emphasis is on radiocarbon and U-series techniques, but one of the papers reviews the potential value of stable isotopes in the correlation of marine terraces, two are primarily concerned with archaeological indicators of tectonic displacement, and another exploits historical records which have long lain unread in the archives; Strombus bubonius emerges reinvigorated in its new tectonic role. Some persistent geodynamic problems are at best highlighted by the work reported in this book, notably the distinction between seismic and aseismic contributions to net tectonic strain; the distinction between stable and unstable coastlines, however, emerges as unhelpful.
Vlll
We thank the following, as well as a few others who wished to remain anonymous, for advice and help with the meeting and with reviewing the manuscripts: F. A. Aberg, N. N. Ambraseys, R. Armijo, K. Berryman, M. Berberian, A. L. Bloom, D. Q. Bowen, J. Coleman, P. E. F. Collins, A. B. Cundy, A. Dawson, M. Evron, R. W. Fairbridge, N. C. Flemming, G. R. Foulger, A. Hull, M. Ivanovich, H. Kelsey, A. J. Long, J. M. McArthur, M. Meghraoui, K. Morris, W. Murphy, D. Neev, A. R. Nelson, P. D. Nunn, J.-L. Ortlieb, Y. Ota, C. Pain, P . A. Pirazzoli, J. P. Platt, G. Roberts, F. Sigmundsson, P. Stewart, F. W. Taylor, A. B. Watts, M. Weinstein-Evron and C. Zazo. We are grateful to A. Hills for editorial assistance. Iain Stewart & Claudio Vita-Finzi
Global glacial isostatic adjustment and coastal tectonics W . R. P E L T I E R
Department of Physics, University of Toronto, Toronto, Ont., Canada M5S 1A7 (e-mail:
[email protected]) Abstract: A global and gravitationally self-consistent model of the process of glacial isostatic adjustment (GIA) has been developed that extremely well reconciles the vast majority of available records of Holocene relative sea-level history, not only from sites that were ice covered at last glacial maximum (LGM) but also from sites that are well removed from such locations. There do exist, however, data that have been construed to constitute a significant challenge to this theory, namely, the long records of relative sea-level history derived on the basis of U/Th-dated coral sequences from the Huon Peninsula of Papua New Guinea and from Tahiti in the central equatorial Pacific Ocean. Following a review of the theoretical model and a discussion of the extent to which it is able to successfully reconcile a very wide range of Holocene shoreline observations, the discussion focuses upon the interpretation of these very important and interesting records, which are subject to different levels and types of tectonic contamination. These analyses suggest that existing estimates of the levels of Holocene tectonic contamination at both locations may require revision. In this context, it is suggested that the global model of the GIA process is sufficiently accurate that the magnitude and form of local tectonic effects during the Holocene period might be sensibly estimated by simply subtracting the GIA prediction for a given site from the observed variation of relative sea level.
The late Pleistocene cycle of glaciation and deglaciation, which has been the dominant contributor to climate system variability for the last 900 000 years of Earth history, is indelibly recorded in the geological record of relative sea-level (rsl) change. As each of these 100000 year quasi-periodic cycles of ice-sheet advance and retreat involved a fall and subsequent rise of eustatic sea level of c. 120 m, it is hardly surprising that the record of these events should be of such high quality. The best proxy recordings of this glacial cycle, from a long timescale perspective, undoubtedly consist of those based upon oxygen isotopic measurements made on the tests of Foraminifera extracted from deep-sea sedimentary cores. Shackleton (1967) demonstrated that the records thereby derived on the basis of benthic species provided a high-quality proxy for the amount of land ice that existed on the continents at the time in the past represented by the depth in the core at which the isotopic measurement was made. It is, of course, on the basis of records of this kind that the important role played by orbital insolation variations in driving the ice-age cycle was first clearly established (Hays et al. 1976). Although the linkage between orbital insolation forcing and ice volume response is not nearly so direct as Milankovitch had envisioned (e.g. see Tarasov & Peltier (1997) for a recent discussion) it was nevertheless clear on the basis of such data that the small changes in the effective intensity of the Sun, caused by temporal variations of the geometric properties of the Earth's orbit, were able to induce significant cryospheric response.
Of primary interest in the present context will be the variations of rsl that are associated with the most recent deglaciation event, which began subsequent to last glacial maximum (LGM) 21000 sidereal years ago. Even though this event had essentially ended by c. 4000 years ago, rsl continues to change owing to this cause (by rsl, in all that follows, I will imply sea level measured with respect to the surface of the solid Earth). This lingering memory of the deglaciation process is essentially a consequence of the fact that the Earth's shape is continuing to deform because of the shift in surface mass load that occurred during deglaciation as the vast Laurentide, Northwest European and southern hemisphere ice complexes disintegrated and the meltwater thereby produced was added to the ocean basins. This continuing deformation is a consequence of the very high value of the effective viscosity of the Earth's mantle, which governs the timescale of the return to gravitational equilibrium of the ice-solid earthocean system subsequent to deglaciation. Because this continuing relaxation of shape depends so strongly upon mantle viscosity, observations of the process may be employed to infer this Earth property and its variation with depth. That such inferences provide information of fundamental importance will be clear by virtue of the fact that knowledge of the steady-state creep resistance of the mantle is required in the construction of mantle convection models of the process of continental drift and sea-floor spreading. In the discussion of these ideas to be presented in what follows, I will begin with a brief
PELTIER,W. R. 1998. Global glacial isostatic adjustment and coastal tectonics. In: STEWART,I. S. & VITA-FINZI, C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 1-29.
2
W . R . PELTIER
review of the theoretical structure of the global model of the glacial isostatic adjustment (GIA) process that has been under continuous refinement at Toronto for some time. The origins of this model lie in the analysis presented by Peltier (1974) of the viscoelastic response of spherically symmetrical models of the planet to variations of surface mass load. Using the impulse response Green function for the perturbation of the surface gravitational potential derived for such models by Peltier & Andrews (1976), Farrell & Clark (1976) discussed the primitive form of a 'Sea-level equation' that could be employed to predict the variations of rsl that should occur as a result of the combined influence of the deformation of the solid Earth caused by the changing surface load and the deformation of the geoid (the surface of constant gravitational potential that is coincident with mean sea level (msl) over the oceans). This equation was constructed by analogy with that introduced by Platzman (1971) to describe the influence of the elastic yielding of the sea floor onto the ocean tides. In the studies by Clark et al. (1978) and Peltier et al. (1978) this equation was more accurately expressed and solved for the realistic model of northern hemisphere deglaciation that had been produced by Peltier & Andrews (1976) and designated ICE-1. Further refinements to the theoretical structure that were thereafter introduced included the additional mathematical analysis required to calculate the rotational response to the glaciation--deglaciation process (Peltier 1982; Wu & Peltier 1984), analysis of the free air gravity anomalies associated with this dynamical forcing (Wu & Peltier 1983; Mitrovica & Peltier 1989; Peltier et al. 1992), more accurate spectral methods for the solution of the sea-level equation itself (Mitrovica & Peltier 1991) and development of a technique with which one might incorporate into the solution the full influence of time dependence of the coastline (Peltier 1994) and of the (rather less important) feedback of the changing rotational state of the planet onto sea-level history itself (Peltier 1998a, b). Various parts of this theoretical structure have been subsequently reproduced by others and the complete structure now serves as basis for the continuing international effort to fully understand the GIA process. These contributions from workers outside the Toronto group include those by Lambeck et al. (1990), who employed a much simplified version of the sea-level equation to investigate the postglacial rebound of Fennoscandia; Lambeck et al. (e.g. 1996), who performed a detailed series of analyses of postglacial rsl histories of the British Isles; Han & Wahr (1995), who have
rederived the viscoelastic normal mode formalism of Peltier (1976, 1985); and Fang & Hager (1995), who have also invested effort to understand the rudiments of the normal mode theory. The implications of this work to our understanding of mantle rheology have also been thoughtfully addressed in the recent literature by Karato & Wu (1993). In the following section of this paper I briefly review the structure of this formal theory of the GIA process. I will then focus upon the problem of tuning the viscosity profile of the model, and will examine the extent to which the theory is able to accurately reconcile a globally distributed set of rsl histories obtained on the basis of a4C dating of various rsl indicators. Given the rather good fit to such data that the model delivers, further analyses are devoted to the investigation of observations that have been suggested to disagree profoundly with the theoretical predictions, Arguments are presented to the effect that these concerns are not particularly well founded, and conclusions are offered.
The global theory of GIA and rsl change The record of sea-level history that is contained in the geological record is a recording of the level of the sea relative to the deforming surface of the solid Earth. It is this fact which makes the interpretation of the record as challenging as it so clearly is. If we define this rsl history to be S(0, A, t), with 0 and A latitude and longitude, respectively, and t time, then we might usefully express rsl history in the following schematic fashion:
s(o, A, t) : c(o, A, t)[G(0, ~, t) - R(0, ;~, t)] (1) in which C(O, A, t) is the so-called 'ocean function', which equals zero over land or land-locked water and unity over the surface of the global ocean. In equation (1), G(O, A, t) is the geoid of classical geodesy which is defined by the surface of constant gravitational potential that is coincident with msl over the oceans and R(O, A, t) is the local radius of the solid Earth. To predict the function S(O, A, t) we are therefore obliged to develop a theory on the basis of which we may compute the triplet of functions (C, G, R). The key ingredient of such a theory, as previously mentioned, was provided by Peltier (1974) who developed a mathematical structure with which one could calculate both G and R assuming C to be fixed to the present-day ocean function. That analysis, which was based upon the application of first-order perturbation theory, led to
GLACIAL ISOSTATIC ADJUSTMENT AND COASTAL TECTONICS
3
S(O, A, t) and to compare these predictions with geologically inferred rsl histories. In this process we would construct the function A~(t) such as to ensure conservation of mass by insisting that
the re-expression of equation (i) in the form:
s(o,A,t) = c(o,;~,t)
• { I dt' J. J dfl L(O', t')
I
c pwX(t ) df~
• +
A~(t) } g
= [ .w{f (2)
in which ~2 is the surface of the Earth, L is the history of variations of surface mass load (mass per unit area) that occur as a result of the glaciation-deglaciation process, q~L and F L are respectively viscoelastic surface load Green functions for the gravitational potential perturbation and radial displacement, and the function A~(t) is constructed so as to ensure that the variation of surface mass load is mass conserving in the sense that only the mass of water produced by melting (or accreting) continental ice appears in (or disappears from) the oceans. The argument "7 in these Green functions is simply the angular separation between source point (0',A') and field point (0, A), a spatial dependence which results from the assumption that the Earth model of interest is spherically symmetrical in its physical properties. From a mathematical perspective the right-hand side of equation (2) is a (triple) convolution integral. The theory required to construct q5L and F L was presented by Peltier (1974) and Peltier & Andrews (1976), and requires knowledge only of the radial viscoelastic structure of the planet. From a technical perspective the challenge posed by equation (2) arises because of the composite property of the surface load L. This may be made explicit by expanding it in the form
c(0, A, t) =
pr1(o, ~, t) + pwS(O, A, t)
(3)
in which pI and pw are the densities of ice and water, respectively, and I and S are respectively ice and water 'thickness'. We consider L to be positive when the net mass per unit area is increasing and negative where it is decreasing. Clearly, when equation (3) is inserted in equation (2) the resulting equation will be seen to constitute an integral equation for the rsl history S(O, A, t), as S now appears not only on the left-hand side but also under the triple convolution integral on the right-hand side. Given an assumed history of ice-sheet thickness variations I(O,A,t) and a radial viscoelastic structure for an assumed spherically symmetrical model of the planetary interior, we could proceed to solve this integral equation to predict
d,' J. J
t')
A ,b ( t_____). ) x [G(t - t') - R(t - t')]~df~ + pwA g J = -Mj(t)
(4)
in which the integral on the left-hand side is the mass that has been added to the oceans by time t, which must equal the mass of water produced by melting ice, here defined as Mi(t). The negative sign affixed to this function on the righthand side of equation (4) is employed to indicate explicitly that Mr(t) itself is negative when this component of the surface mass load is being removed from the surface. By defining {I)(t)
MI(t)
1
g
pwA(t)
A(t)
x ( I r .dt' I~J d~2'L(O',~',t ')
(5)
• [a(t- t')- R(t- t')]~ /
o
in which A(t) is the surface area of the oceans at time t and ()o indicates integration over the oceans, we will then ensure that solutions to equation (2) conserve mass. Because of the form of the A~(t)/g correction to equation (2), one should not expect that the amount of sea-level rise that occurs far from the ice sheets will be well approximated by the first term on the righthand side of equation (5). Although it is precisely the form of the sealevel equation (2) that I have employed as basis for most of my work on the GIA process, there are in fact two potentially relevant physical effects that are not included in this version of the theory. These are, respectively, the influence of the time dependence of the ocean function C(O, A, t) and the feedback onto sea level of the changing rotational state of the planet. I will return to a discussion of the important former effect below. The latter effect turns out to be small but it is difficult to be certain that this is so without actually doing the calculation. To accomplish this we proceed iteratively by first solving equation (2) to determine the global rsl history S by assuming C to be constant and
4
W . R . PELTIER
including a model of the history of ice-sheet loading and unloading before LGM (this may be constructed by employing the SPECMAP 6180 record of Imbrie et al. (1984)). Given the complete history of surface mass loading L, we then solve the Euler equation d dt (gij 6dj) + Eijk6djgkltdl = 0
(6)
in which Jij is the moment of inertia tensor of the planet, wj are the components of its angular velocity vector and eijk is the Levi-Cevita alternating tensor. Assuming a biaxial model for the undeformed shape of the planet (see Peltier & Jiang (1996a) and Peltier (1997) for the complete but unnecessary triaxial theory), highly accurate solutions to equation (6) may be constructed by employing the standard perturbation expansion:
As discussed in detail by Peltier (1982) and Wu & Peltier (1984), equations (8) may be solved most efficiently by using Laplace transform techniques to determine the wi(t) once L(O, A, t) has been fully determined by solving the sea-level equation (2). Given the solution to equation (8) we may simply incorporate the influence of the changing rotation into equation (2) by extending it as:
S(O, A, t ) = C(O, A, t)
{I'
dt'
dgt'
--00
• [L(O', A', t')G~(% t - t') +
9R(0', ~', t')6~(% t- t')]
+
A~(t) } g
(lO)
~Oi = ~(6ij + mi)
J i j = Iij, i r j (7)
Jn = A + Ill J22 = A nt- 122 ,]33 : C -Jr- 133
in which (A, A, C) are the principle moments of inertia, f~ is the angular velocity of the unperturbed Earth, and Iig and mi are assumed small fluctuations away from the unperturbed basic state. On substitution of equation (7) into equation (6) and dropping all terms of higher order than first in the fluctuations we obtain the following decoupled system of equations for polar motion and rotation, respectively: i . -
-
O-r
m + m = ~
(8a)
(8b)
/'~/3 = 9'II3
in which the so-called excitation functions are and t~3, err = ( C - A ) f ~ / A is the Chandler wobble frequency of the rigid Earth, m = mi + im2, ~ = 91 + i92, i = x/Z]- and the ~i are, with the dot indicating time differentiation, respectively, 113
91 - ( C 123
92-(C_A) 93--
/33
C
I23
+ f~(C- A~
(9a)
/r13
fl(C-A)
(9b)
in which the Green function G~ = [0(% t - t')/g - 1-'(% t - t')] is the same kernel as in equation (2), 9g(0 ', A', t') is the variation of the centrifugal potential because of the changing rotational state which, following Dahlen (1976), may be written (to first order in perturbation theory, to be consistent with the approximation employed to solve equation (6)), as +1
ff~R = 900 Yoo(O, A) + Z
92m Y2m(O' "~)
(11)
m=--I
where 900 = ~w3(t )fla 2
92o = - ~w3(t)i2a 2V/4-/5 92-1 -- (Wl -- iw2)(f]a2/2)V/2/15 92+1 = (wl + iw2)(f~aE/Z)x/~/15 and the tidal-loading Green function G~ is expressed in terms of tidal Love numbers h~ and klx as 1
a~(% t) = g ,=o
[1 + kT(t) -- hT(t)]P1 (cos "7)
02) just as the surface-loading Green function G~ is expressed in terms of surface load Love numbers h) and k L as oo
G~(7, t) = a Z [l + k~(t) - hL(t)]PI(cos 7) me l=o
(9c)
(13)
GLACIAL ISOSTATIC ADJUSTMENT AND COASTAL TECTONICS Both sets of viscoelastic Love numbers are calculated using the theoretical ideas and methods developed by Peltier (1974, 1976, 1985). There will be no purpose served by reviewing these technical details here. A further aspect of theory that will be of particular interest, however, is that required to incorporate the full impact of the changing coastline that occurs as land becomes inundated by the sea as sea levels rise because of ice-sheet melting and as land that was once ice covered rises out of the sea as a result of the process of postglacial crustal 'rebound'. To understand how these additional influences may be incorporated into the theory it is helpful to begin by noting that both forms of the 'sea-level equation', namely, those represented by equations (2) and (10), are constructs of first-order perturbation theory that deliver solutions for the history of rsl change with respect to an unspecified and thus arbitrary datum. It is precisely this arbitrariness that may be exploited so as to incorporate the full influence of ocean function time dependence. We simply fix this datum by determining a timeindependent field T'(O, 4) such that
5
solution S(0, A, t), we then determine a new T'(O, A) and thus a new C = C2(0, A, t) using equation (14). We continue this iterative process until the solution for C(O, A, t) converges, which typically occurs in just a few iterations. Very recently, a further refinement of this theory has been developed which has both increased the accuracy of the computation of palaeotopography as defined in equation (14) and improved the understanding of mass balance when this is examined from the perspective of the eustatic sea-level rise expected on the basis of the total ice melt and the net sea-level rise that is predicted by solving equation (2) or (10). This involves a subtle aspect of the theory that has not been explored until recently (Peltier 1998c) which is as follows. We consider the evolution of rsl Sis(P, A, t) at a point on the landscape that is ice covered at L G M but which later comes to be inundated by the sea. Here I employ the subscript IS to denote inland sea. At such points the time series for Sis that is delivered by solving equation (2) or (10) has the following mathematical form: SIs(P, A, t > tD) : ASIs(0, A ) H ( t - tD)
S(O, A, t p ) + T'(O, 4 ) = Tp(0, A) + S~s(O, A, t >_ tD) in which S is a solution to either equation (2) or (10), tp is the present time and Tp(O, A) is the present-day topography of the planet with respect to sea level determined, say, by the ETOPO5 model (or some other higher-resolution model if one is available). If we then construct a time-dependent topography for t h e planet by computing
T(O, A, t) = S(O, A, t) + [Tp(0, A) - S(O, A, tp)] and correct this by adding to T(O, A, t) the thickness of ice 1(8, A, t) to obtain
PT(O, A, t ) = T(O, A, t)+ I(0, A, t) = S(O, A, t) + T'(O, A) + I(0, A, t) (14) it will be clear that where T + I is positive there is (perhaps ice-covered) land that stands above sea level and that where T + I is negative there is ocean. We may then define a 'first estimate' of the time-dependent ocean function as the function C I(o, A, t) that is unity wherever T + I is negative. Given this first estimate we then return to equation (2) or (10) and solve it again incorporating this form of the time dependence C(O, A, t) -- C l(0, A, t). Given the new of
05)
in which to(O, A) is the time of 'deglaciation' when the sea first occupies the region at latitude 0 and longitude A, ASIs < 0 is a spontaneous fall of sea level that is delivered by the solution of equation (2) or (10) at the instant tD, H(t - tD) is the Heaviside step function (+ 1 for t _> to and zero for t < tD) and S]s(O, A, t >__tD) is a function that vanishes at t = tD and thereafter decreases with time so as to represent the fall of rsl that occurs in the inland sea as a consequence of postglacial rebound of the crust. Solutions to equation (2) or (10) deliver an abrupt fall of sea level at t = tD because at that instant there is a marked difference in the gravitational potential between such locations and that which defines the surface of the exterior ocean with this potential being higher than the surface ocean because of the influence of glacial rebound. Solutions to equation (2) or (10) therefore deliver a sharp drop of sea level at the instant of deglaciation t = tD even though the region is being inundated. This is a source of mass to the exterior ocean and constitutes an additional removal of load from the deglaciating region. As this removal of load can only represent, in fact, an additional removal of ice, it is clear that ASIs is actually to be associated with an 'implicit' component of the ice unloading history. I call this 'implicit ice' and its contribution to the loading history, expressed in
6
W . R . PELTIER
terms of an equivalent ice thickness, may be computed from the expression
L(O, A, t)= pIIEx(O, A, t)+ pwASisH(t - tD) = p, I/EX(0, A, t)+ Pw A S I s H ( t - tD)I Pl L J
= pi[I~x(O, A, t)+ I~M(O, A, t)] (16a) in which IEx(O, A, t) is the ice thickness history determined by tuning the model (see below) to fit rsl observations from the ice-covered region, and I~M(O, A, t) is the implicit ice that was also removed to deliver the mass to the exterior ocean that is obtained from the solution of equation (2) or (10). In this solution the ice load IIM 'masquerades' as a fall of sea level. A second contribution to the implicit component of the ice load is connected to the field S(O, A, t) in equation (15). It will be clear that, for rsl to be able to fall in regions that were ice covered at LGM but which are inundated by the sea at t = tD, there must be water in the icecovered region subsequent to tD! Now, in the solution of equation (2) or (10), the unloading of these regions is fully accounted for, partly by IEx and partly by I~M. What is not accounted for, in this equation that derives from the application of first-order perturbation theory, is the net mass of water that fills the depression of the surface that exists at t = tD. This water must, of course, also be delivered by the ice that disappeared from the surface as inundation occurred. This additional contribution to the net implicit ice applies no load to the surface.at t = tD, only subsequently as rebound occurs, but it does 22.5
I
contribute to the thickness of ice that must have existed on the surface at LGM and thus to the palaeotopography defined in equation (14). It may also be computed very accurately as follows. We simply compute the present altitude of the marine limit with respect to msl as a function of geographical position, say ML(O, A, tp), and to this we add the present-day bathymetry, say D(O, A, tp), to obtain, in ice thickness equivalent form, this second contribution to implicit ice as
I~M =PW[ML(O, A, tp)+ D(O, A, tp)] (16b) p~
In computing the full palaeotopography in equation (14) we must therefore also include this second contribution from implicit ice to obtain
PT(O, A, t ) = T(O, A, t)+ IEX(0, A, t) +
:IM(0, ~, t) + I2IM(O, A, t)
Examples of the results obtained from this refinement of the palaeotopography calculation will be provided in what follows. Before employing this theory to illustrate in detail the extent to which postglacial rsl histories may be explained in terms of it, it will be useful to first illustrate the general forms that such solutions possess. To this end I will immediately consider the nature of the solution obtained when the radial variation of mantle viscosity is fixed to that of model VM2 shown in Fig. 1. The origins of this model will be described in the next section. With the radial elastic structure of the model also fixed to that of the Preliminary Reference Earth Model (PREM; Dziewonski &
I
t
I
g 21.5 '~
..........
~7~f
............................
21.0
20.5
20.0
., I
3400
4000
I
I
i
4600
5200
5800
6400
Radius (km) Fig. 1. The viscosity models VM1, VM2 and VM3, which are discussed in detail in the later sections of the text.
GLACIAL ISOSTATIC ADJUSTMENT AND COASTAL TECTONICS
7
Fig. 2. Time slices through the ice-thickness maps that constitute the ICE-4G model of deglaciation (note that these are slightly modified from the fields derived by Peltier (1994, 1996), primarily by an increase of the ice thickness over the Laurentide complex that was centred on Hudson Bay).
8
W . R . PELTIER
Anderson 1981), solutions will be discussed for the ICE-4G deglaciation model of Peltier (1994), examples of the northern hemisphere isopacks for which are shown in Fig. 2. Notable in this figure are the extensive North American and Northwest European ice complexes that existed at LGM in which the thicknesses of the continental ice sheets, were typically of order 4 km. In the southern hemisphere component of ICE4G there was also significantly more ice at LGM than at present over West Antarctica and also significant ice cover over Western Patagonia. In Fig. 3 1 show the present-day predicted rate of rsl rise for these choices of the input fields and for both the version of the sea-level equation that excludes the influence of rotational feedback (equation (2)) and that which includes this effect (equation (10)). The top and middle plates of Fig. 3 show these respective solutions, whereas the bottom plate displays the difference between them. Evident upon inspection of these illustrative results is that the influence of rotational feedback upon this characterization of the sea-level response to deglaciation is extremely weak and is strongly dominated by the degree two and order one pattern that is forced entirely (see equation (11)) by the polar motion component of the rotational response to deglaciation. That the influence of rotational feedback is weak not only from the perspective of the present-day rate of rsl rise driven by the GIA process but also from the perspective of the complete history of rsl change is demonstrated in Fig. 4, where I have compared predicted and observed rsl histories at six different locations. The first two, from Barbados and the Huon Peninsula, are U/Th-dated coral records that will figure prominently in what is to follow. The remaining four are from sites that are as close as possible to the extrema of the degree two and order one pattern that characterizes the contribution to the rsl record by rotational feedback. Even at these locations, where the influence of the feedback is most intense, it is clearly extremely weak and negligible for most purposes. This contradicts the claim to the contrary made recently by Bills & James (1996). One final aspect of the general form of the solution upon which I will comment here concerns the time-dependent topography of the planet with respect to sea level that develops as a result of the deglaciation process. Figure 5 illustrates the northern hemisphere component of this field determined for the ICE-4G (VM2) model by executing the steps described in equations (14)-(16). Evident by inspection of this figure will be the vast land bridges that existed at LGM in both the present-day Bering
Strait, which was then entirely dry land (the continent of 'Beringia'), and the present-day English Channel. At that time, most of the present-day Indonesian Archipelago was dry land, and a vast land bridge also connected Australia to Papua New Guinea. These aspects of the global 'topographically self-consistent' solution to the sea-level equation were first described by Peltier (1994). Figure 6 illustrates the time dependence of the coastline in several of these locations in terms of what I have previously called 'inundation maps', on the basis of which one may infer at a glance the time at which a particular land bridge first became impassible. Also of note in the second plate of Fig. 6, which simply portrays the regions of the Aegean Archipelago and Mediterranean Sea that are predicted to have been dry land at LGM but which are now beneath the sea, is that there are many candidates for Atlantis!
Tuning the model parameters Although some evidence has already been presented to the effect that the ICE-4G (VM2) model successfully fits a considerable range of rsl observations, it will prove useful, before providing a more systematic demonstration of this fact, to discuss the procedure that has been followed to arrive at the two required input components of this model. These components consist respectively of the radial viscosity profile (VM2) and the deglaciation history (ICE-4G). As the model of deglaciation history has been discussed at length by Peltier (1994, 1996), I will focus herein on the viscosity structure. Focusing then upon the radial profile of mantle viscosity, the profile VM2, or rather the family of VM2-1ike profiles, has been inferred (Peltier 1996, 1998b; Peltier & Jiang 1996b, 1997) through application of a formal Bayesian inversion procedure based upon the use of the simple VM 1 profile shown in Fig. 1 as starting model. As a first estimate of the viscosity profile, VM 1 has several properties which strongly suggest it to be highly appropriate. Of these, the most important will be clear on the basis of inspection of Fig. 7, which shows a sequence of predictions of the non-tidal acceleration of planetary rotation (represented by )2, which is the time rate of change of the degree two axial component of the gravitational potential field of the planet) and the speed of polar wander as a function of the viscosity of the lower mantle with the upper mantle and transition-zone viscosity fixed to 1021Pas. In each of these calculations the lithospheric thickness has been held fixed to
G L A C I A L ISOSTATIC A D J U S T M E N T A N D C O A S T A L T E C T O N I C S
9
Fig. 3. Predictions of the present-day rate of rsl rise using the VM2 viscosity model and the ICE-4G deglaciation history. Results are shown for analyses performed that both exclude (top plate) and include (central plate) the influence of rotational feedback and (bottom plate) the difference between these predictions, which isolates the influence of the changing rotation alone.
10
W. R. PELTIER 11018 H U O N PEN. P A P U A N G
5004 BARBADOS
0
I .
.1
. ** r |"
' -21
3.5
' ~
0
I
i
-7
I
I
0
.
.
.
' J[
o
-14
.
.........................................
"~-~l
4055 C L I N T O N CT
-~4
'-~1
11002 TOKYO
BAY JAP
'
10
§
................................................
-21
i
-14
,I -7
,
0
I
-21
0
o ' -21
-14
0
Time (ka BP)
-7
L-
~-
-7
-14
0
11083 A D E L A I D E S A U S T
5038 RIO C O L O R A D O A R G E N T I N A
'1
o. . . . . . . .
' -21
o T ....... i
'
-14
,I
-7
i
i
0
Time (ka BP)
Fig. 4, Relative sea-level histories at six significant locations illustrating the negligibly weak impact of rotational feedback upon the Holocene record. The Barbados and Huon Peninsula sites are the two primary locations from which coral-based records are available. For the latter site only the raw data are shown, uncorrected for any influence of tectonic uplift. The final four sites are located as close as possible to the centres of the degree two and order one structure that characterizes the contribution of rotational feedback on rsl history. It is at these locations that the influence of this feedback is maximum. The continuous curve in each frame represents the rsl history that includes rotational feedback whereas the dashed curve is that obtained excluding this effect. The inset in each plate represents the difference between these two histories on a scale that ranges between - 3 . 5 m and +3.5m. Both sets of calculations were performed with the ICE-4G (VM2) model. 120.6km. In Fig. 7a and b the range of the observed values of these properties of Earth's present-day rotational state (see Peltier & Jiang (1996b) and Peltier (1997), for a full discussion) is shown as the hatched region. Inspection of these
results will demonstrate that both of these rotational observables are fitted by precisely the same viscosity model, namely that labelled VM1 in Fig. 1, as the value of lower-mantle viscosity preferred by both observations is seen to be
G L A C I A L ISOSTATIC A D J U S T M E N T A N D COASTAL T E C T O N I C S
11
Fig. 5. Time-dependent topography of the northern hemisphere of the planet from LGM to present according to the ICE-4G (VM2) model of the GIA process. These results include the contributions from both explicit and implicit ice.
12
W. R. PELTIER
Fig. 6. Inundation maps for the Bering Strait, Australia-Papua New Guinea and the Indonesian Archipelago. Also shown is a map centred on the Greek archipelago showing the regions that would have been dry land at LGM (shown as beige) but which are today sea covered.
2 • 10 21 Pas. Because these rotational observables depend upon entirely independent components of the moment of inertia tensor (see equations (9)) it is highly unlikely, in my view, that they could be reconciled by precisely the same model of the radial viscoelastic structure if both observables were not primarily controlled by the GIA process. Furthermore, these rotational data are sensitive essentially to the average value of the viscosity from the base of the lithosphere to the core-mantle boundary (cmb). This may be seen by computing the Frrchet kernel FKR(r) for either of the rotational data, as this appears in the expression for the perturbation of an arbitrary measure of the response 6R
that is induced by a perturbation in the viscosity model 6 logl0 u(r) as 6R =
r~FKR(r)6 logl0u(r) dr
(17)
in which b and a are the radii of the crab and the Earth, respectively. The Fr~chet kernels for each of the rotational data, computed at the model VM1, are shown in Fig. 8c. Because these functional derivatives for both )2 (J2) and polar wander speed (PW) are very slowly varying functions of radius r it is clear from equation (17) that these data determine essentially the average value of u(r) through the
G L A C I A L ISOSTATIC A D J U S T M E N T A N D COASTAL TECTONICS
13
Fig. 7. (a))2 as a function of lower-mantle viscosity//LM when the upper-mantle and transition zone viscosity is fixed to the value rUM = 1.0 • 1021Pa s and the lithospheric thickness is L = 120.6 km. Results are shown for both of the models of glaciation history for which inertia perturbations are shown in Fig. 3. (b) Polar wander speed as a function of lower-mantle viscosity VLM, with other parameters as in (a).
mantle. Because VM1 fits both observations it seems clear that this m o d e l has the correct value o f this Earth property. However, it is also clear that this m o d e l does n o t provide an acceptable fit to all data related to the G I A process. A n extremely i m p o r t a n t example of a dataset that is n o t fitted by the VM1 m o d e l consists of the relaxation spectrum for F e n n o s c a n d i a n r e b o u n d originally inferred
by M c C o n n e l l (1968). His analysis of the strandline data for the post-glacial recovery o f this region led h i m to infer the variation o f relaxation time (shown as its inverse) as a function of spherical h a r m o n i c degree I shown in Fig. 9, in which the star symbol plotted adjacent to the low-degree asymptote at a relaxation time near 4600 years represents the relaxation time inferred by a M o n t e Carlo derived exponential fit to
14
W. R. PELTIER 0.0 G r-
$ -5.0
•
PW J - ' ~ "
n=15
n=25 -lO.O BATHURST
0
n=40
-15.0
Co) -20.0 I 0.5
I 0.8
I 1.0
(c) ! 0.8
-
0.5
1.0
0.5'
! 0.8
1.0
D i m e n s i o n l e s s radius Fig. 8. Fr6chet derivatives for a representative set of the data related to the GIA process. (a) shows a sequence of kernels for the inverse relaxation times of a number of spherical harmonic degrees of the McConnell relaxation spectrum based upon the analytical formula of Peltier (1976). (b) shows Frchet derivatives for the site-specific relaxation times at sites near the centre of Laurentide rebound (Bathurst Inlet; this is actually a high Arctic site) and at the center for Fennoscandian rebound (the Angerman River site). (e) shows kernels for the non-tidal acceleration of rotation ()2) and polar wander speed (PW) that were determined numerically using the procedure embodied in equations (Sa) and (8b) of the text. Inspection of this suite of kernels, all of which were computed on model VMI, which is employed as starting model in the Bayesian inversions, demonstrates that the observables whose sensitivity to viscosity variations they represent offer the potential of significant resolution from the Earth's surface to the crab. the rsl record at Angerman River, which is located near the centre of Fennoscandian rebound. Also shown in this figure are the theoretically predicted spectra for the VMI, VM2 and VM3 viscosity models shown in Fig. 1. Inspection will show that the relaxation spectrum predicted by VM 1 is such that relaxation times are too long at all spherical harmonic degrees, implying that upper-mantle and transition-zone viscosity is too high. That it is in fact the viscosity over this range of depths to which the Fennoscandian relaxation spectrum is sensitive is demonstrated in Fig. 8a, which shows Fr6chet kernels for these relaxation time data for several values of the spherical harmonic degree, computed on the basis of the exact mathematical formula for them given by Peltier (1976). It will be clear by inspecting the spectrum for the VM2 model, also shown in Fig. 9, that the softer upper mantle and transition zone that characterize VM2 allow this model to fit the McConnell data extremely well.
The final set of data employed in the formal Bayesian construction of VM2 consists of a set of 21 relaxation times inferred on the basis of Monte Carlo fits of an assumed exponential uplift curve to individual rsl histories at sites that were once ice covered and at which the rsl records are distinctly exponential in form. Fifteen of these sites are in Canada and six in Sweden and Norway. These locations and the data from them are discussed in detail by Peltier (1996, 1998b). Examples of the Fr6chet derivatives for such site-specific relaxation time data are shown in Fig. 8b for the Angerman River site in Sweden discussed previously and for the Bathurst Inlet site in the Canadian high Arctic. Inspection of these functions demonstrates that the rsl data controlled by the post-glacial recovery of the Laurentian platform are most sensitive to the viscosity in the upper part of the lower mantle, whereas those from Fennoscandia are most sensitive to shallower transition-zone structure.
GLACIAL ISOSTATIC ADJUSTMENT AND COASTAE TECTONICS
1.2
I
I 7
I
I
15
I
it t o five beaches
1.0
Fit 0.8
Lo
six beaches
#t Angermanland (raw data)
,e,,,i
s
0.6 V
O
U
~
.
~
.
,/"
-el
0.4 ,-..4
0.2
_~"'~*~.~__ - - .~- - - -_.=.~..~..~-~ """
cD
VM1
M. & e'. 1 9 9 7
0.o -0.2 .0
I
I
I
I
1.2
1,4
1.6
1.8
.
2.0
Fig. 9. The relaxation spectrum for Fennoscandian rebound of McConnell (1968) based upon both five and six strandline inversions. Also shown as the star symbol adjacent to the low wavenumber asymptote of the spectrum is the inverse relaxation time inferred from Monte Carlo fit to the rsl history at the Angerman River location, which is located near the centre of Fennoscandian rebound. Theoretical predictions of the spectrum are shown for models VM1, VM2 and VM3 in Fig. 1 as well as for model MF of Mitrovica & Forte (1997).
On the basis of this discussion it should now be clear why the formal Bayesian inversion of the totality of the above-described data deliver VM2 when VM1 is employed as the initial estimate. Because VM1 is too stiff in the upper mantle and transition zone to fit the McConnell spectrum, the viscosity in this region must be reduced. However, this reduction reduces the mean viscosity of the mantle, which is unacceptable to the rotational data. The latter data therefore require that the lower-mantle viscosity be increased to restore the mean value of viscosity to that in VM1. This adjustment occurs primarily in the lower part of the lower mantle because model VM1 fits the site-specific relaxation time data from the Hudson Bay region very well, meaning that the viscosity in the upper part of the lower mantle, to which these data are most sensitive (see Fig. 8), is held fixed near that in VM1, namely 2 x 102a Pa s. It will serve no useful purpose here to review the formal mathematical procedure employed in the Bayesian inversion which delivers VM2 from the VM1 first guess. The interested reader will find detailed discussions of this procedure in the studies by Tarantolla & Valette (1982, 1984), Jackson & Matsu'ura (1985) and Backus (1988). It will be useful, however, to end this discussion of the procedure employed to deduce the radial viscosity profile by illustrating the range of VM2-type models that may be derived by
applying variations on the basic procedure. Examples of such profiles are shown in Fig. 10a and b. In Fig. 10a the two versions of VM2 differ from one another only because the site-specific relaxation time data employed in the inversion are deduced from the envelope sampled form of the rsl curves as in the archive of Tushingham & Peltier (1992) or from the raw age-height pairs directly. The former procedure leads to the version of VM2 shown as the heavy continuous line in Fig. 10a, whereas the latter procedure leads to the thin continuous line. These two variations of VM2 are clearly very close to one another. In Fig. 10b the raw data based version of VM2 from Fig. 10a is compared with a new version in which the forward predictions of the site-specific relaxation times were made using the version of the sea-level equation in which the full influence of time-dependent ocean function was included, an influence that was neglected in the inversions for which results are shown in Fig. 10a. Inspection of Fig. 10b will show that in this most accurate of the inversions the viscosity in the upper part of the lower mantle is somewhat elevated above 2 x 1021pas but only very slightly. In this model the ratio of the viscosity in the upper part of the lower mantle to that in the upper mantle and transition zone is approximately five. It will be useful to end this section with a very brief discussion of the relation between models in
16
W. R. P E L T I E R 22.5
I
I
I
I
I
I
l
22.0
21.5 VM1 21.0
20.5
- raw data
r~
(a) I
20.0 ~ I.,=4
o oO g
Eo0
22.5
22.0
o
-
VM2 C(O
21.5
VM2
21.0
L-----------"
-
f
20.5 (b) raw site data 20.0 3400
I
4000
I 4600
l 5200
I 5800
6400
Radius ( k m ) Fig. 10. Viscosity profiles determined by simultaneous formal Bayesian inversion of the Fennoscandian relaxation spectrum of McConnell (1968), the site-specific relaxation times from 23 ice-covered sites in Canada and Fennoscandia, and the non-tidal acceleration of the rate of axial rotation. In (a) the dashed line is the VM 1 viscosity profile employed as starting model in the inversion process, and two versions of the final model are shown as the dark and light continuous lines, respectively, these being distinct versions of VM2. The former of these two inferred models has been obtained using site-specific relaxation times obtained from fits to the envelope sampled data compiled by Tushingham & Peltier (1992), whereas the latter was obtained on the basis of sitespecific relaxation times deduced from the raw data themselves. In (b), where the dashed line again indicates VMI, the two versions shown are the final VM2 model (that based upon use of the raw data to determine the sitespecific relaxation times) and a further version in which the starting model predictions were made with the version of the model that included the full influence of time dependence of the ocean function. Incorporation of the latter effect in the forward model slightly decreases the forward predictions for the site-specific relaxation times and therefore slightly increases the inferred viscosity in the upper part of the lower mantle, essentially back to the value of 2 • 1021 Pa s that is characteristic of the starting model VM1 in this region. the class V M 2 to those which have recently been a d v o c a t e d by other workers. The closest o f these other models to V M 2 is that derived by L a m b e c k et al. (1990) by trial-and-error fits to a set o f rsl curves from F e n n o s c a n d i a . Their m o d e l is iden-
tical to V M 2 t h r o u g h o u t the u p p e r m a n t l e and transition zone, approximately a factor o f two higher in viscosity in the u p p e r part o f the lowerm a n t l e (4.5 • 1021 Pa s), and essentially equal in the lower part o f the lower m a n t l e (in fact, these
GLACIAL ISOSTATIC ADJUSTMENT AND COASTAL TECTONICS workers quoted a range of allowed lower-mantle viscosities of (2-7) x 1021 Pa s). In the analyses presented herein, it is the data from the Hudson Bay region of Canada that collectively require the viscosity to be somewhat lower in the upper part of the lower mantle. Mitrovica & Peltier (1995) suggested that the site-specific rsl data from the Hudson Bay region exhibit a significant spread in relaxation times ranging from a low near 2000 years (Ottawa Islands) to a high near 7600 years (Richmond Gulf). Advocates of higher values of viscosity in the upper part of the lower mantle than that in VM2 (e.g. Mitrovica 1996; Forte & Mitrovica 1996; Mitrovica & Forte 1997; Simons & Hager 1997; to be referred to collectively in what follows as MFSH) have focused entirely
1.2
I
17
upon the Richmond Gulf record and ignore all the rest, an approach which introduces significant bias as there is no reason to believe that the Richmond Gulf record is superior to any of the others. It has, in fact, become clear in the course of recent analyses (Peltier 1998c) that the previously published high estimate of the relaxation time at Richmond Gulf (Mitrovica & Peltier 1995) is in error. Rather than being near 7600 years as suggested by Mitrovica & Peltier (1995), it is in fact best estimated as 3400 4- 400 years on the basis of a complete reanalysis of all available 14C data from the southeast Hudson Bay region when these are properly transformed onto the sidereal timescale using the Calib. 3.0 software of Stuiver & Reimer (1993). The VM2 model
I
I
I
a) 1.0
Fit to six b e a c h e s *
0.8
Angermanland
v~
(raw data)
,if,' ,'
IQ
0.6
noo
'
'
'
'
'
Depth
o
/
..... ._.._.._. .........
0.4 i
m o.2
Q)
I
o.o E-4
1.2
1.(
*.=4
I
I
1.6
1.B
I
I
2.0
i
L2
I
I
(11
b)
0
n~ 1.0
I Fit to six b e a c h e s
N ~
I
1.4 ,-,
*
0.0
Angermanlemd
(raw data)
~" ~ q t . ~ _ ' ~ "~1 I]--' ~ L li
~
4) m 0.6
,~
191
t* i
eoo
*
I
J
'
Depth
o
,..4 0 4 *
,9. . . . ~ . . ~ . . - . I : ' . . . . - L - . . . 4 - ... . - . . - .I. T . . . -' . . ~ - - - ~t " "
0.2
0.0 1,0
I 1.2
! 1.4
I, 1.6
I 1.8
2.0
Logto(degree) Fig. ll. (a) The inverse relaxation time spectrum for Fennoscandian rebound of McConnell (1968) compared with the prediction of a model in the VM2 class and compared with the predictions for two additional models that differ from VM2 by the presence of a 70 km thick layer immediately overlying the 660 km discontinuity in which the viscosity is reduced either by one or two orders of magnitude from the value near 0.45 x 1021Pa s that otherwise obtains in this region of VM2.
18
W . R . PELTIER
predicts precisely this relaxation time in southeast Hudson Bay near the centre of uplift. It is therefore clear on this basis that the inferences of viscosity presented by MFSH are untenable. Further evidence of this fact is clear on the basis of analysis of the upper-mantle and transition-zone viscosity structures of the MFSH models. Those models of the shallow structure include both extremely soft transition zones and higher-viscosity upper mantles. Structures of this kind would appear to be ruled out entirely by the McConnell (1968) spectrum for Fennoscandia rebound as demonstrated in Fig. 9, where the spectrum of the Mitrovica & Forte model (denoted MF) is compared with those of the VMX models as well as with McConneU's data. If models of this type are to be entertained it is clearly incumbent upon their advocates to prove that the McConnell (1968) spectrum is very significantly in error. Otherwise the VM2 model must be strongly preferred, a model which is similar to that earlier advocated by Lambeck et al. (1990) although with significantly lower contrast in viscosity across the spinel-post spinel phase transition at 660 km depth. It is very important to realize, however, that the VM2 family of models may in no sense be construed to represent a 'uniquely' preferred solution to the one-dimensional mantle viscosity inverse problem. Although one can argue this point formally, it is probably more useful to demonstrate it by providing a specific example. To this end, Fig. 1 l a shows results for the McConnell spectrum obtained using a perturbed version of the VM2 profile in which a thin low-viscosity layer is inserted into the structure immediately above the 660 km seismic discontinuity, with the viscosity in this layer being fixed to either 102oPa s or 1019 Pa s. The presence of such a structure has been inferred to be required in viscosity models derived by inversion of the non-hydrostatic geoid anomalies that are supported by the mantle convection process (e.g. Forte et al. 1993a, b; Pari & Peltier 1995). Even though models of the perturbed kind shown in Fig. 11 a are essentially identical with those required by these data, insofar as the radial variation of viscosity is concerned, it is clear that such models do not fit the McConnell (1968) relaxation spectrum and they are therefore untenable insofar as the GIA data are concerned. However, it is possible to further perturb the structure so as to recover the good fit to the McConnell spectrum as shown in Fig. 11b. The results shown in this figure demonstrate that the presence of the soft layer may be easily accommodated simply by increasing the viscosity of the rest of the transition zone back towards the value in VM1, namely 1021 Pas. Models of this kind would not be ruled out by the data yet they
differ significantly from VM2. As the transition zone is rich in garnet and as this mineral has a high creep resistance, one may be tempted to argue that such models are actually to be preferred. For present purposes, however, these analyses are presented simply to demonstrate the extreme degree of non-uniqueness in the radial variation of viscosity that the data allow.
Model-data intercomparisons for rsl history As discussed in the last section, very few of the rsl data that are actually available have been employed to tune the radial viscosity profile of the model; in fact, the only data used in this way are those from sites located near the centres of the Laurentian and Fennoscandian ice sheets, and from Barbados, where the coral-based record of Fairbanks (1989), which extends to LGM, has been used to constrain the total ice melt in the deglaciation model. All of the remaining data may therefore be employed to verify the quality of the ICE-4G (VM2) model. In the discussion to follow, the focus of the first subsection will be upon rsl data that actually constrain the time-dependent elevation of the shoreline and that usually derive from 14C dating of mollusc shells or wood specimens whose indicative meaning in the landscape suggests that the sample records a former level of the sea. In the second subsection the focus will shift to the U/Th-dated coral records of rsl history mentioned in the Introduction.
Relative sea-level histories beyond the ice sheet margins The 14C data that will be employed in this subsection to test the quality of the ICE-4G (VM2) model constitute an extremely small subset of the c. 600 individual rsl records that are now contained in the database at the University of Toronto. This database has yet to be published and constitutes a considerable improvement upon the reconnaissance collection of Tushingham & Peltier (1991, 1992). Rather than being based upon sampling the envelope of the set of age-height pairs that are derived from each sample, the new data base consists of the raw data themselves and it is these data that will be employed for comparison purposes herein. When these data are compared with the predictions of the theoretical model we must of course transform from 14C time to sidereal time, and for this purpose we employ the Calib. 3.0 program of Stuiver & Reimer (1993), which links the extensive tree ring database for the
GLACIAL ISOSTATIC ADJUSTMENT AND COASTAL TECTONICS Holocene to the coral-based U/Th calibration of Bard et al. (1990) for the late glacial period. One of the most important regions in which the model may be subjected to rigorous test comprises the entire eastern seaboard of the continental USA. This is because even the earliest analyses based upon the VM 1 model (Peltier et al. 1986; Tushingham & Peltier 1992) demonstrated that there were significant misfits between the observations and the predictions such that the rates of sea-level rise predicted to be occurring as a result of the collapse of the proglacial forebulge were much higher than observed in this geographical region. That these misfits are essentially completely eliminated by ICE-4G (VM2) is demonstrated in Fig. 12, which compares observations with theoretical predictions at Montreal, Quebec (which was located just inside the ice sheet margin), Boston, Massachusetts (which was located very close to the ice sheet margin itself), Brigantine, New Jersey (which was located outboard of the ice sheet near the crest of
240.
MONTEEAL, QUE, CANADA I~ !
l
120f BOSTON,MA.USA ,
-50
I
BRIGANTINE. NJ,USA
iif
#
LUIIPUTCREEK.NC. USA
-2o
-',~
-'~0
-o
Time (ka)
Fig. 12. Examples of 14C-dated sea-level curves from four sites on the east coast of the North American continent: Montreal (Quebec) in Canada, and Boston (Massachusetts), Brigantine (New Jersey) and Lilliput Creek (North Carolina) in the USA. The carbon dates for the individual samples have been converted to sidereal years using the U/Th-based calibration of Stuiwer & Reiner (1993) and the raw data, corrected in this way, are compared with the predictions of the ICE-4G (VM2) model.
19
the forebulge) and Lilliput Creek, North Carolina. All of these data are well fitted by the theoretical predictions, demonstrating, as previously documented by Peltier (1996), that the previously evident misfits are eliminated. This is rather important because the modification to VM1 to produce VM2 primarily involves a reduction of the creep resistance in the upper mantle and transition zone that was required to fit the rsl data from Fennoscandia. As this modification to the structure also allows the model to reconcile North American data, this suggests that the upper-mantle and transition-zone viscosity below North America are essentially the same as beneath Northwestern Europe. A global representation of the marked difference in the rsl history predicted by the VM 1 and VM2 viscosity models when the ICE-4G deglaciation history is employed in the calculation is provided in Fig. 13, which shows the present-day predicted rate of rsl rise for both models as well as the difference between them. The difference between these predictions is clearly largest along the US east coast, in precisely the region where the misfits of the VMl-based theory to the observations were largest. Probably the best location in the world from the perspective of the quality of the post-glacial rsl data that are available from it, however, is the British Isles. This region is also especially interesting because it is not only located in the region of forebulge collapse that surrounds the previously glaciated region of Fennoscandia but it was also glaciated in the north, where a significant mass of ice was located over the highlands of Scotland. Furthermore, the coastline of this region experienced significant variation subsequent to LGM, when a vast land bridge connected Britain to France. Figure 14 shows the locations of four sites from which high-quality rsl data are available, superimposed upon the inundation map which illustrates the way in which the coastline is predicted to have evolved based upon the ICE-4G (VM2) model. Figure 15 compares predicted and observed rsl history at Tay Valley and North Solway Firth, Scotland, both sites in the northern region that was once ice covered, and at the Fenlands and Bristol Channel locations in the south, which remained ice free. Inspection of these comparisons clearly demonstrates that the ICE-4G (VM2) model very accurately predicts even the very complex and highly non-monotonic rsl histories that obtained in the northern region, where a complex interplay occurs between the process of post-glacial rebound of the crust that causes sea level to fall and rising sea levels caused by the melting of distant ice sheets. The
20
W.R.
PELTIER
Fig. 13. Present-day rates of sea-level rise predicted using the VM1 and VM2 viscosity models in conjunction with the ICE-4G deglaciation history. The difference between the predictions of these models is maximum along the US east coast, as shown in the final part of the figure.
GLACIAL ISOSTATIC ADJUSTMENT AND COASTAL TECTONICS
21
Fig. 14. Inundation map for the British Isles on which are superimposed the locations of the two sites in Scotland and the two sites in England for which the rsl data are described in the text. TV, Tay Valley; SF, Solway Firth; F, Fenlands; BC, Bristol Channel.
fact that the ICE-4G (VM2) model fits these data extremely well is important because it has been previously argued that the data from this region required a rather high value of the viscosity of the lower mantle, in fact near 1022Pas (Lambeck et al. 1996), and therefore considerably in excess of the viscosity over the same range of depths that had been inferred previously on the basis of rsl records from the immediately adjacent Fennoscandia region (Lambeck et al. 1990). Clearly, models with such large radial viscosity contrasts are not, in fact, required by the data from the British Isles. The interested reader will find a far more detailed discussion of the post-glacial sea-level history of the British Isles in the study by Peltier & S h e n n a n (1998).
Moving further still away from the main centres of glaciation into the region that I have previously referred to as the 'far field' of the ice sheets, Fig. 16 compares predicted and observed rsl histories at a sequence of sites that extends from the Caribbean Sea along the east coast of the South American continent to the northern part of Argentina. The first plate in this sequence once more reproduces the fit to the U/Th-dated coral sequence from Barbados, a record that was actually employed to tune the ice load component of the model (Peltier 1994) as mentioned previously. Moving further south to Recife and Santos-Itanhaem, in Brazil, and Bahia Solano, in Argentina, we note that the most apparent characteristic of the records of rsl history, both observed and predicted, is the existence of a
22
W. R. PELTIER TAY VALLEY. SCOTLAND
BARBADOS "/5 L
I
I
/ -75 _
I I NOR'IN SOLWAYFIRTH. SCOTLAND
[
-150 RECIFE, BRAZIL
0
'~
I~NI.ABDS, I~GIAh~D
'
'
t
o
-10 f -20
I
SANTOS-ITANHAEM. BRAZIL
,/1
-10 f I
-20
BRISTOL CHANNEL, BIq'GLAND
4
BAHIA ~OI~NO, ARGEIfI~A
t
2 -20 0 -20
-10 Time
-
0
(ka)
Fig. 15. Examples of 14C-dated rsl curves from four sites in the British Isles: Tay Valley and Solway Firth in the once ice-covered region of Scotland, and the Fenlands and Bristol Channel, which were both beyond the southernmost extent of the Scottish ice sheet. As discussed in the text, the records from Scotland are highly non-monotonic, because of the superimposition at these locations of the influence of rebound of the crust owing to ice removal and the influence of the collapse of the Fennoscandian forebulge and continuing addition of mass to the oceans caused by the melting of both Laurentide and Fennoscandian ice.
mid-Holocene high stand that occurs at c. 5 ka BP. The fit to the observed highstand is excellent at Santos-Itanhaem, acceptable at Bahia Solano but less so at Recife where a significant phase shift appears to exist. As will be discussed in detail elsewhere, the southernmost east coast of Argentina, Patagonia, appears to be experiencing some tectonic uplift, presumably because of the increasingly close proximity to the Chile trench as one moves further southwards along the east coast. In the final sequence of intercomparisons, shown in Fig. 17, the mid-Holocene highstand of sea level continues to be the most prominent feature of the records of rsl history. These records are all from the Pacific Ocean sector, and from north to south correspond to Osaka Bay, Japan, Rota Island in the Marianas, Balding Bay, Australia, and Christchurch, New
-g -20
15
-10 Time
-5 (ka)
Fig. 16. Examples of 14C-dated sea-level curves from four sites along the east coast of the South American continent from Barbados in the north to Recife and Santos-Itanhaem on the coast of Brazil, and to Bahia Solano in Argentina. It is interesting to note that although the data from the northern sector of the coast are fairly well fitted by the ICE-4G (VM2) based theory, as one moves further to the south the data show increasing evidence of the action of the influence of tectonic uplift (a detailed analysis of this tilting effect will be provided elsewhere).
Zealand. The theoretical prediction of the sealevel history at each of these locations is again dominated by the existence of a mid-Holocene highstand with maximum height above present sea-level occurring at c. 5 ka BP and achieving an amplitude of c. 2 m. As previo,usly mentioned, it was the prediction of this feature which is extremely well expressed across the entire Pacific Ocean, that attracted such attention to this work when solutions of the sea-level equation were first reported by Clark et al. (1978) and Peltier et al. (1978). A more recent discussion of the data that constrain this feature from the archive of Tushingham & Peltier (1992) was presented by Mitrovica & Peltier (1991). All of these records, most of which are constrained by a very small number of data points, are reasonably well explained by the theory. In the next subsection, we will consider a series of records which have been construed to pose a considerable challenge to the global theory.
GLACIAL ISOSTATIC ADJUSTMENT AND COASTAL TECTONICS OSAKA
BAY,JAPAN'
23
Eustatic Sealevel Rise
T 11P/.8
-45-304 NOrAI5~1[
0G
,~ I
, ,//..///'"/"" """"""..................
/,,"
~ oo.4
~IIIlRI~ItI~C]I I , NEW
'2
f -4o -20
-so
J15 -lo
1~ _ ~ - 7 ~ -5
t
_
o
T i m e ~ka)
Fig. 17. Examples of 14C-dated sea-level curves from four sites in and on the margin of the Pacific Ocean, ranging from Osaka Bay in the north to Christchurch, New Zealand, in the south. At the intermediate sites of Rota Island and Balding Bay, Australia, the mid-Holocene highstand of sea level is especially apparent.
Coral-based records of rsl change." the Pacific Ocean sector At several points in the above discussion, I have drawn attention to the fact that the ICE-4G (VM2) model was tuned in terms of the total ice amount so that it would fit the U/Th-dated coral-based rsl record from Barbados. The reason for focusing upon this record is that it is unique from a number of points of view. First, it is the only accurately dated rsl record which extends back to LGM and which therefore can be used to constrain the amount by which sea level rose from that time to the present. Second, however, is the fact that this record is composed almost entirely of a g e - d e p t h measurements on the coral species Aquapora palrnata. As this Caribbean species is known to live at a depth that is within 5 m of sea level, this record may be assumed to constitute a good recording of changing sea level itself. Of course, it is also well known that Barbados is rising at a rate that is usually assumed to be near 0.35 mm per year, implying that to derive actual sea level from the raw a g e - d e p t h data for the Barbados sequence
Time (kaBP)
%, ,, ,, ,, ~ 9 , 0 Fig. 18. The total eustatic sea-level rise is shown from the model that both excludes (T t) and includes (T) the influence of implicit ice as discussed in the text in connection with equations (14)-(16). Also shown are the individual contributions from North America (N, Nt), Eurasia (E, E'), Antarctica (A) and Greenland (G, G').
one needs to make a tectonic correction. The magnitude of this correction is therefore such as to require sea level to have risen at Barbados since LGM by c. 7 m more than suggested by the raw data. The position that I will adopt for present purposes, as I have done in the past, is that if the theoretically predicted rsl history at Barbados is such that the prediction goes through the raw data points themselves then, given the observed range of living depths of A. palmata, no further correction to the data is required. This is particularly true as the best estimate of the rate of tectonic uplift at Barbados is probably closer to 0.25 mm per year than to 0.35 m m per year. The fit of the ICE-4G (VM2) model to the Barbados data has been shown previously in Figs 4 and 14, where the predicted rise of sea level at this location is seen to be just slightly less than 120 m. It is important to note that this is slightly greater than one might expect based upon the net mass of 'explicit' ice that is melted across the glacial-interglacial transition, though one must remember that determination of the rise of sea level at any given site requires a correction to the effect expected on the basis of mass alone (see equations (4) and (5)). The sea-level rise based upon the consideration of ice-mass alone (both explicit and implicit as discussed previously), for each of the main sectors in which melting occurs, is shown in Fig. 18,
24
W. R. PELTIER
inspection of which shows that one would expect sea level to rise by only 106.7 m if only the explicit ice melted in ICE-4G were to be added uniformly to ocean basins of fixed present-day area. On the other hand, when the influence of implicit ice is properly taken into account, the discrepancy is much reduced, as the predicted total eustatic rise increases to 117.8 m. Figure 19 shows the locations of the additional sites in the Pacific basin that will be of interest for the remainder of this subsection, namely, the Huon Peninsula of Papua New Guinea, Tahiti and Sumba and Morley Islands. At each of these locations coral-based records of rsl change are also available for which the age control is based upon U/Th dating (Edwards 1988). It is clearly interesting to enquire as to whether or not the global theory is also able to reconcile these additional observations, the most important of which are probably those from the
Huon Peninsula (Edwards 1988; Chappell & Pollach 1991; Ota et al. 1993). On the basis of their observations of the uplifted Pleistocene interglacial terraces near the Kwambu-Kilasairo location, Chappell & Polach (1991) inferred a late Pleistocene average rate of tectonic uplift at Huon of 1.9mm per year. In Fig. 20a, I show not only the raw data for Huon but also the corrected data when this rate of tectonic uplift is employed to reduce them, along with the theoretical prediction based upon the ICE-4G (VM2) model. Whereas the raw data lie somewhat above the theoretical prediction at young age, it is clear that when the data are corrected by assuming the conventional rate of tectonic uplift they lie considerably below the theory, so far below as to suggest that perhaps the theory may be in error (see comments by Edwards (1995)). Also shown in Fig. 20, however, are comparisons between observations and theoretical prediction
PACIFIC ISLANDS
it
!i
1 ~' r,
4
I
'
'
II,"
I1
"k
:
t
~
I
t
i
I-
i
!"~
.
,,I . . . . . . "
i8
I
9
I
__, ....
,'"7~'~ ~ " 4.,. Ittt:5 I "IPu5 .P"~
/"-
__~
'
41-43"~""
"t~
_
"~
'~
_
ill
i.,i.~_u
i t............. tAWli~lI,
!
1
I
, 1 , I
I
!
1 ~,,l~
.........
~ . 67 " 't, _lio .06,,'] ..... ~. . . . . . !
I
~7
!
~.
!
9 .. . . . .
;
. . .
1
~I,,
' .
i
z
i ., .
i
i
:
I ,
s,
Is,
_~"~
#'L
i
i
~"~.
. . . . . ,! $IJMBtl ~ _
r
~!
.....
.
._
!
1Z
I~'h/j ',"l
(,,J ~,.,~
'in-i
.i
I
i
I
,i" Lo'd,
I ~
h i
.
fo
i
l ,
- - 7 i .............. t
ti l__i
~)~'~ i
9
i I
'
tit 9, ,
,
.-.~ ~
T ....
'rAttrzl
I
o
i !
4~
i ,
i I
!
~,, 100it
i 12lib
14013
16013
180
i I
i
I
160W
140W
Fig. 19. Location map for the equatorial Pacific Ocean region showing the locations of the islands from which information on rsl history is available. The data from those islands for which names are given are discussed in the text. The numbering scheme for the other locations is that employed in the global recompilation of rsl data that has been continuing at Toronto for the past several years.
GLACIAL ISOSTATIC ADJUSTMENT AND COASTAL TECTONICS HUON PEN. PAP&U NG
20,
I,
,I '
I.
- (Vt=l.Omm/yr)
.
.I
,,.
~
'
r
i
Tk,I-II~
20
I
I
~
-40
.40
-00
.60
-OO
.80 I
I
I
L00 I
I
MORLEY IS. I
I
I
I
I
/
~
"~ % -
.20
I, ' /
l
.living depth corrected ! (Bard, pets. comma.)
-20
-100
25
I
I
9
20
1
ki-
Z
,
,
,
I
!
1
SUMBA IS.
]
I
(V,..?)
, I
I
o
(Vt=O.35mm/yr)
0
-
-30
'
,li
~.
--
-20 [.,.
.
.
-40, -BO
-40
f
-80
"
- 8 0 [-100 -24
-80 I -20
-16
-12
-8
I -4
-100 0
-P.AI
I -20
-16
-12
-O
I -4
Time (ka) Fig. 20. The raw and tectonic uplift corrected coral-based records are shown, along with the predicted sea-level histories based upon model ICE-4G (VM2) for (a) the Huon Peninsula, (b) Papeete Harbour, Tahiti, (e) Morley Island and (d) Sumba Island. for the rsl history at Morley Island (Eisenhauer et al. 1993) and Sumba Island (Bard et al. 1996b),
whose locations are also shown in Fig. 19. Inspection of the intercomparisons at these sites demonstrates that at these locations theory and observations agree rather well, although in the case of Sumba Island the data are sparse (essentially only two points). At Sumba there is an estimate, noted in the figure, of the rate of tectonic uplift active at this location, and it will be noted that when this correction is applied to the data the fit of the two data points to the theory is excellent. At Morley Island the observations and theory agree very well if allowance is made for a small living depth correction. Both of these additional datasets are consistent with the existence of a mid-Holocene highstand in the sea-level record but neither data set actually resolves the feature. It is nevertheless clear that when the Huon data are 'corrected' by assuming that the conventional Pleistocene rate of tectonic uplift also
applies during the Holocene period then the fit of the theory to the data is anomalous in that it is so poor. The most apparent anomaly concerns the absence of the mid-Holocene highstand in the rsl record when the 1.9 mm per year tectonic uplift rate is used to make the correction. It seems clear on this basis that the 1.9mm per year average rate of Pleistocene uplift at the Kwambu-Kilasairo site on the H u o n Penninsula is not characteristic of the Holocene period. To determine the rate that has actually been characteristic of this most recent epoch we might best proceed by asking what is the rate that must be assumed to minimize the misfit between observations and theoretical prediction. The results of this analysis are shown in Fig. 21. Depending upon whether one minimizes the misfit over the entire dataset or only over the last 9000 years one infers a best rate of tectonic uplift between 0.3 and 0.65mm per year, although the minimum in variance is rather flat so that the rate is not accurately determined. As suggested by Peltier
26
W. R. PELTIER PEN. P A P A U NO
HUON
4000,
I
I
I
I
I
I
I
I
I
b e s t Vt(all data, solid line)=0.3 m m / y r
/
3000.
0
2000.
v *,-s
1000,
0.0
0.1
0.2
0.3
0.4
0.5
0.8
0.7
0.8
0.9
1.0
Uplift Rate Vt (mm/yr) 35
I
I
I
Vt=O.65 mm/yr o
A
o
E d w a r d s et al---
Chappell & Polach--, Ota et al(Kwambu-Kilasairo)--, -70 -16
,,
-12
I -8
I -4
0
Time (ka) Fig. 21. (a) Variance reduction between the prediction of the rsl history on the Huon Peninsula for the ICE-4G (VM2) model and the observations as a function of the rate of tectonic uplift assumed for the purpose of correcting the observations. Variance reduction is shown based upon the use of all available data and the use of only data younger than 9 ka. With the former assumption the 'best-fitting' rate of tectonic uplift is found to be 0.35 mm per year. (b) Fit of the tectonically corrected data to the ICE-4G (VM2) prediction. (1995), it therefore seems clear that the rate of tectonic uplift at H u o n during the Holocene period has been much less than the Pleistocene average rate. One final piece of analysis that warrants discussion here, concerning the tectonic contribution to Holocene sea-level history, relates to the data for Tahiti (Bard et al. 1996a), which are also shown in Fig. 19. At Tahiti, unlike at Huon, the theoretically predicted rsl history delivered by the ICE-4G (VM2) model lies above the observations rather than below. However, rather than being subject to tectonic uplift, Tahiti is subject to tectonic subsidence. This is perhaps primarily a consequence of the fact that the
island is emplaced in lithosphere that sinks is it cools while moving away from the East Pacific Rise at a rate near 12 cm per year. Now the rate of subsidence that has been suggested to be characteristic of Tahiti is near 0.2 mm per year, but when this correction is applied to the data it does not significantly improve the fit to the observations. When an additional correction is made for the influence of living depth (E. Bard, pers. comm.) the misfit is further reduced (see the figure). However, the data still do not contain the clearly evident mid-Holocene highstand of sea level that is observed to be characteristic of all rsl data from far field locations. Further analysis of the data at Tahiti is clearly warranted.
G L A C I A L ISOSTATIC A D J U S T M E N T A N D COASTAL T E C T O N I C S On the basis of all of the data discussed in this subsection, it is clear that w h e n the model is tuned to fit the history of post-glacial sea-level change at Barbados, as well as those from icecovered sites, then the m o d e l also fits far field rsl histories at locations that were not e m p l o y e d to tune the m o d e l parameters. The model must therefore be considered highly successful.
Conclusions In the previous sections of this paper, I have both reviewed and extended the global theory of post-glacial sea-level change that has been developed over the past two decades. This theory is n o w rather fully articulated and has been shown to reconcile a wide range of geophysical and geological data. One of the most active areas of current application involves the investigation of the extent to which space geodetic data m a y be b r o u g h t to bear to further constrain m o d e l parameters. Three different types of such data have n o w been shown to be useful adjuncts to the geological and astronomical m e a s u r e m e n t s that have been the conventional focus in such work. These consist o f observations of the nontidal acceleration of axial rotation based u p o n the use of L A G E O S satellite laser ranging data (Peltier 1983; Y o d e r et al. 1983), observations o f the rate of radial displacement based u p o n the use of V L B I observations (Argus 1996) and, most recently, observation of tangential motions associated with the G I A process based u p o n global positioning system (GPS) observations ( B I F R O S T 1997). Application of these measurem e n t systems is expected to prove to be especially useful in u n d e r s t a n d i n g the relative contributions of tectonics and glacial isostasy to individual records o f rsl history.
References ARGUS, D. F. 1996. Postglacial rebound from VLBI geodesy: on establishing vertical reference. Geophysical Research Letters, 23, 973-977. BACKUS, G. E. 1988. Bayesian inference in geomagnetism. Geophysical Journal of the Royal Astronomical Society, 92, 125 142. BARD, E., HAMELIN,B., ARNOLD,M., MONAGGIONI,L., CABIOCH, G., FAURE, G. & ROUGERIE, F. 1996a. Deglacial sea level record from Tahiti corals and the timing of global meltwater discharge. Nature, 382, 241 244. , , FAIRBANKS, R. G. & Z1NDLER, A. 1990. Calibration of the 14C timescale over the past 30,000 years using mass spectrometric U-Th ages from Barbados corals. Nature, 345, 405-409.
27
BARD, E., JOHANNIC, C. et al. 1996b. Pleistocene sea levels and tectonic uplift based on dating of corals from Sumba Island Indonesia. Geophysical Research Letters, 23, 1473 1476. BIFROST 1996. GPS measurement to constrain geodynamic processes in Fennoscandia. EOS Transactions of the American Geophysical Union, 35, 377. BILLS, B. G. & JAMES, T. S. 1996. Late Quaternary variations in relative sea level due to glacial cycle polar wander. Geophysical Research Letters, 23, 3023-3026. CHAPPELL, J. • POLACH, H. A. 1991. Post-glacial sea level rise from a coral record at Huon Peninsula, Papua New Guinea. Nature, 276, 602-604. CLARK, J. A., FARRELL,W. E. & PELTIER,W. R. 1978. Global changes in postglacial sea level: a numerical calculation. Quaternary Research, 9, 265-287. DAHLEN, F. A. 1976. The passive influence of the oceans upon the rotation of the Earth. Geophy-
sical Journal of the Royal Astronomical Society, 46, 363-406. DZIEWONSKI, A. M. & ANDERSON, D. L. 1981. Preliminary reference earth model. Physics of the Earth Planetary Interiors, 25, 297-356. EDWARDS, R. L. 1988. High precision Th-230 ages of corals and the timing qf sea levelfluctuations in the late Quaternary. PhD Thesis, California Institute of Technology, Pasadena. - - 1 9 9 5 . Paleotopography of Ice-Age ice sheets. Science, 267, 535-536. E1SENHAUER, A., WASSERBURG, G. J., CHEN, J. H., BONANI, G., COLLINS, L. B., Znu, Z. R. & WYRWOLL, K. H. 1993. Holocene sea level determination relative to the Australian continent: U-Th (TIMS) and a4C (AMS) dating of coral cores from the Abrolhos Islands. Earth Planetary Science Letters, 114, 529-547. FAIRBANKS, R. G. 1989. A 17,000-year glacio-eustatic sea level record: influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation. Nature, 342, 637-641. FANG, M. & HAGER, B. H. 1995. The singularity mystery associated with radially continuous Maxwell viscoelastic structure. Geophysical Journal International, 123, 849 865. FARRELL, W. E. & CLARK, J. A. 1976. On postglacial sea level. Geophysical Journal of the Royal Astronomical Society, 46, 647-667. FORTE, A. M., DZIEWONSKI,A. M. & WOODWARD,R. L. 1993a. Aspherical structure of the mantle, tectonic plate motions, nonhydrostatic geoid and topography of the core-mantle boundary. In: LE MOUEL, J.-L. (ed.) Dynamics of the Earth's Deep Interior and Earth Rotation, Geophysical Monograph Series, American Geophysical Union, 72, 135-166. , PELTIER, W. R., DZIEWONSKI, A. M. & WOODWARD, R. L. 1993b. Earth's dynamics topography: a new interpretation based upon mantle flow models derived from seismic tomography. Geophysical Research Letters, 20, 225-228. HAM, D. & WAHR, J. 1995. The visco-elastic relaxation of a realistically stratified Earth, and a further analysis of postglacial rebound. Geophysical Journal International, 120, 287-311.
28
W. R. P E L T I E R
HAYS, J. D., IMBRIE, J. & SHACKLETON, N. J. 1976. Variations in the Earth's orbit: pacemaker of the ice ages. Science, 194, 1121-1132. IMBRIE, J., HAYS, J. D. et al. 1984. The orbital theory of Pleistocene climate: support from a revised chronology of the marine 5180 record. In: BERGER, A. L. et al. (eds) Milankovitch and Climate, Part 1, D. Reidel Publishing Company, 269-305. JACKSON, D. D. & MATSU'URA, J. 1985. A Bayesian approach to nonlinear inversion. Journal of Geophysical Research, 90, 581-591. KARATO, S.-I. & WU, P. 1993. Rheology of the upper mantle: a synthesis. Science, 260, 771-778. LAMBECK, K., JOHNSTON, P. & NAKADA, M. 1990. Holocene glacial rebound and sea-level change in NW Europe. Geophysical Journal International, 103, 451-468. , SMITHER, C. & NAKADA, M. 1996. Glacial rebound of the British Isles III: Constraints on mantle viscosity. Geophysical Journal International 125, 340-354. MCCONNELL, R. K. 1968. Viscosity of the mantle from relaxation time spectra of isostatic adjustment. Journal of Geophysical Research, 73, 7089-7105. MITROVICA, J. X. 1996. Haskell [1935] revisited, Journal of Geophysical Research, 101, 555-569. & FORTE, A. M. 1997. Radial profile of mantle viscosity: results from the joint inversion of convection and postglacial rebound observables. Geophysical Journal International, 102, 2751-2769. & PELTIER, W. R. 1989. Pleistocene deglaciation and the global gravity field. Journal of Geophysical Research, 94, 13 651-13 671. & 1991. On postglacial geoid subsidence over the equatorial oceans. Journal of Geophysical Research, 96, 20 053-20 071. & 1995. Constraints on mantle viscosity based upon the inversion of post-glacial uplift data from the Hudson Bay region. Geophysical Journal International, 122, 353-377. OTA, Y., CHAPPELL, J., KELLEY, R., YONEKURA, N., MATSUMOTO, E., NISHIMURA,T. & HEAD, J. 1993. Holocene coral reef terraces and coseismic uplift at Huon Peninsula, Papua New Guinea. Quaternary Research, 40, 177-188. PARI, G. & PELTIER, W. R. 1995. The heat flow constraint on mantle tomography-based convection models: towards a geodynamically selfconsistent inference of mantle viscosity. Journal of Geophysical Research, 100, 12 731-12 751. PELTIER, W. R. 1974. The impulse response of a Maxwell Earth. Reviews of Geophysics and Space Physics, 12, 649-669. - - 1 9 7 6 . Glacial isostatic adjustment II. The inverse problem. Geophysicals Journal of the Royal Astronomical Society, 46, 669-706. 1982. Dynamics of the ice-age Earth. Advances in Geophysics, 24, 1-146. --1983. Constraint on deep mantle viscosity from LAGEOS acceleration data. Nature, 304, 434-436. -
-
-
-
-
-
1985. The LAGEOS constraint on deep mantle viscosity: results from a new normal mode method for the inversion of viscoelastic relaxation spectra. Journal of Geophysical Research, 90, 9411-9421. 1994. Ice-Age paleotopography. Science, 265, 195-201. -1995. Paleotopography of Ice-Age ice sheets. Science, 267, 536-538. --1996. Mantle viscosity and ice-age ice sheet topography. Science, 273, 1359-1364. - - 1 9 9 7 . Correction to the paper 'Glacial isostatic adjustment and earth rotation: refined constraints on the viscosity of the lower mantle', by W. R. Peltier and Xianhua Jiang, Journal of Geophysical Research, 102, 10 101-10 103. 1998a. Global sea level rise and glacial isostatic adjustment. Global and Planetary Change, in press. - - 1 9 9 8 b . The inverse problem for mantle viscosity. Inverse Problems, in press. - - 1 9 9 8 c . 'Implicit Ice' in the global viscoelastic and topographically self-consistent theory of glacial isostatic adjustment: implications for ice-age paleo-topography, Geophysical Research Letters, in press. & ANDREWS, J. T. 1976. Glacial isostatic adjustment I. The forward problem. Geophysical Journal of the Royal Astronomical Society, 46, 605-646. & JIANG, X. 1996a. Glacial isostatic adjustment and earth rotation: refined constrains on the viscosity of the deepest mantle. Journal of Geophysical Research, 101, 3269-3290. & 1996b. Mantle viscosity from the simultaneous inversion of multiple data sets pertaining to postglacial rebound. Geophysical Research Letters, 23, 503-506. & -1997. Mantle viscosity, glacial isostatic adjustment and the eustatic level of the sea. Surveys in Geophysics, 18, 239-277. - - - , FARRELL, W. E. & CLARK, J. A. 1978. Glacial isostasy and relative sea level: a global finite element model. Tectonophysics, 50, 81-110. - - - , FORTE, A. M., MITROVICA,J. X. & DZIEWONSKI, A. M. 1992. Earth's gravitational field: seismic tomography resolves the enigma of the Laurentian anomaly. Geophysical Research Letters, 19, 1555-1558. --, HILLAIRE-MARCEL, C. & PARI, G. 1998. Postglacial rebound at Richmond Gulf (Quebec), mantle viscosity, and the free air gravity anomaly over Hudson Bay, Nature, submitted. PLATZMAN,G. W. 1971. Ocean tides and related waves. In: REID, W. A. (ed.) Mathematical Problems in -
-
-
-
-
-
-
-
the Geophysical Sciences 2. Inverse Problems, Dynamo Theory and Tides. Lectures in Applied Mathematics, American Mathematical Society, 14, pp. 239-291, SHACKLETON, N. J. 1967. Oxygen isotope analyses and Pleistocene temperatures re-addressed. Nature, 215, 15-17. --, BERGER, A. & PELTIER, W. R. 1990. Transac-
tions of the Royal Society of Edinburgh. Earth Sciences, 81,251-261.
GLACIAL ISOSTATIC ADJUSTMENT AND COASTAL TECTONICS SIMONS, M. & HAGER, B. 1997. Localization of the gravity field and the signature of glacial rebound. Nature, 390, 500-504. STUIVER, M. & REIMER, R. J. 1993. Extended 14C data base and revised Calib. 3.0 14C age calibration program. Radiocarbon, 35, 215-230. TARANTOLA,A. & VALETTE,B. 1982. Inverse problems = quest for information. Journal of Geophysics, 50, 159-170. & 1984. Generalized non-linear inverse problems solved using the least squares criterion. Reviews of Geophysics and Space Physics, 20, 219-232. TARASOV, L. & PELTIER, W. R. 1997. Terminating the 100kyr ice-age cycle. Journal of Geophysical Research-Atmospheres, 102, D18, 21 665-21 693. TUSHINGHAM, A. M. & PELTIER, W. R. 1991. ICE-3G: a new global model of late Pleistocene deglaciation based upon geophysical predictions of postglacial relative sea level change. Journal of Geophysical Research, 96, 4497-4523.
--
-
29
& 1992. Validation of the ICE-3G model of Wfirm-Wisconsin deglaciation using a global data base of relative sea level histories. Journal of Geophysical Research, 97, 3285-3304. Wu, P. & PELTIER, W. R. 1983. Glacial isostatic adjustment and the free air gravity anomaly as a constraint on deep mantle viscosity. Geophysical Journal of the Royal Astronomical Society, 74, 377-449. & 1984. Pleistocene deglaciation and the Earth's rotation: a new analysis. Geophysical Journal of the Royal Astronomical Society, 76, 202-242. YODER, C. F., WILLIAMS, J. G., DICKEY, J. 0., SCHUTZ, B. E., EANES, R. J. & TAPLEY, B. D. 1983. Secular variation of the Earth's gravitational harmonic J2 coefficient from LAGEOS and non-tidal acceleration of Earth rotation. Nature, 303, 757-762. -
Decoupling post-glacial tectonism and eustasy at Huon Peninsula, Papua New Guinea JOHN
C H A P P E L L 1, Y O K O
OTA 2 & COLIN
CAMPBELL 3
1Research School of Earth Sciences, Australian National University, Canberra, A.C.T., Australia (e-mail:
[email protected]) 2Department of Geography, Senshu University, Kawasaki, Japan 3 Research School of Pacific and Asian Studies, Australian National University, Canberra, A.C.T., Australia Abstract: Late Quaternary uplift of coral terraces varies along the coast at northeast Huon Peninsula, Papua New Guinea, but the uplift rate has been assumed to be constant at any given locality in previous studies. Measurements indicate that rates for the last 7000 and 120 000 years were similar but this may be coincidence, because uplift at Huon Peninsula is dominated by isolated, metre-scale events with recurrence intervals around 1000 years. Using age-height data from 54 corals from the post-glacial reef, we examine the uplift rate over the last 13 000 years near Kwambu, where facies changes in a drill core indicate several uplift events before 7 ka BP. To separate uplift from sea level, a eustatic curve for Kwambu was generated by the global sea-level model described by Lambeck, recalibrated to new, Late Pleistocene sea-level data. With Barbados as a test case, predictions compare well with observations reported earlier, but predicted sea levels for Kwambu cannot be reconciled with the coral data unless the water depth of coral growth at the site was greater than estimated previously.
Two important questions in regions of plate convergence are whether vertical movements are intermittent, and whether the mean rate is constant or varies on 104-105 year time scales. Intermittent uplift indicates that vertical movement is associated with large earthquakes and probably with fault movement; variation of uplift rate can reveal something about the way in which the geometry of relatively shallow structures changes in response to plate motion. Constancy of uplift over 105 years usually is assessed by comparing uplift rates based on raised mid-Holocene and Last Interglacial shorelines, because eustatic sea levels are believed to have been similar, within a few metres, at those times (Chappell & Veeh 1978). Variations over the last 104 years can be evaluated from postglacial shorelines provided that local eustatic changes during this period are known. In fact, it has been shown that Holocene uplift has been intermittent and dominated by metre-scale events, with recurrence intervals of hundreds to thousands of years, in many Pacific countries including Japan, New Zealand, Papua New Guinea and the US west coast (Ota 1991; Berryman 1993; Ota et al. 1991, 1993; Merritts 1996). Similar coseismic uplifts were identified from Late Pleistocene regressive terraces at H u o n Peninsula, Papua New Guinea (Chappell et al. 1996b). In this paper we attempt to separate postglacial uplift and eustatic changes at Huon
Peninsula, and to identify whether uplift rate has varied significantly in the last 13 000 years. Post-glacial sea-level changes at H u o n Peninsula were derived by Chappell & Polach (1991) and Ota & Chappell (1998), on the basis of radiocarbon age measurements from post-glacial coral reefs including observations from a 52m drill core. These studies assumed that the local uplift rate at each sampling site has been constant since the Last Interglacial (119-126 ka). This assumption may be faulty, because Ota et al. (1993) found that the uplift rate at Kwambu on the central Huon coast was about l m p e r l 0 0 0 years for the last 2000 years but averaged about 2 m p e r 1000 years for the last 6000 years. This may be a statistical artefact, however, because uplift at Huon Peninsula is dominated by metre-scale events with an average recurrence interval around 1000 years (Ota et al. 1993; Chappell et al. 1996b). By extending the record to 13 000 years, we attempt to resolve this question.
Methods and data This study is based on age-height and age-depth data from the raised post-glacial barrier reef near Kwambu, Huon Peninsula (Fig. 1). The barrier reef was formed during Post-glacial sea level rise, which culminated locally at about 6-6.514Ckabp (6.8-7.3 cal. kaBP).
CHAPPELL,J., OTA, Y. & CAMPBELL,C. 1998. Decoupling post-glacial tectonism and eustasy at Huon Peninsula, Papua New Guinea. In: SXEWA~a',I. S. & VrrA-FINz~, C. (eds) Coastal Tectonics, Geological Society, London, Special Publications, 146, 31-40.
32
J. C H A P P E L L E T
AL.
Fig. 1. Site locations on raised Holocene coral reef near Kwambu, Huon Peninsula, Papua New Guinea. Drill site is location of drill-hole data reported by Chappell & Polach (1991); base of Holocene reef on section XY is inferred to be at 70 m depth from more recent drilling at same site (F. Taylor, pers. comm.). Numbers 1-7 are sample sites of Ota et al. (1993); K is Kilasairo cliff site of Chappell & Polach (1976).
The reef has emerged in the last 6000 years, in a series of metre-scale coseismic uplift events (Ota et al. 1993). We use age-height and age-depth measurements of corals in the transgressive barrier reef from a 52 m drill core reported by Chappell & Polach (1991) and from surface exposures reviewed by Ota & Chappell (1998), together with data from post-6 ka BP regressive terraces cut into the raised barrier reef (Ota et al. 1993). The average uplift rate at Kwambu, based on thermal ion mass spectrometry scanning (TIMS) U-series age measurements of the crest of the Last Interglacial reef (reef VII), is 1.9mper 1000 years (Stein et al. 1993).
Table 1 lists age-height data from outcrops and regressive terraces; Table 2 lists age, depth and facies information for the drill core. Conventional radiocarbon ages were measured in the ANU Radiocarbon Laboratory and were adjusted for the local 400 year seawater reservoir correction (Chappell & Polach 1991). Adjusted radiocarbon ages were converted to calibrated ages (Stuiver & Reimer 1993) with the OxCal program (Bronk Ramsey 1994). Sample heights were surveyed to tide levels and normalized to mean sea level using tide tables for Dreger Harbour, 80 km southeast of Kwambu, which has a similar tide. Water depths in which the corals grew are discussed later,
T E C T O N I S M A N D EUSTASY, H U O N
PENINSULA
33
Table 1. Holocene outcrop samples from Kwambu area, in ascending age order Sample code*
Context~
A5686 W5 W1 W20 W18 W29 A8670 A5685 W15 W3 W2 W30 A5687 A6119 W14 A1250 W32 W10 W9 W31 A5689 W11 A5683 W19 A1249 A1253 A1248 A1252 A 1251
reg. ter. reg. ter. reg. ter. reg. ter. reg. ter. reg. ter. reg. ter. reg. ter. reg. reef reg. reef reg. reef reg. reef crest crest transg. transg. transg. transg. transg. transg. transg. transg. transg. transg. transg. transg. transg. transg. transg.
Height (m amsl):~
Adjusted 14C age (ka BP)w
Calibrated age
2a 2a 2a 2a 2b 5a 8.5 9.5 4.6 7.4 6.7 7.1 13.2 13.0 4.6 8.0 6.7 6.9 8.7 4.5 6.1 5.1 4.4 5.0 3.1 0.1 1.9 0.8 1.5
2330 2450 2490 2510 3420 3880 4430 5310 5560 5630 5690 5800 5800 6070 6190 6210 6390 6400 6420 6600 6740 6750 6780 7190 7410 7430 7580 7740 7780
2.1-2.6 2.4-2.7 2.4-2.7 2.4-2.8 3.5-3.8 3.9-4.4 4.9-5.3 5.9-6.3 6.2-6.6 6.3-6.6 6.3-6.7 6.4-6.8 6.4-6.8 6.8-7.2 6.9-7.3 6.8-7.2 7.0-7.5 7.0-7.5 7.0-7.5 7.3-7.6 7.4-7.7 7.4-7.8 7.4-7.8 7.8-8.1 7.9-8.4 8.0-8.4 8.1-8.6 8.2-9.0 8.4-9.0
(ka BP)
* Samples were first reported by Chappell & Polach (1976), Ota et al. (1993) and Ota & Chappell (1998); A-codes dated at Australian National University, W-codes dated at Nagoya University. t Context: reg. ter., regressive terrace described by Ota et al. (1993); reg. reef, regressive reef and transg., transgressive reef (last two as defined by Chappell et al. (1996b)). Heights of samples are adjusted because those marked a are from raised surf benches, which form at or above high tide level, and b is from coral heads which grew below low tide level (see Ota et al. 1993). wConventional radiocarbon age with local seawater correction of 400 years subtracted (see Chappell & Polach 1991). Age errors are given in original sources but most are around +80 years at 1 SD.
based on stratigraphic and faunal data summarized in Table 2. To investigate possible variations of uplift, we required a post-glacial eustatic curve for Huon Peninsula that was not based on Huon Peninsula data. High-precision observations from tectonically stable sites in the region do not extend much beyond 6-7kaBP and we have made use of data from Barbados, reported by Fairbanks (1989). Recognizing that eustatic changes at Barbados will differ from those at Huon Peninsula, owing to global post-glacial isostasy, we placed the two sites on an equivalent footing by generating local eustatic curves from a model of Late Quaternary ice volume changes, using the global glacio- and hydro-isostatic procedure described by Nakada & Lambeck (1987). We followed Lambeck (1996) in deriving ice volume changes from isotopic sea levels of Shackleton (1987) but also used
late Pleistocene sea-level data reported by Chappell et al. (1996a) to calibrate the ice volume changes. This task ranged well beyond the scope of the present paper and full details will be published elsewhere. The predicted post-glacial sea-level curve for Barbados matches the observations well (Fig. 2) and thus we expect that sea-level predictions for Kwambu should match values calculated from the coral data, provided that assumptions about uplift rate and estimates of water depth of coral growth are correct.
Stratigraphy The raised post-glacial reef at K w a m b u has a long, n a r r o w lagoon with a barrier reef to seaward
J. CHAPPELL E T AL.
34
Table 2. Kwambu drill-core facies and age data
Depth below core top (m)
1.2 2.5 3.4 4.4 5.7 6.2 7.7 9.0 10.6 11.0 14.2 15.4 17.4 .
.
.
.
.
.
.
.
.
.
.
.
.
.
Dominant taxa*
Coral taphonomyf
Foram guild index:~
Matrixw
Por., Acrop. Acrop Por., Acrop. Pocil. Por. Por., Acrop. Acrop.
b,b&r, C b&r, M broken, C i.s, bored, M
5 4
broken, S i.s, bored, A b, bored, A
Low-rag mud Ar/mgCa sand CorFrag CorFrag Ar/mgCa sand Ar/mgCa mud mgCa cem/mud mgCa/Ar sand
Por. Hal., mol Por., Acrop Por., A. hyac. Acrop., F. pal.
i.s, S broken, S broken, C i.s, bored, C i.s, bored, A
.
.
.
.
.
.
.
.
.
.
Acrop., Pocil Por., F. pal. Gonio. Monti. Monti.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
19.5 20.2 21.3 22.8 23.7 24.5 25.3 26.1 29.2 30.6 32.0 33.1 34.0
Monti. Por., Monti Por., Hal., Monti., Sty. Por., Acrop Por., Acrop Por.
broken, S i.s, bored, M e.s, A i.s, bored, C i.s, bored, C cor. frags, M e.s, bored, A broken, C i.s, bored, C b&b, M broken, M broken, (,M i.s, C
36.8 38.4 39.8 42.5 42.7 45.2 50.0 51.4
Por., F. pal. Por., Pocil Por., Cyph Acrop., F. pal. Monti., Cyph Por., T. mus Pocil. Monti.
e.s, M i.s, C i.s, A broken, M i.s & e.s, M i.s & e.s, C i.s, M e.s, M
2 0.6 1.5 1.5 2.5 1.3
.
.
.
.
.
.
.
1 2 1.5 1.5 1.5 0.6 0.6 0.6 0.8 3 6 0.3 2
1.5 0.7
.
7.6-7.9 8.0-8.3 8.2-8.4 8.4-8.7 8.3-8.6 8.5-9.0
Ar/mgCa sand Ar/mgCa sand mgCa/Ar sand mgCa cem/sand mgCa cem
1 1.4 4 .
Calib. age (ka BP)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Ar/mgCa sand mgCa mud Ar/mgCa mud mgCa mud mgCa cem mgCa/Ar mud mgCa cem mgCa cem Ar/mgCa sand Ar/mgCa mud mgCa cem mgCa cem mgCa cem/sand mgCa mud/sand mud/sand ] / II
8.3-8.7 9.0-9.4 9.4-9.6 9.4-9.8 .
.
.
.
.
.
.
.
.
.
.
.
.
.
9.4-9.8 9.6-10.0 9.8-10.0 9.6-10.0 10.0-10.4 9.9-10.2 10.1-10.6 10.3-10.9 10.3-10.9 10.3-10.9 10.0-10.4 10.5-11.0 10.9-11.7 11.0-11.7 12.5-12.8 12.0-12.5 12.0-12.6 12.4-12 12.8-13.2
* Dominant taxa: Acrop., Acropora; A. hyac., A. hyacinthus; Cyph., Cyphastrea; F. pal., Favia pallida; Monti, Montipora; Pocil., Pocillopora; Por., Porites; Sty., Stylaster; T. mus., Tubipora musica; Hal., Halimeda; mol, molluscs. t Coral taphonomy: b,b&r, bored, broken and rolled; b&r, broken and rolled; i.s, in situ; e.s, ex situ but not obviously broken. Coral abundance scale: A, abundant, >70% of core section; C, common, 50-70%; M, moderate, 20-50%; S, sparse, merely present to 20%; N, none. Foram guild index = ratio of mobile : sedimented species (see text). wMatrix: mgCa, magnesium calcite, Mg typically 10-20%; Ar, aragonite; CorFrag, coral fragments; cem, cement. II Matrix rarely recovered below 39 m owing to change of drilling technique. Horizontal dotted line indicates facies break. and a fringing reef to landward, and laps against Late Pleistocene coral limestone (Fig. 1). The fringing reef platform rises to 13 m above m e a n sea level (amsl) and the barrier crest varies from 6 to 9 m amsl. Drilling in 1988 on the K w a m b u barrier penetrated 5 2 m into post-glacial coral limestone (Chappell & Polach 1991) and fur-
ther drilling at the same site in 1996-1997 showed that the post-glacial reef is about 7 0 m thick (F. Taylor, pers. comm.). Exposures further southeast indicate that the post-glacial structure wedges out against the Pleistocene raised reef sequence and this is expected to be the case at K w a m b u , also.
TECTONISM AND EUSTASY, HUON
PENINSULA
35
Fig. 2. Predicted eustatic curve for Barbados generated by Nakada-Lambeck global glacio-eustatic model (open circles: for details, see text), and uplift-corrected observations from Fairbanks (1989) with time scale in Th-U years from Bard et al. (1990).
Exposures in coastal cliffs, river-cuts and borrow pits show that the post-glacial reef is composed of several facies: coarse coral limestone of the reef crest and upper forereef, coralgal limestone of the reef platform, lagoon deposits of bioclastic grainstone containing dispersed corals, and coral limestone containing arborescent and less robust corals of the lagoon fringing reef. Each facies has its own characteristic fabric and guilds of corals and molluscs (Pandolfi & Chappell 1994). Stacked shallow-water coral deposits exposed in coastal cliffs, particularly at Kilasairo Stream, 2km southeast of the drill-hole site, suggest that reef growth kept pace with rising sea level (Chappell & Polach 1976; Pandolfi & Chappell 1994). Coral limestone throughout the drill core from the Kwambu barier was interpreted by Chappell & Polach (1991) as crest or upper forereef deposits of a shallow water 'keep-up' reef that kept pace with rising post-glacial sea level. However, compared with cliff exposures, the drill core provides only a small sample of reef material and the water depth of reef growth is difficult to gauge. Corals in the core were identified by J. E. N. Veron, who noted that they are common types of the reef crest and upper forereef: dominantly Porites, Acropora, Montipora, Pocillopora and Favids. Transitions occur in the core from assemblages dominated by strong, upward-branching corals, to assemblages with massive and tabular corals surrounded by coral and algal debris,
which suggest changes of habitat and possibly water depth. Variations of foraminiferal assemblages, matrix and cement within the coral limestone also may indicate changes of habitat and water depth. Some 43 taxa of benthic forams were identified in the core (16 Miliolinae, 25 Rotaliinae, 2 Textulariinae). The number of taxa per sample range from six to 19 and the proportions of foraminiferal guilds vary markedly (an ecological guild is a grouping of organisms which employ similar life strategies, regardless of taxonomic relationship). In particular, we found significant variations in the ratio of the guild of mobile species to the guild of species restricted to sediments, which appear to correspond to other changes in the core. Variations of this particular index are listed in Table 2 along with major coral taxa and summaries indicating composition of the coral limestone matrix. Changes of facies at about 10 m, 18 m and 34 m, shown as dotted lines across Table 2 represent simultaneous occurrences of a sharp up-core increase of the foram guild index, an increase of matrix cement and an increased incidence of bored corals with heavy algal rinds.
Reef growth, uplift and sea-level rise On the basis of the drill-core data, Chappell & Polach (1991) estimated that the coral crest was generally 2-5 m below sea level throughout, and considered that the age-depth data define the
36
J. CHAPPELL E T AL.
course of relative sea-level rise at Kwambu. Using seawater-corrected conventional radiocarbon ages, Chappell & Polach (1991) adjusted the age-height data for uniform uplift at 1.9mper 1000 years, and suggested that the results represent the local post-glacial eustatic curve. Ota & Chappell (1998) interpreted post-glacial sea-level changes in the same way but used calibrated radiocarbon ages and a larger dataset, which includes observations from other Holocene sites at Huon Peninsula. These workers included the Chappell & Polach estimates of water depths in deriving local eustatic sea levels, but did not attempt to identify the effects of intermittent coseimic uplift. Age-height data from Tables 1 and 2 are plotted in Fig. 3 using the calibrated radiocarbon time scale, together with eustatic predictions for Kwambu at the same time-points as the coral samples, derived from the global model that generated the Barbados curve shown in Fig. 2. Also plotted are differences between observed heights and predicted sea levels, and a sloping line representing uniform uplift at 1.9m per 1000 years. The height (H) relative to present sea level of a coral depends on sea level (S), the depth of water (W) when it grew, and the uplift history: H=S-
W+*U.d*
(1)
where uplift U is integrated from the present (0) to the calibrated age of the coral (t). Previous workers assumed that U is constant, i.e., *U.d*=Ut and therefore H - S = U t - W . It follows that values (H - S) should lie distance W below the uplift line in Fig. 3. However, most points around 10-12kaBP are offset from the uplift line by over 10 m and range to 18 m, which is much larger than the 2-5 m estimate of W by Chappell & Polach (1991). It is possible that W ranged to 18m, or U was not constant, or the predicted sea levels are wrong. Errors in predicted sea levels cannot be ruled out but we consider it unlikely that they range to 18 m and, to account for the differences between ( H - S) and the uplift line, we examine the effects of variable uplift rate and water depth by holding one or the other constant. Thus, the apparent uplift rate (Uw) at constant water depth, over time interval *t = tl - t2 can be calculated by assuming that the predicted sea levels are correct: Uw = (112 -
$2 -
H=
(2)
+ S1)/*t
To examine variation of apparent uplift over the last 13 000 years we calculated Uw over short intervals by stepping through the age-height
thousand years before present 0
2
4
6
10
8
12
14
o
1~I
9
.
~ "::%.
~
0
-10
-20
E -30
r'l
%un00 i DO
-40
-50
-60
o
-70
Fig. 3. Age-height and age-depth plot showing coral data listed in Tables 1 and 2 (O) compared with predicted eustatic sea levels at Kwambu for the same time points ([~). O, show height differences between corals and predicted sea levels; sloping line represents uniform uplift at 1.9 m per 100 years.
TECTONISM AND EUSTASY, HUON 0
2
4
6
I ....
t ....
PENINSULA
8
37
10
12
14
10
r/)
E
,
~
*1
. . . . .
I.
.;
9
\;-.
....
-5
-10
-15
thousand years B P. Fig. 4. Apparent uplift rates with water depth assumed constant, assessed over short time intervals between paired sample age-height data (smoothed).
data. Samples were ranked by stratigraphic depth and data were smoothed until moving average ages increased with moving average depth. Uw values for adjacent pairs in the smoothed series are plotted in Fig. 4 and show considerable variation, with several negative peaks between 7 and 11 kaBP. Similarly, the
apparent water depth (Wu) with constant uplift rate is calculated: Wu = S -
H+
(3)
Ut
Results plotted in Fig. 5 again show considerable variation, with highest values around 10-11
k a BP.
o
o
14
o1: o
12
;o
o
t0
.g
e
E
o o
o1:
o
F o i
2
.....
~
oo
o
o i
4
o
~o
1"~ 1!
-
i
i
~
6
8
10
.......................
i
12
,,,
14
thousand of years before present
Fig. 5. Apparent palaeo water depths with uplift rate assumed constant, for samples from Tables 1 and 2. Short vertical bars signify episodes when apparent water depth suddenly decreased, perhaps signifying uplift events.
38
J. CHAPPELL E T AL.
Discussion Uplift v. w a t e r depth
Estimates of uplift rate and palaeo water depth obviously are not independent. However, extreme values of Uw and Wu are not equally compatible with stratigraphic data from the study site. Anomalous values of Uw suggest that the assumption of constant water depth is faulty; anomalous Wu values suggest that uplift rate varied. Negative values of Uw in Fig. 4 are anomalous because they imply subsidence, for which no evidence has been found at northeast Huon Peninsula. Uplift has been recorded after historical earthquakes but not subsidence (Pandolfi & Chappell 1994). We are confident that prehistoric subsidence would be recorded in coastal land-forms and reef deposits. Detailed studies at more than 20 sites revealed excellent evidence for repeated uplifts during the last 6000 years but no trace of subsidence was found, nor were any subsidences recognized in the detailed record of uplift events for the interval 3 0 - 5 5 k a a P (Ota et al. 1993; Chappell et al. 1996b). It seems unlikely that processes should have been very different, 6000-10 000 years ago, and we dismiss the assumption that water depth was constant. Turning to apparent water depth, most values of Wu between 2 and 7 ka BP are 2-4 m (Fig. 5). These clearly are anomalous because the samples are of attached corals on regressive terraces, formed as intertidal platforms or surf benches, where the real water depth of coral growth was very close to zero. The anomalies reflect the fact that uplift was not uniform but proceeded in isolated, metre-scale events, which are reflected by sharp vertical steps in the Wu series: three occur within the last 7000 years and correspond to uplift events at Kwambu identified by Ota et al. (1993), namely, KK1 at 2.5kaBP; KK5 at 6.2kaBP and KK6 at 6.9 ka BP. Three similar steps occur earlier in the Wu series, at 8.5, 9.6 and 10.5 ka BP We suggest that these also represent uplift events and note that significant facies changes occur in the drill core, close to these times (Table 2). Furthermore, the recurrence interval between these events is close to 1000 years, similar to that of the last 6000-7000 years on the northeast Huon coast (see Chappell et al. 1996b). The question of whether uplift was statistically uniform, though proceeding by sudden events, cannot be answered from the apparent water depths, because the high Wu values prior to 9.5 ka may be realistic. Water depths of 10-15 m are not incompatible with the observed coral
facies in the lower half of the core, which, except for an apparent event at 35 m, is dominated from 24 to 40 m by ramose growth forms with a uniform matrix of high-magnesium calcite muddy sand (Table 2).
C a t c h - u p or k e e p - u p reefs?
The reef at Kwambu was established before 13kaBP and ceased growth around 7kaaP, when it emerged owing to uplift. According to Fig. 5, water depth over the reef at the drill-hole site increased after the reef was established, reached a maximum about 11 000 years ago and has decreased, apparently stepwise, since 10.5kaBP Perhaps rising sea level at first outpaced reef growth but later the reef later caught up, assisted by either a decrease of the rate of eustatic sea-level rise or an increase of uplift rate. However, our drill hole does not necessarily trace the highest point of the growing reef; therefore we cannot identify whether reef growth really has accelerated since about 10kaBP and, if so, what caused it to do so. The question would be resolved with a transect of drill holes, landwards of the drill site on the Kwambu barrier. By sampling basal deposits from the transgressive reef, which should climb to landward with rising sea level, there would be less uncertainty about palaeo water depths. Sea-level predictions thus would be tested more sharply.
Conclusions The Late Quaternary uplift rate of coral terraces at northeast Huon Peninsula, Papua New Guinea, has been assumed by previous workers to have been constant at any given locality, although the mean rate increases southeastwards along the terraced coast. At several sites, measurements show that the local mean uplift rate for the last 6000-7000 years is very similar to that for the last 120 000 years, but this may be coincidence. The 7000 year rate is a statistical average, because uplift at Huon Peninsula is dominated by sudden events of up to several metres with an average recurrence interval around 1000 years, which are considered to be coseismic and were identified and dated from small regressive terraces (Ota et al. 1993; Chappell et al. 1996b). We have investigated the uplift rate over the last 13 000 years at the Kwambu-Kilasairo site, where the rate is known to have diminished during the last 6000 years (Ota et al. 1993), with age-height (or age-depth) measurements
T E C T O N I S M A N D EUSTASY, H U O N of 54 corals from the post-glacial reef structure. By examining uplift and palaeo water depth separately, we conclude that there is no evidence that the uplift rate during the last 13 000 years was significantly different from the 7000 year or 120000 year averages. Variations of a p p a r e n t water depth indicate coseismic uplift events before 7 ka BP, similar to those previously identified for the last 7000 years, at 8.3, 9.5 and 10.3 ka ~p The analysis rests on a eustatic curve for K w a m b u generated by the global m o d e l described by L a m b e c k (1996), recalibrated to new, Late Pleistocene sea-level data. The m o d e l performs well against observations from Barbados reported by F a i r b a n k s (1989) but should be verified at other sites. N o n e the less, the estimate by Chappell & Polach (1991) of 2 - 5 m water depth t h r o u g h o u t the K w a m b u drill-hole record is both imprecise and too restrictive, and it is clear that better accuracy is necessary in studies of this kind. So far, this appears to have been better achieved for Caribbean than for Indo-Pacific reefs, but we suggest that better definition at K w a m b u w o u l d be obtained by extending a transect of drill holes landwards of the single hole, drilled previously. Finally, we note that the H u o n Peninsula drill-hole data examined here have been used previously to infer variations of the rate of postglacial sea-level rise, possibly related to the Y o u n g e r D r y a s event (Edwards et al. 1993). Given the uncertainties about water depth and the episodic nature of uplift at K w a m b u , together with the likelihood that the drill hole does not pass continuously t h r o u g h the highest growing point of the reef, it appears to us that variations of the rate of sea-level rise should not be derived from these data. We thank K. Lambeck, K. Smithers and K. Fleming for their collaboration with the sea-level model used here, which will be reported in full elsewhere.
R e f e r e n c e s
BARD, E., HAMELIN, B., FAIRBANKS, R. & ZINDLER, A. 1990. Calibration of ~4C time scale over the past 30,000 years using mass spectrometric Th-U ages from Barbados corals. Nature, 346, 456-458. BERRYMAN, K. R. 1993. Age, height, and deformation of Holocene marine terraces at Mahia Peninsula, Hikurangi Subduction margin, New Zealand. Tectonics, 12, 1347-1364. BRONK RAMSEY, C. 1994. Radiocarbon calibration and analysis of stratigraphy: the OxCal Program. Radiocarbon, 37, 425 430.
PENINSULA
39
CHAPPELL, J. & POLACH, H. A. 1976. Holocene sea level change and coral-reef growth at Huon Peninsula, Papua New Guinea. Geological Society of America Bulletin, 87, 235-240. - & POLACH, H. 1991. Post-glacial sea level rise from a coral record at Huon Peninsula, Papua New Guinea. Nature, 349, 147-149. -& VEEH, H. H. 1978. Late Quaternary tectonic movements and sea-level changes at Timor and Atauro Island. Geological Society of America Bulletin, 89, 356-368. --, OMURA, A., ESAT, T., MCCULLOCH, M., PANDOLFI, J., OTA, Y. & PILLANS, B. 1996a. Reconciliation of late Quaternary sea levels derived from coral terraces at Huon Peninsula with deep sea oxygen isotope records. Earth and Planetary Science Letters, 141,227-236. , OTA, Y. & BERRYMAN, K. R. 1996b. Holocene and late Pleistocene coseismic uplift of Huon Peninsula, Papua New Guinea. Quaternary Science Reviews, 15, 7-22. EDWARDS, R. L., BECK, J. W., BURR et al. 1993. A large drop in atmospheric 14C/12C and reduced melting in the Younger Dryas, documented with 23~ ages of corals. Science, 260, 962-968. FAIRBANKS, R. G. 1989. A 17,000 year glacio-eustatic sea level record: influence of glacial melting rates on the Younger Dryas and deep ocean circulation. Nature, 342, 637 642. LAMBECK, K. 1996. Sea-level change and shore-line evolution in Aegean Greece since Upper Palaeolithic time. Antiquity, 70, 588-611. MERRrrTs, D. 1996. The Mendocino triple junction: active faults, episodic coastal emergence, and rapid uplift. Journal of Geophysical Research, 101, 6051-6070. NAKADA, M. & LAMBECK, K. 1987. Glacial rebound and relative sea-level variations: a new appraisal.
Geophysical Journal of the Royal Astronomical Society, 90, 171-224. O-rA, Y. 1991. Coseismic uplift in coastal zones of the western Pacific rim and its implication for coastal evolution. Zeitschrift fiir Geomorphologie, N.F., Supplementband, 81, 163-179 - & CI4APPELL,J. 1998. Holocene sea-level rise and coral reef growth on a tectonically rising coast, Huon Peninsula, Papua New Guinea. Quaternary International, in press. -- - , KELLEY, R., YONEKURA,N., MATSUMOTO, i~., NISHIMURA, T. & HEAD, J. 1993. Holocene coral terraces and coseismic uplift of Huon Peninsula, Papua New Guinea. Quaternary Research, 40, 177-188. --, HULL, A. G. & BERRYMAN, K. R. 1991. Coseismic uplift of Holocene marine terraces in the Pakarae River area, eastern North Island, New Zealand, Quaternary Research, 35, 331 346. PANDOLFI, J. & CHAPPELL, J. 1994. Stratigraphy and relative sea level changes at the Kanzarua and Bobongara sections, Huon Peninsula, Papua New Guinea. In: OYa, Y. (ed.) Study on Coral
Reef Terraces of the Huon Peninsula, Papua
40
J. C H A P P E L L E T AL.
New Guinea-Establishment of Quaternary Sea Level and Tectonic History. Department of Geography, Senshu University, Kawasaki, 119-139. SRACKLETON, N. J. 1987. Oxygen isotopes, ice volume and sea level. Quaternary Science Reviews, 6 183-190. STE1N, M., WASSERBURG, G. J., AHARON, P., CHEN, J. H., ZHU, Z. R., BLOOM, A. L. & CHAPPELL, J.
1993. TIMS U-series dating and stable isotopes of the last interglacial event in Papua New Guinea. Geochimica et Cosmochimica Acta, 57, 2541-2554. STUIVER, M. & REIMER, P. J. 1993. Extended 14C data base and revised CALIB 3.0 14C age calibration program. In: STUIVER, M., LONG, A. & KRA, R. S. (eds) Calibration 1993. Radiocarbon, 35, 215-230.
Holocene uplift and subsidence of the Helike Delta, Gulf of Corinth, Greece STEVEN
SOTER
Department o f Astrophysics, American M u s e u m o f Natural History, Central P a r k West at 79th Street, N e w York, N Y 10024, U S A (e-mail."
[email protected]) Abstract: The southwestern coast of the Gulf of Corinth, known as Aigialeia, lies in a region
of rapid tectonic uplift and extension. Using age and elevation data from raised relic shorelines, and Lambeck's model for local isostatic sea-level rise, I re-examine the uplift of the coastal footwall block in Aigialeia. The average Holocene uplift rate is 2.4 -4-0.8 m ka -1 , significantly higher than the Quaternary uplift rates associated with the raised terraces near Corinth on the southeastern coast. The footwall movement in Aigialeia consists of coseismic uplift events separating periods of relatively aseismic uplift. A footwall uplift of about 2 m apparently accompanied the earthquake that destroyed and submerged ancient Helike 373 8c. The city was built on a Gilbert-type fan delta adjacent to the area of raised relic shorelines. Using dated samples from bore holes drilled in the delta, I estimate that the delta itself subsided by at least 3 m during the earthquake. The opposition between gradual uplift and coseismic subsidence events apparently resulted in a relatively small absolute net displacement of the delta during Holocene time.
The Gulf of Corinth is a marine basin about 105 km long and up to 30km wide, in central Greece (Fig. 1). It occupies an active asymmetrical rift zone (or half-graben), which undergoes north-south extension in connection with uplift of the northern Peloponnesos (Ori 1989; Armijo et al. 1996). The present rate of extension is about 13 mm/year near the western end of the Gulf, and decreases toward the east (Clarke et al. 1997). The southern margin of the rift zone is marked by a series of W N W trending normal extension faults dipping to the north (Doutsos & Poulimenos 1992; Rigo et al. 1996; Roberts & Koukouvelas 1996). A major onshore element of this tectonic system is the Helike Fault, which has been traced for about 40 km (Mouyaris et al. 1992; Poulimenos 1993; Dart et al. 1994; Stewart 1996; Stewart & Vita-Finzi 1996). During the evolution of the rift zone, seismic activity has progressively moved northward (Ori 1989; Dart et al. 1994) and is now concentrated in the Helike and Aigion Faults, and perhaps in other faults located to the north under the Gulf itself (Bernard et al. 1998). This fault system controls the almost linear southern shore of the Gulf of Corinth. In the western sector, modern footwall-derived Gilberttype fan deltas occur wherever rivers have incised northward through the uplifting mountains of the Peloponnesos (Seger & Alexander 1993; Collier & Gawthorpe 1995). The marine delta slopes are eroded by fissures and rotational slumping (Ferentinos et al. 1988). The coastline between the deltas is dominated by limestone conglomerate footwalls. The modern deltas in-
crease in size along this coast from east to west (the direction of increasing tectonic activity) and consist mainly of coarse-grained deposits. The largest of these deltas, here called the Helike Delta, extends for about 13 km between Aigion and Diakopto in the region of Aigialeia (Fig. 1). It is actually a coalesced fan delta, fed by the three rivers Selinous, Kerynites and Vouraikos. The Helike Delta lies on the hanging-wall block of the Helike Fault, which separates it from the footwall mountains to the south. These in turn consist of older (Plio-Pleistocene) uplifted fan deltas of the same kind (Ori 1989). The most prominent of these exhumed structures is the ancient Kerynites fan delta, whose uppermost topset beds have been elevated as much as 1200 m above sea level (Dart et al. 1994). The active Helike Delta has evolved under the influence of long-term tectonic uplift and earthquake-related subsidence. There is good historical evidence for the latter in the delta. In 373BC, a catastrophic earthquake and seismic sea wave destroyed and submerged Helike, then the principal city of ancient Achaea (Marinatos 1960; Soter & Katsonopoulou 1998). According to Pausanias, who visited the site on the Helike Delta around AD 174, the city was located about 7 km southeast of Aigion. In his description of the disaster, Pausanias (7.24.12) wrote that 'the sea flooded a great part of the land and encircled the whole of Helike. Moreover, the flood from the sea so covered the sacred grove of Poseidon that only the tops of the trees remained visible.' Pausanias is generally regarded as a reliable source. Assuming that the grove of
SOTER, S. 1998. Holocene uplift and subsidence of the Helike Delta, Gulf of Corinth, Greece. In: STEWART,I. S. & VITA-FINZI,C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 41-56.
42
S. SOTER
Fig. l. Aigialeia, on the southwestern shore of the Gulf of Corinth. The Helike Delta, between Aigion and Diakopto, is part of the hanging-wall block of the Helike Fault. The Holocene rate of uplift of this block is calculated based on raised shorelines at Diakopto, Trapeza, Paralia Platanou and Aigeira (Table 1). Uplift is also found for similar sites on the Perachora Peninsula at the eastern end of the Gulf. The inset at lower left shows the central part of the Helike Delta with the locations of bore holes B1 B5 and B18. Seismic fault locations adapted from Koukouvelas & Doutsos (1996).
Poseidon was situated at least a metre above sea level and that the trees were at least 3 m high, this account suggests earthquake-related subsidence by at least 3 m. Pausanias goes on to say that in his time the ruined walls of Helike were still visible in the sea. Other ancient writers (Seneca, Strabo, Ovid) also mentioned the submerged ruins. Aigialeia lies in one of the most seismically active areas in the Mediterranean. In the last 300 years, 12 earthquakes with magnitudes estimated in the range 6-7 have occurred within 25 km of Aigion (Ambraseys & Jackson 1997). Macroscopic anomalies have been observed before earthquakes in this area (Soter 1998). On 26 December 1861, an earthquake of estimated magnitude 6.6 struck the Helike Delta (Ambraseys & Jackson 1997). Schmidt (1862, 1875) reported that the earthquake submerged a coastal fringe 13 km long and up to 200 m wide and left a 2 m scarp and fissure of the same length along the base of the foothills. He also found evidence of extensive soil liquefaction, particularly near the mouth of the Vouraikos River. Water and sand erupted from numerous fissures and 'sand volcanoes' up to 20 m across. One of the
eruptions was violent enough to kill a man working in the fields. Schmidt conjectured that the 1861 earthquake destabilized the entire coastal plain, causing it to slip seaward along the steeply inclined basement rock. He suggested that an event of the same nature but greater magnitude had destroyed ancient Helike in the same location. The mechanism he proposed for these earthquakes has some resemblance to the current view of normal faulting. Schmidt's account of what is now called the Helike Fault is one of the earliest scientific descriptions of a seismic fault. Other faults also contribute to the seismicity of the region. On 15 June 1995 an earthquake of magnitude 6.2 seriously damaged Aigion. Its epicentre was near Eratini on the northern shore of the Gulf (Bernard et al. 1998), but the shock caused widespread sediment failures on the Helike Delta, including submarine landslides, shoreline subsidence as a result of shallow rotational sliding, and soil liquefaction with sand blows through fissures and craters (Lekkas et al. 1996; Papatheodorou & Ferentinos 1996). Papatheodorou & Ferentinos (1996) suggested
UPLIFT AND SUBSIDENCE OF THE HELIKE DELTA that the earthquakes of 373 Bc, 1861 and 1995 all induced the liquefaction of a subsurface horizon, setting in motion the translation and subsidence of the overlying sediments. Koukouvelas & Doutsos (1996) identified the east-west trending Aigion Fault as the source of the 1995 earthquake. They traced its surface break on land for 7 km eastward across Aigion until it disappeared about 1.5kin before the shore. Our 1988 sonar survey (Soter 1998) shows that the Aigion Fault continues offshore for at least another 1 km. However, Bernard et al. (1998) suggested that the 1995 earthquake involved a fault located about 10kin ENE of the Aigion Fault. They concluded that the surface breaks on the Aigion Fault in 1995 were secondary features of the earthquake. In view of the historical evidence for subsidence of the Helike Delta, Soter & Katsonopoulou (1998) began a search for the site of ancient Helike in the sea southeast of Aigion. An extensive sub-bottom and sidescan sonar survey showed no unambiguous signs of a city on or under the sea floor. Accordingly, they shifted the search to the subaerial delta, where they drilled 60 bore holes to obtain sediment core profiles. In almost all of the bore holes located in the upper part of the delta between the Selinous and Kerynites Rivers, they found ceramic fragments in occupation horizons dating from Early Bronze Age through Classical and Roman to Byzantine times (Katsonopoulou & Soter 1997). One surprising result was that virtually all the occupation horizons, including the oldest ones, were located above present sea level. If this was in fact the site of ancient Helike, submerged in 373 Rc, then it appears that the delta was subsequently uplifted tectonically. The geological evidence for tectonic uplift along the northern coast of the Peloponnesos is compelling. From Corinth to Xylokastro, the erosion of an uplifting footwall modulated by Quaternary sea-level oscillations has produced an impressive flight of raised coastal terraces. These dated features show that uplift has proceeded at an average rate of order 1.4 m per 1000 years (m ka l) during the last 350 000 years (Kerauden et al. 1995; Armijo et al. 1996). In Aigialeia, on the southwestern shore of the Gulf, elevated erosion notches and emergent marine fauna provide evidence of Holocene elevation of the footwall behind the Helike Fault. Stewart & Vita-Finzi (1996) used the radiocarbon ages of these features to estimate the average rate of footwall uplift there at about 1.5 m ka -l . Here we re-examine this estimate, using Lambeck's (1995) model for local isostatic sea level and radiocarbon dates calibrated for the
43
reservoir effect of the Gulf of Corinth. The analysis invokes an absolute frame of reference to measure the changing elevation of local sea level and of the footwall block. This allows one to visualize the relationship between the present elevation of a relic shoreline and its absolute elevation when created. The data can then be used graphically to reconstruct a Holocene trajectory for the footwall. It then becomes possible, on the basis of the age and depth of selected core samples from bore holes drilled in the Helike Delta, to obtain a very rough estimate for the subsidence of the delta related to the earthquake of 373 BC.
Modelling absolute sea level and coastal uplift The local sea level is the sum of contributions representing the global eustatic sea level (determined by the amount of glacial ice), the local isostatic adjustment (related to the redistribution of mass between glacial ice and ocean water), and the local vertical tectonic movement of the shore. In the absence of vertical tectonic motion, the 'corrected' local sea level ~-c can be written as ~c = ~e + ~i
(1)
where Ce is the global eustatic sea level and ~i is the local glacio-hydro-isostatic correction, both of which are time dependent (see Lambeck 1995). We adopt the model eustatic sea-level curve of Peltier (1994, fig. 3B), shown by r in Fig. 2 for the last 12000 years. The curve shows the deceleration of sea-level rise around 7kaBP, and the cessation of net glacial melting at 5kaBp. The actual eustatic curve may have additional fine structure (see Blanchon & Shaw 1995) but as this has not been firmly established, I will retain the model of Peltier. Lambeck (1995, 1996) modelled the isostatic corrections for local sea level in Greece and Aegean Turkey during the last 20000 years. According to his results, the corrections appropriate for the Gulf of Corinth are almost identical to those for Kavalla in Thrace. As Lambeck provided a time series of r data for Kavalla (1995, figs. 4d and 6d), one can apply this correction factor to the Gulf of Corinth. Adding Ci to the eustatic curve G, one obtains the corrected local sea-level curve ~c shown in Fig. 2. The absolute frame of reference for these curves is fixed relative to the centre of the Earth, with zero elevation corresponding to the present local sea level.
44
S. SOTER
Fig. 2. Global eustatic sea level Ce from Peltier (1994) and the corrected local absolute sea level ~c for the Gulf of Corinth based on the glacio-hydro-isostatic model of Lambeck (1995). The corrected sea-level curve rises steadily during the last 5000 years, despite the flatness of the corresponding eustatic curve. This is due mainly to local isostatic subsidence of the mantle in response to the post-glacial uplift of distant Fennoscandia. Lambeck (1995) showed that in the absence of vertical tectonic movement, all the coastlines in Greece and Aegean Turkey would experience rising sea levels (i.e., coastal submergence) through post-glacial times. To illustrate the effect of vertical tectonic movement on local relative sea level, we can use a simplified schematic model. Let a vertical sea cliff move steadily upward parallel to itself at rate r. Its surface defines a relative frame of reference that moves vertically with respect to the absolute frame. Imagine a baseline scratched on the cliff face at the position that now coincides with sea level. At any time t the absolute elevation ~b of that baseline mark is ~b = - r t
(2)
where t is positive in the past. In Fig. 3 the line CBA represents the trajectory Cb of the cliff for the case r = 2 m k a -1. Point A locates the baseline at the present time t = 0. In reality, of course, the rate of tectonic uplift will not be constant over 10000 yeaers, particularly where abrupt seismic offsets occur, but the long-term average rate of uplift is still a useful quantity. Later, I will consider discontinuities in the rate of uplift. The curve CEA in Fig. 3 is the corrected sealevel trajectory r for the last l0 000 years, taken
Fig. 3. Model of the corrected local sea level s and the footwall baseline elevation Cb (assuming a constant uplift rate of 2 m per 1000 years). The present elevation h of a relic shoreline above sea level equals the difference between the absolute sea level and the footwall baseline at the time the shoreline was created. The curve HKA is thus the difference between the curves for r and Cb.
from Fig. 2 but smoothed with a ninth-order polynomial. Following the sea level and tectonic trajectories in time, we see that the baseline was submerged from about 9 ka BP (point C) to the present (A). Suppose that wave action and biological activity eroded a notch in the cliff at D when sea level was 5 m above the baseline, at B. Then with the steady ascent of the entire cliff, the notch follows the trajectory D F G , parallel to the baseline track CBA. For thousands of years the notch remains below sea level. But with the continued deceleration in sea-level rise, it eventually reappears above water (at F) and today will be 5 m above sea level (at G). The present absolute elevation A G of the notch thus equals the separation BD between baseline and sea level when the notch was created. One can express the present absolute elevation h of any relic shoreline feature on the cliff face (or footwall) as h -- r - Cb
(3)
the difference between the absolute sea level and the baseline at the time the shoreline feature was created. In the upper part of Fig. 3 the changing interval between sea level and baseline (r - ~ b ) is replotted relative to the present sea level (e.g. IJ = BD). The curve H K A represents the function h(t) for the case of constant tectonic
UPLIFT AND SUBSIDENCE OF THE HELIKE DELTA uplift with r = 2 m k a -1. It traces the rise and fall of sea level with respect to the reference frame of the sea cliff. At about 9 ka BP, the rapidly rising sea level overtakes the baseline (at C). Any relic shoreline produced at that time would now have zero elevation (H), and any older relic shoreline features would still be submerged. In this model, the relic shoreline with the highest possible present elevation (at K) was produced when sea level (at E) reached its maximum distance above the baseline. In the example shown in Fig. 3 this occurred about 6000 years ago. It follows that any relic shoreline found today at a given elevation above sea-level (at G for example) can have two possible ages (corresponding to J and L), depending on whether it was produced on the rising or falling leg of the relative sea-level curve (at D or F respectively). This bimodal distribution of age v. height has been observed for dated Holocene shoreline sequences in Japan (Shimazaki & Nakata 1980) and Papua New Guinea (Ota et al. 1993), as well as for the Gulf of Corinth. To interpret the relic shoreline data (elevation v. age) for the Gulf of Corinth, I have plotted a set of h(t) curves for different values of r
Fig. 4. The age and elevation above sea level of raised relic shoreline samples from Aigialeia (filled symbols) and the western Perachora Peninsula (open circles), listed in Table 1. Changing the linear uplift rate r of the footwalt produces a family of curves representing the expected present elevation of relic shorelines as a function of their age. These curves are labelled with the corresponding values of r in m ka -1 . The curve (c represents the corrected local sea level through time and is thus equivalent to setting r = 0.
45
(the model tectonic uplift rate) taken to be time invariable. Figure 4 shows these curves, calculated from equations (2) and (3) and labelled with the corresponding values of r, ranging from 1.0 to 3 . 0 m k a -1. The curve 2.0 is thus equivalent to H K A in Fig. 3. These curves all have bimodal ages for any positive elevation. By plotting the present elevation and age of any relic footwall shoreline for the Gulf of Corinth in Fig. 4 one can estimate its average rate of tectonic uplift from the corresponding background curve of h(t). To do so, however, one must first apply an appropriate calibration to the radiocarbon ages of marine samples from the Gulf of Corinth.
The reservoir effect in the Gulf of Corinth Carbon samples derived from modern ocean surface water have lower 14C/12C ratios than atmospheric CO2, because of vertical mixing with 'older' (14C-depleted) deep water. This 'reservoir effect' increases the radiocarbon age of modern ocean surface water to about 400 years, on average. However, because the magnitude of this effect varies with location, a local correction factor AR is often added to the average (400 year) reservoir effect. Because the Gulf of Corinth is a restricted marine basin, this correction is important here. The Gulf is about 850 m deep, but its connection to the Ionian Sea through the silled Rion Strait is only about 2 km wide and 65 m deep. Nielsen (1912) noted that temperature and salinity below 100m in the Gulf of Corinth were both lower than in the Ionian Sea, and concluded that the deep Gulf water is formed locally by convection during winter. The relatively high and nearly uniform concentration of dissolved oxygen at all depths in the Gulf provides further evidence for efficient vertical mixing (Poulos et al. 1996). Heezen et al. (1966) measured radiocarbon ages in seawater samples taken from the Gulf of Corinth in July 1956. They found the 14C/12C ratio in the surface water to be 4.5% less than in North Atlantic surface water; the comparable figure for water sampled at 800m depth in the Gulf was 4.1%. The equivalent values of AR would be 380 and 350 years, respectively. That is, the waters of the Gulf of Corinth, both deep and shallow, appeared to be nearly twice as 'old' as normal ocean surface water. To Heezen et al. the nearly identical results from both depths suggested 'vigorous and thorough vertical mixing' of the Gulf water. They also noted that radioactively 'dead' carbon
46
S. SOTER
in the surrounding limestone mountains is continuously washed into the Gulf and that the limited exchange of waters through the Rion Straits 'may help to preserve this great apparent age by isolating the Gulf of Corinth waters from the Ionian Sea'. Indeed, it appears likely that carbon dissolved in river and ground water discharging into the Gulf of Corinth is strongly depleted in lac. This is suggested by the anomalous radiocarbon ages we obtained for organic carbon samples from bore holes drilled on the Helike Delta (Maniatis et al. 1996). Samples of wood and marine sediments yielded radiocarbon ages in reasonable agreement with an age-depth profile based on luminescence and archaeological dating of ceramic fragments from the bore holes. However, most of the samples of freshwater organic sediment gave radiocarbon ages 3000-5000 years older than the ceramic age-depth profile (Soter & Katsonopoulou 1998). These anomalously old carbon dates are probably due to the 'hard water effect', involving the dissolution of carbonates by ground water carrying CO2 from the atmosphere. Isotopic reequilibration introduces 'dead' carbon from the carbonates into the CO2. The ground water then feeds lakes where aquatic organisms take up the 'old' CO2 and deposit it as 14C-depleted organic sediments. The hard water effect can increase the apparent radiocarbon age of freshwater organic sediments by as much as the half-life of 14C, nearly 6000 years (MacDonald et al. 1991). Organisms that obtain their CO2 directly from the atmosphere remain unaffected. The case for a hard water effect in the Helike Delta is supported by the presence of carbonate concretion nodules in many of our bore holes. The reservoir effect for the Gulf of Corinth has undoubtedly varied during Holocene time, as a result of changes in climate and sea level. However, the value AR = 380 years measured for the surface water in 1956 (before the global contamination by bomb carbon) is the best available. I will adopt it in preference to the value AR = - 8 0 years estimated for other parts of the Mediterranean (Stiros et al. 1992) and previously applied to the Gulf of Corinth (Pirazzoli et al. 1994; Stewart 1996). Consequently, the calibrated ages of radiocarbon-dated relic shorelines adopted here will be significantly younger than those found previously.
Tectonic uplift bordering the Gulf of Corinth Table 1 lists the dated Holocene samples from relic shorelines on the Gulf of Corinth. The
samples consist of faunal species that lived just below sea level or, in the case of the boring mollusc Lithophaga lithophaga, that produced erosional notches with a distinct upper limit at sea level (Laborel & Laborel-Deguen 1994). One group of samples was collected on the Perachora Peninsula at the eastern end of the Gulf by Vita-Finzi (1993) and Pirazzoli et al. (1994). The other samples are all from Aigialeia, on the southwestern shore, collected at Aigeira, Paralia Platanou, Trapeza and Diakofto (Fig. 1). S. Stiros kindly provided samples of the vermetid gastropod Dendropoma petraeum collected in 1996 from an uplifted shore below the old Trapeza railroad station. The other samples from Aigialeia were collected by Mouyaris et al. (1992), Papageorgiou et al. (1993), VitaFinzi (1993) and Stewart & Vita-Finzi (1996). One sample from Aigeira was dated by the uranium-series method (Vita-Finzi 1993). All other samples in Table 1 were dated by radiocarbon and are calibrated here using the marine calibration curves of Stuiver & Braziunas (1993) with a reservoir correction factor AR = 380 years. Because of unknown variations in the reservoir effect during Holocene time, the actual uncertainties of the calibrated ages must be larger than the errors listed in Table 2. The relic shoreline elevations are plotted against calibrated age in Fig. 4. The ageelevation data for the shorelines of the Perachora Peninsula (open circles) would fall among the curves representing constant tectonic uplift rates of r = 1.35 + 0.25 m ka -1, and their distribution suggests a fairly steady rate of uplift over 7000 years. The data points for the shorelines of Aigialeia (filled symbols) would fall among the curves with rates of r = 2.4 4- 0.8 m km -1 ka -1. The average Holocene uplift rates for both Aigialeia and Perachora are significantly larger than those found by Stewart & Vita-Finzi (1996). The Holocene uplift rates for Aigialeia are also significantly larger than the average Quaternary uplift rates derived from the raised terraces near Corinth (c.l.4mka-1). Noting that footwall topography increases westward along the southern margin of the Gulf of Corinth, Dart et al. (1994) had suggested that 'rates of uplift may increase toward Aigion'. In fact, this appears to be the case. The present extension rate of the Gulf of Corinth also increases to the west (Clarke et al. 1997), which is consistent with the uplift rates and the topography. The distribution of the data for Aigialeia suggests that the rate of uplift there has not been steady but rather has been affected by major discontinuities. Even so, the smooth curves labelled with constant r values are still useful in
U P L I F T A N D S U B S I D E N C E OF T H E H E L I K E D E L T A
47
Table 1. Holocene shoreline data h (m) Aigeira
Fauna
14C age (a BP)
Cal. age (a BP)
@(m)
Ref.
1.0 6.0 6.0 6.5 7.5
D CC L L L
1420 + 60 2965 + 50 U-series 4880 4- 270 8040 4- 85
600 + 50 2245 4- 75 6400 4- 200 4665 4- 335 8055 4- 85
- 1.2 -7.2 -11.5 -8.9 -19.6
2 3 4 2
P. Platanou
2.3 3.7 4.0 6.2 6.5
L L L L L
2785 4- 50 2420 4- 130 8730 4- 340 3285 4- 65 8050 4- 60
2020 4- 70 1590 4- 150 8875 4- 475 2695 4- 45 8065 4- 65
-3.4 -4.7 -20.4 -7.5 -18.6
2 2 4 2 2
Trapeza
3.5
D
6920 4- 50
7030 4- 60
-11.2
5
Diakopto
1.5 3.5
L L
1210 + 100 2190 4- 60
435 4- 85 1340 4- 50
-1.6 -4.3
2 2
Perachora
0.8 1.4 1.7 1.7 1.7 2.2 3.0 3.1
V C N N M L L L
1865 4- 55 1990• 100 6890 4- 90 7100 4- 1300 7200 4- 350 4120 4- 60 4705 4- 50 58204-60
1010 4- 60 11604- 110 6985 4- 115 7095 + 1325 7230 4- 290 3640 4- 70 4450 4- 60 5810+80
- 1.4 -2.1 -9.0 -9.6 - 10.2 -4.0 -5.2 -7.1
6 6 3 3 3 6 6 6
1
h(m) is sample elevation in metres above sea level. Dated shoreline fauna: C, Chthamalus; CC, Cladocora caespitosa; D, Dendropoma petraeum; L, Lithophaga lithophaga; M, Mytilus galloprovincialis; N, Notirus irus; V, Vermetus triqueter. All radiocarbon calibrations are from the marine series of Stuiver & Braziunas (1993) with a reservoir correction factor AR = 380 years. Cb (In), footwall baseline elevation (in metres) corresponding to the calibrated age of each relic shoreline. References: 1, Papageorgiou et al. (1993); 2, Stewart & Vita-Finzi (1996); 3, Vita-Finzi (1993); 4, Mouyaris et al. (1992); 5, this paper; 6, Pirazzoli et al. (1994). For the third Aigeira sample, reference 3 gives an incorrect elevation (C. Vita-Finzi, pers. comm., 1996). The Trapeza sample was collected in 1996 by S. Stiros and dated by AMS at Woods Hole Oceanographic Institution (OS-10187). providing the integrated long-term average rate of tectonic uplift for any raised relic shoreline feature since the time of its formation. To investigate the discontinuities in the tectonic uplift in Aigialeia, we refer again to Fig. 3. In reality, the curve Cb representing footwall elevation is not linear in time, as expressed by equation (2), but has a complex form, with earthquake-related discontinuities and periods of relatively steady aseismic motion. The difference between the smoothly varying sea-level curve ~c and the discontinuous baseline curve ~b will therefore p r o d u c e a discontinuous curve for the present elevation h(t) of relic shorelines as a function o f their age. Conversely, for each relic shoreline plotted in Fig. 4 the c o r r e s p o n d i n g absolute elevation of the footwall baseline w h e n the shoreline was formed is given by Cb = r -- h
(4)
The quantity Cb for each relic shoreline is listed in Table 1 and plotted, for the Aigialeia group, by open symbols in Fig. 5. (The upper part of Fig. 5 repeats the raised shoreline data for
Aigialeia from Fig. 4.) F o r each relic shoreline, the depth below sea level of the footwall baseline (open symbols) at the time of origin equals the present elevation of the feature (solid symbols) above sea level. Figure 5 suggests that the footwall baseline (and hence the entire footwall block) ascended steadily between 7 and 2.2 ka BP. A least-squares linear fit to that segment of the data has a slope of a b o u t 0 . 9 m k a -1. There are p r o n o u n c e d discontinuities at both ends of this segment. Around 2.1kaBp the footwall shoreline ascended rapidly by a b o u t 2 m. This was probably due to the great e a r t h q u a k e of 373 Bc. The discrepancy between the dates is less than 200 years, which is small, considering the uncertainties in the r a d i o c a r b o n ages due to u n k n o w n variations in the ancient reservoir effect for the G u l f of Corinth. Stewart (1996) previously noted this discontinuity in the uplift data for Aigialeia and suggested that it was due to the Helike earthq u a k e of 373Bc, but the present analysis makes the case m u c h stronger. F o r r a d i o c a r b o n
48
S. S O T E R
0
o
_o
~.
~
.~
.~
"~
~
~
~
~
_
_
~i
~
~
o
~
k;
~
i:~
~ o~.~ -~.~
.~
~
~.~ ~
.o.~ .c ~" II o
~=
~,= ~
.~o_.
~o
~<1~
~o~ o
o
~
~
0 e~
-H
-H -H
-H
-H
-H
-H
~l
O0
~
~
~--~
~
44
-H +1
-H
-H
+1
~
O~
re',,
O0
~
~
o6 I
q ~q
oo
oo
o
-H
+t
-H
~.~
'~"
o~
t---
~
--
c~
~-
I
I
I
.o
UPLIFT AND SUBSIDENCE OF THE HELIKE DELTA 373/ BC ] O E ' ' .... ' I I .... I I I ....I [ '....' " .... I
5
....
.... I'"~
#
9
i&
v eO .m
-5
> -10 Ill
-15 /r
-20
_
9
/t~ ,,,I
-25
l0
....
9
P. P l a t a n o u [
: T]tPTpto I
. . . .
8
I ....
I,,i,Ii,,,lll,,I
7
6
5
.... I ....
4
3
2
I,,,
1
0
49
However, this interpretation is sensitive to possible errors in sample dating, in the estimate of sample position with respect to sea level at the time of origin, and in the model of sea-level change. Some of the relic shoreline samples probably experienced an abrupt coseismic uplift or subsidence soon after formation, which selectively preserved them from erosion at the sea-level zone (Laborel & Laborel-Deguen 1994). Such events need not be large, and they are not included in the present model. Finally, the gaps in the distribution of sample dates (possibly reflecting the seismic selection effect just noted) may conceal higher-frequency structure in the uplift trajectory of Fig. 5. Bearing all these caveats in mind, the next section considers the trajectory of the delta block, which depends on the footwall trajectory.
Age (ka BP) Fig. 5. Trajectory of the footwall baseline ~b in Aigialeia (open symbols), obtained by subtracting the present elevation of relic shoreline features (solid symbols) from the corrected sea-level curve ~c. There is a discontinuous uplift at about the time of the Helike earthquake of 373 Be, and another larger pulse of uplift between 7000 and 8000 years ago. Line segments are least-squares fits. calibrations, Stewart assumed a reservoir effect with AR = - 8 0 years, which if applied to the present model would push the apparent age of the discontinuity back to about 2.8 ka aP. The second discontinuity, also noticed by Stewart, is an uplift of about 6-7 m at some time between 7 and 8 ka BP. If gradual, such a tectonic surge implies uplift rates exceeding 6 m per 1000 years. Almost certainly, however, it included one or more major coseismic uplift events. The footwall baseline 9000 years ago was about 20 m below the present absolute sea level. Any relic shoreline features much older than that would still be submerged (Fig. 4). The identification and recovery of any submerged relic shoreline samples might allow us to extend the footwall uplift trajectory back into the preHolocene era. The long-term pattern in Aigialeia appears to be one of gradual footwall uplift from 9 to 8kaaP, followed by a major surge of uplift sometime between 8 and 7kaBP, followed by almost 5000 years of more gradual uplift (mostly aseismic?) until the earthquake of 373 BC, when a footwall uplift of about 2 m occurred. Subsequently, the average uplift rate has been about 2 m k a -1, and the data allow for (but do not establish) a footwall uplift of about 1 m during the earthquake of AD 1861.
Uplift and subsidence of the Helike Delta During an earthquake on a normal fault, uplift of the footwall block is generally accompanied by subsidence of the hanging-wall block. In the case of the Helike Fault, this is represented schematically in Fig. 6 where AB represents the fault, CDE is the surface of the delta just before an earthquake, and FHI shows its position immediately afterwards. During the earthquake
A
footwall~
~~
......
~ ",,
Fig. 6. Schematic model of a coseismic uplift of the footwall block and concomitant subsidence of the hanging-wall delta block. The lines CDE and FHI represent, respectively, the delta surface before and after the earthquake. The earthquake abruptly submerges part of the subaerial delta (GH). The relative block displacement along the fault plane is JCF. Its vertical component is the sum A~b + A(d of the footwall and hanging-wall contributions.
50
S. SOTER
the entire hanging-wall block slips downward and seaward, causing an abrupt marine transgression that drowns the old delta surface from G to H. Simultaneously, the contact of the upper delta surface with the fault plane slips from C to F, and the footwall block moves upward parallel to the fault plane from C to J. Thus the observed coseismic slip JF is the sum of the footwall and hanging-wall displacements. Its vertical component A(s = A~d + z2k~b, where A~a and ACb a r e the contributions of the delta subsidence and the basement uplift, respectively. Analysis of the raised relic shorelines in Aigialeia suggests that the Helike Delta experienced about 2m of earthquake-related uplift during the earthquake of 373 BC. In earthquakes on normal faults, subsidence of the hanging-wall block is usually much larger than uplift of the footwall block (see Stewart 1996). This suggests that the Helike Delta suffered an earthquakerelated subsidence of at least 3 m. A subsidence of such magnitude also appears necessary if one is to accept the description of the Helike disaster given by Pausanias. It is also suggested by the following analysis of sediment core samples from the Helike Delta. The stratigraphy of the Helike Delta, as observed in bore-hole cores, shows evidence of several marine-terrestrial transitions. Such transitional strata are analogous to the relic shorelines on the exposed rocky surface of the footwall to the east of the delta. If these sedimentary transitional strata can be securely dated, then it becomes possible, in principle, to reconstruct the history of uplift and subsidence for the hanging-wall delta block. It requires knowledge of the tectonic trajectory of the footwall block, but such information is available from the raised relic shoreline features on the exposed fault surface. The long-term tectonic uplift of the northern Peloponnesos evidently contributes to the vertical movement of both the Helike Delta and its footwall block. In addition, earthquakes on the Helike Fault should be followed by gradual aseismic uplift of a broad area bordering both sides of the fault, as a result of local isostatic readjustment (King et al. 1988). Post-seismic isostatic uplift thus augments the tectonic uplift of the delta. The long-term uplift from both sources can be visualized in Fig. 6 by displacing the footwall and hanging-wall profiles (AB and FHI) upward together. If sea level rises more slowly than this post-seismic uplift, then the shoreline gradually recedes from G towards H and part of the drowned delta re-emerges. Eventually, there is another earthquake on the Helike Fault, and the entire sequence is repeated.
The Helike Delta and its footwall block may also occasionally experience coseismic uplift as a result of an earthquake on a normal fault north of the Helike Fault. Such a shift in the locus of faulting could transform a footwall into a hanging-wall block, replacing subsidence with uplift (Vita-Finzi 1993; Gawthorpe et al. 1994). Thus, some of the basement uplift inferred from Fig. 4 may be associated with earthquakes not involving the Helike Fault. Sediment samples from bore-hole cores can provide evidence to help reconstruct the palaeoseismology of the Helike Delta. Since 1991, as part of the search for the site of ancient Helike, Soter & Katsonopoulou (1998) drilled 60 bore holes in the Helike Delta. The first five (B1-B5 in Fig. 1) were drilled to an average depth of 44m. Bore hole B2, at 16m surface elevation in the upper part of the delta, contained only terrestrial sediments, whereas B3, drilled near sea level, showed only marine deposits. Bore holes B1, B4 and B5, however, were all drilled in the mid-plain at elevations of 8-9 m and all contained terrestrial, lagoonal and shallow marine deposits, in a manner suggesting that the Holocene shoreline has oscillated across this part of the delta. Maniatis et al. (1996) dated carbon samples from the Helike cores, but many of them consisted of organic sediments deposited in fresh water. As discussed above, freshwater organic sediments on the Helike Delta appear to be highly vulnerable to the hard water effect, so all the carbon dates from such samples are excluded from the present analysis. However, the mid-plain cores yielded two samples of wood from freshwater deposits, one sample of seaweed, and five samples of marine organic silt. Wood is free of the hard water effect (because trees obtain their carbon directly from the atmosphere), and the marine samples can be calibrated with the appropriate reservoir correction for the Gulf of Corinth. Table 2 lists the resulting ages of these eight samples, together with the microfaunal and other indicators of the environments in which they were deposited. Figure 7 plots the calibrated ages of these samples from the Helike bore holes against the absolute elevation h where they were found. This figure also includes the local corrected sea-level curve (~c from Fig. 2) and the approximate trajectory of the footwall baseline (~b from Fig. 5). The footwall discontinuity in Fig. 5 at 2.1kaap is here assumed to be abrupt and ascribed to the earthquake of 373Bc. Noteworthy features are the gap in the footwall trajectory between 7 and 8 ka BP, and the absence of footwall data before 9 ka BP (because of the present submergence of older relic shorelines).
UPLIFT AND SUBSIDENCE OF THE HELIKE DELTA
Fig. 7. Radiocarbon ages and depths with respect to present sea level (horizontal line) for the Helike Delta bore-hole core samples described in Table 2. The stratigraphy suggests that the samples were deposited near sea level. Their present position in the plot, near the Holocene sea-level trajectory (dashed line), suggests that the uplift of the footwall block (dotted line) has counteracted most of the relative subsidence of the hanging-wall delta block. Figure 7 shows that each of the dated core samples from the delta was found in its bore hole at an elevation not far from that corresponding to absolute sea level at the time it was deposited. Assuming that the shallow marine samples were deposited below but within a few metres of sea level and the terrestrial samples were deposited at or above sea level, it follows that the strata containing them have not moved very far in net elevation, despite the large uplift of the underlying footwall block during Holocene time. For example, the wood sample 1B, found in silt containing freshwater ostracods, was presumably deposited above sea level 9800 years ago, hence at an absolute elevation higher than - 2 4 m . The average long-term rate of footwall elevation for the last 9000 years probably exceeded 2 m per 1000 years (Fig. 4). Thus, in the absence of hanging-wall subsidence, sample 1B would have been raised to a present elevation of at least - 4 m. However, it was found at about - 2 4 m as if it had not moved at all. Coseismic subsidence of the delta has apparently cancelled out most of the long-term uplift of the footwall block. It is worth examining this conclusion in more detail for one group of samples. At 8.2 m below sea level in bore hole B4, we found a stratum of fine sand and gravel containing abundant plant fragments, roots and wood, indicating terrestrial
51
deposition. Immediately below this was a stratum of fine sand and silt containing seaweed and abundant brackish to marine foraminiera, indicating deposition in a shallow marine environment. Samples from greater depths in this core showed evidence mainly of marine and brackish environments. The discontinuity at 8.2m thus appears to mark a transition from marine to terrestrial conditions, and we can estimate its age. The wood (sample 4A) in the upper unit was dated at 7.37 ka BP and the marine silt and seaweed (4B and C), sampled at 8.8m below sea level, were dated at 6.34 and 6.28 ka BP, respectively, applying the reservoir correction discussed above for the Gulf of Corinth. The wood appears to be over a thousand years older than the marine horizon upon which it lies. Perhaps it came from a tree that was excavated and deposited by a flood, or perhaps its 14C age is erroneous for unknown reasons. The ages of the two samples from the underlying marine stratum, however, are less likely to be in error, because they are in excellent agreement with each other even though they represent different materials. For the present, I will assume an average age of 6.3kaBP for the marine-terrestrial transition found at about 8.2 m below present sea level in B4. Figure 8 repeats part of the corrected sea-level curve and footwall trajectory from Fig. 7. If one assumes that samples 4B and 4C were deposited near sea level 6300 years ago, one can plot their origin on the sea-level curve at point E in Fig. 8. After its deposition, this reference stratum gradually ascended along the trajectory EF, because of the gradual uplift of the footwall, represented by the parallel track AB. I have neglected possible compaction of the deltaic sediments, which would cause the slope for EF to be somewhat shallower than that of AB. When the reference stratum reached point F the earthquake of 373BC struck. At that time, according to the relic shoreline data, the footwall experienced a coseismic uplift (A~b) of about 2 m from B to C after which it resumed relatively aseismic uplift from C to D. What happened to the Helike Delta during the coseismic uplift of the basement footwall? The reference stratum (for samples 4B and 4C) is at present at 8.8 m below sea level (16.8 m below ground), which corresponds to point I. According to Schmidt (1875), the earthquake of 1861 caused the entire Helike Delta to subside by about 2m. Hence immediately before that earthquake, the reference stratum would have been at point H. And before that, it would have gradually ascended along the trajectory G H because of the (presumably) aseismic uplift of
52
S. SOTER
Fig. 8. Model for the trajectory of the Helike Delta hanging-wall block. The dated samples 4B and 4C from the same bore-hole stratum were evidently deposited near sea level about 6300 years ago (at point E). Gradual uplift (mostly aseismic) of this stratum occurred along trajectory EF (parallel to AB of the footwall). Then the earthquake of 373 Bc abruptly uplifted the footwall from B to C causing subsidence of the hanging-wall delta from F to G. The delta block was then carried from G to H by (aseismic?) uplift of the footwall (represented by CD), until the earthquake of 1861, which caused its subsidence from H to I. The trajectory of the top surface of the delta overlying the dated stratum is represented by EJKLM, which assumes a constant rate of sediment deposition. The subsidence JK of the surface equals that of the dated stratum FG.
the basement, represented by the parallel track CD. Again I ignore compaction. That leaves the vertical interval FG, which according to this model must represent the coseismic subsidence (A~o) of the Helike Delta in the earthquake of 373 BC. It appears here to be about 9 m. This seems excessive, and suggests that samples 4B and 4C were actually deposited a few metres below rather than exactly at sea level. The model also allows one to reconstruct the changing surface elevation of the delta. The present surface at bore hole B4 is 8 m above sea level, represented by point M. Before the 2 m subsidence reported for the earthquake of 1861, the surface would be at L. I assume for the sake of illustration that deposition of the 16.8m of sediment above the reference stratum proceeded at a linear rate. The surface of the delta would then follow the trajectory EJKL, punctuated by the coseismic subsidence JK (equal to FG, the subsidence of the deltaic reference stratum).
In reality, there is no reason to assume a constant rate of sediment accumulation for the last 6000 years. The rate probably depends on the elevation of the deposition surface with respect to sea level. In that case, the deposition rate might increase sharply after the earthquake, owing to the subsidence of the delta. After the earthquake the curve representing the delta surface would at first rise more rapidly than shown by K L in Fig. 8 and later more slowly. This would shift the interval JK downward, perhaps sufficiently that point K dips below the sea-level curve, representing submergence of the delta for a few hundred years. Any marine or brackish deposits from that event would later have been uplifted to be just above the present sea level. We found possible evidence for such a transient marine transgression in bore hole B18, drilled at 7.5m elevation. It contained a sand deposit with brackish ostracods at 1.5 m above the present sea level. This sediment unit remains undated, but it lies directly above an ancient occupation horizon containing ceramic fragments that yielded optical luminescence dates of 3.2 and 3.9 ka BP (Soter & Katsonopoulou 1998). The previously mentioned age discrepancy between the wood sample (4A) and the underlying silt and seaweed samples (4B and 4C) makes it advisable to repeat the above trajectory analysis for the wood, assuming that it was also deposited near sea level and that the linear footwall trajectory AB can be extrapolated back to its age. In that case, the coseismic subsidence of the delta in 373 ac turns out to be about 6 m instead of 9 m. Repeating the analysis for sample 1A, again with the same two assumptions, A(b would be about 4m. However, one cannot assume that different parts of the delta block experienced the same amounts of coseismic subsidence, because the unconsolidated delta sediments are probably disturbed by secondary rotational faults (see Papatheodorou & Ferentinos 1996). All such numerical estimates for the coseismic delta subsidence of 373 Bc must be regarded as illustrative examples only, because the analysis neglects two factors that are difficult to quantify but probably significant. First, there may well have been additional coseismic delta subsidence during mostly unrecorded earthquakes between 373 BC and AD 1861. For example, in the excavation of a Roman building in the Helike Delta, Katsonopoulou (1998) found a destruction layer due to an earthquake in the fourth or fifth century AD. Second, I have neglected sediment compaction and failure. Both factors would reduce the estimates given for coseismic subsidence during the earthquake of 373 Be.
UPLIFT AND SUBSIDENCE OF THE HELIKE DELTA Thus, although the bore-hole results suggest that the delta block subsided by several metres during the earthquake of 373Bc, the exact amount remains very uncertain. However, the magnitude of the apparent coseismic uplift in Aigialeia and the account of the Helike disaster given by Pausanias both suggest that the delta subsided by at least 3m. Given that the less powerful earthquake of 1861 caused 2 m of subsidence, this appears reasonable.
Discussion and conclusions The data for raised relic shorelines in Aigialeia suggest that the footwall block of the Helike Fault has experienced a cumulative uplift of about 20 m in the last 9000 years (Fig. 5). This displacement was apparently due both to coseismic uplift events and to periods of relatively steady aseismic uplift. Two major pulses of uplift occurred during Holocene time. One, probably connected with the earthquake of 373 Bc, showed an uplift of about 2 m. The other occurred some time between 7000 and 8000 years ago, and showed about 6m of uplift, probably involving one or more earthquakes. Between these two major pulses, the footwall may have experienced slow uplift, perhaps mostly aseismic, at a relatively steady rate of about 0.9 m ka -1 . However, it is not certain that steady aseismic uplift occurred during the periods before and after major uplift surges. In particular, such an interpretation for the period from 7 to 2.3 ka BP in Fig. 5 depends on an isolated data point: for Aigeira at 4.7 ka BP. The gaps in the data allow for the possibility that the actual footwall trajectory included many smaller coseismic uplift events separating short periods of aseismic subsidence. Furthermore, there is no direct evidence that the trajectory of footwall uplift measured between Aigeira and Diakopto also applies to the area of the Helike Delta. Bore hole B4 is about 7 k m west of Diakopto and separated from it by a seismic relay zone marked by the Kerynites River (Dart et al. 1994). However, it appears plausible that the two areas share a common footwall trajectory. Stewart (1996) noted the apparent absence of any east-west trend in the average rate of footwall uplift between Aigeira and Diakopto (confirmed by Fig. 4), and suggested that 'the coastal sites can be considered as uplifting as a broadly coherent unit'. This unit may well include the area of the Helike Delta. If the delta was uplifted with the footwall by several metres since the sub-
53
sidence of 373 Bc, as suggested by the data, this might account for the absence of ceramic horizons in the bore holes below present sea level. The estimate of coseismic subsidence for the Helike Delta during the earthquake of 373 Bc neglects the possible effects of subsidence during subsequent earthquakes and of sediment compaction and failure. It also depends on the identification of ancient sea levels in the borehole cores, which is very uncertain. In view of all the uncertainties, the analysis in the previous section must be regarded as merely an example of a method to reconstruct the Holocene tectonic history of the delta block. However, the method offers the potential to produce more secure results when better bore-hole data for the ages of buried shoreline deposits become available. Figure 9 schematically illustrates the proposed long-term motion of the Helike Delta and its footwall block. The footwall trajectory consists of a relatively steady component of aseismic uplift, punctuated by abrupt coseismic pulses of uplift. Most of these pulses are due to earthquakes on the Helike Fault, and are accompanied by an absolute downdrop of the hanging-wall delta block. The figure represents each downward slip of the delta as larger in magnitude than the concomitant uplift of the footwall, because this is generally the case in normal faulting. However, some earthquakes affecting the Helike Delta may occur on the
Fig. 9. Schematic model of long-term tectonic trajectories for the footwall and hanging-wall blocks of the Helike Fault. The footwall motion (dashed line) consists of a steady component of uplift augmented by coseismic pulses. The motion of the hanging-wall delta block (continuous line) generally tracks the uplift due to the aseismic component of footwall motion, but large earthquakes on the Helike Fault cause abrupt subsidence of the delta. When an earthquake occurs on a fault located north of the Helike Fault, however, the Helike Delta is then on the footwall side of an active fault and both the delta (A) and its basement block (B) are uplifted by the same amount.
54
S. SOTER
adjacent Aigion Fault or another fault to the north, and in such cases both the Helike Delta and its footwall will be uplifted together. This is indicated in the figure by the hangingwall uplift at point A which accompanies an equal uplift of the footwall (point By. The long-term uplift of the footwall block is the sum of a relatively steady aseismic component and abrupt coseismic pulses. For the hangingwall delta, however, the coseismic motions are generally downward, and their cumulative effect may cancel out much of the long-term uplift. An occasional coseismic uplift of the delta may also occur, as represented at point A in Fig. 8. The cumulative effect of the various contributions may allow the delta block to maintain its baseline at nearly the same absolute elevation during much of Holocene time. However, on time scales greater than 10000 years, the hanging-wall block of the modern delta must have experienced a large net tectonic subsidence, to accommodate some 500 m of sediment deposition during Quaternary time (Dart et al. 1994). How might the large Quaternary subsidence of the delta be reconciled with the evidence of relatively little net subsidence during Holocene time? The long-term subsidence of the delta may be irregular, containiag shorter periods of reduced net subsidence, of which the Holocene is one. Or perhaps the Holocene marks the final transitional stage between longterm subsidence and long-term uplift, with seismic activity beginning to alternate between neighbouring faults and shifting the status of the Helike Delta block from hanging wall to footwall. Dart et al. (1994) noted the progressive northward evolution of activity on the extensional faults at the southern margin of the Gulf of Corinth, and proposed that successive basin fill deposits, including fan deltas, have been 'transferred from hangingwall to footwall positions, and progressively uplifted and incised'. The Helike Delta block may eventually sustain a long-term net uplift, perhaps carrying it to the elevation of the ancient exhumed delta blocks that form the mountains south of the Helike Fault. Hardy et al. (1994) developed an elegant computer model that illustrates how the sequence stratigraphy of a Gilbert-type fan delta depends on variable rates of extensional subsidence, seaor lake-level change, and sediment supply. They applied it to analyse the sequence stratigraphy of the uplifted Kerynites fan delta, south of the Helike Fault. Their model, with the addition of intervals of hanging-wall uplift, might be applied to describe the stratigraphy of the present Helike Delta (see Collier & Gawthorpe 1995).
There are other places where abrupt coseismic events appear to neutralize or dominate the longterm trend of aseismic coastal tectonic motion. Shimazaki & Nakata (1980) reported evidence for coseismic uplift pulses separated by long periods of stasis or possible subsidence for three sites in Japan. PlaNer et al. (1992) examined the Copper River delta in Alaska, where they found that long-term aseismic subsidence appears to dominate large magnitude coseismic uplift events. Finally, Atwater et al. (1992) and VitaFinzi (1996) have shown that large pulses of coseismic subsidence have occurred in coastal Chile despite net Holocene emergence (at least partly coseismic). The result bears some relation to the trajectory of the Helike Delta block. It should be noted, however, that all three Pacific locations are thrust-fault regimes, unlike the extensional regime of the Gulf of Corinth. I wish to thank D. Katsonopoulou and E. Sotiropoulos for their help in organizing the bore-hole drilling in the Helike Delta. I also thank C, Tziavos (National Centre for Marine Research, Athens) and P. Blackwelder, T. Hood and C. Alvarez-Zarikian (University of Miami), whose analysis of the environmental indicators will appear in a forthcoming paper with the author. The radiocarbon dating of samples in Table 2 was carried out by SY. Maniatis and Y. Facorellis (NCSR Demokritos, Athens) and B. Kromer (University of Heidelberg). SI wish to thank. Stiros (University of Patras) for guidance and for kindly providing the faunal sample from the relic shoreline at Trapeza. That sample was dated by AMS at Woods Hole Oceanographic Institution with support from the NSF (OCE-9301015). The Helike Project is conducted under the auspices of the American School of Classical Studies at Athens.
References AMBRASES, N. N. & JACKSON, J. A. 1997. Seismicity and strain in the Gulf of Corinth (Greece) since 1694. Journal of Earthquake Engineering, 1, 433-474. ARMIJO, R., MEYER, B., KING, G. C. P., RIGO, A. PAPANASTASS1OU,D. 1996. Quaternary evolution of the Corinth Rift and its implications for the Late Cenozoic evolution of the Aegean. Geophysical Journal International, 126, 11-53. ATWATER,B. F., NUNEZ, H. J. & VITA-FINZI,C. 1992. Net late Holocene emergence despite earthquakeinduced submergence, south-central Chile. Quaternary' International, 15-16, 77-85. BERNARD, P., BRIOLE, P., MEYER et al. 1998. The Ms=6.2, June 15 1995 Aigion earthquake (Greece): evidence for low angle normal faulting in the Corinth rift. Journal of Seismology, 1, 131 150. BLANCHON,P. & SHAW,J. 1995. Reef drowning during the last deglaciation: evidence for catastrophic
U P L I F T A N D S U B S I D E N C E OF T H E H E L I K E D E L T A sea-level rise and ice-sheet collapse. Geology, 23, 4-8. CLARKE, P. J., DAVIES. R. R., ENGLAND, P. C. et al. 1997. Geodetic estimate of seismic hazard in the Gulf of Corinth. Geophysical Research Letters, 24, 1303-1306. COLLIER, R. E. L. & GAWTHORPE, R. L. 1995. Neotectonics, drainage and sedimentation in central Greece: insights into coastal reservoir geometries in syn-rift sequences. In: LAMBIASE, J. J. (ed.) Hydrocarbon Habitat in Rift Basins. Geological Society, London, Special Publications, 80, 165-181. DART, C. J., COLLIER, R. E. L., GAWTHORPE, R. L., KELLER, J. V. A. & NICHOLS, G. 1994. Sequence stratigraphy of (?)Pliocene Quaternary synrift, Gilbert-type fan deltas, northern Peloponnesos, Greece. Marine and Petroleum Geology, 11, 545-560. DOUTSOS, T. & POULIMENOS, G. 1992. Geometry and kinematics of active faults and their seismotectonic significance in the western Corinth-Patras rift (Greece). Journal of Structural Geology, 14, 689-699. FERENTINOS, G., PAPATHEODOROU, G. & COLLINS, M. B. 1988. Sediment transport processes on an active submarine fault escarpment: Gulf of Corinth, Greece. Marine Geology, 83, 43-61. GAWTHORPE, R. L., FRASER, A. J. & COLLIER, R. E. L. 1994. Sequence stratigraphy in active extensional basins: implications for the interpretation of ancient basin-fills. Marine and Petroleum Geology, 11,642-658. HARDY, S., DART, C. J. & WALTHAM, D. 1994. Computer modeling of the influence of tectonics on sequence architecture of coarse-grained fan deltas. Marine and Petroleum Geology, 11,561-574. HEEZEN, B. C., EWING, M. & JOHNSON, G. L. 1966. The Gulf of Corinth floor. Deep-Sea Research, 13, 381-411. KATSONOPOULOU, D. 1998. The first excavation in the area of ancient Helike (in Greek). In: KATSONOPOULOU, D., SOTER, S. & SCHILARDI, D. (eds) Ancient Helike and Aigialeia: Proceedings of the Second International Conference. Helike Society, Athens, in press. -& SOTER, S. 1997. Ancient Helike in the light of modern discoveries. American Journal of Archaeology, 101, 378 (abstract). KERAUDREN, B., FALGUERES, C., BAHAIN, J., SOREL, D. & YOKOYAMA, Y. 1995. Nouvelles datations radiometriques des terrasses marines de Corinthie (Peloponnese septentrional, Grece). Comptes Rendus de l'Acaddmie des Sciences, 320, 483-489. KING, G. C. P., STEIN, R. S. & RUNDLE, J. B. 1988. The growth of geological structures by repeated earthquakes. Journal of Geophysical Research, 93, 13307-13319. KOUKOUVELAS, I. K. & DOUTSOS, T. T. 1996. Implications of structural segmentation during earthquakes: the 1995 Egion earthquake, Gulf of Corinth, Greece. Journal of Structural Geology, 18, 1381-1388.
55
LABOREL, J. & LABOREL-DEGUEN, F. 1994. Biological indicators of relative sea-level variations and of coseismic displacements in the Mediterranean region. Journal of Coastal Research, 10, 395-415. LAMBECK, K. 1995. Late Pleistocene and Holocene sea-level change in Greece and south-western Turkey: a separation of eustatic, isostatic and tectonic contributions. Geophysical Journal International, 122, 1022-1044. 1996. Sea-level change and shore-line evolution in Aegean Greece since Upper Palaeolithic time. Antiquity, 70, 588-611. LEKKAS, E., LOzlOS, S., SKOURTSOS,E. & KRANIS, H. 1996. Liquefaction, ground fissures and coastline change during the Egio earthquake (15 June 1995: Central-Western Greece). Terra Nova, 8, 648-654. MACDONALD, G. M., BEUKENS, R. P. & K1ESER, W. E. 1991. Radiocarbon dating of limnic sediments: a comparative analysis and discussion. Ecology, 72, 1150-1155. MANIATIS, Y., FACORELLIS, Y., SOTER, S., KATSONOPOULOU, D. & KROMER, B. 1996. Locating horizons with Iac sediment dating: the case of the lost city of Helike. Radiocarbon, 37, 931-942. MARINATOS, S. 1960. Helice: a submerged town of Classical Greece. Archaeology, 13, 186-193. MOUYARIS, N., PAPASTAMATIOU,O. C. & VITA-FINZI, C. 1992. The Helice Fault? Terra Nova, 4, 124-129. NIELSEN, J. N. 1912. Hydrography of the Mediterranean and adjacent waters. Report on the Danish Oceanographic Expeditions 1908-1910 to the Mediterranean, 1, 79-191. ORl, G. G. 1989. Geologic history of the extensional basin of the Gulf of Corinth (?Miocene-Pleistocene), Greece. Geology, 17, 918-92t. OTA, Y., CHAPPELL, J., KELLEY, R., YONEKURA, N., MATSUMOTO, E., NISHIMURA, T. & HEAD, J. 1993. Holocene coral reef terraces and coseismic uplift of Huon Peninsula, Papua New Guinea. Quaternary Research, 40, 177-188. PAPAGEORGIOU, S., ARNOLD, M., LABOREL, J. & STIROS, S. C. 1993. Seismic uplift of the harbour of ancient Aigeira, Central Greece. International Journal of Nautical Archaeology, 22, 275-281. PAPATHEODOROU, G. ~ FERENTINOS, G. 1996. Submarine and coastal sediment failure triggered by the 1995, M~=6.1 R Aegion earthquake, Gulf of Corinth, Greece. Marine Geology, 137, 287-304 PELTIER, W. R. 1994. Ice age paleotopography. Science, 265 195-201. PIRAZZOLI, P. A., STIROS, S. C., ARNOLD, M., LABOREL, J., LABOREL-DERGUEN, F. & PAPA6EORGIOU, S. 1994. Episodic uplift deduced from Holocene shorelines in the Perachora Peninsula, Corinth area, Greece. Tectonophysics, 229, 201-209. PLAFKER, G., LAJOIE, K. R. & RUBIN, M. 1992. Determining recurrence intervals of great subduction
56
S. SOTER
zone earthquakes in southern Alaska by radiocarbon dating. In: TAYLOR, R. E., LONG, A. & KRA, R. S. (eds) Radiocarbon after Four Decades. Springer-Verlag, Berlin, 436 a53. POULIMENOS, G. 1993. Tectonics and sedimentation in the Western Corinth graben, Greece. Neues Jahrbuch ffir Geologic und Paldeontologie, Monatshefte, 10, 607-630. POULOS, S. E., COLLINS, M. B., PATTIARATCHI, C., CRAMP, A., GULL, W., TSIMPLIS,M. & PAPATHEODOROU, G. 1996. Oceanography and sedimentation in the semi-enclosed, deep-water Gulf of Corinth (Greece). Marine Geology, 134, 213-235. RICO, A., LYON-CAEN, H., ARM[JO, R. et al. 1996. A micro-seismic study in the western part of the Gulf of Corinth (Greece): implications for largescale normal faulting mechanisms. Geophysical Journal International, 126, 663-688. ROBERTS, G. P. & KOUKOUVELAS, I. 1996. Structural and seismological segmentation of the Gulf of Corinth fault system: implications for models of fault growth. Annali di Geofisica, 39, 619-646. SCHMIDT, J. 1862. Sur le grand tremblement de terre qui a eu lieu en GrGce le 26 decembre 1861. Comptes Rendus de l'Acaddmie des Sciences, 54, 669-671. - - 1 8 7 5 . Studien fiber Erdbeben. Leipzig. SEGER, M. & ALEXANDER, J. 1993. Distribution of Plio-Pleistocene and Modern coarse-grained deltas south of the Gulf of Corinth, Greece. In: FROSTICK, L. E. & STEEL, R. J. (eds) Tectonic Controls and Signatures in Sedimentary Successions. Special Publication, International Association of Sedimentologists, 20, 37-48. SHIMAZAKI, K. & NAKATA, T. 1980. Time-predictable recurrence model for large earthquakes. Geophysical Research Letters, 7, 279-282.
SOTER, S. 1998. The Aigion earthquake of 1995: macroscopic precursors. In: KATSONOPOULOU, D., SOTER, S. & SCHILARDI, D. (eds) Ancient Helike and Aigialeia." Proceedings of the Second International Conference, Helike Society, Athens, in press. & KATSONOPOULOU, D. 1998. The search for ancient Helike 1988-1995: geological, sonar and bore hole studies. In: KATSONOPOULOU,D., SOTER, S. & SCHILARDI, D. (eds), Ancient Helike and Aigialeia: Proceedings of the Second International Conference, Helike Society, Athens, in press. STEWART, I. 1996. Holocene uplift and palaeoseismicity on the Eliki Fault, Western Gulf of Corinth, Greece. Annali di Geofisica, 39, 575-588. & VITA-FINZI, C. 1996. Coastal uplift on active normal faults: the Eliki Fault, Greece. Geophysical Research Letters, 23, 1853-1856. STIROS, S. C., ARNOLD, M., PIRAZZOLI, P. A., LABOREL, J., LABOREL, F. & PAPAGEORGIOU, S. 1992. Historical coseismic uplift on Euboea Island, Greece. Earth and Planetary Science Letters, 108, 109-117. STUIVER, M. & BRAZIUNAS, T. F. 1993. Modeling atmospheric X4C influences and ~4C ages of marine samples to 10,000BC. Radiocarbon, 35(11), 137--190. VITA-FINZI, C. 1993. Evaluating Late Quaternary uplift in Greece and Cyprus. In: PRICHARD, H. M., ALABASTER, T., HARRIS, N. B. W. & NEARY, C. R. (eds) Magmatic Processes and Plate Tectonics. Geological Society, London, Special Publications, 76, 417-424. - - 1 9 9 6 . Paleoseismology in coastal Chile. Journal of Geophysical Research, 101, 6109-6114. -
-
Oxygen-isotope correlation of marine terraces and uplift of the Mesa hills, Santa Barbara, California, USA M . A. T R E C K E R ,
L. D. G U R R O L A
& E. A. K E L L E R
Department of Geology and Institute for Crustal Studies, University of California, Santa Barbara, Santa Barbara, CA 93106, USA (e-mail:
[email protected]) Abstract: Resolving the chronology of marine terrace sequences is critical for determining uplift rates along tectonically active coastlines. Unfortunately, lack of suitable dating materials often makes this difficult. We present here oxygen isotopic data from 21 shells of Olivellabiplicatafrom four marine terraces in the Santa Barbara and Ventura area located in southern California, USA. Terraces U-series dated at 47+0.5kaBP at Isla Vista and 70 4- 2 ka Be at Santa Barbara City College (SBCC) provide age control for the isotopic data. Shells from the Isla Vista and SBCC terraces yield average values of 1.117%o and 0.627%o, respectively, and shells from a non-U-series dated terrace at Punta Gorda and an undated terrace at Santa Barbara Point yield average values of 1.010%o and 0.751%o, respectively. The data indicate that stable oxygen isotopic signatures preserved in marine terrace molluscs provide a useful tool for correlating undated terraces with those of known age. Furthermore, we are able to correlate samples collected from offset fragments of the Punta Gorda terrace on either side of the Red Mountain fault, demonstrating the utility of this method for correlating terraces across structural features. Using oxygen isotopic data coupled with the U-series dated wave-cut platform at SBCC we calculate a rate of uplift ranging from 0.62 + 0.03 mm/year (where the elevation of the first emergent terrace is 41 m) to 0.54 + 0.05 mm/year (where the elevation of the first emergent terrace is 36 m) for marine terrace flights preserved on the Mesa hills anticline located in the city of Santa Barbara, California. Coastal areas which have experienced uplift during the Late Pleistocene to present are typically characterized by a series of uplifted marine terraces resembling a flight of stairs. Each terrace forms at sea level as a wave-cut platform overlain by a thin veneer of marine sediments which are capped by more extensive terrigenous sediments, and, often, vegetation. The interface between the wave-cut platform and its sea-cliff is called the shoreline angle, and, as it usually lies within one or two vertical metres of the mean sea level (Lajoie et al. 1979), is the only certain reference to palaeo-sea level. To calculate a rate of uplift, the elevation of the terrace relative to present sea level, palaeo-sea level at the time of terrace formation, and the age of the terrace must be established (Bull 1985). In many areas such as southern California, numerical dating of wave-cut platforms has proved to be frustrating because of the scarcity of materials acceptable for dating. Furthermore, even if a particular wave-cut platform is dated in an area, it is difficult to correlate the dated terrace with others in the area, as faults, folds, and erosional features such as gullies or rivers often obscure geomorphological evidence for lateral continuity (Fig. 1). For example, even if terrace surface X1 (in Fig. 1) is dated, the fault, fold, and gullies make it difficult to confidently correlate this surface with other surfaces several kilometres away.
The hypothesis we test here is that stable oxygen isotopic signatures preserved in marine terrace molluscs can be used to correlate terraces of unknown age with those dated by U-series methods, which will allow for much greater resolution of uplift rates in tectonically active areas.
Dating and correlation of marine terraces U-series analysis of solitary corals is one of the few accepted methods for obtaining a numerical date on a wave-cut platform older than about 50 000 years, as Kaufman et al. (1971) have shown that molluscs remain open systems with regard to uranium after death, and therefore may not yield accurate U-series ages. Unfortunately, solitary corals are extremely rare in marine terrace deposits. Accordingly, a number of methods for correlating undated terraces with those dated by U-series methods have been developed. They include amino-acid racemization, 14C dating of molluscs and charcoal, faunal assemblage analysis, and oxygen isotope stratigraphy. Of these, amino-acid racemization is currently the most commonly applied. The technique is based on the occurrence of amino acids only in the L-enantiomeric configuration in living protein. After death these amino acids undergo racemization, a spontaneous and reversible chemical reaction in which the L-enantiomers convert to an equilibrium mixture (usually 1:1)
TRECKER, M. A., GURROLA,L. D. & KELLER, E. A. 1998. Oxygen-isotope correlation of marine terraces and uplift of the Mesa hills, Santa Barbara, California, USA. In: STEWART, I. S. • VITA-FINZ1,C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 57-69.
58
M. A. TRECKER E T A L .
Beach x,,N~.g~l.~.._ -
VE ~xlO
-~~~ne
10
I
Fig. 1. Typical section of active tectonic coastline with marine terraces. Deformation along faults and folds, as well as erosional processes, makes accurate lateral correlation of terraces based on geomorphological evidence difficult to impossible. of D- and L-enantiomers. The extent of racemization (D " L ratio) is a function of reaction time and ambient temperature (Kennedy et al. 1982), as well as genus, and a variety of diagenetic processes (Muhs et al. 1992). Because the D: L ratio is dependent on factors other than time, this method cannot be used as a numerical dating tool, and owing to regional temperature changes correlations may be limited to a specified area. 14C dating of molluscs and charcoal found in terrace deposits, because of the relatively short half-life of 14C, is applicable only to terraces that are less than 50 000 years old (oxygen isotope stage 3a and younger). The stage 3a terrace is rarely present in terrace sequences because it requires a very rapid rate of uplift to be preserved, hence 14C dating is of limited utility. Correlations are also made, or corroborated, on the basis of palaeontological data. Faunal assemblage analysis is based on the presence in a given assemblage of extralimital species (species which are found either north or south of their present habitat) indicating a shift in isotherms (Valentine & Meade 1963). The presence of extralimital species makes it possible to determine whether a deposit represents former coldor warm-water conditions. It is a relative dating tool which cannot serve on its own to correlate terrace fragments unambiguously. Oxygen isotope stratigraphy relies on the assumption that marine terrace molluscs of different ages preserve distinct isotopic signatures which can be used to correlate undated terraces with those of known age. The method
involves collection of fossil molluscs from dated sites to establish characteristic isotopic signatures for molluscs of different ages (e.g. stage 3a, 5a, and 5e). Samples are then collected from undated sites in the same area, and the stable isotopic signatures of these molluscs are compared with those of molluscs from the dated sites. This method, although currently not widely applied, has proved useful for correlating Pleistocene terraces in the southern California area (Muhs et al. 1992; Trecker et al. 1997a, b).
Study area The Santa Barbara fold belt (SBFB) is developed on the coastal piedmont of the Santa Ynez Mountains south of the 'Big Bend' in the San Andreas fault (Fig. 2). Across the region of the 'Big Bend', north-south contraction and regional shortening is observed (Savage 1983; King & Savage 1984). This shortening produces the weststriking folds and reverse faults in the SBFB. The SBFB is characterized by marine platforms and associated terrace deposits preserved on the flanks of active anticlines (Gurrola & Keller 1997). Previous terrace studies in the SBFB have primarily focused on aminostratigraphy of terrace fossils. Studies by Wehmiller et al. (1978) and Lajoie et al. (1982) suggested an age of stage 3 for the Punta Gorda terrace near Ventura, CA, and identified several uplifted palaeo-shorelines associated with Holocene earthquakes. Aminostratigraphic analysis of terrace fossils from the Santa Barbara area provided preliminary estimates of
M A R I N E T E R R A C E S O F M E S A HILLS, C A L I F O R N I A
59
Fig. 2. Tectonic framework of the onshore Santa Barbara fold belt. Fault systems are shown in heavy lines and folds are shown in lighter lines. Boxes show locations of Figs 3 and 5.
Fig. 3. Location of Isla Vista, Santa Barbara City College (SBCC), and Santa Barbara Point marine terraces. MR, Mission Ridge fault system segments 1 and 2; UCSB, University of California, Santa Barbara. Box shows location of Fig. 4, the Mesa hills study site. Geology simplified after Dibblee (1986).
60
M. A. TRECKER E T A L .
uplift rates of c. 0.12 mm/year (Metcalf 1994). U-series analysis of a terrace coral from the Isla Vista terrace in Goleta, CA, has since established that the terrace formed during stage 3a, yielding a vertical rate of uplift of 1.22 4- 0.13 mm/year (Gurrola et al. 1996).
The site for this study is the Mesa hills (see Figs 3 and 4), a complex faulted anticline which forms an asymmetric ridge with a flight of late Pleistocene marine terraces preserved on the south limb. The steep, north-facing forelimb has been extensively eroded by Mission Creek, and
Fig. 4. Detailed geomorphological map of emergent marine terraces and structures on the Mesa hills, Santa Barbara, California. Lines refer to location of terrace flights in Fig. 6. (For location, see Fig. 3.)
MARINE TERRACES OF MESA HILLS, CALIFORNIA north-dipping beds have been identified in a few locations. The gently south-dipping back limb is well preserved and identified by the Middle to Upper Pleistocene Santa Barbara Formation, which dips c. 15-25 ~ to the south. The Santa Barbara Formation forms the wave-cut platform only at the Santa Barbara City College (SBCC) terrace. The Miocene Monterey shale forms most of the wave-cut platforms on the Mesa hills and is well exposed along the sea-cliffs. In the study area the shale is thin- to thick bedded, highly siliceous and strongly indurated. It has been extensively deformed by pre-Pleistocene tectonics.
Research objectives and scope The primary objectives of the study were: (1) to assess the utility of oxygen isotope stratigraphy
61
as a means of correlating deformed marine terraces; (2) to correlate a U-series dated marine terrace at SBCC with an undated site at Santa Barbara Point; (3) to use this information to estimate rates of uplift and ages of other terraces on the south flank of the Mesa hills anticline.
Correlation using oxygen isotopes Studies in southern California indicate that undated marine terraces can be correlated with numerically dated terraces using the stable oxygen isotopic signature preserved in marine terrace molluscs (Muhs et al. 1992; Trecker et al. 1997a, b). Changes in oxygen isotopic values during late Quaternary time are largely dependent on changing polar ice sheet volume, and to a lesser extent on ocean temperatures. Thus,
Fig. 5. Location of the Punta Gorda terrace and Red Mountain fault. (Modified from Lajoie et al. (1979).)
62
M. A. T R E C K E R E T A L .
molluscs from those terraces formed during periods of lower sea level and therefore relatively higher polar ice volume exhibit higher oxygen isotopic values than those from terraces formed during periods of higher sea level and relatively lower polar ice volume. This relationship offers to provide the basis for oxygen isotope stratigraphy as a correlation tool. It is hypothesized that by establishing characteristic isotopic signatures from U-series dated terraces, it is possible to correlate undated terraces with
these calibration points by rigorous comparison of their stable oxygen isotopic signatures. We have tested this hypothesis and begun to systematically develop the requisite technique on molluscs from marine terraces in the Santa Barbara and Ventura areas. Our test sites include two calibration points: the Isla Vista terrace U-series dated at stage 3a (Gurrola et al. 1996) and the SBCC terrace U-series dated at stage 5a (Gurrola et al. 1997) (Fig. 3). On the basis of isotopic signatures, we correlate the
Table 1. All oxygen isotopic values.from all shells Shell
Oxygen value
Shell
Oxygen value
Shell
Oxygen value
Shell
IVTCSI-1 IVTCS 1-2 IVTCS1-3 |VTCS 1-4 IVTCS1-5
1.352 1.146 1.057 0.877 0.632
PGTD-1 PGTD-2 PGTD-3 PGTD-4 PGTD-5
1.318 1.333 0.884 1.594 0.762
SBPI-1 SBP1-2 SBP1-3 SBP1-4 SBP1-6
-0.012 0.466 0.281 0.659 0.930
SBCCI-1 SBCC1-2 SBCC1-3 SBCC1-4 SBCC1-5
0.531 0.533 1.294 0.809 1.079
IVTCS2-1 IVTCS2-2 IVTCS2-3 IVTCS2-4 IVTCS2-5
0.841 1.245 1.247 1.037 1.075
PGTD2-1 PGTD2-2 PGTD2-3 PGTD2-4 PGTD2-5 PGTD2-6
0.541 0.226 0.659 0.994 0.255 0.127
SBP2-1 SBP2-2 SBP2-3 SBP2-4
0.436 0.324 0.517 1.382
SBCC2-1 SBCC2-2 SBCC2-3 SBCC2-4 SBCC2-5
0.731 0.780 0.166 0.774 0.589
IVTCS3-1 IVTCS3-2 IVTCS3-3 IVTCS3-4 IVTCS3-5
1.449 1.029 1.334 I. 105 1.328
PGTD3-1 PGTD3-2 PGTD3-3 PGTD3-4 PGTD3-5
1.624 1.545 1.861 1.211 1.477
SBP3-1 SBP3-2 SBP3-3 SBP3-4 SBP3-5
1.515 2.329 1.148 0.863 1.835
SBCC3-1 SBCC3-2 SBCC3-3 SBCC3-4 SBCC3-5
0.039 0.354 0.868 0.488 0.506
IVTCS4-1 IVTCS4-2 IVTCS4-3
1.959 2.066 1.992
PGTUI-1 PGTU 1-2 PGTU1-3 PGTU1-4 PGTU1-5
1.193 0.963 1.374 1.012 0.964
SBP4-1 SBP4-2 SBP4-3 SBP4-4 SBP4-5
1.526 1.375 1.389 2.007 1.129
SBCC4-1 SBCC4-2 SBCC4-3 SBCC4-4 SBCC4-5
1.276 0.874 1.150 0.948 0.738
PGTU2-1 PGTU2-2 PGTU2-3 PGTU2-4 PGTU2-5
1.501 0.881 1.010 1.043 1.077
SBP5-1 SBP5-2 SBP5-3 SBP5-4 SBP5-5 SBP5-6
0.944 1.469 1.268 0.566 0.995 1.5
PGTU3-1 PGTU3-2 PGTU3-3 PGTU3-4 PGTU3-5
1.013 0.818 1.432 0.832 1.552
PGTU4-1 PGTU4-2 PGTU4-3 PGTU4-4 PGTU4-5
1.632 0.134 1.577 1.225 0.836
SBCC5-1 SBCC5-2 SBCC5-3 SBCC5-4
Oxygen value
0.601 0.636 -0.898 0.590
All oxygen isotopic values obtained from study area shells. Values shown in Table 2 are averages of these values for each shell. The number of samples is a function of shell size. Machine precision is 0.1%o.
MARINE TERRACES OF MESA HILLS, CALIFORNIA faulted Punta Gorda terrace (Fig. 5) which has been dated by less reliable means at stage 3 (Kaufman et al. 1971; Wehmiller et al. 1978; Lajoie et al. 1982) with the Isla Vista terrace, and the previously undated Santa Barbara Point terrace (Fig. 3) with the SBCC terrace. Selection of the undated sites was based on the availability of well-preserved fossil material and proximity to the calibration points. The gastropod Olivella biplicata was chosen as our test species. It was found at every site sampled, and has been documented as occurring in most fossil localities throughout the California Pleistocene deposits (Valentine 1961). Valentine (1961) also noted that Olivella shells are often the best preserved, and probably suffer the least damage as they are used by hermit crabs after death, which helps preserve them, as well as the fact that they are small, compact, thick shelled, and consist wholly of aragonite. Additionally, our data indicate that Olivella grows throughout the year, so that its isotopic signature is an average value for all the seasons. Lastly, as Olivella lives 8-15 years (Morris et al. 1980), the effects of an E1 Nifio year, or other climate phenomena, do not significantly distort the isotopic signature. Well-preserved specimens were selected for analysis. They were scrubbed in deionized water to remove any matrix, rinsed, ultrasonically cleaned, and then allowed to air dry overnight. The outside layer, possibly containing organic matter, was then removed mechanically. To determine the isotopic variation throughout the life of the organism (as well as to sample from different seasons), multiple samples from each shell were drilled along the growth direction of the shell (Table 1). Powdered aragonite samples were
63
loaded into small copper boats and roasted at 200~ for 1 h under vacuum to remove organic contaminants. The samples were then reacted in orthophosphatic acid at 90~ in an on-line automated carbonate CO2 device. The evolved CO2 was analysed using a Finnigan/MAT 251 light stable isotope mass spectrometer. Instrument precision is 0.1%o or better. Isotopic data are expressed in the standard 6 notation in per mil relative to the Pee Dee Belemnite (PDB) carbonate standard. Isotopic values were related to PDB through the NBS-19 carbonate standard, which has a known isotopic value (Coplen et al. 1983). Oxygen isotopic values from each shell from the four terraces are shown in Table 2. Each shell value represents an average of 5-6 powder samples (Table 1). Values in bold (Table 2) are those which are significantly higher than the rest for that terrace. We suggest that they may represent the rapid (decadal to century scale) cooling events which are known to have occurred repeatedly during late Pleistocene time (Dansgaard et al. 1969, 1993; Taylor et al. 1993; Kotilainen & Shackleton 1995; Thunell & Mortyn 1995). However, this hypothesis would not explain the one warm (lower) value from the downthrown block of the Punta Gorda terrace. It is also possible that these anomalous values are a result of diagenesis, redeposition, or variations in water temperature within the same depositional environment. (The shell IVTCS4 from Isla Vista was badly broken, hence we were only able to obtain three powdered samples from three holes in it compared with the usual five or six. It is very possible that this is the reason this shell yielded a value which is very different from those of the rest of the Isla Vista shells).
Table 2. Oxygen isotopic values for study area terraces Terrace
Shell
6TMO
Terrace
Isla Vista Isla Vista Isla Vista Isla Vista
1 2 3 4
1.013 1.089 1.249 2.006
Punta Punta Punta Punta Punta Punta Punta
SBCC SBCC SBCC SBCC SBCC
1 2 3 4 5
0.849 0.608 0.451 0.997 0.232
SB Point SB Point SB Point SB Point SB Point
Gorda Gorda Gorda Gorda Gorda Gorda Gorda
D D D U U U U
Shell
618O
1 2 3 1 2 3 4
1.178 0.467 1.543 1.101 1.103 1.129 1.081
1 2 3 4 5
0.665 0.465 1.124 1.538 1.485
Oxygen isotopic values for study area terraces. Numbers are a result of averaging values from Table 1 for each shell. Analytical error is 1%oor less. D and U represent the downthrown and upthrown side of the Red Mountain fault at the Punta Gorda terrace. Bold type indicates anomalous values.
64
M. A. TRECKER E T AL. Table 3. Statistical Results Terrace pair
H value
c~
Null hypothesis
IV/SBCC IV/S BPT SBCC/SBPT PGTU/PGTD PG/IV PG/SBCC PG/SBPT
l 3.868 3.881 0.14 0.123 0.11 13.006 3.94
< 0.001 0.05 0.8 0.8 0.8 < 0.001 0.05
Reject Reject Accept Accept Accept Reject Reject
Results of the Kruskal-Wallis one-way analysis of variance. IV, Isla Vista; SBCC, Santa Barbara City College; SBPT, Santa Barbara Point; PGTU, Punta Gorda, upthrown block; PGTD, Punta Gorda, downthrown block; PG, Punta Gorda. We are currently performing X-ray diffraction (XRD) analysis of the shells which yielded anomalous values as well as beginning a detailed study of modern O. biplicata shells to clarify these findings. Preliminary results from the XRD analysis indicates that the shells which yielded anomalous values have no more than c. 1-2% calcite in them. At present, a more detailed study is being undertaken, including scanning electron microprobe examination of thin sections of both modern and ancient shells by (double checked by quantitative X R D analysis) to determine where the calcite is located in the shells, and if it can be avoided during sampling. The data show that there is a distinct difference in isotopic values recorded by molluscs at Isla Vista (stage 3a) and SBCC (stage 5a), our two calibration points. Values recorded by the Punta Gorda molluscs appear to correlate well with those from Isla Vista, and values recorded by the Santa Barbara Point molluscs appear to correlate well with those from SBCC. To evaluate these data quantitatively, we analysed all data points (Table 1), minus shells yielding anomalous values, from all terraces using the Kruskal-Wallis one-way analysis of variance by ranks. The results of this analysis are shown in Table 3. The null hypothesis (H0) that is either accepted or rejected is that there is no significant difference in values, or that the sample pairs came from the same population (i.e. have the same age). Values of H which are so large that the probability associated with their occurrence under H0 is equal to or less than 0.05 are rejected. The results confirm that molluscs from the stage 5a terraces preserve a distinctly different signature from that of molluscs from the 3a terraces. They also show that the Punta Gorda terrace can be correlated with the Isla Vista terrace, and the Santa Barbara Point terrace can be correlated with the terrace at SBCC. Our data corroborate previous studies indicating that the Punta Gorda terrace correlates with oxygen
isotope stage 3 (Kaufman et al. 1971; Wehmiller et al. 1978; Lajoie et al. 1982). Furthermore, the basis of correlation with the SBCC terrace, we are able to assign an age of stage 5a to the first emergent terrace at Santa Barbara Point. (It is important to note that when the anomalous values are included in the statistical analysis the correlations remain the same except that Santa Barbara Point correlates with the Isla Vista values instead of the SBCC values. On the basis of lateral continuity of this terrace with the SBCC terrace, we strongly believe that these are the same terrace; therefore, we feel confident that the anomalously high values recorded at Santa Barbara Point are not representative of the terrace.) Punta Gorda
The results from Punta Gorda provide a significant contribution to this study. At the sample locality, the Punta Gorda terrace is offset about 34 m by a strand of the Red Mountain fault. It was initially unknown whether the two terrace surfaces were the same terrace simply offset by the fault or were two entirely different terraces with the upper surface representing an older terrace. Samples were taken from both the upthrown and downthrown block of the Punta Gorda terrace to determine whether or not these surfaces could be correlated (Fig. 5). Although the results from the downthrown block are somewhat spread out (Table 2), statistical analysis (Table 3) of all data points verifies that these surfaces can be correlated. Correlation of these surfaces across this strand of the Red Mountain fault provides a basis for calculating a slip rate on this fault (Huftile et al. 1997a, b).
Application to the Mesa hills, Santa Barbara The northwest trending Mesa hills are formed as a result of north-verging folding on the hanging
MARINE TERRACES OF MESA HILLS, CALIFORNIA wall of the Mesa fault. To the west the fold is truncated by the Mission Ridge fault system and to the east the fold plunges offshore. Uplifted marine terraces and associated palaeo-shorelines are preserved on the southern and eastern flanks of the fold (Fig. 4). Along the eastern nose of the Mesa fold at Point Castillo, a marine terrace fossil assemblage is exposed in the sea-cliff of the first emergent marine terrace at SBCC (Figs 3 and 4). A marine terrace fossil collection in the Department of Geological Sciences palaeontology collection at University of California, Santa Barbara, contained a single well-preserved, fossil solitary coral Balanophyllia elegans from the sea-cliff exposure at the SBCC fossil site known as 'Bath House Beach' (M-1594). The fossil coral yielded a 23~ age of 70:t:2kaBP with an initial 234u/Z38u activity of 1.164-t-0.015, which is within the same error as that for modern seawater (Gurrola et al. 1997). This age correlates with oxygen isotope stage 4, a sea-level lowstand; however, because marine terraces are rarely preserved at lowstands of sea level, the first emergent marine terrace may be assigned to the highstand at either substage 3c (58 ka) or substage 5a (81 ka). To correlate terraces with the sea-level curve one must assume a constant uplift rate; this means that the lines connecting each terrace to its sea-level highstand will be parallel, on the basis of spacing of the terraces (Bull 1985). Using this assumption we assign the first emergent terrace at SBCC to stage 5a, based on the fact that this yields parallel lines using the Bull (1985) method. If we assume that the terrace formed during stage 3c, the lines connecting each terrace to its highstand would not be parallel. On the basis of these facts, for the following calculations we use 81 ka (stage 5a) as the age of the first emergent terrace at SBCC. There are four uplifted terraces and associated palaeo-shorelines preserved in the same flight at SBCC. The palaeo-shoreline elevations of the four terraces beginning with the first emergent 81 ka terrace are 36, 45, 64, and l l 0 m (Figs 4 and 6a). On the SBCC terrace, on the basis of the U-series age of 81 ka for the first emergent marine terrace and assuming a constant rate of uplift, we have estimated ages of 102, 124, and 212 ka for the 45, 64, and 110 m palaeoshorelines, and calculate a rate of uplift of 0.54 + 0.05 mm/year. To the south at Santa Barbara Point, which we have correlated with oxygen isotope analysis to the 81ka terrace at SBCC, four uplifted terraces are preserved with palaeo-shoreline elevations of 41, 59, 81 and 120 m (Figs 4 and 6b). An additional terrace is preserved along the axis of the Mesa fold at a maximum elevation of c. 137m. Based on the correlation of the first
65
emergent terrace, we have determined that the Santa Barbara Point terraces represent ages of 81,102, 124,212, and 300 ka, and the rate of uplift for this flight is 0.62 + 0.03 ram/year. To the west, at Meigs Road, we use geomorphological evidence (lateral continuity of the terrace) to identify the 81, 102, 124, and 300ka palaeo-shorelines, which occur at elevations of 38, 55, 75, and 143 m (Figs 4 and 6c), respectively. They yield an uplift rate of 0.59 + 0.03 mm/year. Further to the west, on the Mesa Lane flight, we use geomorphological evidence (lateral continuity of the terrace) to identify the 81,102, 124, 212, and 300ka palaeo-shorelines, which occur at elevations of 37, 45, 64, 110, and 124m, respectively (Figs 4 and 6d) yielding a rate of uplift of 0.55 4- 0.05 mm/year. The longitudinal variation in the palaeo-shoreline angle elevations reflects the proximity to the La Mesa anticline and syncline fold axes (Fig. 4). The Mesa anticline and Honda Valley syncline clearly deform the terrace shorelines in the area of the Santa Barbara Point and SBCC terraces. For example, the marine palaeo-shorelines on the SBCC terrace are lower in elevation than the Santa Barbara Point terrace. Additionally, terrace shorelines are folded across the axis of the Mesa anticline, with older terrace shorelines exhibiting greater deformation. Because of the ambiguity of the coral date at SBCC, it is possible that these uplift rates are underestimates if the age of the first emergent terrace at SBCC is, in fact, 58 ka (substage 3c). Taking this into account, the rate of uplift of the SBCC terrace ranges from 0. 54 -+-0.05 mm/year to 1.13 4-0.10 mm/year with an average rate of vertical uplift of 0.84 :t: 0.08 mm/year. However, we restate that using the age of 5a for the first emergent terrace at SBCC best suits the Bull (1985) method of correlating marine terraces and using the age of 3c does not yield the parallel lines required by the method, because of the ages and spacing of the terraces.
Discussion and conclusions The proposed correlation is of value in resolving the uplift rate of the Mesa hills, a young, potentially active anticline located in the city of Santa Barbara. Using the Santa Barbara Point oxygen isotopic data, geomorphological correlations, and the U-series dated terrace at SBCC, we calculate uplift rates ranging from 0.62 + 0.03 mm/year near the crest of the Mesa anticline to 0.54 :t: 0.05 mm/year at the Honda Valley syncline. The range of uplift rates reflects the different rates and styles of deformation along the Mesa hills structures.
Elevation (m) (a) SBCC 916o
12o _ ~ .._ ,1t
/
~ _
-
~ Ii ~ . 80 . . . .
w./I~..~-%"
I
I
i
I
I
I
Uplift
Se --
I
2o0
~ 6o
I
I
|
=
Age (ka)
loo U-series coral age (70 +/- 2 ka) offirst emergent marine terrace
. . . . Santa ~arDara ~'oint
(b)
5e
~5aC'~"~mergent Shorelines I
180 Stages
3oo
/
7c
I Elevation (m) ~,~o L o
~.I
i /
~
I 120"-- ~
~
~
,. ~mergem ~nore,nes
.~ ~ . ~
~
I
.
,
180 Stages
,
300
J
-160
|
200 Age (ka)
,
loo
,
|
U-series coral age (70 +/- 2 ka) offirst emergent marine terrace
Fig. 6. Calculation of uplift rates. (a) Correlation of the first emergent terrace at SBCC with oxygen isotope stage 5a, on the basis of U-series age of 70 • 2. This yields a vertical rate of uplift of 0.54 + 0.05 ram/year at the SBCC terrace. (b) Correlation of the first emergent terrace at Santa Barbara Point with oxygen isotope stage 5a, on the basis of correlation with the U-series dated SBCC terrace using oxygen isotope stratigraphy. This yields a vertical rate of uplift of 0.62 4- 0.03 mm/year at the Santa Barbara Point terrace. (c) Correlation of the first emergent terrace at Meigs Road with oxygen isotope stage 5a, on the basis of geomorphological correlation, with the Santa Barbara Point and SBCC terraces. This yields a vertical uplift rate of 0.59 + 0.03 mm/year for the Meigs terrace. (d) Correlation of the first emergent terrace at Mesa Lane with oxygen isotope stage 5a, on the basis of geomorphological correlation, with the Santa Barbara Point and SBCC terraces. This yields a vertical uplift rate of 0.55 + 0.05 mm/year for the Mesa Lane terrace. (See Fig. 4 for locations.) Sea-level curve from Chappell (1983) and Chappell & Shackleton (1986).
MARINE TERRACES OF MESA HILLS, CALIFORNIA
67
Elevation (m)
(C) Meigs Rd.
160 t
i ~ i
1
t
9
12~
@@l~~J
80
0.~~ t" l - -
1 11 f/
1
~
5~
/
& I uplift
i -~ ~ ' ~ ~=~"~mergent Shorelines I "~ "S-
i/'~
Ir'/l
-80
180 Stages
,
|
,
,
300
,
-160
,
200
,
Age (ka)
,
100
,
,
,
U-series
coral
a.qe (70 +/- 2 ka) of first emergent marine terrace
I Elevation (m)
(d) Mesa Lane
160
9 1 1
, ~
, ~
-
~
r' ' ' " ~
m" ~"
ill
"'-
5e
180 Stages
|
i
300
,
i
l
= 200
,
-160
i
Age (ka)
-
f l f'~
,
,
,
,
~oo U-series coral aqe (70 +/- 2 ka) offirst emergent marine terrace
Fig. 6. (continued) The uplift rates recorded in the Mesa hills are significantly lower than the rate of 1.22 +0.13 ram/year calculated by Gurrola et al. (1996) at Isla Vista, about 10 km away. This large difference in uplift rates over a very short geographical distance demonstrates the utility of oxygen iso-
tope correlations in providing a means for gaining a higher resolution record of uplift rates. Further development of this correlation method will allow calculation of uplift rates for marine terraces where no corals are found, provided a regional calibration point exists.
68
M. A. T R E C K E R ET AL.
This, in turn, will provide a means for obtaining m u c h greater resolution of uplift rates along tectonically active coastlines and allow for lateral differences in uplift rates to be recognized and both rates and styles of d e f o r m a t i o n along local structures to be evaluated. Isotopic analyses were performed in J. P. Kennett's stable isotope laboratory at the University of California, Santa Barbara, where K. Thompson played an integral part in the laboratory work. D. Pierce greatly aided in the XRD analysis. This project was supported in part by the Southern California Earthquake Center (SCEC) Grant USC 572726. SCEC is funded by NSF Cooperative Agreement EAR-8920136 and USGS Cooperative Agreements 14-08-0001-A0899 and 1434-HQ97AG01718. The SCEC contribution number for this paper is 413. This project was also supported by US Geological Survey/NEHRP Grant USGS 1434HQ97GR02978. UCSB Institute for Crustal Studies contribution number 0288-81TC.
References
BULL, W. B. 1985. Correlation of flights of global marine terraces. In: MORISAWA, M. & HACK, J. (eds) Tectonic Geomorphology. Proceedings, 15th Binghamton Geomorphology Symposium. Allen & Unwin, Boston, MA, 129-152. CHAPPELL, J. 1983, A revised sea-level record for the last 300,000 years from Papua New Guinea. Search, 14(3-4), 99-101. - • SHACKLETON, N . J. 1986. Oxygen isotopes and sea level. Nature, 324, 137-140. COPLEN, T. B., KENDALL, C. & HOPPLE, J. 1983. Comparison of stable isotope reference samples. Nature, 302, 236-238. DANSGAARD, W., JOHNSEN, S. J., CLAUSEN,H. B. et al. 1993. Evidence for general instability of past climate from a 250-kyr ice-core record. Nature, 364, 218-220. -& MOLLER, J. 1969. One thousand centuries of climatic record from Camp Century on the Greenland ice sheet. Science, 166, 377-381. DIBBLEE,T. W., JR 1986. In: EHRENSPECK(ed.) Geologic map of the Santa Barbara quadrangle. Dibblee Geological Foundation, Santa Barbara, CA. GURROLA, L. D. & KELLER, E. A 1997. Tectonic geomorphology of the Santa Barbara fold belt, western Transverse Ranges, California. Geological Society of America, Abstracts with Programs, 29(6), A344. , CHEN, J. H. & KELLER, E. A. 1997. Uraniumseries age and rate of uplift of the Mesa marine terraces, Santa Barbara Fold Belt, California. Southern Cal!fornia Earthquake Center Annual Meeting, October 1997, 64. , , -& METCALF, J. G. 1996. Dating of emergent marine terraces and rate of uplift for the western Santa Barbara Fold Belt, California. Geological Society of America, Abstracts with Programs, 28(7), A301.
HUFTILE, G. J., LINDVALL,S. C., ANDERSON,L. W. & GURROLA, L. 1997a. Paleoseismic investigation and tectonic implications of the Red Mountain Fault, Southern California. Los Transactions, American Geophysical Union, Supplement, 78(46), F635. & TRECKER, M. A. 1997b. Paleoseismic Investigations of the Red Mountain Fault at Punta Gorda. Abstracts from Southern California Earthquake Center Annual Meeting, October 1997, Supplement. KAUFMAN, A., BROECKER, W. S., Ku, T.-L. & THURBER, D. L. 1971. The status of U-series methods of mollusk dating. Geochimica et Cosmochimica Acta, 35, 1155-1183. KENNEDY, G. L., LAJOIE, K. R. & WEHMILLER, J. F. 1982. Aminostratigraphy and faunal correlations of late Quaternary marine terraces, Pacific Coast, USA. Nature, 299, 545-547. KING, N. E. 8r SAVAGE, J. C. 1984. Regional deformation near Palmdale, California 1973-1983. Journal of Geophysical Research, 89, 2471-2477. KOTILAINEN,A. T. & SHACKLETON,N. J. (1995) Rapid climate variability in the North Pacific Ocean during the last 95,000 years. Nature, 377, 323-326. LAJOIE, K. R., KERN, J. P., WEHMILLER, J. F. et al. 1979. Quaternary marine shorelines and crustal deformation, San Diego to Santa Barbara, California. In: ABBOT, P. L. (ed.) Geological Excursions in the Southern Calijornia Area. Guidebook for Field Trips, Geological Society of America Annual Meeting, November 1979. Department of Geological Sciences, San Diego State University, San Diego, CA, 3-15. - - , SARNA-WOJCICKI,A. M. & YERKES, R. F. 1982. Quaternary chronology and rates of crustal deformation in the Ventura area, California. In: COOPER, J. D. (ed.) Neotectonics 0l Southern California. GSA, Guidebook for Field Trips, Anaheim, CA. METCALE, J. G. 1994. Morphology, chronology, and deformation of Pleistocene marine terraces, southwestern Santa Barbara County, California. MS thesis, University of California, Santa Barbara. MORRIS, R. H., ABBOTT, D. P. & HADERLIE, E. C. 1980. Intertidal Invertebrates of California. Stanford University Press, Stanford, CA, 290-292. MUHS, D. R., MILLER, G. H., WHELAN, J. F. & KENNEDY, G. L. 1992. Aminostratigraphy and oxygen isotope stratigraphy of marine terrace deposits, Palos Verdes Hills and San Pedro areas, Los Angeles County, California, In: FLETCHER, C. H. I l I & WEHMULLER, J. F. (eds) Quaternary Coasts of the United States: Marine and Lacustrine Systems. Society of Economic Paleontologists and Mineralogists, Special Publications, 48, 363-376. SAVAGE, J. C. 1983. Strain accumulation in western United States. Annual Review of Earth and Planetary Sciences, 11, 11-43. TAYLOR, K. C., LAMOREY, G. W., DOYLE, G. A. et al. 1993. The 'flickering switch' of late Pleistocene climate change. Nature, 361,432-436. THUNELL, R. C., & MORTYN, P. G. 1995. Glacial climate instability in the northeast Pacific. Nature, 376, 504-506.
M A R I N E T E R R A C E S OF MESA HILLS, C A L I F O R N I A TRECKER, M. A., KELLER, E. A., KENNETT, J. P. & GURROLA, L. D. 1997a. Oxygen isotope strati graphy as a means of correlating deformed marine terraces to calculate uplift rates, Santa Barbara Fold Belt, California. Abstracts from Southern
California Earthquake Center Annual Meeting, October 1997, 90. & - - 1 9 9 7 b . Correlation of deformed marine terraces, Santa Barbara Fold Belt, California. Eos Transactions, American Geophysical Union, Supplement, 78(46), F632. VALENTINE, J. W. 1961. Paleoecologic molluscan geography of the Californian Pleistocene. University
of California Publications in Geological Sciences, 34(7), 309-442.
69
& MEADE, R. F. 1963. Californian Pleistocene paleotemperatures, University of California Publications in Geological Sciences, 40, 1-45. WEHMILLER, J. F., LAJOIE, K. R., SARNA-WOJCICKI, A. M., YERKES, R. F., KENNEDY, G. L., STEPHENS, T. A. & KOH, R. F. 1978. Amino-acid racemization dating of Quaternary mollusks, Pacific Coast, United States. In: ZARTMAN, R. E. (ed.) Short Paper of the Fourth International
-
-
Conference, Geochronology, Cosmochronology, Isotope Geology. US Geological Survey OpenFile Report, 78-701, 445-448.
Deformation of the 125 ka marine terrace in Italy: tectonic implications PAOLA
BORDONI
& GIANLUCA
VALENSISE
Istituto Nazionale di Geofisica, Via di Vigna Murata 605, 00143 Rome, Italy (e-mail." bordoni@ ingrm.it) Abstract: In peninsular Italy, marine terraces and shorelines created during oxygen-isotope substage 5e, known historically as the Tyrrhenian highstand, provide a reliable and homogeneous datum of vertical motions during the past 125 000 years. Published accounts of 121 locations on the Tyrrhenian inner edge were reinterpreted and used to calculate both local and regional uplift rates. The data show arching of southern Italy between the Tyrrhenian and the Adriatic coasts and of the Calabrian Arc between the Sangineto and the Longi-Taormina tectonic lineaments at rates of up to 1.2 mm/year for the past 125 000 years. They also suggest that the present configuration of the southern Apennines is the product of post-Early Pleistocene regional uplift superimposed on Tortonian to Early Pleistocene thrusting. Uplift patterns are characterized by different wavelengths in different geodynamic settings, suggesting a close link with the depth of the causative tectonic sources. The occurrence of raised marine deposits and emergent marine terraces is one of the most striking landscape features of a large portion of the Italian coast. For over a century these deposits and the associated landforms have drawn the attention of European Quaternary geologists (see, e.g. La Marmora (1857), Cortese (1895), Gignoux (1913), Issel (1914), among many others), and many stage names, including Calabrian and Sicilian, were derived from the names of coastal areas where representative deposits were well exposed. Nearly all of the early workers focused almost exclusively on the stratigraphy of the uplifted marine deposits and paid little attention to their tectonic significance, and it was only after the pioneering work by Veeh (1966) and MesolelIa et al. (1969) on marine terrace sequences of Barbados that the role of sea-level fluctuations in the generation of marine terraces and elevated coastal features began to be fully appreciated. The ability to quantify the role of climatic changes further promoted interest in marine terraces as indicators of tectonic activity at various scales. In this study we focus exclusively on Italian marine terraces assumed to have formed about 125 000 years ago during the climatic optimum that correlates with substage 5e of the oxygenisotope record. This particular episode of the Pleistocene climatic record has left a clear trace in many parts of the world, partly because, as global sea level stood a few metres above the present level, it ensured the preservation of associated deposits and landforms even in tectonically stable areas. Moreover, unusually warm climatic conditions favoured, in the Mediterranean area, the development of distinctive faunal assemblages.
The index fossil Strombus bubonius Lmk can be taken as the symbol of this highstand, which in Italian geological nomenclature has long been equated with the Tyrrhenian stage first proposed by Depbret (1918). As investigations of the Late Pleistocene progressed, however, it became clear that some of the deposits collectively ascribed to the Tyrrhenian stage had in fact been deposited during successive yet very closely spaced highstands, and the Tyrrhenian level was subdivided into 'Neotirreniano', 'Eutirreniano' and 'Paleotirreniano' (Bonifay & Mars 1959). In the 1970s these three levels were reinterpreted as the stratigraphic counterparts of the three climatic optima corresponding to substages 5c-a, 5e and possibly 7, respectively, of the oxygen-isotope record. More recent work based on extensive aminostratigraphic correlation around the Mediterranean suggests that S. bubonius is restricted to substage 5e, and that its occurrence at multiple levels is due to minor fluctuations within 5e itself (Hearty & Dai Pra 1986, 1987; Hearty et al. 1986a; Hearty 1987; Dumas & Raffy 1993). For this reason, as well as for the sake of simplicity, we have retained the traditional term Tyrrhenian for terraces and deposits ascribed to substage 5e or dated to c. 125ka. During our systematic collection and reevaluation of Italian marine terraces of Tyrrhenian age we have drawn on earlier compilations, including those by Gortani (1950), Selli (1962), Hearty (1986), Parea (1986), Ambrosetti et al. (1987), Cosentino & Gliozzi (1988), and Westaway (1993). In particular, the results presented by Cosentino & Gliozzi (1988) are summarized in Fig. 1.
BORDONI, P. & VALENSISE,G. 1998. Deformation of the 125 ka marine terrace in Italy: tectonic implications. In: STEWART,I. S. & VITA-F1NZl,C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 71-110.
72
P. BORDONI & G. VALENSISE
Fig. 1. Structural sketch of southern Italy and Sicily. The figure shows also the rate of uplift inferred by Cosentino & Gliozzi (1988) from 31 sites along the Tyrrhenian shoreline (in mm/year).
Data distribution Figure 2 shows the complete distribution of our collected Tyrrhenian marine terrace sites. The density of terrace data points is very high only along the southern Tyrrhenian Sea and Ionian
Sea coasts including Sicily, acceptable along the central Tyrrhenian coasts and in western Sicily, and very low along the Adriatic Sea and northern Tyrrhenian Sea coasts. This variable density of information reflects only in part an uneven literature search: as pointed out by Selli
DEFORMATION OF THE 125ka MARINE TERRACE IN ITALY
73
Fig. 2. Distribution of observations of Tyrrhenian marine terraces and shorelines described in this paper. Numbers in small typesize are site numbers, those in larger typesize indicate the conventional elevation assigned to each site within this paper. Both site numbers and elevations correspond to those reported in Table 1. (For coordinates, name and description of each site, refer to Table 1.) (1962), some inner portions of the Mediterranean, including the Adriatic Sea, may have never experienced the warmer conditions that characterized the Tyrrhenian climatic optimum and favoured the northward migration of the warm fauna. Other coastal areas either did not
favour the survival of the evidence, were largely inaccessible, or were simply never investigated in detail. The various sites appear in the description with the same reference number as in Table 1 and Figs 2-7 although not all the localities cited
74
P. B O R D O N I & G. V A L E N S I S E
Table 1. Summary of observations of Tyrrhenian shorelines Site Locality no.
Latitude (~
Longitude (~
Max. elevation of deposits (m asl)
Shoreline elevation (m a s l )
Geomorphological evidence
Conventional Dating method elevation of Tyrrhenian shoreline
1
Buca dei Corvi (Livorno)
43.364
10.455
11
11
Wave-cut platform
ll
Palaeontology and aminostratigraphy
2
Selva Nera, northeast of Lake Burano
42.411
11.406
11
11
Abrasion notch
11
Aminostratigraphy
3
Lasco di Pozzo, south of Pescia Romana
42.384
11.497
7
10
Dune ridge Abrasion notch
10
Aminostratigraphy
4
State Hwy Aurelia, Km 115.5
42.384
11.545
20
Palaeolagoonal surface
20
Aminostratigraphy
5
North bank of Fiora River
42.358
11.561
12
12
Aminostratigraphy
6
Between Arrone River and Mignone River
42.290
11.670
20
20
20
Palaeontology
7
I1 Mandrione, right bank of Arrone River
42.130
11.642
24
28
28
Palaeontology
8
State Hwy Aurelia, Km 103.0
42.289
11.682
25
30
30
Aminostratigraphy
9
Between Arrone River and Mignone River
42.245
11.773
45
45
45
U - T h and aminostratigraphy
10
Northwest of Tarquinia Railway Station
42.230
11.730
14
27
27
Palaeontology and aminostratigraphy
11
Colle Olivastro, north of Civitavecchia
42.192
11.760
16
30
30
Palaeontology and aminostratigraphy
12
Piana del Termine, north of Civitavecchia
42.173
11.794
35
45
Inner edge
45
Aminostratigraphy
13
Monna Felice, near Civitavecchia
42.110
11.811
27
40
Inner edge
40
Palaeontology
14
Santa Severa
42.024
11.940
35
Inner edge
35
Geomorphological correlation
15
Monteroni, near Cerveteri
41.994
12.109
27
30
Dune ridge
30
Palaeontology and aminostratigraphy
16
Casale di Statua between Palidoro and Cerveteri
41.933
12.207
19
35
17
Valle della Sargia, west of Ponte Galeria
41.849
12.262
35
Terrace surface
Palaeontology and aminostratigraphy Inner edge
Geomorphological correlation
DEFORMATION
O F T H E 125ka M A R I N E T E R R A C E IN I T A L Y
75
Significantfossil assemblage
Reference(s)
Uplift rate (ram/year)
Senegalese fauna
Blanc (1953a); Hearty et aI. (1986b); Hearty & Dai Pra (1987)
0.04
Aminozone E 5e (A/I = 0.40 • 0.02)
Hearty & Dai Pra (1987)
0.04
130
Aminozone E 5e (A/I = 0.42 • 0.01)
Hearty & Dai Pra (1987)
0.03
125
Aminozone E 5e (A/I = 0.39 • 0.04)
Hearty & Dai Pra (1987)
0.11
125
Aminozone E 5e (A/I = 0.39 • 0.02)
Hearty & Dai Pra (1987)
0.05
Age (ka)
Aminozone U-Th ratio
130
Aminozone E 5e (A/I = 0.35 • 0.03)
130
Oxygen isotope substage
Cerastoderma
Conus testudinarius; Gignoux (1913); Bonadonna (1967a) Pecten jacobeus
0.11
Strombus bubonius Lmk; Glycymeris
Palieri & Sposato (1988)
0.18
Glycymeris
Hearty & Dai Pra (1986)
0.19
102; 92
Glycymeris; Cerastoderma
Radtke (1986)
0.31
125
Strombus bubonius Lmk
Ambrosetti et al. (1981); Palieri & Sposato (1988)
0.17
125
Aminozone E 5e (A/I = 0.40 • 0.02)
130
Aminozone E 5e (A/I = 0.39 + 0.02)
Strombus bubonius Lmk
Hearty & Dai Pra (1986, 1987)
0.19
125
Aminozone E (A/I = 0.410)
Glycymeris violacescens
Dai Pra (1978); Hearty & Dai Pra (1986, 1987)
0.31
Strombus bubonius Lmk
Dai Pra (1978) and references therein
0.27
Dai Pra (1978) and references therein
0.23
5e
125
Aminozone E 5e (A/I = 0.41 • 0.01)
Strombus bubonius Lmk; Glycymeris
Blanc (1936); Dai Pra (1978); Hearty 0.19 (1986); Hearty & Dai Pra (1986, 1987)
125
Aminozone E 5e (A/I = 0.37 • 0.02)
Strombus bubonius Lmk; Glycymeris
Blanc (1936); Dai Pra (1978); Hearty 0.23 (1986); Hearty & Dai Pra (1986, 1987)
125
5e
Bonadonna (1967b); Dai Pra (1978) and references therein
0.23
76
P. B O R D O N I & G. V A L E N S I S E
Table 1.
(continued) Max. elevation of deposits (m asl)
Shoreline elevation (m a s l )
Geomorphological evidence
Conventional Dating method elevation of Tyrrhenian shoreline
Site Locality no.
Latitude (~
Longitude (~
18
Tacconi quarry, near Pomezia
41.653
12.493
45
19
Ardea, left bank of Fosso Moletta
41.610
12.500
20
20
Borgo Santa Maria, west of Latina
41.460
12.810
12
21
Borgo Sabotino (Pontina Plain)
41.430
12.820
-4
22
Torre Rossa (Mt. Circeo)
41.224
13.065
9.6
Abrasion notch
9.6
Aminostratigraphy
23
'Pisco Montano', near Terracina
41.290
13.260
7.3
Abrasion notch
7.3
Geomorphological correlation
24
Well drilled in Fondi Plain
41.302
13.338
25
Grotta di Tiberio, near Sperlonga
41.256
13.453
7
7
Abrasion notch
7
Aminostratigraphy
26
Torre Capovento, between Sperlonga and Gaeta
41.244
13.471
4.8
7.4
Lithofag&
7.4
Palaeontology and geomorphological correlation
27
Arenauta beach, near Gaeta
41.220
13.539
7.5
Abrasion notch
7.5
Geomorphological correlation with site 28
28
Montagna Spaccata (Gaeta promontory)
41.212
13.575
4.8
Abrasion notch
4.8
Aminostratigraphy
29
Marina di Minturno (Garigliano Plain)
41.268
13.715
8
9.6
Abrasion notch
9.6
Aminostratigraphy
30
Masseria Transitiello (Garigliano Plain)
41.202
13.832
5
8.0
Abrasion notch
8.0
Aminostratigraphy
31
Cancello, left bank of Volturno River
40.994
14.434
32
'S. Marcellino' well, left bank of Volturno River
40.991
14.176
Inner edge
45
Geologicalstratigraphic correlation
20
Geomorphological correlation
12
Aminostratigraphy
Palaeontology and aminostratigraphy
6
-6
-6
borings in sea cave
50
50
-50
Top of dated unit
-50
Aminostratigraphy
Geologicalstratigraphic correlation U-Th
D E F O R M A T I O N OF T H E 125ka M A R I N E T E R R A C E IN I T A L Y
Age (ka)
Aminozone U-Th ratio
125
Oxygen isotope substage
Significantfossil assemblage
Reference(s)
77
Uplift rate (mm/year)
5e
Dai Pra & Arnoldus-Huyzendveld (1984); Malatesta & Zarlenga (1986); Milli & Zarlenga (1991) and references therein
0.31
5e
Dai Pra & Arnoldus-Huyzendveld (1984)
0.11
135-117
Aminozone E 5e (A/I = 0.39 + 0.01)
Glycymeris
Hearty & Dai Pra (1986)
0.05
135-117
Aminozone E (A/ I = 0.037 • 0.02)
Strombus bubonius Lmk
Blanc (1935); Dai Pra & ArnoldusHuyzendveld (1984); Hearty & Dai Pra (1986); Dai Pra (1995)
0.00
Glycymeris
Blanc (1953b); Hearty & Dai Pra (1986); Ozer et al. (11987)
0.03
Antonioli et al. (1988) and references therein
0.01
5e
Aminozone E 5e (A/I = 0.38 + 0.03)
130
Aminozone E 5e (A/I = 0.42 + 0.02)
Cerastoderma
Antonioli et al. (1988)
125
Aminozone E 5e (A/I = 0.37 + 0.02)
Glycymeris
Segre (1949); Ozer et al. (1987); Antonioli (1991)
0.01
5e
Glycymeris
Blanc & Segre (1947); Ozer et al. (1987); Antonioli (1991)
0.01
Antonioli (1991); Dai Pra (1995)
0.01
Aminozone E (A/I = 0.33)
5e
Glycymeris; Ostrea
Antonioli (1991)
-0.10
-0.01
131
Aminozone E 5e (A/I = 0.37 + 0.07)
Hearty & Dai Pra (1986)
0.03
130
Aminozone E 5e (A/I = 0.40 + 0.02)
Segre (1957); Brancaccio et al. (1990)
0.02
125
5e
Romano et al. (1994)
0.35
Romano et al. (1994)
-0.45
126 + 11 23~ = corrected age 0.731 + 0.024 and 0.725 • 0.030
Cladocora caespitosa
78
P. B O R D O N I & G. V A L E N S I S E
Table 1.
(continued) Shoreline elevation (m a s l )
Geomorphological evidence
Conventional Dating method elevation of Tyrrhenian shoreline
Site no.
Locality
Latitude (~
Longitude Max. (~ elevation of deposits (m asl)
33
'Sarno' well driUed between Sarno and Nocera Inf.
40.782
t4.649
34
Massalubrense
40.611
14.348
35
Baia di Ieranto, near Massalubrense
40.582
14.337
36
Vettica Maggiore, near Positano
40.610
14.534
37
Grotta Smeralda, near Amalfi
40.636
14.615
38
Ponte Barizzo, left bank of Sele River
40.488
15.028
39
Agropoli
40.355
15.009
5.5
Wave-cut platform
5.5
Geomorphological correlation with site 42
40
Trentova, near Agropoli
40.343
14.985
7
Wave-cut platform
7
Geomorphological correlation with site 42
41
Acqua di Cesare, near Castellabate
40.264
14.934
Abrasion notch
10
Geomorphological correlation with site 42
42
Ogliastro Marina
40.236
14.961
4
8
Aminostratigraphy
43
Lido Ficoncella, near Palinuro
40.056
15.308
8
8
Geomorphological correlation with site 42
44
Palinuro promontory
40.030
156.300
2.2
Geomorphological correlation
45
Sapri
40.071
15.629
46
Acquafredda
40.036
15.671
5.5
6.9
47
Marina di Maratea
39.963
15.742
9.3
>9.3
48
Punta Iudia, south of Maratea
39.947
15.744
49
Porticello di Castrocucco, north of Praia a Mare
39.933
15.750
-29
-29
8
Geologicalstratigraphic correlation and U-Th Geomorphological correlation
Abrasion notch
U-Th
4
6.8
Abrasion notch
6.8
Geomorphological correlation with site 35
8
Lithophaga
8
Palaeontology
borings Dune ridge
25
10
Inner edge of 5e & 5ac deposits 8
2.2 15
Abrasion notch Inner edge
25
15
Aminostratigmphy
Aminostratigraphy
Inner edge
6.9
Aminostratigraphy
Terrace surface
9.3
Aminostratigraphy
9.5
Wave-cut platform
9.5
Geologicalstratigraphic correlation
9.5
Abrasion notch
9.5
Geomorphological correlation
D E F O R M A T I O N OF T H E 125ka M A R I N E T E R R A C E IN I T A L Y
Age (ka)
+41
217_28
Aminozone U-Th ratio
Oxygen isotope substage
Significantfossil assemblage
Reference(s)
23~ -0.888 + 0.024
5e
Cladocora caespitosa
Cinque et aL (1987); Barra et al. (1991) -0.28
129
Brancaccio (1968)
0.02
5e
Brancaecio et aL (1978); Cinque & Romano (1990)
0.01
5e
Brancaccio et al. (1978); Cinque & Romano (1990)
0.01
Colantoni (1970); Cosentino & Gliozzi (1988)
0.02
Brancaceio et aL (1986, 1987); Russo & BeUuomini (1992)
0.15
5e
Cinque et aL (1994)
0.00
5e
Cinque et al. (1994)
0.01
5e
Cinque e t a l . ( 1 9 9 4 )
0.03
Brancaccio et al. (1990); Cinque
0.02
Strombus bubonius
Lmk
110
Aminozone (A/I = 0.42 + 5%)
Aminozone C
5e
5a-c
Glycymeris
Glycymeris
et al. (1994)
(A/I = 0.31 + 0.02)
100-120
Uplift rate (mm/year)
5e
5e 125
79
5e
Brancaccio et aL (1990)
0.02
5e
Antonioli et aL (1994b)
-0.03
110
Aminozone E 5a-c (A/I = 0.33 + 0.01)
Glycymeris
Brancaeeio et al. (1990)
0.07
126
Aminozone E (A/I =0.437)
5e
Arca
Carobene & Dai Pra (1991)
0.01
Aminozone E (A/I=0318)
5e
Arca
Carobene & Dai Pra (1991)
0.03
Damiani (I970) and references therein; Carobene et al. (1986); Cosentino & Gliozzi (1988); Carobene & Dai Pra
0.03
5e
(1991) 5e
Carobene & Dai Pra (1991)
0.03
80
P. B O R D O N I & G. V A L E N S I S E
Table 1.
(continued)
Site Locality no.
Latitude (~
Longitude (~
50
39.874
15.785
Grotta dei Maiali-Isola di Dino, near Praia a Mare
Max. elevation of deposits (m asl) 8
Shoreline elevation (m a s l ) 9
Geomorphological evidence
Conventional Dating method elevation of Tyrrhenian shoreline Geomorphological correlation
Wave-cut platform Abrasion notch
Lithophaga 51
Torre Talao, near Scalea
39.823
15.792
52
Cirella, south towards Diamante
39.700
15.820
53
Punta la Testa, near Cetraro
39.526
15.926
54
Vibo Valentina Marina
38.708
16.112
55
Capo Vaticano, southwest of Tropea
38.618
56
Capo Vaticano, southwest of Tropea
57
Wave-cut platform
Geologicalstratigraphic corrrelation
Abrasion notch
Geomorphological correlation
12
Sea-cave
Geomorphological correlation
50
Inner edge
50
Palaeontology U - T h and aminostratigraphy
15.840
88
Inner edge
88
geomorphological correlation
38.618
15.840
120
Inner edge
120
Geomorphological correlation with site 54
Contrada Ianni, near Nic6tera
38.550
15.940
80
Inner edge
80
Palaeontology
58
Gioia Tauro
38.530
16.040
Inner edge
65
Palaeontology
59
Scilla
38.248
15.715
Inner edge
150
Geomorphological correlation
60
San Francesco di Archi, north of Reggio Calabria
38.141
15.663
96
96
Aminostratigraphy
61
Ravagnese, south of Reggio Calabria
38.081
15.657
100
Terrace surface
100
Palaeontology and aminostratigraphy
62
Bovetto, south of Reggio Calabria
38.065
15.664
125
Terrace surface
125
Palaeontology and aminostratigraphy
63
Nocella, southeast of Pellaro
38.018
15.658
157
Terrace surface
157
Palaeontology and aminostratigraphy
64
Lazzfiro
37.968
15.677
146
146
Aminostratigraphy
65
Between Fiumara Condoianni and Vallone Notaro, near Ardore Marina
38.185
16.194
86
86
Geomorphological correlation
<8
8-10
38
65 150
Inner edge
D E F O R M A T I O N OF T H E 125ka M A R I N E T E R R A C E IN I T A L Y
Age (ka)
Aminozone U-Th ratio
Oxygen isotope substage
Significantfossil assemblage
5e
1 1 0 f~2 +6694 ..... -6305
Aminozone E (A/I = 0.36); 53~ = 0.677 4- 0.023 and 0.666 4- 0.020
Uplift rate (ram/year)
Blanc & Cardini (1961); Damiani (1970); and references therein
0.02
Damiani (1970); Carobene et al. (1986); Carboni et al. (1988); Cosentino & Gliozzi (1988); Carobene & Dai Pra (1991)
0.02
5e
Damiani (1970) and references therein
0.03
5e
Carobene et al. (1986)
0.05
Pata (1947); Dumas et al. (1991); Dai Pra et al. (1993); Miyauchi et al. (1994)
0.35
Dumas et al. (1987); Dumas & Raffy (1993)
0.66
Miyauchi et al. (1994); Valensise & D'Addezio (1994)
0.91
Bonfiglio et al. (1988); Miyauchi et al. (1994)
0.59
Lmk Strombus bubonius
Cosentino & Gliozzi (1988)
0.47
Valensise & Pantosti (1992)
1.15
Ascenzi et al. (1971); Bonfiglio & Berdar (1986); Hearty (1986)
0.72
0.75
5e
+7875 122.393_734x and
Reference(s)
81
5e
Cladocora; Spondylus
Strombus bubonius Link; Cladocora caespitosa
Strombus bubonius
Lmk
133-117
Aminozone E (A/I = 0.39.05)
135-117
Aminozone E 5e (A/I = 0.45 4- 0.01)
5e
Elephas cfr. antiquus Strombus bubonius
Bonfiglio (1972); Hearty (1986); Hearty
Lmk
et al. (1986b)
Strombus bubonius
Gignoux (1913); Bonfiglio (1972); Hearty (1986); Hearty et al. (1986b)
0.95
5e Glycymeris; or older Strombus bubonius Lmk
Dumas et al. (1987b, 1988)
1.21
120
5e
Dumas & Raffy (1996)
1.21
114
5e or 5d
Dumas et al. (1996)
0.64
135-117
Aminozone E 5e (A/I = 0.42 + 0.04) Aminozone E (A/I= 0.44 + 0.2) (A/I =0.55)
Lmk
82
P. B O R D O N I & G. V A L E N S I S E
Table 1.
(continued)
Site Locality no.
Latitude (~
Longitude (~
66
Between Fiumara Condoianni and Vallone Notaro, near Ardore Marina
38.188
16.194
85
Inner edge
85
Geomorphological correlation
67
Between Fiumara Condoianni and Vallone Notaro, near Ardore Marina
38.188
16.200
87
Inner edge
87
Geomorphological correlation
68
Fiumara Portigliola, east of Portigliola
38.213
16.225
85
Inner edge
85
Geomorphological correlation
69
Fiumara Portigliola, east of Portigliola
38.215
16.221
92
Inner edge
92
Geomorphological correlation
70
Between Fiumara Portigliola and Vallone Castellace
38.216
16.228
85
Inner edge
85
Geomorphological correlation
71
Between Fiumara Portigliola and Vallone Castellace
38.218
16.226
90
Inner edge
90
Geomorphological correlation
72
West of Soverato
38.690
16.543
113
Inner edge
113
Geomorphological correlation
73
West of Soverato
38.691
16.544
104
Inner edge
104
Geomorphological correlation
74
West of Soverato
38.692
16.456
99
Inner edge
99
Geomorphological correlation
75
Crotone Peninsula
38.960
17.140
84
Terrace surface
84
Palaeontology and geomorphological correlation
76
Crotone Peninsula
38.960
17.100
Terrace surface
110
Palaeontology U-Th and aminostratigraphy
77
Left bank of Coscile River
39.760
16.250
120
Inner edge
120
Geomorphological correlation
78
Right bank of Eiano River, near Lauropoli
39.750
16.330
115
Inner edge
115
Geomorphological correlation
79
Torrente Raganello
39.785
16.363
135
Inner edge
135
Geomorphological correlation
80
Villapiana
39.830
16.450
135
Inner edge
135
Geomorphological correlation
81
Trebisacce
39.874
16.525
135
Inner edge
135
Geomorphological correlation
Max. elevationof deposits (m asl)
Shoreline Geomorphelevation o l o g i c a l (m a s l ) evidence
50-100
Conventional Dating method elevationof Tyrrhenian shoreline
DEFORMATION
Age (ka)
Aminozone U-Th ratio
O F T H E 125ka M A R I N E T E R R A C E I N I T A L Y
Oxygen isotope substage
Significantfossil assemblage
83
Reference(s)
Uplift rate (mm/year)
114
5e or 5d
Dumas et al. (1995)
0.63
114
5e or 5d
Dumas et aL (1995)
0.65
114
5e or 5d
Dumas et al. (1995)
0.63
120
5e
Dumas et al. (1995)
0.69
114
5e or 5d
Dumas et al. (1995)
0.63
120
5e
Dumas et al. (1995)
0.67
130
5e
Dumas & Raffy (1996)
0.86
Dumas & Raffy (1996)
0.78
Dumas & Raffy (1996)
0.74
Palmentola et al. (1990) and references therein
0.62
Gliozzi (1988) and references therein; Cosentino & Gliozzi (1988)
0.83
5a-c
Cucci & Cinti (1998)
0.91
5a-c
Cucci & Cinti (1998)
0.87
5a-c
Cucci & Cinti (1998)
1.03
5a-c
Cucci & Cinti (1998)
1.03
5a-c
Cucci & Cinti (1998)
1.03
120 114
5e or 5d 5e
Strombus bubonius
Lmk 123
Aminozone E (A/I = 0.33)
5e
Strombus bubonius Link; Conus ermineus
84
P. B O R D O N I & G. V A L E N S I S E
Table 1.
(continued) Shoreline elevation (m a s l )
Geomorphological evidence
Conventional Dating method elevation of Tyrrhenian shoreline
16.567
105
Inner edge
105
Geomorphological correlation
39.957
16.605
95
Inner edge
95
Geomorphological correlation
Capo Spulico
39.960
16.600
125
Inner edge
125
Geomorphological correlation
85
Piano San Nicola, near Nova Siri
40.133
16.591
115
Inner edge
115
Geomorphological correlation
86
Piano San Nicola, near Nova Siri
40.130
16.598
85
Inner edge
85
Aminostratigraphy
87
Right bank of Sinni River
40.161
16.609
80
Inner edge
80
Geomorphological correlation
88
Left bank of Sinni River
40.204
16.628
70
Inner edge
70
Geomorphological correlation
89
Right bank of Agri River
40.223
16.641
70
Inner edge
70
Geomorphological correlation
90
Left bank of Agri River
40.260
16.676
60
Inner edge
60
Geomorphological correlation
91
Right bank of Cavone River
40.285
16.694
60
Inner edge
60
Geomorphological correlation
92
Left bank of Cavone River
40.335
16.725
50
Inner edge
50
Geomorphological correlation
93
Ponte del Re, left bank of Lato River
40.525
16.954
45; 35
Inner edge
45
Palaeontology and aminostratigraphy
94
Santa Teresiola di Galbe, northwest of Taranto
40.500
17.256
28
Inner edge
28
U-Th
95
Masseria San Pietro, northeast of Taranto
40.503
17.320
26
28
Inner edge
28
Palaeontology and aminostratigraphy
96
I1 Fronte, east of Taranto
40.472
17.317
10
Inner edge
10
Palaeontology, aminostratigraphy and U - T h
97
Torre Castelluccia, southwest of Lizzano
40.383
17.466
Site Locality no.
Latitude (~
Longitude (~
82
Right bank of Avena River, south of Amendolara
39.912
83
Capo Spulico
84
Max. elevation of deposits (m asl)
43
24
28
Inner edge
Palaeontology and aminostratigraphy
D E F O R M A T I O N OF T H E 125ka M A R I N E T E R R A C E IN ITALY
Age (ka)
Aminozone U-Th ratio
Reference(s)
Uplift rate (mm/year)
5a-c
Cucci & Cinti (1998)
0.79
5a-c
Cucci & Cinti (1998)
0.71
5e
Cucci & Cinti (1998)
0.95
5e
Cucci & Cinti (1998)
0.87
Moncharmant Zei (1957); Neboit & Re• (1973); Dai Pra & Hearty (1988); Hearty & Dai Pra (1992)
0.63
5e
Dai Pra & Hearty (1988)
0.59
5e
Dai Pra & Hearty (1988)
0.51
5e
Dai Pra & Hearty (1988)
0.51
5e
Dai Pra & Hearty (1988)
0.43
5e
Dai Pra & Hearty (1988)
0.43
5e
Dai Pra & Hearty (1988)
0.35
Oxygen isotope substage
Significantfossil assemblage
Aminozone C 5a-c (A/I = 0.29 4- 0.02)
i22
Aminozone E 5e (A/I = 0.43 4- 0.03)
i25•
i 35-117
85
5e
Strombus bubonius Lmk; Glycymeris
Boenzi et al. (1985); Dai Pra & Hearty 0.31 (1988)
Cladocora caespitosa
Cotecchia et al. (1969); Cotecchia et al. 0.18 (1971); Dai Pra & Stearns (1977); Hearty & Dai Pra (1985); Dai Pra & Hearty (1988)
Aminozone E 5e (A/I = 0.38 • 0.01)
Cotecchia et al. (1969); Dai Pra & 0.18 Stearns (1977); Dai Pra (1982); Hearty & Dai Pra (1985); Hearty (1986)
~1) 122• Aminozone E 5e ~2) [117-121- (A/I = 0.36 • 0.02) 128] • 7; ~3) 125 • 6
Strombus bubonius Lmk; Cladocora caespitosa; Glycymeris
Gignoux (1913); Cotecchia et al. 0.03 (1969); Dai Pra & Stearns (1977); (1) Hearty & Dai Pra (1985); Hearty (1986); (2) Hearty et al. (1986b); Dai Pra & Hearty (1988); (3) Hearty & Dai Pra (1992)
i35-117
Strombus bubonius Lmk; Glycymeris
Cotecchia et al. (1969); Dai Pra & 0.18 Stearns (1977); Dai Pra (1982); Hearty (1986); Dai Pra & Hearty (1988); Hearty & Dai Pra (1992)
Aminozone E (A/I=0.39)
5e
86
P. B O R D O N I & G. V A L E N S I S E
Table 1.
(continued)
Site Locality
Latitude (~
Longitude (~
Torre Colimena, south of Avetrana
40.290
17.760
Torre Castiglione, near Porto Cesareo
40.280
17.830
100
Ospedale di Gallipoli
40.058
17.996
10
101
Torre San Giovanni, south of Gallipoli
40.051
10.001
102
Grotta Romanelli, near Castro
40.010
18.440
103
Between Torre Mattarelle and Torre Guaceto, near Brindisi
40.680
17.880
104
Between Villanova and Monopoli
40.810
17.450
30
105
Between Villanova and Monopoli
40.860
17.380
106
Fortore floodplain, near Ripalta
41.896
15.299
107
Capo Peloro, northeast of Messina
38.266
15.632
85
108
Capo Milazzo
38.227
15.248
60
109
Capo Tindari, east of Patti
38.130
15.040
85
110
Acquedolci
38.060
14.580
130
111
Capo Zafferano, north of Bagheria
38.100
13.530
6
112
Port of Palermo
38.100
13.500
10
113
Sferracavallo
38.090
114
Trapani
38.020
no.
99
Max. elevation of deposits (m asl)
Shoreline elevation (m a s l )
Geomorphological evidence
Conventional Dating method elevation of Tyrrhenian shoreline 20
Geomorphological correlation
2
Palaeontology and U-Th
Inner edge
10
Palaeontology and aminostratigraphy
8
Inner edge
8
Palaeontology and aminostratigraphy
2-8
Sea cave
8
Palaeontology
4
Palaeontology and radiocarbon
Inner edge
30
Geomorphological correlation
30
Inner edge
30
Geomorphological correlation
25
Inner edge
25
Geologicalstratigraphy
130
Inner edge
130
Palaeontology and aminostratigraphy
60
Palaeontology and aminostratigraphy
85
Palaeontology
130
Palaeontology
20
0.9
Inner edge
2
10
Terrace surface
Inner edge
7
Aminostratigraphy
10
Palaeontology
13.460
8
Palaeontology
12.540
4
Palaeontology
7
Abrasion notch
D E F O R M A T I O N OF T H E 125ka M A R I N E T E R R A C E IN ITALY
Age (ka)
Aminozone U-Th ratio
Oxygen isotope substage
Significantfossil assemblage
5e
156 4- 20
23~ = 0.78 + 0.04
5e
Reference(s)
Dai Pra (1982)
Strombus bubonius
Lmk; Conus,
Dai Pra (1982); Dai Pra & Hearty (1988); Hearty & Dai Pra (1992)
87
Uplift rate (ram/year)
0.11
-0.03
Cladocora caespitosa
135-117
Aminozone E 5e (A/I = 0.41 4- 0.04)
Strombus bubonius Lmk; Glycymeris
Cotecchia et al. (1969, 1971); Hearty & Dai Pra (1985, 1992); Hearty (1986); Cosentino & Gliozzi (1988); Dai Pra & Hearty (1988)
0.03
125
Aminozone E 5e (A/I = 0.38 4- 0.02)
Strombus bubonius Lmk; Glycymeris
Cotecchia et al. (1969); Hearty & Dai Pra (1985, 1992); Hearty (1986); Dai Pra & Hearty (1988)
0.02
Cosentino & Gliozzi (1988); Di Stefano
0.02
5e
et al. (1992) and references therein
(generic)
Di Geronimo (1970); Mastronuzzi & Sans6 (1998) and references therein
5 (generic)
Di Geronimo (1970); Pieri (1975); Iannone & Pieri (1979)
0.19
5 (generic)
Di Geronimo (1970); Pieri (1975); lannone & Pieri (1979)
0.19
5 (generic) or 1
Mastronuzzi et al. (1989)
0.15
Bonfiglio (1981, 1991) and references therein; Bonfiglio & Violanti (1983); Hearty et al. (1986b);
0.99
5
:Dlder than Holocene Jimatic 3ptimum
135-117
Helix
-0.02
Aminozone E (A/I =0.41)
Strombus bubonius
Aminozone E (A/I = 0.41 4- 0.03)
Senegalese fauna
Belluomini (1985); Hearty (1986); Hearty et al. (1986b)
0.43
Stromubs bubonius
Gliozzi & Malatesta (1982); Cosentino & Gliozzi (1988)
0.63
Bonfiglio (1981, 1991) and references therein
0.99
Lmk Arca
Antonioli et al. (1994a)
0.01
Strombus bubonius
Fabiani (1941); Cosentino & Gliozzi (1988)
0.03
Gignoux (1913); Cosentino & Gliozzi (1988) and references therein
0.02
Lmk Strombus bubonius
Ruggieri et al. (1968)
Lmk
Lmk Strombus bubonius
5e
Lmk 5e 5e
Strombus bubonius
Lmk
-0.02
88
P. B O R D O N I & G. V A L E N S I S E
Table 1.
(continued) Max. elevation of deposits (m asl)
Shoreline elevation (m a s l )
Geomorphological evidence
Conventional Dating method elevation of Tyrrhenian shoreline
Site Locality no.
Latitude (~
Longitude (~
115
Torre Scibiliana, southeast of Marsala
37.710
12.520
116
Mazara del Vallo
37.650
12.630
117
Pozallo-Marina di Ragusa
36.720
14.750
15
15
Geomorphological correlation
118
Pachino
36.728
15.122
15
15
Palaeontology
119
Augusta
37.250
15.230
30
Palaeontology and geomorphological correlation
120
Between Punta Bonico and Santa Croce Lighthouse, north of Brucoli
37.296
15.211
34
34
Palaeontology and geomorphological correlation
121
Taormina
37.850
15.270
130
130
Geomorphological correlation
3
8.5
30
Abrasion notch
3
Palaeontology
4
Palaeontology
Fig. 3. Distribution of observations of Tyrrhenian marine terraces for Tuscany and Latium (see Fig. 2 for further details).
D E F O R M A T I O N OF T H E 125ka M A R I N E T E R R A C E IN ITALY
.ge (ka)
Aminozone U-Th ratio
Oxygen isotope substage
5e
Significantfossil assemblage
Strombus bubonius
Lmk 5 5e
Strombus bubonius
89
Reference(s)
Uplift rate (mm/year)
D'Angelo & Vernuccio (1994)
-0.02
Ruggieri & Unti (1974); D'Angelo & -002 Vernuccio (1994) and references therein Carbone et al. (1982)
0.11
Malatesta (1985)
0.07
Cosentino & Gliozzi (1988)
0.19
Di Grande & Scamarda (1973); Carbone et al. (1982); Bonfiglio (1991)
0.22
Bonfiglio (1981, 1991) and references therein
0.99
Lmk 5e
Strombus bubonius
Lmk 5e
Strombus bubonius
Lmk
5e
Fig. 4. Distribution of observations of Tyrrhenian marine terraces for Campania, Basilicata and part of Calabria (see Fig. 2 for further details).
90
P. BORDONI & G. VALENSISE
Fig. 5. Distribution of observations of Tyrrhenian marine terraces for Calabria and Sicily (see Fig. 2 for further details).
in the text could be marked in the relevant figures. All values in metres are elevations above present sea level. It should be noted that largely discordant observations or opinions on the identification of a specific palaeo-shoreline or on the age of a well-established terrace (for instance, when the same shoreline was referred to a different substage of the oxygen-isotope stage 5) were handled by adding new records to Table 1.
Tyrrhenian coast, f r o m northern Tuscany to the Garigliano Plain Our list opens with Buca dei Corvi near Livorno (Fig. 3, site 1), where Tyrrhenian marine deposits crop out at roughly 11 m elevation (Blanc 1953a; Hearty et al. 1986b; Hearty & Dai Pra 1987). Between this site and the north bank of the Fiora River (sites 2-5) Tyrrhenian marine deposits crop out at a maximum elevation of 12 m (Hearty & Dai Pra 1987). As none of the investigators described the actual inner edge, we will use 12 m
as the conventional elevation of the Tyrrhenian shoreline at this site. The Tarquinia Plain is located between the Fiora River and the Mignone River (sites 6-11); this uplifted coastal plain does not exhibit abrupt slope changes, and investigators have commonly placed the Tyrrhenian inner edge somewhere between 20 and 65 m (Gignoux 1913; Blanc 1936; Bonadonna 1967a; Dai Pra 1978; Ambrosetti et al. 1981; Hearty & Dai Pra 1986, 1987; Radtke 1986; Palieri & Sposato 1988). Our preferred elevation of the Tyrrhenian shoreline is 30-45 m, based on absolute dating by Hearty & Dai Pra (1986) and Radtke (1986); this is in reasonable agreement with the morpho-lithostratigraphic reconstruction proposed by Ambrosetti et al. (1981) and Palieri & Sposato (1988), who reported a shoreline culminating at 27 m. It should be noted that the interpretations of Blanc (1936) and Dai Pra (1978), who placed the Tyrrhenian shoreline at 65 m, have not been reported in Table 1, as the associated deposits have been assigned to oxygenisotope stage 7 by Radtke (1986), on the basis on U - T h dating. Similarly, the interpretations of
DEFORMATION OF THE 125ka MARINE TERRACE IN ITALY
91
Fig. 6. Distribution of observations of Tyrrhenian marine terraces for Basilicata and Apulia (see Fig. 2 for further details).
Gignoux (1913) and Bonadonna (1967a) for site 6 will not be used for subsequent analyses. Between the Mignone River and the town of Santa Severa (sites 12-14) Dai Pra (1978) and Hearty & Dai Pra (1986, 1987) described shorelines that are precisely assigned as Tyrrhenian through aminostratigraphic and palaeontological correlation (Hearty & Dai Pra 1987). Maximum elevation is 45 m. South of this area and as far south as the north bank of the Tiber River (sites 15-17), the Tyrrhenian shoreline is traditionally placed at elevations up to 35 m (Blanc 1936; Bonadonna 1967b; Dai Pra 1978; and references therein). This view was further supported by Hearty (1986) and Hearty & Dai Pra (1986, 1987), on the basis of aminostratigraphic dating of deposits from the Cerveteri-Palidoro area (sites no. 15, 16). South of the Tiber River, near Pomezia (site 18), the Tyrrhenian shoreline is seen to climb to 45 m (Dai Pra & ArnoldusHuyzendveld 1984; Malatesta & Zarlenga 1986; Milli & Zarlenga 1991), and it again drops to 20 m at Fosso della Moletta, a few kilometres southeast
of Ardea (site 19) (Dai Pra & Arnoldus-Huyzendveld 1984), and to 12 m at Borgo S. Maria, about 8 km west of Latina (Hearty & Dai Pra 1986) (site 20). The section of the coast that runs from Borgo Sabotino, to the south of Latina (site 21), to the Garigliano Plain (site 30) passing through the Mt. Circeo (site 22), Terracina (site 23) and Gaeta (site 25-28) rocky promontories, is characterized by relative stability. Various investigators consistently reported Tyrrhenian shorelines in the range 5-10m on the basis of the occurrence of S. bubonius, aminostratigraphic correlation and the presence of characteristic notches in steep limestone cliffs (Blanc 1935, 1953b; Blanc & Segre 1947; Segre 1949, 1957; Dai Pra & ArnoldusHuyzendveld 1984; Hearty & Dai Pra 1986; Ozer et al. 1987; Antonioti et al. 1988; Brancaccio et al. 1990; Antonioli 1991; Dai Pra 1995). Limited differential motions, however, have been reported by Ozer et al. (1987) and Antonioli (1991), for example between Mt. Circeo (site 22, at 9.6m) and the Gaeta Promontory (site 28, at 4.8m).
92
P. B O R D O N I & G. V A L E N S I S E
Fig. 7. Simplified contour map of the coast between Piano San Nicola and Torrente Avena, near Amendolara (see also Fig. 6), showing southward increase in the elevation of the Tyrrhenian inner edge with a gradient of 0.8 m/km. The map is also suggestive of the transition from erosion to aggradation of the same terrace surface from south to north in response to a large increase in sediment supply.
DEFORMATION OF THE 125ka MARINE TERRACE IN ITALY Along the same line, Antonioli et al. (1988) pointed out the existence of Tyrrhenian deposits at - 6 m in the Fondi Plain (site 24).
Tyrrhenian coast, f r o m the C a m p a n a Plain to Cetraro, northern Calabria
South of the Garigliano Plain is the Campana Plain, which is divided into Volturno Plain and Sarno Plain by the Vesuvius volcano (Fig. 4). Romano et al. (1994) described beach deposits at 50 m elevation overlying a planation surface carved in Mesozoic limestones near Cancello, at the foothills of the southern Apennines about 8km southeast of Caserta (site 31). The same investigators described U-Th-dated Tyrrhenian deposits at - 5 0 m from a well drilled in the Volturno Plain (site 32); the top of these deposits was described at elevations between - 2 0 and - 9 0 m depending on the site. On the basis of sedimentological and palaeoecological evidence, Romano et al. (1994) proposed a correlation of the Cancello deposits with those found in the subsurface. This implies drastic tectonic control at the northeastern edge of the Campana Plain causing strong subsidence of the plain itself relative to the Apennines. Cinque et al. (1987) and Barra et al. (1991) also reported marine deposits presumably of Tyrrhenian age at - 2 9 m near the south bank of the Sarno river (site 33). To the south of the Campana Plain is the Sorrento Peninsula, where many researchers (e.g. Brancaccio 1968; Colantoni 1970; Brancaccio et al. 1978; Cosentino & Gliozzi 1988; Cinque & Romano 1990) described notches at 7 - 8 m (sites 34-37). The notches are positively assigned a Tyrrhenian age by Brancaccio et al. (1978) through U-Th dating. In the Sele Plain Brancaccio et al. (1986, 1987) and Russo & Belluomini (1992) described Tyrrhenian marine deposits at a maximum elevation of 25 m (site 38) beneath dune ridges at about 4 km from the present shoreline. Along the Cilento promontory the Tyrrhenian shoreline again approaches present sea level, although considerable disagreement still exists on its exact elevation. Brancaccio et al. (1990) and Cinque et al. (1994) placed it at 8-10m (sites 39-43) on the basis of geomorphological correlation with a sample assigned to oxygenisotope substage 5a-c by Brancaccio et al. (1990) using aminostratigraphic correlation. In contrast, Antonioli et al. (1994b) proposed an elevation of 2.2 m (site 44) on the basis of detailed geomorphological evidence. In summary, all the estimates suggest stability of this area during the Late Pleistocene time, in agreement with the large-
93
scale geomorphological evidence and with the absence of raised young marine deposits. At Sapri (site 45) Brancaccio et al. (1990) placed the Tyrrhenian shoreline at 15 m on the basis of geomorphic correlation with a site dated at substage 5a-c by aminostratigraphy. Carobene & Dai Pra (1991) described Tyrrhenian deposits sampled on an abrasion platform with inner edge at 6.9 m near Acquafredda, a few kilometres south of the Basilicata border (site 46). The deposits are dated by aminostratigraphic correlation. According to the same investigators, the abrasion platform rises slowly to over 9 m at Marina di Maratea (site 47). The Tyrrhenian shoreline stays at about the same elevation through the Scalea-Praia a Mare area (sites 48-51), for which no absolute or aminostratigraphy dating is available (Blanc & Cardini 1961; Damiani 1970, and references therein; Carobene et al. 1986, 1988; Cosentino & Gliozzi 1988; Carobene & Dai Pra 1991). Between Cirella and Diamante (site 52) the Tyrrhenian shoreline is found at a similar elevation. Damiani (1970) placed it at 8-10m, on the basis of the presence of volcanic rocks dated by Lirer et al. (1967) near Scalea. According to Carobene et al. (1986), marine deposits found in a sea-cave at 12 m near Punta La Testa-Cetraro (site 53) may represent evidence for the Tyrrhenian shoreline; the hypothesis is based on a morphometric reconstruction of the whole local sequence of marine terraces, one of which is assigned an oxygen-isotope stage 9 age by U-Th dating.
Tyrrhenian coast, f r o m the Savuto River to Scilla, southern Calabria
No specific study on the morphology and chronology of Late Pleistocene terraces could be found in current literature for the stretch of the Calabrian coast between Guardia Piemontese and Vibo Marina (Fig. 5). This is especially unfortunate in view of the evidence that this is the region where the Calabrian Arc joins the rest of the peninsula (Fig. 1) and where Late Pleistocene uplift picks up again after the stability suggested by observations from coastal Cilento and northern Calabria. For the region between Guardia Piemontese and San Lucido (Fig. 5), for instance, Cortese (1895) reported two marine terraces with inner edges at 200 and 300m respectively, which were generically ascribed to the Pleistocene period. For the region south of the mouth of the Savuto River the same investigator described four marine terraces, the lowest of which had its inner edge at about 50 m elevation.
94
P. BORDONI & G. VALENSISE
According to Cortese, this particular terrace surface can be followed southward along the Gulf of Sant'Eufemia, along the Capo Vaticano-Monte Poro promontory, along the western side of the Gioia Tauro Plain and finally along the foothills of the Aspromonte Mountain range, where it is reported to climb to an elevation of 120 m. Even though this terrace is most probably Tyrrhenian in age, Cortese did not mention any fauna that can be used to back up this hypothesis. Similarly, Gu6r6my (1980) described in detail seven marine terraces along the 'Stretta di Catanzaro' (the isthmus located between the S. Eufemia and Squillace gulfs, Fig. 5), the lowest of which had its inner edge between 100 and 124 m but provided no absolute date for any of the terrace surfaces. Neither Cortese's nor Gu@6my's observations have been included in Table 1 or in Figs 2 and 5 but will be considered qualitatively. Starting from an early description of S. bubonius at 38 m elevation made by Pata (1947) at Vibo Valentia Marina (site 54), Dumas et al. (1991) and Dai Pra et al. (1993) dated a sample of the coral Cladocora caespitosa at about 125 ka using U-Th, and placed the associated shoreline at about 50 m. In reconstructing the Tyrrhenian shoreline from Vibo Valentia Marina westward around the Monte Poro promontory, Dumas et al. (1987a) and Dumas & Raffy (1993) reported an elevation slowly climbing from 51 to 88 m, essentially extrapolating up to Capo Vaticano (site 55) the original observation by Pata (1947). Dai Pra et al. (1993) and Miyauchi et al. (1994) also attempted the same reconstruction, but they came to rather different conclusions as they saw the Tyrrhenian shoreline climb to 120 m at Capo Vaticano (site 56) and then drop again to about 80 m along the southern side of the Monte Poro promontory. The latter observation is based on the presence of S. bubonius in deposits associated with an inner edge at 80m reported by Bonfiglio et al. (1988) at Contrada Ianni, near Nicotera (site 57). In marked contrast with all the work described so far, Barrier et al. (1988) assigned to the Tyrrhenian stage marine deposits with Senegalese fauna (that is, a warm fauna) correlated with an inner edge at 287 m observed at Coccorino, about 4 km southeast of Capo Vaticano. According to Dai Pra et al. (1993) and Miyauchi et al. (1994), however, these deposits cannot be of Tyrrhenian age, as they lack S. bubonius, but are nevertheless indicative of a climatic stage as warm as the Tyrrhenian. On these grounds, Dai Pra et al. (1993) and Miyauchi et al. (1994), proposed their correlation with the oxygen-isotope stage 9 (about 330 ka), a conclusion that is also supported by the extrapolation to higher terraces of the rate of uplift inferred from the Tyrrhenian inner edge.
The amount of available information drops again as one enters the lowlands of the Gioia Tauro Plain, a large basin filled with Pleistocene marine sediments. Cosentino & Gliozzi (1988) reported a finding of S. bubonius at 65 m near the town of Gioia Tauro (site 58), in good agreement with the elevation of the Tyrrhenian inner edge at Contrada Ianni (site 57). Miyauchi et al. (1994) and Valensise & D'Addezio (1994) mapped the marine terraces within the Gioia Tauro Plain using geomorphological criteria and dated them by extrapolation of the Gioia Tauro Tyrrhenian site. These two independent groups of investigators described the Tyrrhenian shoreline climbing to 100m at Lido di Palmi before it disappears because of the steepness of the coast. According to Valensise & Pantosti's (1992) geomorphological reconstruction, the Tyrrhenian shoreline again appears for a very short section near Scilla at about 150 m (site 59).
I o n i a n coast, f r o m the M e s s i n a S t r a i t s to C r o t o n e
The highest Tyrrhenian deposits known in the whole Italian peninsula are those that crop out on the western side of Aspromonte facing the Messina Straits (Fig. 5). These deposits have been the object of intense investigations for over a century by generations of geologists. S. bubonius deposits are described at 100m elevation at Ravagnese-Gallina (site 61) (Bonfiglio, 1972), 125 m at Bovetto (site 62) (Gignoux, 1913; Bonfiglio 1972), and 157m at Nocella (site 63) (Dumas et al. 1987b). Deposits at 96m near San Francesco di Archi (site 60) were also assigned a Tyrrhenian age by Ascenzi & Segre (1971) and Bonfiglio & Berdar (1986), respectively based on the presence of a human mandible and of remains of Elephas antiquus cfr. In the framework of an inventory of Late Pleistocene deposits from various Mediterranean countries, Hearty (1986) and Hearty et al. (1986b) presented aminostratigraphic correlations supporting a Tyrrhenian age for all of these sites. Dumas et al. (1988), however, suggested that the general trend of the Tyrrhenian shoreline near the southern end of the Straits may not be as unequivocal as it may appear from palaeontological evidence alone. In particular, they described S. bubonius deposits associated with independent shorelines at 105 and 157 m elevation near Nocella (site 63), and on the basis of aminostratigraphic evidence concluded that the terrace surface culminating at 157 m could have been carved either during the oxygen-isotope substage 5e highstand or during a previous highstand. The same investigators also assigned
DEFORMATION OF THE 125ka MARINE TERRACE IN ITALY to substage 5e deposits culminating with an inner edge at 118 m near Lazzfiro (site 64), but these deposits were later assigned to oxygen-isotope substage 5c by Dumas & Raffy (1996). Current literature offers at least three attempts at lateral reconstruction of the Tyrrhenian shoreline within the Messina Straits based on the above evidence and on geomorphological correlation between adjacent sites. Owing to the strong variability of the inner edge elevation and to the dissection of the terrace surfaces by high-energy streams cutting through the Aspromonte block, however, the three reconstructions share the same range of elevations but are rather different in shape. Dumas et al. (1988) focused specifically on the fluctuations that occur at the southern end of the Straits. In particular, they proposed a correlation of the 105 m shoreline observed around Nocella (site 63) with the similarly high historic Tyrrhenian sites of Ravagnese and Bovetto (sites no. 61 and 62), but were left with the problem of explaining the existence of the much higher (157 m) Tyrrhenian shoreline at Nocella (site 63). To address this question, Dumas et al. (1988) put forward three alternative hypotheses. The first assumes the occurrence of two significant highstands within oxygen-isotope substage 5e, each of which could have carved an individual marine platform (one at 105-125 m the other at 157 m); this hypothesis, however, was rejected by the investigators themselves as the vertical spacing between terraces supposedly created by such highstands would be larger than that between terraces associated with later highstands. The second hypothesis assumes that the 157m Nocella terrace is in fact older than oxygenisotope stage 5e, as is in part suggested by one the samples dated by aminostratigraphy. The third hypothesis simply assumes that there might be a fault perpendicular to the modern coast producing relative dislocation of the Nocella terrace relative to the Bovetto site. Later work by the same group of investigators and based on evidence from the Straits area (Dumas & Raffy 1993) proposes a new model for Mediterranean palaeo-sea levels where oxygen-isotope substage 5e is envisioned as formed by two secondary highstands at 130 and 120 ka, respectively, similarly to what was proposed for the Papua New Guinea sea-level history by Chappell & Shackleton (1986). With the improved resolution offered by this model, Dumas & Raffy (1996) proposed a new aminostratigraphic correlation with the 120ka highstand for a terrace culminating at 141-146m near Lazz~iro and with the 130ka highstand for the 157 m Nocella terrace. Dumas & Raffy (1996) also assigned to oxygen-isotope substage 5c the 105 m terrace observed at Nocella
95
and previously attri-buted to the Tyrrhenian stage by Dumas et al. (1988). This leaves the third of the above hypotheses as the only option available to Dumas et al. (1988) to justify the newly confirmed short-scale variability in elevation of the Tyrrhenian palaeo-shorelines. In marked contrast to the previous hypotheses, Valensise & Pantosti (1992) explained the reported fluctuations entirely as the result of the activity of the large seismogenic normal fault responsible for the catastrophic Messina Straits earthquake of 28 December 1908. The elevation of the terrace would record a combination of localized subsidence associated with this fault and uplift of the whole Calabrian block. Valensise & Pantosti reconstructed a Tyrrhenian palaeoshoreline culminating at about 150 m both off the northern end of the Straits (Scilla, site 59) and at the southern end (Lazzfiro, site 64), reaching a minimum elevation of less than 100m around Reggio Calabria. According to these investigators, the lowest sites coincide with areas where large fan-deltas fed by sediments supplied from the Aspromonte range push the shoreline closer to the Messina Straits fault. Miyauchi et al. (1994) also reconstructed a Tyrrhenian palaeo-shoreline at about 150 m near Scilla, about 100m near Villa San Giovanni, about 140m east of Reggio Calabria, and at 127 m near Bovetto. Although these observations are in fairly good agreement with the reconstruction by Valensise & Pantosti (1992) (for this reason they have not been reported in Table 1 or in the figures), Miyauchi et al. took no stand as to what causes such strong fluctuations in elevation within a relatively short distance. The Ionian side of Calabria between Capo dell'Armi and the Crotone promontory is traditionally far less studied than the Tyrrhenian side. For this region Cortese (1895) described a single terrace surface reaching a maximum elevation of 170 m and extending as far north as the Squillace Gulf. Exactly a century later, Dumas et al. (1995) for the first time investigated in detail the coastal area around Locri and reconstructed a complex flight of marine terraces. The chronological sequence of this terrace flight was entirely based on a correlation with the nearby Messina Straits and with Dumas & Raffy's (1993) model for Mediterranean palaeosea levels. Dumas et al. (1995) extended this model by proposing the existence of an additional highstand at 114 ka, which in the oxygenisotope record would be designated 5d. Within this framework they described marine terrace inner edges in the elevation range 90-92 m (sites 69 and 71) and 85-87 m (sites 65-68 and 70) and assigned them to oxygen-isotope substages 5e
96
P. BORDONI & G. VALENSISE
and 5d, respectively. Dumas & Raffy (1996) used the same reasoning in a study of the Soverato area (sites 72 and 73), where they identified two marine terraces at 113 m and 104m and dated them at 130 ka and 120 ka, respectively. They also described a marine terrace at 99 m (site 74) which they assigned to oxygen-isotope substage 5d (114 ka). Several investigators (e.g. Gubr~my 1972; Selli 1977) proposed reconstructions of the marine terrace sequence of the Crotone promontory, an area that for its wealth of Late Pleistocene deposits and for the clarity of its landforms attracted early workers such as Cortese (1895) and Gignoux (1913). The most recent comprehensive studies are those by Gliozzi (1988) and Palmentola et al. (1990). Gliozzi (1988) placed the Tyrrhenian inner edge at about l l 0 m (site 76) on the basis of a wide range of evidence including the occurrence of S. bubonius, U-Th dating and aminostratigraphy. For this terrace surface the compilation by Cosentino & Gliozzi (1988) lists an elevation of 80-100m (corresponding to an uplift rate of 0.85mm/year; see Fig. 1), but the elevation of its inner edge is not specified. Somewhat in contrast, Palmentola et al. (1990) reported displaced terraces assigned to oxygen-isotope substage 5e in the elevation range 15-84m (site 75), and stressed the role of recent faulting in disrupting the continuity of terrace surfaces in this region.
Ionian coast, f r o m the Calabria-Basilicata border to Taranto Very little is known about the section of coast between the Crotone promontory and the Sibari Plain, which rather sharply separates the Calabrian Arc from the block corresponding to the southern Apennines (Fig. 1). For this area Sestini (1930) reported three separate terrace surfaces at 60-80, 115-145 and 150-200 m elevation, the lowest of which was tentatively assigned to the Tyrrhenian highstand by Vezzani (1967). In view of the uncertainty of this assignment and of the lack of any geochronological markers, we did not include this observation in our database. In contrast, the region to the north of Sibari has been the object of intense investigations for over 40 years (see, e.g. Moncharmant-Zei 1957; Cotecchia & Magri 1967; Vezzani 1967; Brtickner 1980; Parea 1986, and references therein), using both a stratigraphic or a purely geomorpholoical approach (see Fig. 4). All these investigators consistently described a significant drop in elevation of all the identified marine terraces moving from the northern end of the Sibari Plain toward
the northeast and across the Calabria-Basilicata border (see also Fig. 7). Apart from the stratigraphic evidence, the age assignment of the terraces seen in this area rests exclusively on the results obtained by Dai Pra & Hearty (1988) for Piano San Nicola (site 86) and Ponte del Re (site 93) through aminostratigraphic correlation. In particular, the site of Piano San Nicola was the starting point of geomorphological reconstructions of the whole sequence by Brfickner (1980), Dai Pra & Hearty (1988) and Cucci & Cinti (1998). As the age of the marine terraces at this site is crucial for all subsequent analyses, we decided to describe in some detail the basic information available to the different investigators and any differences in their interpretation and conclusions. The first attempts to constrain terrace ages in the Piano San Nicola area are due to Moncharmant-Zei (1957) and Neboit & Reinard (1973), who reported marine deposits 'more recent than Eutyrrhenian', and hence presumably Neotyrrhenian, on the basis of biostratigraphic evidence. The same investigators placed the corresponding inner edge at an elevation of about 60 m in consideration of the palaeoecology of the fossil assemblage. Dai Pra & Hearty (1988) sampled the same deposits at 43m elevation (site 86) and assigned them to oxygen-isotope substage 5c through aminostratigraphic correlation, in agreement with Neotyrrhenian assignment based on stratigraphy. Dai Pra & Hearty (1988) also showed that these deposits could be correlated with an inner edge at 85 m but contended that this inner edge was in fact carved during oxygen-isotope substage 5e. This conclusion is based on two lines of evidence: (1) the observation that the dated deposits are found onlapping on older (undated) deposits, presumably associated with oxygenisotope substage 5e, and (2) the geomorphological extrapolation of the 85 m inner edge towards the northeast through the mouth of the Sinni, Agri and Cavone rivers (sites 87-92) and as far as the Ponte del Re site (site 93), which is positively assigned a Tyrrhenian age (see discussion below). Cucci & Cinti (1998) use Dai Pra & Hearty's (1988) evidence from Piano San Nicola to propose an alternative dating scheme. In particular, they related the oxygen-isotope substage 5c age obtained for the 43 m deposits directly to the 85 m inner edge. As for the Tyrrhenian inner edge proper, they placed it at 115 m (site 85; very close to site 86, but with different age assignment) by extrapolation of the rate of uplift inferred from the 85 m terrace surface. The two dated terrace surfaces (85m and 115 m) were then extrapolated towards the Southwest up to Castrovillari-Coscile River.
DEFORMATION OF THE 125ka MARINE TERRACE IN ITALY Having considered all available information and the arguments used to support the two contrasting opinions, we decided to accept Dai Pra & Hearty's (1988) view as to the age of the 85 m terrace at Piano San Nicola, and Cucci & Cinti (1998) reconstruction for the section of the coastline enclosed between Piano San Nicola and Castrovillari (sites 77-85). This decision is supported by the observation that the 85m terrace and its extension towards the southwest and northeast is consistently wider and better expressed than the 115 m surface, in agreement with the common observation that over terraces created during the oxygen-isotope substage 5e tend to dominate geomorphologically over terraces created during slightly younger or older highstands (e.g. Hearty 1986). Within this scheme, the sediments sampled at 43 m elevation and dated to oxygen-isotope substage 5c by Dai Pra & Hearty (1988) would mark the reoccupation of the lower part of a wave-cut platform already carved during the Tyrrhenian proper highstand. This phenomenon of partial masking of a terrace surface by intervening deposits is commonly observed along relatively slowly uplifting shorelines at the passage between the highstands associated with the oxygen-isotope substages 5e and 5c (an example from central California has been described by Bradley & Griggs (1976)) as a result of the strong difference in the elevation reached by sea level during these two climatic maxima (10-15 m). For these reasons sites 84 and 85 will not be used in subsequent analyses. The Ponte del Re marine terrace site (site 93) is assigned to the Tyrrhenian based on both palaeontological evidence and aminostratigraphic correlation of marine deposits sampled at 24m elevation by Dai Pra & Hearty (1988). Following Brfickner's (1980) work on this area, Dai Pra & Hearty (1988) correlated the dated deposits with a shoreline marking the bottom of a dune ridge at an elevation of 35 m. In contrast, Boenzi et al. (1985) proposed the association of the deposits with a shoreline at 45m which exhibits a better geomorphological expression. Using the same reasoning as expressed above, we prefer the latter interpretation and retain it for further analyses.
Ionian and Adriatic coasts o f Apulia, f r o m Taranto to Gargano Both the Ionian and Adriatic sides of Apulia appear to have been strongly influenced by coastal erosion, suggesting steady uplift of the whole block (Fig. 6). For the palaeoecological reasons described above (recall Selli's (1962) con-
97
clusions), however, the Adriatic side was most probably never colonized by a warm fauna during the warmest highstands, and hence the local fossil assemblages do not allow major climatic changes to be resolved as is possible in other parts of the peninsula. Surprisingly, palaeoecological conditions must have been substantially different on the Ionian side of Apulia and particularly in the innermost portion of the Taranto Gulf, a region that has been investigated for over a century thanks to the diversity and abundance of exposures of stratigraphic significance (Hearty & Dai Pra 1992). For these reasons, dating of ancient shorelines on the Adriatic side of Apulia is accomplished mostly by correlation with similar landforms on the Ionian side. The only exception is represented by dating of Holocene raised beaches (Magri & Zezza 1971; Mastronuzzi & Sans6 1998), which allow some inferences to be made also about the Tyrrhenian shoreline. We will describe the Apulian coast in an anticlockwise fashion, moving from Taranto towards the southeast and then towards the Gargano promontory. In the Taranto-Mar Piccolo area (sites 94-97) most investigators have placed the Tyrrhenian shoreline at about 28 m (Cotecchia et al. 1969, 1971; Dai Pra & Stearns 1977; Hearty & Dai Pra 1985, 1992; Hearty 1986; Dai Pra & Hearty 1988). From these sites the Tyrrhenian shoreline appears to drop slowly to 15-20m at Torre Colimena (site 98) (Dai Pra 1982), and then to 2 m at Porto Cesareo-Torre Castiglione (site 99) (Dai Pra 1982; Dai Pra & Hearty 1988; Hearty & Dai Pra 1992, and references therein). Based on available radiocarbon (Cotecchia et al. 1969) and U-Th (Cotecchia et al. 1971) dating and on aminostratigraphic correlation (Hearty & Dai Pra 1985), Cosentino & Gliozzi (1988) reported a Tyrrhenian shoreline at about 28 m for the Gallipoli area (Fig. 1). However, Dai Pra & Hearty (1988) and Hearty & Dai Pra (1992) revised this elevation down to about 10 m (sites 100 and 101) using a broader spectrum of evidence which includes new U-Th dating and aminostratigraphic correlation. For this reason, we retain this value for subsequent analyses. Overall, these results suggest relative stability of the Ionian side of Salento, the southernmost portion of Apulia. Things are not very different on the Adriatic side of Salento, for which Cosentino & Gliozzi (1988) reported Tyrrhenian marine deposits at about 8 m at Grotta Romanelli, near the village of Castro (site 102). Di Stefano et al. (1992) reported ossiferous gravels between 2 and 8 m of elevation from caves located between Castro and Capo d'Otranto, about 15 km to the northeast.
98
P. BORDONI & G. VALENSISE
In the absence of any datable material, and on the basis of Hearty et al.'s (1986a) results for the whole Mediterranean, Di Stefano et al. (1992) assigned these deposits to the oxygen-isotope substage 5a-c highstand. This correlation, however, is inconsistent with the results obtained for southern Salento by Dai Pra & Hearty (1988) and Hearty & Dai Pra (1992), according to which sea level only reached 10 m in this region during the Tyrrhenian highstand. Owing to this inconsistency, we will retain the estimate proposed by Cosentino & Gliozzi (1988) for this site. For the Brindisi area various workers (e.g. Di Geronimo 1970) have reported at least five marine terraces at various elevations. These terraces (sites 103-105) are carved in marine deposits which Mastronuzzi & Sans6 (1998) ascribed to MidPleistocene time as they contain dated minerals from the Mt. Vulture volcano (La Volpe & Principi 1994; Fig. 1). According to Di Geronimo (1970), three of these terraces (at 65, 40 and 25 m elevation) are of Tyrrhenian age. Pieri (1975) and Iannone & Pieri (1979) assigned to the 'last interglacial' all shorelines at present observed between 30 m and modern sea level, on the basis of correlation with terrace elevations on the Ionian side. Mastronuzzi & Sans6 (1998) used radiometric dating to assign to the Holocene marine deposits sampled close to sea level near Torre Guaceto (site 103), and suggested that slightly higher deposits (4m elevation) could be generically ascribed to the Late Pleistocene. On the basis of this partially contrasting evidence we decided to assign the Tyrrhenian shoreline a conventional elevation of 4 m for the latter site (site 103) and 30 m for the two sites located to the northwest (sites 104 and 105). We could not find any quantitative report of Late Pleistocene coastal deposits and landforms for the section between Brindisi and Manfredonia. Qualitative studies on marine terraces and coastal uplift do exist for the Gargano Promontory (e.g. Ricchetti et al. 1988), a region that impresses for the abundance and clarity of landforms unquestionably generated by wave-action, but no constraints on the age of such terraces are known in the current literature. Finally, for a site north of Gargano near the mouth of the Fortore River (site 106) Mastronuzzi et al. (1989) described an uplifted coastal plain culminating at 25m elevation. A 5-10m erosional cliff separates this remnant from the modern plain below, which in turn is covered by post-Wtirmian dune ridges forming the only age constraint available for the overlying surface. In the absence of any means of precisely dating the 25 m surface we tentatively assign it to the Tyrrhenian stage.
Sicily
In this section we describe the Tyrrhenian shoreline around Sicily starting from the northern tip of the Messina Straits (Capo Peloro) and then moving counterclockwise (Fig. 5). Bonfiglio (1981, 1991) reported a marine terrace with inner edge at 130 m, which can be discontinuously followed from Capo Peloro (site 107) to Acquedolci (site 110) along the Tyrrhenian coast, and from Capo Peloro to Taormina (site 121) along the Ionian coast. On the basis of the occurrence of S. bubonius at 84m (Bonfiglio & Violanti 1983) and of aminostratigraphic dating (Hearty et al. 1986b), Bonfiglio (1981, 1991) assigned this terrace to the Tyrrhenian stage. Somewhat in contrast, Cosentino & Gliozzi (1988) placed the Tyrrhenian shoreline at 85 m elevation at Capo Tindari (site 109; see also Fig. 1) on the basis of the presence of Lithophaga holes correlated with marine deposits at 60m elevation at Capo Milazzo (site 108) (Gliozzi & Malatesta 1982), which are assigned to oxygenisotope substage 5e by aminostratigraphic correlation (Belluomini 1985; Hearty 1986; Hearty et al. 1986b). It should be noted that at least for Capo Milazzo (site 108) the conventional elevation of 60m only reflects the elevation of the terrace deposits, as the inner edge is probably lost to erosion. No quantitative observations are available for a 120 km stretch of coastline between Capo d'Orlando and Capo Zafferano, but several investigators have described a possible extension of the 130m terrace at somewhat lower elevations as far west as Cefalfi. For the Palermo area both early workers (Gignoux 1913; Fabiani 1941) and more recent compilations (Cosentino & Gliozzi 1988) report marine deposits with S. bubonius at about 8-10 m (sites 112 and 113; see also Fig. 1). In contrast, Trevisan (1943) and Ruggieri et al. (1968) reported a Tyrrhenian shoreline at 30m at the foot of Monte Pellegrino (just north of Palermo, not shown in Table 1 and Fig. 5). Antonioli et al. (1994a) sampled marine deposits associated with notches at 7 and 33 m elevation at Capo Zafferano (site 111). The two deposits are assigned respectively to the oxygen-isotope stage 5e and to a generic pre-stage 5 (presumably stage 9 or 11) by aminostratigraphy. We assume that the 30m shoreline seen at Monte Pellegrino correlates with the 33 m shoreline described by Antonioli et al. (1994a), and hence retain the elevation of 8-10m as representative of the elevation of the Tyrrhenian shoreline in this area. Several investigators have placed the Tyrrhenian shoreline at 3 - 4 m throughout the
DEFORMATION OF THE 125ka MARINE TERRACE IN ITALY Trapani-Mazara del Vallo area, on the basis of the presence of marine deposits with S. bubonius (Ruggieri et al. 1968; Ruggieri & Unti 1974) and a morphometric study along selected transects spanning the whole terrace sequence (D'Angelo & Vernuccio 1994) (sites no. 114-116). Only one qualitative observation is available for the whole southern coast of Sicily. Carbone et al. (1982) reported a series of marine terraces culminating at 20 m between Marina di Ragusa and Pozzallo (site 117), which they attributed to the Tyrrhenian stage by correlation with similar terraces in the Augusta area. For the Capo Passero-Pachino area, at the southernmost tip of Sicily, Malatesta (1985) reported S. bubonius deposits at a maximum elevation of 15 m (site 118). According to Carbone et al. (1982), only one terrace surface is clearly visible in this area, which turns into two shorelines separated by a subdued cliff between Fontane Bianche and Siracusa and then into three distinct shorelines between Siracusa and Augusta, which is also suggestive of a progressive increase of the rate of uplift from south to north. Carbone et al. (1982) assigned these terraces to a generic Tyrrhenian stage following Di Grande & Scamarda (1973), who report three separate shorelines at 3-10m below sea level and 5-10m and 10-30m above sea level for the stretch of coastline between Punta Bonico and the Santa Croce Lighthouse (site 120) at the southern end of the Monte Tauro peninsula. Di Grande & Scamarda (1973) assigned to the Tyrrhenian stage proper the deposits associated with the two higher shorelines, owing to the presence of S. bubonius. Cosentino & Gliozzi (1988) also reported marine deposits containing S. bubonius at 8.5m elevation, and correlates them with a notch at about 30m (site 119; see also Fig. 1). According to Bonfiglio (1991), the highest of Di Grande & Scamarda's (1973) shorelines culminates at 34m near Brucoli, which supports on firmer grounds the observation of the 30 m notch at Augusta. Finally, Bonfiglio (1991) described a 'pocket beach' deposit at a maximum elevation of 40 m near the Coste di Gigia tourist resort, a few kilometres southwest of Augusta. Bonfiglio (1991) herself, however, pointed out that these deposits are similar to marine sediments sampled at Tommaso Natale near Palermo and assigned to oxygen-isotope stage 7 by Hearty et al. (1986), and therefore cannot be taken as indicative of the local elevation of the Tyrrhenian shoreline. Nevertheless, this observation indirectly supports a Tyrrhenian shoreline in the elevation range 30-34 m throughout this area.
99
Distribution of uplift Calculating the rate of uplift implied by our observations involves the assessment of the exact age and of the elevation o f formation of the Tyrrhenian highstand. In recent years various investigators have inferred complexities in sea-level history around the peak of the 5e climatic optimum. As for its elevation, existing estimates generally vary between 2 and 10m above sea level. In consideration of the limited resolving power offered by the vast majority of the observations available to us, we chose to use for the Tyrrhenian highstand the 'traditional' estimates of 125 ka and 6 m, respectively, for its age and average elevation. An additional source of uncertainty in any estimates of uplift rate arises from the usually undeclared assumption of constancy of uplift between the time of formation of any given terrace and the present day. This assumption is probably reasonable for Late Pleistocene marine terraces found in areas where the same geodynamic regime has persisted for timescales longer than the age of the terraces themselves, but it may be questionable for areas that have experienced significant geodynamic events during this time span. The rates of uplift inferred for our 121 observations under the assumptions set out above are reported in the last column of Table 1. Figure 8 shows their geographical distribution. Our rates were calculated under the assumption that for each terrace site we could locate and assess the elevation of the inner edge, and must therefore be taken as a minimum for all sites for which our best guess of the inner edge elevation is represented by the maximum elevation of the associated deposits. They may also under- or overestimate the uplift rate where it is based on the elevation of the terrace surface. Cosentino & Gliozzi (1988) used conventional elevations calculated by adding 25 m to the actual elevation of any Tyrrhenian deposits, under the rather crude assumption that the preferred habitat of specifically Tyrrhenian marine faunas was centred around 25 m depth (see our Fig. 1, and figs 1 and 2 of Cosentino & Gliozzi (1988)). Nevertheless their uplift rates compare rather well with ours, particularly along faster uplifting coasts where any assumption about palaeoecology and palaeogeography becomes negligible in comparison with the extent of subsequent uplift. We decided to set at 20 m (corresponding to a rate of uplift of 0.11 ram/year) the elevation at which uplift can be unambiguously identified. The calculation for uplift rate then becomes (X-6)/125, where X is the observed terrace
100
P. BORDONI & G. VALENSISE
Fig. 8. Rate of Late Pleistocene uplift in peninsular Italy. For clarity of representation all uplift values in the range 0.0-0.11 mm/year are shown with symbols of the same height. It should be noted that only three sites (shown in black) located along the Tyrrhenian coast exhibit net subsidence during this time interval. height. Inspection of Figs 2 and 3-8 reveals that there are essentially five areas that have experienced uplift above this level (the corresponding ranges of uplift rate are shown in brackets): (1) the coast of Latium, from the Tuscany border to west of Latina (sites 6-19; Fig. 3), along the seaward flank of the Vulsinian, Sabatinian and Alban Hills volcanoes [0.11-0.31 ram/year];
(2) the inner parts of the Campana Plain and its extension towards the Southeast, represented by the Sele Plain (sites 31 and 38; Fig. 4) [0.15-0.35 mm/year]; (3) the Sila-Aspromonte-PeloritaniNebrodi-Madonie range, from Vibo Valentia (but probably from San Lucido: see discussion in the relevant section) to the alignment between Cefalfi and Siracusa (sites 54-59, 107-110,
DEFORMATION OF THE 125ka MARINE TERRACE IN ITALY 119-121; Fig. 5), including the Messina Straits (sites 59-64), all the Ionian side of southern Calabria (sites 65-74) and the Crotone promontory (sites 75 and 76) [0.11-1.21mm/year]; (4) the southeastern end of the main Apennines belt facing the Gulf of Taranto (sites 77-93; Figs. 6 and 7) [0.31-1.03mm/year]; (5) the central and northern portions of Apulia, from Taranto to Lizzano (sites 94-97; Fig. 6) and from Ostuni to the Gargano promontory (sites 105 and 106; Fig. 6) [0.15-0.19ram/year]. Figure 8 shows that the most of the Tyrrhenian side of the Apennines from Tuscany to northern Calabria has been essentially stable in the Late Pleistocene (sites 1-53; Fig. 2), with two significant exceptions: the section of the coast bordering the three main Pleistocene Latium volcanoes (sites 6-19; Fig. 3) (see description above); the Campana Plain (sites 32 and 33; Fig. 4). Surprisingly, the stable region encompasses coastal areas characterized by steep topography, and sometimes even by older marine terraces that at first glance would suggest active uplift, such as: the section between Mt. Circeo and Gaeta (sites 22-28; Fig. 3); the Sorrento peninsula (sites 34-37; Fig. 4); the coasts of Cilento and northern Calabria as far south as Cetraro (sites 39-53; Fig. 4). As for the remaining parts of the peninsula, stability is observed: in the Salento area, southernmost Apulia (sites 99-103; Fig. 6); in western and southern Sicily, from Capo Zafferano-Palermo anticlockwise to Mazara del Vallo (sites 111-116; Fig. 5), and from Pozzallo to Siracusa (sites 117 and 118; Fig. 5), possibly including the whole southwestern side of Sicily. Observations of Tyrrhenian shorelines along subsiding coasts are obviously limited by the submergence of any erosional or depositional record. Our database contains three significant exceptions, from the Fondi Plain (site 24; Fig. 3) and the Campana Plain (sites 32 and 33; Fig. 4), both areas where post-Tyrrhenian subsidence has been fully compensated by intervening sediment accumulation.
Tectonic implications Our observations may have significant implications for the understanding of the recent geodynamics of the Apennines and may help us to address important open issues. Some of them are briefly outlined in the following summary. The present geological and topographical setting of the Apennines is the result of a tectonic history that started with the collision between the African and European plates between Cretaceous time and Paleogene time. The construction of the
101
range started in Tortonian and was accomplished by progressive thrusting and piling up of oceanic and carbonate platform units towards the east and the northeast and by a parallel northeastward migration of the thrust belt-foredeep-foreland system (Fig. 1). According to most investigators, starting from Late Tortonian time the western margin of the Apennines experienced significant stretching associated with the opening of the Tyrrhenian Sea. Over time the extension also migrated towards the northeast following the front of compression. The most widely accepted mechanism to explain the contemporaneous action of compression and extension and their progressive northeastward shift involves outward migration of the subduction zone and subsequent passive sinking of the underthrusting (African) plate into the mantle. A back-arc basin, corresponding to the present Tyrrhenian Sea, would then form if the overriding plate failed to keep pace with the retreat of the subduction zone (Malinverno & Ryan 1986). The recent evolution of the Apennines is marked by a drastic 'geodynamic change' that is supposed to have taken place in the region at about 700ka, around the beginning of MidPleistocene time. Several independent lines of evidence, which include the beginning of northeastward tilting of Early Pleistocene basin deposits of the Bradanic foredeep (Cinque et al. 1993; Patacca et al. 1997; Fig. 1), the age of the oldest marine terraces (Ciaranfi et al. 1988; Westaway 1993), the youthfulness of seismogenic faulting in the region (Pantosti & Valensise 1990), the analysis of structural relationships (Hippolyre et al. 1994), the analysis of the present-day stress field (Amato & Montone 1997), and even the extrapolation of palaeoseismological estimates (Pantosti et al. 1993), suggest a geodynamic change marking the end of compressional strains and the beginning of a tectonic style dominated by vertical motions. To place the deformation implicit in our findings in the framework of the relevant active tectonic processes we have projected the marine terraces along three transects across portions of the peninsula that have experienced significant tectonic activity in Late Pleistocene time (Fig. 9). All the transects are shown paralleled by a bestfitting polynomial curve drawn with the purpose of smoothing out some of the small-scale irregularities, to emphasize the larger wavelength components forming each trend. Coast of Latium
The first transect (Figs 1 and 3) extends from the Latium-Tuscany border (near site 3) to Mt.
102
P. BORDONI & G. VALENSISE
Fig. 9. Uplift trends along three selected transects (shown in the inset to the far right). Small numbers on white background locate the observations used to construct the transects (see Table 1 and text). For clarity of representation the transect along the Messina Straits uses data from Valensise & Pantosti (1992) in addition to data from our database. The strong differences in net uplift and in the wavelength of the fluctuations of the Tyrrhenian shoreline among different transects should be noted.
Circeo (site 22). The recent evolution of this region is strictly related to the volcanism responsible for the creation of the Vulsinian and Sabatinian complexes and of the Alban Hills. The age of volcanism is conventionally bracketed between about 0.6 Ma (Barberi et al. 1994) and about 0.04 Ma, which is the age of the most recent rocks erupted by the Alban Hills volcano (Alessio et al. 1966). As on the modern coast, the Tyrrhenian shoreline in all three areas runs a few kilometres beyond the maximum extent of volcanic rocks and within 30-40 km of the centre of the volcanic edifice (compare Figs. 1 and 3). In particular, at each site of the shoreline uplift is roughly proportional to the distance from the centre of the closest of the three volcanoes, whereas it tapers gently to zero to the north of the Fiora River and to the south of Anzio. Overall, the Tyrrhenian shoreline describes a large, gently upward convex shield that matches rather closely the general trend of the topography along the transect.
Southern Apennines
The second transect (see also Fig. 6) is very different. By cutting through the extreme southeastern end of the Apennines, this transect bears on three widely distinct sections of the orogen (Fig. 1): the chain, which is essentially a fold and thrust belt forming the highest topography of the region; the foredeep, which is filled with a thick wedge of mildly deformed post-orogenic sediments; and the foreland, formed by a thick sequence of virtually undeformed carbonate shelf deposits.
The uplift signal available along this transect is everywhere well beyond any uncertainties left in the exact location and age of the observed shorelines, and hence the information on the extent and pattern of uplift that can be obtained is especially robust. Nevertheless, uplift of this region has for a long time been virtually ignored by most workers, who focused almost exclusively on the Aspromonte-Peloritani range (see, e.g. Cosentino & Gliozzi 1988; Fig. 1), and is largely unknown to existing large-scale neotectonic compilations (see, e.g. Ambrosetti et al. 1987). Starting from the southwestern end of the transect the trend of the Tyrrhenian shoreline shows at least four distinct uplift domains (the corresponding ranges of uplift rate are shown in brackets): (1) a steep hike from virtually no uplift (around site 53 at 12 m) to some of the largest values observed across the peninsula (about 135m at sites 79-81) [0.05-1.03mm/year]; (2) a 20-30 km-long plateau where the elevation stays nearly constant between 115 and 135m (sites 77-81) [0.87-1.03 ram/year]; (3) a gentle descent from the upper plateau to a wide trough of minimum uplift (from site 81 at 135m to site 94 at 28 m) [1.03-0.18 ram/year]; (4) a second plateau near the minimum of the observed uplift (sites 94 and 95, and 104 and 105) [0.18-0.19 mm/year]. These four domains have an interesting correspondence in the topography. Figure 10 shows a comparison between Transect 2 and the topographical trend along ten successive profiles cut at a distance of 5 km (for the four closest) to 10 km (for the remainder) from each other. It should be noted that the strike of Transect 2 and of all the profiles is 45~ which is taken as the average orientation of the perpendicular to the southern
DEFORMATION OF THE 125ka MARINE TERRACE IN ITALY
103
Fig. 10. A comparison of the uplift trend observed along Transect 2 of Fig. 9 and the topography of the southern Apennines along selected profiles. Topographical data were obtained from a digital elevation model with average grid spacing of about 240 m. The similarity between the trend of the post-Tyrrhenian uplift curve and the general trend of topography, which in itself suggests that the present configuration of the Apennines is strongly controlled by modern uplift, should be noted. Apennines. The profiles were drawn using a digital elevation model of Italy with an average grid spacing of about 240 m (resampled from an archive originally compiled by Carrozzo et al. (1981)). All the profiles are paralleled by a curve showing the best-fitting polynomial approximation. The trend of the post-Tyrrhenian uplift curve and the topographical trends have the same general features, including the strong asymmetry towards the southwest and the existence of a broad trough to the northeast, they share similar height relationships, and they all show a culmination in the distance range 40-70km from the southwestern end. It should be noted that in profiles 4-10 the fit is worsened by the removal of a large amount of sediment by the Bradano, Basento, Cavone and Agri rivers (see Fig. 4). The idea that the mere existence of marine terraces around the Taranto Gulf implies progressive uplift of the Apennines as a whole is not an entirely new one (see, e.g. Cotecchia & Magri 1967; Vezzani 1967). This intriguing concept and its implications, however, are in marked contrast with established geological wisdom, which, at least for this region, envisions a separate role for the chain s e n s u stricto, the foredeep and the foreland (e.g. Ambrosetti et al.
1987; see also Fig. 1). In contrast, our results show that Late Pleistocene uplift violates any pre-existing boundaries between geodynamic domains. In fact, on the basis of the evidence summarized in Figs 9 and 10, we speculate that sustained uplift at the rates inferred from the Tyrrhenian marine terraces explains the general configuration and a large fraction of the topography at present observed in the southern Apennines. Figure 11 illustrates this idea by comparing the topography observed along the transect with an idealized post-Early Pleistocene cumulative uplift curve and with the subsurface structural setting obtained from exploration data. The uplift curve was obtained by multiplying by seven the postTyrrhenian uplift trend (Fig. 9) to reproduce the topography that would accumulate after 870 000 years of uplift at the rates inferred in this paper. Figure 11 shows that positive residuals (shown in dark grey) with respect to the topography predicted by extrapolating post-Tyrrhenian uplift (shown in light grey) correlate faithfully with the hanging-wall of major thrusts along which slivers of Mesozoic carbonate platform rocks have been piled up and overturned. The exercise suggests that the overall shape of this section of the
104
P. B O R D O N I
& G. VALENSISE
~
~ o
~
0
0
.o-~
~9
~
-~
7,,
.~ , ~
0
.~. ~ =
~'~
,0
"~
~ "n
o.~ 0
o
~
-~ = u ~
.
o
~
~ 0
< .~= ~
= .~
~::~ ~ ' - ~
o
~'~
0n,-,.-, ~
o
o'~
o q=l,~
o
~~
~ =~ =r
~=~
"~
,-~
;~
c~ ~ . E
o
~
~ . .
~.~ 9 ~,~ o,..~
~
~ b~
.', . " " ~
~
o~.~
~
~-~
~
~
0
0
..
~.-,--'
~,~
~00
~
~
o'"~ ,.~ ~
~'
~'~
9,~ ~
~
~..
~--.
'~ O ~
DEFORMATION OF THE 125ka MARINE TERRACE IN ITALY Apennines may be seen as a combination of pre-Mid-Pleistocene progressive thrusting and post-Early Pleistocene regional doming. This conclusion is in good agreement with the c. 700 ka 'geodynamic change' and implies that only a small fraction of the present topography of the southern Apennines is the result of conventional mountain building caused by orogenic transport, whereas most of it owes its existence to subsequent regional doming. The overall shape and the large wavelength of the uplift trend (about 200 km: see Fig. 9) also suggest that the post-Early Pleistocene rise of the southern Apennines is driven by a deep-seated mechanism other than the traditional scheme of development of a northeast-verging orogenic system. In recent years various investigators have used seismological, geological and stressfield evidence combined with tectonic modelling (e.g. Westaway 1993) to propose such driving mechanisms, most of which involve changes in the geometry and conditions of Tyrrhenian subduction (for a comprehensive review of the possible lithospheric sources of modern tectonic forcing in the southern Apennines, see Amato & Montone (1997)). Clarification of the uplift driving mechanism may have further important implications, for example, concerning the distribution of present-day extensional stress and the distribution of seismicity along the southern Apennines. Messina Straits
Our third transect (see also Fig. 5) runs along the north-trending axis of the Messina Straits. The history of the investigation of this region is long and fascinating, for this has been the chosen field area for generations of geologists and geomorphologists. The details of the distribution of Tyrrhenian terraces in the Straits and the main published interpretations were discussed in a previous section. Here we wish to emphasize some aspects of the uplift trend that have long been known qualitatively but were made more evident by our systematic re-analysis. To make our case stand out more clearly, in addition to the data from Table 1 we also plotted on this transect some observations on the Tyrrhenian inner edge taken from Valensise & Pantosti (1992) (see their fig. 5). Late Pleistocene uplift in the Messina Straits has been significantly faster than anywhere else in the broader Italian peninsula. When placed in a broader regional perspective (see Fig. 8), this culmination is somehow 'suggested' by a steady increase in the uplift rate both moving southward along the axis of Calabria and moving northeastward along the watershed of the NebrodiPeloritani range. In other words, the Messina
105
Straits are located at the apex of a 250 km long arc, generally referred to as the Calabrian Arc, at both ends of which Late Pleistocene uplift tapers to zero. To judge from this evidence (and in partial rectification of a statement made in the discussion concerning the southern Apennines), one could say that modern uplift does violate pre-existing geodynamic boundaries parallel to the orogen but tends to honour those found along its length. A similar statement certainly applies to the Sangineto Line, which marks the boundary between the Calabrian Arc and the southern Apennines, and to a lesser extent to the Longi-Taormina Line (often referred to also as the Tindari-Giardini Naxos Line; both lineaments are marked with a heavy dashed grey line in Fig. 1). A similar feature within the southern Apennines could be represented by the Ofanto Line, an alignment between the mouth of the Ofanto River and northern Cilento (Fig. 4), which marks a sharp separation in the general trend of the topography possibly resulting from a drop in the rate of modern uplift. Perhaps not surprisingly, the region across the Sangineto Line is characterized by very limited or no seismicity in comparison with adjacent areas (e.g. Valensise & Guidoboni 1995; Cinti et al. 1997). A second striking difference between the uplift trend observed in the Messina Straits and those observed elsewhere is the much shorter wavelength of the fluctuations in the elevation of the Tyrrhenian shoreline observed here. Valensise & Pantosti (1992) related these fluctuations to the combination of large-scale and relatively uniform uplift of the region with more localized subsidence resulting from the fact that all the Calabrian side of the Straits lies in the hanging wall of the Messina Straits fault. In this framework, the observed fluctuations would be the result of the irregularity of the coastline and of the subsequent incursions of the shoreline closer to the fault itself. This process alone would account for 30-40% fluctuations in the resulting elevation of the shoreline within a 30km distance, a condition not seen anywhere else throughout the investigated region (see Fig. 8). The 5-10km wavelength of individual fluctuations would in itself constrain their ultimate source at a comparable depth in the upper crust, well within the seismogenic layer. Indirect support to this hypothesis comes from the observation that no major faults (Valensise & Pantosti 1999) or major historical earthquakes (Boschi et al. 1997) are known for the regions crossed by the two preceding transects, which at least within the resolution offered by the data exhibit a more regular trend. A notable
106
P. B O R D O N I & G. VALENSISE
exception could be represented by the limited anomaly observed near site 78 on Transect 2 which falls near the southern end of the recently discovered Castrovillari n o r m a l fault (Cinti et al. 1997). Again not coincidentally, and in striking contrast to the observation m a d e above concerning the region straddling the Sangineto Line, the Messina Straits fault is c o m m o n l y regarded as one of the fastest slipping faults of the Italian peninsula (Valensise & Pantosti 1999).
References ALESSIO, M., BELLA, F. BACHECHI, F. & CORTESI, C. 1966. University of Rome-carbon 14 dates IV. Radiocarbon, 8, 401-412. AMATO, A. & MONTONE, P. 1997. Present-day stress field and active tectonics in southern peninsular Italy. Geophysical Journal International, 130, 519-534. AMaROSETTI, P., BARTOLINI, C. & BosI, C. 1981. L'evoluzione geologica e morfologica quaternaria dell'area adiacente la bassa valle del Flume Fiora (Viterbo). Geografia Fisica e Dinamica del Quaternario, 4, 104-134. -, BOSI, C., CARRARO, F. et al. 1987. Neotectonic map of Italy, scale 1:500000. Quaderni della Ricerca Scientifica, Consiglio Nazionale delle Ricerche, 114. ANTONIOLI, F. 1991. Geomorfologia subacquea e costiera del litorale compreso tra Punta Stendardo e Torre Sant'Agostino (Gaeta). Il Quaternario, 4, 257-274. , BELLUOMINI,G., FERRANTI, L., IMPROTA, S. & REITANO, G. 1994a. I1 sito preistorico dell'arco naturale di Capo Zafferano (Sicilia) aspetti geomorfologici e relazione con le variazioni del livello marino olocenico. II Quaternario, 7, 109-118. - - , DAI PRA, G. & HEARTY, P. J. 1988. I sedimenti quaternari nella fascia costiera della piana di Fondi (Lazio meridionale). Bollettino della Societal Geografica Italiana, 107, 491-501. - - , PUGLISI, C. & SILENZI, S. 1994b. Morphostratigraphic survey of emerged and submerged coast of the northern portion of Capo Palinuro promontory. A tti del Con vegno GEOS UB '94, Memorie Descrittive della Societfi Geologica Italiana, Vol. LII. ASCENZI, A. & SEGRE, A. 1971. A new Neanderthal child mandible from an Upper Pleistocene site in Southern Italy. Nature, 233, 280-283. BARBERI, F., BUONASORTE, G., CIONI, R. et al. 1994. Plio-Pleistocene geological evolution of the geothermal area of Tuscany and Latium. Memorie Descrittive della Carta Geologica Italiana, 49. BARRA, D., CINQUE, A., GEWELT, M. & HURTGEN, C. 1991. L'ospite caldo Sylvestra seminis (Bonaduce, Masoli & Pugliese 1976) (Crustacea, Ostracoda): un possibile marker dell'ultimo interglaciale nell' area mediterranea. II Quaternario, 4(2), 327-332. BARRIER, P., DI GERONIMO, I., ZIBROWIUS, H. & RAISSON, F. 1988. Faune sOn6galienne du pal~oescarpement du Capo Vaticano (Calabre mOridio-
nale). Implication n6otectoniques. Atti IV Simposio di Ecologia e Paleoecologia delle Comunitd Bentoniche, Sorrento, 1-5 November 1988. Museo di Scienze Naturali di Torino, 511-526. BELLUOMINI, G. 1985. Risultati e prospettive di un nuovo metodo di datazione basato sulla racemizzazione degli aminoacidi. Accademia Nazionale dei Lin-cei, Contributi Centro Linceo Interdisciplinare, Scienze Matematiche e loro applicazioni, 69, 135-171. BLANC, C. A. 1935. Stratigrafia del Canale Mussolini nell'Agro Pontino. Processi Verbali Societd Toscana Scienze Naturali, 54, 52-56. - - 1 9 3 6 . Una spiaggia pleistocenica a 'Strombus bubonius' presso Palidoro (Roma). Rendiconti Accademia Nazionale dei Lincei, 23, 200-204. 1953a. Plage Tyrrhenienne et dunes fossiles de la Buca dei Corvi (Castiglioncello). Livret-guide, I V Congres INQUA, Roma 1953, 7-13. 1953b. Le Quaternaire du monte Circeo. Livretguide, IV Congres INQUA, Roma 1953, 23-108. & CARDINI, L. 1961. Prospezione nei dintorni di Praia a Mare e Scalea (Cosenza). Quaternaria, 5, 58-61. & SEGRE, A. G. 1947. Nuovi giacimenti tirreniani e paleolitici sulla costiera tra Sperlonga e Gaeta. Historia Naturalis, 2. BOENZI, F., CALDARA, M. & PENNETTA, L. 1985. La trasgressione Tirreniana nei dintorni di Castellaneta (Taranto). Geologia Applicata e Idrogeologia, 20, 163-175. BONADONNA, F. P. 1967a. Studi sul Pleistocene del Lazio. III. Linee di costa lungo il litorale di Tarquinia (Lazio Settentrionale). Geologica Romana, VI, 131-135. - - 1 9 6 7 b . Studi sul Pleistocene del Lazio. IV. La linea di costa tirreniana di Ponte Galeria (Roma). Quaternaria, IX, 285-299. BONFIGLIO, L. 1972. I1 Tirreniano di Bovetto e Ravagnese presso Reggio Calabria. Quaternaria, 16, 137-148. 1981. Terrazzi marini e depositi continentali quaternari di Taormina. Quaternaria, 23, 81-98. - - 1 9 9 1 . Correlazione tra depositi a mammiferi, depositi marini, linee di costa e terrazzi medio e tardo pleistocenici nella Sicilia orientale. // Quaternario, 4, 205-214. & BERDAR, A. 1986. Gli elefanti del Pleistocene superiore di Archi (RC): nuove evidenze di insularitfi della Calabria meridionale durante il ciclo Tirreniano. Bollettino della Societd Paleontologica Italiana, 25(1), 9-34. -& VIOLANTI, D. 1983. Prima segnalazione di Tirreniano ed evoluzione pleistocenica del Capo Peloro. Geografia Fisica e Dinamica Quaternaria, 6, 3-15. , BELLOMO, E., BELLOMO, G., BONADUCE, G. & VIOLANTI, D. 1988. Analisi biostratigrafica e paleoambientale dei depositi marini e salmastri del Pleistocene di Contrada Ianni di S. Calogero (Catanzaro, Calabria, Italia). Atti I V Simposio di Ecologia e Paleoecologia delle Comunitd Bentoniche, Sorrento, 1-5 November 1988. Museo di Scienze Naturali di Torino, 527-573. -
-
-
-
-
-
D E F O R M A T I O N OF T H E 125ka M A R I N E T E R R A C E IN ITALY BONIFAY, F. & Mars, P. 1959. Le Tyrrh6nien dans le cadre de la chronologie quaternaire m6diterran6enne. Bulletin de la SociOt~ G~ologique de France, 7, 62-78. BOSCHI, E., GUIDOBONI,E., FERRARI, G., VALENSISE, G. & GASPERINI, P. 1997. Catalogo dei Forti Terremoti in Italia dal 461, a.C. al 1990. Istituto Nazionale di Geofisica and SGA, Bologna 1997 (also CD-ROM). BRADLEY, W. C. & GRIGGS, G. B. 1976. Form, genesis, and deformation of central California wave-cut platforms. Geological Society of America Bulletin, 87, 433-449. BRANCACCIO, L. 1968. Genesi e caratteri delle forme costiere nella penisola Sorrentina. Bollettino della Societ?t dei Naturalisti in Napoli, 77, 247-274. , CAPALDI,G., C1NQUE,A., PECE, R. & SGROSSO,I. 1978. 23~ dating of corals from a Tyrrhenian beach in Sorrentine peninsula (southern Italy). Quaternaria, 20, 175-183. , CINQUE, A., BELLUOMINI, G., BRANCA, M. & DELITALA, L. 1986. Isoleucine epimerization dating and tectonic significance of upper Pleistocene sea-level features of the Sele plain (southern Italy). Zeitschrift fu'r Geomorphologie, Supplement band, 62, 159-166. , D'ANGELO, G., Russo, F., SANTANGELO, ~q. & SGROSSO, I. 1987. Evoluzione tettonica e geomorfologiea della piana del Sele (Campania, Appennino meridionale). Geografia Fisica e Dinamica Quaternaria, 10, 47-55. , Russo, F., BELLUOMINI,G., BRANCA, M. DELITALA, L. 1990. Segnalazione e datazione di depositi marini tirreniani sulla costa Campana. Bollettino della Societd Geologica Italiana, 109, 259-265. BROCKNER, H. 1980. Marine terrassen in Sfiditalien. Eine Quart/ir-morphologische Studie fiber das Kfistentiefland yon Metapont. Da'sseldorfer Geographische Schriften, 14. CARBONE,S., DI GERONIMO,I., GRASSO,M., IOZZIA,S. & LENTINI,F. 1982. I terrazzi marini dell'area Iblea ( Sicilia Sud-Orientale ). Contributi conclusivi per la realizzazione della earta Neotettonica d'ItaIia. Pubblicazioni Consiglio Nazionale delle Ricerche, 506, 1-35. CARBONI, M. G., MALATESTA, A. & ZARLENGA, F. 1988. I1 Quaternario fra Praja a Mare e Scalea (Calabria Settentrionale). Memorie della Societ~ Geologica Italiana, 41, 645-652. CAROBENE, L. & DAI PRA, G. 1991. Middle and Upper Pleistocene sea-level highstands along the Tyrrhenian coast of Basilicata (Southern Italy). II Quaternario, 4, 173-202. & GEWELT, M. 1986. Niveaux marins du P16istoc~ne moyen-sup&ieur de la c6te tyrrh+nienne de la Calabre (ltalie m~ridionale). Datations 23~ et tectonique r6cente. Zeitschrift far Geomorphologie, Supplementband, 62, 141-158. CARROZZO, M. T., CHIRENTI, A., LUZIO, D., MAGIOTTA, G., QUARTA, T. & ZUANNI, F. 1981. Realizzazione di un archivio delle quote medie. Proceedings, I Meeting Gruppo Nazionale Geofisica della Terra Solida, Rome, November
107
1981, 117-129. CHAPPELL, J. & SHACKLETON, N. J. 1986. Oxygen isotopes and sea-level. Nature, 324, 137-140. CIARANFI, N., PIERI, P. & RICCHETTI, G. 1988. Note alla carta geologica deUe Murge e del Salento (Puglia centro-meridionale). Memorie della Societal Geologica Italiana, 41,449-460. CINQUE, A. • ROMANO, P. 1990. Segnalazione di nuove evidenze di antiche linee di riva in Penisola Sorrentina (Campania). Geografia Fisica e Dinamica Quaternaria, 13, 23-36. - - - , ALINAGHI,H. H., LAURETI,L. & Russo, F. 1987. Osservazioni preliminari sull'evoluzione geomorfologica della piana del Sarno (Campania, Appennino meridionale). Geografia Fisica e Dinamica Quaternaria, 10, 161-174. - - , PATACCA,E., SCANDONE,P. & TozzI, M. 1993. Quaternary kinematic evolution of the southern Apennines. Relationships between surface geological features and deep lithospheric structures. Annali di Geofisica, XXXVI, 249-260. - - , ROMANO, P., ROSSKOPF, C., SANTANGELO,N. & SANTO, A. 1994. Morfologie costiere e depositi quaternari tra Agropoli e Ogliastro Marina (Cilento-Italia Meridionale). Il Quaternario, 7, 3-16. CINTI, F. R., CUCCI, L., PANTOSTI,D., D'ADDEZIO,G. & MEGHRAOUI, M. 1997. A major seismogenic fault in a 'silent area': the Castrovillari fault (southern Apennines, Italy). Geophysical Journal International, 130, 322-332. COLANTONI, P. 1970. La grotta dello Smeraldo di Amalfi e la linea di riva tirreniana. Le grotte d'Italia, 4, 45-60. CORTESE, E. 1895. Descrizione geologica della Calabria. Published in 1934 by Regio Ufficio Geologico d'Italia, reprinted by Casa del Libro, Rome 1983. COSENTINO, D. & GLIOZZI, E. 1988. Considerazioni suUe velocit~i di sollevamento di depositi eutirreniani dell'Italia Meridionale e della Sicilia. Memorie della Societh Geologica Italiana, 41, 653-665. COTECCHIA, V. t~ MAGRI, G. 1967. Gli spostamenti delle linee di costa del mare Ionio fra Capo Spulico e Taranto. Geologia Applicata ed ldrogeologia, 2, 1-28. --, DAI PRA, G. & MAGRI, G. 1969. Oscillazioni tirreniane e oloceniche del livello del mare nel Golfo di Taranto, corredate da datazioni con il metodo del radiocarbonio. Geologia Applicata ed Idrogeologia, 4, 93-148. & 1971. Sul Tirreniano della costa ionica salentina (Puglia). Datazione di un campione di coralli col metodo del 23~ Geologia Applicata ed Idrogeologia, 6, 105-112. CuccI, L. & CINTI, F. R. 1998. Regional uplift and local deformation recorded by the Quaternary marine terraces on the Ionian coast of northern Calabria (Southern Italy). Tectronophysics, 292, 67-83. DAI PRA, G. 1978. Linee di costa tirreniane del litorale laziale dal Fiume Tevere a Tarquinia. Geologia Applicata ed Idrogeologia, 13, 1-10. 1982. The Late Pleistocene marine deposits of Torre Castiglione. Geografia Fisica e Dinamica Quaternaria, 5, 115-119.
108
P. B O R D O N I & G. VALENSISE
- - 1 9 9 5 . Geomorfologia e stratigrafia dei depositi del Pleistocene e dell'Olocene nel Lazio meridionale. In: Lazio meridionale. Sintesi delle ricerche geologiche multidisciplinari. ENEA Internal Report, 36-51. & ARNOLDUS-HUYZENDVELD, A. 1984. Lineamenti stratigrafici, morfologici e pedologici della fascia costiera dal flume Tevere al flume Astura (Lazio, Italia Centrale). Geologica Romana, 23, 1-12. - - - & HEARTY, P. J. 1988. I livelli marini pleistocenici del golfo di Taranto. Sintesi geocronostratigrafica e tettonica. Memorie della Societd Geologica Baliana, 41,637-644. & STEARNS, C. E. 1977. Sul Tirreniano di Taranto. Datazioni su coralli con il metodo del 23~ Geologica Romana, 16, 231-242. , MIYAUCHI, T., ANSELMI, B., GALLETTI, M. & PAGANIN, G. 1993. Etfi dei depositi a Strombus bubonius di Vibo Valentia Marina (Italia Meridionale). I1 Quaternario, 6, 139-144. DAMIANI, A. V. 1970. Terrazzi marini e sollevamenti differenziali fra i bacini del Lao e del Corvino (Calabria Settentrionale). Bollettino della Societd Geologica Italiana, 89, 145-158. D'ANGELO, U. & VERNUCCIO, S. 1994. I terrazzi marini quaternari della estremitfi occidentale della Sicilia. Memorie della Societd Geologica Italiana, 51, 585-594. DEPERET, C. 1918. Essai de coordination chronologique g6n6rale des temps quaternaires. Cornptes Rendus de l'Acad~mie des Sciences, 167, 418-422. DI GERONIMO, I. 1970. Geomorfologia del versante Adriatico delle Murge di SE (zona di Ostuni, Brindisi). Geologica Romana, 9, 47-58. DI GRANDE, A. & SCAMARDA, G. 1973. Segnalazione di livelli a Strombus bubonius Lamark nei dintorni di Augusta (Siracusa). Bollettino dell'Accademia Gioenia di Scienze Naturali in Catania, 11, 157-172. D1 STEFANO, G., PETRONIO, C., SARDELLA, R., SAVELLONI, V. & SQUAZZINI, E. 1992. Nuove segnalazioni di brecce ossifere nella costa fra Castro Marina e Otranto (Lecce). I1 Quaternario, 3-10. DuMas, B. & RAFFY, J. 1993. Validit6 et pr6cision de la mesure des mouvements verticaux par r&6rence aux pal6oniveaux marins: exemples en Calabre mbridionale. Bulletin de l'Institut de Gdologie du Bassin d'Aquitaine, 53, 141-150. - - - & - - 1 9 9 6 . Enregistrement gbomorphologique de maxima glacio-eustatiques dans la r6gion soulev6e de Soverato (Italie du Sud). Bulletin de la Socidt~ Gdologique de Fr_Tnce, 167, 285-293. , GUI~RI~MY, P., HAONG, C. T., LHENAFF, R. & RAFFY, J. 1991. Gisement et rivages tyrrhbniens de Vibo Marina (Italie du Sud): datation 23~ 234U, soul+vement diff&entiel de la Calabre m6ridionale. Comptes Rendus de l'Acad(mie des Sciences S6rie II, 312, 785-791. , HEARTY, P. J., LHENAFF, R. & RAFFY, J. '1988. Morphometric analyses and amino acid geochronology of uplifted shorelines in a tectonic 5,
region near Reggio Calabria, Southern Italy. Palaeogeography, Palaeoclimatology, Palaeoecology, 68, 273-289. , LHENAFF, R. & RAFFY, J. 1987a. Rates of uplift as shown by raised Quaternary shorelines in Southern Calabria (Italy). Zeitschriftffir Geomorphologie, Supplementband, 63, 119-132. & - - 1 9 8 7 b . D6converte fi 157m ~t'altitude, pr6s de Nocella (Italie du Sud), du plus haut rivage tyrrh6nien connu fi ce jour. Comptes Rendus de l'Acad~mie des Sciences Sdrie II, 305, 139-142. & - - 1 9 9 5 . Reconstitution morphometrique d'un dispositif de pal6orivag6s soulev6s dans un relief de dissection en Locride (Calabre ionienne, Itatie). Geodinamica Acta, 8, 185-198. FABIANI, R. 1941. Tracce di Tirreniano (strati a Strombus bubonius Link) entro la cittfi di Palermo. Bollettino della Societd di Scienze Naturali ed Economiche di Palermo, 23, 3-8. GIGNOUX, M. 1913. Les formations marines plioc~nes et quaternaires de l'Italie du sud et de la Sicilie. Annales de l'Universitd de Lyon, 36. GLIOZZI, E. 1988. I terrazzi marini del Pleistocene superiore della penisola di Crotone. PhD thesis, University of Naples. & MALATESTA, A. 1982. A Megacerine in the Pleistocene of Sicily. Geologica Romana, 21, 311-395. GORTANI, M. 1950. Gli studi sui terrazzi fluviali e marini d'Italia dal 1938 al 1948. Bollettino della Societd Geografica Italiana, III(6), 297-322. GUEREMY, P. 1972. La Calabre centrale et septentrionale. Guide d'excursion gdomorphologique. Travaux de l'Institut G6ographique de Reims, 10, 1-128. - - 1 9 8 0 . Terrasses marines et n~otectonique dans le fosse de Catanzaro et la p6ninsule de Crotone (Calabre, Italie). In: Niveaux marins et tectonique quaternaires dans l'aire mdditerran~enne. Centre National de la recherche scientifique, Universit6 de Paris I, 231 254. HEARTY, P. J. 1986. An inventory of last interglacial (sensu lato) age deposits from the Mediterranean Basin: a study of isoleucine epimerization and U-series dating. Zeitschrift ffir Geomorphologie, Supplementband, 62, 51-69. - - 1 9 8 7 . New data on Pleistocene of Mallorca. Quaternary Science Reviews, 6, 245-257. & DAI PRA, G. 1985. Aminostratigraphy and uranium dating of Quaternary shorelines in the Puglia region of southern Italy. Proceedings of the 5th International Coral Reef Congress, Papeete, Tahiti, May 1985, 163-169. & - - 1 9 8 6 . Aminostratigraphy of Quaternary marine deposits in the Lazio region of Central Italy. Zeitschr(ft fiir Geomorphologie, Supplement bank 62, 131-140. & - - 1 9 8 7 . Ricostruzione paleogeografica degli ambienti litoranei quaternari della Toscana e del Eazio settentrionale con l'impiego dell'aminostratigrafia. Bollettino del Servizio Geologico Italiano, 106, 189-224.
-
-
-
-
-
-
DEFORMATION
-
OF T H E 125ka M A R I N E T E R R A C E I N I T A L Y
& 1992. The age and stratigraphy of Middle Pleistocene and younger deposits along the Gulf of Taranto (Southeast Italy). Journal of Coastal Research, 8, 82-905. , BONFIGLIO, L., VIOLANTI, D. & SZABO, B. J. 1986b. Age of Late Quaternary marine deposits of Southern Italy determined by aminostratigraphy, faunal correlation, and uranium-series dating. Rivista Italiana di Paleontologia e Stratigrafia, 92, 149-164. , MILLER, G. H., STEARNS, C. E. & SZABO, B. J. 1986a. Aminostratigraphy of Quaternary shorelines in the Mediterranean basin. Geological Society of America Bulletin, 97, 850-858. HIPPOLYTE, J-C., ANGELIER, J. & ROURE, F. 1994. A major geodynamic change revealed by Quaternary stress patterns in the southern Apennines (Italy). Tectonophysics, 230, 199-210. IANNONE, A. & PIERI, P. 1979. Considerazioni critiche sui 'Tuff calcarei delle Murge'. Nuovi dati litostratigrafici e paleoambientali. Geografia Fisica e Dinamica Quaternaria, 2, 173-86. ISSEL, A. 1914. Lembi fossiliferi quaternari e recenti osservati nella Sardegna meridionale dal prof. D. Lovisato. Rendiconti Regia Accademia dei Lincei, 23, 759 750. LA MARMORA, A. 1857. Voyage en Sardaigne. Bocca Imprimeur Royal, Torino. LA VOLPE, G. & PRINCIPI, C. 1994. II Monte Vulture. Guida all'escursione generale precongressuale. 77th Meeting Societfi Geologica Italiana, Bari, 23 Septembe~l October 1994. LIRER, L., PESCATORE, T. & SCANDONE, P. 1967. Livello di piroclastiti nei depositi continentali post tirreniani del litorale Sud tirrenico. Ani dell'Accademia Gioenia delle Scienze Naturali in Catania, serie. 6, 18, 85-115. MAGRI, G. & ZEZZA, F. 1971. I depositi dunari della costa Adriatica tra Monopoli (Bari) e Torre S. Sabina (Brindisi) in rapporto alle oscillazioni del livello marino. Geologia Applicata ed Idrogeologia, 5, 49 54. MALATESTA, A. 1985. Geologia e paleobiologia dell'era glaciale. La Nuova Italia Scientifica, Firenze. & ZARLENGA, F. 1986. Evoluzione paleogeografico-strutturale plio-pleistocenica del basso bacino romano a Nord e a Sud del Tevere. Memorie della Societd Geologica Baliana, 35, 75-85. MALINVERNO, A. & RYAN, W. B. F. 1986. Extension in the Tyrrhenian Sea and shortening of the Apennines as result of arc migration driven by sinking of the lithosphere, Tectonics, 5, 227 245. MASTRONUZZI, G. & SANSO, P. 1998. Late Quaternary evolution of the Adriatic coastline of Southern Apulia (Italy). Journal of Coastal Research, in press. - - , PALMENTOLA,G. & RICCHETTI,G. 1989. Aspetti della evoluzione olocenica della costa pugliese. Memorie della Societd Geologica Italiana, 42, 287-300. MESOLELLA,K. J., MATTHEWS,R. K., BROECKER,W. S. & THURBER, D. L. 1969. The astronomical theory of climatic changes: Barbados data. Journal of Geology', 77, 250-274. -
-
-
109
MILLI, S. & ZARLENGA, F. 1991. Analisi di facies dei depositi tirreniani (Duna Rossa) affioranti nell'area di Castel Porziano-Pomezia (Roma). Una revisione ambientale. II Quaternario, 4, 233-248. M1YAUCHI, T., DAI PRA, G. & LABINI, S. S. 1994. Geochronology of Pleistocene marine terraces and regional tectonics in the Tyrrhenian coast of South Calabria. II Quaternario, 7, 17-34. MONCHARMANT-ZEI, M. 1957. Foraminiferi e molluschi di un livello tirreniano pr6s de Nova-Siri Scalo (Matera). Bollettino della Societd dei Naturalisti in Napoli, 66, 53-68. MOSTARDINI, F. & MERLINI, S. 1986. Appennino centro-meridionale: sezioni geologiche e proposta di modello strutturale. Memorie della Societd Geologica Italiana, 35, 177-202. NEBOIT, R. & REYNARD, P. 1973. t~tude g~omorphologique et 6cologique du gisement quaternaire de la terrasse de 'Piano San Nicola' pr6s de Nova Siri Scalo (Italie du Sud). Geobios, 6, 291-305. OZER, A., DEMOULIN, A. & DAI PRA, A. 1987. Les indices morphologiques t6moins de la stabilit6 tectonique de la bordure littorale du Lazio m6ridional (Italie). Zeitschrift ffir Geomorphologie, Supplementband, 63, 103-117. PALIERI, L. & SPOSATO,A. 1988. Segnalazione di fauna a Strombus Bubonius nell'area compresa tra Tarquinia e Montalto di Castro. Il Quaternario, 1, 161-166. PALMENTOLA,G., CAROBENE, L., MASTRONUZZI,G. & SANSO, P. 1990. I terrazzi marini pleistocenici della penisola di Crotone. Geografia Fisica e Dinamica Quaternaria, 13, 75-80. PANTOSTI, D. & VALENS1SE, G. 1990. Faulting mechanism and complexity of the November 23 1980, Campania-Lucania earthquake, inferred from surface observations. Journal of Geophysical Research, 95, 15 319-15 341. --, SCHWARTZ, D. P. & VALENSISE, G. 1993. Paleoseismology along the 1980 Irpinia earthquake fault and implications for earthquake recurrence in the southern Apennines. Journal of Geophysical Research, 98, 6561-6577. PAREA, G. C. 1986. I terrazzi marini tardo-pleistocenici del fronte della catena appenninica in relazione alla geologia dell'avanfossa adriatica. Memorie della SocieM Geologica Italiana, 35, 913-936. PATA, O. 1947. Su di un nuovo giacimento a Strombus bubonius LMK presso Vibo Valentia. Atti della Societd Toscana di Scienze Naturali, 54, 159-166. PATACCA, E., SCANDONE, P. & MELETTI, C. 1997. Variazioni di regime tettonico nell'Appennino meridionale durante il Quaternario. Proceedings ofAIQ UA Meeting, Parma, 25-27 February 1997, 21-22. PIERI, P. 1975. Geologia della citt/t di Bari. Memorie della Societd Geologica Italiana, 14, 379-407. RADTKE, U. 1986. Value and risks of radiometric dating of shorelines - geomorphological and geochronological investigations in central Italy, Eolian Islands and Ustica (Sicily). Zeitschriftffir Geomorphologie, Supplementband, 62, 167-181. RICCHETTI, G., CIARANFI, N., LUPERTO-SINNI, E., MONGELLI, F. & PIER[, P. 1988. Geodinamica ed
110
P. B O R D O N I & G. V A L E N S I S E
evoluzione sedimentaria e tettonica dell'avampaese apulo. Memorie della Societd Geologica Italiana, 41, 57-82. ROMANO, P., SANTO, A. & VOLTAGGIO, M. 1994. L'evoluzione geomorfologica della pianura del Fiume Volturno (Campania) durante il tardo Quaternario. II Quaternario, 7, 41-56. RUGGIERI, G. & UNTI, M. 1974. Pliocene e Pleistocene nell'entroterra di Marsala. Bollettino della Societd Geologica Italiana, 93, 723-733. , BUCCHERI, G. & RENDINA, M. 1968. Segnalazioni di Tirreniano fossilifero a Trapani. Rivista Mineraria Siciliana, 112-114, 216-219. Russo, F. & BELLUOMINI, G. 1992. Affioramenti di depositi marini tirreniani in destra del Fiume Sele (Campania). Bollettino della Societd Geologica ltaliana, 111, 25-31. SEGRE, A. G. 1949. Tracce di morfologia subaerea sul fondo marino litoraneo del Lazio meridionale. Historia Naturalis, 3, 1-3. - - 1 9 5 7 . Prima relazione del rilevamento geoIogico del Foglio 171. Bollettino Servizio Geologico Italiano, 78. SELLI, R. 1962. Le quaternarie marin du versant Adriatique-Ionien de la phninsule italienne. Quaternaria, 6, 391-413. - - 1 9 7 7 . Excursion in C a l a b r i a - General geological setting of the Crotone-Catanzaro area. Giornale di Geologia, 42, 391-413. SEST1N~,A. 1930. La Piana di Sibari. Proceedings of the XI Congress of Italiano Geographers, 2, 124-131. TREVlSAN, L. 1943. Problemi relativi all'epirogenesi e all'eustatismo nel Pliocene e nel Pleistocene della Sicilia. Atti della Societd Toscana di Scienze Naturali, 51, 11-33.
VALENSISE, G. & D'ADDEZIO, G. 1994. II contributo della geologia di superficie all'identificazione delle strutture sismogenetiche della Piana di Gioia Tauro. Istituto Nazionale di Geofisica Internal Report, 559. -& GUIDOBONI, E. 1995. Verso nuove strategie di ricerca: zone sismogenetiche silenti o silenzio delle fonti? In: BoscHI, E., GUIDOBONI, E., FERRARI, G., VALENSISE,G. &; GASPERINI, P. (eds) Catalogo dei Forti Terremoti in Italia dal 461 a.C. al 1980. Istituto Nazionale di Geofisica and SGA, Bologna (also CD-ROM). -& PANTOSTI, D. 1992. A 125 Kyr-long geological record of seismic source repeatability: the Messina Straits (southern Italy) and the 1908 earthquake (MS 71/2). Terra Nova, 4, 472-483. & --1999. Seismogenic faulting, moment release patterns and seismic hazard along the central and southern Apennines and the Calabrian Arc. In: gAI, G. B. t~; MARTINI, I. P. (eds) Anatomy of a Mountain: the Apennines and Adjacent Mediterranean Basins. Chapman & Hall, London, in press. VEEH, H. H. 1966. 23~ and 234U/238U ages of Pleistocene high sea-level stands. Journal of Geophysical Research, 71, 3379-3,386. VEZZANI, L. 1967. I depositi Plio-Pleistocenici del litorale ionico della Lucania. Atti dell'Accademia Gioenia delle Scienze Naturali in Catania, serie 6, 18, 159-179. WESTAWAY, R. 1993. Quaternary uplift of southern Italy. Journal of Geophysical Research, 87, 21 741-21 772.
Neotectonic implications of a lineament-coplanarity analysis in southern Calabria, Italy YVES CORNET 1 & ALAIN
DEMOULIN
1'2
1Department of Physical Geography and Quaternary Geology, University of Li@e, Sart-Tilman, Bdt. Bll, B-4000 Li@e, Belgium (e-mail:
[email protected]) 2Research Associate NFSR (e-mail:
[email protected]) Abstract: A synthesis of the geodynamical interpretations proposed in the literature shows
that the respective role of different possible driving mechanisms in the Neogene and Quaternary tectonic evolution of southern Calabria is still under discussion. Owing to different methodological approaches and scales of observation, the same holds true for marine terrace levels recognized in the area of Reggio Calabria. Their number, elevation, spatial correlation and age, as well as their tectonic significance, are still debated, resulting in some uncertainty about the timing and rate of the regional uplift and its local variations. To complete the mapping of fault zones potentially dislocating the marine terraces, a threedimensional (3D) morphometrical analysis of lineaments has been carried out. From a dataset composed of linear structures recognized on an illuminated digital elevation model (17km x 18kin in size) centred on Reggio Calabria, the method aims at identifying lineaments pertaining to a single plane. Assuming that faults may be assimilated to planes at the local scale, the coplanar lineaments allow the 3D attitude and location of fault or fracture planes to be determined. The planes retained as tectonically significant (on the dip value and number of coplanar lineaments) have been successfully compared with the existing faults as well as with geomorphological data (long river profiles) of southern Calabria. Beyond mapping the main faults of the study area, the coplanarity analysis highlights the existence of conjugate faults within the NW-SE fault zone of Terreti-Monte Embrisi. It also locates some E-W striking faults which were previously known only at the outcrop scale. The main object of this research was the analysis of a three-dimensional (3D) set of lineaments measured by use of a digital elevation model (DEM) to determine the regional fracture pattern of the southern Calabrian upper crust. This approach has been suggested by Eliason (1992), who analysed the coplanarity of lineaments extracted from topographical maps so as to map fracture planes along a complex segment of the San Andreas Fault, including its intersection with the Sargent Fault. Assuming that the lineaments are straight features, the method is based on the 3D location of their tips within an orthogonal reference adjusted to the U T M (Universal Transverse Mercator) grid. This allows four planes to be computed from the four possible triplets of tips yielded by each vector pair. When these four planes are identical at a given tolerance, the two vectors are called 'coplanar'. Planes defined by at least four coplanar lineaments have been supposed to be tectonically significant. Such an analysis allows a 3D definition of plane attitude (strike and dip) and location, better than that given by the traditional diagrams showing only the map-view (planimetric) orientation of lineaments. The quality of the results gained by this lineament-coplanarity analysis strongly relies on a clear definition of what a 'lineament' is, which is closely related to its geomorphological
expression. According to Koch & Mather (1997), lineaments are expressed by 'geomorphological features such as aligned ridges and valleys, straight drainage channel segments, linear scarp faces, or pronounced breaks'. No length criterion can be given to define a lineament. If we assume that the lineaments are basically developed in relation to planar geological discontinuities, long and short lineaments will be equally useful to our analysis. The former will mainly influence a plane's strike definition whereas the latter (which are often steep reaches of low-order streams) will predominantly determine a plane's dip. Especially for the short features, the coplanarity method precisely infers their tectonic significance from their common affiliation to given planes. Beyond the accuracy of the data acquisition (discussed below), a further limitation of this method lies in the assumption of homogeneity of geological fracture and fault planes. This restricts its use to areas of about 15 km x 15 km. Such areas will generally include fault segments no longer than 10-15km, which may reasonably be assumed to be plane within the upper 2-3 km of the crust. The 3D analysis of lineament coplanarity is restricted to areas of steep slope gradients, as high relief energy is required to satisfactorily determine the third dimension (dip) of the fracture planes. The method is also best applied to
CORNET, Y. & DEMOULIN,A. 1998. Neotectonic implications of a lineament-coplanarity analysis in southern Calabria, Italy. In: STEWART,I. S. & VITA-FINZI,C. (eds) CoastalTectonics,Geological Society, London, Special Publications, 146, 111-127.
112
Y. CORNET & A. DEMOULIN
areas of intense tectonic activity. The Reggio Calabria Basin in southern Italy satisfies both conditions, although with the exception of the major faults (Armo, Calanna, Martiniti, Scilla Faults), the fault systems defined by Ghisetti (1981a) and Atzori et al. (1983) are not always expressed by unequivocal morphological features. Nevertheless, the Quaternary uplift of Southern Calabria, estimated at 1 mm/year by Westaway (1993), has produced very steep slopes (the Monte Alto, the highest summit of the Aspromonte at 1955m above sea level, is only 20km distant from the Straits of Messina). Furthermore, the morphology of the western slopes of the Aspromonte results from the combined influence ofeustatic variations and regional tectonic uplift, therefore making the distinction between sea-cliffs, fault scarps and even faultrelated shorelines often difficult. A comparison between the geological map ofAtzori et al. (1983) and the upper Quaternary coastlines mapped by Dumas & Raffy (1993) in the Diminiti-San Nicola area clearly illustrates this difficulty, the marine terraces either being limited in their landward extension by a pre-existing fault scarp or being dislocated after their formation. The coplanarity method can help to partly solve the problem by focusing exclusively on planar features, which are mostly of tectonic origin. In this paper, the contribution of the lineamentcoplanarity analysis to the structural mapping of southern Calabria and to the understanding of its neotectonic evolution will be discussed. The results of this morphometric method are also compared with data obtained by fieldwork and more traditional remote-sensing techniques.
The study area Tectonic
setting
The study area is located in the CalabroPeloritan Arc, which is the frontal sector of the Kabilo-Calabrian Chain formed by the Eocene delamination of the European margin of the Tethys (Fig. 1). This structural domain was thrust over the Apenninic-Maghrebian Chain from Late Oligocene to Early Miocene times, probably in relation to the clockwise rotation of the Corso-Sardinian Block (Finetti et al. 1996; Lentini et al. 1996). Both chains constitute the orogenic domain that overlies the External Thrust System, bounding to the south and southeast the Apulian Foreland, the Pelagian Foreland (both belonging to the African Plate), respectively, and the Ionian Basin located between them.
The driving force for this deformation was the post-Serravalian opening of the Tyrrhenian Basin, probably as a result of the subduction of the Ionian oceanic lithosphere slab (Malinverno & Ryan 1986). The detachment of this slab 0.7 • 106 years ago probably stopped the subduction, thus accelerating the extension of the Calabro-Peloritan arc (Westaway 1993). This opening of the Tyrrhenian Basin produced the eastward motion of the Apenninic Arc in Central Italy and the southeastward advance of the orogenic front in Calabria. The Plio-Pleistocene extension responsible for the formation of the Reggio Calabria and Gioia Tauro Basins probably was also related to the same geodynamics (Finetti et al. 1996). The southward roll-back of the Ionian subduction hinge explains the uplift of the thick and relatively rigid Apulian area (Doglioni et al. 1996). The convergence of the orogenic front on one hand and the presence of the African continental margin and the Ionian Basin on the other hand give rise to the 'spreading bending' of the chains (Lungo 1988; Lungo et al. 1991), resulting in the expansion of the orogen's outer arc and the development of transform faults (e.g. the Palinuro and Vulcano-Sisifo Faults) that compensate for the variation of motion velocity along the orogenic belt (Finetti et al. 1996; Neri et al. 1996). According to Westaway (1993), the overall extension rate of the Messina Strait is c. >0.1 mm/year since the Tortonian. However, it was probably slower from the Tortonian to 0.9-0.7 Ma BP; after which it increased from c. 0.1 to 0.2-0.3 mm/year when vertical slip rate on the active normal faults of the Aspromonte range increased. Surprisingly, in spite of the strong earthquakes that have shaken the area during the last two centuries (Bottari & Lo Giudice 1987), no clear geomorphological evidence of Plio-Pleistocene dislocation has been found, except for a few very local signs of Pleistocene deformation (Ricchetti & Richetti 1991, 1996; Dumas & Raffy 1993; Miyauchi et al. 1994). Many workers (e.g. Ghisetti 1979; Atzori et al. 1983; Ghisetti 1981a, 1981b, 1984; Ghisetti & Vezzani 1982; Ott d'Estevou et al. 1987; Vachard et al. 1987) have nevertheless collected numerous structural data that indicate Plio-Pleistocene brittle deformation of the upper crust of southern Calabria. Ghisetti (1981a) suggested that, independent of their strike, the dislocations along the main fault systems may be coeval, developed under an extensional stress field with horizontal o-2 and o-3 axes of nearly the same magnitude. Ott d'Estevou et al. (1987) recognized three main fault systems striking NE, NW and ENE, and two secondary systems striking N and E. According to them, the stress-field evolution is
LINEAMENT-COPLANARITY ANALYSIS, SOUTHERN CALABRIA
113
Fig. 1. Simplified geological map of the study area with the main faults after Ott d'Estevou et al. (1987). I, Upper Pleistocene and Holocene deposits of the coastal plain; II, Tortonian and Plio-Pleistocene deposits; III, crystalline basement; IV, unmapped geology; V, main faults. 1-3, topographical cross-sections of Fig. 2. The inset sketches the geodynamical map of the central Mediterranean Sea after Finetti et al. (1996). 1, Tyrrhenian Basin; 2, Kabilo-Calabride Domain; 3, Apenninic-Maghrebian Chain; 4, Southern Apennines; 5, Pelagian Foreland; 6, Ionian Abyssal Plain; 7, Apulian Foreland; PF, Palinuro Fault; VF, Vulcano Fault. characterized by three main phases: (1) an Early Pliocene E - W extensional phase; (2) a late Early Pliocene N N E - S S W to N E - S W compressive phase; and (3) a Pleistocene E - W to E N E WSW extensional phase with a minor component ofmultidirectional compression (recorded only at the outcrop scale). From macroseismic data (Bottari & Lo Giudice 1987), the present-day tectonic activity on the Calabrian side of the Straits of Messina is generally attributed to N E - S W striking faults. In the
Straits of Messina, the focal mechanism of the 28 December 1908 Messina earthquake (Bottari and Lo Giudice 1987) corresponded to a normal dip-slip motion with a small dextral strike-slip component on a N N E - S S W striking fault. As for the Reggio Calabria earthquake of 16 January 1975, it results o f a sinistral motion along a E N E WSW fault plane (Bottari & Lo Giudice 1987). Whereas the latter event responded to a stress regime with o1 and o-3 respectively striking N N E and ESE, the former one developed within an
114
Y. CORNET & A. D E M O U L I N
:,,,,,.
%
o~
LINEAMENT-COPLANARITY
~.=>
ANALYSIS,
SOUTHERN
"~=~
.o ~ =
o~"~ 9 o ~.~ o o . O~. ~ ~ ~ o
=-~= ~ ~
-~ 'C = ~
9
0a .,..~ ...~
~_~ eO
-~
;~
a 0
.,
>.>-,
~
.~.~
~~ m
9~
~_~
0
~-~ .g=>~
~,-~I
~
Oa
~ ~
t--:. ,.--ZOO
~
~ ~o
~.~.~
,a::l
"~cs~
0a
~
~o=
~a,.--
- >,
~- =~
~ = ~
~,o~
~,>
~
= o
=~
.=~
,-'.~
c~
,-----,
= ~a ~.-~ o , ~ =
~'a
~.~
"~
Ca
Ca
9
~,.~., ,~ .~ ~
~,~- o o = ~
" ,~2)
(D
,~ o
~
t,%
~
0:8
*-~
8
ca
_Y~ " ~
_p.
~-~c~
;;~
~
~
=~-=g'= . =_
9
~
,~
I
U ~ g" Y:
~.= ~ . o ~
*" ~a
,.~
I::I
9
~.-~ I::I ..~
~
.~
~
~
,..
.
"CI
,-~
ca
ca
0
1: l"~ ca ---J o.0 e.i_,
~
CALABRIA
115
116
Y. CORNET & A. DEMOULIN
extensional stress field of similar o3 orientation. On the Ionian side of the Aspromonte, the Ferruzzano event of 16 March 1978 was associated with a sinistral strike-slip motion along a N-S striking fault, compatible with a NW-SE compression.
G e o m o r p h o l o g i c a l setting." m a r i n e terraces a n d f a u l t scarps
Within this extensional context, the Pleistocene uplift of the Calabro-Peloritan Arc is demonstrated by several structural and geomorphological studies (Ghisetti 1981b; Dumas & Raffy 1993; Westaway 1993; Miyauchi et al. 1994). On the basis of the elevation distribution and some age determinations of marine terraces, Westaway (1993) has proposed a geodynamical model that explains the uplift as the isostatic response to the detachment of the old subducting Ionian slab and its replacement by an asthenospheric bulge located under Calabria. He suggested that the mean regional uplift rate seems to be the same (c. 1 ram/year) all over Southern Calabria since 0.7MaBP. Furthermore Westaway (1993) speculated that this rate was probably negligible (e. 0.07-0.1 ram/year) from 5.3 to 0.7MaaP (Table 1), thus explaining the very large extension of the upper terraces of the Piano Don Rocco (600m), Piani di Melia and Piano Chiusa (600m), Bosco di Moio (700m), Piani d'Aspromonte (1120m), Campi di Reggio (l150m), Campi di Santa Agata (1200m), Portella Zagaria (ll00m) and Campo San Antonio (1000 m) - Monte Embrisi (1050m) (Fig. 1). These two last levels are probably parts of a surface dislocated by the Armo normal fault. They are covered by bioclastic sands described by Barrier (1987), who dated them to the later part of the Early Pleistocene: they would thus correspond to the onset of the uplift of the area. Moreover, the highest marine deposits have been described by Barrier (1987) in the area ofPietra di Molino and Embrisi (Campo San Antonio). In Pietra Molino (1000 m), transgressive, sandyconglomeratic coastal deposits are covered by deeper marine sediments. The Early Pleistocene age of the sands is testified by the presence of Globorotalia inflata, Pseudamussium septemradiatum and Arctica islandica. In Monte Embrisi (1051m) and Case Embrisi (990m), shallowwater carbonates have also been dated from the Lower Pleistocene by their fauna (Barrier 1987; Montenat & Barrier 1987). The observations of Westaway (1993) between Scilla and Monte Alto led him to calculate local
components of uplift rate linked to faulting in the area. In particular, he interpreted locally varying uplift rates as the consequence of movements of the Scilla, Calanna and Armo faults and rotation of the intervening crustal blocks. According to Westaway, such rotations could be responsible for local uplift rates of 0.31 mm/year in the footwall of the Scilla Fault and 0.67 mm/year in the footwall of the Armo Fault, this latter area thus having been raised at a combined (isostatic + fault-related) rate of 1.67mm/year. However, palaeomagnetic and isotopic dating of recent deposits (Aifa et al. 1987; Fontes et al. 1987) suggests that uplift rate could reach 2.7 mm/year during the last 30 000 years and at least 1 mm/year since 0.9 Ma aP. Dumas & Raffy (1993) recognized 26 marine terrace levels between the present-day sea level and an altitude of 556m, the spatio-temporal distribution of which allowed them to propose an uplift rate of 1.02-1.19m/year since 120kaaP. Miyauchi et al. (1994), however, described only 12 levels between present-day sea level and an altitude of 1360 m and concluded a high uplift rate in the Early Pleistocene, decreasing during the Mid-Pleistocene and again increasing in the Late Pleistocene. In agreement with Cosentino & Gliozzi (1988), they also suggest a N to NW tilting of the Calabro-Peloritan Arc. Unfortunately, field data from different workers (Ghisetti 1981b; Dumas & Raffy 1993; Westaway 1993; Miyauchi et al. 1994) are often contradictory (Table 1), mainly as a consequence of lacking a clearly defined morpho-sedimentological reference datum for the marine platform elevation determination. The only paper that gives such a definition is that of Dumas & Raffy (1993), who used the inner edge of the marine platform as the datum. This definition is of practical use in areas of coherent sediments or bedrock where marine platforms would form as erosional features only during periods of rising sea level (Westaway 1993). However, in areas ofunlithified deposits, terraces may have formed during regressive episodes, making the definition of the palaeo-shorelines more elusive. In fact, the continental or marine character of the deposits must be demonstrated (Dumas & Raffy 1993), which is difficult, particularly in areas such as Villa San Giovanni, Catona, Gallico and Archi, where the rate of fluvial sediment discharge is high. Indeed in this area, Miyauchi et al. (1994) recognized an alluvial fan morphology where Dumas & Raffy (1993) have mapped palaeo-shorelines penetrating inside the Catona and Gallico valleys. According to Dumas & Raffy (1993), some of the platforms are also cut into folded colluvial deposits. Moreover, numerous surfaces near
LINEAMENT-COPLANARITY ANALYSIS, SOUTHERN CALABRIA Favazzina, Scilla, Ferrito and the Gullina river terminate against a fault scarp in their upper parts (Atzori et al. 1983). In such circumstances, the distinction between fault scarp and ancient cliff becomes almost impossible. However, Dumas et al. (1993) acknowledge only one case of a fault scarp having possibly determined the location of a cliff. Here at Rosali, they described faulted marine deposits, but they did not observe any dislocated platforms along the new Archi-Orti road in the Gullina river area. It should also be stressed that in some cases the uplift rate estimates of Westaway (1993) are biased by erroneous aerial photograph interpretations of Valensise & Pantosti (1992) on which they are based. For example, along the ScitlaNucillari-Piani d'Aspromonte line. Valensie & Pantosti (1992) identified morphologically a Eutyrrhenian marine terrace at 170 m above sea level, whereas coastal gravels actually rest on the crystalline basement at 80-90 m above sea level and are overlain by a 30-40 m thick sequence of colluvial deposits whose top surface gently climbs up to 170 m (Fig. 2). Finally, a further problem in correctly interpreting marine terrace sediments is the presence of large amounts of gravels and pebbles within the regional bedrock (Messina gravels, Tortonian pebble layers), which, in addition to having been reworked several times, are available for slope and fluvial transport and deposits as well as for coastal accumulation on marine terraces. This throws doubt on the identification of some marine terraces and highlights the difficulty of obtaining reliable terrace elevation data. In summary, it appears that well-dated deposits and unequivocally defined marine terrace elevation data are rather scant in the study area. Diverging points of view are still debated, most results having been obtained mainly or exclusively by height correlation of often ill-defined morphologies. However, more accurate dating and height definition of Late Pleistocene terraces may allow better estimates of the uplift rate of southern Calabria in the last 120000 years (Table 1). The intensity of Quaternary fault activity remains a matter of discussion, with some workers minimizing it (Dumas et al. 1993) whereas others present evidence of Early Pleistocene dislocation (Vachard et al. 1987; Westaway 1993; Miyauchi et al. 1994). Within this geodynamical and morphostructural context, this study involves the development, application and validation of a morphometric method that allows a multiscalar analysis of morphological features potentially caused by fracturing and faulting and appar-
117
ently not spatially correlated to be performed. Moreover, in the case of southern Calabria, by combining information from river terraces and long profiles the method can detect faults of varying orientations rather than simply those E-W structures that disrupt the N-S continuity of marine terraces.
The lineament-coplanarity analysis Cop&nar lineaments
The analysis of lineament coplanarity is intended to improve classical lineament maps by taking advantage of the fact that, though planimetrically non-collinear, some lineaments can, however, pertain to the same plane. The attitude (strike and dip) and location of planes including several lineaments may be determined if the lineament tips are defined in three dimensions. Several relative topographical positions may be occupied by linear morphological features which are coplanar. Figure 3 illustrates various cases in plan view. The cases shown in Fig. 3e-g are most evident, showing coplanar features (e.g. straight river reaches) that are recognized by a traditional 2D analysis too. Whereas the two first examples correspond to vertical planes, the third one may be accounted for by an infinity of planes rotating around the axis passing through the four endpoints of the lineaments, provided these points are aligned in the three dimensions. The examples shown in Fig. 3d, h and 1 represent parallel lineaments usually believed to mark individual fractures of the same system, but which can also pertain to a single plane more or less perpendicular to this system. The examples shown in Fig. 3d and 1, involving only stream segments, will generally yield almost horizontal planes against which the incising rivers have abutted. Although corresponding sometimes to thrust faults or more frequently to bedding planes, their significance will always remain highly questionable. The other examples (Fig. 3a-c and i-k) involve two or more lineaments (either stream or ridge segments or breaks of slope) of markedly different orientations. Traditionally, such lineaments are referred to different sets of fractures although they may also belong to a single inclined plane (this is obvious in Fig. 3a). For instance, the examples in Fig. 3i could be explained either as the surface trace of a fault or as a bedding plane on both sides of an incised valley, whereas that in Fig. 3j corresponds to a faceted spur identified by its linear edges.
118
Y. C O R N E T & A. D E M O U L I N
L I N E A M E N T - C O P L A N A R I T Y ANALYSIS, S O U T H E R N CALABRIA
119
Fig. 3. Various examples of lineament coplanarity. 1, Rectilinear river reach; 2, other linear features (ridge, break of slope, etc.); 3, river; 4, watershed. (See text for explanation.)
Database The lineament data have been collected in analogue form from the screen image of a digital elevation model ( D E M ) and immediately digitized. Unfortunately the Calabrian high relief results in a loss of accuracy of the generalized contour lines of the 1/25 000 topographical maps
(with a contour interval of 25m) which were used to produce the D E M . The interpolation of this D E M has been performed using the Hutchinson (1989) algorithm. Neglecting the effect of a variable angular height o f illumination, we have illuminated the D E M from the NE, SE, SW and N W with an angular height of 45 ~ ( M u r p h y 1993). A n y two perpendicular perspectives of
Fig. 2. Topographical cross-sections transverse to the Tyrrhenian coast of Calabria showing the interplay of marine terraces and fault scarps between Villa San Giovanni and Favazzina (see Fig. 1 for location). (a)/, More than 30 m thick, roughly stratified colluvial deposits (alternating silts and pebbles); at least one palaeosol has been observed. 2, Marine deposits resting on the crystalline basement (a, b) and hardly visible marine terrace inner edge (c). The heights have been measured with a barometric altimeter. Locations are approximate. 3, Casa Gulli-San Giovanni deposits. 4 and 5, Differential vertical motion on both sides of the Scilla Fault scarp. (b) 1, Height of the Eutyrrhenian marine terrace after Valensise & Pantosti (1992). 2, Height of the Nucillari wavecut platform. 3 and 4, Local Scilla fault footwall uplift (0.31 ram/year) v. local Calanna fault hanging-wall uplift (0.27 ram/year). Both values have to be added to the regional uplift rate of 1 mm/year. (c) 1, Marine deposits resting on the crystalline basement. 2, Hardly visible marine terrace inner edge on the crystalline basement, overlain by a more than 15 m thick, roughly stratified colluvial sequence (alternating silts and pebbles). The heights have been measured with a barometric altimeter. Locations are approximate. 3, Differential vertical motion on both sides of the Scilla fault scarp.
120
Y. CORNET & A. DEMOULIN
these four directions yield about 90% of the whole dataset, clearly showing that no additional direction of illumination is required. The final location of each lineament and position of their endpoints have been adjusted from their mapping on the differently illuminated images. Every lineament has been mathematically represented by a vector defined by the (x, y, z) cartographic coordinates of its tips, each of which can be indifferently considered as the startpoint or the endpoint of the vector. The lower hemisphere Schmidt-Lambert net of Fig. 4a shows the plunge distribution of the 928 mapped lineaments. This distribution is slightly anisotropic, more lineaments plunging in the western half of the stereogram, i.e. in the downstream direction of the drainage network of the study area. The mean lineament plunge is small, considerably less than 40 ~ with maximum values of about 50 ~ In the Fig. 4b, the lineament length is plotted against the plunge direction. Although no correlation
(a)
North
~',.-:.":
:..
i~!'~':": ~..z,
"..
,
.
.....
,,
'
..,
.
..r
,
k" C:"-~':';."
..
'
' ...
", ; "...,'.~..
(b)
928
exists between the two parameters, the longest lineaments are more frequently plunging westwards. This again is related to the orientation of the overall slope of the study area towards the Messina Straits.
Single coplanarity analysis Coplanarity analysis is based on a two-step processing of the lineament data. In a first step called 'single coplanarity analysis', all vectors have been combined two by two to determine whether they pertain to a single plane. Four plane equations can be computed for each vector pair by using the four possible combinations of three of their four tips. For each of these planes we further calculate the angle between its axial vector and the lineament, of which only one endpoint participates in its definition. In Fig. 5 the two combined vectors are defined by their endpoints, respectively labelled 1 and 2 for the first vector and 3 and 4 for the second. The axial vector of the plane containing points 1, 2 and 3 is marked N4. The angle between this vector and the lineament of which one tip (point 4) has been left aside is computed through the scalar product of these vectors. The two lineaments are called coplanar when the angle is equal to 90 ~ with a tolerance of 2 ~ for each of the four planes associated with the vector pair. At the end of this first step, a mean plane equation is statistically computed for each pair of coplanar vectors by applying the least-squares method to their 4 endpoints.
lineaments
2500 2000
E
,..
1500
.~ 1000 _j
. i
500
'*'.-
o L'2."' 0
.. .
, 9
. .~ . . . .
"
9
.,'.
r
"L'.":,.
.
. . ..,~
....' -
.:.
..,,.t
%"."
,, ".~
,~.
" . . : "':', -'.': ?.'. ' ; ' ? " . ~ 90
180
Direction
."
"~ 9
.., .
.
.'.'.~'.-~.~". .'3 .:,w
~ : . ~ . . ~ . ' ~ . " "; 270
360
in ~
Fig. 4. (a) Distribution of lineament plunges in a lower hemisphere Schmidt-Lambert stereonet plot. The radius of the dotted circle is 0.5, for a slope value of c. 49~ (b) Distribution of lineament length against plunge direction (counted from the north in a clockwise direction).
Fig. 5. Geometric principles of lineament coplanarity analysis.
LINEAMENT-COPLANARITY ANALYSIS, SOUTHERN CALABRIA
Multiple coplanarity analysis The second step of the processing consists of comparing the position of all the lineaments which are not involved in the original definition of a plane with the plane's location and attitude, to determine how many lineaments are actually contained in this particular plane. It is thus clearly a 'multiple' coplanarity analysis which is carried out on the basis of fixed distance and angle tolerances for all mean planes defined in the previous step of the processing. The distance tolerance is a parameter which allows a third or ith potential coplanar lineament to be detected when each of its endpoints is closer than a fixed value to the plane. This value (70 m in this case study) has been arbitrarily defined as the greatest distance measured between one mean plane and the furthest of the four points used to compute its equation in the first step of the analysis. In addition to satisfying the distance condition, a lineament must be adequately oriented before its identifier is attached to the plane. Again, this is verified by computing the angle Oi between the lineament and the axial vector of the plane and allowing it to depart from 90 ~ by a maximum of 2 ~. Having carried out this procedure for all planes, a new set of several coplanar lineaments is thus defined for each of them. A new plane equation is then adjusted by using the vector endpoints corresponding to all coplanar lineaments and reapplying the least-squares method. Finally, the set of improved plane equations resulting from the second step is submitted to a cleaning process to eliminate redundant cases.
Error propagation We have analysed the consequences of two error types. First, the error on the z coordinate of the lineament tips as a result of the inaccuracy of the DEM does not affect the results provided it is spatially continuous at the lineament scale, which seems to be generally the case. If this error was randomly distributed in xy space, however, the results would however be significantly changed. The second type of error results from subjectivity in mapping the lineaments and especially in positioning the lineament tips. We have experimentally demonstrated that this imprecision is significantly higher in the strike direction of the lineament than in the cross-direction. Simulating such an error on our dataset, we have produced a new data set with different x, y and z coordinates. Provided the imprecision in the cross-direction remains small, the results are unaffected. Finally, the geomorphological and geological validation
121
of the computed fracture planes independently confirms the quality of their positioning.
Results of the coplanarity analysis The results of the coplanarity analysis are presented in Figs 6 and 7. Despite the final cleaning, the processing of the 928 lineaments of the dataset has yielded 33 598 planes containing two or more coplanar lineaments. Obviously, only a small part of these planes are tectonically significant. The selection of the most probable fracture and fault planes will be conducted on the basis of the dip of the plane and the number of lineaments defining it. The polar diagrams of Fig. 6a show that, when considering the whole set of planes, the vast majority dip towards the Straits of Messina. Moreover, most of them are also almost flat-lying planes, with only 9% dipping more than 40 ~ and 3 % more than 60 ~ The gently sloping planes are generally computed from drainage segments directed towards the Straits and, though possibly corresponding to some thrust faults, they are mainly related to bedding planes. In the extensional context of southern Calabria, we have thus decided to remove all planes dipping less than 60 ~. The remaining steeply dipping planes show a strongly oriented distribution with two maxima for planes striking NW-SE and N-S and indifferently dipping on either side of their strike. Generally speaking, this strike clustering regionally conforms to the fault orientations recognized by Ott d'Estevou et al. (1987). However, the main NE-SW fault system identified by those workers is not represented in the diagram. The symmetrical distribution of the dip directions shows that several of the main fault systems of the Reggio Calabria basin are composed of conjugate faults or fractures. For instance, most of the NW-SE striking planes containing at least three lineaments dip either 40-60 ~ SW (e.g. the Terreti fault) or 80-90 ~ NE. Similar conjugate systems have already been described for various regions of the Southern Apennines and the Calabro-Peloritan Arc (Boschi et al. 1989; Pantosti & Valensise 1989; Valensise & Pantosti 1992; Pantosti et al. 1993; Valensise & D'Addezio 1994). The poles of planes successively containing 2-9 and >9 coplanar lineaments are represented in the lower hemisphere stereonet plots of Fig. 6b. They highlight the decrease of the plane dip average and the correspondingly increasing anisotropy of the distribution when a higher number of coplanar lineaments is imposed to keep a plane significant. The decreasing dip value is correlated with the predominance of the horizontal
122
Y. C O R N E T & A. D E M O U L I N
Fig. 6. (a) Distribution of dip orientations for planes of varying minimal dip. (Note the drastic decrease of the plane number for dip >30~ (b) Lower hemisphere Schmidt-Lambert stereonet plots of the computed planes containing different numbers of coplanar lineaments (2-9 and >9 as indicated in the upper-right corner of each plot). (Note that the x and y values of the last plot range from -0.5 to +0.5 (the corresponding dip is c. 41~
LINEAMENT-COPLANARITY ANALYSIS, SOUTHERN CALABRIA
123
Fig. 6. (continued) dispersion of lineament tips over the vertical one when the number of lineaments per plane increases. As in the first diagram of Fig. 6a, the centre of gravity of the set of poles in the plots of Fig. 6b is clearly shifted toward the east. This is again a consequence of the nature of the input data, chiefly related to the drainage system (Fig. 4).
Comparison of the fracture planes with existing faults In the following comparison, to maximize the probability that the mapped planes are really of tectonic origin, we shall retain only those containing more than three lineaments and dipping
more than 60 ~. They are 34 planes whose surface traces are mapped in Fig. 7, assuming that they are continuous throughout the area covered by the DEM. This map shows that, unlike the NW-SE and NE-SW fault systems recognized by Ott d'Estevou et al. (1987), their N-S system is hardly identified by the coplanarity analysis. This could be due either to the fracture planes pertaining to that system having too low a dip or, more probably, to the scarcity of morphologic evidence in the area covered by the DEM. However, some aligned flat-lying poles point to planes moderately dipping WNW, which could be assigned to this N-S system (Fig. 6b). Nevertheless, the location of most computed planes is fairly consistent with that of faults identified by Ghisetti (198 la), Atzori et al. (1983) and Ott d'Estevou et al. (1987). For instance, the
124
Y. CORNET & A. DEMOULIN
Fig. 7. Results of the coplanarity analysis applied to 928 lineaments in the Reggio Calabria area. An angular tolerance of 2~ and a distance tolerance of 70 m have been applied (see text for explanation). The surface trace of the 34 planes dipping more than 60~ and containing at least four lineaments is superimposed on a shaded DEM. 1, Faults after Ghisetti (1981b) and Ott d'Estevou et al. (1987); 2, planes determined by the coplanarity analysis, with dip sense and value; 3, coplanar lineaments; 4, river; 5, revived river downcutting.
fault zone of Terreti-Mosorrofa is clearly expressed by a suite of N W - S E striking planes dipping 63-82 ~ towards the southwest. According to the geomorphological observations of Miyauchi et al. (1994), this fault has been active since the formation of the Terreti marine terrace (at about 950 ka aP) and should be responsible for the 30 m high scarp dislocating the terrace. The N E - S W
striking faults of Oliveto-Monte San DemetrioCardeto (i.e. the eastern segment of the Armo fault zone) and Calopinace-Annunziata zone, which were active in the early, and perhaps also the mid-Pleistocene (Miyauchi et al. 1994), are also identified by the coplanarity analysis as a NW-dipping plane defined by several lineaments in the crystalline basement.
LINEAMENT-COPLANARITY ANALYSIS, SOUTHERN CALABRIA Three new sets of fracture planes are mapped, with strikes which correspond to those described by Ott d'Estevou et al. (1987) at the outcrop scale but which do not coincide with any known fault of regional character. One set comprises several NNW-SSE striking planes extending from Oliveto to Archi and from Orti to Cardeto and Monte Embrisi. The lineaments defining the fracture planes between Orti and Monte Embrisi are spatially scattered, justifying the mapping of the fault traces throughout the outcropping crystalline basement. East of Orti, one of these traces coincides with a major fault reported by Atzori et al. (1983) at the contact between the basement and the Tortonian deposits. In contrast, planes running from Oliveto to Archi are defined by lineaments concentrated mainly in the Archi-Vito-Monte Riga area and cutting Tortonian and Plio-Pleistocene deposits as well as the crystalline basement. Their mapping should thus be restricted to this area, where they coincide with a normal W-dipping fault mapped by Vachard et al. (1987) across the Annunziata valley. Moreover, not very far from there, in Pozzi, Atzori et al. (1983) identified similarly striking local faults. The almost E-W trending set of fractures located south of Oliveto-Monte Embrisi was also previously unrecognized. However, Atzori et al. (1983) mapped several E-W striking faults in the Plio-Pleistocene of the nearby Reggio Calabria basin, through which another E-W plane goes through. The general dip of the E-W planes, defined by lineaments that are predominantly located in the crystalline area, is towards the south. Finally, three ENE-WSW striking planes have been mapped in the areas of Terreti and Cataforio. Though unknown in the study area, this fracture system may again be related to similarly oriented structures described by Murphy (1993) just to the north.
Conclusion Although the fracture planes mapped in Fig. 7 represent only the very marginal part of the set of computed planes, the consistency of the results of the coplanarity analysis with existing faults demonstrates the reliability of this geomorphological technique. It appears to be a complementary approach to more traditional field and remote sensing studies, as the 3D treatment of morphological features generally permits faults ignored from field work to be mapped and provides the third dimension lacking in most lineament maps obtained from aerial photographs or satellite imagery.
125
In southern Calabria, apart from identifying some new faults which now wait to be confirmed by field data, the coplanarity analysis has highlighted the existence of conjugate faults within several of the previously recognized fault systems (especially the NW-SE system). Securing the third dimension, and thus fault dip information, also makes it possible to suggest the type of movement that can be linked to the development of the lineaments. For example, normal fault motion may be inferred for the 60-80 ~ dipping NNW-SSE striking faults (Fig. 7), with a possible right-lateral strike-slip component indicated by stream offsets southwest of Pozzi. As for the E-W striking, S-dipping faults south of Oliveto, they could represent transpressive features inherited from Early Pliocene NE-SW to NNE-SSW directed compression (Ott d'Estevou et al. 1987). Overall our results are furthermore consistent with the 'spreading bending' of the two chain domains proposed by Finetti et al. (1996) as well as with the recent E-W extensional phase inferred by Ott d'Estevou et al. (1987). Other geomorphological evidence, notably river channel profiles, further confirm the structural patterns inferred from the coplanarity procedure. For example, oversteepened stream reaches indicate that recent movements have taken place along the Terreti-Mosorrofa fault zone (Fig. 7). The influence of the NE-SW Monte San Demetrio Fault is also recorded by the profiles of rivers crossing it. Likewise, the existence of two N-S striking fault planes east of Cardeto and Monte Embrisi is confirmed by irregularities in some long profiles. The spatial relationship between steep river reaches and a plane pertaining to the NNW-SSE striking suite that extends from Armo to Oliveto is also obvious. Furthermore, as stated above, two planes of this set could also be responsible for offsetting the lower valleys of the San Agata, Menga and Armo rivers. Such a close relation between river profile irregularities and the fault planes emphasizes the very recent activity of most faults in the area. Although the highly variable rates of erosion prevent any precise dating of fault movement, this suggests simultaneous activity in differently oriented fault systems, in agreement with Ghisetti's (1981a) interpretation of a multidirectional extension phase. Finally, it should be stressed that a number of conditions are required for the coplanarity processing of lineaments to yield significant results. First of all, the assumption of evenness of the fault 'planes' is rather unrealistic at the regional scale. It is thus necessary to limit the size of the study area so that fracture planes no longer than 10-15km are described. This implies also that
126
Y. CORNET & A. DEMOULIN
the method should more appropriately apply to tectonic settings characterized predominantly by strike-slip faulting. A second limitation is that the coplanarity analysis can only work in areas of high relief. This is essential for good definition of the third dimension of the computed planes. Methodological constraints result from the low vertical dispersion of lineament endpoints with respect to their horizontal dispersion, and from the leastsquares method used to determine the plane equation. A problem arising from this typical lineament distribution is that the greatest part of the computed planes are almost flat-lying planes without tectonic significance. These planes must therefore be removed, and thus potential information on reverse and thrust faults is lost. Moreover, such an anisotropic 3D distribution prevents any extrapolation of the fracture planes down to more than a few kilometres, so that any comparison with earthquake hypocentres is unreliable. The coplanarity method is also very sensitive to the accuracy of the data and should be improved by a better morphometric definition of the small-scale linear features it exploits. Such a clear definition could lead to an automatic or semiautomatic detection of these features, which would increase the number of input data, reduce the random effect of outliers on the construction of plane equation and ensure a more reliable determination of the tectonic significance of these planes. In the same way, another refinement of the method would be to use the (x, y, z) coordinates of not only lineament endpoints but also of points distributed along its whole length to compute fracture planes, therefore taking into account the actual 3D non-rectilinearity of the lineaments. P. Ozer is acknowledged for the digitization of the topographical maps contour curves. C. Vita-Finzi and I. Stewart are thanked for numerous suggestions for the draft version of this paper before its submission to the referees. We also wish to thank W. Murphy and two anonymous referees, for their helpful comments, which greatly improved the manuscript.
References A]'FA, T., BARRIER, P., FEINDBERG, H. • POZZ1, J. P. 1987. Pal6omagn6tisme des terrains s6dimentaires plio-quaternaires du d~troit de Messine. Documents et Travaux de l'Institut Gdologique Albert Lapparent, 11, 83-90. ATZORI, P., GHISETTI, F., PEZZINO,A. & VEZZANI,L. 1983. Carta geologica del bordo occidentale dell'Aspromonte. Lavoro eseguito con finanziamento CNR, GTM, Rome.
BARRIER, P. 1987. Stratigraphie des d6p6ts plioc~nes et quaternaires du d~troit de Messine. Documents et Travaux de l'Institut G~ologique Albert Lapparent, 11, 59-81. BOSCHI, E., PANTOSTI, D. & VALENSISE, G. 1989. Modello di sorgente per il terremoto di Messina del 1908 ed evoluzione recente dell'area dello Stretto. Atti VIII Convegno, GNTS, Rome, 1989, 245-258. BOTTARI, A. & Lo GIUDICE, E. 1987. Structural studies on the Strait of Messina. The seismotectonic data. Documents et Travaux de l'lnstitut Gdologique Albert Lapparent, 11, 115-125. CHAPPELL, J. & SHACKLETON, N. J. 1986. Oxygen isotopes and sea-level. Nature, 324, 137-140. COSENTINO, D. & GLIOZZI, E. 1988. Considerazione sulle velocitfi di sollevamento di depositi eutirreniani dell'Italia meridionale e della Sicilia. Memorie della Societd Geologica Italiana, 41, 653-665. DOGLIONI, C., TROPEANO,M., MONGELLI, F. t~ PIERI, P. 1996. Middle-Late Pleistocene uplift of Puglia: an 'anomaly' in the Apenninic Foreland. Memorie della Societgl Geologica Italiana, 51, 101-117. DUMAS, B. & RAFFY, J. 1993. Validit6 et pr6cision de la mesure des mouvements verticaux par r6f6rence aux pal6oniveaux marins: exemples en Calabre m~ridionale. Bulletin de 17nstitut G~olqgique du Bassin d'Aquitaine, 53, 141 150. - - , GUI~REMY,P., LHI~NAFF, R. & RAFFY, J. 1987. Rates of uplift as shown by raised Quaternary shorelines i. southern Calabria (Italy). Zeitschrift fu'r Geomorphologie Neuefolgo, 63, 119-132. & - - 1 9 9 3 . Rapid uplift, stepped marine terraces and raised shorelines on the Calabrian coast of Messina Strait, Italy. Earth Surface Processes and Landjbrms, 18, 241 256. ELIASON, J. R. 1992. Mapping fractures remotely for earthquake hazard assessment by the use of topographic and seismic hypocenter data. Episodes, 15, 75-82. FINETTI, I., LENTINI, F., CARBONE, S., CATALANO, S. & DEC BEN, A. 1996. I1 sistema appennino meridionale-arco calabro Sicilia nel Mediterraneo centrale: studio geologica-geofisico. Bolletino della Societd Geologica Italiana, 115, 529 559. FONTES, J. C., BARRIER, P., Dl GERONIMO, I. & JEHENNE, F. 1987. Datations carbone 14 sur des bivalves pl~istoc6nes sup6rieurs du D6troit de Messine. Documents' et Travaux de I'Institut G~ologique Albert Lapparent, 11, 101-104. GHmETTI, F. 1979. Evoluzione neotettonica dei principale sistemi di faglie della Calabria centrale. Bolletino della Societgt Geologica Italiana, 98, 387 430. - - 1 9 8 1 a . L'evoluzione strutturale del bacino pliopleistocenico di Reggio Calabria nel quadro geodinamico dell'arco calabro. Bolletino della Societ~i Geologica Italiana, 100, 433-466. 1981b. Upper Pliocene-Pleistocene uplift rates as indicators of neotectonic pattern: an example from southern Calabria (Italy). Zeitschriftj~ir Geomorphologie, Neue Folge, Supplementband, 40, 93 118. - - 1 9 8 4 . Recent deformations and the seismogenetic source in the Messina Strait (Southern Italy). Tectonophysics, 109 191-208.
LINEAMENT-COPLANARITY
ANALYSIS, SOUTHERN CALABRIA
& VEZZANI, L. 1982. Different styles of deformation in the Calabrian arc (Southern Italy): implications for a seismotectonic zoning. Teetonophysics, 85, 149-165. --, LOMBARDO, G. & VEZZANI, L. 1979. Dati preliminari sulla neotettonica dell'Aspromonte. Fogli 254 (Messina-Reggio Calabria), 255 (Locri), 263 (Melito di Porto Salvo) e 264 (Palizzi). Nuovi contributi preliminari alla realizzazione della Carta neotettonica d'Italia. Publ. 251 del, Centro Nazionale delle Richerche, Progitto Finalizzato Geodinamica. HAQ, B. U., HARDENBOL, J. & VAIL, P. R. 1987. Chronology of fluctuating sea levels since the Triassic. Science, 235, 1156 t 167. HUTCHINSON, M. F. 1989. A new procedure for gridding elevation and stream line data with automatic removal of spurious pits. Journal of Hydrology, 106, 211-232. KOCH, M. & MATHER, P. M. 1997. Lineament mapping for groundwater resource assessment: a comparison of digital Synthetic Aperture Radar (SAR) imagery and stereoscopic Large Format Camera (LFC) photographs in Red Sea Hills, Sudan. International Journal o[" Remote Sensing, 1 8 , 1465-1482. LENTINI, F., CARBONE, S., CATALANO,S. & GRASSO, M. 1996. Elementi per la ricostruzione del quadro strutturale della Sicilia orientale. Memorie della Societdt Geologica Italiana, 51, 179-195. LUNGO, G. 1988. Tettonica globale dell'Italia meridionale: subduzione o bending? Memorie della Societd~ Geologica Italiana, 41, 159-163. , CUBELLIS, E., FERRI, M., OBRIZZO, F. & TORTORA, A. 1991. E falsificabile il modello di tettonica globale? I1 caso dell'Italia meridionale. Memorie della Societh Geologica Italiana, 47, 333-340. MALINVERNO, A. & RYAN, W. B. F. 1986. Extension in the Tyrrhenian Sea and shortening in the Apennines as result of arc migration driven by sinking of the lithosphere. Tectonics, 5, 227-245. MIYAUCHI, T., DAI PRA, G. & SYLOS LABINI, S. 1994. Geochronology of Pleistocene marine terraces and regional tectonics in the Tyrrhenian coast of South Calabria, Italy. ll Quaternario, 7, 17-34. MONTENAT, CHR. & BARRIER, P. 1987. Approche quantitative des mouvements verticaux quaternaires dans le d6troit de Messine. Documents et Travaux de I'Institut Gdologique Albert Lapparent, 11, 185-190. MURPHY, W. 1993. Remote sensing of active faults: case studies from southern Italy. Zeitschrift ffir Geomorphologie, Supplement-Band, 93, 1-23. - -
127
NERI, G., CACCAMO,D., COCINA, 0. 8~;MONTALTO,A. 1996. Geodynamic implications of earthquake data in the southern Tyrrhenian Sea. Tectonophysics, 258, 233-249. OTT D'ESTEVOU, P., BARRIER, P., BOUSQUET, J. C. et al. 1987. Evolution structurale plioc4ne et quaternaire du D6troit de Messine. Documents et Travaux de l'Institut Gdologique Albert Lapparent, 11, 105-114. PANTOSTI, D. & VALENSISE,G. 1989. Faulting mechanism and complexity of the November 23 1980, Campania-Lucania earthquake, inferred from surface observations. Journal of Geophysical Research, 95, 15319-15341. --, SCHWARTZ, D. P. & VALENSISE, G. 1993. Paleoseismology along the 1980 surface rupture of the Irpinia fault: implications for earthquake recurrence in the Southern Apennines, Italy. Journal of Geophysical Research, 9 8 , 6561 6577. RICCHETTI, E. & RICHETTI, G. 1991. Aspetti della morfogenesi pleistocenico-olocenica sul versante tirrenico della Calabria. Memorie della Societh Geologica ltaliana, 47, 655 663. - & 1996. Morpho-structural study of southwestern Calabria (Italy) using GIS. Memorie della Societh Geologica Italiana, 51, 403-417. SHACKLETON, N. J. • OPDYKE, N. D. 1973. Oxygen isotope and paleomagnetic stratigraphy of equatorial Pacific Core V28 238. Oxygen isotope temperature on a l0 s and 106 year time scale. Quarternary Research, 3, 39 55. VACHARD, D., BARRIER, P., MONTENAT, C. & OTT D'ESTEVOU, P. 1987. Dykes neptuniens, br6ches internes et 4boulis cimenths des escarpements de faille du dhtroit de Messine au Plio-Quaternaire. Documents et Travaux de l'Institut Gdologique Albert Lapparent, 11, 127-141. VALENSISE, G. & D'ADDEZIO, G. 1994. I1 contributo della geologia di superficie all'identificazione delle strutture sismogenetiche della pianura di Gioia Tauro. Pubblicazione dell' Istituto Nazionale di Geofisica, 559. -& PANTOSTI, D. 1992. A 125 Kyr-long geological record of seismic source repeatability: the Messina Strait (Southern Italy) and the 1908 earthquake (Ms 7.5). Terra Nova, 4, 472-483. WESTAWAY, R. 1993. Quaternary uplift of Southern Italy. Journal of Geophysical Research, 9 8 , 21 741-21 772.
Archaeological evidence for vertical movement on the continental shelf during the Palaeolithic, Neolithic and Bronze Age periods N. C. F L E M M I N G
Southampton Oceanography Centre, Empress Dock, Southampton S014 3ZH, UK (e-mail:
[email protected]) Abstract: Submarine archaeological sites of human occupation on the continental shelf
provide evidence of palaeo-shorelines during the last 50000 years, with possibly longer periods of accessible data. Sites are widely distributed globally, and provide a rich source of peripheral information on landscapes, climate, water supply, sediment transport, and soil types, which are relevant to coastal tectonics and processes, in addition to location, elevation, and date. More than 500 sites have been discovered world-wide where there are the remains of human occupation under the sea. Artefacts embedded in a few metres of sediments can survive multiple marine transgression, given local topography that creates suitable wave and current conditions. The raw data have been analysed to separate glacial eustatic fluctuations of global sea level and local or regional tectonic processes. The number of sites in the Mediterranean permits the detection of regional vertical tectonic displacement with a spatial resolution of a few tens of kilometres in many regions. In other parts of the world the density of sites is not yet sufficient to achieve this, but indicators of vertical movements, or lack of vertical movement, provide valuable point data. Continuing discovery of new sites indicates that the technique will provide data with improved spatial resolution, and, outside the Americas, over a timescale extending to hundreds of millennia, and possibly one million years The purpose of this paper is to show that archaeological material on the continental shelf older than the classical Roman, Greek, and Phoenician harbours of the Mediterranean makes a contribution to understanding the vertical tectonic movements of coastlines during late Quaternary oscillations of sea level. Archaeological data are a part of the armoury of sea-level indicators, comprising also beach deposits, terraces, cliff-terrace junctions, sea caves, peat, coral reefs, molluscs, coastal dunes, shallow-water deposits, and other indicators. There is an extensive literature on the subject of deriving palaeo-sea levels and regional Earth movements from classical archaeological sites (e.g. Flemming 1969, 1978; Blackman 1973; Pirazzoli 1976; Flemming & Webb 1986; Lambeck 1995). The present paper seeks to show that it is possible to work backwards from the proven examples of classical structures and relative sea levels derived with an accuracy of the order of 20-50 cm, into the less certain era of Bronze Age, Mesolithic, Neolithic, and Palaeolithic sites on the continental shelf. From an archaeological perspective the sites have an added value because organic materials are often preserved in marine sediments, where the same materials would not survive on land. The sites have also usually been protected from subsequent human disturbance. The preserved state of the sites, including datable organic materials, results in improved accuracy of determination of related sea levels and dates.
The paper concludes with a prediction of the future rate of discovery, an analysis of the value of archaeological data in analysing vertical relative movements of the land and sea level, and hence the value of the data in deriving late Quaternary tectonic movements of the coast. Conversely, study of human migration between land masses is influenced by the width of channels and the height of adjacent coasts. Thus a knowledge of vertical coastal movements, whether tectonic or isostatic, is a factor in understanding the archaeology of the last few hundred thousand years. Of the 500 known submarine sites world-wide containing in situ remains of buildings, structures, harbour works, quarries, or lithic artefacts, approximately 100 are older than 3 ka BP, that is, in European archaeological terminology, Bronze Age or older. During the last 25 years there has been a steady rate of discovery of older and deeper archaeological sites, with some of them many tens of kilometres out to sea. Submerged sites older than 3 ka BP have been found in the Mediterranean, Baltic, North Sea, English Channel, Gulf of Mexico, and on the coasts of California, Florida, South Africa, Japan, and Australia. Several sites have been found in the time range of 20-45 ka BP, earlier than the Flandrian glacial minimum, but only one submarine archaeological site so far which pre-dates the last glacial cycle. The earliest date of human subsistence based on the exploitation of the sea for food, salt, and
FLEMMING, N. C. 1998. Archaeological evidence for vertical movement on the continental shelf during the Palaeolithic, Neolithic and Bronze Age periods, hi: STEWART, I. S. & VJTA-FINZl, C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 129-146.
130
N. C. FLEMMING
transport has not been established. It is reasonable to assume that shellfish, fish, seaweed, and salt were exploited for many millennia before any kind of floating craft were used, but even this is not certain. It does not require much ingenuity or ambition to observe a floating log and realize that it could be used to cross a river, or reach a nearshore island. The most indisputable evidence for early seafaring is the date of first occupation of Australia-Papua New Guinea. This land mass was never joined to Indonesia during the Pleistocene low sea levels, and hence the occupation must have taken place as the result of a substantial sea crossing (Birdsell 1977). Somewhat similar evidence from Gibraltar or Sicily is ambiguous because of the parallel migration route through the Middle East in each case. Evidence from Australia (Allen et al. 1977; Jones 1977; Masters & Flemming 1983a; Webb 1995) indicates progressively earlier crossings from Indonesia to the Australian-Papua New Guinea land mass, with estimates now in the range 50-100kaBe. The Australia crossing requires a capability to transport hundreds of people across many tens of kilometres of open sea. This implies fairly advanced seafaring competence. Evidence from land sites such as Terra Amata near Nice show that seafood was exploited as early as 250kaBe. Nevertheless, the sea level during the last million years has been below present sea level for most of the time, and hence most coastal sites where the debris of seafood might have been found by archaeologists are now underwater. Archaeological sites above present sea level where shell middens have been found pre-dating 10ka B~' tend to be associated with previous high sea levels (e.g. Klasies River Mouth, Deacon et al. 1986). The continuous history of Palaeolithic and Neolithic marine food subsistence and maritime technology is largely missing, in the sense that those artefacts and deposits which do still exist are mostly under the sea, unless the coast itself has been uplifted (e.g. St61ting 1997). Flemming (1996) has analysed the reasons for expecting coastal human occupation sites and hunting areas to be extended seawards as sea level drops. There are positive advantages in living on the coastal plain, close to estuaries, lagoons, and seafood. To understand the methods for location and survey of sites older than 3 ka BP it is necessary to consider how more recent sites have been discovered, mapped, and interpreted. Submarine archaeological remains of the last 3000 years are mapped and excavated by archaeologists so as to understand the marine technology and culture of
past ages, the construction of harbours, lighthouses, slipways, mooring stones, fish tanks, salt pans, warehouses and weapons sheds, breakwaters, quays, and the military defences and trade of coastal cities. For earlier periods, excavation and the recovery of surface artefacts on the sea bed reveal early settlement of offshore islands, fishing and seafaring abilities, and the distribution of occupation sites on the submerged continental shelf. As the sea level rose about 100m between 18 and 5kaBp, the study of coastal settlements earlier than 5kaBp can only be carried out by underwater research, except in a few areas of rapid land uplift for tectonic or isostatic reasons. Submerged human occupation sites and the ruins of submerged cities have been known and described reliably by antiquarians, travelling scholars, archaeologists and scientists for over 200 years (e.g. Hamilton 1776-1779; Gfinther 1903; Frost 1963; Flemming 1969; Pirazzoli 1976; Blackman 1982; Stright 1990). Analysis of the vertical distribution of coastal sites, whether uplifted or submerged relative to the sea level since the time of occupation, is used to plot the distribution of vertical Earth movements, and to measure the probable rate of sealevel change over centuries to millennia (Flemming & Webb 1986). Archaeological methods of measuring relative change of sea level can be used in conjunction with tide gauges, and other systems of measurement, to make estimates of the probable rate of change in the next 10-100 years at each site (Flemming 1993). Some 500 submarine occupation (Fig.l) sites have been discovered on the coasts of 32 countries, excluding lake dwellings. (The terms submerged or submarine will be used interchangeably and always refer to the sea.) Many more sites have been surveyed on or close to the coast which give evidence of relative stability or uplift of the coast. Flemming & Webb (1986) compiled an inventory of over 1000 sites for the Mediterranean, and this has been extended in an unpublished report compiled by Blue & Flemming (1994) for UNEP (United Nations Environment Programme) Mediterranean Action Plan. In a similar vein, Stright (1990) and Johnson & Stright (1992) have provided analyses of the known submerged human occupation sites on the coasts of the Americas.
The number and rate of discovery of submarine human occupation sites The abbreviation SHOS will be used to mean submarine human occupation site or sites. Freshwater lake sites, although important in
ARCHAEOLOGICAL EVIDENCE FOR TECTONIC MOVEMENTS
131
Fig. 1. The map shows sites where submerged artefacts and buildings occur as the remains of human occupation activities, not shipwrecks. The areas where many sites occur close together are shown by a generalized stipple. (For a listing of the numbers of sites per country, see Table 1.)
their own right, will not be included in this study. Table 1 shows approximate totals for each country, based on key references cited, personal correspondence, and unpublished conference reports and preprints. Some sites show remains of structures below the sea surface which were built in the sea in the first place, and are not evidence of a relative change of sea level. Others also show evidence of relative submergence. The two columns of figures show the two types of sites, where they can be separated. The numbers in parentheses refer to the bibliography below the table. The figures in the first column for known sites include the data in the next column. The types of material referred to in Table 1 are very varied, and include extensive wellpreserved areas of ruined cities of the Medieval or later periods, Roman and Greek Classical cities, Bronze Age ruins, scattered single houses, villas, fish tanks, olive presses, storage silos, and other signs of work and habitation, harbour works, quays, slipways, tombs, steps, salt pans, Neolithic, Palaeolithic, and Palaeo-Indian sites, quarries of all ages, fish traps, channels and canals, wooden trackways, middens, hearths, stone tool assemblages, stone tool d~bitage, wells, occupied caves, stone tools in sedimentary
context, and assemblages of tools and bone in submarine peat and alluvial deposits. The methods of interpretation of the data at typical sites are given in the literature. Each occurrence of anthropogenic artefacts is counted as a separate site at scales of single towns, villages, occupied caves, and isolated dwellings, structures, kill sites, work or debris sites. Sites therefore vary in size from a few metres square to several hectares. Smyth (1854) described a few tens of SHOS in the Mediterranean, and similar numbers were known (with many in common) to GUnther (1903), N6gris (1904), Zeri (1905), and Gnirs (1908). Properly documented and surveyed SHOS were added to the records at a rate of much less than one per year up to the time of the rather personal inventory compiled by Diol6 (1952). Shepard (1964) summarized the findings of the previous decades regarding stone mortars and metates retrieved from offshore San Diego county by divers. By 1969 Flemming had identified 73 sites with submerged materials, and 24 definitely displaced downwards relative to present sea level in the Western Mediterranean, and Frost (1963) described about 20 from the Eastern Mediterranean. In the 1970s sites
132
N. C. F L E M M I N G
Table 1. Distribution of submerged human occupation sites ( S H O S ) Country
Sweden Denmark Germany France (North and West) UK Spain France (Mediterranean) Italy Yugoslavia Albania Greece Turkey Cyprus Syria Lebanon Israel Egypt Libya Tunisia Malta Algeria Morocco Caribbean Bahamas USA (E Coast and Gulf) USA (W Coast) Japan India Australia Bulgaria Former USSR (Black Sea) Canada South Africa Total
Known sites with SHOS remains
Known sites showing relative rise of sea level
2 50 3 20 40 9 20 83 20 65 60 18 2 7 34 4 6 14 1 9 4 2 3 30 6 30 2 3 2 2 1
2 50 3 20 40 14 65 5 42 24 8 2 3 17 4 3 7 1 3 1 1 3 30 6 30 2 3 2 2 1
552
394
Key references (1) (1,14, 26) (1,2) (3, 22, 23) (4, 5) (4, 5, 28) (4, 5, 6, 7) (5, 8) (5, 9, 27) (10, 11) (11) (12) (12) (13) (5) (5, 7, 15) 5 5 (25) (16, 20, 21,25) (17, 21,25) (18, 24) (19) (16)(30) (29)
References: (1) Masters & Flemming (1983a); (2) Scuv~e & Verague (1988); (3) N A S Newsletter; (4) Flemming (1969); (5) Flemming & Webb (1986); (6) Schmiedt (1972); (7) Pirazzoli (1976); (8) Marasovic et al. (1976); (9) UNEP database; (10) Flemming (1973a); (11) Flemming (1978); (12) Frost (1972); (13) Flemming et al. (1978); (14) Skaarup (1980); (15) Paskoff& Sanlaville (1983); (16) Cockrell (1980); (17) Masters (1983); (18) Osaki (1973); (19) Flemming (1971); (20) Dunbar (1988); (21) Stright (1990); (22) Bell (1991); (23) Akeroyd (1972); (24) Araki & Ishihara (1989); (25) Johnson & Stright (1992); (26) Fischer (1993); (27) Tina (1990); (28) Clottes et al. (1992); (29) Werz & Flemming (pers. comm. 1998); Josenhans et al. (1997). were discovered in the Baltic (Skaarup 1980), and in Japan (Osaki 1973), and by the early 1980s numerous sites had been confirmed off the coast of California and Florida (Cockrell 1980; Masters 1983; D u n b a r et al. 1992). By the mid1980s, sites were being found in the Bahamas, Australia, and India. The rate of discovery in the Mediterranean may have reached a plateau by the 1990s, as the Classical and later sites have been almost completely catalogued, but there are still several hundred sites of these periods which have not been fully explored underwater. The discovery
of Bronze Age and lithic SHOS in the Mediterranean continues slowly but steadily. The phase of rapid discovery in the Danish Baltic may also have flattened out, although other parts of the Baltic, especially the eastern Baltic and the G u l f of Finland, have not yet been so intensively searched. As the availability of cheap diving methods becomes widespread, there is likely to be a rapid increase in discoveries of all ages off the coasts of Asian countries, coral or sediment permitting. The number of sites with evidence of submarine remains has grown as shown in Fig. 2.
ARCHAEOLOGICAL EVIDENCE FOR TECTONIC MOVEMENTS
133
Fig. 2. The number of known submerged human occupation sites has increased rapidly since 1950. The bold line shows the number of all istes where ancient structures, inhabited caves, and lithic artefacts have been found in situ beneath the sea or below sea level in caves close to the sea. The dashed line shows the number after excluding those sites where the buildings, such as breakwaters, were build in the sea originally. The fastest increase in the number of documented SHOS occurred in Europe and the Mediterranean with the increasing popularity of scuba diving after 1950 and the availability of this cheap method of work to universities, museums, and amateur research groups. Most of the sites discovered during the rapid period of growth were stone or brick constructions from 1000 to 2800 years old. This phase is likely to be worked out in Europe and the Mediterranean in the next decade or so in terms of discovery and preliminary mapping, although there will be a limited number of important sites which continue to attract attention for more serious excavation. Very few sites warrant the massive effort of excavation which has been carried out at Caesarea (Oleson et al. 1984; Raban 1992) or on a smaller scale at Kenchreai (Scranton & Rammage 1964). The coasts of India and South East Asia were developed by technically advanced cultures on the order of 2000 years ago, and thus equivalent maritime structures and SHOS should be discovered there as scuba diving becomes more widely practised. The possibility of finding lithic artefact sites in the Indonesian archipelago shelf is particularly interesting, because of the early occurrence of seafaring in the region, and because it is the transit area for the first human occupation of Australia. In earthquake-prone
areas such as the Caribbean and Indonesia, SHOS may be discovered from fairly recent periods within the last 1000 years, and examples are already known for the Caribbean such as Port Royal (Hamilton 1984). Older sites are usually less substantial in structure, consisting of midden heaps, tool sites, shallow stone foundations of wooden houses, etc., and are therefore both difficult to find and difficult to recognize. The rate of discovery of these sites is slower than that for the Classical sites, which furthermore can be located approximately from the historical literature. The most complete attempts to detect regional and local vertical coastal tectonic movements from archaeological data are those of Pirazzoli (1976), Flemming (1978) and Flemming & Webb (1986). Lambeck (1995) has computed separate eustatic, glacio-hydro-isostatic, and tectonic contributions to the vertical movements of the coast of the Aegean, and compared the results with the data published by Flemming (1973a, 1978). This will be discussed more fully later. Sites older than 5000 years were constructed or occupied at a time of lower Pleistocene sea level, and therefore the relative altitude of the sea level computed for each site always contains components of displacement caused by glacial eustatic sea-level change, post-glacial isostatic compensation, and local or regional tectonics. SHOS of
134
N. C. FLEMMING
this age are being found slowly off all coasts, and will not be restricted to Europe and the Mediterranean (Masters & Flemming 1983b), Josenhans et al. (1997).
The process of discovering sites Statistical derivation of coastal tectonic movements, and the separation of causes of relative vertical change of level, eustatic, glacio-isostatic, or tectonic, depends upon the compilation of a large consistent dataset of many tens of sites for which the technique of dating and measurement of level is standardized. There is therefore a premium in promoting the discovery of the maximum number of good sea-level indicator sites of all kinds. Sports divers have been active in the location of Mesolithic SHOS in the Baltic (Skaarup 1980; Fischer 1993, 1995; Pedersen et al. 1997), and in the subsequent supervised excavation. Discoveries in Israel (Raban 1983), and off the coast of Florida (Dunbar 1988) were made as a result of continuous dialogue and interaction between professional archaeologists, student diving groups, and amateurs including fishermen. Mesolithic bone and antler implements trawled from banks in the North Sea were documented systematically by Louwe Kooijmans (1970-1971) as a result of regular discussions with fishermen. Professional searches in new areas and deeper water where discoveries would be academically significant have been carried out by Van Andel & Llanos (1983, 1984) off the Argolid, Greece; Flemming (1986) off the Cootamundra Shoals, northern Australia; and Flemming & Kazianis (1987) off Kerkyra, Greece. Garrison (1992) conducted a precision acoustic survey of palaeoshorelines and submerged sinkholes out to a depth of 200 m off the coast of Alabama, Gulf of Mexico, but regarded this only as the precursor to a search for archaeological materials. Numerous searches in oil concession areas have been conducted off the USA under the requirements of the Outer Continental Shelf (OCS) legislation (Flemming 1981; Patterson 1981; Pearson et al. 1986; Johnson & Stright 1992). Of these searches the only one to produce lithic artefacts from below sea level was that off the coast of Kerkyra, and in this case the underwater finds were within 1-2 km of prolific areas of artefacts on the mainland of Kerkyra, or the adjacent offshore islets of Laghoudia and Diaplo. The field experience shows that intensive searching in prospectively rich areas using acoustic surveys and diving over several months stands only a marginal chance of locating artefacts. The most
effective method of improving the chances of success is to make very close comparison between artefact locations on shore and the geomorphological environments being searched underwater. Excavations conducted on shipwrecks less than 1000 years old have resulted in the discovery of prehistoric SHOS at the same location. This first occurred in excavation of the 1715 Plate Fleet on the coast of Florida (Cockrell & Murphy 1978). The excavation of Dutch East India wrecks off Cape Town, South Africa, produced Acheulean hand axes from the underlying sediments (Werz & Flemming 1998). The discovery of the submerged Jomon culture site off Tokonami was the result of a survey which started as a search for shipwrecks, but found no shipwreck at all (Araki & Hayashida 1993, p. 124). Professional search time for submarine lithic artefact sites is expensive, and is only likely to amount to a few tens of person-months per year in any technically developed country. It follows that finds are more probable if amateurs are enlisted.
Taphonomy of submarine occupation Kraft et al. (1983) analysed the probable sequence of events as a coastal occupation or kill site is submerged by rising sea level for different shelf environments. Flemming (1983b) and Geddes et al. (1983) considered the effects of the rising sea level on prehistoric SHOS, and the effects of fetch, wave action, and alluvium. Discovery is most probable when a site which has been buried in sediment is slowly re-eroded as a result of a change in sediment supply or oceanographic conditions. If it is exposed by slow erosion, there is a chance that a diver will observe the critical indicators during the few decades before the SHOS is destroyed. This sequence seems to have occurred at Asopos in southern Greece, where Flemming photographed submarine standing Bronze Age or Neolithic walls in 1967 which had been destroyed by 1979; at Atilt, Israel, where organic materials are exposed and eroded on a seasonal basis; and at Fermanville, where flint tools are being eroded out of the peat. In ideal circumstances the SHOS may achieve a stable state which combines preservation from natural forces and the possibility of discovery. This has occurred in the Florida sinkholes such as Warm Mineral Springs (Cockrell 1980) and Little Salt Springs (Clausen et al. 1979), and possibly in Wakulla Springs (Stanton, pets. comm.). The Neolithic site at Aghios Petros appears to be stable because of the extreme degree of shelter behind a small island in a deep
ARCHAEOLOGICAL EVIDENCE FOR TECTONIC MOVEMENTS bay (Flemming 1983b). Lagoon and archipelago locations also can provide stable environments with sufficient sediment to protect artefacts, and low wave energy. High-energy beaches can be stable preservation sites provided that the sand supply is in equilibrium with the seasonal fluctuations of transport. Thus well-preserved SHOS have been found in the surf zone on the coast of Israel at Neve Yam (Wreschner 1977), and on the east Florida coast near Cape Canaveral (Cockrell & Murphy 1978). The Acheulean tools found by Werz in Table Bay (Werz & Flemming pers. comm. 1998) were located on or close to bedrock under several metres of sand on a lowgradient coast exposed to the Atlantic swell from the west. In this case, the tools have survived several marine transgressions with protection from abrasion or transport, as the original cutting edges are still sharp in one case, and only slightly worn in others.
Age distribution of submarine occupation sites The first documented SHOS were Roman and Greek harbour structures and city buildings in the Mediterranean (Hamilton 1776-1779; Shaw 1808, p. 59; Beechey 1827; Smyth 1854). In the 1930s a Mesolithic fish spear was trawled up from the Leman bank in the North Sea (Godwin 1941, p. 350; Steers 1948, p. 492), but this was a chance find, and could not be exploited in the field, although its significance was fully appreciated. By the 1960s Frost (1963, 1972) had identified pre-Classical harbour works in Crete and the Levant. Blanc (1940) recognized that calcified artefact encrustations and breccia on the walls of sea caves at Palinuro, Italy, proved that the caves must have been occupied when the sea level was much lower, and similar deductions could be made for Gorham's Cave, Gibraltar, which was occupied about 40000 years ago (Waechter 1964; Clark 1969, p.46). At both Palinuro and Gibraltar the caves can now only be entered from the sea, although the surveyed deposits are now at or above sea level and no discoveries were made below sea level. In 1967, Flemming found a submerged Bronze Age harbour town in southern Greece (Flemming 1968, 1971, p. 49-69; Harding et al. 1969), and various submerged tombs in shallow water were known from the Bronze Age before 1967. In the 1970s, Raban and others surveyed and partially excavated Bronze Age SHOS off the coast of Israel (Raban 1983). In the Baltic, Skaarup and others surveyed and partially excavated Meso-
135
lithic SHOS in the 1970s and 1980s (Skaarup 1980; Flemming 1983a; Fischer 1993), with an age of the order 6kaBP. In 1981, Efstratiou & Flemming surveyed and conducted test pit excavation on a submerged Neolithic SHOS in northern Greece (Flemming 1983b). During the 1980s, Galili surveyed and excavated in detail a Neolithic site of about 10 ka BP off the coast of Israel, near Atlit (Galili 1987). In the USA, Murphy (1978), Clausen et al. (1979), and Cockrell (1980, 1986) excavated SHOS in the sinkholes and springs of Florida, dating from 10 to 11 kaBP. Masters (1983) compiled the most detailed documentation on the coastal grindstone and mortar finds off the coast of Florida, some of which are associated with SHOS dating from about 6kaBp. Riccardi et al. (1987) excavated a Neolithic deposit in a coastal marine cave in Italy, to a depth of 22 m. Araki & Hayashida (1993) described the excavation by diving and airlift dredging of a Jomon culture pottery and lithic site dated at 8630 4- 105 a BP at a depth of 25m submerged off the Tokonami river mouth in the Tsushima Straits of southern Japan. The difficult frontier now is that of making the transition to fully Palaeolithic sites, with no structures or pottery. The most complete excavation of this kind is that at Fermanville, near Cherbourg, France (Scuv6e & Verague 1988). The site shows a Mousterian layer of tools of the Acheulean tradition in the peat banks of a creek bed at a depth of 18m below present sea level. The date is approximately 40 ka BP, with a very rich tool assemblage. For more than 20 years there has been a search for prehistoric material in submerged caves in the south of France. In 1991 wall paintings and charcoal dated to 15kaBp were found by divers in a submarine cave in the south of France, and reported in the popular press in many countries. The finds have been checked by J. Courtin, a professional archaeologist who reports regularly on submarine work in progress (Clottes & Courtin 1992; Clottes et al. 1992). The cave entrance is at a depth of 40m, although the reported artefacts are above present sea level, with a maximum age of 27 ka BP. Off the coast of Kerkyra (Corfu) Greece, Sordinas (1983) reported the finding of numerous Levallois-Mousterian tools in shallow water, dragged up by fishermen, lifted by tourist divers, or picked up in the surf zone. Flemming visited the area with Sordinas in 1984 and located areas where flint tools and d6bitage could be observed on the sea bed in 5 m depth and 200 m offshore. A joint team of Greek and British divers worked on the site, and other
136
N. C. FLEMMING
locations along the coast, during 1987, and located a number of tools and areas of probable d6bitage, with a maximum depth of 30m (Flemming 1985; Flemming & Kazianis 1987). Hopkins (pers. comm. 1976, reported by Stright (1990)) described percussion flakes of possible human manufacture dredged from a depth of 38m c. 10km off the eastern shore of St Lawrence Island in the Bering Sea. Kershaw et al. (1993) attributed charcoal in the core from Ocean Drilling Program (ODP) site 820 to anthropogenic forest burning as early as 140kaaP, although this was disputed by Webb (1995) and other commentators. In February 1995, the excavation of a Dutch East Indian wreck in Table Bay, South Africa, also produced an Acheulean hand axe from sediments immediately overlying the bedrock at a depth of 8 m below present sea level. The tool dates to the range 0.3-1.4MaBP. The detailed circumstances of the find, the stratigraphy of the sediments, the condition of the hand axe, and the analysis of the stability or vertical movement of the coast, have been described by Werz & Flemming (unpublished field report 1998). Sub-
sequently, two further hand axes have been found at the same depth adjacent to other wreck excavations. These discoveries by Bruno Werz constitute a 10-20-fold increase in the age of the oldest known human artefacts found in situ on the continental shelf. The evidence indicates that lithic artefacts buried in a few metres of coastal sediments can survive multiple marine transgression without significant transport or abrasion. The potential for discovery of submarine tools in the age range 50-500 ka BP is therefore confirmed. Figure 3 shows in outline the increase in maximum age of SHOS discovered and excavated. The exponential increase has occurred post-1950, and is not flattening out. It is probable that submerged Lower Palaeolithic sites will be found with dates of 150-250 ka BP, within the next 10 years. Older sites are not necessarily deeper (although deeper sites must be older during the last 20000 years) and a site dating from 100-200kaBp might be found in 5 - 1 0 m of seawater, just as the finds in Table Bay were found in 8 m depth. Once SHOS with ages of the order of 100 ka BP can be located and studied as routine,
Fig. 3. In recent decades divers have discovered much older inundated occupation sites. The dashed lines represent the time lag between discovery, survey, and publication. WMS, Warm Mineral Springs; LSS, Little Salt Springs. (Note non-linear scale on vertical axis.)
ARCHAEOLOGICAL EVIDENCE FOR TECTONIC MOVEMENTS the archaeological objective will be to develop our knowledge of the resource base represented by the occupational sequence of human activity on the continental shelf, and the location of sites in the major inter-continental straits.
Depth distribution of submarine occupation The maximum depth of known SHOS has increased rapidly during the last 30 years. Before AD 1900 all known SHOS were in shallower depths than 3m of seawater, with the exception of a few buildings on the northern shore of the Bay of Naples (G~nther 1903) displaced by the adjacent volcanism (Dvorak & Mastrolorenzo 1990). During the early twentieth century a few harbour sites were surveyed in which the foundations of breakwaters and other harbour works were in water depths of 5-10 m (Jondet 1916; Poidebard 1939), and in 1958 Flemming surveyed the harbour of Apollonia in Libya, with the foundations of breakwaters at 8 m. In 1959, Gargallo & Flemming worked on the harbour of Agathocles at Syracuse, Sicily, with the breakwater footing at 15 m. The large scale of Greek and Roman breakwaters necessitated survey to these depths, although the relative change of sea level at such sites was usually only 1-2m (Flemming 1969). In 1959, palaeo-Indian materials were found in Warm Mineral Springs, Florida, and subsequent finds led to a professional excavation being developed in the 1970s (Murphy 1978; Cockrell 1980). Burials were discovered at a depth of 13 m below surface water level. The spring is 15 km from the sea in karst topography, and the water table correlates with mean sea level through karstic underground channels. This site was the first archaeological discovery to indicate a necessary sea level change greater than 10m. Little Salt Springs a few kilometres southeast of Warm Mineral Springs provided similar finds (Clausen et al. 1979). During the 1960s and 1970s several site surveys indicated necessary sea-level or seismic land-level changes of the order of 5-10 m. These include Caesarea, Israel (Oleson et al. 1984), Franchthi Cave, Greece (Gifford 1983), and the mortar finds off the coast of California (Tuthill & Allanson 1954; Marshall & Moriarty 1964; Masters 1983). The evidence from the southern California coast seemed at first to indicate ages and depths correlated with eustatic sea-level changes of the order of 20-30m, but careful review of the provenance of finds suggests that many of the deeper ones were lost from sea craft. In 1981, Efstratiou & Flemming surveyed the submerged Neolithic village of Aghios Petros in
137
northern Greece, at a maximum depth of 10m (Flemming 1983b). Galili has been working for several years on the 10m depth site at Atlit (Galili 1987). In 1986, Ruckley recovered a Late Palaeolithic tool from a geological core taken in the northern North Sea from a water depth of 145m (Long et al. 1986). The unconsolidated sediment stratigraphy showed that the tool was in situ. In 1987, Flemming & Kazianis retrieved materials which are possibly Palaeolithic d6bitage from a depth of 30 m off the north coast of Kerkyra, and the late Rob Palmer found palaeoIndian burials in caves at about 25m on the coast of the Bahamas (R. Palmer, pers. comm.). The finds from 20m at Fermanville, France, were published in 1988, although the diving survey of the site started in 1970 (Scuv6e & Verague 1988). The maximum depth of tool materials was 20 m. Riccardi's work on the Italian coast has revealed Neolithic artefacts also at depths of the order of 20m (Riccardi et al. 1987), and Araki & Hayashida (1993) have found pottery and tools from the ninth millennium BP at a depth of 25 m off southern Japan. Figure 4 illustrates the trend of depth of discovered sites with date of discovery or principal publication. There is a considerable scatter, partly because of the early discovery of the remains in the Florida springs which are not in open seawater, although they are so close to the sea that the water table is determined by sea level. Archaeologists have known for decades that the continental shelf could, in principle, have been occupied to a depth of 100-200m during the Pleistocene (Blanc 1940; Godwin 1941; Flemming 1962, p. 171, 1985; Emery & Edwards 1966) but there has been very little evidence until recently that cultural materials older than 5 ka BP would survive or be discoverable in situ. Underwater discoveries of the last two decades show that submarine lithic artefact sites can survive in some circumstances, and that they can be found down to depths of several tens of metres. The chance find of the single tool by Long et al. at 145m proves finally that human occupation and hunting did extend towards the maximum depth of low sea level during the Pleistocene. The majority of submarine archaeological research is carried out by diving with compressed air scuba, and with this equipment it is very difficult to work below 40-50m. It is probable that the current maximum working depth of 20-30 m for SHOS survey and excavation will increase slowly to 40-50m over the next 10-15 years, but it is unlikely to progress to greater depth without a change in technology. The steady increase in the use of self-contained
138
N. C. FLEMMING
Fig. 4. The maximum depth of known SHOS has increased erratically. The footings of tower and breakwaters were surveyed in deeper water during the 1950s and 1960s. Submerged caves, and lithic artefact sites were discovered and surveyed during the 1970s and 1980s. WMS, Warm Mineral Springs. heliox and tri-mix diving equpment will extend the working depth to 80-100 m. Lithic SHOS are not like wrecks and cargoes which show up on underwater photographs or video, and nor do they give unambiguous signals on acoustic sub-bottom profilers. Clues and indicators can be obtained by remote sensing methods (Stright 1980; Rupp~ 1981; Gagliano et al. 1982; Johnson & Stright 1992; Joly pers. comm. 1993; Quinn et al. 1997), but in deep water it would be necessary to core extensively through a suspected site, or dive for many hours at depth using industrial techniques, to prove the existence of human artefacts or debris. Flemruing (1986) used remote sensing and diving on the Cootamundra Shoals, northern Australia, to try and identify occupation sites, and although suitable environments were located in depths of 30-60m, it was impossible to prove human occupation.
Geographical distribution of sites An advantage of prehistoric SHOS as a sea-level indicator is that sites are potentially distributed in a similar way off the coasts of all continents. It is unlikely that material older than 10ka BP will be found off the Americas, but for the rest of the world dates to 100kaBP are possible for all continents even including Australia, and older dates excluding Australia. The approximate numbers and locations of sites are listed in Table 1 and Fig. 1. The major-
ity are in the Old World, and in areas where hundreds of thousands of people have access to scuba diving equipment. Areas of high sedimentation, rapid coral growth, or recent ice scour are unsuitable for the survival or discovery of underwater sites. The last two decades have been notable for the rapid growth in numbers of SHOS outside the Mediterranean, with the most discoveries in the Baltic, Caribbean, Florida, California, and Japan. Lithic artefact sites under the sea are likely to be discovered by chance, or by logical extension of professional land excavations. However, as most non-European or non-Mediterranean SHOS are likely to be lithic artefact or midden sites, they will not be found quickly or recognized by amateur groups, or be attractive to souvenir hunters. Discovery of SHOS off the coasts of Africa, Asia, and the Pacific margin and islands will therefore tend to be slow and gradual, with an emphasis on planned searches and targeted discoveries in key areas. Areas of maximal interest are likely to be the migration routes between continents, Indonesia, northern Australia, Siberia, Alaska (Hoffecker et al. 1993; Josenhans et al. 1997) and locations where low sea level would significantly alter the nature of a migration route to small islands, such as the route to Fiji, or the Caribbean chains. The late John Gibbons had developed a range of testable hypotheses concerning migrations between low sea-level land masses in Melanesia, and was planning extensive diving research at the time of
ARCHAEOLOGICAL EVIDENCE FOR TECTONIC MOVEMENTS his unfortunate death in 1986. (J. Gibbons, pers. comm., and proposal submitted to the National Geographic Society in 1986). Araki & Ishihara (1989) reported 30 occurrences of lithic artefact sites in the shallow coastal waters around the Japanese islands. The original publications are in Japanese. The recent excavation of Acheulean tools in Table Bay, South Africa (Werz & Flemming 1998) has demonstrated for the first time the potential for discovery of submarine artefacts south of the Sahara, and this was foreshadowed by Van Andel (1980).
Site survey, assessment and derivation of relative sea level change Terrestrial survey and excavation methods have been adapted from archaeology on land for application underwater. The bibliography is vast and only a few references will be given to works which provide an overview and further broad bibliography. For survey of classical or more modern harbour sites the techniques have been described by Frost (1963, 1972), Scranton & Rammage (1964), Jameson (1969), Flemming (1971), Blackman (1973), Oleson et al. (1984) and Oleson & Raban (1989), and for measurement of palaeo-sea levels at archaeological sites, by Schmiedt (1972) and Flemming (1978). Vitali et al. (1992) have described sophisticated sedimentological techniques for identifying the location of a harbour basin at Carthage. For SHOS of lithic technology, the most complete excavations are those at Tybrind Vyg (Andersen 1980), Fermanville (Scuv6e & Verague 1988), the Florida sinkholes (Clausen et al. 1979; Cockrell, 1980), Atlit (Galili 1987), and Franchthi Cave (Gifford 1983). A general review of lithic SHOS and methods has been provided by Masters & Flemming (1983a), and for North America and the Caribbean by Johnson & Stright (1992). Many researchers have stressed that acoustic systems provide indispensable geomorphological evidence from the sea bed, and thus indicate where occupation sites, quarries, springs, and kill sites are likely to occur (Stright 1980, 1986; Rupp~ 1981; Gagliano et al. Van Andel & Llanos 1983, 1984; 1982; Flemming 1986; Pearson et al. 1986; Flemming & Kazianis 1987; Garrison 1992) in areas where SHOS are not already known. On land an archaeologist seldom searches at random, but uses topographical knowledge to select streambeds, fertile soil, spring lines, flint beds, defensible locations, etc., which are known to be associated with occupation sites in a given culture. Acoustic
139
remote sensing, photography, and coring from a surface vessel provide the marine archaeologist with similar data. Experimental surveys with sub-bottom profiling and frequency-modulated Chirp sonar show that very subtle changes in sub-bottom reflecting layers can be detected in the sediments surrounding the wreck of the M a r y R o s e (Quinn et al. 1997). This approach might reveal distinct characteristics in submerged shell middens. No researcher has succeeded in making the step from detecting a probable site by remote methods through to survey and excavation of an indisputable SHOS, without the assistance of extrapolation from known shoreline sites within l km, or arbitrary luck. This paradox was examined critically by Patterson (1981), who maintained that the OCS legislation required commercial companies to undertake pointless acoustic surveys which would never reveal archaeological data. Sixteen years later, the paradox remains. Divers off the coasts of Denmark, France, italy, Israel, Greece, Japan, the Bahamas, South Africa and Florida, are regularly excavating known lithic artefact sites, and occasionally finding new ones, without the aid of acoustic systems to speed discovery. The implications of this paradox were discussed by Masters & Flemming (1983b), and the reader is referred to that publication. The present author strongly supports the continued surveys by acoustic methods and coring, and particularly recommends trials of acoustic technology over known prehistoric SHOS and midden sites to identify unique signals. For sites younger than 40 ka BP the dating of archaeological artefacts by technological type, combined with radiocarbon, pollen, and stratigraphical indicators, has provided adequate dating comparable with that of other methods. Only one site, Table Bay, has produced older submarine artefacts, and these can only be dated very roughly by their technology and comparison with sites on shore to the period 0.7-1.2MaBP. Submarine archaeological sites associated with caves contain speleothems in several cases, and these can be dated by uranium-thorium for periods older than 40kaBP (e.g. Richards et al. 1994). Neolithic and Mesolithic sites contain sufficiently complex structures to define the relationship of the settlement to the contemporaneous waterline, as in the case of Atlit (Galili 1987), Tybrind Vig (Andersen 1980) and Aghios Petros (Flemming 1983b). Such features include alignments of wooden stakes to form primitive jetties or to retain banks, foundations which follow a coastal contour in a very sheltered locality, fish
140
N. C. FLEMMING
traps, cisterns and fresh-water wells, and freshwater springs within a site area, combined with evidence of marine exploitation. Given marine sediment deposits adjacent to and contemporary with the site boundary, the sea level can be deduced. In these cases the relative sea level can be determined with an error of the order of + 1 m in 10000 years. Older Palaeolithic sites contain no structures of habitation, and so can only be categorized as defining a point which was dry land at a certain date.
The value of submarine human occupation sites and detection of relative sea-level change Changes of coastal land level and sea level in the last 100000 years have been caused by the formation and melting of polar ice-caps (Fairbridge 1961; Chappell & Thorn 1977; Clark & Lingle 1978; Marcus & Newman 1983; Pirazzoli 1987; Tushingham & Peltier 1991; Watters et al. 1992; Lambeck 1995), minor climatic fluctuations (Jelgersma 1961; Gornitz et al. 1982; Flemming & Webb 1986), and Earth movements of various origins (Fairbridge 1961; Ambraseys 1971; Kraft & Aschenbrenner 1977; Le Pichon & Angelier 1981; Flemming & Webb 1986; Flemming & Woodworth 1988; Dvorak & Mastrolorenzo 1990; Erol & Pirazzoli 1992). The global sea level at the maximum of the last polar glaciation was about 100m lower than at present. Minor climatic fluctuations with durations of a few hundred to a few thousand years can cause sea-level changes from a few centimetres to about 10-20m. Vertical earth movements caused by earthquakes, vulcanism, aseismic accumulation of strain between earthquakes, isostatic loading from ice or sediments, and sediment compaction can have rates of the order of several millimetres per year or several metres per thousand years. It is generally not possible to tell whether the observed displacement of a single site is due to earth movements or sealevel change. Each observation provides the sum or difference of the two processes at that point. The effects of post-glacial hydro-isostasy have been discussed by Clark & Lingle (1978) and Tushingham & Peltier (1991). Early papers analysing responses of different continental margins to deglaciation and rise of sea level showed that the Mediterranean was in a zone where the sea approached present sea level at about 5 ka BP, and then remained approximately constant. For this reason, Flemming & Webb (1986) discounted the effect of glacial hydroisostasy. Lambeck (1995) has computed regional
trends in hydro-isostatic adjustment during the last 5000 years within the Aegean with a horizontal spatial resolution of 75km, and compared the results with publications on the region by Flemming and co-workers. Several of the broader trends revealed by fitting contoured surfaces to the empirical data by Flemming and co-workers, especially the apparent doming of the Peloponnese, were computed by Lambeck to be caused by hydro-isostasy. This is a convincing explanation. The finer-scale tilting of blocks such as western Crete or the island of Rhodes are still shown to be strongly controlled by tectonism. It is curious that the larger-scale hydro-isostatic model for the Mediterranean shows a slow rise of 1 mm per year for the last 6000 years (Lambeck 1995, p.1042), or a rise of 6 m in total, when the great majority of sites throughout the Mediterranean show no submergence. Lambeck (1995) did not cite the studies by Flemming (1969) or Flemming & Webb (1986), both of which show that the most common vertical displacement of archaeological sites throughout the Mediterranean in the last 5000 years is zero. It is understandable, though not necessarily correct, to suggest that some of the submerged sites in the Aegean are submerged because of hydro-isostasy rather than tectonics. However, this trade-off cannot be achieved in the western Mediterranean, or on the south coast of Turkey, or the coast of Israel, where no general submergence has occurred in the last 5000 years. The study of Mediterranean coastal sites, both submerged and not submerged, provides an example of the state of progress over the last 90 years. About 500 sites have only been catalogued or sketched very briefly (Flemming & Webb 1986; Flemming 1993; Blue & Flemming 1994). About 400 sites have been studied with proper archaeological surveys and analysis of technology of structures by researchers such as Yorke & Davidson (1969), Schmiedt (1972), Blackman (1973, 1982), Flemming (1973a), McCann (1987), Ximenes & Moerman (1988), Tina (1990), and others. The only city sites extensively surveyed and excavated stratigraphically underwater have been Halieis (Jameson 1969), Kenchreai (Scranton & Rammage 1964) and Caesarea (Oleson et al. 1984; Raban 1992). To obtain an estimate of relative sea-level change the second level of research and analysis, including outline mapping and dating of visible artefacts or structures, is necessary. With regard to lithic artefact SHOS, there is less experience of known sites, and a greater potential for new discoveries of an unexpected kind. To give the best picture of relative sea level change and
ARCHAEOLOGICAL EVIDENCE FOR TECTONIC MOVEMENTS coastal tectonics, future objectives are: (1) discovery and survey of lithic artefact SHOS in situations where identifiable levels provide the most information on periods for which there is little dated evidence from other sources, and from the middle and outer shelf regions; (2) survey and derivation of relative sea-level change at SHOS located in areas where measurement of coastal tectonics would be of particular value, such as in straits and channels which have changed in cross-sectional area or maximum sill depth during the late Pleistocene; (3) discovery and excavation of lithic SHOS which will extend the range and pattern of known occupation, hunting, subsistence, and trade in a region; (4) extension of SHOS analysis to sites which were based on the submerged shoreline and exploited marine resources using technology which is not yet typified for the Neolithic and Palaeolithic; (5) identification and excavation of SHOS in locations which are critical to the migration and distribution of the human race, and occupation of continents and islands. These objectives have been discussed broadly by Masters & Flemming (1983b). During recent years the work of Cockrell (1986), Galili (1987), Riccardi et at. (1987), Dunbar (1988), Scuv~e & Verague (1988), Clottes & Courtin (1992), Araki & Hayashida (1993) and Werz & Flemming (pers. comm. 1998) suggests that the successful achievement of these objectives is more likely rather than less. Dunbar et al. (1992) stressed that subaquatic sites have been shown to preserve a greater range of organic materials than land sites, which are valuable for dating. In the Mediterranean there are over 1200 known coastal archaeological sites (Flemming & Webb 1986) and the relative change of sea level has been measured at 335 of them. Recent fieldwork by S. Stiros, O. Erol, A. Pirazzoli, Haniotes, D. J. Blackman, E. Hadjidaki (pers. comm.) and others is continuously adding to the catalogue of Mediterranean palaeo-sea levels. The archaeological data give an estimate of the average rate of vertical change over centuries, and during such a time span the Earth movements can fluctuate, reverse or become quiescent. The result is that the net rate of change measured archaeologically is usually less than the rate measured by a tide gauge over 10-20 years. However, the long-term average may be more useful in predicting what is likely to happen over the next 50-100 years. The relatively high rates of vertical tectonic movement indicated by archaeological data for the last 5000-10 000 years, compared with the longterm Quaternary average rates, were used by Lambeck (1995) to suggest that the difference is
141
accounted for by hydro-isostasy. Although a hydro-isostatic component is inevitably present in the Mediterranean, however small, this factor does not explain why the neo-tectonic rates of movement measured by archaeological data are faster for both local uplift and local subsidence. Compared with the archaeological dataset of 335 sites, the tide gauge dataset, with 35 active gauges in the Mediterranean, is very small, with stretches of coastline of hundreds of kilometres between gauges where the Earth movements are undetected. Taking the two methods together it is possible to use the tide gauges for fine temporal structure and high-frequency events, while using the archaeological data to develop spatial interpolation and long-term trends. This method has been developed in connection with a project of the Mediterranean Action Plan of UNEP to provide information on the most probable future rates of change on the coast caused by the combined effects of climate change and local Earth movements (Flemming 1988, 1993).
Conclusions The study of submerged terrestrial human occupation sites has made rapid progress in the last two decades, with rapid increase in the depth of studied sites, increase in the number of known sites, and increase in the maximum age of known sites. Sites of all ages provide useful information on relative changes of land and sea level, from which the tectonic component can be extracted by statistical methods. The explorations by divers working seawards from the shore are beginning to overlap in depth and age with the geomorphological sites which are being mapped by remote sensing on the OCS. In general, there is not yet co-occurrence of the two approaches, because the diving surveys tend to be within 5-10 km of the shore, and the OCS remote sensing surveys tend to be farther offshore. Within the next 10 years the two techniques should become mutually reinforcing in the same site areas, with lithic artefact sites being discovered through a combination of acoustic remote sensing and diving. When this synthesis has been achieved it will be possible to consider extending the knowledge of the Palaeolithic occupation of the continental shelf to the greatest depths of the Pleistocene low sea levels, and to the deepest sills on human migration routes. The accumulated data on submerged and uplifted coastal archaeological sites can be exploited to calculate palaeo-sea levels, past rates of vertical Earth movements, and to predict probable vertical changes of the coast in the future.
142
N. C. F L E M M I N G
References
AKEROYD, A. V. 1972. Archaeological and historical evidence for subsidence in southern Britain. Philosophical Transactions of the Royal Society of London, Series A, 272, 151-169. ALLEN, J., GOLSON, J. & JONES, R. (eds) 1977. Sunda and Sahul, Prehistoric Studies in Southeast Asia, Melanesia and Australia. Academic Press, London. AMBRASEYS, N. N. 1971. The value of historical records of earthquakes. Nature, 232, 375-379. ANDERSEN, S. H. 1980. Tybrind Vyg, a preliminary report on a submerged Ertebolle settlement in the Little Belt. Antikvariske, 4, 7-22. ARAKI, S., & HAYASHIDA,K. 1993. The archaeological materials .from the Takashima seabed (Part II). Takashima Town cultural property research report. Educational Committee of Takashima Town, Nagasaki Prefecture. - & ISHIHARA, W. 1989. Distribution of seabed archaeological remains in Japan, and future problems. Proceedings of the Congress of the ConfOdOration Mondiale des Activit~s Subaquatiques in Nagoya, Vol. 3, 33-38. BEECHEY, F. W. 1827. Proceedings of the Expedition to Explore the North Coast of Africa. John Murray, London. BELL, M. 1991. The Severn Estuary Levels Research Committee, Annual Report 1990. BIRDSELL, J. B. 1977. The recalibration of a paradigm for the first peopling of Greater Australia. In: ALLEN, J., GOLSON, J. & JONES, R. (eds) Sunda and Sahul, Prehistoric Studies in Southeast Asia, Melanesia and Australia. Academic Press, London, 113-167. BLACKMAN, D. J, 1973. Evidence of sea level change in ancient harbours and coastal installations. In: Colston Symposium Papers, 23. Butterworth, London, 114-137. - - 1 9 8 2 . Ancient harbours in the Mediterranean. Part 1 and Part 2. International Journal of Nautical Archaeology, 11, 79-104 and 185-211. BLANC, A. C. 1940. Industrie musteriane e paleolithiche superiore nelle dune fossile e nelle grotte litorannee del Capo Palinuro. Rendiconti della Reale Accademia d'Italia, 10, 1. BLUE, L. & FLEMMING, N. C. (eds) 1994. UNEP-MAP catalogue of Mediterranean coastal archaeological sites, 2 vols, plus maps, photographs and disk database. (Unpublished, held by UNEP-MAP, Athens.) CHAPPELL, J. & THOM, B. 1977. Sea levels and coasts. In: ALLEN, J., GOLSON, J. & JONES, R. (eds) Sunda and Sahul, Prehistoric Studies in Southeast Asia, Melanesia and Australia. Academic Press, London. CLARK, G. 1969. Worm Prehistory. Cambridge University Press, Cambridge. CLARK, J. A. & LINGLE, C. S. 1978. Predicted relative sea-level changes (18,000BP to present) caused by late glacial retreat of the Antarctic Ice Sheet. Quaternary Research, 9, 265-287.
CLAUSEN, C. J., COHEN, A. D., EMILIANI, C., HOLMAN, J. A. & ST1PP, J. J. 1979. Little Salt Spring, Florida: a unique underwater site. Science, 203, 609-614. CLOTTES, J. ~ COURTIN, J. 1992. La Grotte Cosquer un an apr~s. Archeologia, 283, 14-19. - - - , BELTRAN,A., COURTIN, J. & COSQUER, H. 1992. The Cosquer cave on Cape Morgiou, Marseilles. Antiquity, 66, 583 598. COCKRELL, W. A. 1980. Drowned sites in North America. In: MUCKELROY, K. (ed.) Archaeology Underwater. McGraw-Hill, New York, 138-145. - - 1 9 8 6 . The Warm Mineral Springs archaeological research project: current research and technological applications. In: MITCHELL, C. T. (ed.) Diving for Science. American Academy of Underwater Sciences, 63-68. -& MURPHY, L. 1978. 8 SL17: methodological approaches to a dual component marine site on the Florida Atlantic coast. In: BARTO-ARNOLD,J. (ed.) Beneath the Waters of Time, 9th Conference on Underwater Archeology. Texas Antiquities Committee Publication, 6, 175-182. DEACON, H. J., GELEIJNSE, V. B., THACKERAY, A. I., THACKERAY, J. F. & TASENIUS, M. L. 1986. Late Pleistocene cave deposits in the Southern Cape. Current research at Klasies River. Paleoecology of Africa, 17, 31-38. DIOLI~, P. 1952. Promenades d'Arch~ologie sousmarine. Albin Michel, Paris. Translated as 4000 Years under the Sea. Sidgwick & Jackson, London. DUNBAR, J. S. 1988. Archaeological sites in the drowned Tertiary karst region of the eastern Gulf. The Florida Anthropologist, 41, 177-181. --, WEBB, S. D., & FAUGHT, M. 1992. Inundated prehistoric sites in Apalachee Bay, Florida, and the search for the Clovis shoreline. In: JOHNSON, L. L., & STR~GHT, M. (eds) Paleoshorelines and Prehistory: an Investigation of Method. CRC Press, Boca Raton, FL, 117-146. DVORAK, J. J. & MASTROLORENZO, G. 1990. History of vertical movement in Pozzuoli Bay, southern Italy': the result of regional extension related to evolution of the Tyrrhenian Sea and of local volcanic' activity. Geological Society of America Special Paper, 263. EMERY, K. O., & EDWARDS, R. L. 1966. Archaeological potential of the Atlantic continental shelf. American Antiquity, 31,733-737. EROL, O., & PIRAZZOLI,P. A. 1992. Seleucia Pieria: an ancient harbour submitted to two successive uplifts. International Journal of Nautical Archaeology, 21, 317-327. FAIRBRIDGE, R. W. 1961. Eustatic changes in sea level. In: Physics and Chemistry of the Earth, 4. Pergamon Press, London, 99-185. FlSCHER, A. 1993. Stenalderbopladser i Smalandsfarvandet (Stone Age settlements in the SmaIand Bight. a theory tested by diving). Miljoministeriet, Skov- og Naturstyrelsen, Kobenhavn. (ed.) 1995. Man and Sea in the Mesolithic. Coastal settlement above and below present sea level. In: Proceedings of the International Sympo-
ARCHAEOLOGICAL
EVIDENCE FOR TECTONIC MOVEMENTS
sium, Kalundborg, Denmark, 1993. Oxbow Monograph, 53, Oxford. FLEMM~NG, N. C. 1962. Sunken cities and forgotten wrecks. In: DEACON, G. F. (ed.) Oceans. Paul Hamlyn, London, 96-106. - - 1 9 6 8 . Holocene earth movements and eustatic sea level change in the Peloponnese. Nature, London, 217, 1031 1032. - - 1 9 6 9 . Archaeological evidence .for eustatic change in sea level and earth movements in the Western Mediterranean in the last 2000 years. Geological Society of America, Special Paper, 109. - - 1 9 7 1 . Cities in the Sea. Doubleday, New York. - - 1 9 7 3 a . Archaeological evidence for eustatic and tectonic components of relative sea level change in the South Aegean. In: Colston Symposium Papers, 23. Butterworth, London, 1 66. -(ed.) 1973b. Science Diving International. British Sub Aqua Club, London. --1978. Holocene eustatic changes and coastal tectonics in the north-east Mediterranean: implications for models of crustal consumption. Philosophical Transactions of the Royal Society of London, Series A, 289, 405-458. - - 1 9 8 1 . Preservation and rescue: challenges and controversies in the protection of archaeological resources. Journal of Field Archaeology, 8, 505. - - 1 9 8 3 a . Survival of submerged lithic and Bronze Age artefact sites: a review of case histories. In: MASTERS, M. & FLEMMING, N. C. (eds) Quaternary Coastlines and Marine Archaeology. Academic Press, London, 135-174. 1983b Preliminary geomorphological survey of an early Neolithic submerged site in the SPORADHES, N. Aegean. In: MASTERS, M. & FLEMMING, N. C. (eds) Quaternary Coastlines and Marine Archaeology. Academic Press, London, 233 268. - - 1 9 8 5 . Ice Ages and human occupation of the Continental Shelf. Oceanus, 28, 18 26. - - 1 9 8 6 . A survey of the late Quaternary landscape of the Cootamundra Shoals, North Australia: a preliminary report. In: FLEMMING, N. C., MARCHETTI, F. & STEFANON, A. (eds) Proceedings of the 7th International Diving Science Symposium of CMAS 1983, Padova, Italy, 149-180. --1988. Predictions of relative coastal sealevel change in the Mediterranean based on archaeological, historical and tide-gauge data. In: UNEP Joint Meeting on the Implications of Climatic Change, Split, Yugoslavia. (UNEP(OCA)/WG.2/13. --1993. Predictions of relative coastal sea-level change in the Mediterranean based on archaeological, historical and tide-gauge data. In: JEFTIC, L., MILLIMAN, J. D. & SESTINI, G. (eds) Climatic Change and the Mediterranean. Edward Arnold, London, 247-281. - - 1 9 9 6 . Sea level, neotectonics, and changes in coastal settlements: threat and response. In: RICE, E. E. (ed.) The Sea and History. Sutton, Stroud, 23 52. & KAZIANIS, D. 1987. Preliminary report on the Otranto Straits project. Report submitted to
143
the Department of Antiquities, Underwater Archaeology Section, Athens, Greece. & WEBB, C. O. 1986. Tectonic and eustatic coastal changes during the last 10,000 years derived from archaeological data. Zeitschrift j~ir Geomorphologie, Supplementband, 62, 1-29. & WOODWORTH, P. L. 1988. Monthly mean sea levels in Greece 1969-83 compared to relative vertical land movements measured over different timescales. Tectonophysics, 148, 59-72. - - , RABAN, A. & GOETSCHEL, C. 1978. Tectonic and eustatic changes on the Mediterranean coast of Israel in the last 9000 years. In: BARTO-ARNOLD, J. (ed.) Beneath the Waters of Time, 9th Conferenee on Underwater Archeology. Texas Antiquities Committee Publication, 6. FROST, H. 1963. Under the Mediterranean. Routledge and Kegan Paul, London. - - 1 9 7 2 . Ancient harbours and anchorages in the eastern Mediterranean. In: Underwater Archaeology, a Nascent Discipline. UNESCO, Paris, 95-114. GAGLIANO, S. M., PEARSON, C. E., WEINSTEIN, R. A., WISEMAN, D. E. & MCCLENDON, C. M 1982. Sedimentary studies of prehistoric archaeological sites. Coastal Environment Inc., Baton Rouge, LA. GALILI, E. 1987. A late Pre-Pottery Neolithic B site on the sea floor at Atlit. Mitkufat Haeven, 20, 50-71. GARRISON, E. G. 1992. Recent archaeogeophysical studies of the paleoshorelines of the Gulf of Mexico. In: JOHNSON, L. L. & STRIGHT, M. (eds) Paleoshorelines and Preshistory: an Investigation of Method. CRC Presss, Boca Raton, FL, 103 116. GEDDES, D. S., GUILAINE, J. & MONACO, A. 1983. Early Neolithic occupation on the submerged continental plateau of Roussillon, France. In: MASTERS, M. & FLEMMING, N. C. (eds) Quaternao' Coastlines and Marine Archaeology. Academic Press, London, 175-188. GIFFORD, J. A. 1983. Core sampling of a Holocene marine sedimentary sequence and underlying Neolithic cultural material off Franchthi Cave, Greece. In: MASTERS, M. & FLEMMING, N. C. (eds) Quaternary Coastlines and Marine Archaeology. Academic Press, London, 269-282. GNIRS, A 1908. Beobachtung fiber den Fortschrift einer s/ikularen Niveauschwankung des Meeres w~hrend der letzten swei Jahrtausend. Geographische Gesellschaft Wien Mitteilungen, 51, 1-56. GODWIN, N. 1941. Pollen-analysis and Quaternary geology. Proceedings of the Geologists' Association, 52, 328 361. GORNITZ, V., LEBEDEFF, S., & HANSEN, J. 1982. Global sea level trend in the past century. Science, 215, 1611-1614. G~NTHER, R. T. 1903. The submerged Greek and Roman foreshore near Naples. Archaeologia, 58, 1-62. HAMILTON, D. L. 1984. Preliminary report on the archaeological investigations of the submerged remains of Port Royal, Jamaica 1981-82. International Journal of Nautical Archaeology, 13, 11-25. HAMILTON, W. 1776-1779. I Campi Phlegraei. Peter Fabris. Naples. -
-
144
N. C. F L E M M I N G
HARDING, A., CADOGAN, G. & HOWELL, R. 1969. Pavlopetri, an underwater Bronze Age town in Laconia. Annual of the British School at Athens, 64, 113-142. HOFFECKER, J. F., POWERS, W. R. & GOEBEL, T. 1993. The colonization of Beringia and the peopling of the New World. Science, 259, 46-53 JAMESON, M. H. 1969. Excavations at Porto Cheli and vicinity, preliminary report I: Halieis. 1962-68. Hesperia, 38, 311-342. JELGERSMA, S. 1961. Holocene sea level changes in the Netherlands. Geologische Stichting Medelinge (Serie C), 6. JOHNSON, L. L., & STRIGHT, M. (eds) 1992. Paleoshorelines and Prehistory: an Investigation of Method. CRC Press, Boca Raton, FL. JONDET, G. 1916. Les ports submergOs de l'ancienne lle de Pharos. Memoires de l'Institut Egyptien, 9. JONES, R. 1977. Man as an element of a continental fauna: the case of the sundering of the Bassian Bridge. In: ALLEN, J., GOLSON, J. & JONES, R. (eds) Sunda and Sahul. Academic Press, London, 316-386. JOSENHANS, H. FEDJE, O., PIENITZ, R. & SOUTHON, J. 1977. Early humans and rapidly changing Holocene sea levels in the Queen Charlotte Islands, Hecate Strait, British Columbia, Canada. Science, 277, 71-74. KRAFT, J. C. & ASCHENBRENNER, S. E. 1977. Paleogeographic reconstruction in the Methoni embayment in Greece. Journal of Field Archaeology, 4, 19-44. - - , BELLKNAP, O. F. & KAYAK, I. 1983. Potentials of discovery of human occupation sites on the continental shelves and nearshore coastal zones. In: MASTERS, M. & FLEMMING, N. C. (eds) Quaternary' Coastlines and Marine Archaeology. Academic Press, London, 87-120. LAMBECK, K. 1995. Late Pleistocene and Holocene sea-level change in Greece and south-western Turkey: a separation of eustatic, isostatic and tectonic controbutions. Geophysical Journal International, 122, 1022-1044. LE PICHON, X. & ANGELIER, J. 1981. The Aegean Sea. Philosophical Transactions of the Royal Society of London, Series A, 300, 357-372. LONG, D., WICKHAM-JONES,C. R. & RUCKLEY,N. A. 1986. A flint artefact from the northern North Sea. In: Studies in the Upper Palaeolithic of Britain and North Western Europe. British Archaeological Reports International Series, 296. LOUWE KOOIJMANS, L. P. 1970/71. Mesolithic bone and antler implements from the North Sea and from the Netherlands. Bericht van het Rijksdienst Oudheidkundig Bodemonderzoek, Jaargang 22/21. MARASOVIC, J., MARASOVIC, T., MCNALLY, S. & WILKES, J. 1972. Diocletian's Palace, Part One. Smithsonian Institution and Yugoslav Institute for International Technical Co-operation, Belgrade. MARCUS, L. F. & NEWMAN, W. S. 1983. Hominid migrations and the eustatic sea level paradigm, a critique. In: MASTERS, P. M. & FLEMMING, N. C. (eds) Quaternary Coastlines and Marine Archaeology. Academic Press, London, 63-85.
MARSHALL, N. F. & MORIARTY, J. R. 1964. Principles of underwater archaeology. Pacific Discovery, 17, 18-25. MASTERS, P. M. 1983. Detection and assessment of prehistoric artefact sites off the coast of southern California. In: MASTERS, M. & FLEMMING, N. C. (eds) Quaternary Coastlines and Marine Archaeology. Academic Press, London, 189-214. -& FLEMMING, N. C. (eds) 1983a. Quaternary Coastlines and Marine Archaeology. Academic Press, London. 1983b. Towards the prehistory of land bridges and continental shelves: summary and conclusions. In: MASTERS, M. & FLEMMING, N. C (eds) Quaternary' Coastlines and Marine Archaeology. Academic Press, London, 601-609. MCCANN, A. 1987. The Roman Port and Fishery of Cosa. Princeton University Press, Princeton, NJ. MURPHY, L. 1978. 8 SLO 19: specialised methodological, technological, and physiological approaches to deep water excavation of a prehistoric site at Warm Mineral Springs, Florida. In: BARTO-ARNOLD, J. (ed.) Beneath the Waters of Time, 9th Conference of Underwater Archaeology. Texas Antiquities Committee Publication, 6, 123-128. NEGRIS, PH. 1904. Vestiges antiques submerg6s. Athenischer Mitteilungen. 29, 340-363. OLESON, J. P. & RABAN, A. 1989. The Harbours of Caesarea Maritima." Results of the Caesarea Ancient Harbour Project 1980-85. Vol. 1. The Site and Excavation, Parts I & II. BAR, Oxford. - - , HOHLFELDER,R. L., RABAN, A., & VANN, R. L. 1984. The Caesarea Ancient Harbour Excavation Project (CAHEP): preliminary report on the 1980-1983 seasons. Journal of Field Archaeology, 11, 281-305. OSAKI, E. 1973. Discovery of the remains of 17th century Japanese harbour works. In: FLEMMING, N. C. (ed.) Science Diving International. British Sub Aqua Club, London, 66-69. PASKOFF, R. & SANLAVILLE,P. 1983. Les Cdtes de la Tunisie, variations du niveau marin depuis le Tyrrhbnien. Maison de l'Orient Mediterranben, Lyon. PATTERSON, L. 1981. Outer Continental Shelf: preservation and rescue. Journal of Field Archaeology, 8, 231-232. PEARSON, C. E., KELLEY, D. B., WEINSTEIN, R. A. & GAGLIANO, S. M. 1986. Archaeological investigations on the outer continental shelf." a study within the Sabine River Valley, offshore Louisiana and Texas. US Department of the Interior, Minerals Management Service, OCS Study MMS 86-0019. PERDERSEN, L., FISHER, A. & AABY, B. (eds) 1997. The Danish Storebaelt since the Ice Age, Man, Sea, and Forest. A/S Storebaelt Fixed Link. Copenhagen. PIRAZZOLI, P. A. 1976. Les variations du niveau marin depuis 2000 ans. Memoire du Laboratoire GOomorphologique, lEcole Pratique des Hautes Etudes, 30, 1-421. 1987. Sea level changes in the Mediterranean. In: Sea Level Changes. Basil Blackwell, Oxford, 152-181.
ARCHAEOLOGICAL
EVIDENCE FOR TECTONIC MOVEMENTS
, AUSSEIL-BADIE, J., GIRESSE, P., HADJIDAKI, E. & ARNOLD, M. 1992. Historical environmental changes at Phalasarna harbour, West Crete. Geoarchaeology, 7, 371-392. POIDEBARD, A. 1939. Tyr, un grand port disparu. Paul Geuthner, Paris. QUINN, R., BULL, J. M., DIx, J. K. & ADAMS, J. 1997. The Mary Rose s i t e - geophysical evidence for palaeo-scour marks. Interntional Journal of Nautical Archaeology, 26, 3-16. RABAN, A. 1983. Submerged prehistoric sites off the Mediterranean coast of Israel. In: MASTERS, M. & FLEMMING, N. C. (eds) Quaternary Coastlines and Marine Archaeology. Academic Press, London, 215-232. - - 1 9 9 2 . Sebastos: the royal harbour at Caesarea - a short-lived giant. International Journal of Nautical Archaeology, 21, 111-124. RICCARDI, E., CHIZZOLA, S. & VICINO, G. 1987. Considerazioni sul rilievo della parte immersa della caverna a mare di Bergeggi (SV). In: Rivista di Studi Liguri, Anno LII, Numero 1-4, Bordighera, Forma Maris Antiqui, 403-411. RICHARDS, D. A., SMART, P. L. & EDWARDS, R. L. 1994. Maximum sea levels for the 1st glacial period from U-weries ages of submerged speleotherms. Nature, 367, 357-360. RICKMAN, G. E. 1988. The archaeology and history of Roman ports. International Journal of Nautical Archaeology, 17, 257-267. RUOFF, U. 1980. Alpine villages on stilts. In: MUCKELROY, K. (ed) Archaeology Underwater. McGraw-Hill, New York, 148-155. RUPPE, R. 1981. The archaeology of drowned terrestrial sites: a preliminary report. In: Bureau of Historical Sites and Properties Bulletin No. 6. Division of Archives, History and Records Management, Department of State, Tallahassee, FL, 33-45. SCHMIEDT, G. 1972. Il livello antico del Mar Tirreno. Leo Olschki, Firenze. SCRANTON, R. L., & RAMMAGE, E. S. 1964. Investigations at Kenchreai 1963. Hesperia, 33, 134-145. SEUVEE, F. & VERAGUE, J. 1988. Le gisement sousmarin du Pal~olithique Moyen de l'anse de la Mondr~e h Fermanville (Manche). Minist+re des Affaires Culturelles, Autorisation No. 001740 1971, CEHP-Littus, 306, Cherbourg. SHAW, T. 1738. Travels in Barbary. Johnstone, Edinburgh, 2 vols. SHEPARD, F. P. 1964. Sea level changes in the past 6000 years: possible archaeological significance. Science, 143, 574-576. SKAARUP,J. 1980. Undersoisk stenhalder. Tiddskriftet, 1, 3-8. SMYTH, W. H. 1854. The Mediterranean. John Murray, London. SORDINAS, A. 1983. Quaternary shorelines in the region of Corfu and adjacent islets, Western Greece. In: MASTERS, M. & ELEMMING, N. C. (eds) Quaternary Coastlines and Marine Archaeology. Academic Press, London, 325-334. STEERS, J. A. 1948. The Coastline of England and Wales. Cambridge University Press, Cambridge.
145
STOLTING, S. 1997. The boats of Slettne: sources of Stone Age shipbuilding in Northern Scandinavia. International Journal of Nautical Archaeology, 26, 17-25. STRIGHT, M. J. 1980. Federal cultural resources management on the OCS." problems and potential. Report of the US Department of the Interior, Bureau of Land Management. New Orleans Outer Continental Shelf Office, presented at 1 lth Annual Conference on Underwater Archaeology, New Mexico. 1986. Human occupation of the outer continental shelf during the late Pleistocene/early Holocene: methods for site location. Geoarchaeology, 1, 347-364. 1990. Archaeological sites on the North American continental shelf. Geological Society of America, Centennial Special Volume, 4, 439-465. TINA, M. N. 1990. Palaeoenvironmental history of submerged ruins on the northern Euboean Gulf coastal plain, central Greece. Geoarchaeology, 5, 323 -347. TUSHINGHAM, A. U., & PELTIER, W. R. 1991. ICE 3-G: a new global model of late Pleistocene deglaciation based upon geophysical predictions of post glacial relative sea level change. Journal of Geophysical Research, 96, 4497-4523. TUTHILL, C. & ALLANSON, A. A. 1954. Ocean bottom artefacts. The masterkey. (Southwest Museum, Los Angeles) 28, 222-232. VAN ANDEL, T. 1980. Late Pleistocene sea levels and human exploitation of the shore and shelf of southern South Africa. Journal of Field Archaeology, 16, 133 155. - & LLANOS, N. 1983. Prehistoric and historic shorelines of the southern Argolid peninsula; a sub-bottom profiler study. International Journal of Nautical Archaeology, 12, 303-324. - & 1984. High resolution seismic reflection profiles for the reconstruction of postglacial transgressive shorelines: an example from Greece. Quaternary Research, 22, 31-45. VITALI,V., GIFFORD, J. A., DJINDJIAN, F. & RAPP, G. JR 1992. A formalized approach to analysis of geoarchaeological sediment samples: the location of the Early Punic harbour at Carthage, Tunisia. Geoarchaeology, 7, 545-581. WAECHTER, J. D'A. 1964. Excavation of Gorham's Cave, Gibraltar 1951-54. Institute of Archaeology Bulletin, London, 4, 189. WATTERS, D. R., DONAHUE, J. & STUCKENRATH, R. 1992. Paleoshorelines and the prehistory of Barbuda, West Indies. In: JOHNSON, L. L. & STmGHT, M. (eds) Paleoshorelines and Prehistory: an Investigation of Method. CRC Press, Boca Raton, FL, 15-52. WEBB, R.E 1995. ODP Site 820 and the initial human colonisation of Sahul. Quaternary Australasia, 13, 13-18. WRESCHNER, E. E. 1977. Newe Yam, a submerged late Neolithic settlement near Mount Carmel. In: ARENSBURG, B . & BAR-YOSEF, O. (eds) EretzIsrael, 13. Israel Exploration Society, Institute of
146
N. C. F L E M M I N G
Archaeology of the Hebrew University, Jerusalem, 260-271. XIMENES, S., & MOERMAN, M. 1988. The Roman harbour of Laurons: buildings and structures. In: RABAN, A. (ed) Archaeology of Coastal Changes, Proceedings of the First International Symposium Cities on the Sea, Past and Present. Haifa, Israel,
September 1986. BAR International Series, 404, 229 252. YORKE, R. A., & DAVIDSON, D. P. 1969. Roman harbouts of Algeria, Mahgreb Project 1968. Expedition field report. ZERI, A. 1905. Monographia storica degli porti dell'antichita nelle peninsola Italia. Roma.
Ancient coastal installations and the tectonic stability of the Israeli coast in historical times EHUD
GALILI
& JACOB SHARVIT
Israel Antiquities Authority, Marine Archaeology Branch, P.O. Box 180, Atlit 30350, Israel (e-mail. udi@ isran tique, org. il) Abstract: Coastal archaeological structures are often used to identify coastal displacements
and relative land-sea changes. Settlement of foundations and collapse of coastal stone-built structures, however, can sometimes lead to misinterpretations of sea-level changes or neotectonic activity. Coastal rock-cut installations, on the other hand, because they are cut into the bedrock, are not threatened by settlement or collapse and can serve as reliable indicators for detecting land sea changes. A detailed survey of rock-cut installations along the Israeli coast shows that coastal pools, channels and quarries cut within the last 2500 years are today found at present-day sea level or very close to it and are still able to function. Archaeological and geological indications of major neotectonic activity along the Israeli coast during historical times were re-examined in this study, and were found to be inconsistent with the new data presented. The study concludes that the Israeli coast is relatively stable and that no significant neotectonic movement (> 0.2 m per 1000 years) has occurred during historical times.
Underwater and coastal research carried out in recent decades along the Israeli coast has revealed a vast amount of diverse data that contribute to our understanding of geological and geomorphological processes in the coastal region, as well as shedding light on the ancient navigation practices, commerce and material cultures of coastal communities. Despite this, major disagreements have emerged as to the extent of the neotectonic contribution to recent coastal changes. Relying on archaeological evidence, this study will reexamine the issue of the tectonic stability of the Israeli coast during historical periods. Scholars have argued that destruction of ancient harbours and coastal installations in Israel is associated with neotectonic activity that occurred in historical times (Raban 1976, 1981, 1985; Neev et al. 1978, 1987; Adler 1986; Neev & Emery 1991). This was based mainly on archaeological structures found today at elevations that prevent proper functioning, and on geological and geomorphological studies. However, ancient coastal structures may be displaced from their original elevation relative to sea level by various mechanisms, such as sea-level changes, tectonic activity, erosion and collapse, and settlement of foundations in unconsolidated sediments (Fig. 1). Various combinations of these mechanisms are also possible. In using archaeological records to re-assess the tectonic stability of the coast, this study has tried to disregard equivocal evidence, and instead uses only reliable data that allow the contribution of tectonics to be isolated from the impact of erosion, settlement and eustatic change.
Stone-built coastal structures, as distinct from rock-cut ones, are constantly weathered, eroded, and undermined, and are often displaced from their original elevation by waves and tidal currents. Also, many of the stone-built structures were originally constructed on the sea bottom in antiquity and, therefore, cannot serve as reliable markers for vertical coastal displacement. Coastal rock-cut installations, in contrast, are better indicators of the tectonic contribution to relative sea-level changes. This is because (a) they are cut into the bedrock and, therefore, are not threatened by settlement or collapse, and (b) although there are difficulties in dating the rock-cut installations (Blackman 1973), most scholars agree that the majority of them along the Israeli coast were hewn between 500BC and AD 1300. It is also widely agreed that in the last 2500 years there were no significant eustatic sea-level changes in the Mediterranean (Flemming et al. 1978). Many rock-cut installations have been discovered along the Mediterranean coasts (Blackman 1973). Figure 2 illustrates the main types of these installations along the Israeli coast and their association with sea level. Their use in identifying neotectonic activity can be relatively reliable if one is aware of the limitations of relying on these indicators alone. In particular, it is worth noting that the functions of many rock-cut installations are not yet clear. Also, rock-cut pools, channels or quarries located above present-day sea level do not necessarily indicate vertical coastal displacement, because many of them could have functioned a few metres above sea level. Pools at elevations of up to 2.5m above sea level, for
GALILI, E. & SHARVIT, J. 1998. Ancient coastal installations and the tectonic stability of the Israeli coast in historical times. In: STEWART,I. S. & VITA-FINZI,C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 147-163.
148
E. G A L I L I & J. S H A R V I T
Fig. 1. Main types of coastal changes affecting man-made structures: (a) sea-level changes; (b) tectonic activity; (c) erosion and collapse; (d) settlement of foundations in unconsolidated sediments.
Fig. 2. Schematic cross-section depicting the main types of rock-cut installations along Israel's Mediterranean coast and their typical relation to sea level.
TECTONIC STABILITY OF ISRAELI COAST
Fig. 3. Map of the study area showing the location of sea-lex;el markers (rock-cut installations, abrasion platforms) discussed in the paper.
149
150
E. GALILI & J. SHARVIT
example, can be fed naturally by wave splash, and can be manually maintained at even higher elevations. Coastal pools that operated with seawater through gravity (wash tanks) and still function today, however, are good indicators of stable tectonic conditions. Many of the pools situated along the Israeli coast (e.g. Akhziv, Yasaf River, Shiqmona, Caesarea, Fig. 3) have retained their functioning conditions: the floors of the pools lie about 1.5m below present-day sea level, their upper rims are very close to present-day sea level, and their feeder channels still supply seawater. Rock-cut seawalls (commonly called wave traps) were left behind by ancient quarrymen to protect the working area from water penetration and wave splash (Fig. 2). The rock was cut at sea level and the seaward edge was left at its original height, forming a seawall. These rockcut seawalls can thus be good indicators for sea level or tectonic movements, if significant differences in elevations are observed. Within the quarries, building stones were excavated down to the lowest level that economically justified the invested effort. Theoretically, building stones could also have been quarried as deep as 1.5 m below sea level (employing techniques similar to those used in excavation of the coastal pools). However, it seems reasonable to assume that the ancient quarrymen cut the stones as far down as sea level, rarely to a level slightly lower than low tide, and certainly no lower than 0.5 m below sea level. Excavation of stones deeper than 0.5m below sea level probably required a greater investment that did not justify the effort. Thus, a quarry or any other rock-cut installation that is today submerged at a considerable depth that prevents functioning can be used as a reliable indicator for vertical coastal subsidence.
Rock-cut installations along the Israeli coast Having taken into account the limitations and reliability of some of the archaeological evidence, and after clarifying the main principles of using rock-cut installations as shoreline markers, we now present an overview of rock-cut installations found along the Israeli coast, discussed from north to south (Fig. 3). Western Galilee The coastal kurkar (aeolian calcareous sandstone) ridge and offshore islands in western Galilee underwent extensive quarrying for building stones in antiquity. Large areas were levelled, leaving behind wide rock-cut abrasion platforms. In some cases, a rock-cut seawall was left in the
western side of the quarries for protection from waves. In some of the quarries, secondary rock cuttings were made to create pools for various purposes. The quarries and rock-cut installations were described in detail by various scholars (Flemming et al. 1978; Raban 1986a; Spier 1993). To exploit the quarry efficiently, the kurkar and beachrock were usually excavated to sea level at the time of functioning or slightly lower. Today, the lowest levels in the quarries are in most cases located at sea level, or very close to it, and rarely 0.3-0.5 m below sea level. Rock-cut pools were reported from Nahlieli Island (33~ 35~ ' 50"E), the Liman coast (33~ 35~ Akhziv (33~ 35~ Akhziv south, Yasaf River (32~ 35~ (Spier 1993), and also on Segavion Island located about 1500m off the Akhziv coast (33~ 35~ In all cases, the elevations of the pools are close to present-day sea level and in some cases the floors of the pools are lower than sea level, to allow water penetration by gravity. At Akhziv, for example, 50m northwest of the ancient tell (a local term for a large, artificial mound resulting from settled ancient site), a trapezoid-shaped pool (20 m long, 12 m wide and 0.9 m deep) with subdivisions (Raban 1986a) still functions today, and two rock-cut channels lying at present-day sea level continue to supply it with seawater. Elsewhere, some pools located at elevations of up to 1 m above present-day sea level were probably fed by seawater manually or by wave splash. Acre Demand for building stones resulted in an almost total destruction of the available exposed kurkar ridge in the vicinity of ancient Acre (32~ 35~ to 32~ 35~ ' 45"E). In particular, almost all of the kurkar coastal exposures from the western shores of the ancient city to the Yasaf River (a few kilometres to the north) were extensively quarried. A levelled platform (5-100 m wide) was thus formed and lies today at sea level (A in Fig. 4). In places, a rockcut raised seawall on the seaward edge of the quarry prevented water from penetrating into the working area (C in Fig. 4). Rock-cut pools of various sizes and shapes are found on the hewn sandstone platform near the ancient city walls and west of the modern city. The foundations of the Crusader city walls were laid on top of the platform (B in Fig. 4, Fig. 5). All the rock-cut quarries, pools and city-wall foundations are located at modern sea level or very close to it. In a few places, the floors of the pools and the lower surfaces of the quarries lie 0.3-0.7 m below present-day sea level.
T E C T O N I C STABILITY OF ISRAELI COAST
151
Fig. 4. Aerial view of Acre showing (A) the levelled platform of the ancient kurkar quarry, (B) the Crusader city wall's foundations, and (C) a rock-cut seawall. All the archaeological rock-cut installations are today found at elevations that allow them to function.
Fig. 5. Detailed view of city wall at Acre.
Bat Galim In a few places along the 600 m long stretch of beach at Bat G a l i m (32~ 34~ extending from the p r o m e n a d e in the south-west
to the municipal beach in the northeast, limestone quarries are found at present-day sea level. In some places the excavations reach as far as 0.3m below present-day sea level ( F l e m m i n g et al. 1978).
152
E. GALILI & J. SHARVIT
Shiqmona-Haifa
About 80m west of the Shiqmona tell (32~ ' 7"N, 34~ a round pool, 4.5 m in diameter (Fig. 6) is cut into the limestone abrasion platform (Spier 1993). The two channels that convey seawater by gravity into the pool still function today. The bottom of the pool is about 1.5m below present-day sea level and its upper rim is at the elevation of the abrasion platform. Traces of purple dye found at Shiqmona were reported by Elgavish (1994) and Karmon & Spanier (1988). Hence, the pool may have been used in the Roman-Byzantine periods for storing live Murex shells for the manufacture of purple dye.
Atlit
At Atlit (32~ 34~ foundations of the Crusader city walls that were constructed on the levelled kurkar rock are found close to modern sea level (Flemming et al. 1978; Ronen & Olami 1978). The western quay of the Phoenician harbour is at present-day sea level, as are sections of the northern breakwater (Flemming et al. 1978; Ronen & Olami 1978 (sites 80/4 and 80/6), Raban 1985). In the western section of the Atlit penin-
sula, pre-Crusader rock-cut pools and quarries were reported (Ronen & Olami 1978 (site 82)). The elevations of these installations are slightly (0.5-0.8m) higher than present-day sea level. About 1 km southwest of the Crusader fortress, on Melah Island (32~ 34~ channels, rock-cuttings and pools are found (Ronen & Olami 1978 (site 99)). Eight hundred metres south of Melah Island, rock cuttings and channels are also found (Ronen & Olami 1978 (site 100)).
M e ' a r o t River outlet
Two parallel channels (c. 30m long, 0.6m wide, 0.6 m deep) trending east-west are found on the northern side ofa kurkar peninsula at the Me'arot River outlet (32~ 34~ A few metres north of the channels, a row of squareshaped rock-cut depressions (c. 0.4 m x 0.4 m x 0.3m) are aligned parallel to the channels. The seaward outlets of the channels are located 0.2-0.3 m above present-day sea level and during storms they still supply seawater to the backshore (Galili & Sharvit 1994). In antiquity, therefore, they may have been used for salt production.
Fig. 6. An ancient rock-cut Murex shell piscina at Shiqmona-Haifa, located at an appropriate elevation for modern functioning.
TECTONIC STABILITY OF ISRAELI COAST
Habonim
153
and on the southern bank of Taninim River north of the tell, a few rock-cut pools are found (Spier 1993). All of the pools are located at elevations close to present-day sea level.
Two small (c. 1.5 m x 2.5 m x 0.2 m) rectangular pools are cut into an abrasion platform north of Habonim (32~ 34~ The elevation of the upper parts of the pools lies at present-day sea level, and their floors lie at 0.2m below sea level. Two rock-cut channels (c. 25m x 0.4m x 0.3m), whose function is unclear, occur at an elevation of 0.5-1.8 m above present-day sea level. The coastal kurkar ridge at Habonim was extensively quarried in antiquity for building stones and a quarry extends a few hundred metres along the coast (32~ 34~ 15"E to 32~ 34~ Some of the huge stones (2m x 1 m x 1 m) are still in situ. The elevations of the quarried areas range from 3 m above present-day sea level to the present-day sea level. In some places inside the quarries, pools of various shapes and sizes are observed at elevations of 0.5-2.5 m above present-day sea level; these pools are still fed by wave splash.
On an abrasion platform west of the Roman theatre at Caesarea (32~ 34~ ~15"E), lies an entire complex of pools and channels (Flinder 1976, 1985; Flemming et al. 1978). A large (35 m x 17m x 1.5m), rectangular rock-cut pool (Fig. 7), believed to be a piscina (Flinder 1976), and associated smaller pools and channels, are found today at present-day sea level or very close to it. The large pool and the channels that supply seawater to it by gravity still function today. To the south of Caesarea (32~ 34~ ~ 10"E), on an abrasion platform nicknamed 'the horseshoe rock', two rock-cut pools (5.5m x 9 m x 0.3m and 6 . 2 m x 3 . 7 r e x l m ) are found (Spier 1993). The upper parts of the pools lie at present-day sea level.
Dor
Tel Baruch
The site of Dor has been surveyed by various scholars, and the coastal archaeology and geology has been described in detail (Flemming et al. 1978; Sneh 1981; Sneh & Klein 1984; Raban & Galili 1985; Spier 1993). Along a 1800m long stretch of coast (32~ 34~ to 32~ 34~ located to the south, west and north of the ancient tell, many rock-cut installations and quarries are found. The installations consist of pools, channels and seawalls. In most cases, the installations lie at or above present-day sea level. The floors of a few pools and channels lie 0.1-1.5 m below present-day sea level (Sneh 1981).
Yonim Island On Yonim Island, offshore of Ma'agan Michael (32~ ~12"N, 34~ there are rock-cuttings (quarries?) and two rock-cut pools (5 m x 3 m x 0.25 m and 7 m x 6 m x 0.6 m). The floors of the pools are c. 1 m above present-day sea level (Spier 1993). According to Flemming et al. (1978), the rock cuttings similarly correlate relatively well with present-day sea level.
Caesarea
At Tel Baruch (32~ 34~ two rock-cut pools (2m x 3 m x 0.5 m) are found at elevations c. 0.5m above present-day sea level (Spier 1993).
Jaffa south Quarries of kurkar for building stones are found on a abrasion platform south of Jaffa (32 ~l'50"N, 34~ at present-day sea level. Small pools ( i m x 2 m x 0 . 3 m ) are located inside the quarries.
Yavneh Yam At Yavneh Yam (31 ~ 34~ (Fig. 8), kurkar quarries are situated on the coastline at, and up to 0.5 m above, present-day sea level. Small (1 m x 0.7 m x 0.2 m), rock-cut pools within the quarries also lie at present-day sea level.
Ruqeish A beachrock quarry at Ruqeish (31~ 34~ is situated at present-day sea level (Oren 1992).
Tel Qatifa Tel Taninim On abrasion platforms west of the archaeological site of Tel Taninim (32~ 35~
At Tel Qatifa (31~ 34~ a beachrock quarry is located at present-day sea level (E. Oren, pers. comm.).
154
E. GALILI & J. SHARVIT
Fig. 7. A Roman rock-cut pool at Caesarea situated at an elevation suitable for use at present-day sea level.
Fig. 8. An ancient kurkar quarry at Yavneh-Yam, located at present-day sea level.
Observations from other parts of the southeastern and eastern Mediterranean coasts Coastal rock-cut quarries and seawalls situated close to present-day sea level were reported by
Frost (1973) at Sidon (Lebanon) and Arwad (Syria). Rock-cut plateaux (probably quarries) lying at present-day sea level and slightly lower ( - 0 . 2 m ) were observed by the present authors at Alexandria (Egypt) (31~ 29~ However, a comprehensive underwater and
TECTONIC STABILITY OF ISRAELI COAST coastal study is needed in these regions before any conclusions concerning neotectonic activity can be reached.
155
Other coastal indicators along the Israeli coast Coastal wells
Summary
This review of rock-cut installations along the Israeli coast shows that ancient pools and channels, originally operated by gravity with seawater, are still found today at elevations suitable for operation. The lowest parts of coastal quarries and rock-cut seawalls in the study area are found at sea level, or very close to it. No rock-cut installation (quarry, pool, seawall, or channel) has been found in the study area submerged at a depth that prevents its functioning. Although the floors of some pools and channels are slightly below presentday sea level or below the tide limit, such an elevation is needed for proper functioning of the installations and does not indicate tectonic subsidence. Because no submerged rock-cut installations have been found on the sea bottom in Israel, we should consider the possibility that such installations have been eroded and, therefore, have disappeared. From our observations, however, it is evident that preservation of archaeological sites on the shallow continental shelf and foreshore zone along the Mediterranean coast of Israel is generally very good. Although high-energy coastal erosion processes are active in the foreshore region, thick sand deposits, originating from the Nile delta, appear to have covered the sites for thousands of years and thereby protected them from marine erosion and. destruction by marine organisms. However, sand quarrying along the coast, the construction of the Aswan High Dam and the development of other, marine projects (ports, marinas, etc.) have all artificially reduced the availability of sands. As a result, bedrock is being exposed on the sea-bottom and in the foreshore for the first time in thousands of years, and many archaeological sites are being discovered in a perfect state of preservation. Given these circumstances, if significant vertical neotectonic movements had occurred in the last 2500 years (as previous studies have suggested), we might expect to find some of the rock-cut installations well preserved on the sea floor, yet this is not the case. Instead, it seems that the ancient rock-cut installations more or less coincide with present-day sea level. This is also evident from other coastal installations that appear to have retained their original elevation with respect to sea level, and have not settled, eroded or collapsed.
The Pleistocene aquifer of Israel's coastal plain drains westward from adjacent foothills to the Mediterranean, which controls its level in a steady-state condition of flow (Galili & Nir 1993). The initial groundwater table slope is in the order of 1:1000 (Kafri & Arad 1978). The water level within the coastal wells, therefore, is close to sea level, or is slightly higher. Because the water table is tilting seaward at a gradient of 1:1000, the further inland a well is located, the higher the water level is within it (c. 0.1m elevation for every 100 m distance). Assuming that, in antiquity, the water table in the coastal region was also slightly above sea level (though by how much would depend on the distance of a well from the ancient coastline), it is possible to reconstruct former sea levels by studying the ancient wells. Recent studies of ancient coastal wells along the Israeli coast (Nir & Eldar 1986; Nir 1997; Tuweg 1997; S. Eliezer pers. comm.) have shown that there has been no significant neotectonic activity in the coastal zone during historical time. Instead, all vertical variations of water levels recorded in the coastal wells are within the range of global eustatic fluctuations.
Coastal stone-built structures
In spite of the potential for erosion and collapse along the Israeli coast, many stone-built coastal structures have retained their position relative to sea level, and could still function today. Examples include: the protective Crusader (eleventh-twelfth century AD) seaward city wall at Ashkelon (Mazor 1974); the foundations of Moslem (thirteenth century AD) fortifications at Herzlia-Appolonia (Mazor 1974); the high aqueduct and the Roman quay on the southern breakwater at Caesarea (Olami & Peleg 1977; Raban 1981, 1984); the northern breakwater and western quay of the Phoenician (third-sixth century AD) harbour and the foundations of Crusader city walls at Atlit's north bay (Flemming et al. 1978; Ronen & Olami 1978; Raban 1984, 1985); and the foundations, of the Crusader city walls (B in Figs 4 and 5), the Roman southern breakwater and head of the eastern Moslem (ninth-tenth century AD) breakwater ('Tower of Flies') (Fig. 9) at Acre (Flemming et al. 1978; Raban 1985). At Acre, stone-built structures (as distinct from rock-cut ones) that
156
E. GALILI & J. SHARVIT
Fig. 9. Map of the Acre peninsula showing the location of archaeological features that have retained their elevation relative to sea level and those that underwent settlement. P.D.S.L., present-day sea level.
did not retain their original elevations are usually those that were constructed on unconsolidated sediments and later underwent settlement and erosion.
Geomorphological m a r k e r s A survey of geomorphological features cut into bedrock (notches and abrasion platforms) in the study area (Fig. 3) reveals no evidence of submerged or emerged rock-cut notches (see Flemming et al. 1978). In contrast, numerous well-developed notches are found at present-day sea level, and similarly, all observed abrasion platforms in the study area coincide remarkably well with present-day sea level (see Lipkin & Safriel 1971). These data suggest relatively stable sea-level conditions in the last few thousand years (see Flemming et al. 1978) and indicate that the study area is tectonically stable.
Discussion In this section, we re-examine the geological and archaeological basis of previous indications of neotectonic activity at key sites within the study area and discuss possible alternative interpretations based on the results of our observations presented above.
Caesarea Previous scholars have identified several archaeological indications of neotectonic activity. In particular, ruins of the western section of the Herodian-Roman (first century Bc-first century AD) harbour are found submerged at a depth of 5 - 6 m below sea level (Raban 1981), though most of the ancient harbour constructions that were built on top of the kurkar ridge, have
TECTONIC STABILITY OF ISRAELI COAST maintained their original elevation (Raban 1976, 1981). R o m a n shipwrecks from the second century AO are found on top of the submerged Herodian breakwater, indicating that it was already ruined some 200 years after its construction (Raban 1976). In addition to the submerged portion of the harbour, a platform made of hewn stones was also found at a depth of 5 m below sea level (Raban 1976).
157
West of the modern harbour, offshore subbottom sounding profiles revealed a stair-like feature on the sea floor (Neev et al. 1973; Mart 1996b). Several boreholes in and around the harbour (Fig. 10) reveal a kurkar horizon that appears to be at a different elevations offthe coast and on land (Neev e t al. 1973). Both these geological observations have long been interpreted to indicate that a post-Roman neotectonic fault,
Fig. 10. Map showing the main offshore archaeological features of the Caesarea region, locations of boreholes and proposed fault mentioned in the text (modified after Neev et al. (1973)). (1) submerged western part of the Roman harbour built in the open sea on top of sand; (2) central part of the Roman harbour built on top of the kurkar ridge; (3) eastern onshore part of the Roman harbour.
158
E. GALILI & J. SHARVIT
trending N-S, is responsible for the subsidence of the western section of the Herodian harbour (Neev et al. 1973, 1987; Raban 1976, 1981; Nir 1985; Mart 1995a; Raban & Mart 1995). This archaeological and geological evidence for neotectonic activity, however, is open to reinterpretation. First, the existence at Caesarea of a stone-built platform made of hewn stones 5 m below present-day sea level does not necessarily indicate tectonic subsidence. Ancient port builders had the knowledge and capability to build structures on the sea bottom. Some of these structures were constructed of huge stones that had been laid carefully on the sea floor, as is the case with Acre's ancient 'Tower of Flies' breakwater head (Raban 1985). Indeed, the Roman historian Josephus Flavius described in detail the construction of the Caesarea harbour, noting: 'He (Herod) first lowered into 20 fathoms of water blocks of stone mostly 50 feet long, 9 deep and 10 broad, sometimes even bigger. When the foundations had risen to water level, he built above the surface a mole 200 feet wide' (Whiston 1960, p.453). Thus, it should not be surprising to find the stone-built platform on the sea bed at Caesarea. Second, although the stair-like feature observed in the seismic profiles seaward of the modern harbour at Caesarea has been identified as the tectonic fault responsible for the Herodian harbour's destruction (Neev et al. 1978; Mart 1996b), the coincidence of this feature with the western boundaries of the submerged kurkar ridge (Fig. 10) is highly suspicious. We suggest that this feature may represent a typical subbottom topographic profile of the western edge of the kurkar ridge, rather than a tectonic fault (see Arad et al. 1978). However, even if we accept the assumption that the seismic profile shows a fault rather than the ridge edge, the only conclusion that can be inferred from this is that faulting occurred after the formation of the sandstone ridge. The ridge was formed thousands of years before the Roman period and so any association of the fault with the destruction of the Herodian harbour is highly speculative. Third, geologists have shown that there is a good correlation and continuity between offshore and onshore drill cores, with no indication of intervening faulting (Arad et al. 1978). Although the kurkar horizon does appear at different elevations in the boreholes, it does not necessarily indicate faulting, as a westward tilting of the kurkar layer is also possible. Finally, as noted above, ancient coastal installations in the Caesarea region have maintained their original elevation with respect to sea level. These include the rock-cut pool of the
palace (Cleopatra pool; Fig. 7), the rock-cut pools south of Caesarea (see above), port installations (Raban 1981) and an aqueduct (Olami & Peleg 1977). On Yonim Island, north of Caesarea, rock-cut installations reported by Raban (1976) indicate the stability of this offshore zone. Relying on the same data, however, Raban (1986b) suggested that the offshore region of Caesarea and Ma'agan Michael island ('Hayonim' Island) is tectonically unstable. Overall, it seems that both archaeological and geological evidence support tectonic stability of the Caesarea region over the last 2000 years. If this is the case, what then caused the destruction of the western section of the Herodian harbour? Judging by the available data, we suggest the following sequence of events (Fig. 11). Originally, the western section of the Herodian harbour in Caesarea was built in the open sea on unconsolidated sediments (Fig. l lb), but subsequent marine erosion, followed by settlement of the foundations into the sediments, then caused subsidence of the west section of the Caesarea harbour (Fig. 1 l c) (Galili & Sharvit 1995a, b). A similar mechanism was also responsible for the destruction of the southern breakwater of the Phoenician port of Atlit (Galili & Inbar 1986). Acre
At Acre, subsidence at a rate of 0.6m per 1000 years is thought to have occurred since about AD 1300 (Flemming et al. 1978). A study of rockcut installations, however, suggests no indications of neotectonic activity in the western Galilee coastal region (Raban 1986a). Despite a submerged Crusader vault (arch) (Fig. 9) (Flemming et al. 1978) that was previously associated with post Mameluke neotectonic subsidence (Neev et al. 1987), as discussed above, other rock-cut installations (mainly quarries and seawalls) (A and C in Fig. 4) and stone-built structures at Acre (Fig. 9) have more or less retained their original elevations. Furthermore, a Crusader well recently excavated a few hundred metres inland of Acre coast has revealed a water-table level 0.6m above present-day sea level. Considering the distance of the well from the coastline, this elevation coincides fairly well with present-day sea level (see observations above). Hence, if any post-Crusader tectonic submergence had occurred in the region, we would expect to find all these archaeological features below present-day sea level, but this is not the case. The submerged Crusader vault is not a reliable indicator of neotectonic subsidence, because it could originally have been constructed on the sea bottom as the foundation
TECTONIC STABILITY OF ISRAELI COAST
159
Fig. 11. Schematic cross-sections of Caesarea port: (a) the region before the Herodian period; (b) the Herodian harbour installations; (c) present-day features in the harbour.
for a building. In any case, it is most unlikely that most of the archaeological features in the vicinity of this vault retained their elevation with respect to sea level and only the vault underwent tectonic subsidence. Although onshore archaeological and geomorphological features in the Acre region and in western Galilee indicate no significant neotectonic activity in the past thousands of years, E-W trending submarine faults and canyons have been observed offshore of Acre, Yasaf River, Shavai
Zion, Ein Sara and Akhziv (Galili & Eytam 1988; Galili & Sivan 1998). Here, 1-2m high, 300-600m long submarine scarps indicate that E-W striking faults observed on land (Kafri 1972; Mero 1983; Sivan 1996) continue offshore. Considering that the kurkar which the faults cut is likely to have formed in prehistoric times (possibly early Holocene or before), 2 m of vertical displacement within the Holocene period corresponds to a vertical displacement rate of only 0.2m per 1000 years. This is the same order of
160
E. GALILI & J. SHARVIT
magnitude as changes in eustatic sea level in the last 2500 years and tidal fluctuations (Flemming et al. 1978; Goldsmith & Gilboa 1986), so it is likely that such a tectonic contribution would be hard to identify within our present study.
than the archaeological layers, years of differential erosion by wave splash is likely to have eroded the ancient buildings and structures and left behind a levelled platform.
A tlit Dor
Archaeological indications of neotectonic activity at Dor were reported by Flemming et al. (1978), Neev & Bakler (1978), Sneh (1981) and Sneh & Klein (1984). However, in a review of archaeological and geomorphological studies, Sneh (1981) noted that any tectonic movement that had occurred in the Dor region during the last 3000 years was not obvious, and in any case did not exceed a few tens of centimetres (c. 0.1 m per 1000 years). Flemming et al. (1978) reported solution notches 0.3-0.9m above present-day sea level as an indication of neotectonic activity at Dor. These features are not typical wave-cut notches that can serve as a reliable markers for former sea levels, but are instead small horizontal grooves cut in a man-made structure. Sneh (1981) suggested that these grooves could be marks cut at a later stage of the structures. Alternatively, they could be traces of preferential weathering in the k u r k a r layers and, therefore, not suitable sea-level markers. Another alleged indication of neotectonic vertical displacement at Dor is a series of pools on Shehafit Island found at an elevation of 1 m above present-day sea level (Flemming et al. 1978). The function of many of the rock-cut pools is not yet clear; some of them may have been used for short-term fish storage and could have functioned above sea level (Spier 1993). Others could have been used for salt production, and as such should have been originally located above sea level to prevent waves from interfering with the dehydration process. Thus, the emergent pools at Dor cannot be reliable indicators for uplift. Our observations at Dor demonstrate that rock-cut coastal installations are found at elevations consistent with present-day sea level. Furthermore, no coastal abrasion platforms or notches were observed on the sea bottom or above sea level. An uplifted plateau described by Sneh (1981) as a Flandrian (mid-Holocene) marine terrace and reported at 3-4 m above present-day sea level at Dor (Neev et al. 1987) was examined in the course of this study. We suggest that this feature is a product of differential erosion of the archaeological site. The bedrock must have been artificially levelled in antiquity, and the installations and buildings then erected on top of it. With the bedrock being more resistant to weathering
The most ancient archaeological feature located along the Israeli coast is the early sixth millennium BC well uncovered at the submerged Neolithic site of Atlit-Yam (Galili et al. 1993). The reconstructed ancient water level in the well indicates that sea level at the time of occupation was about 16 m lower than present-day sea level (Galili & Nir 1993). This well, and a series of fifth millennium BC. Pottery Neolithic sites submerged today at depths of 1-7 m, allow reconstruction of the Holocene sea-level curve for the Carmel coast (Galili et al. 1988). This resulting curve more or less coincides with the global eustatic changes proposed elsewhere (e.g. Van Andel 1990; Stanley 1995). As a result, it seems reasonable to conclude that during the last 8000 years no major vertical tectonic displacement has taken place in the region (Galili et al. 1988). Several scholars (e.g. Gvirtsman et al. 1986; Neev et al. 1987; Mart 1996a) have suggested that neotectonic faults, with generally E - W or N-S trends, are responsible for subsidence in the area north of Atlit. For example, Neev et al. (1987) claimed that the Neolithic Atlit-Yam village, situated in the north bay of Atlit, subsided as a result of neotectonic activity. Mart (1996a), however, supported the interpretation by Galili et al. (1988) that the Neolithic well at Atlit-Yam indicates tectonic stability of the region during the last 8000 years. Adler (1986) stated that the Phoenician harbour installations in the north bay of Atilt are today submerged as a result of neotectonic activity that occurred in the last few thousand years. As mentioned above, however, the western quay of the Phoenician harbour in the northern bay and sections of the northern breakwater are found today at sea level. Foundations of the Crusader wall remain today situated at a proper elevation relative to sea level (Flemming et al. 1978, Ronen & Olami 1978). It is most unlikely that the whole section north of the Atlit peninsula underwent subsidence during the historical period, as suggested by Adler, whereas the western quay and the northern breakwater retained their original elevations. Today, the uppermost sections of the southern breakwater of the Phoenician harbour at Atlit are submerged at a depth of 1-1.5 m. Given the stability of associated features in the area, we suggest that
TECTONIC STABILITY OF ISRAELI COAST its foundations underwent settlement because they were constructed on unconsolidated sediments (Galili & Inbar 1986). In summary, we conclude that no significant neotectonic activity had taken place during the last 8000 years in the Atlit region. Yavneh- Yam ( Palmachim )
Neev et al. (1987) suggested that the western sec tion of the early to middle Bronze Age (thirtiethfifteenth century BC) rampart at Yavneh-Yam is now submerged under the sea as a result of neotectonic subsidence and showed the approximate location of drowned man-made structures that were previously observed by Z. Ben-Avraham (pets. comm. in Neev et al. (1987, fig. 23)). However, extensive (and continuing) underwater surveys carried out in the course of this study failed to reveal offshore man-made structures in the location proposed by Neev et al. (1987). In contrast, as discussed above, rock-cut quarries at Yavneh-Yam are located at suitable elevations for functioning (Fig. 8), and abrasion platforms on offshore reefs similarly coincide with present-day sea level.
Conclusions and implications The available archaeological coastline markers existing along the Israeli coast were re-examined in this study, which has focused particularly on the possibility that significant tectonic activity occurred in the region during the last 2500 years. Minor vertical displacements of rock-cut installations identified in the study area (on the coast and on offshore islands) appear to vary within the range of eustatic and tidal variations. None of this archaeological evidence indicates rates of tectonic movement exceeding 0.2m per 1000 years in the coastal region during historical times. Geological and geomorphological observations support our claim that the coast and adjacent seafloor in the region have been relatively stable in the last few thousand years. However, because by its nature the accuracy of this study is limited, minor tectonic displacement (less than 0.2m per 1000 years) is possible. We suggest that the main reason for the destruction of coastal installations and harbour constructions along the Israeli coastline is not neotectonic activity, but differential settlement and marine erosion. The potential tectonic instability of the Israeli seaboard during historical times, as suggested by previous studies (e.g. Neev et al. 1973, 1978, 1987; Flemming et al. 1978; Neev & Bakler 1978;
161
Gvirtzman et al. 1986; Levy et al. 1986; Neev & Emery 1991; Mart 1996a), is a crucial consideration for planning large-scale coastal development projects such as nuclear power stations. However, the results of this study suggest that the Israeli coast is tectonically relatively stable. However, planners should take into consideration the possibility of exceptional events, such as earthquakes and tsunamis (Nir 1985; Amiran et al. 1994), and structures should be properly constructed to guard against the effects of marine erosion and differential settlement. Coastal archaeological sites, in particular, should be protected from weathering and erosion, as many are currently undergoing rapid destruction. Sand quarrying and the construction of marinas, ports and other coastal developments, are the main reasons for accelerated marine erosion in historical times, so the availability of sand along the coastal zone is crucial to the survival and protection of modern coastal installations, and the preservation of ancient coastal sites. A large-scale action plan is needed, therefore, to prevent the reduction of sand resources along the coastline and shallow continental shelf. We wish to thank M. Weinstein-Evron, D. Neev, Y. Nit, D. Sivan, V. Spier, A. Kotser, E. Stern, N. Flemming and an anonymous referee for their assistance and helpful comments, and C. Vita-Finzi and I. Stewart for their useful remarks.
References ADLER, E. 1986. The subsidence of the kurkar ridges in the Carmel coast (in Hebrew). Teva VaAretz, 4,
13 16. AMIRAN, D. H. K., ARIEH, E. & TURCOTTE, T. 1994.
Earthquakes in Israel and adjacent areas: macroseismic observations since 100Bc. Israel Exploration Journal, 44, 260-305. ARAD, A., ECKER,A. & OLSHINA,A. 1978. The young (post lower Pliocene) geological history of the Caesarea structure. Israel Journal of Earth Sciences, 27, 142-146. BLACKMAN, D. J. 1973. Evidence of sea level changes in ancient harbour and coastal installation. In: BLACKMAN, D. J. (ed.) Marine Archaeology. Colston Papers, 23. Butterworths, London, 115 139. ELGAVlSH,J. 1994. Shiqmona on the Seacoast of Mount Carmel (in Hebrew). Hakibbutz Hameuhad Publishing House, Tel Aviv. FLEMMING, N., RABAN, A. & GOETSCHEL, C. 1978. Tectonic and eustatic changes on the Mediterranean coast of Israel in the last 9000 years. In: GAMBLE, J. C. & YORKE, R. A. (eds) Progress in Underwater Science, Vol. 3 (New Series) of the Report of the Underwater Association, Proceedings of the 11 th Symposium of the Underwater Association, 18-19 March 1977. British Museum (Nat-
ural History), London, 33-93.
162
E. G A L I L I & J. S H A R V I T
FLINDER, A. 1976. A piscina at Caesarea a preliminary Report. Israel Exploration Journal, 26, 77-80. - - 1 9 8 5 . The piscinas at Caesarea and Lapithos. In: RABAN, A. (ed.) Proceedings of the First International Workshop on Ancient Mediterranean Harbours, Caesarea Maritima. BAR International Series 25"7, 173-178. FROST, H. 1973. The offshore island harbour at Sidon and other Phoenician sites in light of new dating evidence. International Journal of Nautical Archaeology and Underwater Exploration, 2, 75-94. GALILI, E. & EYTAM, Y. 1988. Young faulting in the shallow continental shelf: evidence from northern Israel. Proceedings of the Annual Con['erence of the Israel Geological Society, Ein Boqeq, 30-31 (Hebrew/English). & INBAR, M. 1986. Underwater clay exposures along the Israeli coast and submerged settlements off the Carmel coast (in Hebrew). Proceedings of the 5th Coastal Symposium. Technion Israel Institute of Technology, Haifa, 39-48. -& NIR, Y. 1993. The Submerged Pre-Pottery Neolithic water well of Atlit-Yam, northern Israel, and its palaeoenvironmental implications. The Holocene, 3, 265-270. & SHARVVr,J. 1994. Classification of underwater archaeological sites along the Mediterranean coast of Israel: finds from underwater and coastal archaeological research. In: ANGELOVA, C. (ed.) Actes du Symposium International Thracia Pontica V, Sozopol 1991, 269-296. & 1995a. The destruction of ancient coastal installations and the stability of the Israeli coast during historical times. In: GALIL, B. S. & MART, Y. (eds) Proceedings of the 7th Annual Symposium on the Mediterranean Continental Margin of Israel, National Institute of Oceanography, Haifa, 28-32 (Hebrew/English). & - - 1 9 9 5 b . The destruction of ancient coastal installations and the stability of the Israeli coast during historical times. In: ARKJN, Y. & AWGAD, D. (eds) Proceedings of the Israel Geological Society Annual Meeting, Zikhron Ya'aqov, 27, 34 (Hebrew/English). & SWAN, D. 1998. Young tectonic activity in the Galilee coasts - the rate of vertical displacement according to underwater and coastal archaeological and geological finds. In: WEINBERGER,R. GAVRIELI, I., YECH1ELI, Y., PORAT, N. & AYALON, A. (eds) Proceedings of Annual Meeting, Mizpeh Ramon, Israel Geological SocieO' (Hebrew/English). WEINSTEIN-EVRON, M., HERSHKOVITZ, I., GOPHER, A., KISLEV, M., LERNAU, O., KOLSKA HOROWITZ, L. & LERNAU, H 1993. Atilt-Yam: a prehistoric site on the sea floor off the Israeli coast. Journal of Field Archaeology, 20, 133. & RONEN, A. 1988. Holocene sea-level changes based on submerged archaeological sites off the northern Carmel noast in Israel. Quaternary Research, 29, 36-42. GOLDSMITH, V. & GILBOA, M. 1986. Tides in Israel (in Hebrew). Horizons in Geography, 15, 21-46.
-
-
-
-
-
-
-
-
-
-
,
GVIRTZMAN, G., KLANG, A., ADLER, E., MICHAELSON, H. & KASHAI, E. 1986. The Atlit faults system: left latteral strike-slip faults and rotation of southern Carmel (in Hebew). Proceedings of the Israel Geological Society Annual Meeting, Ma'alot, 58-59. KAFRI, U. 1972. Nahriya, geological map, scale 50 000. Geological Survey of Israel, Jerusalem. & ARAD, A. 1978. Paleohydrology and migration of the ground-water divide in regions of tectonic stability in Israel. Geological Society of America Bulletin, 89, 1723-1732. KARMON, N. & SPANIEl, E. 1988. Remains of a purple dye industry found at Tel Shiqmona. Israel Exploration Journal, 38, 184-187. LEVY, Z., NEEV, D. & PRAUSNITZ, M. W. 1986. Late Holocene tectonic movements at Akhziv, Mediterranean coastline of northern Israel. Quaternary Research, 25, 177 188. LIPKIN, Y. & SAFRIEL, U. 1971. Intertidal zonation on rocky shores at Mikhmoret (Mediterranean Israel). Journal of Ecology, 59, t-30. MART, Y. 1995. Caesarea Maritima: unique evidence for earthquake patterns and sea level fluctuations in the Late Holocene. In: HOLUM, G. (ed.) Proceedings of International Symposium Caesarea Maritima - Retrospective after Two Millennia, 23-25. 1996a. Faults at the proximal continental shelf off Atlit, central Israel, and their neotectonic significance. Geo-Marine Letters, 16, 41-48. MART, Y. 1996b. Destructive earthquakes in historic times on the continental shelf off Caesarea., CMS News, Report 23, University of Haifa, Centre for Maritime Studies, 7-10. MAZOR, E. 1974. On the stability of the Mediterranean coast of Israel since Roman times: a discussion. Israel Journal of Earth Sciences, 23, 149-151. MERO, D. 1983. Subsurface Geology of Western Galilee and Zevulun Plain. Tahal Consulting Engineerings Ltd, Tel Aviv. NEEV, D. & BAKLER, N. 1978. Young Tectonic Activities along the Coast of Israel (in Hebrew). Hof Ve-Yam, Ha Kibbutz Hameuhad Publishing House, 9-30. -& EMERY, K. 0. 1991. Three oscillatory tectonic movements at Caesarea: one Byzantine and two Post Byzantine. Report GSI/3/90, Ministry of Energy and Infrastructure, Geological Survey of Israel, Jerusalem, 1-25. --, BAKLER, N. & EMERY, K. O. 1987. Mediterranean Coasts of Israel and North Sinai. Holocene Tectonism from Geology, Geophysics and Archaeology. Taylor and Francis, New York. MOSHKOVITZ, S., KAUFMAN, A., MAGARIZ, M. & GOPHNA, R. 1973. Recent faulting along the Mediterranean coast of Israel. Nature, 245, 254-256. --, SHACHNAI, E., HALL, J. K., BAKLER, N. & BENAVRAHAM, Z. 1978. The young (post Lower Pliocene) geological history of the Caesarea structure. Israel Journal of Earth Sciences, 27, 43-64. Nm, Y. 1985. The destruction of the Roman high level aqueduct and the Herodian harbour at Caesarea. In: RA~AN, A. (ed.) Proceedings of the First International Workshop on Ancient Mediterranean 1
-
-
:
T E C T O N I C S T A B I L I T Y OF I S R A E L I C O A S T
Harbours Caesarea Maritima. BAR International Series, 257, 185 194. - - 1 9 9 7 . Middle and Late Holocene sea-levels along the Israel Mediterranean c o a s t - e v i d e n c e from ancient water wells. Journal of Quaternary Science 12(2), 143-151. -& ELDAR, I. 1986. Ground water table levels in ancient wells as indications for ancient sea levels', and for neotectonic changes in the Mediterranean coastal region of Israel. GSI/34/86 Report, Geological Survey of Israel, Jerusalem, 1-28 (in Hebrew). OLAMI, Y. & PELEG, Y. 1977. The water supply system of Caesarea Maritima. Israel Exploration Journal, 27, 127 137. OREN, E. 1992. Ruqeish. In: STERN,E. (ed.) TheNewEncyclopediajbr Archaeological Excavations in Israel, Vol. 4. The Israel Exploration Society, Ministry of Defence, Carta Press, Jerusalem, 1293 1294. RABAN, A. 1976. Marine archaeological research at Caesarea: location of evidence for level changes of ancient building remnants (in Hebrew). Final Report 2/76, Maritime Studies, University of Haifa, 7-58. - - 1 9 8 1 . The Ancient Harbours of Caesarea (in Hebrew). Qadmoniot, 55-56, 80 88. - - 1 9 8 4 . The ancient harbours of Erez-Israel in Biblical times (in Hebrew). In: SCHILER, A. (ed.) Zeev Vilnai Book. Jerusalem, 241-253. - - 1 9 8 5 . The ancient harbours of Israel in Biblical times (from the Neolithic period to the end of the Iron Age). In: RABAN, A. (ed.) Proceedings of the First International Workshop on Ancient Mediterranean Harbours, Caesarea Maritima. BAR International Series, 257, 11-44. - - 1 9 8 6 a . Rock cut installations off the western Galilee coast (in Hebrew). In: YEDAYAH, M. (ed.) Kadmoniot HaGalil Hama'aravi. Ministry of Defence Publication, 209 234. - - 1 9 8 6 b . Archaeological evidence for ancient sea levels in the Mediterranean coast of Israel. In: Pro-
-
163
ceedings of the Israel Geological Society Annual Meeting, Ma'alot, 102 105. & GALILI, E. 1985. Recent maritime archaeological research in Israel. International Journal of Nautical Archaeology, 14, 321-356. & MART, Y. 1995. Caesarea, field trip no. 5 (in Hebrew). In: ARKIN, Y. (ed.) Israel Geological SocieO' Annual Meeting, Zikhron Ya'aqov, Field Trips Guidebook, 89-103. RONEN, A. & OLAM1, Y. 1978. Atlit Map. Archaeological Survey of Israel, Jerusalem, 37-58 (Hebrew/English). SWAN, D. 1996. Paleogeography of the Galilee coastal plain during the Quaternary. (in Hebrew) PhD thesis, Hebrew University, Jerusalem. SNEH, Y. 1981. The paleogeography and the history of the coast of Dor in the Holocene period. (in Hebrew). MA thesis, Haifa University. & KLEIN, M. 1984. Holocene sea level changes at the coast of Dor, Southeast Mediterranean. Science, 226, 831-832. SP1ER, V. 1993. Methods of growing and holding fish in fish ponds" (piscinas) and in other facilities along the coasts of Israel and Italy in the Roman period. (in Hebrew). MA thesis, University of Haifa. STANLEY, O. J. 1995. A global sea-level curve for the late Quaternary: the impossible dream? Marine Geology, 125, 1-6 TUWEG, R. 1997. Level of the well bottoms in the area of the internal harbour at Caesarea as an indirect index for locating sea level changes (in Hebrew). Proceedings of the Annual Meeting Mediterranean Continental Margins of Israel, 6 7. VAN ANDEL, T. H, 1990. Addendum to 'Late Quaternary sea-level changes and archaeology', Antiquit)', 64, 151-152. WHISTON, W. 1960. The Complete Works of Josephus Flavius, Wars of the Jews I xxi 6. Cregel Publications, MI. -
Regional vertical motion in Iceland 1987-1992, determined using GPS surveying G. R. F O U L G E R
& M. A. H O F T O N
Department of Geological Sciences, University of Durham, South Road, Durham DH1 3LE, UK (e-mail.
[email protected]) Abstract: Vertical motion in Iceland may arise from tectonic sources, e.g. large earthquakes
and spreading episodes, isostatic movements owing to the change in mass of icecaps, and secular sinking of the volcanic pile under its own weight. Regional Global Positioning System (GPS) surveys covering the eastern half of Iceland have revealed the vertical motion for the period 1987-1992. Although the errors associated with the vertical component are large, exceptionally good agreement between independent data processing and agreement with isolated terrestrial measurements lends credence to the somewhat surprising results. These are dominated by relative uplift of c. 12 cm in the northern part of the Northern Volcanic Zone and subsidence of up to c. 12cm elsewhere. This motion can be partially explained by recent tectonic and glacio-isostatic events. The most significant of these are the 1975-1984 Krafla spreading episode and isostatic uplift around the currently melting Vatnaj6kull icecap of c. 1 cm/year. However, a substantial amount of the observed motion cannot be explained by any known processes. It is concluded that hitherto unknown, largescale processes associated with the hotspot and plume, possibly continuing at depths of c. 75 km, may be responsible. Behaviour of this kind has been observed associated with the Yellowstone hotspot, which is broadly comparable with the Iceland hotspot. Future remeasurements of the GPS network in Iceland will be valuable for refining these observations. The spreading plate boundary in the North Atlantic is co-located with a large hotspot. This has resulted in locally excessive extrusive activity that has built a pile of basaltic volcanic rocks several kilometres thick over the last few tens of millions of years. This pile is sufficiently thick to rise to heights of up to c. 2 k m above sea level where it forms the 4 5 0 x 350km island of Iceland. Approximately 600km of spreading plate boundary, including over 20 spreading segments, are thus elevated above sea level, along with up to 350 km of the flanking plates on either side. The spreading segments are generally composed of a single central volcano through which passes a swarm, several tens of kilometres in length, of normal faults and fissures. The time-averaged half-spreading rate obtained from global plate models (e.g. DeMets et al. 1990) is c. 2 cm/year. Probably because of the unusually large thickness of the extruded layer, the plate boundary in Iceland is very complex compared with its submarine counterpart. Several phases of ridge migration are known to have occurred on a variety of spatial scales. M a n y off-ridge active volcanoes are recognized, along with a propagating-dying rift pair in south Iceland, a microplate and two major seismic zones that appear to perform the kinematic function of transform zones. The plate boundary in Iceland is associated with a wide variety of tectonic events that cause
substantial horizontal and vertical movements. These include earthquakes of up to magnitude 7.5 in the fracture zones (e.g. Einarsson et al. 1981) and magmatic intrusions along the spreading plate boundary (e.g. Bj6rnsson 1985). Horizontal and vertical movements of up to several metres result. Cyclic vertical motion, sometimes also of the order of metres, is associated with the inflations and deflations of magma chambers underlying active volcanoes (e.g. Tryggvason 1986). Large tectonic events, in addition to causing instantaneous deformation of the Earth's surface, also contribute to secular motion. To a first approximation, the structure of the Earth may be modelled as an elastic lithosphere overlying a viscoelastic asthenosphere. Large events, including earthquakes and volcanic intrusions, cause immediate, local elastic deformation of both the lithosphere and the asthenosphere with attendant loading of stress in both elements. Unlike the lithosphere, the asthenosphere is capable of flow and subsequently releases the stress within it by slow deformation. This deformation is known as post-event stress relaxation and it is transmitted to the surface of the Earth through traction at the lithosphere-asthenosphere boundary. The spreading plate boundary and the mantle plume that is presumed to underlie the Iceland hotspot cause asthenospheric temperatures to be elevated, which in turn causes the viscosity of the
FOULGER, G. R. & HOFTON, M. A. 1998. Regional vertical motion in Iceland 1987-1992, determined using GPS surveying. In: STEWART, I. S. & VITA-FINZI, C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 165-178.
166
G. R. FOULGER & M. A. HOFTON
asthenosphere to be 2-3 orders of magnitude lower than it is beneath the continents (e.g. Sigmundsson 1991; Hofton & Foulger 1996a). For this reason, the pattern of plate movements and post-event stress relaxation contrasts greatly with that in most other parts of the world. The rate of motion is very high shortly after a large event, but this decays to become immeasurably low within the space of a few decades. Because the motion from events is thus highly variable, depending on the time of occurrence, each must be modelled explicitly to assess its contribution to the total motion at a given time. A simple assumption of a time-averaged rate of plate motion is not adequate to model motion observed in Iceland on a decade time-scale. Other sources of motion in Iceland include predominantly vertical movements that result from isostatic readjustments to the expansion and contraction of the several icecaps in Iceland. The entire country was covered with an icecap during the Weichselian glaciation and isostatic rebound following the melting of that icecap between about 10 and 9kaBp was at least c. 100m (Sigmundsson 1991). More recently, melting during the twentieth century of small vestigial
icecaps is thought to be sufficient to be causing observable local isostatic uplift (Sigmundsson and Einarsson 1992). Added to this is the continuous subsidence of the volcanic pile under it's own weight (Palmason 1980). The subsidence rate is greatest close to the spreading plate boundary, where frequent volcanic activity adds rapidly to the volcanic pile, and it is least in distal areas that are volcanically extinct.
Regional GPS surveys in the eastern half of Iceland Since geodetic global positioning system (GPS) surveying was first applied to tectonic research in 1986, many surveys have been conducted in Iceland because its unique geological setting presents numerous research targets that cannot be addressed anywhere else in the world. The first survey was conducted in 1986, when a broad network of 51 points distributed over the whole country was measured by an international consortium of scientists (Foulger et al. 1993). In 1987, a much denser network was established and measured that was focused on the eastern
Fig. 1. Displacements of GPS points, 1987-1992, relative to a point arbitrarily fixed in the centre of the neovolcanic zone, calculated using the Bernese software. The ellipse at each arrowhead indicates the scaled formal errors at the 68% (1.5a) confidence level. The arrow at lower left gives the scale. The Krafla and Askja volcanic systems are shown schematically. Inset shows the neovolcanic zone in Iceland. NVZ, Northern Volcanic Zone; TFZ, Tj6rnes Fracture Zone.
REGIONAL VERTICAL MOTION IN ICELAND 1987 1992 half of Iceland only, including the Northern Volcanic Zone (NVZ) and the icecap Vatnaj6kull (e.g. Foulger et al. 1992) (Fig. 1). The objectives were to measure the deformation field over the whole region by differencing repeat surveys, and to relate it to physical processes. The 1987 network was remeasured in 1990, 1992 and 1995, during which time it was expanded and redesigned as constraint of the amplitude and extent of the deformation field improved. The fieldwork, data processing and interpretation of the results were conducted cooperatively by the University of Durham, UK, and the University of Hannover, Germany, with contributions to the fieldwork by Icelandic institutions for some surveys (Foulger et al. 1992; Heki et al. 1993; Jahn et al. 1994; Voelksen et al. 1995; Hofton 1995; Hofton & Foulger 1996a, b). Because GPS surveying was in its infancy in the 1980s and early 1990s, in particular the data processing, extreme care was taken at all stages of the work to ensure the integrity of the results. All points were measured independently multiple times in the field, to assess the repeatability of the calculated point coordinates (i.e. the variability of the results within a set of independent
167
measurements). Data processing was conducted independently with the Bernese software (Rothacher et al. 1990) at the University of Durham and with G E O N A P software (Wubebbena 1989) at the University of Hannover. Although this was a major undertaking, as the processing of data from each survey typically take several months, the close agreement of the results from completely independent processing using different software packages lent strong support to the results. The repeatabilities of the results obtained using the Bernese software were at or below the 1 cm level in the horizontal and below the 2 cm level in the vertical for all the surveys (Table 1).
Table 1. Repeatability of station coordinates at the lcr level in the three separate coordinates for the 1987, t990 and 1992 GPS surveys Survey
North south (cm)
East-west (cm)
Up-down (cm)
1987 1990 1992
0.64 1.01 0.71
0.92 1.10 0.62
1.67 1.93 1.90
Fig. 2. Vertical displacements of GPS points, 1987-1992, relative to a point arbitrarily fixed in the centre of the neovolcanic zone, calculated using the Bernese software. The vertical motions of the points, whose positions are at the corners of the triangles, are shown and bilinear interpolation and shading has been applied between them. It should be borne in mind that there are no observations between the points. Scale bar give the vertical displacement.
168
G. R. FOULGER & M. A. HOFTON
The error in the vertical component is unavoidably greater than in the horizontal because the GPS space vehicles occupy only the upper hemisphere around the observation points. This causes a trade-off during data processing of arrival time of the signal at the GPS receiver and receiver height, which results in the vertical component of point position being less well constrained than the horizontal position by a factor of c. 2-3. Processing of the 1995 survey data is incomplete at the time of writing.
Results Differencing the results of the surveys revealed significant relative horizontal and vertical motions (Figs 1 and 2). Only the 1987-1992 epoch yielded a statistically significant vertical deformation field. This was because the 1990 survey data were degraded by ionospheric noise, as 1990 was at the peak of a sunspot cycle. Furthermore, because of the inevitably greater error in the vertical component, larger movements, usually requiring a longer inter-survey period, are required to obtain a satisfactory signal-to-noise ratio.
In the horizontal plane (Fig. 1), rift-normal expansion of the network occurred around the NVZ in 1987-1992. The expansion rate was greatest at distances of about 20 km from the axis of the plate boundary, where it was about 4.4cm/year on average. In more distal parts of the network, and out to the edge of the network at about 150kin from the plate boundary, the expansion rate was lower, about 3cm/year for 1987-1992. Examination of the rates of horizontal motion for the epochs 1987-1990 and 1990-1992 showed that the expansion rate had slowed with time. The observed expansion rate was thus considerably greater than the timeaveraged plate rate. In the case of the vertical deformation field, substantial relative uplift (up to c. 12cm) occurred in and around the northern part of the NVZ. Little vertical motion occurred in a broad zone extending from the northern part of the network and covering much of the NVZ and it's eastern flank. Relative subsidence (up to c. 8 cm) occurred in the western parts of the network and the far eastern areas close to the coast (Fig. 2). The most extreme vertical motion is seen south and west of the Vatnaj6kull icecap at two points which subsided by c. 12 cm relative to the NVZ in general.
Fig. 3. Map of Iceland showing place names and the dates, locations and inferred dimensions of historic spreading events in the Northern Volcanic Zone.
R E G I O N A L VERTICAL MOTION IN I C E L A N D 1987-1992
169
Fig. 4. Predicted (a) horizontal and (b) vertical displacement including the elastic and the post-event stress relaxation motion caused by an infinitely long dyke, assuming an Earth model of a 10 km thick elastic lithosphere overlying a viscoelastic asthenosphere (a half-space). The dyke is assumed to rupture the whole of the lithosphere. The elastic response is shown separately, along with post-event cumulative motion at a suite of times following the spreading event. ~-~ is the relaxation time, where "ra = 2~7/#n, ~j is viscosity and #h is the elastic modulus of the asthenosphere. For Iceland the relaxation time is c. 1.7 years.
Significant sources of contemporary deformation in Iceland
Introduction The main tectonic element in the eastern half of Iceland, the NVZ, is composed of an array of five en echelon spreading segments that is
truncated at its n o r t h e r n end by the Tj6rnes Fracture Zone (TFZ) at the coast and vanishes beneath the Vatnaj6kull icecap at its southern end (Figs 1 and 3). Historically d o c u m e n t e d spreading events have activated three segments since the beginning of the seventeenth century and major earthquakes occur regularly in the TFZ. A d d e d to this are possible isostatic m o v e m e n t s caused by
170
G. R. FOULGER & M. A. HOFTON
recent changes in the volume of the Vatnajakull icecap in the southern part of the study area.
The pattern of deformation associated with spreading episodes Spreading episodes typically last a few years and cause crustal widening of the order of metres by the intrusion of dykes along the axis of the spreading plate boundary. They are the mechanism by which episodic motion occurs at spreading plate boundaries and are thus analogous to large earthquakes at transform plate boundaries or subduction zones. The Earth responds immediately elastically to large earthquakes and spreading episodes, and major horizontal and vertical movements result. In the case of spreading events, horizontal motion is everywhere away from the locus of the dyke intrusions whereas the vertical motion is dominated by uplift (Fig. 4). Post-event stress relaxation causes continued motion in the same sense in the horizontal plane, and has been dealt with in detail elsewhere (e.g. Hofton & Foulger 1996a). In the vertical plane a variable pattern of uplift and subsidence occurs, which affects a progressively broadening area.
Historical spreading episodes in the Northern Volcanic Zone Phenomena that are interpreted as indicating spreading episodes have been reported from the Theistareykir system in 1618, the Krafla system in 1724-1729 and the Askja system in1874-1879 (Fig. 3). It is naturally difficult to gauge the thickness of the dykes intruded and so estimates were made assuming that the average widening in the NVZ has been 2 cm/year over the last few centuries and that the dyke bundles intruded taper where the spreading segments overlap. As these events occurred a long time ago, their contributions to contemporary motion are very small and these assumptions are thus adequate for the present study. The vertical motion calculated for these historical spreading episodes for the period 1987-1992 is shown in Figs 5-7. A thickness of 10 km was used for the elastic lithosphere and a viscosity of 1.1 • 10 is Pa s for the viscosity of the asthenosphere. These values were obtained from geological evidence and from modelling of the horizontal deformation field (Hofton & Foulger 1996a). In all cases the maximum amplitudes of relative motion are less than 1.5cm. The width of the deformation field increases with the time
Fig. 5. Vertical motions, 1987-1992, as a result of post-dyking stress relaxation following the 1618 Theistareykir spreading episode.
R E G I O N A L V E R T I C A L M O T I O N IN I C E L A N D 1987-1992
Fig, 6. Same as Fig. 5 except for the 1724-1729 Krafla spreading episode.
Fig. 7. Same as Fig. 5 except for the 1874-1879 Askja spreading episode.
171
172
G. R. F O U L G E R & M. A. HOFTON
since the event, whereas its amplitude decreases. The Askja episode, which was relatively far inland, caused negligible relative movement at the coasts but episodes closer to the coast and more recent in time have a larger influence.
Contemporary activity in the Askja magma chamber Substantial vertical motions have been detected in and around the Askja central volcano in recent decades and have been interpreted as activation of a magma chamber beneath the volcano (Rymer & Tryggvason 1993). Vertical motion for the period 1987-1992 because of activity in the Askja magma chamber was calculated using a simple Mogi model (Mogi 1958) (Fig. 8). The effect of this source was very local, as the centre of inflation was estimated to be very shallow (2.8 km).
Large earthquakes in the Tj6rnes Fracture Zone Four earthquakes with magnitudes of 6-7 are known to have occurred this century, in 1910, 1934, 1963 and 1976. The TFZ is a right-lateral
strike-slip zone and the earthquakes were assumed to have such a mechanism, although reliable focal mechanisms are available only for the more recent events. The contribution to vertical motion for the period 1987-1992 was calculated using the same Earth model as was used to model the spreading events. The combined deformation field for all four earthquakes is shown in Fig. 9. Because the earthquakes are assumed to be strike-slip, relative vertical motion was vanishingly small.
The 1975-1984 spreading episode in the Krafla spreading segment The most important tectonic event to have occurred in Iceland recently is the 1975-1984 spreading event in the Krafla segment. Unlike the historic spreading episodes, this episode was very well researched and documented and detailed quantitative observations are available (see Bj6rnsson (1985) for a summary). During the period of activity the c. 80km long fissure swarm that marks the locus of spreading in the Krafla segment widened by 2-8 m at the surface as a result of lateral dyke injections from an activated magma chamber beneath the Krafla central volcano. In total, some 0.6 km 3 of magma
Fig. 8. Same as Fig. 5 except for deflation of a point source at 2.8 km depth beneath the Askja volcano.
R E G I O N A L V E R T I C A L M O T I O N IN I C E L A N D 1987-1992
173
Fig. 9. Same as Fig. 5 except for post-seismic stress relaxation following four large, recent earthquakes in the Tj6rnes Fracture Zone.
Fig. 10. Same as Fig. 5 except for the 1975-1984 Krafla spreading episode.
174
G. R. FOULGER & M. A. HOFTON
was injected laterally into the crust along the whole length of the segment, from which it may be estimated that the height of the dykes was 2-3 kin. Modelling suggests that dyking below this also occurred, supplied by magma migrating vertically up along ruptures that opened in response to the concentration of extensional stress at the bottom of the shallow dykes. Widening along the axis of the segment was accompanied by contraction of the flanks out to a few tens of kilometres by an almost equal amount in the manner predicted for elastic behaviour. The majority of the c. 22cm of relative horizontal motion measured using GPS surveying in 1987-1992 could be modelled as resulting from post-event stress relaxation following this spreading episode and this work also yielded estimates of the viscosity of the asthenosphere and the dimensions of the dyke bundle (Foulger et al. 1992; Heki et al. 1993; Hofton & Foulger 1996a, b). The corresponding predicted vertical deformation field for 1987-1992 exhibits uplift of the centre of the Krafla segment and subsidence of the flanks with relative amplitudes of c. 12cm (Fig. 10). It is instructive to compare Fig. 10 with Figs 5-7. The progressive broadening and decrease in amplitude with time of the vertical motions associated with spreading episodes is clearly seen. The width of the vertical deformation field calculated to result from the c. 10-year-old Krafla spreading episode is less than c. 100kin because of the shortness of the period of time that had elapsed since the spreading. The only coastal areas to be affected were those to the immediate north.
Isostatic motion around the VatnajOkull icecap There is thought to have been substantial melting of the Vatnaj6kull icecap during the twentieth century, amounting to a decrease in average thickness of c. 20 m. This is predicted to be sufficient to cause measurable local isostatic uplift (Sigmundsson and Einarsson 1992). Terrestrial surveying measurements were made at Lake Langisjdr in 1991 in an effort to confirm this (Fig. 11). Langisj6r is 20 km long, located at the SW edge of Vatnaj6kull, and oriented radially with respect to the centre of the glacier. In 1991 benchmarks originally surveyed in 1959 were resurveyed, which revealed that the northern end of the lake had been elevated by 12.4-t-3cm relative to the southern end in the intervening 32 year period. This tilt of 0.19 ram/ km per year is consistent with isostatic rebound if the viscosity of the asthenosphere is 10 is to
5 x 10 is Pas (Sigmundsson & Einarsson 1992). Further evidence for isostatic rebound is the reported shallowing of the harbour at the fishing village H6fn, southeast of Vatnaj6kull, estimated to be c. 2 cm/year (Einarsson et al. 1996) (Fig. 11). For the case where asthenosphere viscosity is 10 is Pa s relative vertical motion across the GPS network is predicted to be up to c. 20 cm for the period 1987-1992 (Fig. 11). For the case where asthenosphere viscosity is 1019pas relative vertical motion is only predicted to be c.10cm (Fig. 12). Regardless of what viscosity is used, the contours of vertical motion are predicted to be broadly concentric about the centre of Vatnaj6kull.
Subsidence o f the volcanic pile The continuum model of Palmason (1980) suggests that the Icelandic crust is gradually subsiding under its own weight as new lava is piled onto the surface. Detailed geological mapping in eastern Iceland has provided relatively accurate estimates of the deposition rate and dip of lavas there, from which the subsidence rate may be estimated. The deposition rate is estimated to be 620m per 106 years and, making the approximate assumption of constant topography, a subsidence rate of 0.062cm/year is obtained, kavas in eastern Iceland with an approximate age of 10Ma dip towards the NVZ as a result of subsidence at a mean angle of 3.9 ~ (Watkins & Walker 1977). These lavas are c. 100kin from the NVZ, which implies a subsidence rate at the neovolcanic zone of c. 0.068 cm/year. A rough subsidence rate of 0.065cm/year is thus implied for lavas in the NVZ, decreasing to zero in distal areas where volcanic activity is extinct.
Discussion The combined effects of all of the tectonic sources of vertical motion modelled above cannot account for the observed field. After subtracting the effects of all the spreading episodes, large earthquakes in the TFZ and motion related to activity in the Askja magma chamber, residual relative vertical motions of up to c. 20cm still remained (Fig. 13). The residual field (Fig. 13) bears little resemblance to the fields predicted for isostatic uplift around Vatnaj6kull for different viscosities of the asthenosphere (Figs 11 and 12).
REGIONAL
VERTICAL MOTION IN ICELAND
1987-1992
175
Fig. 11. Vertical displacements, 1987-1992, of points of the GPS network as a result of isostatic uplift following the removal of ice load during the twentieth century. An asthenospheric viscosity of 10 L8Pas is assumed. The icecap is assumed to be circular, centred at 64.40~ 16.68~ and to have thinned by an average of 20m.
Fig. 12. Same as Fig. 11 except using a viscosity of 1019 Pa s for the asthenosphere.
176
G. R. FOULGER & M. A. HOFTON
Fig. 13. Residual vertical motions, 1987-1992, after subtracting the effects of known spreading episodes, large earthquakes in the Tj6rnes Fracture Zone and activity in the Askja magma chamber.
Whereas circular contours centred on Vatnaj6kull are predicted, the observed residual motion in the neighbourhood of Vatnaj6kull is dominated by tilt up to the NE at a rate of c. 0.15 ram/kin per year. Although they do not show the regional pattern of deformation expected, the GPS results do agree with both terrestrial observations cited in support of the isostatic uplift model, e.g. tilt up to the NE at a rate of c. 0.19 ram/kin per year at Langisjdr, and uplift of c. 2 cm/year at H6fn, SE of Vatnaj6kull, relative to surrounding areas. This agreement lends confidence to the vertical GPS results, a valuable additional check in view of the relatively poor quality of vertical GPS results. Secular subsidence of the lava pile under it's own weight is predicted to result in sinking in the NVZ relative to distal areas. The residual vertical field shows the opposite, i.e. uplift in general of the NVZ of typically c. 5-8 cm relative to distal areas for the period 1987-1992. This implies a rate of uplift of c. 1-1.5 cm/year compared with the predicted rate of subsidence of c. 0.065cm/year. Gravitational sinking of the lava pile thus cannot explain the observations either, and a process hitherto unconsidered must be responsible. Errors in the GPS-derived vertical deformation field, or an unmodelled and large-scale pro-
cess are candidate explanations for the motion observed. Despite the large errors inherent in the vertical component of GPS results, the former explanation is unlikely. In addition to the agreement with terrestrial measurements at Langisj6r and H6fn, the results obtained from processing the GPS data independently and using a different software package at the University of Hannover show a strikingly good agreement with the Bernese results interpreted here (Fig. 2). If the vertical deformation field calculated is indeed in error, and it is never possible to entirely rule this out, then the error source must be one that is common to both processing systems. This rules out the most obvious sources, e.g. the troposphere model used, and strengthens the case for assuming that the vertical deformation results are real. Hitherto unknown processes associated with the hotspot and plume thus provide the most likely explanation for the vertical deformation observations that defy explanation by the known processes discussed above. The general uplift of the centre of the network relative to the flanks could be explained by a regional inflow of material at depth beneath the neovolcanic zone in the period 1987-1992. The full lateral extent of the deformation field may not have been detected, but it appears to be at least the width
REGIONAL VERTICAL MOTION IN ICELAND 1987-1992 of the network, i.e.c. 300 kin. The half-width of one flank is then c. 75 km, which gives a rough idea of the depth scale of a possible regional influx of material. The volume of the uplift bulge indicates the volume of the possible influx, and is 2-3 km 3. Such an explanation for the observations is an ad hoe model. Clearly, the sense and rate of vertical deformation could not continue indefinitely, as it is opposite in sign to and an order of magnitude more rapid than the known secular rate averaged over periods of the order of 1 • 106 years (Palmason 1980). However, because before the advent of GPS the tools for measuring such fields were not available, very little is known about the sense and rate of vertical motion in tectonic provinces over regional areas and decade time-scales. There are thus few other case histories with which to compare the observations in Iceland. An exception to this is the Yellowstone hotspot, which has been surveyed repeatedly since 1923 using terrestrial and, more recently, GPS methods (see Smith & Braile (1994) for a summary). Yellowstone is a continental hotspot marked by a 600 km wide topographical bulge 600 m high and a 150 mGal Bouguer gravity low centred on the 70 km wide Yellowstone caldera. The Iceland topographical high is at least 900 km wide and 3000m high, and is marked by a 100 mGal Bouguer gravity low. The scales of the two hotspots are thus very broadly comparable. For the period 1923-1977 uplift occurred over the Yellowstone hotspot at an average rate of 1.4cm/year, increasing to 3.1cm/year for the period 1977-1985, and reversing to subsidence at a rate of 3.1 cm/year for the period 1985-1990. The deformation was centred on the Yellowstone caldera, but the deformation field was at least 120km in diameter. These observations were interpreted by Smith & Braile (1994) as the migration of melts and/or hydrothermal fluids into the mid- and upper crust at Yellowstone and to characterize a giant volcanic system at unrest. The observations made at Yellowstone show that regional vertical motions, at substantial rates, and sustained for periods of the order of decades may occur at hotspots. Viewed in this light, the motion observed in Iceland is not implausible and perhaps not unusual. As is the case at Yellowstone, it is not necessarily related to any imminent increase in volcanic activity. Vertical motion associated with plume processes may, however, obscure motion from other processes. In particular, the disappointing correlation of vertical motion observed around Vatnaj6kull with predictions from isostatic uplift models may be explained in this way.
177
Implications for coastal vertical motion Contributions to coastal vertical motion in Iceland may be expected from many sources, all of which must be taken into consideration when modelling observations. Large earthquakes in the fracture zones of Iceland contribute little to vertical motion as they are predominantly strike-slip, with normal-faulting earthquakes in the spreading segments being of lesser magnitude. Spreading episodes make major contributions to vertical deformation, on the other hand, and where the accretionary plate boundary approaches the coast, this must make a substantial contribution to vertical coastal movements. The Reykjanes Peninsula (RP in Fig. 1) in the extreme southwest and the coast of north Iceland are the zones most affected. These are also the areas most affected by sinking of the volcanic pile under it's own weight, which is greatest in the neovolcanic zones. Isostatic rebound resulting from the melting of glaciers during the twentieth century is significant, but likely to affect sea level only in southeast Iceland, where Vatnaj6kull, the largest icecap in Iceland, lies close to the sea. Observations in the period 1987-1992, on the other hand, show that the largest relative vertical movements to occur on a scale of a few years may come from large-scale, poorly understood, and currently unpredictable processes in the hotspot-plume system as a whole. Coastal tilt during this observation period was as large as 0.7 #tad/year along the coast south of Vatnaj6kull, and much of this probably arose from such processes. The possibility of such behaviour should therefore be taken into account when attempting to model relatively short-term observations of vertical deformation. F. Sigmundsson provided an invaluable review and K. Atkinson wrestled tirelessly with the figures.
References BJORNSSON,A. 1985. Dynamics of crustal rifting in NE Iceland. Journal of Geophysical Research, 90, 10 151-10162. DEMETS, D., GORGON, R. G., ARGUS, D. R. & STERN, S. 1990. Current plate motions, Geophysical Journal International, 101, 425-478. EINARSSON, P., BJORNSSON, S., FOULGER, G., STEFANSSON, R. & SKAFTADOTTIR, T. 1981. Seismicity pattern in the South Iceland seismic zone. Earthquake Prediction-An International Review. AGU, Maurice Ewing, Series 4, 141-151. - - , SIGMUNDSSON,F., HOFTON, M. A., FOULGER, G. R. & JACOBY, W. 1996. An experiment in glacio-isostacy near Vatnaj6kull, Iceland 1991. J6"kull, 44, 29 39.
178
-
G. R. F O U L G E R & M. A. H O F T O N
, BEUTLER, G., BILHAM, R. et al. 1993. The Iceland 1986 GPS geodetic survey: tectonic goals and data processing results. Bulletin Geodetica, 67, 148 172. --, JAHN, C.-H., SEEBER, G., EINARSSON, P., JULIAN, B. R. & HEKI, K. 1992. Post rifting stress relaxation at the accretionary plate boundary in Iceland, measured using the global positioning system. Nature, 358, 488-490. HEKI, K., FOULGER, G. R., JULIAN, B. R. & JAHN, C.-H. 1993. Plate kinematics near divergent boundaries: geophysical implications of posttectonic crustal deformation in NE Iceland detected using the Global Positioning System. Journal of Geophysical Research, 98, 14 279-14 297. HOFTON, M. A. 1995. Anelastic deformation in Iceland studied using GPS. PhD. thesis, University of Durham. -& FOULGER, G. R. 1996a. Post-rifting anelastic deformation around the spreading plate boundary, north Iceland, 1: modeling of the 1987-1992 deformation field using a viscoelastic Earth structure. Journal of Geophysical Research, 101, 2043 2422. & 1996b. Post-rifting anelastic deformation around the spreading plate boundary, north Iceland, 2: Implications of the model derived from the 1987-1992 deformation field. Journal of Geophysical Research, 101, 25 423-25 436. JAHN, C.-H., SEEBER, G., FOULGER, G. R. & EINARSSON, P. 1994. GPS epoch measurements spanning the mid-Atlantic plate boundary in northern Iceland 1987-1990. In: Gravimetry and Space Techniques Applied to Geodynamics and Ocean Dynamics. Geophysical Monograph 82, IUGG, 17, 109-123. MOGI, K. 1958. Relations between the eruptions of various volcanoes and the deformation of the ground surfaces around them. Bulletin of the Earthquake Research Institute, 36, 99-134. -
PAkMASON, G. 1980, A continuum model for crustal generation in Iceland; kinematic aspects. Journal of Geophysics, 47, 7 l 8. ROTHACHER, M., BEUTLER, G., GURTNER, W., SCHII_DKNECWr, T. & WILD, U. 1990. Documentation for Bernese GPS software version 3.2. University of Bern. RYMER, H. & TRYGGVASON, E. 1993. Gravity and elevation changes at Askja, Iceland. Bulletin of Volcanology, 55, 362-371. SIGMUNDSSON, F. 1991. Post-glacial rebound and asthenosphere viscosity in Iceland. Geophysical Research Letters, 18, 1131-1134. -& E1NARSSON, P. 1992. Glacio-isostatic crustal movements caused by historical volume change of the Vatnaj6kull ice cap, Iceland. Geophysical Research Letters, 19, 2123 2126. SMITH, R. B. & BRAILE, L. W. 1994, The Yellowstone hotspot. Journal of Volcanology and Geothermal Research, 61, 121-187. TRYGGVASON, E. 1986. Multiple magma reservoirs in a rift zone volcano: ground deformation and magma transport during the September 1984 eruption of Krafla, Iceland. Journal of Voh'anology and Geothermal Research, 28, 1-44. VOELKSEN, C., JAHN, C.-H., SEEBER, G. & FOULGER, G. R. 1995. Determination of recent crustal movements along the plate boundary in northern Iceland between 1987 1990 and 1992. Proceedings, IUGG Meeting, Istanbul, Turkey, September 1994, 971-982. WATKINS, N. D. & WALKER, G. P. L. 1977. Magnetostratigraphy of eastern Iceland. American Journal of Science, 227, 513 584. WUBBEbaA, G. 1989, The GPS adjustment software package - GEONAP - concepts and models. In: Proceedings of the Fo"th International Geodetic Symposium on Satellite Positioning, Las Cruces, NM, 452-461.
Late Quaternary tectonism along the Pacific coast of the Californias: a contrast in style ANTONY
R. O R M E
Department of Geography, University of California, Los Angeles, CA 90095, USA (e-mail:
[email protected]) Abstract: Marine terraces and associated features are examined along 1500 km of coast at the seaward margin of that part of the North American plate that has been captured by the Pacific plate. This coast embraces parts of three morphotectonic provinces, each with its own distinctive tectonic style, namely the southern Coast Ranges, western Transverse Ranges, and northern Peninsular Ranges. Based primarily on the 125ka shoreline, four magnitudes of late Quaternary deformation are recognized: (1) high uplift rates exceeding 1.0 m per 1000 years; (2) moderate uplift rates between 1.0 and 0.1 m per 1000 years; (3) low uplift rates less than 0.1 m per 1000 years; and (4) variable net subsidence. The highest rates of uplift and subsidence are associated with the intense fold-thrust tectonics of the western Transverse Ranges nearest the Big Bend in the San Andreas fault, and diminish in magnitude to east and west. Moderate uplift rates occur in both the southern Coast Ranges and the northern Peninsular Ranges adjacent to the western Transverse Ranges, commonly associated with block uplift and subsidence between high-angle reverse faults within a broader strike-slip scenario. This suggests that compression and transpression are significant features of these areas and that the continued rotation and relative westward motion of the western Transverse Ranges, within the broader context of shear distributed across the San Andreas fault system, continue to have an impact on regions immediately to the north and south. Uplift diminishes with distance from the Transverse Ranges and is low for Baja California south of Punta Banda. The style of late Quaternary deformation is similar to that defined by earlier Quaternary marine limits, but intensity varies, most notably in the western Transverse Ranges.
As indicators of former sea levels, the shoreline angles of marine terraces provide excellent measures of the nature and extent of relative sea-level change attributable to tectonism and eustasy. When known eustatic signatures are removed, the magnitude of tectonic deformation can be assessed. Where marine terraces are datable, they also provide information on the rate of deformation. Other coastal features such as relict barrier beaches, palaeodunes and wetland deposits, though less durable, may also offer useful evidence. Emergent features are of course more readily observed but evidence from beneath present sea level, though less easily studied, must where possible be incorporated into any comprehensive model of coastal evolution. Along the California coast, emergent marine terraces have long been invoked as measures of tec-tonic deformation (e.g. Lawson 1893; Smith 1898). Much early work was, however, confounded by misinterpretation, uncertain eustatic paradigms, poor age constraints, and inappropriate structural models. For example, palaeodunes on Point Sal were misinterpreted as marine deposits (Fairbanks 1896); glacio-eustatic models were often preferred (Davis 1933); long-term absolute dating techniques were unavailable until the 1950s and for some years thereafter
were restricted by the 40 000 year limit on radiocarbon dating and the large errors linked with other isotopic methods; and structural interpretations awaited the emergence of modern plate tectonic models. Nevertheless, several marineterrace sequences were carefully mapped and, through a combination of stratigraphic, geomorphological, palaeontological and chemical techniques, assigned a relative place on the Quaternary time scale (e.g. Woodring et al. 1946). The tectonic implications of deformed fluvial and marine terraces in the Ventura area were also reaffirmed (Putnam 1942). Marine-terrace studies received a boost some 30 years ago with the application of more versatile numerical and correlative dating techniques, notably U-series methods (Veeh & Valentine 1967; Ku & Kern 1974), aminostratigraphy (Wehmiller et al. 1977), and soil and tephra chronologies. These methods in turn led to more refined palaeoclimatic and palaeoecological analyses of invertebrate faunal assemblages, as warm-water and cold-water faunas identified by Valentine (1961) and Addicott (1966) were correlated with distinct late Quaternary sea levels (Kennedy et al. 1982). The remarkable but hitherto ignored terrace sequences in Baja California also began to receive attention (e.g. Orme
ORME, A. R. 1998. Late Quaternary tectonism along the Pacific coast of the Californias: a contrast in style. In: STEWART,I. S. & VrrA-FINZI, C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 179-197.
180
A . R . ORME
1972, 1974). Simultaneously, the emergence and refinement of plate tectonic theory, especially its remarkable expression in California (e.g. Atwater 1970; Yeats 1977, 1978), refocused attention on marine terraces and associated features as tectonic indicators. However, because older terrace assemblages are more eroded and lack fossils, and because of limitations inherent in various dating techniques, most recent attention has focused on the late Quaternary record. The Ventura River terraces have also been revisited (Rockwell et al. 1984) and a better understanding of tectonism has been sought for the southern Coast Ranges (e.g. Alterman et al. 1994) and the Continental Borderland offshore (e.g. Vedder & Howell 1979; Legg 1985). This paper discusses the nature and rate of late Quaternary tectonic deformation along 1500 km of coast from south-central California to Bahia Sebastifin Vizcaino in central Baja California. This coast encompasses parts of three distinct morphotectonic provinces, the Coast Ranges, the Transverse Ranges, and the Peninsular Ranges, together with their offshore expression in the Continental Borderland (Fig. 1). The paper first outlines the tectonic setting and reviews longer-term deformation suggested by the 'Quaternary marine limit'. It then focuses on the evidence for deformation provided by the 125 ka, 105 ka, 80 ka, and earlier Holocene shorelines, related respectively to marine oxygen-isotope stages 5e, 5c, 5a and 1, together with what is known about features beneath present sea level. The paper concludes by interpreting this evidence in terms of present tectonic theory, particularly with respect to late Quaternary motion in and near the western Transverse Ranges. The following questions are specifically addressed: (1) What is the spatial distribution of coastal deformation within and between each morphotectonic province? (2) Do these rates and patterns of deformation vary through time? (3) Ignoring local tectonic noise (see Trecker et al. this volume), what do the above spatial and temporal distributions imply with respect to regional tectonism and its prognosis?
Tectonic setting To evaluate the evidence for late Quaternary coastal tectonism in the Californias, it is necessary to review briefly the evolving Cenozoic framework of the region. In essence, Cenozoic tectonism has involved a wide variety of rocks which have been variably deformed and rearranged by the interaction of three principal tectonic plates: the continental North American
plate to the east and, farther west, the oceanic Farallon and Pacific plates forming opposing limbs of the East Pacific Rise spreading centre. Before 30Ma, variably oblique convergence saw the gradual subduction of the NE-moving Farallon plate beneath the W-moving North American plate, a process associated with complex arc-forearc-accretionary wedge systems and displaced terranes. Around 30Ma, the North American plate reached the East Pacific Rise and began overriding the NW-moving Pacific plate, intensifying lateral shear along the plate boundary and shifting convergence from oblique transpression to a transform margin. The Rivera and Mendocino triple junctions formed at points of contact between these three plates. Before subduction of this spreading centre, however, fragments of the subducted Farallon plate had already been captured by, and assumed the motion of the Pacific plate (Lonsdale 1991). This created complexities in the subsequent evolution of the region that are still being resolved (Crouch & Suppe 1993; Nicholson et al. 1994; Bohannon & Parsons 1995). Over the past 30 x 106 years, as subduction of the Farallon plate has continued, the Rivera and Mendocino triple junctions have parted along a lengthening transform boundary which has stepped eastward, generating the San Andreas and related fault systems in the overriding North American plate. Around 5 Ma, this transform boundary jumped eastward into the Gulf of California, which opened in response to crustal extension within the Basin and Range Province. Baja California thus transferred to the Pacific plate, tilted, and began separating from mainland Mexico at an average rate of 6cm/year (e.g. Stock & Hodges 1989). Thus, in the south, Neogene tectonism evolved from subduction, active until 12.5 Ma, through crustal extension with the opening of the protogulf of California around 10 Ma, to the transtensional regime associated with the present Gulf of California. Meanwhile, farther north, as the San Andreas transform system migrated eastward across the California borderland, oblique strike-slip faulting intensified and a succession of initial transtensional and later transpressional forces generated numerous deep sedimentary basins and intervening ridges, triggered in part by the NW motion of those fragments of the Farallon plate captured by the Pacific plate. Though later segmented and rearranged, these basins and ridges have continued to form prominent features of coastal California throughout later Quaternary time. The marine terraces discussed in this paper thus lie at the seaward margin of a captured portion of the North American plate whose
CALIFORNIA COASTAL TECTONISM
181
Fig. 1. Morphotectonic provinces of the Californias with selected faults and localities in text.
eastern boundary between the Mendocino and Rivera triple junctions, now 2500km apart, is represented by the East Pacific Rise and associated transforms beneath the Gulf of California and by the dextral San Andreas fault system farther north. The western boundary of this cap-
tured plate is represented by an inactive eastdipping low-angle overthrust fault, representing a fossil subduction zone (Bohannon & Parsons 1995), reflected in the Patton Escarpment and above the Cedros Trench. At the coast, the alignments of the Coast Ranges and Peninsular
182
A . R . ORME
Ranges reflect a prolonged relationship between the North American plate and the plates farther west. Along the western margin of northern Baja California, for example, there is a close juxtaposition of Cretaceous to Quaternary shorelines (Orme 1980; Johnson et al. 1996). However, mesoscale structures which cross the coast obliquely, such as the Santa Maria Basin in the southern Coast Ranges, indicate a more complex relationship. In contrast, the western Transverse Ranges seem to have been rotated 90 ~ clockwise over the past 20 x 106 years as the subducting Farallon plate fragmented and, as its Monterey microplate became attached to the Pacific plate, basal shear affected the overriding North American plate (Luyendyk & Hornafius 1987; Nicholson et al. 1994). Further, palaeomagnetic studies of late Pliocene strata in the Ventura Basin indicate 20 ~ of clockwise rotation in the western Transverse Ranges over the past 3 x 106 years (Liddicoat 1992). Intensified compression is thought to have been initiated by distinct changes in the relative motions of the Pacific and North American plates around 3.9-3.4 Ma and 2.5 Ma (Harbert 1991). Clearly, these ranges owe much of their present form to relatively recent crustal shortening within a massive fold-thrust system, bounded to the south by a complex system of east-west trending reverse, oblique-slip and leftlateral strike-slip faults (Yeats & Rockwell 1991; Dolan et al. 1997). Within the coastal zone, marine terraces are well preserved in late Cretaceous and Cenozoic marine clastic rocks, notably the Rosario Formation (Upper Cretaceous) in Baja California and the Delmar (Eocene), Monterey (Miocene), and Sisquoc (Mio-Pliocene) formations in California. They are variably preserved in the older metasedimentary and metavolcanic rocks that were involved in Cenozoic subduction and/ or subsequently accreted to the North American plate, notably the Franciscan m~lange and various ophiolitic terranes (Jurassic-Miocene), and poorly preserved where Cretaceous granitoid plutons and their metamorphic host rocks reach the coast, notably in the Salinian block of the Big Sur coast.
The Quaternary marine limit The Quaternary marine limit is the highest elevation to which marine terraces and/or coastal deposits have been observed. It is a useful indicator of relative deformation along the coast against which late Quaternary tectonism may be evaluated. However, its recognition is not without
limitations, notably because older marine terraces have been much modified by subsequent erosion, more so in the humid north than in the arid south, and older coastal sedimentary sequences may have been deeply buried. Further, because terrace forms often lack adequate age constraints, this limit may range in age from late Pliocene to midPleistocene and indeed its various expressions may be asynchronous. In some circumstances, the age of the marine limit may be approximated by extrapolating uplift rates defined for lower shorelines to the elevation of the highest shoreline in a sequence. However, where deformed Plio-Pleistocene marine deposits occur, as in the Ventura Basin, age constraints provided by tephrochronology, magnetostratigraphy, fissiontrack dating, isotope analysis, and faunal correlation provide valuable estimates of the marine limit. A dozen or more marine terraces flank the southern Coast Ranges (Fig. 2). Though much dissected, the marine limit reaches over 200m above present sea level near Point Estero, 247 m in the San Luis Range east of Point Buchon (Hanson et al. 1994), and 265 m in the Casmalia Hills east of Point Sal (Woodring & Bramlette 1950; Clark 1990). Within the Santa Maria Basin, however, marine terrace continuity is broken at Morro Bay and the Santa Maria and Santa Ynez rivers where the presence of earlier Pleistocene fluvial and estuarine deposits (Paso Robles Formation) well below sea level implies prolonged basinal subsidence between bounding faults. In the western Transverse Ranges, a flight of 15 benches of presumed marine origin rises to 330m above sea level along the south flank of the Santa Ynez Mountains between Point Conception and Gaviota (Rockwell et al. 1992). Farther east, flanking the Ventura Basin, a spectacular suite of Plio-Pleistocene marine deposits, ranging from deep-water turbidites (Fernando Formation) to nearshore sands and muds (San Pedro Formation), has been raised to at least 600 m along the Ventura Avenue anticline and to between 320 and 700m on Oak Ridge, where fossiliferous estuarine deposits and coquinas interfinger eastward with non-marine fluvial deposits (Saugus Formation). These two localities are separated by a massively faulted synclinal trough which carries the Plio-Pleistocene marine boundary down 5000 m below the lower Santa Clara River and nearby Oxnard Plain. The upper part of this marine sequence has been placed by tephrochronology, magnetostratigraphy, amino acid racemization, and microfaunal biostratigraphy between 2.01Ma and 0.2-0.4Ma (Sarna-Wojcicki et al. 1980; Yeats & Rockwell 1991). This, together with geometric
CALIFORNIA COASTAL TECTONISM
183
Fig. 2. Quaternary marine limit in the Californias (references in text).
reconstructions of the Ventura Avenue anticline, suggests an uplift rate of around 10-15 m per 1000 years for the past 2 x 106 years. The Ventura Basin contains one of the world's thickest known sequences of Pleistocene sediment: nearly 5 km deep beneath the Santa Clara
valley (Yeats 1977, 1978). The marine limit is less well defined along the south flanks of the Santa Monica Mountains but geomorphological evidence again suggests rapid uplift across the path of former distributaries of the proto-Santa Clara River.
184
A . R . ORME
At the northern end of the Peninsular Ranges, Plio-Pleistocene deep-water marine deposits of the Fernando Formation are variably present beneath the Los Angeles basin to a maximum depth of 1.5-2km (Yeats & Rockwell 1991). They are overlain by the mid-Pleistocene San Pedro Formation, conformably in the centre of the basin, unconformably around its margins, and all are mantled unconformably by late Pleistocene and Holocene aeolian, estuarine and fluvial deposits. Along the southwest margin of the Los Angeles Basin, 13 marine terraces rise to elevations of411 m on the Palos Verdes peninsula (Woodring et al. 1946; Muhs et al. 1992a). Up to 16 terraces occur in Orange and San Diego counties, often supporting impressive shoreparallel beach ridges (Kern & Rockwell 1992). The terraces broaden near San Diego but the marine limit reaches only 172m suggesting that marine planation was more effective in a setting of reduced tectonic uplift. Based on uplift rates for the 125 ka terrace, the 172 m marine limit may be 1.29 x 106 years old (Kern & Rockwell 1992). Farther south, in Baja California, the marine limit on Punta Banda reaches 345 m above the Agua Blanca fault (Orrne 1972, 1980; Rockwell et al. 1994) and maintains high elevations across the southern strand of this fault in the Santa Tomas valley. The limit descends southward to only 100120m between Colnett and San Quintin, then rises to 357m above Valle del Rosario before descending below 185m at Punta San Carlos (Orme 1980). Terraces reappear onshore south of Mesa San Carlos and reach 196m behind Los Morros. Farther south, between Punta Prieta and Bahia Sebastifin Vizcaino, a terrace sequence rises to a maximum elevation of about 150 m on the eastern flanks of the Vizcaino syncline (Woods 1978), disappears beneath aeolian dunes along the axis of this syncline, and emerges again atop a dozen or so terraces to between 100 and 150 m on Punta Eugenia and the offshore islands of Cedros and San Benito (Ortlieb 1987, 1991). Offshore, in the Continental Borderland, San Clemente Island and San Nicolas Island yield impressive marine terrace sequences. On San Clemente, a tilted fault block rising to 600 m, a suite of 12 terraces reaches at least 300 m (Muhs 1983), and as many as 20 terraces may lie below a 580 m marine limit (Lawson 1893). Extrapolating from a mean uplift rate of 0.21 m per 1000 years for the 125 ka shoreline gives a tentative age of 2.8Ma for this limit, but in view of regional evidence for higher uplift rates during the earlier Quaternary, the limit could be significantly younger. Some 14 terraces reach the crest of San Nicolas, implying that the island lay beneath sea level in earlier Quaternary times (Muhs 1985).
Other offshore islands yield less complete or more ambiguous evidence, some of which suggests that certain islands, notably Santa Cruz and Santa Catalina, may have been submerging rather than rising for part of the Quaternary. In summary, earlier Quaternary seas covered the present coastal area to a variable distance inland, ranging up to 40 km in the Ventura Basin and to 20 km near San Diego and Punta Baja. The marine limit was subsequently deformed, massively so across the Ventura Basin with lesser uplift focusing on the San Luis Range, Casmalia Range, Palos Verdes, Punta Banda, and Valle del Rosario, as well as on San Clemente and San Nicolas islands offshore. The discussion now evaluates late Quaternary tectonism in the context provided by this marine limit and with the benefit of better age constraints.
Deformation of the 125 ka and later shorelines The 125ka shoreline related to the high stillstand of the Last Interglacial (marine oxygen isotope stage 5e) is the best preserved shoreline along the California coast and affords an excellent reference datum for subsequent deformation (Fig. 3). The 105 ka and 80ka shorelines of stages 5c and 5a, which formed during eustatic highstands just below present sea level, have emerged where significant tectonic uplift has occurred but are normally less developed than the 125 ka feature. The 105 ka shoreline is the least developed of the three and is missing over long stretches of coast, either because the stillstand was of short duration or because of subsequent erosion. Local names for these terraces have been omitted to avoid confusion. The elevation of the shoreline angle where shore platforms intersect old seacliffs is the principal datum for establishing former sea level. This angle is sometimes visible in gullies breaching former shorelines but, more commonly, must be sought by trenching, boring or seismic refraction, or simply extrapolated from surviving shoreline geometries. Along the modern coast, the shoreline angle approximates mean high water, about l m above mean sea level. Evidence from relatively stable coasts distant from plate boundaries indicates that eustatic sea levels during the 125 ka, 105 ka and 85ka stages were approximately + 6 m , - 5 m and - 5 m relative to present mean sea level, Rockwell et al. (1989) and Muhs et al. (1992b) for local discussion). Thus any departure of these shorelines from their eustatic elevations implies tectonic deformation.
CALIFORNIA COASTAL TECTONISM
185
Fig. 3. Late Quaternary deformation of the Californias based on the 125 ka shoreline (references in text).
The ages of these shorelines have been confirmed by U-series methods and calibrated amino acid ratios derived from molluscs within associated beach and nearshore deposits, and extended by correlative stratigraphic, geomorphological and soil properties. The most satis-
factory U-series ages (23~ have been obtained from rare corals associated with the 125 ka and 80 ka shorelines, notably the solitary coral Balanophyllia elegans and, farther south, the hermatypic coral Pocillopora guadelupensis, and from the hydrocoralline Stylaster californicus
186
A . R . ORME
(formerly Allopora californica) (Muhs et al. 1992b). Although U-series ages from fossil molluscs are unreliable, minimum ages for these organisms have been estimated where 23~ and 231pa/Z35u values are concordant (Muhs et al., 1992b). Amino acid ratios have been derived from such molluscs as the gastropods Tegula and Haliotis, and the pelecypods Epilucina, Macoma, Mytilus, Protothaca, Saxidomus and Tivela.
Southern Coast Ranges The 125 ka shoreline is broadly preserved for over 60 km along the western flank of the Santa Lucia Mountains, sloping generally southward from around 90 m north of Ragged Point to around 6 m at Cayucos (Fig. 4) (Weber 1983; Hanson et al. 1994). At San Simeon, this shoreline is dislocated by the San Simeon fault, an onshore splay of the right-lateral strike-slip Hosgri fault zone, and rises from 21 m to 38m. This suggests that uplift rates over the past 125ka have varied from 0.75m per 1000 years north of Ragged Point to 0.13-0.27 m per 1000 years across the San Simeon fault, to zero at Cayucos. Similar uplift is indicated by a lower composite terrace, related to
either the 105 ka or 80 ka shorelines, which rises from 5 m to 24 m across the fault zone. From Morro Bay to San Luis Bay, around the SW margin of the San Luis Range, the 125ka and 80ka shorelines descend from 34m to 13m and from 14m to 5m, respectively (Fig. 5) (Hanson et al. 1994). This indicates an uplift rate of between 0.23 and 0.06m per 1000 years. Around Point Sal, at the western end of the Casmalia Range, the 125 ka and 80 ka shorelines lie at 21-26 m and 7-10 m, respectively, suggesting similar uplift rates of between 0.12 and 0.19m per 1000 years, but probably closer to 0.14 to 0.17m per 1000 years (Clark 1990). These shorelines reveal little internal deformation, suggesting that both the San Luis Range and the Casmalia Range have suffered reasonably uniform block uplift and tilting. Late Quaternary marine terraces are either absent or obscured by aeolian and alluvial deposits in the basins flanking these blocks. Whereas subsidence is likely, it is possible that the late Quaternary shorelines were not given bedrock expression in the larger Santa Maria and Santa Ynez valleys and that their estuarine equivalents did not survive subsequent fluvial erosion. In Morro Bay, however, subsidence
Fig. 4. The Coast Ranges near Ragged Point showing a composite marine terrace rising northward and pinching out as the dextral strike-slip San Simeon fault zone passes offshore west of the Santa Lucia Mountains. The terrace was shaped mainly by 125 ka seas whose prominent cliff continues northward as a bevelled slope above the modern seacliff. (Photo: Spence Collection, UCLA, November 1947.)
CALIFORNIA COASTAL TECTONISM
187
Fig. ft. Marine terraces on the SW flank of the San Luis Range. The low terrace with its prominent sea stacks relates mostly to the 125 ka sea whose shoreline angle reaches 23-33 m but the 7-11 m shoreline angle of the 80 ka terrace occurs just above the modern seacliff (Hanson et al. 1994). Both shorelines step down to the south, probably across a fault similar to other high-angle reverse faults bounding the range. Inland, fragments of older terraces rise to 247 m, and Green Peak on the skyline reaches 431 m. (Photo: Spence Collection, UCLA, November 1947.) between bounding thrust faults is probable, in part because the 80ka terrace ends abruptly against the Los Osos fault, a high-angle reverse fault, and in part because Holocene sedimentation rates beneath Morro Bay indicate sudden increases in estuarine accommodation space probably related to seismic activity within the same fault zone (Orme & Gallagher 1994; Gallagher 1996). Furthermore, the presence of up to 600m of early Pleistocene fluvial and estuarine deposits, the Paso Robles Formation, lying unconformably on late Pliocene marine sediment at depth beneath both Morro Bay and the Santa Maria valley, suggests prolonged subsidence of these basins, presumably aided by displacement along bounding faults.
Transverse Ranges
Late Quaternary marine terraces are well preserved along the south flanks of the Santa Ynez
and Santa Monica mountains in the western Transverse Ranges, although their continuity is broken massively where the subsiding Ventura Basin crosses the coast. In the western Santa Ynez Range, between Point Conception and Gaviota, the variably preserved 125ka, 105ka and 80 ka shorelines have been well defined and correlated using U-series ages, amino acid ratios and molluscan palaeoecology (Kennedy et al. 1992; Rockwell et al. 1992). All three shorelines have been differentially deformed, most notably across the Government Point syncline, and the lower two are dislocated by the South Branch of the Santa Ynez fault. The 80 ka shoreline varies in elevation from 16 m at Point Conception to 10m across the synclinal hinge, rises to 31m across the fault and falls to 20m at Arroyo Hondo, a distance of 32 kin. This indicates uplift rates over the past 80 ka that range from 0.05 m per 1000 years in the syncline to 0.30m per 1000 years across the fault. Average rates of late Quaternary uplift for this region are between
188
A . R . ORME
0.15 and 0.23m per 1000 years, significant but lower than those observed farther east. The most dramatic indicators of later Quaternary tectonism along the California coast, perhaps anywhere, are the marine and river terraces of the Ventura area. Between the Red Mountain and Ventura faults, high-angle north-dipping reverse faults to the north and south of the Ventura Avenue anticline, late Quaternary marine terraces are notched unconformably into the strongly deformed mid-Pleistocene San Pedro Formation and older marine sediments. Dating of the 80 ka, 4 0 - 6 0 k a and Holocene shorelines by U-series methods, amino acid ratios and, for the Holocene features, 14C ages, yields late Quaternary uplift rates of between 1 and 10 m per 1000 years (Lajoie et al. 1991). The 80 ka shoreline has been raised 175m and tilted 9 ~ south near Ventura, for an uplift rate of 1.7-2.2m per 1000 years. More spectacularly, the 40-60 ka shoreline rises from
below sea level at Carpinteria to 100m atop Holocene seacliffs above Punta Gorda 10km farther east and to 360 m inland from Pitas Point, for an uplift rate ranging from zero to 9 m per 1000 years (Fig. 6) (Lajoie et al. 1982). Between Punta Gorda and Ventura, four modest Holocene strandlines, whose fossil molluscs give 14C ages between 1.8 and 5.5 ka, rise to 37 m above sea level for a maximum uplift rate of 6.7 m per 1000 years (Fig. 6). Although some uncertainty surrounds the dating of these Holocene terraces (they form a composite coastal bench that has been much disturbed by rail and road construction and was occupied previously by pre-Columbian Chumash peoples and their seafood middens) the evidence combines to establish a remarkable range of uplift rates. Holocene terraces may be inferred elsewhere (e.g. western Santa Cruz Island and northern Baja California), and Holocene displacement of earlier terraces is well documented, but only in
Fig. 6. The Rincon-Ventura Uplift in the western Transverse Ranges, looking east along the axis of the Ventura Avenue anticline which bisects the photograph from the oil wells at the coast. The high marine terrace (bottom left) is 40-60 ka in age. Below the abandoned seacliff, a Holocene terrace complex, dated at 1.8-5.5 ka, rises as much as 37 m above sea level from Punta Gorda (bottom) to beyond Pitas Point (middle distance). This Holocene sequence has been much disturbed by subsequent road construction. (Photo: Spence Collection, UCLA, November 1933.)
CALIFORNIA COASTAL TECTONISM the Ventura region are such features raised clearly beyond the range of predictable storm wave and tsunami activity. Assuming that the above chronology is correct, it is thought that the minimum average uplift rate along the axis of the Ventura Avenue anticline has decreased from 14 to 2 m per 1000 years over the past 200 000 years (Rockwell et al. 1988). Uplift rates of 20 m per 1000 years, 9 m per 1000 years and 5 m per 1000 years are inferred for the respective intervals 200 ka to 105 or 80 ka, 105 or 80 ka to 30 ka, and 30 ka to the present. The anticline has been shortening at a relatively constant rate of about 9 m per 1000 years since its midPleistocene inception. The 125ka and 80ka shorelines reappear intermittently along 30 km of coast flanking the Santa Monica Mountains between Little Sycamore Canyon and Malibu Creek (Szabo & Rosholt 1969; Birkeland 1972). Tilted slightly westward, they give average uplift rates of about 0.3 m per 1000 years, which, because they either straddle or lie south of the Malibu Coast fault, must represent a minimum rate of uplift for the Santa Monica Mountains to the north. Various splays of this high-angle, north-dipping reverseleft-lateral strike-slip fault show signs of late Pleistocene and Holocene movement (Rzonca et al. 1991; Treiman 1994). The 125ka shoreline reappears north of the Malibu Coast fault near Topanga Canyon and in the Pacific Palisades reaches 46 m for a minimum uplift rate of 0.32 m per 1000 years (Heron & Shaller 1997). Peninsular Ranges
South of the Santa Monica Mountains, along the northern edge of the Peninsular Ranges in the Los Angeles Basin, late Quaternary marine terrace deposits containing a warm-water fauna occur along raised portions of the Newport-
189
Inglewood fault zone, a complex dextral shear that extends southeast towards San Diego. For example, marine terrace deposits containing a warm-water fauna probably related to the 125ka shoreline occur at 70m on the flanks of the Cheviot Hills, for an uplift rate of 0.52 m per 1000 years, and on the Baldwin Hills (Willett 1937; Rodda 1957; Lajoie et al. 1991). Similar terrace deposits with beach ridges and mantling dunes also occur farther west, between this fault zone and Santa Monica Bay, but not to the east, suggesting that the uplift was already an effective barrier to the sea by 125 ka. The remarkable flight of 13 marine terraces on the Palos Verdes peninsula, formed as a doubly plunging anticlinorium rose from the sea, contains the 125 ka and 80 ka shorelines towards its base, both offset by faulting (Woodring et al. 1946). Around the southwest corner of the peninsula, the 125 ka shoreline lies at 4 6 m giving an average uplift rate of 0.32m per 1000 years (Figs 7 & 8). At the southeast corner, on the upthrown side of the Cabrillo fault, uplift rates increase to 0.56 m per 1000 years for the 125 ka shoreline and 0.64m per 1000 years for the 80 ka shoreline (Muhs et al. 1992a). Uplift rates may be as high as 0.72m per 1000 years at the northwest corner close to the Palos Verdes fault. Indeed, uplift of the entire peninsula is closely related to upthrow on the south side of this 100 km long, mainly SW-dipping, reverse-strikeslip structure which slices southeast from Santa Monica Bay to near San Clemente. Late Quaternary sediments are also faulted offshore where Holocene vertical slip rates of 0.1-0.4m per 1000 years and 0.4-0.7m per 1000 years have been estimated for the Palos Verdes and Cabrillo faults respectively (Fischer et al. 1987). Late Quaternary marine terraces occur more or less continuously along the seaward margins of the Peninsular Ranges from the southeast
Fig. 7. The Palos Verdes Peninsula with its sequence of 13 Quaternary marine terraces, viewed from the north. Splays of the Palos Verdes fault, a high-angle reverse and dextral strike-slip structure bounding the NE margin of the uplift, are expressed as shadows trending SE from the coast. The terrace sequence is now largely masked by suburban growth. (Photo: Spence Collection, UCLA, November 1927.)
190
A . R . ORME
corner of the Los Angeles Basin to the flanks of the major structural downwarp at Bahia Sebastifin Vizcaino. Certainly, terrace continuity is broken by numerous stream valleys, by landslide terrain near Ensenada, and more broadly by Bahia de Todos Santos, but essential shoreline angles are readily traced, especially as masking colluvium thins southward as aridity increases. In the San Diego and Tijuana areas, the 125 ka and 80ka shorelines are in the range 22-23m and 9-11 m, respectively, giving a low, reasonably uniform uplift rate of 0.13-0.14 m per 1000 years (Ku & Kern 1974; Kern & Rockwell 1992). Slightly higher rates occur along the dextral Rose Canyon fault at Mount Soledad (Kern 1977) and subsidence may be indicated beneath San Diego Bay. On the Punta Banda peninsula, immediately south of Bahia de Todos Santos, uplift related to the dextral Agua Blanca fault system has raised and dislocated a flight of some 12-14 marine terraces (Allen et al. 1960; Orme 1972, 1974, 1980). The 125 ka shoreline varies from 27m to 43m for an uplift rate of 0.16-0.29m per 1000 years; the 80ka shoreline varies from 14m to 18m for an uplift rate of 0.11-0.16m per 1000
years (Rockwell et al. 1989). Deformation varies as a function of proximity to the Agua Blanca fault whose main strand passes north of the peninsula, to the Maximinos fault to the south, and to intervening subsidiary faults. Farther south, marine terraces over long segments of the northern Baja California coast have yet to be precisely dated but correlative warm-water faunas and amino acid racemization indicate that the 125 ka shoreline is widely present. Initially, it maintains a relatively high elevation across the Santo Tomfis-Bahia Soledad fault zone, then declines gently to beyond Cabo Colnett for an uplift rate diminishing from 0.29 to 0.10m per 1000 years (Orme 1980). It is then locally uplifted between bounding faults on Mesa San Jacinto, for an uplift rate of 0.18m per 1000 years, but descends towards present sea level across Valle Camalfl and Rio Santo Domingo and is largely obscured by Holocene cinder cones and basalt flows, and by aeolian sand in Valle San Quintln, where uplift rates must approximate zero. It then rises southward to around 30m near E1 Rosario, for a maximum uplift rate of 0.18 m per 1000 years, before falling again beyond Punta Baja (Orme 1980).
Fig. 8. The western slopes of the Palos Verdes Peninsula looking north from Lunada Bay to beyond Flatrock Point, showing the marine terrace sequence before suburban development. The broad low terrace is related to the 125ka sea whose shoreline angle lies 46m above sea level, for an uplift rate of 0.32m per 1000 years (Muhs et al. 1992a). At Flatrock Point the terrace reaches 53 m and has been related to the 80 ka sea for an uplift rate of 0.72 m per 1000 years; correlation with the 125 ka sea would yield a lesser uplift rate of 0.38 m per 1000 years. (Photo: Spence Collection, UCLA, September 1921.)
CALIFORNIA COASTAL TECTONISM Flights of marine terraces and abandoned barrier-lagoon complexes reappear south of Mesa San Carlos and continue more or less to Morro Santo Domingo in the Vizcaino syncline (Woods 1978, 1980; Orme 1980). As farther north, the 125ka feature is the lowest visible shoreline between E1 Marron and Santo Domingo, ranging in elevation from 6 to 14m and thus barely above its last interglacial limit. This shoreline has been invoked at a maximum elevation of 5 m around Laguna Ojo de Liebre inland from Bahia Sebasti~in Vizcaino, suggesting mild subsidence of less than - 0 . 0 1 m per 1000 years (Ortlieb 1991). On the Vizcaino peninsula, the 125ka shoreline reaches maximum elevations of 17 m in Bahia San Cristobal and Bahia Tortugas, and 13m and 18m on Cedros and San Benito islands offshore for an uplift rate of around 0.10m per 1000 years, (Ortlieb 1987, 1991). A lower shoreline, probably the 80 ka stillstand, occurs locally at 5-8 m for an uplift rate around 0.15m per 1000 years. Offshore, the 125 ka and 80 ka shorelines form the lowest steps in the 14-terrace sequence on San Nicolas Island (Vedder & Norris 1963; Valentine & Veeh 1969; Muhs 1985). The 125 ka shoreline reaches 33 m for an average uplift rate of 0.22 m per 1000 years. The 80ka shoreline is locally present at l l m for a similar rate. On San Clemente Island, the 125ka shoreline reaches 27-35 m for an uplift rate of 0.17-0.23 m per 1000 years (Muhs & Szabo 1982). On Isla Guadelupe, some 300 km NW of Punta Eugenia, deposits between present sea level and 6 m have been dated at around 125 ka by U-series methods on the coral Pocillopora guadelupensis, indicating little or no net deformation of the 125 ka shoreline (Lindberg et al. 1980; Muhs et al. 1994b). This implies that the island, a 4-7 Ma alkalic basalt seamount on a fossil ridge crest beyond the captured North American plate boundary, has been essentially stable during the late Quaternary.
Tectonic implications As a result of heightened interest and improved techniques over the past 30 years, sufficient is now known about the location of late Quaternary shorelines along the California coast to support the search for explanation in a tectonic context. A comparison of the Quaternary marine limit (Fig. 2) with the pattern of late Quaternary deformation (Fig. 3) reveals considerable similarity. Areas where the marine limit is highest or most deformed are also areas where late Quaternary uplift rates are highest, notably around Ventura in the western Transverse Ranges but also on other fault-bound blocks in the southern
191
Coast Ranges, the northern Peninsular Ranges, and certain islands within the Continental Borderland. Conversely, areas of prolonged Neogene subsidence, such as the Ventura Basin and the Vizcaino syncline, are also areas where late Quaternary shorelines either do not occur or have not survived later erosion, or where correlative estuarine and marine deposits descend to beneath present sea level. In short, the pattern of deformation established in the earlier Quaternary appears to have persisted through the later Quaternary. Only locally does the resolution of these data allow for definitive statements to be made about regional acceleration or deceleration during the late Quaternary, notably in the Ventura area (Yeats 1978; Rockwell et al. 1988). With respect to late Quaternary deformation, if one removes the noise caused by localized folding and faulting from the regional signature, four magnitudes of deformation emerge for the past 125 ka: (1) high uplift rates in excess of 1.0m per 1000 years and locally much higher; (2) moderate uplift rates of between 1.0 and 0.1 m per 1000 years; (3) low uplift rates of less than 0.1 m per 1000 years; and (4) areas of variable subsidence (negative uplift) rates (Fig. 9). High uplift rates occur in the western Transverse Ranges, centred on the Ventura region, and have been the focus of much speculation. Moderate uplift rates characterize much of the remaining coast but increase markedly towards the Transverse Ranges. Low rates represent minimal uplift, implying that the 125 ka shoreline has not been raised more than 18 m above present sea level (or 12 m above its Stage 5e limit), a feature of much of the Pacific coast of Baja California south of Punta Banda. Areas where little or no net deformation has occurred, such as central Baja California, should, however, be distinguished from areas of subsidence because in some localities, such as the Oxnard Plain, the latter are similar in magnitude, in a negative sense, to the higher uplift categories. The unusually high uplift rates of the western Transverse Ranges in the Ventura area have been explained in terms of crustal convergence resulting in intense compression within a foldthrust belt constrained to the north by the Big Bend in the San Andreas fault, only 60 km away, and favoured internally by d6collement at shallow depth beneath the mountains (Yeats 1981, 1983; Namson & Davis 1988a). Following Neogene rotation of the Transverse Ranges after 20Ma (Schneider et al. 1996), intensified compression began between 5 and 2.5 Ma, probably as the Pacific plate further adjusted its motion relative to the North American plate (Harbert 1991). Rates of convergence have been estimated at between 17 and 23 m per 1000 years (Yeats
192
A . R . ORME Southern Coast Ranges
TransverseRanges
10
Northern Peninsular Ranges
x
A
A
A
SLM
SLR
CR
A
A
SYM
A
A
PVH
"T
1.0 2
Uplift Rate
(m/ka) go
0.1
9
_
Y. ?,
0.01
0 N
CP
SC L
SD
LA
I 500 Shoreline Distance
i
V
TS 1
so 1
!
9
1000 k m S
Fig. 9. Late Quaternary uplift rates based on the 125 ka, 80 ka and 40-60 ka shorelines over 1000 km of coast between Ragged Point and Punta Baja. Uplift rates: e, 125 ka shoreline; -, 80 ka shoreline; • 40-60 ka shoreline. Uplifts: SLM, Santa Lucia Mountains; SLR, San Luis Range; CR, Casmalia Range; SYM, Santa Ynez Mountains; RV, Rincon-Ventura Uplift; SMM, Santa Monica Mountains, PVH, Palos Verdes Hills; PB, Punta Banda. Basins: MB, Morro Bay; SM, Santa Maria Valley; SY, Santa Ynez Valley; CP, Carpinteria Basin; SC, Santa Clara Valley-Oxnard Plain; LA, Los Angeles Basin; SD, San Diego Bay; TS Todos Santos Bay; SQ, San Quintin. 1983; Rockwell et al. 1988). This convergence has been attributed in part to the transfer of slip from right-lateral faults in the southern California borderland to faulting off the coast of the southern Coast Ranges (Weldon & Humphreys 1986). The lower uplift rates of 0.15-0.30 m per 1000 years farther west may be explained by reduced rates of crustal shortening with increasing distance from the San Andreas fault. Farther east, the situation is less clear. Certainly, uplift rates decrease eastward along the Santa Monica Mountains into the Los Angeles Basin, but at 0.30-0.52m per 1000 years they are still relatively impressive and there is good evidence for late Pleistocene and Holocene left-lateral and oblique-reverse movement along the Malibu Coast and Santa Monica faults and their continuation east as the Hollywood fault (Rzonca et al. 1991; Treiman 1994; Dolan et al. 1997). The San Andreas fault zone remains within 70-80 km of this coast, so it is likely that much of the compression is accommodated in the foldthrust belt whose primary axis extends more or less due east from Ventura, along Oak Ridge and the Santa Susana Mountains, the focus of recent
seismic activity in the 197l M 6.4 San Fernando and 1994 M6.7 Northridge earthquakes. The record of coastal uplift here is complicated by rapid coastal erosion and mass movement and by encroaching urbanization. Beyond the Transverse Ranges, areas of moderate uplift and local subsidence characterize both the southern Coast Ranges and the northern Peninsular Ranges. Although these latter provinces are dominated by dextral strike-slip tectonics, much of the accompanying vertical motion is accommodated along faults which also have a significant high-angle reverse component. In the southern Coast Ranges, the Los 9 Orcutt Frontal and Lion's Head faults, which converge offshore into the dextral Hosgri fault system, are among those high-angle reverse faults that define a succession of fault-bounded structural blocks sub-parallel to the present coast. Both the San Luis and Casmalia blocks have risen significantly during late Quaternary times, carrying the 80ka shoreline and a sequence of deformed late Pleistocene fluvial deposits well above original levels (Orme 1992; Hanson et al. 1994). Meanwhile, the adjacent Morro Bay, Santa
CALIFORNIA COASTAL TECTONISM Maria Valley and Santa Ynez Valley blocks have subsided, as shown by the depth to which early Pleistocene fluvial deposits have dropped (Hall 1982). In Morro Bay, coseismic subsidence around 2-2.5ka seemingly increased accommodation space for Holocene estuarine sedimentation and generated fault scarplets, closed depressions and deranged drainage in the nearby Los Osos Valley (Orme & Gallagher 1994; Gallagher 1996). A similar situation occurs in the northern Peninsular Ranges, where the Palos Verdes terrace sequence has been raised on the upthrow side of the Palos Verdes fault and further dislocated across the Cabrillo fault. The Palos Verdes fault has been interpreted primarily as a strike-slip fault (Nardin & Henyey 1978) but vertical uplift on the south side of its high-angle reverse plane is clearly important (Fischer et al. 1987), and recent seismic-reflection studies confirm its Holocene movement beneath Los Angeles harbour (Clarke et al. 1997). From elevations and suggested ages for the entire terrace sequence, Ward & Valensise (1994) have generated a fault model with 3.0-3.7m per 1000 years of oblique, dextral reverse slip on a Palos Verdes fault dipping at 67 ~ to 6-12 km beneath the peninsula. They further estimated that, since initiation of uplift around 3Ma, the crest of the Palos Verdes anticlinorium has been raised more than 1300 m from a depth of around 850m to over 400m above sea level, at an average rate of 0.43 m per 1000 years, a value bracketed by the uplift rates of 0.32-0.72 m per 1000 years inferred for the 125 ka and 80 ka shorelines (Muhs et al. 1992a). Modest uplift and subsidence also occur across the dextral Rose Canyon fault near the San Diego area and in Holocene deposits beneath San Diego Bay (Kennedy & Clarke 1997). Farther south, the shoreline sequence on Punta Banda owes its elevation and tilting to vertical motion within the dextral Agua Blanca fault system. Extensions to the NNW of this and adjacent systems are probably responsible for both horizontal and vertical dislocation in the Coronado Bank and San Diego Trough off San Diego, and for the massive uplift of San Clemente Island farther west. It is tempting to believe that a continuing rotation of the western Transverse Ranges within the context of a complex transform margin is at least partly responsible for late Quaternary vertical motion along high-angle reverse faults in the adjacent southern Coast Ranges and northern Peninsular Ranges. It is also likely that part of the 20 ~ clockwise rotation inferred by Liddicoat (1992) for the western Transverse Ranges over the past 3Ma has
193
occurred during the late Quaternary. To the northwest this would generate compression and crustal shortening in the southern Coast Ranges. Namson & Davis (1988b) have estimated 33 km of crustal shortening for these ranges during post-Miocene time, concomitant with continuing right slip within the coastal Hosgri fault zone and other strands of the San Andreas system farther inland. Such compression would generate significant folding and high-angle reverse and reverse-oblique faulting such as that responsible for raising the 125 ka and 80 ka shorelines across the San Luis and Casmalia ranges while dropping them downward beneath Morro Bay and the Santa Maria valley (Clark et al. 1994). To the southeast the continuing movement northwest of the Peninsular Ranges and adjacent Continental Borderland would cause similar compression and transpression, increasing in magnitude as their various structural slices approached the western Transverse Ranges. Compression and westward displacement of the western Transverse Ranges relative to the northern Peninsular Ranges and Continental Borderland are expressed in the system of oblique-reverse, left-lateral strike-slip faults, such as the Malibu Coast fault, that form the >200km boundary between the two provinces. Compression is further emphasized by recent seismic events, notably the 1987 M6.0 Whittier Narrows earthquake and the 1994 M6.7 Northridge earthquake, which occurred on blind thrust faults at depth near the southern margins of the western Transverse Ranges. The tectonic style implied by the Palos Verdes uplift may also be similar to other fault-bounded anticlinal structures within the Continental Borderland, whose several islands and shallow banks may express reverse slip associated with left-lateral bends of regional right-lateral strike-slip faults (Ward & Valensise 1994). Farther south in Baja California, south of the Agua Blanca fault system, the tectonic regime appears to change as marine terraces are deformed gently by broad aseismic warping and offset by NE-SW trending faults (Orme 1980). Movement ceased on many of these faults before late Quaternary time but the faults bounding Mesa San Jacinto appear to dislocate the 125ka shoreline. Similar NE-SW faults occur in California, notably along the San Diego River, and in the Continental Borderland where NNW-trending structures are disrupted by cross-faulting (Krause 1965; Legg 1985). For example, the sea floor drops about 450 m to the south of the extension of the Santa Toms fault, with a left-lateral offset of 15 km, so that basins south of the fault abut against ridges to the
194
A . R . ORME
north and vice versa. It is reasonable to infer that differential movement between crustal slivers is creating tension along the peninsula's western margin, tension expressed in shore-normal faults that cross the borderland to reach the coast. These dislocations apart, the limited deformation of the 125ka shoreline south of Punta Banda supports the belief that the Agua Blanca fault system forms a boundary between two distinct crustal blocks (e.g. Ortleib 1991). Conclusion
Marine terraces and associated features along the Pacific coast of the Californias provide valuable evidence for evaluating the style and magnitude of late Quaternary deformation and for testing models of crustal behaviour in the region. The western Transverse Ranges have long been recognized as a region of massive Quaternary uplift related to convergence and crustal shortening for which various models have been invoked (e.g. Yeats 1983). However, adjacent parts of the Coast Ranges and the Peninsular Ranges, long regarded as dominated by right-lateral strikeslip tectonics west of the San Andreas fault zone, have also emerged from recent studies with evidence for significant compression and crustal shortening (Clark et al. 1994; Ward & Valensise 1994). Such motion is reflected by the deformation and dislocation of late Quaternary shorelines in uplifted blocks and subsiding basins bounded by high-angle reverse and obliquereverse faults. Although late Quaternary vertical movements in these latter areas do not compare in magnitude with those observed in the Ventura area within the western Transverse Ranges, they are significantly greater than those noted in the more stable parts of central Baja California (Ortlieb 1991). Continuing rotation of the western Transverse Ranges is probably responsible in part for the style and pattern of coastal deformation within these adjacent areas by triggering both compression and lateral motion within and between discrete fault-bounded blocks. Accordingly, tectonic models for this coastal region must accommodate (1) dextral strike-slip motion within the constraining geometry of the San Andreas system, (2) sinistral strike-slip motion and strong compression within the western Transverse Ranges, and (3) high-angle reverse faulting and associated folding related to compression within the adjacent Coast Ranges and Peninsular Ranges, regions also subject to dextral slicing. As a result of these motions, and excluding localized subsidence, the 1500km coast of the Californias studied here has significantly increased its freeboard during late Quaternary time.
The author gratefully acknowledges the constructive advice of three anonymous referees. Historical aerial photographs, useful because so many marine terraces have been masked by later development, are provided from the Spence Collection (1920-1970) by permission of the Department of Geography, UCLA.
References
ADDICOTT,W. O. 1966. Late Pleistocene Marine Paleoecology and Zoogeography in Central California. US Geological Survey Professional Paper, 523-C. ALLEN, C. R., SILVER, L. T. & STEHLh F. G. 1960. Agua Blanca fault - a major transverse structure in northern Baja California, Mexico. Geological Socie O' of America Bulletin, 71, 457-482. ALTERMAN, I. B., MCMULLEN, R. B., CLUFF, L. S. & SLEMMONS, D. B. (eds) 1994. Seismotectonics of the Central California Coast Ranges'. Geological Society of America Special Paper, 292. ATWATER, T. 1970. Implications of plate tectonics for the Cenozoic tectonic evolution of western North America. Geological Society of America Bulletin, 81, 3513-3535. BIRKELAND, P. W. 1972. Late Quaternary eustatic sea-level changes along the Malibu coast, Los Angeles County, California. Journal of Geology, 80, 432-448. BOHANNON, R. G. & PARSONS, T. 1995. Tectonic implications of post-30Ma Pacific and North American relative plate motions. Geological Society of America Bulletin, 107, 937-959. CLARK, D. G. 1990. Late Quaternary tectonic deformation in the Casmalia Range, coastal southcentral California. In: LETTIS, W. R., HANSON, K, L., KELSON,K. I. & WESHNO, J. R. (eds) Neotectonics of South-Central Coastal California. Friends of the Pleistocene, Pacific Cell, Fieldtrip Guidebook, 349-383. --,
SLEMMONS,D. B., CASKEY, S. & DEPOLO, D. M.
1994. Seismotectonic framework of coastal central California. In: ALTERMAN,I. B., MCMULLEN, R. B., CLUFF, L. S. & SLEMMONS,D. B. (eds) Seismotectonics of the Central California Coast Ranges. Geological Society of America Special Paper, 292, 9-30. CLARKE, S. H., KENNEDY, M. P. & PONTI, D. J. 1997.
Seismic-reflection study of the Palos Verdes fault zone in the vicinity of Los Angeles harbor, California. Geological Society of America Cordilleran Section, Abstracts with Program, 29, 85. CROUCH, J. K. & SUPPE, J. 1993. Late Cenozoic tectonic evolution of the Los Angeles basin and inner California borderland: a model for core complex-like crustal extension. Geological SocieO, of America Bulletin, 105, 1415-1434. DAVIS, W. M. 1933. Glacial epochs of the Santa Monica Mountains, California. Geological Society of America Bulletin, 44, 1041-1133. DOLAN, J. F., SIEH, K., ROCKWELL,T. K., GUPTILL, P.
& MILLER, G. 1997. Active tectonics, paleoseismology, and seismic hazards of the Hollywood fault, northern Los Angeles basin, California. Geological Society of America Bulletin, 109, 1595-1616.
CALIFORNIA COASTAL TECTONISM FAIRBANKS, H. W. 1896. The geology of Point Sal. University of California Publications in the Geological Sciences, 2, 1 92. FISCHER, P. J., MESA2 INC., PATTERSON, R. H., DARROW, A. C., RUDAT, J. H. & SIMLA, G. 1987. The Palos Verdes fault zone: onshore to offshore. In: FISCHER, P. J. & MESA2 INC. (eds) Geology of the Palos Verdes Peninsula and San Pedro Bay. Pacific Section, Society of Economic Paleontologists and Mineralogists and American Association of Petroleum Geologists, Los Angeles, 91-133. GALLAGHER, J. 1996. Late Holocene evolution of the Chorro delta, Morro Bay, California. PhD dissertation, University of California, Los Angeles. HALL, C. A. 1982. Pre-Monterey Subcrop and Structure Contour Maps, western San Luis Obispo and Santa Barbara Counties, South-Central California. US Geological Survey Map MF-1384. HANSON, K. L., WESLING,J. R., LETTIS,W. R., KELSON, K. I. & MEZGER, L. 1994. Correlation, ages, and uplift rates of Quaternary marine terraces: southcentral coastal California. In: ALTERMAN, I. B., MCMULLEN, R. B., CLUFF, L. S. & SLEMMONS, D. B. (eds) Seismotectonics of the central California Coast Ranges. Geological Society of America Special Paper, 292, 45 71. HARBERT, W. 1991. Late Neogene relative motions of the Pacific and North American plates. Tectonics, 10,1 15. HERON, C. W. & SHALLER, P. J. 1997. Reinterpretation of wave-cut marine terraces west of Castellammare Mesa, Pacific Palisades, California. Geological Society of America Cordilleran Section, Abstracts with Program, 151. JOHNSON, M. E., LEDESMA-VAZQUEZ,J., CLARK, H. C. & ZWEIBEL, J. A. 1996. Coastal evolution of late Cretaceous and Pleistocene rocky shores: Pacific rim of northern Baja California, Mexico. Geological Society of America Bulletin, I08, 708-721. KENNEDY, G. L., LAJOIE, K. R. & WEHMILLER, J. F. 1982. Amino-stratigraphy and faunal correlations of late Quaternary marine terraces, Pacific coast, USA. Nature, 299, 545 547. , WEHMILLER, J. F. & ROCKWELL, T. K. 1992. Paleoecology and paleozoogeography of late Pleistocene marine-terrace faunas of southwestern Santa Barbara County, California. In: FLETCHER, C. H. & WEHMILLER, J. F. (eds) Quaternary Coasts of the United States." Marine and Lacustrine Systems. Society for Sedimentary Geology Special Publication, 48, 343-361. KENNEDY, M. P. & CLARKE, S. H. 1997. An analysis of the Rose Canyon fault zone in San Diego Bay, California. Geological Society of America Cordilleran Section, Abstracts with Program, 29, 112. KERN, J. P. 1977. Origin and history of upper Pleistocene marine terraces, San Diego, California. Geological Society of America Bulletin, 88, 1553-1566. & ROCKWELL, T. K. 1992. Chronology and deformation of marine shorelines, San Diego County, California. In: FLETCHER, C. H. & WEHMILLER, J. F. (eds) Quaternary Coasts of the
195
United States. Marine and Lacustrine Systems. Society for Sedimentary Geology Special Publication, 48, 377 382. KRAUSE, D. C. 1965. Tectonics, bathymetry, and geomagnetism of the southern continental borderland west of Baja California, Mexico. Geological Society o/'America Bulletin, 76, 617-650. Ku, T.-L & KERN, J. P. 1974. Uranium-series age of the upper Pleistocene Nestor terrace, San Diego, California. Geological Society of America Bulletin, 85, 1713- 1716. LAJOIE, K. R., PONTI, D. J., POWELL, C. L. II, MATHIESON, S. A. & SARNA-WOJCICKI, A. M. 1991. Emergent marine strandlines and associated sediments, coastal California: a record of Quaternary sea-level fluctuations, vertical tectonic movements, climatic changes, and coastal processes. In: MORRISON, R. B. (ed.) Quaternary Nonglacial Geology. Conterminous US. Geological Society of America, Boulder, Colorado, The Geology of North America, K-2, 190 203. - - , SARNA-WOJCICKI, A. M. & YERKES, R. F. 1982. Quaternary chronology and rates of crustal deformation in the Ventura area, California. In: COOPER, J. D. (ed.) Neotectonics in Southern California. Geological Society of America Cordilleran Section Annual Meeting Guidebook, 43-51. LAWSON, A. C. 1893. The post-Pliocene diastrophism of the coast of southern California. UniversiO' of California Department qfl Geological Sciences Bulletin, 1, 115-160. LEGG, M. R. 1985. Geologic structure and tectonics of the inner continental borderland, offshore northern Baja California. PhD dissertation, University of California, Santa Barbara. LIDDICOAT, J. C. 1992. Paleomagnetics of the Pico Formation, Santa Paula Creek, Ventura basin, California. Geophysical Journal International, 110, 267-275. LINDBERG, D. R., ROTH, B., KELLOGG, M. G. & HUBBS, C. L. 1980. Invertebrate megafossils of Pleistocene (Sangamon interglacial) age from Isla de Guadelupe, Baja California, Mexico. In: POWER, D. M. (ed.) The California Islands. Santa Barbara Museum of Natural History, Santa Barbara, 41-62. LONSDALE, P. 1991. Structural patterns of the Pacific floor offshore of peninsular California. In: DAUPHIN, J. P. & SIMONEIT, B. T. (eds) Gulf and Peninsula Province o[ the CaliJornias. American Association of Petroleum Geologists, Memoir, 47, 87-125. LUYENDYK, B. P. & HORNAFIUS, J. S. 1987. Neogene crustal rotation, fault slip, and basin development in southern California. In: INGERSOLL, R. V. & ERNST, W. G. (eds) Cenozoic Basin Development in Coastal California. Prentice-Hall, Englewood Cliffs, NJ, 259 293. MUHS, D. R. 1983. Quaternary sea-level erects on northern San Clemente Island, California. Quaternary Research, 20, 322-341. 1985. Amino acid age estimates of marine terraces and sea levels on San Nicolas Island, California. Geology, 13, 58-61.
196
A.R.
& SZABO, B. J. 1982. Uranium-series age of the Eel Point terrace, San Clemente Island, California. Geology, 10, 23-26. , MILLER, G. H., WHELAN, J. F. & KENNEDY, G. L. 1992a. Aminostratigraphy and oxygen isotope stratigraphy of marine-terrace deposits, Palos Verdes Hills and San Pedro areas, Los Angeles, County, California. In: FLETCHER, C. H. & WEHMILLER, J. F. (eds) Quaternary Coasts of the United States. Marine and Lacustrine Systems. Society for Sedimentary Geology Special Publication, 48, 363-376. , ROCKWELL,T. K. & KENNEDY, G. L. 1992b. Late Quaternary uplift rates for late Pleistocene marine terraces: southern Oregon to Baja California Sur. Quaternary International, 15--16, 121-t 33. NAMSON, J. & DAVIS, T. 1988a. Structural transect of the western Transverse Ranges, California: implications for lithospheric kinematics and seismic risk evaluation. Geology, 16, 675-679. & 1988b. Seismically active fold and thrust belt in the San Joaquin Valley, central California. Geological Society of America Bulletin, 100, 257-273. NARDIN, T. R. & HENYEY, T. L. 1978. PliocenePleistocene diastrophism of Santa Monica and San Pedro shelves, California continental borderland. Bulletin, American Association of Petroleum Geologists, 62, 247-272. NICHOLSON, C., SORLIEN, C. C., ATWATER, T., CROWELL, J. C. & LUYENDYK, B. P. 1994. Microplate capture, rotation of the western Transverse Ranges, and initiation of the San Andreas transform as a low-angle fault system. Geology, 22, 491495. ORME, A. R. 1972. Quaternary deformation of western Baja California, Mexico, as indicated by marine terraces and associated deposits. Tectonics, 24th International Geological Congress, Montreal, 3, 627--634. - - 1 9 7 4 . Quaternary deformation of marine terraces between Ensenada and E1 Rosario, Baja California, Mexico. In: GASTIL, R. G. (ed.) Geology of Peninsular California. American Association of Petroleum Geologists, Society of Economic Paleontologists and Mineralogists, and Society of Economic Geologists, 67-79. - - 1 9 8 0 . Marine terraces and Quaternary tectonism, northwest Baja California, Mexico. Physical Geography, 1, 138-161. - - 1 9 9 2 . Late Quaternary deposits near Point Sal, south-central California: a time frame for coastaldune emplacement. In: FLETCHER, C. H. & WEHMILLER, J. F. (eds) Quaternary Coasts of the United States: Marine and Lacustrine Systems. Society for Sedimentary Geology Special Publication, 48, 309-315. & GALLAGHER, J. 1994. Relative sea-level change in the Chorro delta, Morro Bay, south-central California. Geological Society of America, Abstracts with Programs, 26, 256. ORTUEB, L. 1987. Neotectonique et Variations du Niveau Marin au Quaternaire dans la R~gion du Golfe de Californie, Mexique. Institut Frangaise de
-
-
-
-
ORME Recherche Scientifique pour la D6veloppement en Coop6ration, Collection Etudes et Th6ses, Paris. 1991. Quaternary vertical movements along the coasts of Baja California and Sonora. In: DAUPmN, J. P. & SIMONEIT,B. R. T. (eds) The Gulf and Peninsular Province o[" the Californias. American Association of Petroleum Geologists, Memoir, 47, 447-480. PUTNAM, W. C. 1942. Geomorphology of the Ventura region, California. Geological Society of America Bulletin, 53, 691 754 ROCKWELL, T. K., KELLER, E. A., CLARK, M. N. & JOHNSON, D. L. 1984. Chronology and rates of faulting of Ventura River terraces, California. Geological Society of America Bulletin, 95, 1466-1474. & DEMBROFF,G. R. 1988. Quaternary rate of folding of the Ventura Avenue anticline, western Transverse Ranges, southern California. Geological Society of America Bulletin, 100, 850-858. --, MUHS, D. R., KENNEDY, G. L., HATCH, M. E., WtLSON, S. H. & KUNGER, R. E. 1989. Uraniumseries ages, faunal correlations and tectonic deformation of marine terraces within the Agua Blanca fault zone at Punta Banda, northern Baja California, Mexico. In: ABBOTT, P. L. (ed.) Geological Studies in Baja California. Pacific Section, Society of Economic Paleontologists and Mineralogists, Los Angeles, Book, 63, 1-16. - - , NOLAN, J., JOHNSON, D. L. & PATTERSON, R. H. 1992. Ages and deformation of marine terraces between Point Conception and Gaviota, western Transverse Ranges, California. In: FLETCHER, C. H. & WEHMILLER, J. F. (eds) Quaternary Coasts of the United States." Marine and Lacustrine Systems. Society for Sedimentary Geology Special Publication, 48, 333-341. RODDA, P. U. 1957. Paleontology and stratigraphy of some recent marine Pleistocene deposits in northwest Los Angeles basin, California. Bulletin, American Association of Petroleum Geologists, 41, 2475-2492. RZONCA, G. F., SPELLMAN,H. A., FALL, E. W. & SHLEMON, R. J. 1991. Holocene displacement of the Malibu Coast fault zone, Winter Mesa, Malibu, California: Engineering geologic implications. Association o[ Engineering Geologists Bulletin, 28, 147-158. SARNA-WOJCICKI, A. M., BOWMAN, H. R., MEYER, C. E., et al. 1980. Chemical Analysis, Correlations, and Ages of Late Cenozoic' Tephra Units of EastCentral and Southern California. US Geological Survey Open-File Report, 80-231. SCHNEIDER, C., HUMMON, C., YEATS, R. S. & HUFT1LE, G. J. 1996. Structural evolution of the northern Los Angeles basin, California, based on growth strata. Tectonics, 15, 341-355. SMITH, W. S. T. 1898. A geological sketch of San Clemente Island. US Geological Survey 18th Annual Report, II, 459-496. STOCK, J. M. & HODGES, K. V. 1989. Pre-Pliocene extension around the Gulf of California, and the transfer of Baja California to the Pacific plate. Tectonics, 8, 99-115.
-
-
_
-
-
CALIFORNIA COASTAL TECTONISM
197
WEHMILLER, J. F., LAJOIE, K. R., KVENVOLDEN, K. A. et al. 1977. Correlation and Chronology of
SZABO, B. J. & ROSHOLT, J. N. 1969. Uranium-series dating of some Pleistocene molluscan shells from southern California- an open system model. Journal of Geophysical Research, 74, 3253-3260. TRECKER, M. A., GURROLA, L. D. & KELLER, E. A. 1998. Oxygen-isotope correlation of marine terraces and uplift of the Mesa hills, Santa Barbara, California, USA. This volume. TREIMAN, J. A. 1994. The Malibu Coast Fault. California Division of Mines and Geology Fault Evaluation Report, FER 229. VALENTINE, J. W. 1961. Paleoecologic molluscan geography of the California Pleistocene. Univer-
WILLETT, G. 1937. An upper Pleistocene fauna from the Baldwin Hills, Los Angeles County, California. San Diego Society of Natural History Transactions, 8, 379-406. WOODRING, W. P. 8z BRAMLETTE, M. N. 1950.
sity of California Publications in the Geological Sciences, 34, 309-442.
Geology and Paleontology of the Santa Maria District, California. US Geological Survey Profes-
& VEEH, H. H. 1969. Radiometric ages of Pleistocene terraces from San Nicolas Island, California. Geological Society of America Bulletin, 80, 1415- 1418. VEDDER, J. G. & HOWELL, D. G. 1979. Topographic evolution of the southern California borderland during late Cenozoic time. In: POWER, D. M. (ed.)
sional Paper, 222. & KEW, W. S. W. 1946. Geology and Paleontology of the Palos Verdes Hills, California. US Geological Survey Professional Paper, 207. WOODS, A. J. 1978. Marine terraces between Playa El
The California Islands. Proceedings of a Multidisciplinary Symposium. Santa Barbara Museum
of California, Los Angeles. - - 1 9 8 0 . Geomorphology, deformation and chronology of marine terraces along the Pacific coast of central Baja California, Mexico. Quaternary Research, 13, 346-364. YEATS, R. S. 1977. High rates of vertical crustal movement near Ventura, California. Science, 196, 295-298. - - 1 9 7 8 . Neogene acceleration of subsidence rates in southern California. Geology, 6, 456-460. - - 1 9 8 1 . Quaternary flake tectonics of the California Transverse Ranges. Geology, 9, 16-20. 1983. Large-scale Quaternary detachments in Ventura basin, southern California. Journal of Geophysical Research, 88, 569-583. -& ROCKWELL, T. K. 1991. Quaternary geology of the Ventura and Los Angeles basins, California. In: MORRISON, R. B. (ed.) Quaternary Nonglacial Geology: Conterminous US. Geological Society of America, Boulder, Colorado. The Geology of North America, K-2, 185-189.
of Natural History, 2-31. & NORRIS, R. M. 1963. Geology of San Nicolas Island, California. US Geological Survey Professional Paper, 369. VEEH, H. H. & VALENTINE, J. W. 1967. Radiometric ages of Pleistocene fossils from Cayucos, California. Geological Society of America Bulletin, 78, 547-550. WARD, S. N. & VALENSISE,G. 1994. The Palos Verdes terraces, California: bathtub rings from a buried reverse fault. Journal of Geophysical Research, 99, 4485- 4494. WEBER, G. E. 1983. Geologic Investigation of the
--
Marine Terraces of the San Simeon Region and Pleistocene activity of the San Simeon Fault Zone, San Luis Obispo County, California. US Geological Survey Technical Report, Contract No. 14-08-001-18230. WELDON, R. & HUMPHREYS, E. D. 1986. A kinematic model of southern California. Tectonics, 5, 33-48.
Pacific Coast Marine Terrace Deposits of continental United States by Fossil Amino Acid Stereochemistry- Technique Evaluation, Relative Ages, Kinetic Model Ages, and Geological Implications. US Geological Survey Open-File Report, 77-680.
Marron and Morro Santo Domingo, central Baja California, Mexico. PhD dissertation, University
Convergent-margin deformation of Pleistocene strata on the Olympic coast of Washington, USA GLENN
D. T H A C K R A Y
Quaternary Research Center, University of Washington, Seattle, WA 98195, USA Present address." Department of Geology, Idaho State University, Pocatello, ID 83209, USA (e-mail: thacglen @ isu.edu)
Abstract: Quaternary tectonism on the central Olympic coast is dominated by the Kalaloch syncline, a broad (>32 km wavelength), low-amplitude (c. 45 m) fold deforming the Pleistocene glacial-nonglacial stratigraphic sequence. The stratigraphy includes two sequences of glacial and nonglacial sediments, one overlying and one underlying a prominent wave-cut surface. Stratigraphic and chronological data indicate that the buried wave-cut surface formed during isotope stage 7 or 5 sea-level highstands, probably during substage 5e or 5c. The elevation pattern of the wave-cut surface expresses the syncline most clearly. The outcrop patterns of overlying strata also reflect the syncline, as does outcrop-scale deformation in the underlying strata. Geological uplift rates for the wave-cut surface on the syncline's north limb contrast in magnitude with, but are similar in pattern to, geodetically measured 1931-1974 uplift rates. Geological uplift rates range from -0.03 to 0.7 mm/year. The difference at each location between geological and geodetic uplift rates ranges from 0.4 to 1.0 ram/year, and is inferred to reflect interseismic strain accumulation. For each assumed age for the wave-cut surface, the inferred interseismic strain accumulation is nearly constant across the syncline's north limb. Thus, the syncline accounts for most of the spatial variability in the 1931-1974 uplift rates. Comparison of the geological uplift rates with alternative representations of 1931-1974 uplift does not result in the same patterns, but does reveal a significant component of interseismic strain accumulation. The Cascadia Subduction Zone has been intensively scrutinized since Heaton & Kanamori (1984) used plate-age and convergence-rate comparisons to suggest that the margin's relative seismic quiescence might belie a capability of producing infrequent but large earthquakes. Several workers (e.g. Atwater 1987, 1992; Darienzo & Peterson 1990; Darienzo 1991; Nelson 1992) subsequently found geological evidence of past coseismic subsidence and tsunami in coastal marshes of Washington and Oregon, and Adams (1990) attributed Holocene turbidites in the Cascadia channel to recurrent earthquakes along the subduction zone. Geodetic evidence (e.g. Holdahl et al. 1987, 1989; Savage et al. 1991; Mitchell et al. 1994) and geophysical modelling of the plate interface (e.g. Hyndman & Wang 1993, 1995) suggest that the interface is locked and that a great-earthquake hazard exists. Studies of Quaternary coastal deformation (e.g. Kelsey et al. 1994, 1996; McCrory 1996) have contributed data and interpretations that augment understanding of, and constrain geophysical models pertaining to, tectonic processes such as those at work in the Cascadia Subduction Zone. Tectonic deformation on the central Olympic coast of Washington is recorded in a Pleistocene glacial-interglacial stratigraphic sequence. Structures deforming that sequence are a component of accretionary wedge processes, and provide for
instructive comparisons of coincident, geologically and geodetically determined uplift rates. This paper documents deformation of the Pleistocene sedimentary sequence, and uses the resulting structural data to distinguish uplift rates and patterns reflective of net, permanent, structureand landform-producing deformation from those reflective of interseismic strain accumulation.
Tectonic and geologic setting The Olympic coast lies above the Cascadia Subduction Zone (Fig. 1) and at the western edge of the Olympic Mountains. Interaction of the Juan de Fuca and North American plates has produced the Cascade volcanic arc, the PugetWillamette forearc basin, and prominent coastal ranges. On the Olympic Peninsula, the coastal mountains reach nearly 2500m elevation and expose pervasively deformed and mildly metamorphosed rocks of the Olympic accretionary complex. Long-term uplift rates in the Olympic Mountains are c. 1 ram/year (Brandon & Vance 1992). The Olympic uplift is an east-plunging domal structure, consisting of mildly metamorphosed accretionary wedge sandstone and shale inset structurally within a horseshoe-shaped belt of Eocene basalt. Within the accretionary complex, the rocks exhibit a general east-to-west trend of age (Eocene to Miocene), metamorphic grade
THACKRAY, G. D. 1998. Covergent-margin deformation of Pleistocene strata on the Olympic coast of Washington, USA. In: STEWART, I. S. & VITA-FINZl, C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 199-211.
200
G . D . THACKRAY This paper focuses on the Kalaloch syncline (Baldwin 1939), a gentle structure similar in style and dimensions to other coastal folds noted above. The Kalaloch syncline is unusual for the Cascadia margin, in that it is expressed in a glacial-interglacial stratigraphic sequence, rather than in a marine terrace sequence. The stratigraphic sequence is very well exposed in sea-cliffs and provides a detailed, time-transgressive view of the deformation. A possible correlative syncline has been identified on seismic reflection images 20-30km offshore (L. McNeill, pers. comm., 1997). That offshore syncline, which is similar in style and dimension to others on the Cascadia continental shelf, offers the potential to correlate onshore and offshore deformation.
Coastal stratigraphy Stratigraphic f r a m e w o r k
Fig. 1. Sketch map of the Cascadia Subduction Zone, showing major plates, plate boundaries, and the location of the study area.
(chlorite-epidote to pumpellyite) and pervasiveness of tectonic deformation (extreme to mild; Tabor & Cady 1978a, b). On the Olympic coast, the bedrock consists of folded and faulted, but stratigraphically intact, Miocene sandstone and shale with minor conglomerate and basalt. The Olympic Mountains have experienced extensive glaciation, and glacial sediments and landforms dominate the lower valleys and coastal plain. The areal distribution and chronology of the glacial sequence has been described by Thackray (1996). Quaternary coastal deformation on the Cascadia Subduction Zone is most prominently expressed as a series of faults and gentle folds deforming wave-cut marine terraces and terrestrial sediments (e.g. McInelly & Kelsey 1990; Kelsey et al. 1996; McCrory 1996). Some of the documented coastal structures have been correlated with numerous structures described on the Oregon and Washington continental shelf through marine geophysical methods (e.g. Goldfinger et al. 1992; McNeill et al. this volume). McCrory (1996) described subaerially exposed faults and folds on the southern and central Olympic coast, some that reflect the dominant SW-NE plate convergence direction and others (including the syncline described below) that reflect northward translation of the Siletz terrane against stable North American crust lying to the north.
The stratigraphy of the central Olympic coast (Fig. 2) consists of two principal components: strata lying below and strata lying above a prominent wave-cut surface. The sequence of strata underlying the wave-cut surface consists of glacial and nonglacial sediments and forms the exposed base of stratigraphic sections (except where Tertiary bedrock outliers predominate) in the northern and southern portions of the study area. This sequence of older strata is hereafter called the Steamboat Creek Formation, an informal time-stratigraphic unit. In the central portion of the coastal study area, the Steamboat Creek Formation lies below beach level, having subsided tectonically in the axial zone of the Kalaloch syncline. The sequence overlying the wave-cut surface consists of glacial and nonglacial sediments assigned to three informal units of formation rank. Within several kilometres of the Queets and Hoh river mouths, this sequence consists principally of outwash. The outwash is assigned to two time-stratigraphic units, the Lyman Rapids outwash and Hoh Oxbow outwash, corresponding to allostratigraphic units of the same names mapped and described by Thackray (1996). Between the areas dominated by Hoh and Queets valley outwash (Fig. 2), the sequence is dominated by silt, clay, and peat deposits (the Browns Point Formation, an informal lithostratigraphic unit). The buried wave-cut surface is a prominent feature in the stratigraphy, exposed continuously in several areas tbr as much as 2 km and discontinuously along most of the coastline in the limbs of the Kalaloch syncline. The wave-cut
DEFORMATION OF PLEISTOCENE STRATA, WASHINGTON, USA
201
Fig. 2. Composite stratigraphic profile of the central Olympic coast showing major stratigraphic units, location of stratigraphic sections in Fig. 3a-d and geographical locations noted in the text. Simplified from plate 2 of Thackray (1996). surface truncates the Steamboat Creek Formation, as well as Miocene- and Pliocene-age bedrock. Where cut across bedrock, the surface is marked by ancient pholad (rock clam) borings. In some areas, most notably in the northern portion of the study area, the wave-cut surface is marked by a cobble-boulder lag. The wave-cut surface is everywhere overlain by horizontally bedded sand and gravel deposited in a beach and/or intertidal environment. The wave-cut surface can be correlated, via stratigraphic exposures in stream cuts, with a palaeo-sea-cliff lying 0.5 1.5km inland from the modern coastline.
turn overlain by two outwash units (Lyman Rapids and Hoh Oxbow units) and loess. Fossil pollen spectra from the peat-bearing unit indicate interglacial vegetation similar to the modern coastal forest (Florer 1972). The unit appears to have accumulated during a period of very slow sedimentation, probably adjacent to small creeks that drained a higher outwash terrace lying 1.5kin landward. The unit probably represents a period of late-interglacial landscape stabilization following marine regression and preceding glacial-fluvial inundation. The significance of this unit for wave-cut surface age inferences is discussed below.
Stratigraphic sections
Beach Trail 4. The Beach Trail 4 area (Fig. 3b) typifies stratigraphy in the area isolated from Hoh and Queets valley outwash deposition. Basal Miocene bedrock is truncated by the wave-cut surface at only 2.7 m altitude (sloping gradually from l l.5m altitude at Destruction Island Viewpoint) and is marked by ancient pholad borings. Sand and gravel beach sediments (4.3 m) fill irregularities in and cover the erosional surface. The overlying sequence consists of the Browns Point Formation, here dominated by clay, silt, and peat, with minor gravel. Most of the gravel consists of lithologies similar to bedrock in adjacent hills, but one gravel interbed correlates with Lyman Rapids outwash from the Hoh valley. The sequence is capped by loess. The section indicates that sediment deposited in this area following formation of the wave-cut surface was dominated by detritus from adjacent uplands. The area was isolated from the Hoh
Four stratigraphic sections (Fig. 3a-d) exemplify the coastal stratigraphy and reflect the synclinal deformation. The locations of the four sections are shown in Fig. 2. Destruction Island Viewpoint. The Destruction Island Viewpoint section (Fig. 3a) typifies the outwash-dominated stratigraphy of the syncline's north limb. Outwash and till of the Steamboat Creek Formation form the base of the section. The wave-cut surface truncates the outwash, forming a clearly defined unconformity c. 11.5 m above sea level. The wave-cut surface is marked by a coarse gravel lag, and ancient pholad borings indent a nearby bedrock knob at the same altitude. Nine metres of horizontally bedded sand and gravel beach deposits cover the wave-cut surface and lag. On top of the beach deposits is a unit consisting of 2.5 m of organicrich sand and silty sand, peat, and palaeosols, in
202
G. D. THACKRAY
Fig. 3. Stratigraphic sections showing major stratigraphic units and chronological data in four portions of the coastal sequence (locations shown in Fig. 2). Radiocarbon dates denoted 'AA' were determined by the NSFArizona AMS Facility; those denoted 'Beta' are conventional radiocarbon dates determined by Beta Analytic, Inc. (a) Destruction Island Viewpoint; (b) Beach Trail 4; (e) Kalaloch; (d) Whale Creek. valley sediment source except during the Lyman Rapids glaciation, and the numerous peat beds indicate that clastic sedimentation was probably sporadic. North Kalaloch. A stratigraphic section 2 km north of Kalaloch (Fig. 3c) typifies stratigraphy near the synclinal axis. The Steamboat Creek Formation, bedrock, and the wave-cut surface are absent, presumably lying below beach level. Lyman Rapids outwash from the Queets valley comprises the basal 1.4m, overlain by a 5.6m thick sequence of silt, peat, clay, and sand. That sequence, in turn, is overlain by 3.7 m of pebble gravel (local lithologies), fluvial sand, and silty clay. Excepting the outwash, the sediment is dominated by detritus from nearby uplands, as at Beach Trail 4. Whale Creek. A stratigraphic section 0.4km north of Whale Creek (Fig. 3d) includes typical Steamboat Creek Formation sediments for the area, as well as the wave-cut surface and over-
lying sequence. The Steamboat Creek Formation strata include silt and clay (both laminated and massive), peat, and fluvial sand and fine gravel, deposited in terrestrial and/or shallow marine environments. The strata dip approximately ll ~ north. The wave-cut surface (5m altitude) rises from beach level 0.1 km to the north, and marks an angular unconformity in this area. Sand and gravel beach deposits cover the wave-cut surface to a thickness of 7.8 m. Less than 3 km north, the beach deposits are cut by a channel of 4 m depth, filled with gravel and capped by 6 m of peat. The overlying sequence at Whale Creek consists of Lyman Rapids outwash (6.3m), interbedded fine sand, silt, peat, and minor gravel (6.3m), and loess (1.5 m). As a whole, this section and surrounding exposures reflect coastal erosion during an interglacial sealevel highstand, followed by late-interglacial fluvial activity and peat deposition and, finally, glacial-fluvial inundation. The Hob Oxbow outwash, present at Destruction Island Viewpoint, is absent in this portion of the study area,
DEFORMATION OF PLEISTOCENE STRATA, WASHINGTON, USA and post-Lyman Rapids sedimentation was limited to local deposition of slackwater clastic sediments, peat, and loess.
Summary o f chronological data and age inferences Palaeomagnetic polarity determinations, radiocarbon dates, thermoluminescence dates, and stratigraphic relationships allow age estimation for the stratigraphic units described above. Data and inferences are here summarized to support the use of specific ages in wave-cut surface uplift rates calculations. Quaternary stratigraphic units and features are correlated with marine oxygenisotope stages or events (e.g. isotope stage 2 isotope substage 5e), as designated by Imbrie et al. (1984) and Martinson et al. (1987).
more than c. 50 14Cka BP, and probably represents a glacial event correlative with oxygenisotope stage 4, 5b, or 5d. Radiocarbon dates of 28.4+0.514CkaBP and 36.8+0.814CkaBP below Hoh Oxbow outwash in the coastal sequence (Fig. 3a), coupled with several radiocarbon dates below, within, and above correlative Hoh Oxbow sediments in the Hoh River valley, indicate that the outwash was deposited between c. 29 and 2714CkaBP. The Browns Point Formation has yielded numerous radiocarbon dates ranging from 16.7+0.214CkaBp (Heusser 1972) to >4914CkaBp. Those dates support the inference from stratigraphic observations (e.g. as shown generally in Fig. 2) that deposition of the Browns Point Formation spanned a time period substantially longer than that represented by the two outwash units.
Age of the wave-cut surface.
Age of stratigraphic units above and below the wave-cut surface. Chronological data for the Steamboat Creek Formation are limited to palaeomagnetic polarity determinations and thermoluminescence dates. At Destruction Island Viewpoint (Fig. 3a) and an additional location near the Hoh River, Steamboat Creek Formation sediments yielded magnetically reversed samples, suggesting that they were deposited during the Matuyama reversed magnetic polarity chron. Near Whale Creek (Fig. 2), the unit yielded two normally polarized samples, suggesting deposition during either the Brunhes or Gauss normal magnetic polarity chrons. The Steamboat Creek Formation includes several types of glacial and nonglacial sediment (e.g. outwash, till, glacial-marine drift, estuarine mud), and stratigraphic relationships are unclear; it is conceivable that the unit includes sediments of widely ranging ages as suggested by the magnetic polarity data. Thermoluminescence dates of 261 :t:50kaBP and < 1.08 Ma BP were also obtained from the Steamboat Creek Formation at Whale Creek (Fig. 3d; see discussion below). The age of the stratigraphic sequence lying above the wave-cut surface is constrained, in part, by radiocarbon dates. Samples from sediments below and within Lyman Rapids outwash in the coastal sequence exceed the upper age limit of the radiocarbon method, yielding five dates ranging from >33.7 14CkaBP (Fig. 3a; Florer 1972) to >4914CkaBP. Five minimum limiting dates range from 28.4+0.5~4C to 46.7 + 3.414CkaBP. Coupled with additional dates on correlative sediments in the Hoh and Queets river valleys (Thackray 1996), these dates indicate that the unit was deposited at
203
The wave-cut surface has not been dated directly, but stratigraphic relationships with chronologically constrained, overlying and underlying units allow age inferences. Radiocarbon dates associated with the overlying Lyman Rapids outwash indicate that the wave-cut surface is older than c. 5014CkaBP. Thus, the surface must have been cut during sea-level highstands during the last or previous interglaciations (i.e. isotope stage 5 or older). Thermoluminescence dates, obtained on an experimental basis for this sequence, suggest an isotope stage 7 age. Dates on samples collected from exposures at and near the Whale Creek section (Fig. 3d) are 238 4- 31 ka BP (above wavecut surface, 3 km north of Whale Creek section), 273 + 31 ka BP (above wave-cut surface; Fig. 3d), and 261+50kaBP (below wave-cut surface; Fig. 3d). These dates suggest a broad correlation of the wave-cut surface with isotope stage 7. Stratigraphic relationships, revealed in several kilometres of excellent sea-cliff exposures in the vicinity of the dated strata, do not support such a correlation. If the wave-cut surface formed during isotope stage 7, then isotope stage 6 outwash capped by an isotope stage 5 weathering horizon would be expected in the overlying sequence. Neither has been observed. Interglacial gravel, silt, and peat cap the wave-cut surface and give way to a continuous outwash body broken only by silty interbeds. It is conceivable that a last-interglacial weathering horizon did form on isotope stage 6 outwash, but was eroded before deposition of the Lyman Rapids outwash. However, the expected unconformity has not been observed, and all available chronological and stratigraphic evidence supports most strongly the inference that
204
G . D . THACKRAY
the overlying outwash correlates entirely with the Lyman Rapids outwash. In the Destruction Island Viewpoint area (Fig. 3a) the sequence overlying the wave-cut surface is very consistent, with little lateral variability, and also clearly lacks an unconformity. Additionally, the thermoluminescence dates themselves suggest incomplete zeroing of the thermoluminescence signal at the time of deposition. The similarity of thermoluminescence dates from both above and below the wave-cut surface suggests that beach sediments above the wavecut surface merely represent sediments that were recycled, without having their thermoluminescence signal reset, as pre-existing sediments were eroded by wave action. That is, the dates on sediments overlying the wave-cut surface may reflect the age of an immediate sediment source rather than the timing of their own deposition. Forman & Ennis (1992) and Kaufman et al. (1996) found that the variability of bleaching in waterlain sediments, caused by the light-filtering effects of water and turbidity, can lead to substantial thermoluminescence age overestimates. At best, the thermoluminescence dates from the Whale Creek area suggest a maximum wave-cut surface age of c. 260 ka BP. In summary, the stratigraphic and chronological evidence can most logically be interpreted to indicate an isotope stage 5 wave-cut surface origin. Radiocarbon dates from overlying sediments indicate an age greater than c.5014C ka BP, and stratigraphic evidence, albeit largely negative evidence, argues against an origin during the isotope stage 7 or earlier sea-level highstands. Stratigraphic evidence also can be used to argue against wave-cut surface formation during isotope substage 5a. At Destruction Island Viewpoint (Fig. 3a), a 2.5 m thick unit of palaeosols, peat, and fluvial overbank sediments overlies the beach sediments that cap the wave-cut surface. That unit yielded interglacial pollen spectra (Florer 1972). Similarly, within 3 km north of the Whale Creek section (Fig. 2), the beach sediments covering the wave-cut surface are cut by a shallow, gravel-filled channel (see above), which is itself capped by silt, palaeosols, and peat. The latter sediments thicken northward to two beds, of 3-4 m thickness, separated by fine gravel and containing peat, palaeosols, and in situ stumps. Those beds are overlain by Lyman Rapids outwash. In both cases, substantial terrestrial sedimentation occurred subsequent to sea-level regression and before outwash deposition. Heusser (1972) derived a sedimentation rate of c. 2500 years/m from a 17m thick, radiocarbon-dated sequence of similar sediments in the Browns Point Formation near Kalaloch. Application of this
sedimentation rate to the interglacial sediments in question implies that they represent c. 600020 000 years. Although the amount of time required to deposit these post-highstand interglacial sediments cannot be directly determined, the figures suggest that the time of deposition was probably longer than the time available following the isotope substage 5a highstand and preceding isotope stage 4 glaciation. Therefore, I infer that the wave-cut surface probably does not correlate with the isotope substage 5a highstand.
Summary. I interpret the available stratigraphic and chronological evidence as indicating that the wave-cut surface most probably formed during isotope substage 5e or 5c. However, because the arguments against possible isotope stage 7 or isotope substage 5a ages are equivocal, as they are based upon inferences from stratigraphic relationships and sedimentation rates and not upon direct chronological data, uplift rates are calculated below assuming ages corresponding to all four sea-level highstands (isotope stage 7 and substages 5e, 5c, and 5a) suggested by the available chronological and stratigraphic data.
Quaternary tectonic deformation The Kalaloch syncline The overall pattern of the coastal stratigraphy (Figs 2 and 3a-d) reflects the Kalaloch syncline (Baldwin 1939), a gentle, asymmetric fold reflected in the wave-cut surface, in overlying strata, and in the coastal terrace. Dipping strata in the Steamboat Creek Formation also reflect deformation associated with the syncline. The Kalaloch syncline is evident along the 31 km coastline traverse between the Hoh and Raft rivers (Fig. 2). The traverse largely reflects the actual cross-sectional geometry of the structure because the coastline is nearly perpendicular to the ENE synclinal axis trend inferred from attitudes of Steamboat Creek Formation strata. The synclinal structure is most perceptible in the pattern of the buried wave-cut surface and its overlying beach sediments. The surface descends from 30m altitude to beach level (1-2m altitude) in the 14km between the Hoh River and Browns Point, lies below beach level in the 18 km between Browns Point and the Whale Creek area, and rises from beach level to 31m altitude in the 4km between the Whale Creek area and the Raft River. The southern limb of the syncline dips more steeply than the northern limb, indicating a SSE-dipping axial plane. Steamboat Creek Formation strata also exhibit deformation, evident at outcrop scale,
DEFORMATION OF PLEISTOCENE STRATA, WASHINGTON, USA reflective of the Kalaloch syncline. In the Destruction Island Viewpoint area (Fig. 3a) the outwash contains a 1 m thick bed of overbank sand and silt, the upper surface of which dips 8~ south. Sea-cliffs within 1.6 km south of the Hoh River (Fig. 2) expose Steamboat Creek strata that dip as much as 30 ~ south. However, those exceptional dips may be at least partially related to palaeo-landsliding. A structural component may also exist, but distinction of the two components is not possible. Steamboat Creek Formation strata exposed on the south limb of the Kalaloch syncline are more strongly deformed, forming clear angular unconformities with the wave-cut surface and overlying strata. North of Whale Creek (Fig. 2), the Steamboat Creek Formation strata dip 11 ~ north. Dips in the Whale Creek area range from 15~ north to 26 ~ south and appear to reflect a gentle, WNW-trending anticline in the Steamboat Creek Formation. Finally, an exposure 0.8km south of Whale Creek displays N N W dips of Steamboat Creek Formation strata that increase progressively from 18~ to nearly 90 ~ over a distance of 80m. These relationships suggest an apparent high-angle fault (McCrory 1996) or thrust-fault propagation fold (Thackray 1996), which may also control the relatively steep dip of the wave-cut surface in that area. The syncline is also reflected in the exposure pattern of the strata overlying the wave-cut surface. Lyman Rapids outwash is exposed on both limbs of the syncline near the axis, but the outwash lies below beach level in the immediate vicinity of the axis (Fig. 2). Gradients of the tops of outwash units south of the Hoh River are probably steepened by deformation on the north limb of the syncline because the depositional gradient and tectonic tilt are additive. In contrast, the outwash surface is probably backtilted south of the Queets River, where the southerly depositional gradient is opposed to the northward tilt of the south limb of the syncline. Finally, the coastal terrace, in places underlain by outwash and in others by Browns Point Formation sediments, lies at its lowest elevation near the synclinal axis and rises on either limb.
Deformation
rates
The stratigraphic and chronological data and inferences discussed above provide for the calculation of uplift rates for the Kalaloch syncline. Those rates allow interpretation of geodetically measured uplift rates, and of the nature of coastal deformation along this portion of the Cascadia margin.
205
Fig. 4. Uplift rates (mm/year) inferred from tide gauges and levelling. Contours are after Holdahl et al. (1987) with 1.2ram/year added to all rates by Savage et al. (1991) to account for refined eustatic sea-level rise estimates. Note uplift minimum (arrow) coincident with the synclinal axis at Kalaloch. Tide gauges used by Holdahl et al. (1987) include Neah Bay (NB) and Toke Point (TP). Contemporary deformation associated with the Kalaloch syncline is apparent in one representation of the regional pattern of geodetically measured uplift. Contouring of geodetically measured uplift rates (1931-1974) by Holdahl et al. (1987; revised slightly by Savage et al. 1991) shows areas of uplift minima and maxima in coastal Washington state (Fig. 4). Maximum uplift rates of more than 3.2mm/year were measured on the northwestern Olympic Peninsula, whereas uplift minima (<0.08ram/year) encompass Willapa Bay, Gray's Harbor, and a portion of the central Olympic coast. The latter uplift minimum is centred on Kalaloch, and is coincident with the Kalaloch syncline. If these detailed contours of geodetic data are accurate, they show that the Kalaloch syncline is a contemporary structure. The coincidence of geologically and geodetically recorded data permits useful comparisons
206
G. D. THACKRAY
between deformation on widely differing time scales. The Holdahl et al. (1987) contours are chosen for these comparisons, in part, because they are detailed enough to reflect the syncline. Contrasting representations of the 1931-1974 geodetic data (Mitchell et al. 1994; Hyndman & Wang 1995) are discussed subsequently. The buried wave-cut surface serves as the most reliable datum for uplift rate calculations. The elevation of the wave-cut surface serves as a reliable datum because the modern sea-cliff exposures are nearly parallel (<0.5 km variation) to the palaeo-sea-cliff corresponding to the wave-cut surface. For uplift rate calculations, the buried wave-cut surface is projected to the palaeo-shoreline angle, accounting for the minor, modern- to palaeo-sea-cliff distance variations. Furthermore, I infer that the broad pattern of the wave-cut surface is not affected significantly by variations in the seaward slope of the uplifted wave-cut surface, as the overall synclinal pattern varies little across variable substrate. Minor, local elevational irregularities may reflect such variations in the surface's seaward slope. The quality of geological and geodetic data is higher for the north limb of the Kalaloch syncline than for the south limb. The quality of geological data is enhanced between the Hoh River and Browns Point (Fig. 2) by extensive exposures and an easily recognized and measured wave-cut surface. Numerous wave-cut surface altitudes have been measured. The quality of coastal geodetic data is also likely to be highest in this area. The relevelling survey route (Holdahl et al. 1987) follows US Highway 101 through the western Olympic Peninsula; only between the Queets and Hoh Rivers is the highway near the coastline. Thus, the north limb of the Kalaloch syncline between Browns Point and the Hoh River is the only portion of the structure in which the geodetic survey route coincides closely with well-exposed and easily measured portions of the wave-cut surface; therefore, only the north limb is treated here. Calculation of uplift rates requires estimation of the altitude of the buried palaeo-shoreline angle. The altitude of the shoreline angle at the time of platform formation approximates the altitude of the sea-level highstand to within +2 m (Wright 1970; Trenhaile 1980). The palaeo-shoreline angle altitude is estimated through projection of the buried wave-cut surface landward to the palaeo-sea-cliff, using the measured altitude in the modern sea-cliff (0-30m), the distance landward to the associated palaeo-seacliff (0-1.5 km), and assumed platform gradients. Gradients of wave-cut platforms measured on
the central California coast (Bradley & Griggs 1976) are used for the projection: 0.02 for the 600 m wide, inner portion and 0.005 for the remainder of the platform width (outer portion). Uplift rate calculation also requires estimates of the difference between modern eustatic sea level and palaeo-sea level for the respective highstands. Interpretation of marine terraces on stable coastlines suggests that early last-interglacial sea level (isotope stage 5e, c. 125kaBP) was 3-11m higher than modern sea level (Mesolella et al. 1969; Veeh & Chappell 1970; Bloom et al. 1974; Dodge et al. 1983). The average of these estimates, c. 6 m above modern sea level, is supported by palaeo-sea-level data from uplifted terraces in New Guinea (Chappell & Shackleton 1986) and by an apparently in situ terrace on Guadalupe Island, 330 km west of Baja California, Mexico (Muhs et al. 1992). Rockwell et al. (1989) determined from uplifted marine platforms in Baja California that isotope substage 5c sea level (105 kaaP) was c. 2m below modern sea level. For isotope substage 5a (80kaaP), Muhs et al. (1992) summarized evidence from coastal California, USA, suggesting that sea level was c. 5 m below modern. The California model is used here for uplift rate calculations. Uplifted terraces in New Guinea suggest substage 5c and 5a sea levels of c. 9 and 19 m below modern, respectively (Chappell & Shackleton, 1986); use of the New Guinea model, rather than the California model, would increase corresponding uplift rates c. 0.07 and 0.19 mm/year, respectively. For isotope stage 7, Chappell & Shackleton (1986) reported a sea-level highstand of c. 7m below modern sea level at c. 212 ka BP, as well as highstands of 7 m below modern at c. 200 000ka BP and c. 10 m below modern at c. 238 ka ap. The former is used here; use of either of the other two highstands would result in uplift rate differences of less than 0.01 mm/year. Potential errors in this analysis arise principally from three sources: errors in measuring the wave-cut surface altitude (+2m); errors in the estimated interglacial eustatic sea level (• m); and divergence of the actual buried platform gradient from the assumed gradients ( i 6 m ) . The latter estimate allows for c. 40% error because the wave-cut surface was cut largely on marginally consolidated Steamboat Creek Formation outwash, rather than on mudstone and sandstone as in the case of the California platforms measured by Bradley & Griggs (1976). Together, these potential errors (c. 11 m) equal uplift rate errors of 0.05 mm/year (isotope stage 7 age) to 0.14 mm/year (isotope substage 5a age). Figure 5a shows uplift rates calculated for the north limb of the syncline and compares them
D E F O R M A T I O N OF PLEISTOCENE STRATA, WASHINGTON, USA
~
1.0
o
f
E E ~0.5
5c
s~ 0.0 Distance south of Hob River (krn)
E E _~1.o ,= .o 5c
8
5a
?o.s
o~
0
' g . . . . lb Distance south of Hoh River (krn)
Fig. 5. Uplift rate comparisons for the north limb of the Kalaloch syncline. (a) Profiles of geological and geodetic rates. Geological uplift rates are calculated for the four sea-level highstands that represent possible age interpretations for the wave-cut surface; as discussed in the text, isotope substage 5e and 5c highstands are most strongly supported by stratigraphic data. (b) Difference between geologically and geodetically derived uplift rates, interpreted as the amount of interseismic strain accumulation. with the profile of contoured, geodetically determined uplift from Fig. 4. The northward increase in calculated deformation rates reflects the tilting of the north limb of the syncline. The shape of the uplift rate profile varies with the sea-level highstand used, the isotope stage 7 profile being the gentlest and the isotope substage 5a profile the steepest. For all possible wave-cut surface ages, uplift rates are generally low, ranging from -0.03 to 0.7 mm/year. For the isotope substage 5e and 5c sea-level highstands, which I consider from stratigraphic evidence to most probably have formed the wave-cut surface (see above), all calculated uplift rates are less than 0.5mm/year. In the axial zone of the syncline (for which rates are not calculated and are not represented in Fig. 5 owing to lack of wave-cut surface exposure), uplift rates are probably negative. The wave-cut surface lies below modern sea level and thus lies completely or largely below isotope substage 5e, 5c, and 5a sea levels. The most notable aspect of the deformation rate comparison is the difference in magni-
207
tude between the geological and geodetic rates (Fig. 5b). Because differencing removes the permanent component of the contemporary deformation, I interpret that the residual rates (Fig. 5b) represent the interseismic strain accumulation included in the geodetic uplift rates. On the basis of this interpretation, I infer that this portion of the coastline is accumulating 0.4-1.0ram/year of interseismic strain that should be recovered during the next seismic event. This inference assumes that permanent uplift rates have been constant over the corresponding time period (212000-80000 years) since wave-cut surface formation. Similarly, the significance of the interseismic strain accumulation inferred from the 1931-1974 geodetic uplift depends on the variability of uplift through the interseismic cycle. Geodetic data pre- and post-dating subduction zone earthquakes elsewhere (e.g. southwest Japan; Savage & Thatcher 1992) demonstrate that overall interseismic deformation varies in magnitude through the earthquake cycle, particularly shortly before and after earthquakes. The 1931-1974 Olympic data represent a short time span in comparison with the inferred length of the Cascadia Subduction Zone seismic cycle (hundreds of years; e.g. Atwater 1987), and it is not known whether the data represent middle or latter portions of the interseismic period. Therefore, it cannot be known how representative these estimates are of overall interseismic strain accumulation through the entire seismic cycle. The spatial variability of the geological and geodetic uplift rates is very similar. This is particularly apparent with respect to geological uplift rates corresponding to isotope substage 5e, 5c, and 5a sea-level highstands. I infer that their similarity with the geodetic uplift pattern suggests that most of the spatial variability in coastline geodetic rates reflects the geological structure. Additionally, I suggest that the similarity in spatial variability may suggest that the synclinal deformation lacks a strain-accumulation component. That is, I infer that although this section of the coastline is accumulating interseismic strain at nearly uniform spatial rates, the structure itself appears not to be accumulating significant strain.
Discussion Sources o f error in geodetic uplift rates and influences on rate comparisons Errors in the geodetically derived uplift rates (Figs 4 and 5) contoured by Holdahl et al. (1987) and revised slightly by Savage et al. (1991),
208
G. D. THACKRAY
naturally are reflected in the uplift rate comparisons presented above and affect interpretations of those comparisons. Potential errors may be either uniform or nonuniform with respect to the uplift pattern. Nonuniform types would include levelling errors, errors introduced by unstable benchmarks, and variability in glacio-isostatic rebound. Levelling errors include systematic measurement error, solar refraction error, and magnetic error. The latter two errors are corrected empirically. The former should be minimal along the Olympic levelling circuit, which traverses low-relief coastal areas (Holdahl et al. 1989). Unstable benchmarks (e.g. those affected by landslides) might introduce uplift or subsidence patterns unrelated to tectonic processes. If landslides affected benchmarks along the US Highway 101 levelling route, they might be partly or wholly responsible for the apparent uplift minimum coincident with the axial zone at Kalaloch. However, observed landslide movements, largely in the Browns Point Formation, have narrow (metre-scale) inland penetration. Deep-seated movements that would affect benchmarks along the highway, lying tens to hundreds of metres inland, have not been observed and would not be anticipated in this stratigraphic setting. Glacio-isostatic rebound may be reflected in the geodetic uplift rates. The Juan de Fuca lobe of the Cordilleran ice sheet advanced to and retreated quickly from a position 20-30 km north of the study area at about 14 ka BP (Heusser 1973; Waitt & Thorson 1983). Because of its short residence time, glacio-isostatic effects of the ice sheet were probably short lived. However, the study area may have lain in the forebulge zone of thicker and longer-lived portions of the ice sheet occupying the Strait of Georgia and Vancouver Island (Clague et al. 1980) and continuing glacioisostatic subsidence might be expected. Although such subsidence may affect the 1931-1974 uplift field (Fig. 4) in a nonuniform fashion, the differential effect across the relatively small area occupied by the Kalaloch syncline would probably be minimal. Inaccurate rates of eustatic sea-level rise would be represented as spatially uniform error in the geodetic uplift rates. Savage et al. (1991) used the sea-level rise estimates of Peltier & Tushingham (1989) to revise the uplift rates of Holdahl et al. (1987). Further refined sea-level rise estimates would uniformly affect the geodetic uplift rates (Fig. 4), at least at the scale of the syncline. Thus, the magnitude of the difference between geological and geodetic uplift rates would be affected, but the trend of the differenced values would not.
A l t e r n a t e g e o d e t i c uplift r e p r e s e n t a t i o n s
More recent studies of contemporary uplift have resulted in different uplift field patterns, which compare differently with the geological deformation pattern. The uplift rate comparisons presented above, and the resulting interpretations, rely on the uplift field pattern as calculated by Holdahl et al. (1987), using 1974 relevelling of c. 1931 benchmarks. Mitchell et al. (1994) used 1988 resurveys of Oregon benchmarks to constrain the uplift field along that portion of the margin, and recalculated and recontoured the existing 1931-1974 data in Washington. The result is a more generalized uplift pattern for Washington than that of Holdahl et al. (1987). Mitchell et al. (1994) depicted uplift rates increasing uniformly across the area occupied by the Kalaloch syncline. That is, the axial zone of the syncline is not reflected as a minimum in the geodetic uplift rates. As with the Holdahl et al. (1987) uplift rate contours, however, the uplift rate contours of Mitchell et al. (1994) are substantially larger than the geological uplift rates reflected by the syncline, indicating interseismic strain accumulation ofc. 0.3-1.3 mm/year (as much as 1.6mm/year for geological uplift rates corresponding to the inferred substage 5e and 5c wave-cut surface ages). Hyndman & Wang (1995) depicted a yet more generalized geodetic uplift field that results in similar interpretations. They used the same 19311974 Washington dataset as Holdahl e t al. (1987) and Mitchell et al. (1994). However, they filtered out most coastal variations in geodetic uplift, arguing that the variations are more probably caused by measurement errors or unstable benchmarks than by tectonic processes. Their geodetic uplift contours, representing relevelling and tidegauge data fitted to a viscoelastic model of strain accumulation, imply uniform 1931 1974 uplift of 3-4 ram/year along the central Olympic coast. Within this view, the Kalaloch syncline bears no spatial correspondence to the 1931-1974 uplift field. As with the comparisons between geological uplift rates and the geodetic rate contours of Mitchell et al. (1994), the 1931-1974 geodetic uplift implied by the Hyndman & Wang (1995) contours is substantially greater than the geological uplift. Comparison of geological uplift with the latter implies 2-4 mm/year of inferred interseismic strain accumulation. R e l e v a n c e to other aspects o f Cascadia Subduction Zone tectonism
The relative proportions of permanent uplift and interseismic strain accumulation inferred for the
DEFORMATION OF PLEISTOCENE STRATA, WASHINGTON, USA central Olympic coast are similar to those inferred for portions of the Oregon coast. Through the uplift rate comparisons above, I inter that the geodetically measured uplift is dominated by interseismic strain accumulation, amounting to 0.4-1.0mm/year (as much as 4mm/year if alternative geodetic uplift interpretations are considered)- short-term uplift that should be recovered as coseismic subsidence in a megathrust earthquake. In a broader study covering much of the Oregon portion of the Cascadia margin, Kelsey et al. (1994) also inferred that geodetic uplift rates largely reflect short-term strain accumulation. They calculated geological uplift rates ranging from -0.04 to 0.87mm/year, very similar to the range of Olympic geological uplift rates described above. On portions of the Oregon coast, however, the inferred interseismic strain component is as high as 4.5ram/year,as much as an order of magnitude larger than the interseismic strain component on the central Olympic coast. The inferred Olympic interseismic strain rates are comparable with only one section of the Oregon coast. On the basis of this comparison, I infer that although the Olympic coast is experiencing substantial interseismic strain accumulation, the rate is much less than that along portions of the Oregon coast. These contrasts in coastal interseismic uplift may reflect a broader locked zone between the North American and Juan de Fuca plates in the Olympic coast region. Hyndman & Wang (1995) used a thermal model, fitted to horizontal and vertical geodetic data, to estimate the widths of the locked zone on the plate boundary. They concluded that the locked zone is wider across the Olympic Peninsula (85 kin) than across other portions of the margin (35-60km). The wider locked zone is related to a gentler dip on the subducting plate and consequent higher heat flow. The broad locked zone in the Hyndman & Wang (1995) model corresponds to an inland zone of geodetically constrained uplift (Mitchell et al. 1994) that is greater than the coastal uplift (Fig. 4).
Conclusions Quaternary tectonic deformation on the central Olympic coast is dominated by a syncline deforming the Pleistocene stratigraphic sequence. The Kalaloch syncline is a broad ( > 3 2 k m wavelength), low-amplitude structure (c. 45m) reflected in the outcrop pattern of a wave-cut surface (isotope stage 7 or 5) and outwash (isotope stage 4), as well as in observable dips of older strata. The stratigraphic setting of the syncline
209
permits detailed analysis of the character of the deformation and of corresponding geodetic data. Interseismic strain accumulation appears to dominate geodetically measured uplift. Geological uplift rates determined from the buried wave-cut surface are low (<0 to 0.7mm/year). Comparisons of the geological rates with 19311974, geodetically determined uplift rates indicate that recoverable interseismic strain is accumulating, at rates of 0.4-1.0 mm/year (as much as 4 ram/year if alternative geodetic interpretations are utilized). Comparison of spatial uplift patterns, using the most detailed geodetic uplift field representations (Holdahl et al. 1987), suggests that interseismic strain uplift is nearly constant across the Kalaloch syncline, and that the syncline accounts for most of the spatial variability in 1931-1974 uplift. This project has been funded by National Science Foundation Grant EAR-9405659 to S. C. Porter and G.D.T., and by research grants to G.D.T. from the Geological Society of America, the Sigma Xi Scientific Research Society, and the University of Washington Department of Geological Sciences. The project has been aided greatly, both in and out of the field, by S. C. Porter, F. J. Pazzaglia, W. J. Gerstel, B. F. Atwater, and A. R. Gillespie. M. Burrill, J. E. Thackray, A. M. Wiens, and J. Williams assisted with field work. I also thank L. Workman of the Quinault Indian Nation and personnel of the Hoh Tribal Center and Olympic National Park for facilitating land access. Radiocarbon dating was performed by the NSF-Arizona AMS facility, by the University of Washington Quaternary Isotope Laboratory, and by Beta Analytic, Inc. Thermoluminescence dates were determined by G. Berger of the Desert Research Institute, University of Nevada, and palaeomagnetic polarities by H. Rowe and J. Geissman of the University of New Mexico. Reviews by H. M. Kelsey and K. Berryman generated significant improvements in the manuscript.
References ADAMS, J. 1990. Paleoseismicity of the Cascadia subduction zone: evidence from turbidites off the Oregon-Washington margin. Tectonics, 9, 569-583. ATWATER, B. F. 1987. Evidence for great Holocene earthquakes along the outer coast of Washington State, Science, 236, 942-944. - - 1 9 9 2 . Geologic evidence for earthquakes during the past 2000 years along the Copalis River, southern coastal Washington: Journal of Geophysical Research, 97 1901-1919. BALDWIN, E. M. 1939. Late-Cenozoic diastrophism along the Olympic Coast. MS thesis, State College of Washington. BLOOM,A. L., BROECKER,W. W., CHAPPELL,J. M. A., MATTHEWS, R. K. & MESOLELLA, K. J. 1974. Quaternary sea level fluctuations on a tectonic coast. Quaternary Research, 4, 185-205.
210
G. D. T H A C K R A Y
BRADLEY, W. C. & GRIGGS, G. B. 1976. Form, genesis, and deformation of central California wavecut platforms. Geological Society of America Bulletin, 87, 433-449. BRANDON, M. T. & VANCE, J. A. 1992. Tectonic evolution of the Cenozoic Olympic subduction complex, Washington State, as deduced from fission track ages for detrital zircons. American Journal of Science, 292, 565-636. CHAPPELL, J. & SHACKLETON, N. J. 1986. Oxygen isotopes and sea level. Nature, 324, 137-140. CLAGUE, J. J., ARMSTRONG, J. E. & MATHEWS, W. H. 1980. Advance of the late Wisconsin Cordilleran ice sheet in southern British Columbia since 22 000 yr b.p. Quaternary Research, 13, 322-326. DAR[ENZO, M. E. 1991. Late Holocene paleoseismicity along the northern Oregon coast. PhD dissertation, Portland State University. --& PETERSON, C. D. 1990. Episodic tectonic subsidence of late Holocene salt marshes, northern Oregon coast, central Cascadia margin, USA. Tectonics, 9, 1-22. DODGE, R. E., FAIRBANKS, R. G., BENNINGER, L. K. & MAURRASSE, F. 1983. Pleistocene sea levels from raised coral reefs of Haiti. Science, 219, 1423-1425. FLORER, L. E. 1972. Quaternary paleoecology and stratigraphy of the sea cliffs, western Olympic Peninsula, Washington. Quaternary Research, 2, 202-216. FORMAN, S. L. & ENNIS, G. 1992. Limitations of thermoluminescence to date waterlain sediments from glaciated fiord environments of western Spitsbergen, Svalbard. Quaternary Science Reviews, 11, 61-70. GOLDFINGER, C., KULM, L. D., YEATS, R. S., APPLEGATE, B., MACKAY, M. E. & MOORE, G. F. 1992. Transverse structural trends along the Oregon convergent margin implications for Cascadia earthquake potential and crustal rotations. Geology, 20, 141-144. HEATON, T. H. & KANAMORI, H. 1984. Seismic potential associated with subduction in the northwestern United States. Bulletin of the Seismological Society of America, 74, 933-941. HEUSSER, C. J. 1972. Palynology and phytogeographical significance of a late Pleistocene refugium near Kalaloch, Washington. Quaternary Research, 2
189-201. --
1973. Environmental sequence following the Fraser advance of the Juan de Fuca lobe, Washington. Quaternary Research, 3, 284-306. HOLDAHL, S. R., FAUCHER, F. & DRAGERT, H. 1989. Contemporary vertical crustal motion in the Pacific northwest. In: COHEN, S. C. & VANICEK, P. (eds)
Slow Deformation and Transmission of Stress in the Earth. Geophysical Monograph, American Geo--,
physical Union, 40, 17 29. MARTIN, D. M. & STONEY,W. M. 1987. Methods for combination of water level and leveling measurements to determine vertical crustal motions. In:
Proceedings of Symposium on Height Determination and Recent Crustal Movement in Western Europe. Dumler Verlag, Bonn, 373-388.
HYNDMAN, R. D. & WANG, K. 1993. Thermal constraints on the zone of major thrust earthquake failure: the Cascadia subduction zone. Journal of Geophysical Research, 98, 2039 2060. -& 1995. The rupture zone of Cascadia great earthquakes from current deformation and the thermal regime. Journal of Geophysical Research, 100, 22133-22154. IMBRIE, J., HAYS, J. D., MARTINSON,D. G. et al. 1984. The orbital theory of Pleistocene climate: Support from a revised chronology of the marine 180 record. In: BERGER, A. L. et al. (eds) MiIankovitch and Climate, Part 1. D. Reidel, Dordrecht, 26%305. KAUFMAN, D. S., FORMAN, S. L., LEA, P. D. & WOBUS, C. W. 1996. Age of pre-late-Wisconsin glacialestuarine sedimentation, Bristol Bay, Alaska. Quaternary Research, 45, 59-72. KELSEY, H. M., ENGEBRETSON,D, C., MITCHELL, C. E. & TICKNOR, R. L. 1994. Topographic form of the Coast Ranges of the Cascadia Margin in relation to coastal uplift rates and plate subduction. Journal of Geophysical Research, 99, 12245 12255. --, TICKNOR, R. L., BOCKHEIM, J. G. & MITCHELL, C. E. 1996. Quaternary upper plate deformation in coastal Oregon. Geological Society of America Bulletin, 108, 843 860. MARTINSON, D. G., PISIAS, N. G., HAYS, J. D., IMBRIE, J., MOORE, T. C. & SHACKLETON, N. J. 1987. Age dating and the orbital theory of the ice ages: development of a high-resolution 0 to 300000-year chronostratigraphy. Quaternary Research, 27, 1-29. MCCRORY, P. A. 1996. Tectonic model explaining divergent contraction directions along the Cascadig subduction margin, Washington. Geology, 24, 929-932. MCINELLY, G. W. & KELSEY, H. M. 1990. Late Quaternary tectonic deformation in the Cape Arago Bandon region of coastal Oregon as deduced from wave-cut platforms. Journal of Geophysical Research, 95, 6699 6713. MCNEILL, L. C., GOLDFINGER, C., YEATS, R. S. & KULM, L. D. 1998. The effects of upper plate deformation on records of prehistoric Cascadia subduction zone earthquakes. This volume. MESOLELLA,K. J., MATTHEWS,R. K., BROECKER,W. S. & THURBER, D. L. 1969. The astronomical theory of climatic change: Barbados data. Journal of Geology, 77, 250-274. MITCHELL, C. E., VINCENT, P., WELDON, R. J. & RICHARDS, M. 1994. Present-day vertical deformation of the Cascadia margin, Pacific Northwest, USA. Journal of Geophysical Research, 99, 12257-12278. MUHS, D. R., ROCKWELL, T. K. & KENNEDY, G. L. 1992. Late Quaternary uplift rates of marine terraces on the Pacific coast of North America, southern Oregon to Baja California Sur. Quaternary International, 15-16, 121-133. NELSON, A. R. 1992. Holocene tidal-marsh stratigraphy in south central Oregon-evidence for localized sudden submergence in the CascadiG subduction zone. In: FLETCHER, C. H. &
D E F O R M A T I O N OF P L E I S T O C E N E STRATA, W A S H I N G T O N , USA WEHMILLER, J. F. (eds) Quaternary Coasts of the United States." Marine and Lacustrine Systems. Society for Sedimentary Geology Special Publication, 48, 287-301. PELTIER, W. R. & TUSHINGHAM, A. M. 1989. Global sea level rise and the greenhouse effect: might they be connected? Science, 244, 806-810. ROCKWELL, T. K., MUHS, D. R., KENNEDY, G. L., HATCH, M. E., WILSON, S. H. & KLINGER, R. E. 1989. Uranium-series ages, faunal correlations and tectonic deformation of marine terraces within the Agua Blanca fault zone at Punta Banda, northern Baja California, Mexico. In: ABBOTT, P. L. (ed.) Geologic Studies in Baja Calijornia. Society of Economic Paleontologists and Mineralogists, Pacific Section, 1-16. SAVAGE, J. C. & THATCHER, W. 1992. Interseismic deformation at the Nankai trough, Japan. Journal of Geophysical Research, 97, 11 117-11 135. --, LISOWSKI, M. & PRESCOTT, W. H. 1991. Strain accumulation in western Washington. Journal of Geophysical Research, 96, 14 493-14 507. TABOR, R. W. & CADY, W. M. 1978a. The Structure of the Olympic Mountains, Washington- Analysis
--
211
ofa Subduction Zone. US Geological Survey Professional Paper, 1033. & - - 1 9 7 8 b . Geologic Map of the Olympic Peninsula. US Geological Survey Map, 1-994, 1 : 125 000.
THACKRAY, G. D. 1996. Glaciation and neotectonic deformation on the western Olympic Peninsula, Washington. PhD dissertation, University of Washington, Seattle. TRENHAILE, A. S. 1980. Shore platforms: a neglected coastal feature. Progress in Physical Geography, 4, 1-23.
VEEH, H. H. & CHAP~ELL, J. 1970. Astronomical theory of climatic change: support from New Guinea. Science, 167, 862 865. WAITT, R. B. & THORSON, R. M. 1983. The Cordilleran ice sheet in Washington, Idaho, and Montana. In: PORTER, S. C. (ed.) Late Quaternary Environments of the United States. University of Minnesota Press, 53-70. WRIGHT, L. W. 1970. Variation in the level of the cliff/shore platform junction along the south coast of Great Britain. Marine Geology, 9, 347-353.
Variable rates of Late Quaternary surface uplift along the Banda Arc-Australian plate collision zone, eastern Indonesia DOROTHY
M E R R I T T S 1, R E B E C C A
R. L A W R E N C E
E B Y 2, R O N
EDWARDS 4 &HAI
H A R R I S 3,
CHANG 4
1Department of Geosciences, Franklin and Marshall College, Lancaster, PA, 17604-3003, USA (e-mail.
[email protected]) 2 Department of Earth and Planetary Sciences, Washington University, St Louis', MO 63130-4899, USA 3 Department of Geology, Brigham Young, Provo, UT 84602, USA 4 Department of Geology and Geophysics, University of Minnesota, Minneapolis, M N 55455, USA Abstract: Radiometrically dated emergent coral terraces from southeastern Indonesia pro-
vide estimates of differential vertical strain in the Banda Arc-continent collision complex. At Semau island, two samples from the lowest emergent reef (5-7 m) yield 23~ dates that correspond to the 5a (c. 83 ka) sea-level highstand and a low surface uplift rate of 0.2-0.3 m per 1000 years. At Rote island, samples from the lowest emergent reef (c. 1-2 m) on both north and south sides of the island yield late Holocene ages and an average short-term uplift rate of c. 1-1.5 m per 1000 years. Similarity of ages from different samples on the north coast of Rote suggests possible coseismic emergence. Survey data from nine emergent reefs and marine notches up to 170m in altitude on the south side of Rote indicate that uplift rates may have been c. 1-1.5m per 1000 years for c. 120000-130000 years. Combined with previous studies, these results indicate that late Quaternary surface uplift rates vary an order of magnitude along the strike of the Banda orogen. Vertical displacement rates are greatest in young parts of the orogen where the shelf-slope break recently has been underthrust beneath the orogenic wedge, as at Rote, and in older parts of the orogen where retroarc thrust faulting occurs, as at Alor island.
This paper presents new 23~ ages and associated surface uplift rates for emergent coral reefs on two islands near West Timor in southeastern Indonesia (Fig. 1). On Rote island, no previous work has been done on emergent corals. On Semau island, only one coral sample has yielded a previous age estimate (Jouannic et al. 1988). Both islands are located in an important zone of transition from subduction to collision along the plate boundary between the Indian-Australian and Eurasian plates. Whereas the transition from subduction to collision is poorly preserved in most other collision zones, it is well preserved in the Timor region. Many lines of evidence indicate that both horizontal and vertical rates of deformation are greater in this area than elsewhere in the orogen. This paper also compiles all previous estimates of surface uplift rates from coral dating on islands throughout southeastern Indonesia. We use this compilation to consider regional variability in vertical strain rates in relation to recent models of tectonic processes in the Banda Arc-Australian continental margin collision zone.
Cenozoic deformation in the Banda ArcAustralian continental plate collision zone The Indonesian region contains an active subduction zone, accretionary arc, volcanic island arc, and emergent collision zone. In western Indonesia, oceanic crust of the Indian-Australian plate is subducting northward beneath the Eurasian plate along the Java trench, forming the Sunda volcanic arc and accretionary wedge, which includes the volcanic islands of Sumatra and Java (Fig. 1). In eastern Indonesia, where oceanic crust has been subducted completely, underthrusting of buoyant Australian continental lithosphere since about 3 Ma changes the Java Trench into a collisional foredeep and the locally inactive Banda Arc into an arc-continent collision zone along the Timor Trough (Hamilton 1979; Bowin et al. 1980; Silver et al. 1983; Breen et al. 1986; Karig et al. 1987). Underthrusting of the mature Australian continental margin significantly increases the influx of accretionary material and the frictional resistance to slip along the plate boundary (Harris 1991). Collisional strain
MERRITTS, D., EBY, R., HARRIS, R., EDWARDS,R. L. & CHANG,H. 1998. Variable rates of Late Quaternary surface uplift along the Banda Arc-Australian plate collision zone, eastern Indonesia. In: STEWART,I. S. • VITAFINZI, C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 213-224.
214
D. MERRITTS E T A L .
Fig. 1. Plate tectonic map of the Banda Arc orogen (upper right) and enlarged view of eastern Indonesia (lower panel). Rote and Semau are accretionary islands situated in the collision complex. The Timor Trough separates the Indian-Australian plate from the Eurasian plate. Arrows in upper panel represent relative motion of the Indian-Australian plate. Bathymetric contour interval in lower diagram is 1000m.
in this region is increasingly distributed away from the deformational front as continental underthrusting progresses (Harris 1991). Because of the obliquity of convergence, the transition from subduction to collision propagates to the southwest along the plate boundary at an estimated rate of 110 km/106 years (Harris 1991) and currently is south of the island of Sumba at about 120~ longitude (Breen et al. 1986). Recent work in the Banda Arc-continent collision complex has focused on estimating late
Cenozoic deformation rates in response to plate collision and shortening (see Bowin et al. 1980; Audley-Charles 1986; Karig et al. 1987; DeSmet et al. 1989). Geologists have used structural and stratigraphic data, in particular from seismic reflection profiles, in conjunction with biostratigraphic data to estimate rates of horizontal and vertical displacement of rock. Few reliable quantitative estimates of vertical strain rates have been made, however, because of lack of precise indicators of depositional depths of uplifted strata.
LATE QUATERNARY SURFACE UPLIFT, EASTERN INDONESIA In addition, it is difficult to estimate surface uplift rates from elevated columns of rock and sediment unless the amount of material exhumed (eroded) during uplift also is known (England & Molnar 1990). Global positioning system (GPS) surveys have yielded short-term (years) velocity vectors of horizontal displacement throughout the region (Genrich et al. 1996), but again these do not provide rates of vertical surface uplift, which commonly are an order of magnitude smaller than horizontal rates of plate motion and thus are more difficult to measure over short periods of time. From previous work, it is known that rates and styles of horizontal and vertical deformation vary markedly along the Indonesian archipelago as the nature of the plate boundary changes from a zone of subduction to collision. Near 120~ longitude, where subduction is nearly perpendicular to the orientation of the trench (Genrich et al. 1996), little to no horizontal shortening occurs along the Java Trench (Fig. 1). Net surface uplift along the outer edge of the upper plate, as indicated by late Quaternary coral reefs, is minimal and Holocene rates of emergence are less than 1 m per 1000 years (Vita-Finzi & Situmorang 1989). To the east, in contrast, collisioninduced changes along the Timor Trough result in increased rates of horizontal shortening and vertical displacement, as well as greater distribution of strain away from the deformation front into a broad, mountainous orogenic zone. Offshore seismic-reflection and bathymetric profiles examined in the Timor collision zone by Karig et al. (1987) show that relative rates of late Cenozoic horizontal deformation vary markedly throughout the orogenic complex. In addition, Karig et al. qualitatively assessed the relation between horizontal and vertical displacement rates, noting that the vertical component is much smaller than the former, but still substantial. Using structural and depositional information from Pliocene to Quaternary age sediments, they concluded that the most intense horizontal shortening, caused by accretion and thrusting, has occurred along the inner slope of the Timor Trough, referred to as the deformation front. Cumulative horizontal and vertical deformation are greatest between the deformation front and an arch known as the outer arc high that forms the central spines of Savu, Rote, and Timor islands (Fig. 1). Karig et al. estimated that as much as 3-5kin of rock uplift has occurred in parts of the outer arc high since late Miocene time, and DeSmet et al. (1990) determined that rapid uplift in West Timor began about the time that subduction ceased, about 200 000 years ago. Northward from this
215
region of emergence of bathyal sediments, rates of deformation decrease rapidly. Seismic data from the forearc basin region north of the Timor collision zone show that the sedimentary cover is mostly undeformed and tilted to the north, which Karig et al. (1987) interpreted as evidence for no shortening between the forearc ridge and volcanic arc (with exception of a small arch known as the North Sumba Ridge; Fig. 1). However, offshore seismic reflection profiles presented by Silver et al. (1983) and Breen et al. (1986) and onshore field studies conducted by Harris (1991) indicate northward displacement of the accretionary wedge along the Savu thrust and other south-dipping thrusts at the rear, or northern front of the accretionary wedge (Fig. 1). Balanced sections indicate that back-arc convergence may account for as much as 2530% of the total convergence in the collision zone (Harris 1991). These observations are consistent with regional GPS measurements which indicate that parts of Timor are moving essentially northward with the Australian plate relative to parts of the Banda Arc (Genrich et al. 1996).
E m e r g e n t coral r e e f terraces, sea-level change, a n d crustal uplift
Some coral species grow close to sea level, providing precise markers from which to estimate the amount of surface uplift that has occurred since coral death (Bloom et al. 1974; Chappell 1974; Taylor & Jouannic 1985; Chappell & Shackleton 1986; Lajoie 1986; Ota et al. 1993). Coral reefs grow upwards in response to a relative rise in sea level, and are abandoned during times of relative sea-level fall. Reef crests indicate the approximate height of each transgression (Chappell 1974). As a result, flights of reefs formed during periodic sea-level highstands are stranded above sea level in areas where net land mass uplift is occurring. Relative sea-level transgressions and regressions comprise two components of change: eustatic sea-level fluctuation (caused primarily by changing glacial ice volume) and vertical crustal displacement as a result of tectonism. If the former is known, the latter can be deduced from the present altitudes and ages of emergent reefs in a given area. Surface uplift rates for late Quaternary time can be obtained from emergent, unrecrystallized aragonitic corals dated by methods of 23~ isotope analysis. This basic procedure has been applied in different parts of the world with much success (see Bloom et al. 1974; Chappell 1974; Taylor & Jouannic 1985), but early work was hindered by the difficulty of obtaining sufficient
216
D. MERRITTS E T AL.
amounts of unrecrystallized aragonite, which readily converts to calcite in the presence of diagenetic fluids in porous coral structures. Until recently, 23~ dating relied on alpha-counting techniques that required sizeable samples and yielded dates with substantial uncertainties. Because late Quaternary sea level has oscillated at cycles as short as 20000 years, the dating uncertainties limited correlation of certain age reefs with specific sea-level highstands. Recent improvements in methods of 23~ dating that utilize thermal ionization mass spectrometric techniques require only small sample sizes and have improved analytical error, so that uncertainties in final age estimates are smaller (Edwards et al. 1987; Chen et al. 1991). As a consequence, the timing and altitudes of global sea-level highstands now are well constrained for about the past 200000 years (Edwards et al. 1987; Gallup et al. 1994).
Previous estimates o f late quaternary surface uplift in the Banda A r c collision zone
The first to analyse emergent coral reefs in the Timor region were Chappell & Veeh (1978), who worked at four sites along the north coast of East Timor, as well as along the southern coast of a
small island to its north, Atauro (Fig. 2). Their U - T h ages (alpha counting) of coral samples indicated that late Pleistocene surface uplift rates along the northeast coast of East Timor are 0.5 m per 1000 years, but decrease westward to less than 0.03 m per 1000 years near central Timor. The sites in northeastern Timor are close to the axis of the outer arc high, or structural culmination zone where DeSmet et al. (1989) determined that substantial structural relief has developed in Quaternary time. In contrast, the central Timor site is well north of the arch's axis, at the edge of a zone of subsidence known as the Savu Basin (Fig. 1). Along the north coast of West Timor, Chappell & Veeh noted that no emergent reefs occur, and other workers have inferred that little to no net uplift, and instead possibly even subsidence, is occurring in this area (e.g. Karig et al. 1987; VitaFinzi & Hidayat 1991). West of the city of Kupang, U - T h dating (alpha counting) of one coral head in growth position on a reef 44 m above sea level yielded an age of 152+ 10kaBP and a surface uplift rate of 0.3m per 1000 years (Jouannic et al. 1988). Radiocarbon dating of shell material from a low, wave-abraded beach sand deposit 1.3 m above sea level just east of Kupang yielded a mid-Holocene age, and indicates possible local and slight emergence since
Fig. 2. Generalized tectonic map of Timor region, showing compilation of surface uplift rates from this study (bold) and previous work (italics) on emergent coral reefs (see text for references to sources of uplift rate data). K, Kupang; PD, Point Dombo; PT, Point Termanu.
LATE QUATERNARY SURFACE UPLIFT, EASTERN INDONESIA that time if sea level has not risen since then (Vita-Finzi & Hidayat 1991). However, if sea level was higher during mid-Holocene time, the deposit could indicate crustal stability since then. Northwest of Kupang, on the island of Semau, Jouannic et al. (1988) dated a fossil coral from a terrace on the southeast coast at an altitude of about 7 m and obtained a U - T h age estimate of 124+~ ka nP. If this terrace represents oxygen-isotope stage 5e, the average late Pleistocene uplift rate in southern Semau is <0.2 m per 1000 years. Chappell & Veeh (1978) also studied a flight of emergent reefs to the north of the Savu forearc basin, on Atauro island (Figs 1 and 2). This island is part of an inactive segment of the Banda Arc that is surrounded by active faults. Sonar mapping of the sea floor in this region by Breen et al. (1986) revealed that some horizontal shortening and possible vertical deformation are occurring. Indeed, Chappell & Veeh (1978) determined from coral dating that late Pleistocene surface uplift rates are ~ 0 . 5 m per 1000 years at Atauro. More recently, others have investigated emergent reefs on Sumba island, which is an uplifted segment of the Savu forearc basin, and on the volcanic island of Alor, west of Atauro (Fig. 1). A flight of six major reefs with many subterraces and notches spans the coastline from sea level to nearly 500m on the north coast of Sumba, where the North Sumba Ridge crosses the island (Figs 1 and 2). Uranium-series and electron spin resonance dating methods yielded a late Quaternary surface uplift rate of 0.5m per 1000 years (Pirazzoli et al. 1991, 1993), the same as that obtained by Chappell & Veeh (1978) for the outer-arc high in East Timor (Fig. 2). On the north coast of Alor, which is bounded by a welldeveloped back-arc thrust zone to the north (Silver et al. 1983), six major reefs with many subterraces and notches up to 700 m in altitude were dated with carbon-14, uranium-series, and electron spin resonance methods, and yielded a late Pleistocene surface uplift rate of 1-1.2 m per 1000 years (Hantoro et al. 1994).
Results of this study S i t e descriptions a n d s u r v e y m e t h o d s
Until the work described in this paper, no studies of Pleistocene reefs had been carried out on islands between West Timor and Sumba, with exception of a single coral sample from Semau dated by Jouannic et al. (1988). These islands are located at the important transition
217
zone along the Indian-Australian and Eurasian plate boundary, where subduction ceases and collision begins. We worked on Semau and Rote islands offshore of West Timor to assess flights of emergent reefs for evidence of vertical displacement. Semau is located within the Banda ArcAustralian plate collision zone just west of Timor, north of the axis of the outer arc high (Figs 1 and 2). Emergent reefs are ubiquitous along the northeastern coast of Semau. Reefs commonly encircle mud diapirs capped by mud volcanoes associated with localized release of fluid over-pressure in the collision zone. Sites are accessible only by boat, as few dirt roads and motorized vehicles exist on this remote island. Reef altitudes were obtained using a Lietz hand level mounted on a bracket fixed to a telescoping rod (cumulative errors i 1 m). Indicators of high tide, such as the zone of highest green algae, were used to provide an absolute vertical datum relative to mean sea level. Tidal range is about 2.2 meters in the area. Reefs were surveyed and sampled at Aikalui Point as well as Bahansalit Beach (Fig. 3), where reefs occur at similar altitudes, but only those from Aikalui Point yielded unrecrystallized aragonite (Table 1). At Aikalui, sample A4 was collected from a coral (genus Faviidae) exposed in a prominent marine notch 2-3 m below a reef crest at c. 8 m above mean sea level. Sample A5 (genus Faviidae) was collected from the reef crest. Flights of emergent reefs are prominent on most of the island of Rote, the southernmost island in Indonesia (Figs 1 and 2). Rote is located along the axis of the outer arc high, closer to the deformation front than any other island. Warped reefs arch upwards over the island's central ridge, rising to altitudes of nearly 400m. Quaternary bathyal sediments from deep in the accretionary prism are exposed just below the capping of coral limestone, indicating substantial amounts of vertical displacement. An extensive flight of reefs was surveyed on the southern coast of the island, at Point Dombo (Fig. 4), using a US Paulin System Altimeter (range of 0-5000 feet, or 0-1524m), with supplemental hand-level surveying as described for Semau. A base camp near sea level was used to monitor barometric changes over the duration of the survey period (4 days) and to correct altimeter readings for diurnal fluctuations. Hand-level surveys were used in addition to altimeter readings to assess the uncertainty of the latter, which was determined to be + 3 - 5 m. As on Semau, the zone of highest green algae was used to provide an absolute vertical datum relative to mean sea level. Only samples from the
218
D. MERRITTS ET AL.
Fig. 3. Eustatic sea-level curve from Huon Peninsula, Papua New Guinea (Chappell & Shackleton 1986; Gallup et al. 1994) and correlation of interglacial and interstadial sea-level highstands with emergent coral terraces on Semau. Combined topographical transect from Aikalui Point and Bahansalit Beach was obtained with a hand level (see text); horizontal distances are shown diagramatically. The total horizontal distance shown is about 400 m. Vertical error bars for survey altitudes are + 1 m. Two samples dated from reef II correlate with stage 5a (83 ka), and are used to estimate an uplift rate of 0.2-0.3 m per 1000 years. lowest reefs immediately along the coast yielded unrecrystallized aragonite. Aragonite was obtained from the lowest reef at Point Dombo, just below a prominent marine notch, as well as from a broad reef in an embayment on the north coast of the island, at Point Termanu (Fig. 2; Table 1).
X-ray diffraction and petrographic analysis Each sample was divided into four parts: one for archival purposes, a second for X-ray diffraction (XRD) analysis, a third for thin section preparation, and a fourth for U - T h dating. For X R D analysis, samples were submersed in acetone and ground with a mortar and pestle, then sieved through 230 mesh to eliminate particle-size effects during X-ray diffraction. A cavity mount of the sample was prepared and a full scan from 2 ~ to 60 ~ 20 was conducted. To determine the per cent aragonite in each sample, we used the Gibbs (1971) working curve of the ratio of the height of the calcite fundamental peak to the height of the
aragonite fundamental peak (h3.4o/h3.o3) v . per cent aragonite. Those samples which indicated large amounts of aragonite (Table 1) then were used for further petrographic analysis and radiometric dating. Standard thin sections were prepared using a clear epoxy resin. The thin sections were etched in l % hydrochloric acid and stained with an alizarin red S and potassium ferricyanide stain to aid in identification of carbonate minerals (Dickson 1966). Thin sections were examined under a petrographic microscope to characterize the state of preservation, the mineralogy as indicated by the stain, and the crystal form and degree of neomorphism. Samples selected for 23~ age dating exhibited excellent preservation of the original coral skeleton and yielded less than 5-10% calcite during X R D analysis. Unique petrographic features include pore spaces that essentially are void of cement and sediment, and aragonite needle cements along primary pore edges. In addition, thin rims of intermediate iron calcite were noted on several samples, indicating that
LATE QUATERNARY
0 m
~%._
SURFACE
UPLIFT, EASTERN
INDONESIA
r-i
~
44
44
r-i 44
..4 44
~ 44
-H
tth
,-~
~
~
oo
44
44
44
44
-/4
219
.'~ ..--..
44 R
06
o ~
~s 44
44
,~
oo
-H
~
44
r'--
-H
-H
-H
~
.
44
.
.
-t4
t'~
',~
-H
-H
~=
&-
sx
9~
. X ,,,.~
-H
44
-H
44
44
""
,4T
44
~:
.~ .= .~
~
~
S
S
S .~
4-t o.
-H m.
-4-1 o.
44 ~
-H o-:.
44
-H
-H
t'-,l
~2,
4-t
-H
4-1
t"-,I
r
~
t'4
s
t"-I
44
4t
44
-H
44
-H
<
u_. <e
~
-H
'~
,..~ ,o
2~'
. ~
=_,9.
#
>
.,,,.
o ~.~ x~
.~ -
II .s S" o <
.<
.<
.,~
b 0
"CI r~
O m~
~_,, ~
,..~1 ----.~
~au ~
~'~
~
NO
.~_ ~- ~ ~ ~ ~ :.-:~ ~,_..
W=
~Z
~,
0,20 .,-~ ~
~_~
~'~
-'4-
;~
4---
e.a,~ ~,,.o
~ ~ ~ ~
-~.-- c.o~ ~
~
~
220
D. MERRITTS ET AL.
Fig. 4. Eustatic sea-level curve from Huon Peninsula, Papua New Guinea (Chappell & Shackleton 1986; Gallup
et al. 1994) and preliminary correlation of interglacial and interstadial sea-level high stands with emergent reefs at Point Dombo, southern Rote. A sample dated from reef ! was used to estimate an uplift rate of 1-1.2m per 1000 years. However, different uplift rates were used to test the degree of fit between older reefs and the times and altitudes of sea-level highstands. That shown here is the most consistent with all reefs surveyed. It should be noted that reefs correlate with times of sea-level peaks, but steps and notches continue to form even as sea level regresses. A similar process is occurring today. As the lowest reef at Point Dombo is raised above a relatively stable sea level, a broad surface is formed, with notches and steps that span several metres in altitude. At the time of the next eustatic sea-level regression, more notches and steps might form before a subsequent sea-level highstand.
reducing and oxidizing pore fluids passed through the system and some neomorphism occurred (Aikalui 4 and 5 Termanu 2). The diagenetic environment may have been in the deep sea or on a shallow-water platform; however, given the geological context, the most likely scenario involves a shallow platform with storm waves carrying marine waters into the supratidal regime.
U - T h sample preparation and analytical procedures Isotope analysis for radiometric dating was done at the University of Minnesota Isotope Laboratory. Detailed analytical procedures for measurement of Th and U isotopes have been described by Edwards et al. (1993) and Gallup et al. (1994), and are modifications of those described by Chen et al. (1986) and Edwards et al. (1987). 23~ dating of corals is based on the decay of 238U through a series of relatively shortlived intermediate daughters to 2~ In this chain, 234U decays to 23~ with a mean life of 352 700 years, which makes these nuclides suita-
ble for dating Quaternary corals. The initial amount of 23~ in corals is extremely low and can be assumed to be zero (Edwards et al. 1988). If corals are assumed to represent closed systems from which no 23~ 234U,or 238Uescapes or into which no 23~ 234U, or 238U enters, equations for radioactive production and decay can be used to solve for coral age as a function of the decay constants for each nuclide and the measured values of the ratios of 234U/238Uand 23~ (Chen et al. 1986; Edwards et al. 1987). In addition, the initial value of ~5234U when the system was isolated from seawater can be calculated for each sample and compared with the present seawater value as an independent check on the assumption of closed system behaviour.
N e w coral ages and estimates o f Late Quaternary surface uplift rates in the Banda A r c collision zone At Aikalui Point, northeastern Semau island, two coral samples (A4 and A5) from a low emergent reef yield dates that correlate with the 5a (c. 83kaBP) interstadial sea-level highstand
LATE QUATERNARY SURFACE UPLIFT, EASTERN INDONESIA (86.5 • 88.3 +0.5kaBP; Table 1). Although these samples contain some calcite (<5-10%), the estimated initial value of 6234U when the system was isolated from seawater is similar to the modern value (c. 150-165), indicating that diagenesis has not been extensive and the nominal age is close to the true age. At this time, there is increasing effort to re-evaluate the timing and altitude of the interstadial highstand that corresponds to oxygen isotope stage 5a. Most sea-level estimates range from as high as at present to as low as - 1 9 + 5 m (see review by Ludwig et al. (1996)). Using an altitude of formation of - 1 3 to - 1 8 m for the 5a highstand (Gallup et al. 1994) and the present reef crest altitude of c. 5-8 m to compute a total amount of uplift ((+5 to 8m) to ( - 1 3 to - 1 8 m ) = 1826 m), we estimate a very low surface uplift rate of c. 0.2-0.3 m per 1000 years. On the basis of this uplift rate and estimates of sea-level highstand altitudes (Chappell & Shackleton 1986; Gallup et al. 1994), we predict that the 5c terrace (c. 100kam') should be located at present altitudes of c. 8 - 1 8 m above sea level, and the 5e terrace (c. 120-130 ka m') at c. 3 0 - 4 5 m above sea level. Many higher emergent reefs occur at Aikalui Point, including a prominent reef complex at an altitude of about 35-40 m but no higher reefs have yet been dated. Comparison of our surveyed reef crest and marine notch altitudes (from sea-level to 1 9 + l m ) with the eustatic sea-level curve (Chappell and Shackleton 1986; Fig. 3) indicates that multiple crests and notches are associated even with single sea-level highstands, including the present highstand. This same phenomenon has been noted on the islands of Sumba (Pirazzoli et al. 1991, 1993) and Alor (Hantoro et al. 1994) in Indonesia, and on the Huon Peninsula, Papua New Guinea (Ota et al. 1993). Ota et al. (1993) argued that in Papua New Guinea the multiple benches and notches on a given reef complex are the result of episodic coseismic emergence, not just a continuously changing relative sea level. It might be possible with more detailed surveying and sampling to test this hypothesis for Semau, although it is easier at a site with higher uplift rates and younger emergent reefs, such as Rote (discussed below). By assuming that uplift rates in northeastern Semau have been relatively constant for the past 120 000 years, we tentatively correlate reefs I, II, and III with the 5a interstadial sea-level highstand, and IV and V with the 5c interstadial sealevel highstand (Fig. 4). We propose that notches associated with surfaces I and IV formed at the end of times of interstadial highstands, as falling sea level undercut highstand surfaces. The uplift
221
rate at Semau is low enough that all reefs from highstands younger than oxygen isotope stage 5a have been submerged by the present 0 - 6 k a interglacial highstand. This highstand is responsible for the prominent notch forming at the base of the 5a reef (reef I), and indicates the modern high-water mark. On Rote island, coral samples from two sites yield late Holocene reef ages and relatively high short-term uplift rates. At Point Termanu, on the north side of Rote (Fig. 2), two samples from the lowest emergent reef (c. 1 -t-0.5 m above sea level) yielded ages of 1008 -4- 13 yrs and 1035:5 8 years (Table 1). An average short-term uplift rate of c. 1 + 0.5 m per 1000 years is calculated for Point Termanu, assuming that sea level has changed little in altitude during that time (see discussion below). About 20 km to the south, at Point Dombo, a sample from the lowest emergent reef(c. 2 + 0.5 m above sea level) yielded ages of 1319 + 18 years and 1300 + 17 years (duplicate analyses), and an average short-term uplift rate of c. 1.5 + 0.4 m/ky (Table 1). Based on these average short-term uplift rates, we predict that the 5a terrace would be located at present altitudes of c. 60-110m above sea level. Higher emergent reefs occur in this altitudinal range on the island, but have not yet been dated. All samples collected from these terraces contained more than 5-10% calcite. Until more age control is available, it is possible to use the eustatic sea-level curve derived from Papua New Guinea (Chappell & Shackleton 1986) and elsewhere (Gallup et al. 1994) to infer possible ages of each reef crest and erosional marine notch surveyed with altimeter and hand level (Fig. 4). Assuming that uplift rates have been fairly constant and that the late Holocene uplift rate is a reasonable value for longer-term rates, we identified a plausible best-fit solution for reef ages. This solution matches the most prominent reef complexes at 170 i 3 to 180 :t: 3 m 139 + 3 m and 102 + 3 to 119 -t- 3 m with the 5e, 5c, and 5a sea-level highstands, respectively. All lower reefs and notches are correlated with interstadial highstands younger than stage 5. It should be noted that uplift rates are high enough at Rote, as compared with Semau, that the 6 ka marine notch is well above sea level and is not submerged during high tide. Multiple notches and reef crests spaced closely together on each broad reef surface suggest that episodic coseismic emergence might occur on Rote, as at Papua New Guinea (Ota et al. 1993). Because each reef crest spans time periods of several thousands of years, the individual steps notched into them probably represent events separated by durations of several hundred to a thousand years or so.
222
D. MERRITTS E T AL.
Evidence o f possible coseismic emergence on R o t e island
Preliminary evidence that episodic coseismic emergence might have occurred on Rote comes from the 23~ dates of the late Holocene reef complexes. Recent estimates of late Holocene sea level for the Timor Sea indicate that sea level was about + 1 m at 6 ka, but has decreased to c. 0 m since then (Nakada & Lambeck 1987; Hantoro et al. 1994). No significant sea-level change has been documented for the past 1000 years, so emergent reefs with ages of about 1000 years can be explained only by crustal deformation, be it coseismic or aseismic. If multiple dates from corals at the same site yield similar ages, then it is possible that death was sudden and widespread, as during coseismic emergence (e.g. Taylor & Jouannic 1985; Ota et al. 1993). At Point Termanu on Rote, two samples (T2 and T3) collected about 10 m from one another yield age estimates (+1 SD) that differ by only 6 years for samples that aremore than 1000 years old (Table 1). The close similarity in ages suggests that the north coast of the island might have experienced sudden, coseismic uplift, in a manner similar to recent and historic events elsewhere in Indonesia (VitaFinzi & Hidayat 1991; K. Sieh, unpub, data). In contrast to Rote, the other places in Indonesia with historic coseismic emergence, or evidence of prehistoric coseismic emergence, occur to the west, where subduction of oceanic crust still is active, as along the coast of Sumatra. On the south side of Rote, at Point Dombo, a single sample yielded replicate ages that are about 300 years greater than at Point Termanu. If this reef were raised by coseismic deformation, then different earthquake sources generated uplift at different times on the north and south sides of the island.
Conclusions regarding vertical strain in the Banda Arc collision zone Compilation of results from new 23~ ages of emergent coral reefs on islands near Timor and from previous work throughout the Timor region indicates that late Quaternary surface uplift rates vary more than an order of magnitude along and across the emergent Banda Arc orogenic complex (Fig. 2). In general, vertical displacement rates are greatest near the deformation fronts in both fore- and retro-wedge parts of the orogen, where rates of horizontal shortening also are largest. Surface uplift rates decrease northward toward the forearc basin, but increase again where horizontal shortening is accommo-
dated by back-arc thrusting in the retro-wedge region along the Banda volcanic arc, north of the forearc basin. GPS measurements indicate that rates of horizontal shortening (convergence) are low in Timor (c. 20 mm/year), but increase westward to 46mm/year in Sumba (Genrich et al. 1996). Between these areas are the zone of transition from collision to subduction and the islands of Semau and Rote. Just as GPS results indicate that the rates of horizontal displacement increase, the results of our new dates indicate that rates of vertical strain might increase, too. Diapirism clearly influences the pattern of deformation on Semau island, as evidenced by the circular reef fringes around Aikalui Point and many other mud diapirs on the island. Nevertheless, rates of vertical deformation from the past 125000 years have been very low (c. 0.2-0.3m per 1000 years or less) along the north and northeastern coast of the island. The diapiric zone that forms the island is located near a major lateral ramp zone that bounds the western edge of Timor and separates it from the most submerged and narrow part of the outer arc ridge to the southwest (Harris 1991); however, its structural significance is poorly constrained. Southwest of Semau and this lateral ramp zone, near the deformation front in Rote, uplift rates are much higher, with late Holocene rates of 1-1.5m per 1000 years. In addition, 23~ dating of a reef l + 0 . 5 m above sea level suggests that coseismic emergence might have occurred about 10084-13 to 1035-t-8 years BP. Multiple notches and reef crests on higher terraces in southern Rote indicate that episodic coseismic emergence in the area has occurred possibly for at least the past 120000-130000 years. It is suggested here that coseismic emergence in Rote is associated with vertical accommodation of movement along contractional faults in the collisional wedge, as documented at other convergent plate boundaries (Carver et al. 1994; Merritts 1996). Although coseismic emergence has occurred during historic time to the west along the Java Trench, this is the first time evidence of possible coseismic emergence has been identified for the region of arc-continental margin collision east of Sumba. Emergence near Java is associated with subduction and results in little long-term net surface uplift at the plate boundary (because of post-seismic crustal relaxation), whereas in Rote, net uplift during Quaternary time is substantial, probably as a result of underthrusting of buoyant continental lithosphere. We are grateful for sponsorship and field support from the University Pembangunan Nasional (UPN Veteran)
LATE QUATERNARY
SURFACE UPLIFT, EASTERN INDONESIA
of Indonesia. R. O'Connell and O. O'Connell provided invaluable field assistance for collecting the coral samples. Petrographic expertise was provided by C. DeWet of Franklin and Marshall College. We also thank A. Bloom for advice regarding how to sample corals, and J. Muller for assistance with the XRD analysis. We are most appreciative of financial support from NSF grant EAR-9118151 to R. Harris and from Petroleum Research Fund grant 27865-B2 to D. Merritts.
References AUDLEY-CHARLES, M. G. 1986. Rates of Neogene and Quaternary tectonic movements in the southern Banda Arc based on micropalaeontology. Journal of the Geological Society, London, 143, 161-175. BLOOM, A. L., BROECKER, W. S., CHAPPELL, J. M. A., MATTHEWS, R. K. & MESOLELLA, K. J. 1974. Quaternary sea level fluctuations on a tectonic coast: New 23~ dates from the Huon Peninsula, New Guinea, Quaternary Research, 4, 185-205. BOWIN, C., PURDY, G. M., JOHNSTON, C., SHOR, G., LAWYER, L., HARTONO, H. M. S. & JEZEK, P. 1980. Arc-continent collision in the Banda Sea region. Bulletin, American Association of Petroleum Geologists Bulletin, 64, 868~15. BREEN, N. A., SILVER, E. A. & HUSSONG, O. M. 1986. Structural styles of an accretionary wedge south of the island of Sumba, Indonesia, revealed by Seamarc II side scan sonar. Geological Society of America Bulletin, 64, 868-915. CARVER,G. A., JAYKO,A. S., VALENTINE,D. W. & LI, W. H. 1994. Coastal uplift associated with the 1992 Cape Mendocino earthquake, northern California, Geology, 22, 195-198. CHAPPELL, J. M. 1974. Geology of coastal terraces, Huon Peninsula, New Guinea, a study of Quaternary tectonic movements and sea level changes. Geological Society of America Bulletin, 85, 553-570. -& SHACKLETON,N. J. 1986. Oxygen isotopes and sea level. Nature, 324, 137 140. -& VEEI-I, H. H. 1978. Late Quaternary tectonic movement and sea-level changes at Timor and Atauro Island. Geological Society of America Bulletin, 89, 356-358. CHEN, J. H., CURRAN, H. A., WHITE, B. & WASSERBURG, G. J. 1991. Precise chronology of the last interglacial period: 234U-23~ data from fossil coral reefs in the Bahamas. Geological Society of America Bulletin, 103, 82-97. , EDWARDS, R. L. & WASSERBURG, G. J. 1986. 238U, 234U and 232Th in sea water. Earth and Planetary Science Letters, 80, 241-251. DESMET, M. E., FORTUIN, A. R., TJOKROSAPOETRO,S. & VAN HINTE, J. E. 1989. Late Cenozoic vertical movements of nonvolcanic islands in the Banda Arc area. Netherland Journal of Sea Research, 24, 263-275. k , TROELSTRA, S. R., VANMARLE, L. J., ARMINI, M. & HADIWASATRA,S. 1990. Detection of collision-related vertical movements in the Outer
223
Banda Arc (Timor, Indonesia), using micropaleontological data. Journal of Southeast Asian Earth Science, 4, 337-356. DICKSON, J. A. D. 1966. Carbonate identification and genesis as revealed by staining. Journal of Sedimentary Petrology, 36, 491-505. ENGLAND, P. & MOLNAR, P. 1990. Surface uplift, uplift of rocks, and exhumation of rocks. Geology, 18, 1173-1177. EDWARDS, R. L., BECK, J. W., BURR, G. S., DONAHUE, D. J., DRUFFEL, E. R. M. & TAYLOR, F. M. 1993. A large drop in atmospheric 14C/12C and reduced melting during the Younger Dryas documented with Th-230 dating of corals. Science, 260, 962-968. --, CHEN, J. H. & WASSERBURG, G. J. 1987. 238U234U-23~ systematics and the precise measurement of time over the past 500 000 years. Earth and Planetary Science Letters, 81, 175-192. , TAYLOR, F. W. & WASSERBURG, G. J. 1988. Dating earthquakes with high precision Th-230 ages of young corals. Earth and Planetary Science Letters, 90, 371-381. GALLUP, C. D., EDWARDS, R. L. & JOHNSON, R. G. 1994. The timing of high sea levels over the past 200,000 years. Science, 263, 796-800. GENRICH, J. F., BOCK, Y., MCCAFFREY, R., CALAIS,E., STEVENS, C. W. & SUBARYA,C. 1996. Accretion of the southern Banda arc to the Australian plate margin determined by global positioning system measurements. Tectonics, 15, 288-295. GraBS, R. D. 1971. X-ray diffraction mounts. In: CARVER, R. E. (ed.) Procedures in Sedimenztao' Petrology. Wiley-Interscience, New York, 552-563. HAMILTON,W. 1979. Tectonics of the Indonesian Region. US Geological Survey Professional Paper, 1078. HANTORO, W. S., PIRAZZOLI,P. A., JOUANNIC, C. et al. 1994. Quaternary uplifted coral reef terraces on Alor Island, East Indonesia. Coral Reefs, 13, 215-223. HARRIS, R. A. 1991. Temporal distribution of strain in the active Banda orogen: a reconciliation of rival hypotheses. In: HALL, R., NICHOLS, G. & RANGIN, C. (eds) Journal of Southeast Asian Earth Science, 6, 373-386. JOUANNIC, C., HOANG, C. Y., HANTORO, W. S. & DELINOM, R. M. 1988. Uplift rate of coral reef terraces in the areas of Kupang, West Timor: preliminary results. Palaeogeography, Palaeoclimatology, Palaeoecology, 68, 259-272. KARIG, E. E., BARBER, A. J., CHARLTON, T. R., KLEMPERER, S. & HUSSONG, D. M. 1987. Nature and distribution of deformation across the Banda arc-Australian collision zone at Timor. Geological Society of America Bulletin, 98, 18-32. LAJOIE, K. R. 1986. Coastal tectonics. In: ROBERT & WALLACE Active Tectonics. Studies in Geophysics. Geophys. Res. Forum. National Academy Press, Washington, DC, 95-124. LUDWIG, K. R., MUHS, D. R., SIMMONS,K. R., HALLEY, R. B. & SHINN, E. A. 1996. Sea-level records at ~80 ka from tectonically stable platforms: Florida and Bermuda. Geology, 24, 211-214.
224
D. M E R R I T T S ET AL.
MERRITTS, D. J. 1996. The Mendocino triple junction: active faults, episodic coastal emergence, and rapid uplift. Journal of Geophysical Research, 101, 6051-6070. NAKADA, M., & LAMBECK, K. 1987. Late Pleistocene and Holocene sea-level change in the Australian region and mantle rheology. GeophysicalJournal, 96, 497-517. OTA, Y., CHAPPELL, J. M., KELLEY, R., YONEKURA, N., MATSUMOTO, E., NISHIMURA, T., & HEAD, J. 1993. Holocene coral reef terraces and coseismic uplift of Huon Peninsula, Papua New Guinea. Quaternary Research, 40, 177-182. PIRAZZOLI,P. A., RADTKE,U., HANTORO,W. S., JOUANNIC, C., HOANG, C. T., CAUSSE,C. & BOREL BEST, M. 1991. Quaternary raised coral-reef terraces on Sumba Island, Indonesia, Science,252, 1834-1836. ---, . . . . & - - 1 9 9 3 . A onemillion-year-long sequence of marine terraces on
Sumba Island, Indonesia. Marine Geology, 109, 221-236. SILVER, E. A., REED, D. R., MCCAFFREY, R. & JOYODIWIRYO, Y. 1983. Back arc thrusting in the eastern Sunda arc, Indonesia: a consequence of arc-continent collision. Journal of Geophysical Research, 88, 7429-7448. TAYLOR, F. W. & JOUANNJC, C. 1985. Quaternary uplift history of the Torres Islands, northern new Hebrides frontal arc: comparison with Santo and Malekula Islands, central New Hebrides frontal arc. Journal of Geology, 93, 419-438. VITA-FINZI, C. & HIDAYAT, S. 1991. Holocene uplift in West Timor. Journal of Southeast Asian Earth Science, 6, 387-393. ~; SITUMORANG, B. 1989. Holocene coastal deformation in Simeulue and Nias, Indonesia. Marine Geology, 89, 153-161.
Quaternary marine terraces and tectonic uplift rates on the south coast of Iran J. L. R E Y S S l, P. A. P I R A Z Z O L I 2, A. H A G H I P O U R 3, C. H A T T I ~ 1 & M. F O N T U G N E 1
1 Centre des Faibles Radioactivit~s ( C N R S - C E A ) , Avenue de la Terrasse, 91198 Gif-sur- Yvette Cedex, France (e-mail."
[email protected]) 2 CNRS-Laboratoire de GOographie Physique, 1 Place Aristide Briand, 91190 Meudon-Bellevue, France 3 Persian Gulf International Centre for Biotechnology Research, Qeshm Island, h'an Abstract: A survey of 1500km of the Iranian coastline, from Bushehr (Persian Gulf) to the Pakistani border, made it possible to visit several sequences of raised marine terraces, some of which (especially on the Makran coast) had already been described by previous workers. As many as 18 marine terraces, up to 220 m in altitude, were identified on Qeshm Island and as many as 19 levels, up to 246m in altitude, near Chah Bahar. Though in situ coral heads and fossilized marine shell deposits exist in many uplifted areas, only a few samples, mainly from elevations lower than 30m, appeared to be unrecrystallized and thus suitable for radiometric dating. Six uranium-series analyses of aragonitic corals from Haleh, Jazeh and Qeshm Is. gave apparent ages between 100 and 140 ka BP, in some cases demonstrating that previous radiocarbon dates between 20 and 40 ka BP had to be considered minimum ages. This made it possible to estimate that average uplift rates since the last Interglacial period had been about 0.2 mm/year, i.e. one order of magnitude less than most previous estimates. Radiocarbon dates suggest almost uniform emergence of 2-4m since mid-Holocene times along most of the coastal sectors investigated. This is in agreement with the 2 3 m emergence predicted by glacio-hydro-isostatic models in this area if an uplift rate of about 0.2mm/year is also taken into account. Near a salt dome on Qeshm Is., however, a much faster uplift rate of about 6 mm/year was documented for the period between 6 and 5 ka BP but only over a limited area.
Marine terraces along the south coasts of Iran have long been used as evidence of recent coastal movements but uncertainty over their age has fuelled debate about the rate, duration and magnitude of tectonism in the region. The existence of marine terraces in the Persian Gulf and Makran coasts of Iran has been discussed by Blanford (1872), Pilgrim (1908) and Falcon (1947). A m o n g others Falcon identified terrace levels at 76-90m, 30m and 15m along the Makran coast and concluded that movement had been spasmodic. The first radiocarbon determinations at these sites were published by Little (1970), who obtained several dates between 20 and 30 ka BP Discrepancies between the height and the age of the terraces led him to the conclusion that uplift and warping, as well as faulting, caused pronounced tectonic activity during this interval. Several additional radiocarbon dates older than 20kaBp on the Makran coast of Iran were reported by Vita-Finzi (1981). Page et al. (1979) used different methods to date subsets of two shell samples collected a few metres above sea level from a marine terrace near Jask (Fig. 1). They obtained dates of
34.31 4- 3.0 ka BP and 26.025 + 1.05 ka BP with the radiocarbon method, 133 + 13 ka ~P and 136 414 ka BP with the Th/U method, and 1Aa+s~ -v_32 ka BP and 11A+S0 ..7_32 kaBP with the Pa/U method, respectively. Similar results were obtained from a marine terrace at an elevation of c. 20m at Konarak, near Chah Bahar. Page et al. (1979) concluded that radiocarbon dates greater than 20 ka ~P of shell material had to be considered minimum ages, and favoured uranium-series determinations. They concluded that rates of uplift along the Makran coast of Iran ranged between 0 and 0.1 mm/year since the last interglacial, whereas Holocene rates were estimated at between 0.3 and 2 mm/year, the rates of uplift increasing eastward towards Pakistan. Vita-Finzi (1980), however, doubted 'that uranium series dates on shells are any more dependable than 14C dates'. According to this researcher, 'the 14C ages would seem to deserve provisional acceptance' until the uranium-series technique is applied to associated corals. An average uplift rate of 2.7 ram/year during the past 30 000 years was proposed by Vita-Finzi (1980) for those parts of the Makran coast not affected by major faulting. Snead (1993), who ignored the
REYSS, J. L., PIRAZZOLI, P. A. HAGHIPOUR, A. 1998. Quaternary marine terraces and tectonic uplift rates on the south coast of Iran. In: STEWART,I. S. & VITA-FINZI, C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 225-237.
226
J. L. REYSS E T A L . !
~"
'
!
52~
" ~
'l . . . . .
54~
o ':
56~
1oo '
9
2ookm ~
0, 10 20k , , m / ~ ~
_ ~ ;'S.7
-2 "~ Qcsh~ Islaad A h . . J . ~
i
~%e
9
?o'~
,
9
58"E
60*E
6bE
. .
Fig. 1. Location and simplified structural map of the area investigated. Locality names on Qeshm Island: 1, Laft-e-Kuhneh; 2, Kuvehei; 3, Tula; 4, Tourgan; 4b, Berkeh Khalaf (Persian Gulf Biotechnology Research Centre); 5, Suza; 6, Namakdan salt dome; 7, southwest of Dustekuh; 8, west of Basaidu.
conclusions of Page et al. (1979), believed that marine terraces were formed during two time periods, at c. 30 14C ka BI, and 23 laCka BP, and that they were subsequently uplifted by renewed faulting. On Qeshm Is., uplift rates ranging from 1.1 to 2.6 mm/year were deduced by Haghipour & Fontugne (1993) from marine terraces radiocarbon dated between 4.7 and 26.7kaBP Such rates are, however, increased by overestimation of the depth of deposition of the dated samples and by the assumption that the apparent date of 26.7 ka BP is not a minimum age. Discussions arising from the last paper stimulated the organization of a new survey of raised marine terraces on the southern Iranian coasts, which was carried out by three of us (M.F., A.H. and P.A.P.) in May 1994, from Bushehr (Persian Gulf) to the Pakistani border (Oman-Makran Sea), a distance of about 1500 km (Fig. 1). Qeshm Is. was surveyed in greater detail. The samples collected during the survey were subsequently analysed at the Centre des Faibles Radioactivit6s and at the underground laboratory at Modane (LSM). Preliminary conclusions drawn from the results obtained during the 1994 survey have been presented by Fontugne et al. (1997). Geological background
The northern Persian Gulf forms the southeastern part of the Zagros geological-structural province. It is characterized by NW-SE trend-
ing fold-thrust belts folded and deformed by the latest phase of the Alpine orogeny in PlioPleistocene times. Geological formations may range in age from Palaeozoic to Late Tertiary time and include diapirs related to the early Palaeozoic Hormuz series, which have been active up to recent times. This area has been tectonically active since late Tertiary time, as part of the southern deformational front or convergence belt (Mesopotamian foreland and Persian Gulf basin) and as part of the Arabia-Iran continentcontinent compressional and collision plate margins (Haghipour & Aghanabati 1985-1989; Haghipour 1989; Alavi 1994; Bushara 1995; Edgell 1996). The Makran coast is part of an accretionary wedge of deformed sediments, of late Cretaceous to Holocene age, which have accumulated near an oceanic subduction margin. The Makran province is bordered by two major fault zones of mainly recent strike-slip motion, the right-lateral Zendan Fault zone on the west and the leftlateral Chaman Fault zone on the east (Haghipour et al. 1984), which are under the influence of the interplate collision system. The frequent occurrence of uplifted marine terraces is evidence of active tectonism in this well-exposed and young accretionary margin. Reverse faults, steeply dipping to the north or north-west, are roughly parallel to the fold axes, which generally strike east-west; the density of reverse faults generally decreases towards the coast.
QUATERNARY UPLIFT, SOUTH COAST OF IRAN Located at the south-eastern end of the Persian Gulf in the Strait of Hormuz, Qeshm Is. is characterized by the same geological, sedimentary and structural features as the mainland, from which it is separated by a strait only 2.5 km wide at its narrowest point. The emergence of Tertiary sediments on Qeshm is., partly associated with Quaternary deposits, thus results from compressive tectonic forces related to the latest phase of the Alpine orogeny (Haghipour 1989). Both the Zagros and Makran provinces are seismically active. Frequent historical and instrumental earthquakes, occasionally ranging up to M 7, have been registered in the Zagros-Qeshm regions and even larger events, though less frequent, from the Makran province (Ambraseys & Melville 1982; Ni & Barazangi 1986; Byrne et al. 1992; Haghipour 1992; Berberian 1995; Jackson et al. 1995).
Marine terraces Quaternary marine terraces are often well developed, with horizontal or slightly seaward-sloping stepped surfaces. Older terraces, especially those that overlie bedrock anticlinal structures, may be more steeply sloping, gently folded and faulted.
227
The terraces are made of reef and beach deposits laid down on the shore or in shallow water and varying in thickness from a few metres to about 10 m. Lithologically, they consist of coral and lumachelle limestone-gritstone and sandstone deposits, generally with a duricrusted surface, which overlie unconformably thicker but weaker bedrock formations, commonly Miocene-Pliocene marly-silty and sandy red beds. A maximum inland extension of the marine terraces to about 10 km from the present-day shoreline has been observed on Qeshm Is. and in Makran province. Erosion of the underlying bedrock often leaves impressive landscapes, with disordered piles of broken Quaternary slabs filling eroded depressions (Fig. 2). As many as l 8 terrace levels have been observed in a single transect on Qeshm Is. (Fig. 3) and as many as 19 terrace levels in the Chash Bahar area (Fig. 4), reaching an altitude of 220 m in the former and 246 m in the latter. Only a few of them have been displaced as the result of faulting. Several marine terraces displaying a gradual slope have been shaped by more than one sea-level stand. According to local harbour and hydrographic authorities and to the National Oceanic and Atmospheric Administration (NOAA) tide tables, the local tidal range varies from about
Fig. 2. Spectacular landscape near the edge of a marine terrace at about 60 m near Tang, 75 km west of Chah Bahar (photograph by P.A.P., no. E94).
228
J. L. REYSS E T AL.
\ x
\
) MIO-PLIOCENE I ANTICLINE STRUCTURE k20m:>31 ka<"C); 127:s
/
4zo
_
~
+rim: ~39-~6k~( ~ r h ) ~
\
/--
,.
N%15:
+16.5 m.">40 ka (~C) !" ~ .... MSL +5.5 m: 4.7 ka (~4C) "7 +I m
EASTERN QESHM ISLAND N ofTula (Lot.3)
Tourgan (Loc.4)
Fig. 3. Schematic profile of the marine terraces along a cross-section between Locality 3 and Locality 4, eastern Qeshm Island. The continuous line represents Quaternary marine deposits capping Mio-Pliocene deposits. Arrows correspond to locations where the altitude was measured. Dating results are also summarized.
1.5 m near Bushehr to a maximum of 5 m in the narrow strait between Qeshm Is. and the mainland, falling to between 1.8 and 2.7m along the coasts of the O m a n - M a k r a n Sea. Indian Ocean cyclones and atmospheric depressions during the southwesterly monsoons may cause strong swells on the Makran coast.
Methods In the field, geographical coordinates were determined using a pocket global positioning system (GPS) device, with an estimated accuracy of the
order of 50m. Altitudes are with reference to mean sea level (MSL), if not otherwise stated. Elevations were generally measured with reference to the sea surface for Holocene samples (and then corrected according to tide predictions), or to high-water marks (and then corrected for the local tidal range), by means of a spirit level and a folding rule. For higher terraces, elevations were estimated by averaging readings on two pocket altimeters. Because interpolated corrections for barometric changes were often not possible, altitudes of inland higher terraces are probably accurate to no better than 4-5 m.
246 m
+210 m 00m
@,-~.
m
+16o m
+14o
m.-?,~".~" x~10
m
' ~
....
Fig. 4. As for Fig. 3 but near Chah Bahar.
+36 m '/ +28 m z +20m ~ t ~ / + 4 m: 3.7ka (NC) -- ~ - -
MSL
QUATERNARY UPLIFT, SOUTH COAST OF IRAN To reconstruct former MSL positions, an additional vertical uncertainty equalling half the local tidal range was ascribed to sampled beach deposits. More precise estimations were possible when fossilized organisms collected in growth position (e.g. barnacles) could be measured with reference to the position of their modern counterparts. Corals are known to develop below the low-tide level, reaching depths that vary according to the species considered, though generally not exceeding 15-20 m in the area considered because of limited water transparency. Accordingly, it has been assumed that coral samples collected in growth position generally indicate minimum former low-tide level elevations, and are thus usually associated with the next slightly higher terrace level. Well-preserved horizontal reef flats are thought to correspond closely to former low-tide positions. Reconstructions of former sea levels are more difficult on sloping terraces where a polycyclic origin seems likely. In these cases, the stratigraphical position and altitude of the sample to be dated remain the most objective altitudinal criterion, whereas the elevation of the sloping terrace surface must be interpreted on a case-bycase basis. Samples selected for absolute age determination consisted of shells and corals which had already been investigated by X-ray diffraction. In spite of the aridity of these regions, most of the corals were recrystallized, suggesting more humid conditions in the past. Fourteen 14C dates were estimated by 13-counting on shell samples from the lowest terrace. Six additional dates obtained from a previous survey were also considered (Table 1). Conventional ages were expressed as recommended by Stuiver & Polach (1977) and were calibrated using the marine calibration curve proposed by Stuiver & Braziunas (1993), with an apparent age of 190 years for the Persian Gulf (J. Southon & M. Fontugne, unpubl, data). Except for Gif-8622 and -8624 the ~13C values are in the range of variation observed for modern marine carbonates and do not indicate recrystallization. Shell samples consist of aragonite or calcite. A content of <1% calcite in Holocene aragonitic samples will not affect radiocarbon ages significantly. For samples older than 20 000 years, however, even minor recrystallization can have serious effects, and the ages for such samples are therefore considered minimum values. In the case of aragonitic coral, samples containing more than 2% calcite were discarded. Only six samples were suitable for 23~ dating of terraces between 10 and 26m above sea level (Table 2). The radiochemical procedure
229
used for thorium and uranium isotope separation was similar to that described by Ku (1976), using a 232U/228Th spike. Alpha counting was carried out either in a grid chamber or with a solid-state detector. Errors are expressed as 1 SD based on counting statistics.
Results At Oli, 4 k m west of Dayyer, a reef flat with corals and shells in growth position that crops out near MSL (Fig. 5) is indicative of a relative sea-level fall of about 1.5 + 0 . 5 m . A shell sample collected from the reef flat was dated to 3270 + 45 a BP (Gif-9840). Near the same location, another coral reef flat at about 10m in altitude could not be dated owing to recrystallization of the samples collected. Near Haleh, a sequence of four Pleistocene coral reef terraces have developed above the Miocene bedrock (Fig. 6). The highest reef, at about 100-110m was rich in fossil corals which again had been strongly recrystallized. The second terrace, at about 85 m, was narrower and appeared to be of abrasional origin; it did not yield any coral. The third terrace sloped between 33 and 23m. Aragonitic coral in growth position near its outer margin has been dated by 23~ to 111 i 4 ka BP. The lower part of this terrace may therefore be ascribed to oxygen-isotopic stage 5e or 5c. The fourth terrace, between 5 and 2.5m was bordered at its seaward margin by a cliff which is notched at present sea level. The only date obtained from this site indicated uplift at about 0.2 ram/year. At this rate, the lowest elevated terrace would have last been reached by the sea during isotopic stage 5c or 5a. Several terrace levels including beach deposits at various elevations could also be seen near Jazeh (Fig. 6). The two lower steps, at about 21 m and 9 m, may date from isotopic stage 5e (Table 2), suggesting a local uplift rate of no more than 0.2ram/year. On the shore at Pol, where construction of the Persian Gulf Bridge between Qeshm Is. and the mainland is planned, beach deposits including marine shells collected at 2.5 m have been dated to 3 6 1 0 ~ 3 5 a B P (Gif-9841). As the local tidal range here reaches 5m an uncertainty range of +2.5 m has to be ascribed to this sample for the reconstruction of the former MSL position. On Qeshm Is. (Fig. 1 Laft-e-Kuhneh, Locality 1), lightly cemented shelly marine deposits can be found behind the present beach up to 1.5 m above high-tide level (Fig. 7). They have been dated to 5 6 5 0 + 5 0 a B p (Gif-9847). West of Kuvehei (Locality 2), an elevated beachrock is
230
E T AL.
J. L. R E Y S S
v v ~ 8_ _ v v
~ ~ v v _ _ v8 ~
~
v ~
vv
b
0
2. ~
0
/.-,,
,....,
0 0 0 0 ~:~0
<&
~
~
A
AAAA
.
~
.
.
~
~
AA
.
A
AAA
~
Z
~
~
9
8~ 0 .,,,~ c~ 0
,..a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
QUATERNARY UPLIFT, SOUTH COAST OF IRAN
A
A
~D
v
9---,
t',l
~t"xl
~D
0
0
ggg 0
O
231
232
J . L . REYSS E T AL.
Fig. 5. Late Holocene exposed coral reefs now located above MSL provide evidence of 1.5 4, 0.5 m of emergence near Dayyer. They were dated by radiocarbon to 3270 4-45 a Be (photograph by P.A,P., no. E35). visible about 2.5m above beachrock which has developed in the present intertidal range. On the surface of the elevated beachrock, barnacle shells are still preserved in growth position (Fig. 8). They are 2.5 m above the level of similar barnacles living at present, and have been dated to 6210+ 50aBP (Gif-9848). In the same location marine shells collected from beach deposits at 1.5 m above the present high-tide level have been dated to 4400 + 50 a BP (Gif-9849). A transect across the easternmost part of Qeshm Is. is shown in Fig. 3. The Mio-Pliocene anticlinal structure appears capped almost everywhere by a thin layer of Quaternary coral reef
+llOm
sediment. North of Tula (Locality 3), where a marine terrace at 21m had been radiocarbondated to >31 kaBp by Haghipour & Fontugne (1993), renewed sampling in 1994 yielded coral from the cliff face at 12m (i.e. 8 m below the terrace edge at 20 m but stratigraphically continuous with the terrace) which was dated by 23~ to 1 3 9 + 6 k a a p , whereas another coral sample taken from the surface of the terrace edge was dated to 127 + 6 ka BP. In the middle of the transect, undated sloping terraces rise by a series of steps to about 170 m. Similar terraces, sloping downwards, were found on the other side of the erosional valley that cuts the anticlinal
+lOOm
I03~85 m
\
~ ~
+23 m: 11.1+4 ka (z-~~
MSL HALEH
Fig. 6. As for Fig. 3 but near Haleh and Jazeh.
+54m
~
9_ ' " ~ "~:x.. '"x~'Ar~.
+38 m x " ~ +20m" ]lS**~ " ., . ', J J '~-~,,~ +9 m: 130 + -12 lea ( ~ ] 1 )
' JAZEH
~
-
-
MSL
QUATERNARY UPLIFT, SOUTH COAST OF IRAN
Fig. 7. Beach deposits reaching about 4m and dated to 5650 + 50 a BP on the north coast of Qeshm Island (Locality 1) (photograph by P.A.P., no. E153).
233
structure. Near Tourgan (Locality 4), a coral head in growth position at about 26 m (Fig. 9) was dated by 23~ to 104_+4 kaBp, suggesting that the coral developed during isotopic stage 5c. Three other terraces occur at lower elevations: between 15 and 19 m, at about 6 m, and at about 1 m (within the present intertidal range). In a previous study (Haghipour & Fontugne, 1993), a coral sample at 16.5m within the highest of the three terraces, was radiocarbon-dated to >40kaBP, and a coral at 5.5m, from the 6 m terrace, was dated to 4680 + 50 a Be. The lowermost intertidal terrace appears to be an abrasional feature cut at present sea level into an older coral-reef formation. Near Suza (Locality 5) a terrace at about 6.5 m was radiocarbon-dated to 26.73+0.52kaBP by Haghipour & Fontugne (1993). This is considered to be a minimum age. Two formations containing shell debris, separated by a thin marl layer, are visible below the 6.5 m terrace, where marine erosion is cutting two small platform levels, at about 2 m and near the present MSL. Both of these lower formations are of Pleistocene age (Table 1). In the western part of Qeshm Is., near a salt dome (Locality 6), halokinetic processes have uplifted Holocene shorelines. There is here a sharp peripheral upwarping of salt with very steep folding and overturning of the Neogene bedrock and sharp tilting or upwarping of Quaternary caprocks. Shell deposits at about 11 m have been
Fig. 8. Fossilized barnacles dated to 6210 i 50 a BP. about 2.5 m above their modern counterparts, on the north coast of Qeshm Island (Locality 2). Scale is 4cm. Local tidal range is about 5 m (photograph by P.A.P., no. E159).
234
J. L. REYSS E T AL.
Fig. 10. Remnants of a quay built by the Portuguese near the northwestern boundary of Qeshm Island (Locality 8). A mooring ring can still be seen still in place (arrow). Water level is 0.2 m above MSL. High water is 1.8 m above the ring (photograph by P.A.P., no. E217).
Fig. 9. Coral head collected in growth position at about 26 m near Tourgan (Qeshm Island, Locality 4) and dated to 104_+4 ka (photograph by P.A.P., no. E139).
dated to 5380 4-100 a BP by Haghipour & Fontugne (1993). New samples, collected in 1994, provided further dates of 5680 4- 70 a 13p (Gif-9852) at about 12 m and of 5040 4- 50 a BP (Gif-9855) at about 8 m suggesting that most of halokinetic uplift occurred in mid-Holocene times. Marine terraces outcropping at about 4 m above the high-tide level on the southwesternmost tip of Qeshm Is. (Locality 7) and at about 4.5 m above the high-tide mark on the northwesternmost tip (Locality 8) are both of Pleistocene age (Table 1). Near Locality 8 remnants of a quay built by the Portuguese can be seen (Fig. 10). A mooring ring is still preserved there, proving that the quay surface has not been eroded. As the ring and the quay surface are now below high tide level, a slight relative rise in sea level since the time the quay was constructed can be inferred. Similar evidence was reported by Stein (1937) for Hormuz Is., where the remains of the Old Hormuz harbour may be traced on ground which is liable to flooding at high tide. Near Jask, again on the mainland, a marine terrace ascribed by Page et al. (1979) to the last Interglacial (see above) has developed about 2 m
above the present high-tide level. The poorly cemented nature of the sediments may suggest relatively recent deposition. However, this terrace is clearly older than Holocene time. Three radiocarbon dates of shell samples gave apparent ages between 23 and 3 1 k a a P (Vita-Finzi 1981). In the absence of coral heads in living position, only one other radiocarbon age estimate was attempted here, using marine shells collected from close to the terrace surface. The date obtained (24.8kaBP, Gif-9842), is considered to be a minimum age and supports previous estimates. Near Chah Bahar, at least 19 marine terraces could be distinguished between the present sea level and 246m (Fig. 4). Fossilized corals were rare, however, and only the lowest bench, 5-6 m wide, could be dated. It is situated at around 4 m (i.e. about 2 m above the high-tide level) and yielded shells which were dated to 3670 4- 50 a BP (Gif-9843). At higher elevations, three Ostrea samples from the 20 m terrace had been radiocarbon dated to between 28 and 30 ka or >34 ka by Vita-Finzi (1981). Lastly, at Gwater, near the Pakistani border, a marine terrace developed at 3.5m gave an apparent age of 2 5 . 9 7 + 0 . 3 k a a P (Gif-9844), which we consider as a minimum age. A similar date (23.6+0.65kaBP) was obtained from the same terrace by Snead (1993), who considered it finite. From a 29 m high marine terrace near the
QUATERNARY UPLIFT, SOUTH COAST OF IRAN same location, two minimum radiocarbon dates of about 27 ka BP were obtained by VitaFinzi (1981).
Discussion and conclusions For late Pleistocene time, only four outcrops, at altitudes from 9 to 26 m could be dated by 23~ with ages scattered between 104 and 139kaBp. Four of the six samples analysed, however, contained some calcite (though less than 2%) and their 234U/238U ratios were not strictly representative of that of seawater (1.14), facts which could provide an explanation for the scatter in the age values that are often encountered in studies. Nevertheless, all of these formations clearly date from the last Interglacial period and can be associated with isotopic stages 5e or 5c. In both cases, they would indicate an average uplift rate of about 0.2 ram/year.
235
For Holocene time, the available information is summarized in Fig. 11. Two contrasting patterns appear. Most of the samples, from Dayyer to Chah Bahar, are consistent with a maximum sea-level elevation of 2-4 m, which was reached in mid-Holocene time and decreased gradually after 4 ka BP. Sea-level oscillations might have occurred during this period, but were unlikely to have exceeded + 1.0 m in amplitude, within the resolution of the relative sea-level band obtained. Such consistency over a distance of more than 1000 km suggests an absence of major tectonic deformation during the period considered. Other elevated beach deposits which have been radiocarbon-dated from the Holocene include a shell sample from the continental Iranian coast (26~ 55~ facing Locality 8 on Qeshm Is., at 3.0 m above high-water mark and dated to 5040 + 115 a ~P, two shell samples from the Iranian coast (27~ 56~ facing the northern part of Qeshm, at 2.3 and 4.0 m above
Fig. 11. Time-elevation diagram of the Holocene sea-level data obtained from the Iranian coast. Dotted lines delimit an interpreted band of relative sea-level change for the whole region when areas locally affected by major tectonism are disregarded. A-A, relative sea-level change caused by halokinetic processes near a salt dome, western Qeshm Island.
236
J. L. REYSS E T AL.
high-water mark and dated to 3960 :k 100 and 5410~ 110a BP respectively, and two shell samples from near Jask (25~ 57~ ~) at 1.3 m above high-water mark and dated to 6595 • 125 and 4870+ 100aBP (Vita-Finzi 1979, 1981). All these data are consistent with the regional sealevel band proposed in Fig. 11 if adequate uncertainty margins are considered, that is, if the influence of the tidal range and storm waves on the level of deposition of beach material is taken into account. Using a mathematical model to reconstruct the glacio-hydro-isostatic effects of the last deglaciation, Lambeck (1996) recently concluded that sea level rose to its present position shortly before 6000 years ago and then exceeded it by 2-3 m. Such a prediction is perfectly consistent with most of the Holocene field data obtained from the Iranian coast if the effects of an average uplift rate of about 0.2 ram/year (i.e. of about a metre since mid-Holocene times), calculated from Late Pleistocene terraces dated with the uranium series method, is taken into account. In comparison with the area parallel to the coast in the northern Zagros the areas discussed here have been active tectonically during the Alpine orogenic period. Nevertheless, the radiometric dates reported here suggest that uplift rates have not exceeded 0.2mm/year during Late Quaternary time. It is possible that higher rates can be found further inland, in Zagros as well as in the Makran, for instance around salt domes and near the axes of anticlinal folds. Near active faults vertical displacements may also have been faster, but uplift appears in general to have remained rather limited, at least in the Late Quaternary. According to the morphology, number and elevation of older, undated marine terraces, no marked difference can be observed between Qeshm Is. (Fig. 3) and the Chah Bahar area (Fig. 4). In both the Zagros and Makran provinces the rate of uplift may decrease not only with distance from a salt dome or an anticlinal structure, but also generally approaching the coast. In the Mesopotamian Plain and near the northeastern Persian Gulf coasts foreland (southwestern Zagros), the intensity of Neogene-Quaternary deformation is clearly much gentler than in the strongly folded and imbricated zones of the northeastern Zagros; it is probably the same for the accretionary margin of coastal Makran relative to the northern Makran ranges. On the other hand, there is evidence of rapid, localized uplift of tectonic origin. The rate of 6 ram/year, obtained for the period 6-4 ka BP at Locality 6 on Qeshm Is. (similar to the 6-7 mm/ year reported for a diapir in the Dead Sea Basin by Frumkin (1996)), is close to the uplift rate
of 7 ram/year between 6.5 and 10 ka BP inferred by Vita-Finzi (1979, 1980) from a sequence of Holocene raised beaches reaching an altitude of about 30m near Tujak, 55 km N W of Jask. And, as we have seen, archaeological remains indicate a slight rise in high-tide level in the Strait of Hormuz area during the last seven centuries. We are grateful to Qeshm Free Area authorities for their logistical and financial support, and to L. Froget, S. Martin and C. Noury for preparing samples for 14C dating. Thoughtful reviews by anonymous reviewers, and additional comments from I. Stewart, helped to improve this paper. This is CFR Contribution No. 1996.
References ALAVI, M. 1994. Tectonics of the Zagros orogenic belt of Iran: new data and interpretations. Tectonophysics, 229, 211-238. AMBRASEYS,N. N. & MELVILLE,C. P. 1982. A History of Persian Earthquakes. Cambridge University Press, Cambridge. BERBERIAN, M. 1995. Master 'blind' thrusts hidden under the Zagros folds: active basement tectonics and surface morphotectonics. Tectonophysics, 241, 193-224. BLANFORD, W. T. 1872. A note on the geological formations seen along the coasts of Baluchistan and Persia from Karachi to the head of Persian Gulf, and on some of the Gulf Islands. Records' of the Geological Survey of India, 4, 41-45. BUSHARA, M. N. 1995. Subsurface structure of the eastern edge of the Zagros basin as inferred from gravity and satellite data. Bulletin, American Association ~?/'Petroleum Geologists, 79, 1259 1274. BYRNE, D. E., SYKES, L. R. & DAVIS, D. M. 1992. Great thrust earthquakes and aseismic slip along the plate boundary of the Makran subduction zone. Journal of Geophysical Research, 97, 449 478. EDGELL, H. S. 1996. Salt tectonism in the Persian Gulf basin. In: ALSOP,G. I., BLUNDELL,D. J. & DAVIDSON, I. (eds) Salt Tectonic,;, Geological Society, London, Special Publications, 100, 129-151. FALCON, N. L. 1947. Raised beaches and terraces of the Iranian Makran coast. Geographical Journal, 109, 149-151. FONTUGNE, M., REYSS, J. L., HATT/~, C., PIRAZZOL1, P. A. & HAGHIPOUR, A. 1997. Global sea level changes as indicated by 14C and 23~ datings of marine terraces of the Persian Gulf and the Makran coast (Iran). In: Earth Processes in Global Change - Climate of the Past. Proceedings of the Lanzarote Fuerteventura UNESCOIUGS Meeting, 1 6 June 1995, Universidad de Gran Canaria, Las Pahnas, 81-88. FRUMKIN,A. 1996. Uplift rate relative to base-levels of a salt diapir (Dead Sea Basin, Israel) as indicated by cave levels, ln: ALSOP, G. I., BLUNDELL,D. J. & DAVtDSON, I. (eds) Salt Tectonics. Geological Society, London, Special Publications, 100, 41-47.
QUATERNARY
UPLIFT, SOUTH COAST OF IRAN
HAGHIPOUR, A. 1989. Geology of the Qeshm area. Qeshm Free Area, 5 vols (unpublished). - - 1 9 9 2 . Seismotectonic Map o[ the Middle East, 1/5 000 000 scale. Commission for the Geological Map of the World and Geological Survey of Iran, Tehran. -& AGHANABATI, A. 1985-1989. Geological Map of Iran, 1 : 2 500 000. Geological Survey of lran, Tehran. & FON'rUGNE, M. 1993. Quaternary uplift of Qeshm Island (Iran). Comptes Rendus de l'AcadOmie des Sciences, S~rie II, 317, 419 424. --, GHORASHI, M. & KADJAR, M. H. 1984. Seis-
motectonic map of Iran, Afghanistan and Pakistan, 1:5000000, with explanatory text. Commission for the Geological Map of the World, Subcommission for the Middle East. UNESCO, Paris. JACKSON, J., HAINES, J. & HOLT, W. 1995. The accommodation of Arabia-Eurasia plate convergence in Iran. Journal of Geophysical Research, 100, 15205-15219. Ku, T. L. 1976. The uranium series methods of age determination. Annual Reviews of Earth and Planetary Sciences, 4, 347-379. LAMBECK, K. 1996. Shoreline reconstructions for the Persian Gulf since the last glacial maximum. Earth and Planetary' Science Letters, I42, 43-57. LITTLE, R. D. 1970. Terraces of the Makran coast of Iran. In: SNEAD, R. E. (ed.) Physical Geography of the Makran Coastal Plain of lran. University of New Mexico, Albuquerque, 318-372.
237
NI, J. & BARAZANGI, M. 1986. Seismotectonics of the Zagros continental collision zone and a comparison with the Himalayas. Journal of Geophysical Research, 91, 8205-8218. PAGE, W. D., ALT, J. N., CLUFF, L. S. & PLAFKER, G. 1979. Evidence for the recurrence of large-magnitude earthquakes along the Makran coast of Iran and Pakistan. Tectonophysics, 52, 533 547. PILGRIM, G. E. 1908. The geology of the Persian Gulf and the adjoining portions of Persia and Arabia. Memoirs of the Geological Survey of India, 34, 1 177. SNEAD, R. J. 1993. Uplifted marine terraces along the Makran coast of Pakistan and Iran. In: SHRODER, J. F., JR (ed.) Himalaya to the Sea. Routledge, London, 327-362. STEIN, A. 1937. Archaeological Reconnaissances in
North-Western
India and South-Eastern Iran.
London. STUIVER, M. & BRAZIUNAS, T. F. 1993. Modeling atmospheric 14C influences and 14C ages of marine samples to 10,000 BC. Radiocarbon, 35, 137-189. -& POLACH, H. A. 1977. Discussion: reporting of 14C data. Radiocarbon, 19, 355 363. VITA-FINZI, C. 1979. Rates of Holocene folding in the coastal Zagros near Bandar Abbas, Iran. Nature, 278, 632-633. 1980. 14C Dating of recent crustal movements in the Persian Gulf and Iranian Makran. Radiocarbon, 22, 763-773. - - 1 9 8 1 . Late Quaternary deformation on the Makran coast of Iran. Zeitschrift fffr Geomorphologie, Supplementband, 40, 213-226.
Coastal Tectonics across the South Atlas Thrust Front and the Agadir Active Zone, Morocco MUSTAPHA ABDELMAJID
MEGHRAOUI
1, F A T I M A
CHOUKRI 3 & DOMINIQUE
O U T T A N I 2,
FRIZON
DE LAMOTTE 2
1CNR-GNDT, Research Institute of Recent Tectonics, via del Fosso del Cavaliere, Roma-Tor Vergata, Italy (e-mail:
[email protected]) 2D~partement des Sciences de la Terre, Universit6 de Cergy Pontoise, France 3D~partement de Physique, Universit~ Ibn Tofail, Kenitra, Morocco Abstract: Quaternary marine terraces of the South Atlas Thrust Front near the Agadir seismogenic region are investigated by detailed geological mapping, precise height measurement and tectonic analysis. A coastal step-like morphology with four main benches (QI-Q4) crosses the Ait Lamine-Kasbah anticlines, 10 km north of Agadir. U-Th dating of 12 fossil samples (mainly molluscs) yields two main groups of ages for marine terraces Q4 and Q3 and can be correlated with oxygen-isotope stages 5 and 7 respectively. Terrace height changes significantly near the Kasbah fold, and attains a maximum of 18-28 m for Q4 and 35.5 m for Q3. Estimates and measurements of terrace elevation yield an average late Quaternary uplift rate of 0.1-0.2 mm/year. Modelling of anticlinal structures suggests that the Kasbah fold may have developed as a fault-propagation fold with a low dip angle of 25~ or a listric geometry as implied by flexural slip faulting. Surface ruptures associated with the 1960 (Ms5.9) earthquake coincide with flexural slip faulting showing 4-5m offset of terrace Q3. Incremental movements and uplifted marine terraces on the Kasbah fold are likely to occur during large earthquakes related to the 25 km long flexural slip fault. In zones of compressional tectonics, fold-related faults can produce earthquakes with coseismic deformation large enough to be recorded by recent deposits at the surface. Although this deformation may not be clearly visible during a single seismic event, long-term cumulative tectonic movements may provide information on the late Quaternary evolution of a seismogenic structure (Yeats 1986; Stein & Yeats 1989). Surface ruptures produced during a moderatesized seismic event, particularly in the compressional tectonic domains of north Africa, are difficult to identify mostly because of reduced coseismic displacement and interference with preexisting fold structures (Meghraoui 1991; Aoudia and Meghraoui 1995). The 1960 Agadir earthquake (Ms5.9), which was one of the major seismic events of the fold and thrust belt of north Africa, occurred on the coastal Air LamineKasbah fault-related fold, and it revealed a complex pattern of surface faulting which was not immediately interpreted as an emergent active fault (Roth6 1962). Nevertheless, recent studies of incremental movements on active anticlines which can be related to blind, hidden or emergent active thrusts emphasize the usefulness of tectonic geomorphology for interpreting coseismic deformation and determining the seismic potential of an active zone (Stein & King 1984; Rockwell et al. 1988; Meghraoui & Doumaz 1996).
In this paper, we present an analysis of recent tectonic movements of the Ait L a m i n e - K a s b a h fault-related fold, which shows faulted and folded late Quaternary deposits associated with prominent topographical relief. Uplifted marine terraces of the northern Agadir shoreline provide an indication of the extent, size and velocity of active deformation during the late Quaternary along the South Atlas Thrust Front. The growth of an anticline the and related behaviour of a seismogenic structure are illustrated by means of a fault propagation fold model. Finally, a reappraisal of the 1960 earthquake surface ruptures highlights the importance of coseismic movements on flexural folds and their implications for seismic hazard assessment.
Geological setting of marine terraces The Agadir region is located near the South Atlas Thrust Front (SATF), which is the southern deformation front of the High Atlas fold and thrust belt (i.e. the structural boundary between the Atlas Mountains and the Sahara platform). South of the SATF, the Souss basin constitutes a weakly deformed foreland basin. In the Atlas, Neogene and Quaternary deposits are deformed; on the basis of a kinematic analysis of faults, E1 Maamar (1988) determined
MEGHRAOUI, M., OUTTANI,F., CHOUKRI,A. & FRIZON DE LAMOTTE,D. 1998. Coastal Tectonics across the South Atlas Thrust Front and the Agadir Active Zone, Morocco. In: STEWAgT, I. S. & VITA-F~NZ~, C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 239-253.
240
M. MEGHRAOUI E T A L .
a NNW-SSE shortening component across the SATF. Major compressional structures in the High Atlas consist of a large-scale folding system which affects Mesozoic and Cenozoic formations generally with east-west trending fold axes and thus orthogonal to the Atlantic coastline; Ambroggi (1963) noted the existence of folded Quaternary marine terraces along the coast between Casablanca and Agadir. In contrast to the generally accepted interpretation of transcurrent faulting adopted for the SATF (Weijermars 1987; Cherkaoui et al. 1991), it is now established that this major structure, which runs continuously from Agadir to southern Tunisia, is a thrust front (Frizon de Lamotte et al. 1990; Outtani et al. 1995; Bracene et al. 1998; Saint Bezar et al. 1998). There have been several studies of the marine terraces along the Atlantic coastline. Stearns & Thurber (1965) determined with precision the existence of beaches corresponding to what is now termed isotopic stage 5 and dated to 115-140 ka BP. Weisrock (1980) examined the coastal geomorphology of the High Atlas, established a detailed terrace chronostratigraphy based on the marine molluscan fauna, and reported several phases of river incision and terrace uplift. Finally,
using precise U-Th dating, E1 Gharbaoui et al. (1994) identified two highstands, at 275kaBp and 120 ka BP, which they correlated with interglacial highstands 9 and 5, respectively (Imbrie et al. 1984; Chappell & Shackleton 1986). In the western part of the SATF, immediately north of Agadir, marine terraces are preserved near prominent fold structures. The Ait Lamine anticline, which has an ESE-WNW trending axis and is about 10 km wide immediately north of Agadir, corresponds to an asymmetrical boxfold structure with a gently dipping northern flank and ENE-WSW striking vertical bedding planes on its southern flank (the Kasbah fold, Figs 1 and 2). The fold structure, which is bounded to the north by the Oued Assersif syncline and to the south by the Oued Souss valley, is composed primarily of upper Cretaceous limestone with intercalated clay units. Between the northern and southern flank, the anticline corresponds to a flat plateau that displays the Moghrebian marine terrace of Calabrian or lower Quaternary age (2-3Ma; Ambroggi 1963). The terrace deposits border a bay which emerged during Quaternary time (Figs 2 and 3).
Fig. 1. General tectonic framework of the 1960 Agadir earthquake (Choubert & Faure-Muret 1962; Roth~ 1962). Marine terrace development is associated with the Kasbah and Ait Lamine anticlines. Coseismic surface ruptures mapped by Roth6 (1962) are related to a flexural slip fault. Diagram of fault planes with striations in the fault zone of Founti and indicating the main horizontal shortening direction 1 (crl = N 166) with a slight right-lateral component of movement.
COASTAL TECTONICS ACROSS SOUTH ATLAS THRUST FRONT
241
Fig. 2. Landsat image showing High Atlas geological structures near the Atlantic Ocean coast and the SATF near Agadir. (Note the uplifted flat morphology of marine terrace deposits along the coast north of Agadir.) Rectangle shows area of Fig. 3.
Marine terrace development and active folding Coastal morphology Detailed mapping of the coastal zone near the Ait Lamine and Kasbah anticlines revealed a continuous succession of Quaternary marine deposits (Fig. 3). The field survey consisted of several parallel cross-sections normal to the coast, and aimed at a complete inventory of marine terraces with related deposits and their fossils (Fig. 4a). Altitudes were measured using an Alti Plus-Pretel altimeter, which allows for atmospheric pressures fluctuations and gives elevations with an estimated uncertainty of • m. The results were checked by reference to the 1/2000 scale topographical maps for the area around Agadir harbour and for the Kasbah anticline.
A stepped terrace morphology with four main successive benches is preserved along the littoral and can be observed along the 10 km shoreline north of Agadir (see also Fig. 4b), the total topographical relief across the Kasbah anticline and Souss valley being about 200 m. The marine terraces are preserved probably in response to successive uplift movements and to a low erosional rate during late Quaternary time. The uppermost marine terrace Q1, which is attributed to the Moghrebian stage, corresponds to hardened calcarenite levels (with a rich fauna of Ostrea cucullata, Pecten planornedius and Lissochlamys excisa among other species), which cover the top of Ait Lamine and Kasbah anticlines. The Moghrebian and associated early Quaternary deposits form a flat morphology, with a dip to the north of less than 10~. The same marine formation dips south (c. 25 ~) on the southern
242
M. MEGHRAOUI E T AL.
Fig. 3. Marine terrace distribution along the coastline and across the anticlines. The geological distribution of the Moghrebian deposits is from Ambroggi (1963). Cross-sections approximately normal to the coastline to locations a-h are shown in Fig. 4a. The main terrace deformation and uplift is between Arhesdis, Kasbah and Founti. The flexural fault is coincident with the 1960 surface ruptures of Fig. 5.
flank of the Kasbah anticline and is transgressive on the Mesozoic units. Moghrebian marine terrace Q~ constitutes the western edge of the uppermost bench on the Ait Lamine anticline. It reaches 190m altitude on the north and 200m on the south of the Kasbah anticline (Figs 3 and 4a).
The second flat bench, 4 5 - 6 0 m in height between Assif Tamraght and Anza, rises to a maximum of 70 m at the Kasbah site. The flat bench corresponds to the uppermost surface of marine terrace Q2 and consists of conglomerates overlain by calcarenites which contain mainly
COASTAL TECTONICS ACROSS SOUTH ATLAS THRUST FRONT
(a)
WSW
200 m
ENE
QI~~'~~1
a
100
Q 4 , ~
ENE Q1 ~ - _ _ _ ~
e
ou Q1
0
WSW
243
Q372~ /
Arhesdis1
Q 4 ~ ~ r
200 m 100
0
200 m
200 m
100
Q3 Q2
Use 1
0
Arhe~is2
100
0
QI
200 m
200
100
Use2
O2--._
/
Kasbah1
0
rn
100
0
'OOr.I d
Q2
100
" Anza
0
Q3 f ' Q 4 ~
10o
Kasbah2 0
U
0.'5
l~km
Fig. 4. (a) Cross-sections and related marine terrace altitudes at locations shown in Fig. 3. (Note terrace elevation beginning from Arhesdis 1 section.) Vertical exageration 5 • (b) Step-like morphology of marine terraces Q2, Q3 and Q4 near Cap Bouzellou (see also Fig. 3 and cross section (a). (c) Uplifted terrace Q4 (18 +0.5m) near the lighthouse at Arhesdis.The maximum height of Q4 is 28 m across the Kasbah anticline.
M. M E G H R A O U I ET AL.
244
Fig. 4. (continued)
Trochatella trochiformis, Conus mediterraneus, Chlamys varia, Turritella triplicata and Anomia ephipium. On the basis of the fauna, Ambroggi (1963) attributed this terrace to the 'Anfatian' stage, which can be correlated with the Sicilian stage (0.3-0.7 Ma). Several marine terraces which exhibit a steplike morphology are found below 40 m, but the most prominent surfaces and related deposits are units Q3 and Q4 (Fig. 4). Q3 consists of gravels covered by sand dunes, containing continental silty deposits with no fauna. The marine terrace surfaces are commonly between 20 and 30m in height and the related deposits contain mainly Patella safiana, Mytilus perna and
Turritella flammulata; they may correspond to the Harounian stage (0.2-0.3Ma, lower Tyrrhenian stage). Near the shore, Q4 covers the Mesozoic units showing a clear unconformity. It includes fossil-rich sandy and calcarenitic units displaying conglomerates at the base and dune deposits at the summit (Fig. 4c). The fossilrich units, often characterized by Ostrea edulis, Pecten jacobeus, Patella intermedia and Arca lactea (among others), are at 5-7m; the terrace deposits rise to 10-13 m and form a flat surface near the major beaches. The abundance of large species suggests that Q4 may corresponds to the Ouljian transgression and thus isotopic stage 5.3 (Chappell & Shackleton 1986).
Table 1. U-Th dating of fossil samples collectedfrom marine terraces Q3 and Q4 north of Agadir Location Cap Bouzellou Cap Bouzellou Anza Anza Anza Arhesdis Arhesdis Arhesdis Founti Founti Founti Founti Founti
Sample CB 1 CB 3 Anza 1 96-14-1 96-14-2 ARH 1 96-15-1 96-15-2 Found 2 96-23 Found 3 96-22-1 96-22-2
U (ppm) 1.889 4- 0.023 1.8924-0.057 1.226 4- 0.035 0.5554-0.022 2.264-0.12 3.619 4- 0.108 2.272 4- 0.060 5.438 4- 0.270 0.280 4- 0.010 0.231 + 0.018 5.7504-0.133 3.020 4- 0.056 1.864 4- 0.056
234U/238U (initial)
U/Th age (ka)
1.230 4- 0.039 Ind. 1.414 4- 0.038 1.1024-0.057 1.2624-0.047 1.265 4- 0.027 1.339 + 0.030 1.301 4- 0.042 1.481 4- 0.076 1.996 4- 0.227 0.9304-0.028 1.824 4- 0.035 1.864 4- 0.056
92 (+5.5; -5.2) Old age 119 (+9; -8.3) 101 (+12; -11) 108(+11; -10) 115 (+7.5; -7) 105 (4-6) 105 (+11; - 10) 116.6 (+9.7; -9) 158 (+34; -27) 212(+18; -16) 234 (+18; - 16) 267 (+41; -31)
Correlation with eustatic highstands after Imbrie et al. (1984) and Chappell & Shackleton (1986).
Climatic stage
5.3 5(?) 5(?) 5.3 5 (?) 5 (?) 5.3 7.3 7 (?) 7 (?)
COASTAL TECTONICS ACROSS SOUTH ATLAS THRUST FRONT U - T h dating
A set of 37 of mollusc samples were taken from marine terraces Q4 and Q3 for U - T h dating, to estimate the uplift rate of the Kasbah fold. A preliminary selection of 12 samples was made assuming a chemically closed system (Table 1). Although the dating of molluscs by U-series remains problematic (Stearns & Thurber 1965; Kaufman et al. 1971), the assumption of reliable ages is supported by consistent results after several chemical analyses of the same sample, no evidence of recrystallization indicated by X-ray analysis, and the fact that the ages (Table 1) are consistent with the conventional succession of Quaternary marine deposits. A first group of nine samples gave ages between 101 and 158 ka BP, and a second group of three samples gave ages between 212 and 267kaBP. Because samples A R H 1 96-15-1 and 95-15-2 collected from the Arhesdis location belong to the same level and the same site (Fig. 4c), and uncertainties in radiometric ages also depend on the possibility of environmental exchange, the Arhesdis terrace was correlated with the highstand transgression of stage 5.3. The same observation can be made for the samples from Anza. On the other hand, Founti 3 96-22-1 and 96-22-2 collected in the Founti zone in the same fossil-rich level are correlated with the highstands of stage 7.3 or 7.5 (Table 1).
Terrace elevation changes
The height measurements for each marine terrace show variations in altitude from Assif Tamraght to Anza across the Ait Lamine anticline which fall within the limits of altimeter error
245
(Fig. 3). In contrast, from the Arhesdis lighthouse to the Kasbah and Founti zone marine terraces Qz-Q4 exhibit significant changes in altitude (Figs 3 and 4a, Table 2), which coincide with the strongly folded and faulted Mesozoic units of the Kasbah box fold structure. On the anticline itself, terrace Q4 attains a maximum height of 28 m, Q3 one of 35.5 m, terrace Q2 one of 70m and terrace Q1 one of 210m. Terraces Qz and Q2 are, however, gently folded across the anticline structure, implying either low shortening and uplift rates on the anticline during the Quaternary, or recent folding in middle or late Quaternary time. Indeed, the distribution and folding of terraces Qa and Q2 across the South Atlas flexure structure (Fig. 3 and according to the geological map of Ambroggi (1963)), in conjunction with the extent and uplift of terrace Q3 and Q4 on the western shoreline of the Kasbah anticline, indicate an episode of folding younger than Q2 and older than Q3.
Modelling fold development and terrace deformation The Kasbah and Ait Lamine anticlines are the result of a complex folding process that took place mainly during Neogene time (Outtani 1996). The Quaternary tectonic episodes have been strongly influenced by inherited structures and earlier deformational processes. Structures observed at the surface can be correlated with the geometry at depth because field investigations were combined with seismic profiles and borehole logs across the Souss basin and the SATF near Agadir (Mustaphi 1997). Detailed analysis of stratigraphic sections and fold dimensions have allowed a consistent structural study of the fold
Table 2. Measured and maximum estimated elevations and uplift rates of marine terraces related to the growth of the Kasbah anticline Location
Age (a B.P.)
Marine terrace
Sea level (m)
Measured height (m)
Estimated maximum height (m)
Estimated uplift rate (mm/year)
Arhesdis Arhesdis Founti Founti Arhesdis Kasbah Kasbah
115(+7.5; -7) Harounian + 0.2-0.3 Ma 212(+18; -16) 234(+18; -16) Anfatian (0.3-0.7 Ma) Anfatian (0.3 0.7 Ma) Moghrebian (c. 2 Ma)
Q4 Q3 Q3 Q3 Q2 Q2 Q1
+6 - 10 -10 -10 <present <present <present
18-4-0.5 38 -4-0.5 27+0.5 27+0.5 56 4- 0.5 180 + 2
28• 35.5+0.5 35.54-0.5 70 + 0.5 210 + 2
0.1-0.19 0.16-0.24 0.17-0.21 0.15-0.19 0.08-0.18 0.1-0.23 0.09-0.1
According to the geological succession of Quaternary marine deposits (Fig. 3) and U-Th dating (Table 1), marine terraces Q4 and Q3 are correlated with sea-level altitudes at +6m and -10m that correspond to eustatic highstands 5.3 and 7.3, respectively (Chappell & Shackleton 1986). The average uplift rate during Quaternary time is 0.15 mm/year.
246
M. MEGHRAOUI E T A L .
and thrust system east of Agadir by means of balanced cross-sections and forward kinematic modelling of successive displacements on faultrelated folds (Outtani, 1996; Mercier et al. 1997). Figure 5a and b shows a comparison between a geological cross-section and the final result of forward modelling of a fold using a variety of scenarios of ramps (with different dips) and flats (with different depths). The proposed faultrelated fold structure showing a 25 ~ N N W dipping ramp that levels off at about 6 km depth may not be unique and another geometric possibility not allowed by the model is the existence of listric fault plane structure as shown in Fig. 5a. A similar listric geometry with a relatively highangle fault plane is imaged from the tomography of the E1 Asnam thrust fault which is also associated with a complex folding pattern (Chiarabba et al. 1997).
The proposed model (Fig. 5b) is a 'generalized fold-propagation fault model' (Chester & Chester 1990) which combines a fault-bend fold at depth (here in the Jurassic stiff layers) and a fault propagating in the overlying strata (Outtani 1996). It is worth noting that this scenario is capable of explaining the existence of the double anticline structures of Kasbah and Ait Lamine. Additionally the model takes into account the influence of inherited structures (early Jurassic normal faults) on the development of faultrelated folds. The superimposition of fold structures is a common tectonic feature associated with thrust tectonics and, as discussed earlier, the sharp transgression of Quaternary deposits onto the folded Cretaceous formation of the Ait Lamine-Kasbah anticlines suggests subsequent, post-Moghrebian, reactivation. The flat-lying, transgressive Moghrebian and
Fig. 5. (a) Cross-section of the Ait Lamine-Kasbah anticlines (modified after Duffaud et al. (1962)). The listric geometry of the thrust fault is suggested as a possible alternative configuration of the fault plane dip of (b). (b) Forward kinematic model based on field geology, boreholes and seismic profiles across the SATF near Agadir. The fault plane dip is 25~. According to both cross-sections and their extreme values of the mean fault dip (25~ to 45~ the hypocentral depth of the 1960 event may range between 3.3 and 5.6 km.
COASTAL TECTONICS ACROSS SOUTH ATLAS THRUST FRONT Quaternary marine terraces with no visible bedding plane displacement across the Air Lamine and Kasbah anticlines (except along the active flexural fault) attest to the lack of recent beddingplane movements (see also Fig. 4c) and to localized recent deformation at the thrust front. The pattern of marine terrace distribution along the coast also reflects the localized deformation at the thrust front and on the Kasbah fold (see also Fig. 10, below). Further analysis of growth strata during Neogene time using restored deformation on balanced cross-sections shows a total amount of shortening of 2.3 kin, which implies a shortening rate of 0.2-0.5ram year, consistent with the 0.1-0.2ram year uplift rate of the marine terraces (Table 2) if we assume a low dip angle (25 ~) for the thrust fault at depth.
Coseismic surface faulting in the Agadir area Reappraisal of 1960 Agadir earthquake surface breaks On Monday 29 February 1960, a damaging earthquake occurred along the SATF near the coastal city of Agadir. The epicentre was located at 30.57~ 9.43~ the peak magnitude was Ms 5.9 + 0.2, and the maximum intensity Io was X (MKS), which apparently reflects a shallow mainshock depth (Fig. 1; Roth6 1962; Cherkaoui et al. 1991; Benouar 1994). The region was severely damaged, and a complex pattern of surface breaks produced by bedding plane movements, transverse ruptures on roads, rockfalls and landslides, and liquefaction features were observed and mapped after the mainshock (Duffaud et al. 1962). Although Roth6 (1962) suggested a possible link with the High Atlas fold-thrust system, the analysis of tectonic features of the seismogenic zone remained incomplete and no clear relationship with an active structure was established. We conducted detailed field investigations on the basis of the numerous observations and descriptions mentioned in the earthquake volume of Dufaud et al. (1962), and using aerial photographs and topographical maps at 1/50 000 scale, and locally (north of the city and along the coast) at 1/2000 scale. Although coseismic surface ruptures were not likely to have survived, the aim of the field campaign was to look for any correlation between geological structure and the deformation of young deposits as observed during field investigations after the 1980 E1 Asnam earthquake (King & Vita-Finzi 1981; Philip & Meghraoui 1983). To judge from the description ofcoseismic surface breaks given by Choubert & Faure-Muret
247
(1962), and the mapping of about 1 km of surface ruptures on the southern flank of the Kasbah anticline (Roth6 1962) made immediately after the mainshock (Fig. 6), we may distinguish three types of earthquake-induced features: (1) surface ruptures of tectonic character cutting roads transverse to the slope; (2) sinuous cracks parallel to slopes corresponding to landslides and cracks on road surface probably caused by compaction of loose material underneath; (3) sand blows corresponding to liquefaction phenomena observed south of Agadir near the seashore at the Souss river mouth (Fig. 1). According to the above workers, surface ruptures visible along the flexural bedding presented a minimum of 5-8 cm and a maximum of 30cm of vertical displacement across the road surface along the southern flank of the Kasbah anticline at Founti (Figs 3 and 6). Though sceptical about surface faulting, Choubert & Faure-Muret (1962) made a 'mention speciale" of the coseismic ruptures of Founti, where they observed two main alignments, the westernmost (near the military camp, Fig. 6) showing slight or no displacement, and the easternmost with a maximum displacement of 30 cm on the road.
Recent st#face faulting and cumulative displacements Recently enlarged road-cuts on the southern flank of the Kasbah anticline expose, at two sites (A and B in Fig. 6), faults affecting late Quaternary marine deposits and showing bedding plane movements.
Military camp site. At the first site, faults are visible on the same alignment along the two roads that border the military camp (now abandoned) and they perfectly coincide with the 1960 coseismic ruptures (Fig. 6; Roth6 1962). The easternmost fault plane strikes N 080, dips 75S and shows an uplifted southern block. The fault is visible because it juxtaposes terrace Q3, ~hich shows a gentle southward dip (15 ~ of brown calcarenitic units overlain by loose coastal dunes, with Mesozoic limestones. The westernmost fault affects marine terrace Q2, which is composed of a succession of coarse gravels and sandy deposits with intercalated shell units: it strikes N085, dips 30~ and shows 0.5 m of vertical displacement with an upthrown southern block (see Fig. 7a). Displacement here is recent (Holocene?) as the fault extends into the palaeosol and the recent soil above. A few metres to the south and on the same exposure,
248
M. M E G H R A O U | ET AL.
Fig. 6. Surface faulting associated with the 1960 Agadir earthquake (Choubert & Faure-Muret 1962; Roth6 1962). A refers to the main fault plane visible in Fig. 8b and B to a minor thrust fault in Fig. 7a.
an east-west trending open vertical fissure with 0.30m of vertical displacement accompanied by minor fissures also affects marine terrace deposit Q2 (see Fig. 7b). Founti-Agadir harbour site. At Founti and along the road parallel to Agadir harbour, the fault zone corresponds to a 3-5 m wide gouge zone, located between bedding planes within strongly folded, upper Cretaceous white limestones. Here again, the fault gouge coincides perfectly with the 1960 coseismic ruptures as mapped by Roth6 (1962; Fig. 6), and shown in a
photograph published in a report by Choubert & Faure-Muret (1962, see "also Figs 8a and b). The gouge zone, which can be represented by a single main fault plane striking N 080 and dipping 70~ exhibits striations that seem to indicate a normal faulting geometry (Fig. 1). The change in faultplane geometry is caused by flexural slip and implies that the compressional tectonic character is conserved along the fault; we consider the striations consistent with a thrust mechanism (Fig. 9). A comparable observation was made at E1 Asnam by Philip & Meghraoui (1983). On the other hand, marine terrace Q3 displays 4-5 m of
COASTAL TECTONICS ACROSS SOUTH ATLAS THRUST FRONT
249
Fig. 7. a ) Thrust fault affecting terrace Q2 west of Kasbah. The dip slip is about 0.10 m. (b) Minor normal faults affecting terrace Q2 and 10m away on the hanging wall of the thrust fault of (a). The vertical offset is 0.30m. vertical offset across the fault zone at the site where the 1960 coseismic rupture produced 0.30 m of vertical movement (Fig. 8b), suggesting that it represents the cumulative effect of successive coseismic movements in the past. The active fault at Founti is not a single branch, as a few tens of metres to the south, a second frontal branch displaces the Q3 fossil horizons which do not appear in the nearby footwall block, implying 15m of minimum offset. This second fault branch is not directly visible, as slope debris covers the road-cut, but a
fault scarp is indicated by the morphology of the slope. Moreover, the 1960 fault ruptures were also at the toe slope inflection. Although bedding plane movements associated with flexural movements of the fold cannot be ruled out, the fault trace and related gouge zone, unique within the strongly folded Cretaceous limestones, is taken to represent the main active fault of the Kasbah anticline. Hence, the coincidence between the 1960 surface ruptures and displacements on the marine terraces suggests that flexural slip movements of possible earthquake
Fig. 8. a ) View of 1960 coseismic ruptures across a road near Founti (from Choubert & Faure-Muret (1962)). (b) Displaced marine terrace Q3 above the main fault zone at the same location as the 1960 coseismic ruptures shown in (a).
COASTAL TECTONICS ACROSS SOUTH ATLAS THRUST FRONT
251
origin may have occurred repeatedly on the active fold-related fault.
Conclusions: uplift rates and implications for seismic hazard
Fig. 9. Striations on the hanging wall of the fault plane visible at Founti (see Fig. 1). Although the general geometry at the surface seems to indicate a normal faulting mechanism, the flexural slip faulting and associated striations result from shortening.
Elevation
The study of marine terraces across the SATF at Agadir provides some constraints on the behaviour of a seismogenic fold-related fault and its evolution during the Quaternary period. On the basis of our structural analysis of folds along the SATF, we have seen that the tectonic morphology is necessary but not sufficient to explain the evolution of an active fold-related fault. This is because the geometry and evolution of preexisting structures have a profound influence on the growth of active folds. It follows that correct interpretation of tectonically active landscapes, drainage patterns and terrace sequences calls for detailed knowledge of the local geological structure and palaeogeography. Our measurements of coastal uplift demonstrate that the Kasbah anticline is an active structure built up mainly from middle to late Quaternary time at an average rate of 0.15 mm/
Cap B0uzellou
200m-
Kasbah::
~
__~.~i
.............
150 ................................. -........................................................................................................... Assif Tamraght 100
.Arhesdis
................................................................................................
i ,
i
50-
0-
_:
"
'Anza
2.5
South Anza :.
Q2
F0unti
i ........................................
i
,,
i. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
o
,~
-o-, 5
:
.
.
.
.
.
Q3
.
.
.
.
.
.
.
.
.
-
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
AGADIR
o 7.5
10 km Distance along the shoreline
Fig. 10. Marine terrace distribution across the Ait Lamine-Kasbah anticlines. Each point corresponds to a terrace elevation with a clear bench morphology and continuity along the shoreline. Uplifted terraces near the Arhesdis-Kasbah-Founti sites (see also Figs 3 and 4a) can be correlated with incremental movements on the Kasbah fault-propagation fold.
252
M. M E G H R A O U I E T AL.
year (Fig. 10 and Table 2). It appears from fold evolution that the rate of deformation is not uniform and that the folding process may have resulted from successive tectonic pulses. This interpretation is supported by three observations: the existence of well-preserved marine terraces, the topographical difference of 100-150m between bench-terraces Ql and Q2, and the gentle folding of terrace Q1 and Q2 compared with the uplift of terraces Q3 and Q4. However, marine terraces in the hanging wall of the SATF have relatively low heights and a low rate of deformation with respect to other regions of interplate tectonics. For instance, the vertical movement of marine terraces Q2 and Q3 at the Kasbah and Arhesdis sites, with maximum up-lift rates of 0.23mm/year and 0.24ram/year respectively (Table 2), is low compared with the 0.6-1 ram/ year maximum uplift rate obtained from palaeoseismic investigations on the E1 Asnam thrust fault in the Tell Atlas of Algeria (Meghraoui & Doumaz 1996). The distribution of Quaternary marine deposits and related recent tectonic episodes can be combined with structural analysis to constrain the behaviour of the Agadir seismogenic zone. Uplifted marine terraces in the hanging block and the evidence for coseismic ruptures in 1960 indicate that the growth of the Kasbah anticline occurs on a seismogenic fault (Figs 3 and 10). In general, an earthquake size with M s = 5 , 9 , similar to the 1960 seismic event, implies a seismic moment of Mo c. 1017 (Kanamori & Anderson 1977), and taking into account a possible length of 8 km for the Kasbah fault and about 0.15 m of average coseismic displacement, we may obtain a width of 6-8 km for the seismogenic fault. The 1960 hypocentral depth will then depend on the fault dip, and a dip angle between 25 ~ and 45 ~ yields a depth 3.3-5.6km. A shallow mainshock could help to explain the severe damage in the epicentral area on both northern and southern flanks of the Ait LamineKasbah anticlines (see also Fig. 1). Episodic tectonic uplift associated with the formation of successive marine terraces may result from coseismic movements related to large or moderate-sized earthquakes (Rockwell et al. 1988; Berryman et al. 1992). An earthquake of comparable size to the 1980 E1 Asnam seismic event (Ms 7.3) would generate an average 2 m of vertical uplift, which could be recorded by the marine terraces along the Agadir shoreline at Founti. The historical earthquake at Agadir in 1731, reported to have destroyed the Portuguese Santa Cruz-Kasbah fortress (Benouar 1994), can serve as an analogue to the 1960 earthquake. Although a component of aseismic deformation
on the Kasbah anticline cannot be ruled out, the 4-5 m offset of terrace Q3 on the gouge zone and the 1960 earthquake ruptures suggest the occurrence of repeated coseismic movements in the past. Seismic hazard assessment must also take into account the fact that the Kasbah fault extends to the east as far as the Tildi tear fault and to a similar active structure at Anounfeg, equivalent to a total length of 25km for the active structure (Fig. 1). This work benefited from the assistance of the Division de la G6ologie G6n6rale, Minist6re de l'Energie et des Mines, Royaume du Maroc, and we thank Mohammed Dahmani for help and support during the field campaigns. The field-work was also partly supported by the Peri-Tethys Programme (South Atlas Front Project). Fruitful discussions with A. Faure-Muret clarified some points concerning the 1960 coseismic surface ruptures, D. Sorel (University of Orsay) provided the Alti-Plus Pretel altimeter together with relevant discussions on coastal tectonics, and A. Sposato (IRTR, Rome) provided useful comments on the marine terrace classification. We are grateful to A. E1 Maoukour (Regional Geological Survey of Agadir) and M'H. Aberkane (D~partment de G6ologie, Universit6 Mohammed V Rabat, Morocco) for their help during the field investi-gations. We thank F. Gomez (Cornell University) for his constant interest in this work and for providing the Landstat image. U-Th dating and sample analysis benefited from the assistance of J. L. Reyss (Centre des Faibles Radioactivit~s, Gif-surYvette, France). We thank I. Stewart and an anonymous reviewer for useful comments and suggestions.
References AMBROGGI, R. 1963. Etude gdologique du versant m&idional du Haut Atlas occidental et de la plaine du Souss. Notes et M6moires du Service G6ologique du Maroc, 157. AOUDIA, A. & MEGHRAOUI,M. 1995. Seismotectonics in the Tellian Atlas of Algeria, the Cavaignac (Abou E1 Hassan) earthquake of 25 August 1922. Teetonophysics, 248, 263-276. BENOUAR,D. 1994. Materials for the investigation of the seismicity of Algeria and adjacent regions during the twentieth century. Annali di Geofisica, 37. BERRYMAN, K. R., OTA, Y. & HULL, A. G. 1992. Holocene evolution of an estuary on a tectonically rising coast: the Pakarae River locality, eastern North Island, New Zealand. Sedimentary Geology, 80, 151-165. BRACI~NE, R., BELLAHCI~NE, D., BEKKOUCHE, D., MERCIER, E. & FR~ZONDE LAMOTTE, D. 1998. The thin-skinned style of the South Atlas Front in central Algeria. In: MACGREGORD. S., MOODY, R. T. J. & CLARK-LOWES,D. D. (eds), Petroleum geology of North Africa. Geological Society, London, Special Publications, 133, 391-400. CHAPPELL, J. & SHACKLETOY, N. J. 1986. Oxygen isotopes and sea level. Nature, 324, 137-140.
COASTAL T E C T O N I C S ACROSS SOUTH ATLAS T H R U S T F R O N T CHERKAOUI,T. E., MEDINA, F. & HATZFELD,D. 1991. The Agadir earthquake of February 29 1960: examination of some of the parameters. Publicaei6n Instituto Geogr~fico Nacional, Spain, Serie Monografica, 8, 133-148. CHESTER, J. S. ~ CHESTER, F. U. 1990. Fault propagation folds above thrusts with constant dip. Journal of Structural Geology, 12, 903-910. CHIARABBA, C., AMATO, A. & MEGHRAOUI, M. 1997. Tomographic images of the El Asnam fault zone and the evolution of a seismogenic thrust-related fold. Journal of Geophysical Research, 102, 24 485-24 498. CHOUBERT, G. 8r FAURE-MURET, A. 1962. Le s~isme d'Agadir, ses effets et son interpr&ation g6ologique. In: DUFFAUD, F. ROTHI~, P., DEBRACH, J., ERIMESCO,P., CHOUBERT,G. & FAURE-MURET,A. (eds.) Le s~isme d'Agadir du 29 f~vrier 1960. Notes et M6moires du Service G~ologique du Maroc, 154, 53-68. DUFFAUD, F., ROTHE, P., DEBRACH,J., ERIMESCO,P., CHOUBERT, G. & FAURE-MURET, A. 1962. Le sdisme d'Agadir du 29 f~vrier 1960. Notes et M6moires du Service G6ologique du Maroc, 154. EL GHARBAOUI, A., CHOUKRI, A., BERRADA, M., FALAKI, H. & REYSS,J. L. 1994. Datation de deux niveaux matins sur la cote du Haut Atlas atlantique fi 275 000 ans et fi 120 000 ans. Cahiers de Gdographie du Quebec, 38, 104, 241-247. EL MAAMAR, K. 1988. Etude tectonique et microtectonique de la bordure sud du Haut Atlas occidental (r~gion d'Agadir, Maroc). Th6se d'Universit6, Montpellier II. FRIZON DE LAMOTTE, D., GHANDRICHE, H. & MORETTI, I. 1990. La flexure saharienne: trace d'un chevauchement aveugle plurikilom&rique au nord du Sahara (Aur&, Alg6rie). Comptes Rendus de l'Acad~mie des Sciences, S&ie II, 310, 1527-1532. IMBRIE, J., HAYS, J. D., MARTINSON, D. G., MCINTYRE, A., MIX, A. C., MORLEY, J. J., PISIAS, N. G., PRELL, W. L. & SHACKLETON,N. J. 1984. The orbital theory of the Pleistocene climate: support from a revised chronology of the marine 6180 record. In: BERGER,A. et al. (eds) Milankovitch and Climate, Reidel, Part L Dordrecht, 269-305. KANAMORI, H. & ANDERSON, D. L. 1977. Theoretical basis of some empirical relationships in seismology. Bulletin of the Seismological Society of America, 65, 1073-1095. KAUFMAN, A., BROECKER, W. S., Ku, T. L. & THURBER, D. L. 1971. The status of U-series methods of mollusk dating. Geochimica et Cosmochimica Acta, 35, 1155-1183. KING, G. & VITA-FINZI, C. 1981. Active folding in the Algerian earthquake of October 10 1980. Nature, 292, 22-26. MEGHRAOU1, M. 1991. Blind reverse faulting system associated with the Mont Chenoua-Tipaza earthquake of 29 October 1989 (north-central Algeria). Terra Nova, 3, 84-93.
253
& DOUMAZ,F. 1996. Earthquake-induced flooding and paleoseismicity of the E1 Asnam (Algeria) fault-related fold. Journal of Geophysical Research, 101, 17 617-17 644. MERCIER, E., OUTTANI, F. & FRIZONDE LAMOTTE, D. 1997. Late evolution of fault propagation folds: principles and examples. Journal of Structural Geology, 19, 185-193. MUSTAPHI, H. 1997. dpport des donn~es de subsurface (forages et profils sismiques) et de la modOlisation d l'~tude de l'~volution du bassin du Souss du Trias d l'Actuel. Th+se de 3~me cycle, Universit+ Mohammed V Rabat, Morocco. OUTTAN1, F. 1996. Cin~matique, modOlisation et bilan dnerg~tique des plis de rampes: approche th~orique et application d deux rOgions du front sud atlasique. Th6se en Sciences, Universit6 de CergyPontoise. OUTTANI , F., ADDOUM, B., MERCIER, E., FRIZON DE LAMOTTE, D. & ANDRIEUX, J. 1995. Geometry and kinematics of the South Atlas Front, Algeria and Tunisia. Tectonophysics, 249, 233-248. PHILIP, H. & MEGHRAOUI, M. 1983. Structural analysis and interpretation of the surface deformations of the E1 Asnam earthquake, Tectonics, 2, 17-49. ROCKWELL,T. K., KELLER, E. A. & DEMBROFF, G. R. 1988. Quaternary rate of folding of the Ventura anticline, western Transverse Ranges, southern California. Geological Society of America Bulletin, 100, 850-858. ROTHE, P. 1962. Le s~isme d'Agadir et la s6ismicit~ du Maroc. In: DUFFAUD, F. ROTHI~,P., DEBRACH,J., ERIMESCO, P., CHOUBERT, G. & FAURE-MURET, A. (eds.) Le s~isme d'Agadir du 29 fdvrier 1960. Notes et m6moires du service g6ologique du Maroc, 154, 7-30. SAINTBEZAR,B., FRIZONDE LAMOTTE,D., MOREL, J-L. & MERCIER, E. 1998. Kinematics of large scale tip line folds from the High Atlas thrust belt (Morocco). Journal of Structural Geology, in press. STEARNS, C. E. & THURBER, D. L. 1965. 23~ dates of late Pleistocene marine fossils from Mediterranean and Moroccan littoral. Quaternaria, 7, 29-41. STEIN, R. & KING, G. 1984. Seismic potential revealed by surface folding: 1983 Coalinga, California, earthquake. Science, 224, 867-872. - - & YEATS, R. 1989. Hidden earthquakes. Scientific American, 260, 48-57. WEIJERMARS, R. 1987. A revision of the EurasianAfrican plate boundary in the western Mediterranean. Geologisches Rundschau, 76, 667-676. WEISROCK, A. 1980. G6omorphologie et pal6oenvironnements de l'Atlas atlantique (Maroc). Th~se de Doctorat d'l~tat, Universit6 de Paris I. YEATS, R. S. 1986. Active faults related to folding. In: WALLACE, R. E. (ed.) Active Tectonics. National Academy Press, Washington, DC, 63-79.
Quaternary neotectonism and intra-plate volcanism: the Coorong to Mount Gambier Coastal Plain, Southeastern Australia: a review C. V. M U R R A Y - W A L L A C E
l, A. P. B E L P E R I O 2 & J. H. C A N N 3
1 School of Geosciences, University of Wollongong, N.S.W., 2522, Australia (e-mail:
[email protected]) 2 Minotaur Gold, la Gladstone Street, Fullarton, S.A., 5063, Australia 3 School of Engineering (Applied Geology), University o f South Australia, The Levels Campus, S.A., 5095, Australia Abstract: The Coorong to Mount Gambler Coastal Plain in southeastern South Australia
preserves a lengthy record of Quaternary temperate carbonate sedimentation in the form of high wave energy barrier shoreline deposits and associated back-barrier lagoon facies. The barriers occur sub-parallel to the modern coastline and to each other and increase in age landwards. The age of the barriers is now generally well established within a framework of radiocarbon, thermoluminescence, uranium-series disequilibrium, amino acid racemization dating and the position of the Brunhes-Matuyama boundary between the East and West Naracoorte Ranges. This framework is in accord with the deep-sea 6180 record of global ice volume change and the inferred timing of sea-level highstands for this interval. Individual barriers such as the last interglacial Woakwine Range (oxygen-isotope substage 5e) may be traced laterally for distances up to 300 km and record a history of neotectonism in the form of regionally varying epeirogenic uplift associated with Quaternary volcanism. Covering c. 24 000 km 2 (just under a fifth the area of England) in the far southeast of South Australia and extending into western Victoria, the Coorong to Mount Gambier Coastal Plain represents a unique feature in the Australian coastal landscape as well as in global terms (Fig. 1). Across an otherwise broad, featureless coastal plain with a gradient of approximately 0.5% is preserved a series of calcareous to siliceous barrier shoreline deposits, which formed during successive Quaternary interglacials and interstadials (Sprigg 1952; Huntley et al. 1993; Figs 1 and 2). The relict coastal barriers are typically up to 30 m above the general level of the coastal plain and up to 10 km apart, occur sub-parallel to the modern coastline and to each other and increase in age landwards. They have been referred to as 'dune ranges' by the local inhabitants, for in places the features appear imposing in an otherwise flat landscape. The barriers coalesce to the far north and south of the coastal plain away from the centre of inferred maximum uplift. Individual coastal barriers such as the last interglacial Woakwine Range (oxygen-isotope substage 5e) may be traced laterally up to 300 km parallel to the modern coastline (Murray-Wallace et al. 1996). These landforms are essentially coastal barriers of temperate bioclastic carbonate sand with associated subtidal, intertidal and aeolian dune sand facies of similar composition, and sandy to muddy lagoonal facies on their landward side. Pedogenesis of the Quaternary carbonate dunes and associated facies has led to
the formation of extensive surficial calcretes, which blanket the regional landscape and have effectively preserved the original morphology of the coastal barriers (Fig. 3). An older series of coastal barriers occurs further inland (Fig. 1). They are more siliceous and only poorly preserved (Belperio & Bluck 1990), and are generally ascribed to Early Pleistocene time. The age of the younger barriers is now generally well established within a framework of radiocarbon, thermoluminescence, uranium-series disequilibrium, amino acid racemization dating and magnetic reversal stratigraphy (Fig. 2). The coastal barrier succession younger than 780 ka BP has been examined in the greatest detail. Thirteen barriers younger than the Brunhes-Matuyama geomagnetic polarity reversal (780 ka BP) have been identified on the basis of morphostratigraphic and dating evidence (Sprigg 1952; Huntley et al. 1993, 1994). Some of the barriers represent composite features formed during more than one sea-level highstand, with older calcarenite cores acting as anchors on which younger sediment has accumulated (Schwebel 1984). This framework agrees with the deep-sea 5180 record of global ice volume change (Martinson et al. 1987) and the inferred timing of Quaternary sealevel highstands. In this work, the general geology of the coastal plain succession and the geochronological results leading to a model of Quaternary neotectonism for the region are reviewed. We go on to suggest
MURRAY-WALLACE, C. V., BELPERIO, A. P. & CANN, J. H. 1998. Quaternary neotectonism and intra-plate volcanism: the Coorong to Mount Gambier Coastal Plain, southeastern Australia: a review. In: STEWART,I. S. & VITA-FINZl, C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 255-267.
256
C. V. M U R R A Y - W A L L A C E
E T AL.
Fig. 1. The Coorong to Mount Gambier Coastal Plain showing location of Pleistocene coastal barriers and palaeo-shoreline altitudes in metres (circles with dots) above present sea level.
QUATERNARY NEOTECTONISM, SOUTHEASTERN AUSTRALIA
257
Fig. 2. Schematic section across the Coorong to Mount Gambier Coastal Plain. The shore normal transect through the dune range system runs from the present coastline near Robe (37~ 139~ 90kin inland to Naracoorte (36~ 140~ also shown are the thermoluminescence ages of Huntley et al. (1993, 1994) and correlation with the 6180 record of global ice volume change. Figure modified after Schwebel (1984) and Huntley et al. (1993).
that spatial separation of the successive Quaternary shorelines is due to steady regional epeirogenic uplift, the product of Quaternary volcanism and associated igneous intrusions. In addition, as the Coorong to Mount Gambier Coastal Plain represents one of the few places in the world where a lengthy, on-land record of Quaternary sea-level highstands for at least the
last 800 000 years may be delineated on the basis of unambiguous morphostratigraphic evidence, we take the opportunity to evaluate critically inferences from the 5180 record about eustatic sea level for numerous interglacials and interstadials older than the penultimate interglacial (oxygen-isotope stage 7; c. 220kaBP), and thus the validity of the Milankovitch hypothesis.
258
C . V . MURRAY-WALLACE E T A L . exceeded that of the (Gostin et al. 1988).
Great
Barrier
Reef
Previous investigations: benchmark studies
Fig. 3. The penultimate interglacial Burleigh Range (c. 237 ka BP) 2 km east of Mount Schank. The original morphology of the dune range is preserved by a thick calcrete carapace.
The Coorong to Mount Gambier Coastal Plain: Australian geomorphological context Set within a tectonic plate and remote from plate boundaries, Australia is characterized by low continental gradients, antiquity of landscapes and relative tectonic stability, enhanced by the widespread occurrence of Precambrian cratons. Only minor seismic activity is experienced, although earthquakes may occur. Australia experienced only limited Quaternary volcanism, apart from intraplate volcanism in southeastern Australia and northern Queensland. Few coastal areas in Australia display the magnitude of uplift documented for plate boundaries (Ota 1994), but in places remote from stable cratons, such as the Coorong to Mount Gambier Coastal Plain, evidence for neotectonic activity can be confidently distinguished from smaller-scale hydro-isostatic feedback effects (Murray-Wallace & Belperio 1991). Australia also experienced limited ice cover during Quaternary glaciations, mainly confined to the Snowy Mountains and the highlands in Tasmania (Colhoun et al. 1994). Accordingly, glacio-isostasy does not figure in the coastal evolution of this continent. Extensive endorheic drainage to Lake Eyre, the lowest point on the continent (15m below present sea level), and generally low topographical relief at a continental scale, combined with extensive aridity, ensure that few perennial streams debouch to the Southern Ocean. In southern Australia, including the Coorong to Mount Gambier Coastal Plain, this has enhanced temperate carbonate sedimentation to the extent that carbonate productivity in this region during Quaternary time has far
Numerous studies have examined the Quaternary history of the Coorong to Mount Gambier Coastal Plain. In the following discussion, the most significant works are briefly summarized as a prelude to reviewing the current state of knowledge. The first documented observations were reported by Reverend Julian Edmund Woods in 1862, resulting from his missionary travels (he subsequently published under the name Tennison-Woods). Woods noted the pronounced lateral extent of the dune ranges, that each range corresponded to a former coastline and that the flats on the leeward side of each barrier represent former estuaries. He also attributed the origin of the raised beaches to continuing uplift of the coastal plain. At the time Woods published his work, little was known about the nature of Quaternary sea-level change, and accordingly, processes associated with sealevel change could not be distinguished from neotectonic uplift. Other benchmark studies that significantly improved the understanding of the Quaternary history of the coastal plain include the work of Hossfeld (1950), Sprigg (1952, 1979), Cook et al. (1977), Schwebel (1984) and Huntley et al. (1993, 1994). Hossfeld (1950) presented a comprehensive overview of the Late Cenozoic history of the south-east of South Australia. He identified 18 distinct ranges across the coastal plain and noted that the complete succession is not encountered in any single traverse normal to the ranges because of variations in their length and in places a thick cover of drift sands. Sprigg (1952) published the now classic work on the geology of the south-east province and arguably presented the most detailed map of the dune ranges for the southern part of the coastal plain. He assigned ages to each of the dune ranges within the framework of the Milankovitch hypothesis of global climate and sea-level change. Cook et al. (1977) described the late Cenozoic sedimentary successions in the southeast of South Australia and reported the results of a drilling transect from Naracoorte to Robe. They noted that the sediments of the East Naracoorte Range are reversely polarized, whereas the Pleistocene succession from the West Naracoorte Range to the modern coastline shows normal magnetic polarity. The magnetostratigraphy of the coastal plain was subsequently reported in greater detail by Idnurm & Cook (1980).
QUATERNARY NEOTECTONISM, SOUTHEASTERN AUSTRALIA Sprigg (1979) reappraised the Quaternary stratigraphy and geomorphology of the coastal plain, and concluded that two associations of dune ranges may be identified. The siliceous inland association and the calcareous nearcoastal association he termed the 'Wimmera' and 'Bridgewater' Catenas, respectively. Sprigg suggested that the Bridgewater Catena, located closer to the present coastline and at a lower altitude, represented beaches of glacial ages, whereas the Wimmera Catena, situated inland and at about 100m above the level of the Bridgewater Catena, represented interglacial dune range complexes. This represented a surprising and fundamental departure from his original ideas (Sprigg 1952), which are now regarded as a more correct interpretation of the relative ages of the dune ranges. Sprigg did not address the issue of the preservation potential of the dune ranges closest to the sea, or that they would have been submerged during successive interglacial marine transgressions were his subsequent interpretation correct. Similarly, he did not consider the apparently anomalous position of sea level in the present interglacial in relation to the position of the alleged glacial age dunes. Schwebel (1984) summarized aspects of the Quaternary stratigraphy for part of the coastal plain near Robe. He identified several sedimentary facies genetically related to each of the emergent barrier shorelines and reported results of uranium-series disequilibrium dating for estuarine shells (125+20kaBP) and aragonitic mud (100+30kaB1,) from the back-barrier lagoon facies of the last interglacial Woakwine Range. Schwebel's findings supported the general evolutionary framework for the coastal plain as initially outlined by Sprigg (1952). Belperio & Cann (1990) assigned ages to the dune ranges based on known chronological benchmarks and Milankovitch-oxygen-isotope cyclicity. Huntley et al. (1993, 1994) presented a comprehensive summary of their thermoluminescence dating of the barrier dune facies across the coastal plain, from the modern coastline to the East Naracoorte Range. Their geochronological results corroborated inferences about the ages of many of the dune ranges previously assigned on the basis of the Milankovitch hypothesis.
Geochronology The thermoluminescence studies of Huntley et al. (1993, 1994) have provided an important framework for evaluating earlier attempts to assign ages to the dune ranges based on the Milankovitch orbital-forcing model (e.g. Sprigg 1952;
259
Schwebel 1984; Belperio & Cann 1990). Additional thermoluminescence dates for the Burleigh and Caveton Ranges have been reported by Murray-Wallace et al. (1996). Before the work of Huntley et al. the following results had been obtained from geochronological studies of the coastal plain succession: (1) the BrunhesMatuyama geomagnetic polarity reversal had been identified between the East and West Naracoorte Ranges (i.e. East Naracoorte Range showing reversed polarity and therefore and age >780kaBP: Idnurm & Cook 1980); (2) a last interglacial age had been determined for the Woakwine Range near Robe, on the basis of uranium-series disequilibrium dating of aragonitic muds (100 i 30 ka Be) and marine molluscs (125 9 20 ka BP; Schwebel 1984) and amino acid racemization dating of fossil molluscs from the lee of the Woakwine Range (Von der Borch et al. 1980; Murray-Wallace et al. 1991). On the basis of the age of these two barrier shorelines and assumptions about uplift rates, interglacial and interstadial ages were assigned to the barrier shorelines between these two ranges in the context of models of glacio-eustatic sea-level change. Thus, the studies by Huntley et al. (1993, 1994) have provided vital information demonstrating the increasing ages of the dune ranges between the Woakwine and Naracoorte Ranges and confirming that the barrier shorelines formed during successive interglacial highstands. The thermoluminescence chronology of Huntley et al. (1993, 1994) is based on the dating of quartz sand from the aeolian facies of the barriers. Although some of their thermoluminescence dates do not relate precisely to palaeo-sea level (i.e. they date not 'fixed' but 'relational' sea-level indicators such as coastal dunes in the sense of Chappell (1987)), it is unlikely that the uncertainty associated with this exceeds the analytical precision of the individual thermoluminescence measurements. Revised ages for the oldest part of the Robe Range (Robe III in the stratigraphic nomenclature of Schwebel 1984)) and for the Naracoorte Range were reported by Huntley et al. (1994). A summary of the currently accepted thermoluminescence ages for the dune ranges for the coastal plain is summarized in Table 1 and Fig. 2. Samples from the Woakwine Range yielded thermoluminescence ages of 132 + 6, 132 + 9 and 118 + 4 ka B~',consistent with the currently accepted duration of the last interglacial maximum (oxygen-isotope substage 5e; c. 135-115 ka BP; Chert et al. 1991; Lambeck & Nakada 1992; Zhu et al. 1993; Stirling et al. 1995). Huntley et al. (1994) attributed this variation in the dates to analytical precision, but as the youngest of the three dates falls just outside
260
C. V. MURRAY-WALLACE E T A L .
Table 1. Summary of thermoluminescence (TL) dates (ka) from the Coorong-Mount Gambier Coastal Plain Dune range
TL age (ka)
~5180 stage
Reference
Robe I Robe III Woakwine I
0 116 4- 6 132 + 6 118 + 4 132 + 9 230 + 11 237 + 16 258 + 25 320 + 22 342 4- 32 414 + 29 456 + 37 585 4- 40 800 + 100 720-t-70
modern 5a or 5c 5e 5e 5e 7a or 7c 7e 7e 9 9a 9c 1lc 13.13 <780ka 21c or 25
Huntley et al. (1994) Huntley et al. (1994) Huntley et al. (1993) Huntley et al. (1994) Huntley et al. (1994) Huntley et al. (1993) Murray-Wallace et al. (1996) Huntley et al. (1993) Murray-Wallace et al. (1996) Huntley et al. (1993) Huntley et al. (1993) Huntley et al. (1993) Huntley et al. (1993) Huntley et al. (1994) Huntley et al. (1993)
Woakwine II-III Burleigh Reedy Creek Caveton West Avenue East Avenue Baker Harper West Naracoorte East Naracoorte
the range of the other two dates, there is a possibility that this age relates to a later phase of the last interglacial maximum. The extent of racemization as determined for several amino acids in the fossil molluscs Anadara trapezia and Katelysia rhytiphora from the back-barrier lagoon facies of the Woakwine Range yielded results consistent with a last interglacial age (Substage 5e; Murray-Wallace et al. 1996). The fossil molluscs were collected from widely separated sites along the length of the dune range over a distance of 60 kin. The amino acid data support previous assumptions about the uniformity of age along the strike of individual barriers.
Pleistocene depositional events and coastal facies Fossiliferous, Lower Pleistocene marine sediments (Coomandook Formation) underlie much of the coastal plain and can be equated with the dune barrier complexes furthest inland (East Naracoorte and Coonalpyn Ranges). An Early Pleistocene age is assigned to these barriers on the basis of the presence of fossils of the pelagic marine gastropod Hartungia (dennanti) chavani, a diagnostic indicator species (Ludbrook 1983). Palaeomagnetic studies (Idnurm & Cook 1980) revealed that the coastal barriers east of Naracoorte are of reversed polarity, indicating a minimum age of 780 ka BP. The East Naracoorte Range has been correlated with oxygen-isotope stage 25 sea-level highstand with an inferred age of 870kaBP (Huntley et al. 1993). The West Naracoorte Range extends as a distinct linear
shoreline for over 100km before it recurves in a tombolo-like arc and loses its distinct morphology behind the Mount Burr volcanic complex. As the West Naracoorte Range is of normal magnetic polarity (i.e. <780 ka BP), it has been correlated with oxygen-isotope stage 19 highstand (Huntley et al. 1993). The successive dune ranges in a seaward direction (Stewart, Woolumbool, Baker, Ardune and East Avenue) show a progressive decrease in altitude of the barriers and associated back-barrier flats. The last interglacial shoreline of the coastal plain is represented by the Woakwine Range. Morphologically, the Woakwine Range complex is largely a single linear form about 30m high, rising up to 60 m near Cape Jaffa where it bifurcates to form an easterly trending bar or 'hook'. The facies architecture of the Woakwine Range is revealed by good exposures in the walls of deep drainage cuttings. In McCourt's Cutting near Beachport, littoral and back-barrier facies indicate a major depositional event with a sea level that regressed from 10 to 8 m (relative to present sea level, and uncorrected for uplift). In Drain L, east of Robe, three successive depositional events are recognizable, on the basis of lithostratigraphic evidence, at between 5 and 10m above present sea level (Sprigg 1952; Schwebel 1984) but only two events have been resolved by thermoluminescence dating. North of Kingston, the Woakwine Range equivalent backs the Holocene Coorong Lagoon and may be traced parallel to Younghusband Peninsula as far as Lake Alexandrina. The elevation of the shoreline facies associated with this barrier decreases to 3 m near Salt Creek and to just 1 m above present sea level before it disappears beneath Lake Alexandrina near the mouth of the Murray River.
QUATERNARY NEOTECTONISM, SOUTHEASTERN AUSTRALIA Robe Range represents the youngest emergent Late Pleistocene dune range of the coastal plain and has been assigned an interstadial age (oxygen-isotope substage 5c; 116 :t: 6 ka BP), on the basis of thermoluminescence dating (Huntley et al. 1994) and amino acid racemization (Von der Borch et al. 1980). The crest of the Robe Range structure crops out at the present shoreline giving rise to rocky coastal cliffs and shore platforms (Fig. 4). To judge from the accordant surface of erosion on offshore residual sea stacks and shore platforms, close to 1 km of coastal erosion of Robe Range is likely to have occurred since the culmination of the post-glacial marine transgression 7000 years ago. The toe of the dune range, which rests unconformably on the Miocene Gambier Limestone, is some 10-15 m below present sea level. Transgressive dunes of the Holocene barrier drape the erosional remnants of the Late Pleistocene dune range and provide a model for interpreting the stacked arrangement of depositional units within the older barriers. Each barrier shoreline from Robe to Naracoorte, although representing a different time interval and varying degrees of preservation and exposure, displays a common suite of sediment facies. Analogous features and evolutionary pattern are evident for the modern, Holocene barrier, Younghusband Peninsula-Coorong Lagoon barrier complex. The principal facies include the following: (1) Shallow subtidal facies consisting of seaward-dipping, coarse carbonate sands with abundant broken shells of marine gastropods and bivalves together with Foraminifera and coralline algae.
26I
(2) Littoral sediments, well-sorted, mediumto coarse-grained, quartzose and skeletal carbonate sand, occurring as shoestring sand bodies subparallel to the present coast. The beach deposits, up to 3 m thick, extend into the core of the ranges as distinctive bidirectional trough cross-bedded strata. (3) Transgressive dune or aeolianite facies form the largest component of the dune ranges. They comprise weakly cemented, fine- to medium-grained skeletal sand, of well-sorted and rounded molluscan, foraminiferal and algal grains. Reactivation surfaces, marked by weakly developed palaeosols and rhizolith concentrations, are common. They separate the cosets of very large-scale cross-beds which dip consistently landward at 3-30 ~. (4) Poorly sorted estuarine or lagoon facies are preserved in back-barrier settings of the interdune corridors. These sediments are typically very shelly and muddy calcarenite with numerous gastropods and large articulated intertidal to subtidal molluscs. The intertidal fauna provide a particularly useful datum for determining shoreline elevation and reconstructing former sea levels. (5) Lacustrine sediments were deposited in interdune corridors, not necessarily coeval with the formation of adjacent barrier aeolian facies. They consist of dense, white calcitic and dolomitic mudstone with interbedded greenish clay and clayey quartz sand. The calcilutite is amorphous to indistinctly bedded, often pelletal, and contains remains of a variety of fresh- to brackish-water molluscs, Foraminifera, ostracods and charophytes. The lake deposits that overlie marine lagoonal facies reflect closure of the barriers to the open ocean. Preservation and exposure of the complete spectrum of facies is greatest for the youngest ranges (namely, Holocene and last interglacial), but elements of each have been recognized in most of the ranges between Robe and Naracoorte.
Holocene coastal sedimentation
Fig. 4. Cross-section of an erosional remnant of the Late Pleistocene (oxygen-isotope substage 5c; c. 105 ka BP) Robe Range at Cape Northumberland, south of Mount Gambier. Landward migrating aeolian crossbeds and major erosional truncation surfaces are clearly visible.
Sea level rose rapidly after the last glacial maximum, transgressing the continental shelf and reaching the present coastline c. 7000 years ago (Belperio 1995). The Holocene barrier Younghusband Peninsula (180 km long) and Coorong Lagoon formed as a result of this sea-level rise (Fig. 5). Much of the Holocene barrier complex that makes up Younghusband Peninsula consists of former transgressive dune sheets now stabilized by vegetation. Palaeosol horizons
262
C. V. MURRAY-WALLACE E T AL. corridor between the Late Pleistocene Robe and Woakwine Ranges represented an open seaway during the early Holocene, after the post-glacial marine transgression flooded the depression between these two calcreted dune ranges. Extensive deposition of intertidal shell beds reflects an open seaway within this corridor and record the subsequent change to more brackish lagoonal sedimentation (Cann et al. 1991; Fig. 6).
Quaternary volcanism Fig. 5. Oblique aerial view taken in May 1996 looking southeast along the Holocene coastal barrier, Younghusband Peninsula and the back-barrier Coorong Lagoon. Landward migrating transgressive aeolian dunes are visible in the middle distance. This setting represents a modern analogue for the Quaternary barrier shoreline succession of the Coorong to Mount Gambier Coastal Plain.
exposed in present-day dune blowouts indicate periods of former dune stability, vegetation and soil development. Aboriginal remains, including middens composed largely of the shoreface cockle Plebidonax deltoides, are commonly associated with the palaeosols (Bourman & Murray-Wallace 1991). Further south, near Robe, the general pattern of Holocene coastal evolution is slightly different in that Holocene aeolian sands are draped over the Late Pleistocene Robe Range owing to the partial emergence of this feature, probably because of neotectonic uplift. Accordingly, the
Fig. 6. A transported assemblage of disarticulated bivalve molluscs from the Holocene infill between the calcreted Late Pleistocene Robe and Woakwine Ranges.
Quaternary volcanism on the Coorong to Mount Gambier Coastal Plain relates to the northward movement of the Australian Plate over a mantle hot region and evidence for genetically related volcanism, represented by the Portland Volcanic Province, may be traced into Victoria (Joyce 1975; Belperio 1995; Fig. 1). Two distinct associations of Quaternary volcanism have been identified (Sheard 1990, 1995). The Early Pleistocene Mount Burr volcanic province includes 15 major volcanic centres (Sheard 1990). The eruptive centres with lava flows, composite domes, cones and maars occur northwest of Mount Gambier and pre-date the earliest Pleistocene shoreline deposits (Naracoorte Range; c. 800 ka BP) as evidenced by tombolo-like recurvature of shoreline bodies in the lee of what would have been a volcanic archipelago. In places, eruptive centres of the Mount Burr volcanic group exhibit erosional modification by Pleistocene high sea levels and are in part mantled by shoreline sands. The Holocene volcanoes Mount Gambler and Mount Schank are constructional features produced by basaltic phreatomagmatic explosions. These two eruptive complexes are the youngest of a province of several hundred centres distributed over a very large area of southeastern Australia. Some of the volcanic conduits across the coastal plain have exploited structural weaknesses such as joint sets as well as karst caverns within the Tertiary Gambier Limestone. Several attempts have been made to determine the age of the volcanism at Mount Gambier and Mount Schank. On the basis of radiocarbon dating of charcoal associated with tuff from several sites around Mount Gambier, Blackburn et al. (I 982) concluded that significant volcanism at Mount Gambier occurred between 4.3 and 4kaBP. Barton & McElhinny (1980) concluded from magnetic remanence and susceptibility measurements that volcanic episodes in Valley Lake, part of the Mount Gambler volcanic maar complex, had ceased before the onset of organic mud deposition some 6000-5000 years
QUATERNARY NEOTECTONISM, SOUTHEASTERN AUSTRALIA
263
ago and that the episodes are therefore older than reported by Blackburn et al. (1982). Radiocarbon dating of inorganic and organic carbon fractions of sediments from the Blue Lake led Leaney et al. (1995) to question the midHolocene age previously assigned to the volcanism at Mount Gambier. They suggested that Mount Gambier is considerably older than previously indicated on the basis of dating ejecta from the volcano (Blackburn 1966; Blackburn et al. 1982), and reported a minimum age of 28 ka BP for it. Despite their careful work, the possibility remains that Leaney et al. (1995) underestimated the amount of contamination by detrital and authigenic calcium carbonate, depleted in 14C, from external sources such as the Miocene Gambier Limestone, which thus yielded an apparently older radiocarbon age. Robertson et al. (1996) derived an age of 4.2 4- 0.5 ka BP based on thermoluminescence dating of baked tuff underlying a lava flow at Valley Lake, and noted the possibility of an earlier event about 7000 years ago. Smith & Prescott (1987) obtained an age of 4.93+ 5.4 ka BP for the volcanism at Mount Schank by thermoluminescence dating of quartz sand from a raised beach overlain by a lava flow. Calculation of a z-score for these data reveals that the thermoluminescence ages for the volcanism at Mount Gambier and Schank are not significantly different (z = 0.99) and collectively indicate a mid-Holocene age for the volcanism. The Mount Burr Group of volcanoes are primarily of alkaline basalt affinity. Mantlederived xenoliths such as lherzolite inclusions in lava flows or as volcanic bombs and lapilli are also present (Sheard 1995). Non-igneous country rock fragments such as the Miocene Gambier Limestone are also present in the ejecta. The basaltic lithologies of the younger group are nepheline hawaiite for Mount Gambier and K-rich nepheline hawaiite or nepheline trachyandesite for Mount Schank (Sheard 1995). Collectively, these magmas are considered to have been derived by partial melting of basanite parent materials in deep crustal or upper mantle situations (Sheard 1995).
cally related episodes of volcanism may be traced well into Victoria (Joyce 1975). The last interglacial shoreline (oxygen-isotope substage 5e) represents a particularly useful datum for mapping spatial variation in the amount of neotectonic uplift, as sufficient time has elapsed since this event for uplift to be detected even for coastlines that have experienced modest uplift. Deposits of this age also crop out extensively along the Australian coastline, have not been denuded to the degree of older interglacial shoreline deposits, and provide a relatively complete picture of Late Quaternary neotectonism. Murray-Wallace & Belperio (1991) documented an Australian datum of 2 m above present sea level (APSL) as a eustatic sea level during the last interglacial along stable crustal regions of Australia. This local datum has been used to quantify differential coastal neotectonism on a regional scale around the Australian continent (MurrayWallace & Belperio 1991; Bryant 1992). The last interglacial shoreline can also be used to document higher-resolution neotectonics, for instance on the Coorong to Mount Gambier Coastal Plain. The height of the back-barrier lagoon facies of the Woakwine Range rises progressively in a south-southeasterly direction from 1 m APSL at Lake Alexandrina to 3 m APSL at Salt Creek, 8 m APSL at Robe and 18 m APSL at the southern extremity of this dune range, closer to the Pleistocene and Holocene centres of volcanism (Figs 1 and 7). Similar systematic variations in palaeo-shoreline altitude are evident for the older Pleistocene shorelines (Sprigg 1952; Fig. 7). Local datum levels for stage 7 and stage 9 palaeo-sea levels have been documented by Belperio (1995). In addition to gross rates of neotectonic upwarp, the chronological framework allows the calculation of site-specific uplift rates as well as
Neotectonics
14.1
50, ~3o. 20-
WoakwineRange
10
Regional epeirogenic uplift of the Coorong to Mount Gambier Coastal Plain is thought to be related to doming in response to Quaternary volcanism and associated crustal emplacement of magma chambers (Sheard 1990). The volcanism is in the form of a hot region rather than a single mantle hotspot, and evidence for geneti-
0
(
S
,
,
9
20
40
60
~ .
80
,
,
,
100
120
140
,
160
Fig. 7. Height above present sea level (APSL) of intertidal back-barrier estuarine lagoon facies of the last interglacial Woakwine Range (oxygen-isotope substage 5e), Reedy Creek-Burleigh Range (substage 7e) and West Avenue-Caveton Range (Stage 9).
264
C. V. MURRAY-WALLACE E T AL.
their variations in time and space. Near Robe, the lagoon facies of the Woakwine Range occurs 8 m APSL, indicating an average uplift rate of 50 mm per 1000 years for this portion of the coastal plain. Nearby inland, the Reedy Creek Range (oxygen-isotope stage 7 at 18 + 1 m APSL) and West Avenue Range (oxygen-isotope stage 9 at 24 + 1 m APSL) indicate slightly higher, longerterm rates of uplift of 80mm per 1000 years. However, the uncertainties inherent in measuring palaeo-shoreline altitude and inferring height of sea level at the time of deposition of the older barriers could mean that the longer-term uplift rate in the Robe region is not significantly different from that calculated for the last interglacial Woakwine Range in the Robe area. Thus, an uplift rate of 70ram per 1000 years, based on all these data as previously determined by Belperio & Cann (1990), is considered to be representative for a shore-normal transect across the coastal plain from Robe. At the southernmost extremity of the Woakwine Range near Mount Gambier, shoreline facies occur up to 18 m APSL, indicating 16 m of uplift since the last interglacial maximum and an average uplift rate of 130mm per 1000 years. This calculation is based on a thermoluminescence age of 132 4- 9 ka BP for the last interglacial Woakwine Range (Huntley et al. 1993). The altitude of the Woakwine Range shell beds near Mount Gambler represents the highest occurrence of last interglacial coastal strata on the Australian mainland. Longer-term uplift rates for the nearby Burleigh and Caveton Ranges are similar, 140 and 120mm per 1000 years, respectively (Murray-Wallace et al. 1996). Thus, the uplift rates at least since late Mid-Pleistocene time have always been greater near Mount Gambier and Mount Schank than near Robe,
3O
2o
~
|
120
24Q
3120
Age (ka)
Fig. 8. Uplift curves (lines of best fit) for Mount Gambier and Robe, South Australia, from the altitude of dated palaeo-shoreline indicators. A local datum of 2 m APSL for the last interglacial sea surface (MurrayWallace & Belperio 1991) has been adopted in calculation of uplift rates,
although the crustal emplacement of magma chambers associated with mid-Holocene volcanism may have enhanced the amount of uplift in the latter area (Fig. 8). In a global context, uplift of this magnitude is considered to rank as slow to moderate (Bowden & Colhoun 1984). In summary, the Coorong to Mount Gambier Coastal Plain has undergone regional epeirogenic uplift at a rate sufficient for each interglacial of the last 800 000 years to be represented in the form of a prominent, laterally persistent coastal barrier. Intraplate Quaternary volcanism and associated emplacement of magma chambers, as well as dislocation along the Kanawinka Fault (Figs 1 and 9) near the Naracoorte Range, appear to represent the main mechanism for the regional uplift. In a global context, other neotectonic mechanisms responsible for uplift over the Quaternary time scales relevant to this review include hydroisostasy, glacio-isostasy, erosional unloading and isostatic compensation, and crustal movements at plate boundaries. It is unlikely, however, that any of these variables explain the magnitude of uplift for the Coorong to Mount Gambier Coastal Plain. The area is located in an intraplate setting remote from plate boundaries and was not subject to ice cover during Quaternary glaciations. Similarly, the low relief of southern Australia precludes uplift associated with erosional unloading and crustal isostatic compensation. Hydroisostatic readjustments on the coastal plain are unlikely to exceed 0.5m by analogy with other open ocean coastal settings in South Australia that are located closest to the continental shelf break (Belperio 1995). Further evidence for genetic association between uplift of the coastal plain and Quaternary volcanism was reported by Cook et al. (1977), who presented a subcrop map of the altitude, relative to present sea level, of the upper surface of the Miocene Gambier Limestone, which acts as basement to the Quaternary succession (Fig. 9). The map reveals a discrete, elongate dome that trends NW-SE and is coincident with the known distribution of volcanic centres (Fig. 9). The dome represents the highest point of the upper surface of the Gambler Limestone beneath the coastal plain, apart from its occurrence on the upthrown block of the Kanawinka Fault west of Naracoorte (Fig. 9). When traced in a north-northwesterly direction from Millicent (Fig. 9), the gradient of the upper surface of the Gambler Limestone in subcrop is similar to that noted for the palaeoshoreline indicator facies of the last interglacial Woakwine Range. Collectively, these observations suggest that the primary uplift mechanism relates to Quaternary volcanism.
QUATERNARY NEOTECTONISM, SOUTHEASTERN AUSTRALIA
265
Fig. 9. Subcrop map of the upper surface of the Miocene Gambier Limestone in the southeast of South Australia, modified after Cook et al. (1977). The distribution of volcanic centres is after Sheard (1990). A well-defined dome is evident near the known centres of Quaternary volcanism suggestive of a genetic relationship in uplift of the coastal plain and volcanism.
Conclusions (1) The Coorong to Mount Gambier Coastal Plain preserves a remarkable record of barrier shoreline and associated back-barrier lagoon deposition spanning at least the last 800000 years. Older, siliceous barrier shoreline deposits occur further inland but their age has yet to be accurately resolved. (2) Significant attributes of the coastal plain include: (a) the extensive area it covers (c. 24 000 km 2) and lateral persistence of individual barrier shorelines, that extend uninterrupted for up to 300 km sub-parallel to the modern coastline; (b) the number of interglacial and interstadial barriers preserved across the coastal plain; (c) the reduced influence within the region of processes which frustrate accurate determination
of former eustatic sea levels, such as isostatic processes; (d) the wide range of materials from the coastal successions suitable for analysis by a variety of Quaternary dating methods. (3) A coast-parallel transect of the coastal plain reveals significant differences in uplift rate which relate to proximity to the centres of volcanism. Across the coastal plain from Robe to Naracoorte the rate of uplift has been calculated at 70 mm per 1000 years. Near Mount Gambier a longer-term uplift rate of 130ram per 1000 years has been determined. Closer to the mouth of the Murray River at the northern extremity of the coastal plain, the dune ranges coalesce, indicating minimal uplift and possibly subsidence in this setting. (4) The region provides an opportunity to examine lengthy records of Quaternary sea-level
266
C. V. M U R R A Y - W A L L A C E
change and evaluate the predictive capacity of the marine ~80 record in determining the altitude of interglacial and interstadial sea-level highstands older than stage 7 (i.e. for intervals of time less commonly represented by coastal deposits around the world, because of erosion during sea-level lowstands). This research was financially supported by the Australian Research Council and the Quaternary Environments Research Centre of the University of Wollongong. We thank D. Bowen and C. Pain, and an anonymous reviewer for their constructive reviews of this work. The figures were prepared by P. Johnson and P. WiUiamson.
References BARTON, C. E. & MCELHINNY, M. W. 1980. Ages and ashes in lake floor sediment cores from Valley Lake, Mt Gambier, South Australia. Royal Society o f South Australia, Transactions, 104, 161-165. BELPERIO,A. P. 1995. The Quaternary. In: DREXEL, J. F. & PREISS, W. V. (eds) The Geology of South Australia, Vol. 2. Geological Survey of South Australia, Bulletin, 54, 218-281. -& BLUCK, R. G. 1990. Coastal palaeogeography and heavy mineral sand exploration targets in the western Murray Basin, South Australia. Austra-
--
lian Institute of Mining and Metallurgy, Proceedings, 295, 5-10. & CANN, J. H. 1990. Quaternary Evolution of the Robe-Naracoorte Coastal Plain. an Excursion Guide. South Australian Department of Mines
and Energy, Report Book 90[27. BLACKBURN, G. 1966. Radiocarbon dates relating to soil development, coast-line changes, and volcanic ash deposition in south-east South Australia. Australian Journal of Science, 29, 50-52. , ALLISON, G. B. &LEANEY, F. W. J. 1982. Further evidence of the age of tuff at Mt Gambier, South Australia. Royal Society of South Australia, Transactions, 106, 163-167 (and erratum 1984,
108, 130). BOURMAN, R. P. & MURRAY-WALLACE, C. V. 1991. Holocene evolution of a sand spit at the mouth of a large river system: Sir Richard Peninsula and the Murray Mouth, South Australia. Zeitschrift ffir Geomorphologie, Supplementband, 81, 63-83. BOWDEN, A. R. & COLHOUN, E. A. 1984. Quaternary emergent shorelines of Tasmania. In: THOM, B. G. (ed.) Coastal Geomorphology in Australia. Academic Press, Sydney, 313-342. BRYANT, E. A. 1992. Last interglacial and Holocene trends in sea level maxima around Australia: implications for modern rates. Marine Geology, 108, 209-217. CANN, J. H., DE DECKKER, P. & MURRAY-WALLACE, C. V. 1991. Coastal aboriginal shell middens and their palaeoenvironmental significance, Robe Range, South Australia. Royal Society of South Australia, Transactions, 115, 161-175.
ET AL.
CHAPPELL, J. 1987. Late Quaternary sea-level changes in the Australian region. In: TOOLEY, M. J. & SHENNAN, I. (eds) Sea-level Changes. Basil Blackwell, London, 296-331. CHEN, J., CURRAN, H. A., WHITE, B. & WASSERBURG, G. J. 1991. Precise chronology of the last interglacial period: 234u-z34Th data from fossil coral reefs in the Bahamas. Geological Society of America Bulletin, 103, 82-97. COLHOUN, E. A., VAN DE GEER, G. FITZSIMONS, S. J. & HEUSSER, L. E. 1994. Terrestrial and marine records of the Last Glaciation from Western Tasmania: do they agree? Quaternary Science Reviews, 13, 293-300. COOK, P. J., COLWELL,J. B., FIRMAN, J. B., LINDSAY, J. M., SCHWEBEL,D. A. 8r VON DER BORCH, C. C. 1977. The late Cainozoic sequence of southeast South Australia and Pleistocene sea-level changes.
Bureau of Mineral Resources Journal of Australian Geology and Geophysics, 2, 81-88. GOSTIN, V. A., BELPERIO, A. P. & CANN, J. H. 1988. The Holocene non-tropical coastal and shelf carbonate province of southern Australia. Sedimentary Geology, 60, 51-70. HOSSFELD, P. A. 1950. The Late Cainozoic history of the south-east of South Australia. Royal Society of South Australia, Transactions, 73, 232-279. HUNTLEY, D. J., HUTTON, J. T. & PRESCOTT, J. R. 1993. The stranded beach-dune sequence of south-east South Australia: a test of thermoluminescence dating, 0-800ka. Quaternary Science Reviews, 12, 1-20. --& 1994. Further thermoluminescence ~tates from the dune sequence in the southeast of South Australia. Quaternary Science Reviews, 13, 201-207. IDNURM, M. & COOK, P. J. 1980. Palaeomagnetism of beach ridges in South Australia and the Milankovitch theory of ice ages. Nature, 286, 699-702. JOYCE, E. B. 1975. Quaternary volcanism and tectonics in southeastern Australia. In: SUGGATE, R. P. & CRESSWELL, M. M. (eds) Quaternary Studies. Bulletin, Royal Society of New Zealand, 13, 169-176. LAMBECK, K. & NAKADA, M. 1992. Constraints on the age and duration of the last interglacial period and on sea level variations. Nature, 357, 125-128. LEANLY, F. W. J., ALLISON, G. B., DIGHTON, J. C. & TRUMBORE, S. 1995. The age and hydrological history of Blue Lake, South Australia. Palaeogeography, Palaeoelimatology, Palaeoecology, 118, 111-130. LUOBROOK, N. H. 1983. Molluscan faunas of the Early Pleistocene Point Ellen Formation and Burnham Limestone, South Australia. Royal Society of South Australia, Transactions, 107, 37-49. MARTINSON, D. G., PISIAS, N. G., HAYS, J. D., IMBRIE, J., MOORE, T. C. & SHACKLETON, N. J. 1987. Age dating and the orbital theory of the ice ages: development of a high-resolution 0 to 300000-year chronostratigraphy. Quaternary Research, 27, 1-29. MURRAY-WALLACE,C. V. • BELPERIO,A. P. 1991. The last interglacial shoreline in Australia - a review. Quaternary Science Reviews, 10, 441-461.
QUATERNARY NEOTECTONISM, SOUTHEASTERN AUSTRALIA , , CANN, J. H., HUNTLEY, D. A. & PRESCOTT, J. R. 1996. Late Quaternary uplift history, Mount Gambier region, South Australia. Zeitschrift ffir Geomorphologie, Supplementband, 106, 41-56. , , PICKER, K. & KIMBER, R. W. L. 1991. Coastal aminostratigraphy of the last interglaciation in southern Australia. Quaternary Research, 35, 63-71. OVA, Y. 1994. Stud)' on coral reef terraces of the Huon Peninsula, Papua New Guinea. A preliminary report on project 04041048, supported by Monbusho International Research Program. ROBERTSON, G. B., PRESCOTT, J. R. & HUTTON, J. T. 1996. Thermoluminescence dating of volcanic activity at Mount Gambier, South Australia. Royal Society of South Australia, Transactions, 120, 7-12. SCHWEBEL, D. A. 1984. Quaternary stratigraphy and sea-level variation in the southeast of South Australia. In: T~OM, B. G. (ed.) Coastal Geomorphology in Australia. Academic Press, Sydney, 291-311. SHEARD, M. J. 1990. A guide to Quaternary volcanoes in the lower south-east of South Australia. Mines and Energy Review, 157, 40-50. --1995. Quaternary volcanic activity and volcanic hazards. In: DREXEL, J. F, & PREISS, W. V. (eds) The Geology of South Australia, Vol. 2. Geological Survey of South Australia, Bulletin, 54, 264-268.
267
SMJTrL B. W. & PRESCOTT, J. R. 1987. Thermoluminescence dating of the eruption at Mt Schank, South Australia. Australian Journal of Earth Sciences, 34, 335-342. SPRIGG, R. C. 1952. The Geology of the South-East Province, South Australia, with Special Reference to Quaternary Coast-line Migrations and Modern Beach Developments. Geological Survey of South Australia Bulletin, 29. 1979. Stranded and submerged sea-beach systems of southeast South Australia and the aeolian desert cycles. Sedimentary Geology, 22, 53-96. STIRLING, C. H., ESAT, T. M., MCCULLOCH, M. T. & LAMBECK,K. 1995. High-precision U-series dating of corals from Western Australia and implications for the timing and duration of the Last Interglacial. Earth and Planetary Science Letters, 135, 115-130. VON DER BORCH, C. C., BADA,J. L. & SCHWEBEL,D. L. 1980. Amino acid racemization dating of Late Quaternary strandline events of the coastal plain sequence near Robe, southeastern South Australia. Royal Society of South Australia, Transactions, 104, 167-170. WOODS, J. E. 1862. Geological Observations in South Australia. Longman, London. ZHu, Z. R., WYRWOLL, K.-H., COLLINS, L. B., CHEN, J. H., WASSERBURG, G. J. 8r EISENHAUER, A. 1993. High precision U-series dating of last interglacial events by mass-spectrometry: Houtman Abrolhos Islands, Western Australia. Earth and Planetary Science Letters, 118, 281-293.
Late Cenozoic emergence of the islands of the northern Lau-Colville Ridge, southwest Pacific P. D. N U N N
Department of Geography, The University o f the South Pacific, P.O. Box 1168, Suva, Fiji Abstract: The Lau-Colville Ridge is part of an island arc abandoned during Pliocene time as a result of development of the back-arc Lau Basin. Throughout much of Plio-Pleistocene time, the ridge subsided, and its volcanic peaks were submerged and cloaked with (reef) limestone. Uplift, mostly during Pleistocene time, caused the northern part of the ridge to emerge by at least 315 m. Islands along the northern Lau-Colville Ridge exhibit ten welldefined terraces above c. 10 m whose ages are constrained by those of anomalously young lavas and avian phosphates. Low-level emerged shorelines may be all of Holocene age and exhibit a recurrence interval of coseismic uplift events of 1500 years. The areas of lithospheric convergence in the southwest Pacific provided a testing ground for many models of crustal evolution (Raitt et al. 1955; Isacks et al. 1967). The study of this region continues to raise questions about such models, particularly the applicability of large-scale global models to local situations. This paper outlines two models of lithospheric behaviour which have arisen from the study of island
tectonics along the northern part of the elongate north-south Lau-Colville Ridge in the southwest Pacific (Fig. 1).
Area of study An understanding of the region between the large Pacific and Indo-Australian Plates in the
Fig. 1. The islands rising from the northern Lau-Colville Ridge showing the geotectonic context of the LauColville Ridge within the southwest Pacific.
NUNN, P. D. 1998. Late Cenozoic emergence of the islands of the northern Lau-Colville Ridge, southwest Pacific. In: STEWART,I. S. • VITA-FINZI,C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 269-278.
270
P . D . NUNN
southwest Pacific which is fully consistent with observations on its geological evolution has long eluded researchers (Kroenke 1984; Pelletier & Louat 1989). Part of this understanding hinges on viable interpretations of back-arc basins and microplate behaviour, the study of both of which has been hindered here by a paucity of information from the region itself and lack of sufficient communication between marine geologists and their land-based counterparts. Interest in the region's hydrocarbon potential led to a spate of studies which have aided research into geological evolution, resulting in a number of volumes which are regarded as providing the best available interpretations of this (Greene & Wong 1988; Stevenson et al. 1994). The Pacific and Indo-Australian Plates converge obliquely in the southwest Pacific. Their boundary is marked by the development of microplates of generally unknown origin, size and number. Some microplates may not be completely independent but are developing through the propagation of particular boundary types. Some microplates may be fusing with others as formerly active plate boundaries become progressively inactive (Nunn 1995a, 1998). The islands of eastern Fiji, the Lau group, have long been recognized as having an origin distinct from that of the rest of Fiji (Agassiz 1899; Davis 1920). Whereas most of the rest of Fiji is of dominantly igneous composition, formed from the accretion of island arcs upon one another, most Lau islands are composed largely of limestone overlying a volcanic basement. The islands are clustered in the northern part of the north-south trending, southward plunging, elongate Lau-Colville Ridge (Fig. 1), which was an active island arc during late Tertiary time. At that time the Lau-Colville Ridge was joined to an ancestral Tonga-Kermadec Ridge, with a trench marking the axis of plate convergence along its eastern side. Longitudinal splitting of the composite ridge at c. 5 Ma was associated with development of a divergent plate boundary (the Valu Fa Ridge in the south, the Central Lau Spreading Centre (CLSC) and Northern Lau Spreading Centre (NLSC) further north) in the centre of the modern Lau-Havre Basin and led to the relative eastwards displacement of the Tonga-Kermadec Ridge and Trench. This movement removed the Lau-Colville Ridge from a convergent plate boundary and it became an abandoned island arc. Subsidence marked much of the early post-displacement history of the Lau-Colville Ridge but its northern part experienced considerable uplift during mid- and late Quaternary time (Nunn 1987, 1996; Parson et al. 1990).
Late Cenozoic evolution of the Lau islands The conspicuous evidence of late Cenozoic uplift in the Lau islands proved enigmatic for many researchers before the advent of plate-tectonic theory. For example, Davis (1927) regarded the Lau islands as having been elevated by a migrating anticline, whereas Ladd & Hoffmeister (1945) regarded the islands as rising from an upfold of ocean floor which was conceivably part of a continent in the making. An understanding of the dynamics of plate convergence along ocean trenches such as the Tonga-Kermadec Trench led to an appreciation of the significance of adjoining island arcs, and the eventual recognition of back-arc basins such as the Lau-Havre Basin and abandoned island arcs such as the Lau-Colville Ridge (Milsom 1970; Katz 1978). Island-arc volcanism, which was responsible for the construction of the volcanic foundations of most modern Lau islands, occurred at 14.556.06 Ma. Subsequent development of limestones suggests that subsidence occurred thereafter, as a consequence of the initial removal of this part of the Lau-Colville Ridge from the convergent plate boundary at c. 5Ma. Limestone development continued during Pliocene and perhaps early Pleistocene time although volcanism occurred on a very few islands at 4.01-2.46Ma and 2.53-0.28Ma in different parts of the group. The significance of these localized phases of volcanism, which were superficial compared with the earlier island-building phase, is discussed below (Woodhall 1985; Rodda 1994). Uplift of the some of the islands in the northern part of the Lau-Colville Ridge probably began in Pliocene time and had extended to most of the remainder by the end of the early Pleistocene around 0.78 Ma. Uncertainty about the age of initial uplift reflects the difficulty of dating the highest, most altered limestones in Lau (Nunn 1996). Uplift continued for most of Quaternary time, although it may have slowed or even stopped by its latest part.
Nature of pre-Holocene uplift The evidence for Plio-Pleistocene uplift of the northern part of the Lau-Colville Ridge comes principally from emerged limestones which reach a maximum 315m above modern reefs (on Vatuvara; Fig. 2). The form of these limestones has engendered much debate about the nature of uplift. The uppermost parts of the limestone-capped islands generally exhibit a basin-and-rim morphology which several early workers, notably Andrews (1916) and Davis
EMERGENCE OF THE LAU-COLVILLE RIDGE,
271
Fig. 2. Vatuvara viewed from the north. The staircase of terraces visible on the left is upfaulted with respect to the highest part of the island, represented by the surface at 315 m above sea level. (1920, 1928), regarded as evidence for uplifted atolls, the basin and rim being the subaerial equivalents of the atoll lagoon and reef, respectively. Other researchers, notably Agassiz (1899) and Ladd & Hoffmeister (1945), found little geological evidence for this, and preferred, as does the present author (Nunn 1996), to regard the basin-and-rim morphology as a wholly solutional effect. This agrees with the idea that many of these limestones formed so far below sea level that coral reef could not grow (see above), and also implies that initial uplift was either so rapid that reef did not have a chance to become established when the rising surface of the island passed through the photic zone or that surface waters at the time of initial island uplift were not
suitable for the establishment of coral reef. Either suggestion is compatible with the conspicuous lack of emerged coral reef in the highest parts of most of the highest limestones. Apart from the morphology of the islands' highest parts, that of their flanks, which are commonly terraced, has also attracted attention from geomorphologists. Most terraces investigated in detail (Table 1) are clearly emerged coral reefs which have grown outwards from the base of older emerged limestone or volcanic rocks. A few terraces may be wholly erosional but it is difficult to know whether they appear this way because they are truly so or because a veneer of emerged reef has been wholly removed since emergence. A good example comes from Yacata,
Table 1. Data referring to emerged terraces in Lau (after Nunn 1996) Surface High terraces Koromakawa Terrace Vesikana Terrace Naseicake Terrace Delaisalia Terrace Middle to low terraces Vagadra Terrace* Nativativa Terracet Koroqara Terrace Nalami Terrace Maruna Terrace Bureta Terrace Vunirewa Terrace
Number of observations
Mean height (m)
Height range (m)
2 1 3 3
255 160 128 103
50 0 15 12
8 11 13 7 9 10 8
70 55 33 25 20 14 10
* Shoreline Group 4 of Nunn (1987). t Shoreline Group 3 of Nunn (1987).
19 21 13 6.2 5.4 2.5 2.3
272
P . D . NUNN
Chronology of pre-Holoeene uplift
Fig. 3. View along an emerged notch lying above the assumed Last Interglacial notch at Skeleton Cave on the east of Kaibu island. This notch may be of Penultimate Interglacial age.
where a terrace cut in limestone continues onto volcanic rocks. Unsurprisingly, the lower terraces are the best preserved and best dated, and those whose origin can be best understood. Last Interglacial and Holocene emerged reefs can be readily located on most islands in the northern part of the Lau islands. On some islands the cliff faces display vertical series of emerged notches. Emergence may carry a notch into a protected location where its essential form will be preserved, particularly when covered by tufa, moved inland from the shoreline and cloaked with vegetation. A good example found on the island Kaibu (Fig. 3) is inferred to date from the Penultimate Interglacial. Analyses of emerged shoreline levels from Lau islands (Fig. 4) show that emergence was greatest in the north and declined, apparently uniformly, southwards (Nunn 1987, 1996). There is no evidence to suggest a consistent change in emerged shoreline elevations from west to east (cross-ridge).
As no dates are yet available from late Cenozoic emerged reefs on Lau islands, a chronology has been constructed using dates from avian phosphate and anomalously young lavas. Avian phosphates developed on the Vunirewa Terrace and higher surfaces on Vanua Vatu have been dated by uranium series to 111 + 15 ka BP; phosphates on the Vagadra Terrace on Tuvuca had developed by at least 300 ka BP (Roe et al. 1983). As these phosphates could have developed only when the (youngest) surface had emerged above sea level, their ages are minima for the initial emergence of the respective terraces. Young lavas on Mago island have been eroded at the same level as the Maruna Terrace (Fig. 5). Field mapping suggests that the erosional volcanic terrace is of the same age as the surface of the constructional reef terrace. The minimum age of the youngest volcanic rocks here (0.28 ~: 0.12 Ma; Whelan et al. 1985) can thus be regarded as a maximum date for this terrace. Using these dates and correlating terrace formation and emergence with sea-level maxima, a tentative chronology for middle to low terraces is suggested (Table 2). Of particular note are the multiple Last Interglacial terraces, which support the idea that there was more than one sea-level maximum at this time, a view originating with Chappell's (1974) recognition of a disconformity between emerged Reefs VIIb and ViIa on the Huon Peninsula in Papua New Guinea, and suggested by similar studies elsewhere in the Pacific (Sherman el al. 1993). Also shown in Table 2 are uplift rates for each of the seven low terraces based on the inferred age of the terrace, the sea level at which it formed and its mean elevation. Rates vary from 0.11 to 0.35 ram/year (mean 0.22ram/year) and appear to have been slightly more rapid at the time of formation of the older terraces. Not surprisingly, these rates are considerably lower than most from active frontal arcs although they are close to long-term rates from Atauro island in Timor (Chappell & Veeh 1978) and Hateruma island in the Ryukyus (Pirazzoli & Kawana 1986).
Outline of model for pre-Holocene uplift Following its abandonment some 5 x 106 years ago, the Lau-Colville Ridge began sinking. This is manifested most clearly by the build-up of limestone around and over the subaerial volcanic peaks of the ridge. This limestone covered many peaks; islands such as Vatoa and Vatuvara (which reaches 315m above the modern reef;
E M E R G E N C E OF T H E L A U - C O L V I L L E R I D G E ,
273
Fig. 4. Middle to low terraces on islands of the northern Lau-Colville Ridge (after Nunn 1996). (Note the greater elevations of contemporary shorelines in the north compared with the south of the region.)
Fig. 5. Profile of the 20 m Maruna Terrace on the north coast of Mago island. This terrace is both cut across young volcanic rocks (pictured) dated to 0.28 Ma and represented by reef of presumed Last Interglacial age.
274
P.D. NUNN
Table 2. Suggested chronology of Lau island terraces Terrace
Known age (ka BP)
Suggested Sea level age (ka BP) (m)
Uplift rate (mm/year)
Vagadra Terrace Nativativa Terrace Koroqara Terrace Nalami Terrace Maruna Terrace Bureta Terrace Vunirewa Terrace
> 300 (minimum)
325* 240* 215" 175" 124t l18t 106t
0.22 0.35 0.18 0.28 0.11 0.12 0.27
280 (maximum) 111 (minimum)
0 -30 -6 -24 +6 0 -19
It is assumed that emergence took place shortly after formation so that the ages of the two events are the same. * Sea-level highstand recognized on the Huon Peninsula in Papua New Guinea (derived graphically from Chappell 1983). t Last Interglacial sea-level highstand recognized on the Huon Peninsula in Papua New Guinea and reconciled with the deep-sea chronology (Chappell & Shackleton 1986).
Fig. 2) are wholly limestone, others like Nayau and Yacata are limestone in their upper parts yet have volcanic basements protruding in lower parts. A few islands such as Lakeba are typical makatea types, as defined by N u n n (1994), with a high volcanic peak surrounded by lower ridges of emerged reef limestone. Post-subsidence uplift commencing in late Pliocene or early Pleistocene time appears to have been largely responsible for the islands' subsequent emergence. This emergence was not uniform along the ridge. It appears to have exclusively affected its northern part, and to have decreased uniformly in magnitude from north to south. Given that the highest emerged limestone in Lau is 315 m above modern sea level, and assuming that it was raised at a rate of 0.22 ram/year, it may be of only early Pleistocene (c. 1.43 Ma) age. Using maximum and minimum rates of uplift (as in Table 2), this limestone may date from late Pliocene (2.86 Ma) or late early Pleistocene (0.9 Ma) time. However, these calculations take no account of the depth below sea level at which this limestone formed, the amount of subaerial lowering of the emerged limestone, or of possible temporal variations in uplift rate.
Notches slightly larger than present ones are believed to have formed when the coast was not reef fringed (most are reef fringed today) and thus wave erosion was proportionately greater; in Hopley's (1984) term, a 'high-energy window' was open. Typical views of low-level (<10m) emerged shorelines are shown in Fig. 6. Unlike emerged shorelines within this elevation range on other islands in the Lau group, which are thought to be pre-Holocene in age (e.g. Fig. 3), most of the low-level emerged shorelines on Moce and its satellites are free of debris and regularly doused with sea spray, which suggests that they may be only of Holocene age (see below). Analysis of heights between islands (Table 4) suggests that some low-level emerged shorelines have already been obliterated from certain islands. However, the three lowest emerged shorelines are clear on each of the islands studied and correlation between their heights shows a pattern consistent with increasing emergence away from a centre in the M o c e - O n e a t a nexus (Fig. 7).
Nature of Holocene uplift
Moce
Oneata
Lakeba
Nayau
Namuka
Evidence for low-level recent uplift of Moce and four satellite islands involves mostly emerged notches etched into the cliffs in various places. Below 10 m there are 4-7 emerged shorelines on most islands (Table 3). Most emerged notches have forms which are comparable with those of modern notches, which are the usual cliff-foot landforms found around the modern limestone coasts of the islands,
0.28 1.02 1.69 4.59 7.60
0.40 0.85 1.52
0.47 1.46 2.68 4.18 5.14 7.60 9.30
0.57 2.25 3.70 4.90
0.72 1.79 2.85 4.53 5.68 7.71
Table 3. Data referring to evidence oflow-level (<10 m) shoreline emergence (m) on Moce and its satellite islands in central Lau (after Nunn 1995b)
5.70 9.58
Suggested age correlations on the basis of mean elevation as shown in Fig. 4 are indicated.
EMERGENCE OF THE LAU-COLVILLE RIDGE,
275
Fig. 6. Typical cliff sections from central Lau islands showing multiple emerged shorelines.
Chronology of uplift On account of their generally exposed locations and excellent state of preservation, it is suspected that all the low-level emerged shorelines on Moce and its satellites formed since the sea surface reached close to its present level, around at 6 ka BP in this region (Nunn 1995c). On Moce and Lakeba, the two lowest shorelines have been dated to around 2200 cal. a BP and 3800cal. aBP (dates from Nunn (1995b) calibrated using the OxCal program of Stuiver et al. (1993)). Ifcoseismic uplift was responsible for the uplift of this shoreline series, a crude recurrence time of around 1600 years is calculated, a significant increase from that of 1045 years obtained by Nunn (1995b) from uncalibrated dates. Most events involve uplift of 1-2 m, which is similar to that associated wth coseismic uplift elsewhere in the island arcs of the southwest and west Pacific (Ota 1991; Ota & Kaizuka 1991).
The proposed rate is comparable with that obtained for similar situations in Fiji, principally Vatulele island (1375-1406 years: Nunn 1998), nearby Beqa and Yanuca islands (1125 years; Nunn 1998); the Ryukyu forearc (1200-1400 years; Sieh 1981) and the Huon Peninsula in Papua New Guinea (970-1165 years; Ota & Chappell 1996). Shorter intervals have been proposed for the outer sedimentary arc in Tonga (869 years; Nunn & Finau 1995), and for active forearcs in Vanuatu and Solomon Islands (Taylor et al. 1980; Hopley 1987).
Outline of model Nunn (1996) recently argued that the long-term tectonic history of the Lau islands has alternated between periods of slow monotonic aseismic uplift, which produced the broad terraces flanking islands, and periods of coseismic uplift
276
P.D.
NUNN
Table 4. Suggested ages for the times of formation and initial uplift of the three lowest of at present emerged shorelines on Moce and its satellite islands assuming a mean recurrence interval of coseismic uplift of 1045 radiocarbon years (after Nunn 1995b) Island
Mean shoreline elevation (m)*
Time of formation (a BP)
Time of initial uplift (a BP)
Moce
0.28 1.02 1.69
2695 3740 4785
Not uplifted 2695 3740
Oneata
0.40 0.85 1.52
1045 2090 3135
Not uplifted 1045 2090
Lakeba
0.47 1.46 2.68
3800 4845 5890
Not upli~ed 3800 4845
Nayau
0.57 2.25 3.70
2240 3285 4330
Not uplifted 2240 3285
Namuka
0.72 1.79 2.85
2360 3405 4450
Not uplifted 2360 3405
* Data from Table 3. responsible for comparatively rapid emergence associated with cliffs. It n o w seems m o r e plausible to suppose that rapid coseismic uplift occurs only on islands close e n o u g h to y o u n g volcanoes to have been affected by d e f o r m a t i o n associated
with their loading of the lithosphere. M c N u t t & M e n a r d (1978) explained the elevation of m a n y limestone islands in the South Pacific by the loading effects of nearby volcanoes. In Lau, both volcano size and uplift m a g n i t u d e are m u c h
em
12 ~D
"v"
Nayau
v
Lakeba
v
Oneata
,~v
Moce
"V"
Namuka
Fig. 7. Mean shoreline elevations and elevation ranges plotted for Moce and its satellite islands (after Nunn 1995b). Only those emerged shorelines with elevations below 4 m are shown.
E M E R G E N C E OF T H E L A U - C O L V I L L E R I D G E , smaller but this is not t h o u g h t to invalidate the analogy. A n intriguing possibility is that whereas slow m o n o t o n i c aseismic uplift could be the n o r m for the islands of the n o r t h e r n L a u - C o l v i l l e Ridge, where this process is exacerbated by the loading o f y o u n g volcanoes, slow uplift becomes fast, and aseismic uplift becomes coseismic. This w o u l d certainly explain the enigma, which is not confined to the L a u - C o l v i l l e Ridge, of why aseismic uplift occurs in some places and coseismic uplift in others, and why the two varieties of uplift occasionally alternate t h r o u g h time. E. Silver (University of California at Santa Cruz) was the first person to suggest to me that heating of a detached slab of lithosphere might have been responsible for uplift of the northern part of the Lau-Colville Ridge. J. Chappell (Australian National University) has given me considerable feedback regarding some of the ideas expressed in this paper. P. Rodda (Fiji Mineral Resources Department) provided many thoughtful comments on the first draft of this chapter. The University of the South Pacific funded much of the research on which this paper is based, principally through Grant 6839-1431-70766-15.
R
e
f
e
r
e
n
c
e
s
AGASSIZ, A. 1899. The Islands and Coral Reefs of Fiji. Bulletin of the Museum of Comparative Zoology, Harvard, 33. ANDREWS, E. C. 1916. Relations of coral reefs to crust movements in the Fiji Islands. American Journal of Science, 41, 135-141. CHAI'PELL, J. 1974. Geology of coral terraces, Huon Peninsula, New Guinea: a study of Quaternary tectonic movements and sea level changes. Geological Society of America Bulletin, 85, 553-570. - - 1 9 8 3 . A revised sea-level curve for the last 300,000 years from Papua New Guinea. Search, 14, 99-101. & SHACKLETON,N. J. 1986. Oxygen isotopes and sea level. Nature, 324, 137-140. & VEEH, H. H. 1978. Late Quaternary tectonic movements and sea-level changes at Timor and Atauro Island. Geological Society of America Bulletin, 89, 356-368. DAVIS, W. M. 1920. The islands and coral reefs of Fiji. Geographical Journal, 55, 34-45, 200 220, 377-388. - - 1 9 2 7 . A migrating anticline in Fiji. American Journal of Science, 83, 333-351. - - 1 9 2 8 . The Coral Reef Problem. American Geographical Society Special Publication, 9. GREENE, H. G. & WONG, F. L. (eds) 1988. Geology and Offshore Resources of Pacific Island ArcsVanuatu Region. Earth Science Series, Vol. 8. Circum-Pacific Council for Energy and Mineral Resources, Houston, TX. HOPLEY, D. 1984. The Holocene 'high energy window' on the central Great Barrier Reef. In: THOM, -
-
-
-
277
B. G. (ed.) Coastal Geomorphology in Australia. Academic Press, London, 135-150. - - 1 9 8 7 . Holocene sea-level changes in Australasia and the southern Pacific. In: DEVOY, R. J. N. (ed.) Sea Surjace Studies: A Global View. Croom Helm, London, 375 408. ISACKS, B. L. SYKES, L. R. & OLIVER, J. 1967. Spatial and temporal clustering of deep and shallow earthquakes in the Fiji-Tonga Kermadec region. Bulletin of the Seismological Society of America, 57, 935-958. KATZ, H. R. 1978. Composition and age of Lau Basin and Ridge volcanic rocks: implications for evolution of an interarc basin and remnant arc: discussion and reply. Geological Society of America Bulletin, 89, 1118 1120. KROENKE, L. W. 1984. Cenozoic tectonic development of the southwest Pacific. United Nations ESCAP, CCOP/SOPAC Technical Bulletin, 6. LADD, H. S. & HOFFMEISTER, J. E. 1945. Geology of Lau, Fifi. Bernice P. Bishop Museum (Honolulu), Bulletin, 181. McNuTT, M. & MENARD, H. W. 1978. Lithospheric flexure and uplifted atolls. Journal of Geophysical Research, 83, 1206 1212. MILSOM, J. S. 1970. The evolution of the Lau Ridge, Fiji Islands. Earth and Planetary Science Letters, 8, 258-260. NUNN, P. D. 1987. Late Cenozoic tectonic history of Lau Ridge, South-West Pacific, and associated shoreline displacements: review and analysis. New Zealand Journal of Geology and Geophysics, 30, 241-260. - - 1 9 9 4 . Oceanic Islands, Blackwell, Oxford. - - 1 9 9 5 a . Lithospheric flexure in southeast Fiji consistent with the tectonic history of islands in the Yasayasa Moala. Australian Journal of Earth Sciences, 42, 377-389. - - 1 9 9 5 b . Holocene tectonic histories for five islands in the south central Lau group, Southwest Pacific. The Holocene, 5, 160-171. - - 1 9 9 5 c . Holocene sea-level changes in the south and west Pacific. Journal of Coastal Research, 17, 311-319. - - 1 9 9 6 . Emerged Shorelines of the Lau Islands. Fiji Mineral Resources Department Memoir, 4. - - 1 9 9 7 . Pacific Island Landscapes. Institute of Pacific Studies, The University of the South Pacific, Suva. & FINAU, F. T. 1995. Late Holocene emergence history of Tongatapu island, South Pacific. Zeitschrift for Geomorphologie, 39, 69 95. O'rA, Y. 1991. Coseismic uplift in coastal zones of the western Pacific rim and its implications for coastal evolution. Zeitschrift fiir Geomorphologie, Supplementband, 81, 163 179. & CHAPPELL,J. 1996. Late Quaternary coseismic uplift events on the Huon Peninsula, Papua New Guinea, deduced from coral terrace data. Journal of Geophysical Research, 101, 6071-6082. -& KAIZUKA, S. 1991. Tectonic geomorphology at active plate boundaries- examples from the Pacific Rim. Zeitschrift ft~'r Geomorphologie, Supplementband, 82, 119-146. -
-
-
-
278
P.D.
PARSON, L. M., PEARCE, J. A., MURTON, B. J., HODKINSON, R. A. & R.R.S. Charles Darwin Scientific Party 1990. Role of ridge jumps and ridge propagation in the tectonic evolution of the Lau back-arc basin, southwest Pacific. Geology, 18, 470-473. PELLETIER, B. & LOUAT, R. 1989. Mouvements relatifs des plaques dans le sud-ouest Pacifique. Comptes Rendus de l'Acad~mie des Sciences, S~rie II, 308, 123-130. PIRAZZOLI, P. A. & KAWANA, T. 1986, DGtermination de mouvements crustaux quaternaires d'aprGs la d&ormation des anciens rivages dans les ]les Ryukyu, Japon. Revue de G~ologie Dynamique et de Gdographie Physique, 27, 269-278. RAITT, R. W., FISHER, R. L. & MASON, R. G. 1955. Tonga trench. Geological Society of America, Special Paper, 62, 237-254. RODDA, P. 1994. Geology of Fiji. In: STEVENSON,A. J., HERZER, R. H. & BALLANCE,P. F. (eds) Geology and Submarine Resources of the Tonga-Lau-Fiji Region. SOPAC Technical Bulletin, 8, 131-151. ROE, K. K., BURNETT, W. C. & LEE, A. I. N. 1983. Uranium disequilibrium dating of phosphate deposits from Lau group, Fiji. Nature, 302, 603-606. SHERMAN, C. E., GLENN, C. R., JONES, A. T., BURNETT, W. C. & SCHWARCZ, H. P. 1993. New evidence for two highstands of the sea during the last interglacial, oxygen isotope stage 5e. Geology, 21, 1079 1082.
NUNN SIEH, K. E. 1981. A review of geological evidence for recurrence time of large earthquakes. In: SIMPSON, D. W. & RICHARDS, P. G. (eds). Earthquake Prediction: an International Review. American Geophysical Union, Washington, DC, 181-207. STEVENSON, A. J., HERZER, R. H. & BALLANCE,P. F. (eds) 1994. Geology and Submarine Resources of the Tonga-Lau-Fiji Region. SOPAC Technical Bulletin, 8. STUIVER, M., LONG, A. & KRA, R. S. (eds). 1993. OxCal v 2.0. Radiocarbon 35. TAYLOR, F. W., ISACKS, B. L., JOUANNIC,C., BLOOM, A. L. & DuBoIs, J. 1980. Coseismic and Quaternary vertical tectonic movements, Santo and Malekula islands, New Hebrides island arc. Journal of Geophysical Research, 85, 5367-5381. WHELAN,P. W., GILL, J. B., KOLLMAN,E., DUNCAN,R. & DRAKE, R. 1985. Radiometric dating of magmatic stages in Fiji. In: SCHOLL, D. W. & VALLIER, T. L. (eds) Geology and Offshore Resources of Pacific Island Arcs- Tonga Region. Circum-Pacific Council for Energy and Mineral Resources, Houston, TX, 415-440. WOODHALL, D. 1985. Geology of the Lau Ridge. In: SCHOLL, D. W. & VALLIER, T. L. (eds). Geology and Offshore Resources of Pacific Island Arcs-Tonga Region. Circum-Pacific Council for Energy and Mineral Resources, Houston TX, 351-378.
Holocene coastal tectonics in NE Brazil FRANCISCO
H. R. B E Z E R R A 1'2, F R A N C I S C O
RICARDO
F. A M A R A L 1, L U C I A N O
P. L I M A - F I L H O L,
H. O. C A L D A S 1
& L E A O X. C O S T A - N E T O 3
1Departamento de Geologia, Universidade Federal do Rio Grande do Norte ( UFRN), Natal-RN, 590 72-9 70, Brazil (e-mail: bezerrafh @geologia. ufrn.br) 2 Department o f Geological Sciences, University College, Gower Street, London, WC1E 6BT, UK 3 Escola T~cnica Federal do Rio Grande do Norte ( E T F R N ) , Av. Senador Salgado Filho, no. 1559, Nova Descoberta, Natal-RN, 59056-000, Brazil Abstract: We have identified two beachrock facies in NE Brazil which can be used as reliable sea-level indicators. Facies (a) represents the lower foreshore and the upper shoreface, being mainly characterized by coarse unsorted sandstones with trough crossstratification, whereas facies (b) corresponds to sediments deposited on the middle to lower foreshore and is usually characterized by medium to coarse, better sorted sandstones which form seaward-dipping, swash cross beds. The age of the beachrock bodies ranges from c. 7000 to 1150cal. a BP. Although the relative sea-level record is complicated by oscillations which are probably due to minor climatic changes, it was possible to identify a general rise at c. 7000 cal. a Bp which reached its maximum of +2 mc. 5000 cal. a BP and then started to fall to its present level some 300 years later. Our sea-level data are consistent with glacio-hydroisostatic models for the region but dates of 4080cal. a BP on shells in growth position at +5 m above sea level and 2780 cal. a ~p on a beachrock at sea level east of the Carnaubais fault point to rapid, possibly coseismic, late Holocene emergence.
Identifying and quantifying modest and local, coastal emergence or submergence caused by tectonics at passive continental margins may be hampered by eustatic and isostatic effects. There are two main problems: how to distinguish variations in sea level from local emergence and submergence of tectonic origin, and whether there is any widespread sea-level indicator which may be used to identify such variations on a tropical coast. The Earth's isostatic adjustment caused by deglaciation associated with ocean loading is now seen as a key explanation of sea-level changes along many coasts. Glacio-hydroisostatic models explain such adjustments and have provided curves which have been used as trends of sea-level changes all over the world (e.g. Clark et al. 1978; Newman et al. 1980; Peltier 1988). The Brazilian coast is distant from the main Pleistocene glaciation centres. However, it is generally accepted (e.g. Walcott 1972; Newman et al. 1980) that such distant regions have also shown an isostatic response to deglaciation and the resulting sea-level rise: continental regions such as South America and Africa were flexed upwards as a consequence of the depression of oceanic areas and the transfer of mass from regions below the oceanic lithosphere to beneath
the continents. Sea-level studies in Brazil (e.g. Fairbridge 1976; Suguio et al. 1985) have shown that the Brazilian coastline has indeed undergone emergence in Holocene time and that raised beaches are a common feature. However, there is little information on how local coastal chronologies match the increasingly sophisticated glacio-hydro-isostatic models that are emerging and whether it is possible to distinguish the glaciohydro-isostatic component from the tectonic component in relative sea-level changes. The second problem, that of palaeo-shoreline indicators, demands a local solution. In tropical regions such as northeastern Brazil, beachrock is one of the most common coastal deposits. Beachrock may extend for several kilometres along the littoral zone and is, in places, the only available material for dating. Many studies quote beachrock as a reliable sea-level indicator in microtidal coastlines (e.g. Inden & Moore 1983; Hopley 1986; Cooper 1991; Kindler & Bain 1993; Ramsay 1995). However, little is known about how beachrock may be used as a sea-level indicator in mesotidal coasts such as those of northeastern Brazil. This study presents a description of the major recent coastal sedimentary deposits, mostly beachrock, along the littoral zone between Macau and
BEZERRA, F. H. R., LIMA-FILHO, F. P., AMARAL, R. F., CALDAS, L. H. O. & COSTA-NETO, L. X. 1998. Holocene coastal tectonics in NE Brazil. In: STEWART,I. S. 8~ VITA-FINZI,C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 279-293.
280
F. H. R. BEZERRA E T A L .
Fig. 1. General geological map of northeastern Brazil and study-area location. Sea-level curves proposed for the region: (a) Suguio et al. (1985); (b) Clark et al. 1978; (e) Peltier 1988; (d) W. R. Peltier (pers. comm., 1997).
Cunhati, northeastern Brazil (Fig. 1), and the corresponding radiocarbon chronology in an attempt to identify crustal movements in the Holocene. The occurrence of well preserved raised coastal deposits along such a littoral zone makes it an ideal site for a sea-level investigation. This paper emphasizes sedimentary structures as a basis for facies subdivision in the transitional (beach) environment where beachrock is formed. By comparing beachrock with modern beaches, we propose its use as a valid sea-level indicator in mesotidal coasts. We also compare relative sealevel changes in the area with known models of glacio-hydro-isostatic behaviour predicted for the region, to identify the local sea-level history and isolate tectonic movements. Tectonic analysis is confined to what the coastal data reveal about the local history of emergence and submergence.
Tectonic and coastal settings The study area is located in the Brazilian continental margin, 5~ to the south of the Equator (Fig. 1). A simplified stratigraphy for the region starts with a deformed crystalline basement of Precambrian age; Cretaceous basins formed by reactivation of Precambrian shear zones during the break-up of the South Atlantic overlie this basement, followed mainly by Cenozoic sedimentary rocks (Almeida et al. 1981; Matos 1992). Previous evidence of late Quaternary coastal tectonics in the region was presented by Martin et al. (1986a), who recognized tectonic vertical movements of up to 3m along the Salvador area (Fig. 1). Much of the seismicity of northeastern Brazil is less than 12 km deep. The region experiences half of Brazil's known seismicity
COASTAL TECTONICS IN NE BRAZIL
281
Fig. 2. Geological map and sampling sites in the study area.
and has been subject to a series of earthquake swarms (Ferreira et al. 1987; Takeya et al. 1989). An active fault (Samambaia fault, Fig. 2), well defined by earthquake epicentres and focal mechanisms (Takeya et al. 1989), is found near Jogo Cfimara at a depth of 5-10 km but evidence that it reaches the surface or can be traced to the coast has not yet been presented. Northeastern Brazil has a tropical climate, with average temperatures of c. 30~ The rainfall rate is about 600-1000mm/year (Nimer 1989). The coast has a mesotidal regime, where normal tides attain a maximum of 1.0-2.0 m and spring tides have a range of 3.2m (Hayes 1979). Recent erosion and weathering have left the Quaternary sedimentary record relatively unscathed, although oysters and algae obscure some outcrops along the littoral zone.
Sea-level investigation and glacio-hydroisostatic models on NE Brazil Darwin (1841) was perhaps the first to describe the coastal geomorphology of this region, in particular using the physiography of the reef off Recife (or Pernambuco, as it was then called) to infer a change in relative sea level. Since then,
many of the major physiographical units such as coastal barriers, sandstone reefs, and marine terraces have been analysed by other workers (e.g. Bigarella et al. 1961). Only a few deposits have been dated, however, and the resulting chronology for this coast depends largely on relative dating based on geomorphological, sedimentary and stratigraphical attributes. The use of radiocarbon analyses in Brazilian sea-level studies began in the 1960s. The first results were published by van Andel & Laborel (1964), who reconstructed the sea-level history of the Recife area (Fig. 1). Nevertheless, it was not until 1971 that Delibrias & Laborel studied a coast more than 3000 km long from Recife (Fig. 1) to Santo Amaro (southeastern Brazil) and made an initial attempt to understand the processes responsible for the last 7000 years of sea-level change in Brazil. Fairbridge (1976) also proposed a single sea-level curve for the Brazilian coast, which was based mainly on shell middens from the southeastern and south coasts. Five periods of transgression with amplitudes of 1-5 m were proposed. More localized time-depth diagrams have been presented in the last two decades. The most complete investigation was published by Suguio et al. (1985), who established different mid- to
282
F. H. R. BEZERRA E T AL.
late Holocene sea-level curves. The best example described by them was the 'Salvador curve', which is the best constrained sea-level curve for the Brazilian coast so far (diagram (a) in Fig. 1). The Salvador curve is based on 60 14C ages of geological, biological, pre-historical (shell middens) sea-level indicators collected along 50kin of littoral zone to the north of Salvador (Fig. I). It indicates that a relative sea-level rise began shortly before 7000 a Be, reached 5 m above the current sea level briefly before 5000aBe, and then fell to its present position. Two regressive periods around 4000 a BP and 2700 a BP were also indicated. More recently, a study by Dominguez et al. (1990) concluded that the Salvador curve was also valid for the Recife area (Fig. t). They dismissed samples which did not fit that curve as contaminated. Glacio-hydro-isostatic models predict a major post-glacial sea-level rise along the Brazilian coast. The first relevant scheme to be published was by Clark et al. (1978), who divided the Earth's surface into six regions of different isostatic patterns and used a numerical model to calculate sea-level changes. Emerged beaches were predicted in four regions, including zone IV (continental shorelines). The 'eustatic' sea-level rise was assumed to be 75.6m during the last 16000aBP, and ocean water volume was presumed to be constant since 5000 a Be. According to this model, most continental coasts, including that of Brazil, rose because of crustal tilting caused by loading of the ocean basins by the addition of meltwater. The sea-level curve of Clark et al. for northeastern Brazil (Recife area, diagram (b) in Fig. 1) predicted 3 m of land rise since 6000 a BP. Later glacio-hydro-isostatic models have been refined in the light of new facts including additional sea-level data. In the first version of his model, Peltier (1982) predicted raised beaches in the Recife area up to 2 m above present sea-level elevation from about 8000aBe and reaching their maximum elevation about 7000 a BP. In his second version, Peltier (1988) used three different melting chronologies: (a) melting chronology of ICE-2; (b) a 5000 year delay of the Antarctic component of melting; (c) a 7000 year delay of the Antarctic component of melting. These three hypotheses predicted unimodal sea-level curves with (a) 3 m (b) 1.6m and (c) less than 0.5 m sealevel rise at about 6750 a aP for the Recife area (diagram (c) in Fig. 1). More recently, the refined model by Peltier & Jiang (1996) indicates that northeastern Brazil experienced a sea-level rise of 4.6-4.8 mm/year at 20 000 cal. a BP, which increased to about 10 mm/year at 10 000 cal. a Be, and eventually decreased to 0.1-0.2 mm/year at
5000cal. a Be. A curve by W. P. Peltier (pers. comm., 1997) for the Touros area is presented in diagram (d) in Fig. 1 and will hereafter be called the 'Touros curve'.
Holocene coastal deposits in the Macau to Cunhafi littoral zone Beachrock
Beachrock is a sedimentary rock commonly formed in the intertidal zone although it can also develop in the sublittoral zone. The mineralogy of beachrock may vary from pure silica sands to biogenic carbonate sands, whereas beachrock cement may vary from aragonite to Mg-calcite (Stoddart & Cann 1965; Alexandersson 1972). More recent reviews of beachrock cement have been given by Inden & Moore (1983), Beier (1985), Hopley (1986), Amieux et al. (1989) and Strasser et al. (1989). Beachrock is found mainly in tropical regions but occurrences up to 45~ and 30~ latitude have also been reported by Alexandersson (1972) and Ramsay (1995), respectively.
Fig. 3. Oblique aerial view of the Guaraira beachrock (foreground) and Barreta beachrock (background).
COASTAL TECTONICS IN NE BRAZIL The location of the beachrock bodies investigated during this study is shown in Fig. 2. They display a great range of dimensions and shapes. In the majority of cases, they are present as elongated bodies, ranging in length from some kilometres (e.g. Perobas, Barreta, Guaraira, Cunhafi, Jacum~,, Farol de Sto. Alberto, Fig. 3) to some dozens of metres (e.g. Via Costeira, Guajiru, Galinhos). Beachrock also occurs in patches (e.g. Pedra Grande, Recuado, Macau, Lagoa do Sal, Fig. 4). The beachrock bodies range in width from 50 cm to 3 m and present tabular sets whose boundaries are generally erosive. They usually show gentle seaward-dipping bedding surfaces (<10 ~ and are parallel to the present coastline. The sedimentary structure, set geometry, petrographic characteristics and fossil content of the deposits were compared with modern beaches to identify their palaeogeography. In the study area, beaches are mainly characterized by swash cross-stratification in the middle to lower foreshore and trough cross-stratification in the lower foreshore to upper shoreface (Fig. 5). Some of the beaches are backed by cliffs in the Barreiras Formation (late Tertiary), whereas others are gentle strandline shores with recent sand dunes to the rear. Other mesotidal coasts on which beachrock is developed present similar physical processes and structures. Inden & Moore (1983) concluded that small to medium tabular festoon cross-beds are very common in the shoreface zone, whereas the foreshore is characterized, from bottom to top, by fine-grade beds, cross-bedding and parallel bedding. Reading & Collison (1996) stated that the beach progradation sequence comprises cross-lamination interbedded with and eventually passing to cross-bedding, capped by parallel bedding in the foreshore zone.
Fig. 4. Recent mangrove-swamp site and the Macau beachrock (lower right) (MC on Fig. 2).
283
Local descriptions of mesotidal beaches agree with these general models. Dabrio (1982) described the sedimentary structures typical of mesotidal beaches in southern Spain as crosslaminated sand and cross-bedded sands overlain by cross-bedding, and finally by parallel-bedded sands. Semeniuk & Johnson (1982) showed that in Western Australia the shoreface is characterized by trough-bedded sand and gravel, whereas the foreshore is distinguished by parallelbedded sand and laminated-bubble sand. All these studies agree that physical processes do not change sharply at low water but extend to the shoreface zone. Despite that, the transition between the lower foreshore and the upper shoreface is marked by the accumulation of the coarsest available grain size, which is therefore a sea-level indicator (Bourgeois 1980; Inden & Moore 1983; Shipp 1984; Dupr6 1984). Two beachrock facies were identified in the present study on the basis mainly of sedimentological features and comparison with modern beaches. They are hereafter informally called facies (a) and facies (b). Beachrock facies (a) represents the lower foreshore and the upper shoreface zones. Facies (a) is a medium to coarse, sometimes conglomeratic, sandstone. Its terrigenous constituents are quartz, limonite, fragments of marine shell and fragments of the underlying rocks. It presents a great variety of textural types and mineralogical maturity indicating different formative processes and sources. The most common sedimentary structures are trough cross-stratifications 0.2-1.5 m in thickness, interpreted as the result of migration of sinuous crested bedforms (Figs 5 & 6), which points to an important traction flow mechanism during transportation. Palaeocurrents show transport predominantly to the N N W on the N-S-trending coast as far as Touros city, where the coast bends to the W. The palaeocurrents shift predominantly to the W and to the NW on the E-W-trending coast. Both palaeocurrent patterns are similar to those observed on the present coastline, which are mainly influenced by longshore currents. Although the shoreface zone can extend from the low-tide level to the fairweather wave base (Reading & Collison 1996), the low-water level can in places be identified and used as a sea-level indicator with a precision of +0.5 m. The lowwater level is characterized by the coarsest texture associated with trough cross-stratification, sometimes capped by facies (b) (Fig. 6). Facies (a) corresponds to the lower foreshore beachrock of Flexor & Martin (1979) and the upper foreshore to lower shoreface beachrock of Oliveira et al. (1990) described in the Salvador and CunhafiNatal littoral zones, respectively (Fig. 1).
284
F. H. R. BEZERRA E T AL.
Fig. 5. Sedimentary structures of foreshore and upper shoreface. (a) swash cross-stratification in modern beach (middle to lower foreshore); (b) trough cross-stratification in modern beach (lower foreshore); (e) trough crossstratifications produced by the migration of sinuous crested bedforms in modern beach (lower foreshore); (d) swash cross-stratification intercalated with trough cross-stratification in the Barreta beachrock; (e) and (f), trough cross-stratification in the Barreta beachrock (BR1 on Fig. 2).
Beachrock facies (b) corresponds to sediments deposited on the foreshore. They are usually medium to coarse sandstones, which form tabular beds and sheets from 0.1 to 1.0m in thickness. Facies (b) is chiefly composed of quartz grains, heavy minerals (ilmenite, magnetite, zircon, tourmaline, staurolite, and rutile), and fragments of marine shells. The grain size increases on the lower foreshore, where the low-water level can sometimes be identified (Fig. 6). The most common sedimentary structure of this facies is sea-
ward dipping, swash cross-stratification (parallel bedding of previous studies), which is a sea-level indicator of the middle to lower foreshore with a precision of + 1.0 m. Other common structures on the foreshore are ripple marks, subcritically climbing translate strata, aeolian waves, and thin deflation pavements. Important palaeoecological implications may be drawn from correlation between beachrock fossils and modern beach fauna. There is no quantitative or qualitative difference between the fossil
COASTAL TECTONICS IN NE BRAZIL
285
N-S-trending coast. The complete diagenetic sequence observed in facies (a) and in facies (b) indicates five different stages of coastal emergence or submergence. The primary cement consists mainly of acicular crusts of aragonite (Fig. 7a and b) formed in the marine phreatic zone (e.g. Moore 1971; Tietz & Mfiller 1971). It has been replaced locally by late micrite grains of calcite which formed in the meteoric phreatic environment and indicates coastal emergence (Fig. 7c and d) (e.g. Beier 1985; Amieux et al. 1989). The third diagenetic phase is characterized by porosity reduction because of the growth of aragonite and Mg-calcite cement (Fig. 7c and d), which implies submergence (e.g. Tietz & Mtiller 1971; Meyers 1986; Amieux et al. 1989; Strasser et al, 1989). The fourth phase is marked by the overgrowth of crystalline calcite (Fig. 7f) from Mgcalcite in the meteoric phreatic zone, indicating another phase of coastal emergence (see Meyers 1986). The last diagenetic phase is characterized by cement dissolution and porosity growth still in the meteoric phreatic zone (Fig. 7e and f).
Other coastal deposits
Fig. 6. Modern beach cross-sections and beachrock facies: (a) Guajiru beach showing sample sites and cross-sections of the Guajiru and Recuado beachrocks; (b) Guaraira beach showing sample site and cross-section of the Guaraira beachrock. content of beachrock and death assemblages of marine shells which occur in modern beaches in northeastern Brazil (Maury 1934; Campos e Silva et al. 1964; Mendon~a 1966). Death assemblages of shells tend to concentrate in the lower foreshore on modern beaches. In beachrock, they are similarly more abundant in the coarsest part of facies (a). The most common species in order of abundance are Donax striata, Divaricella quadrisulcata, Tivela mactroides, Anomalocardia brasiliana, Anadara ovalis, and Ostrea sp. Our observations of beachrock diagenesis both in the E-W- and N-S-trending coasts match results presented by Oliveira et al. (1990) for the
Other coastal deposits were used as complementary evidence of sea-level changes. Coral reefs overlie the Barreiras Formation bedrock. The Pirangi and Jacumfi coral-reef banks (Fig. 2) form lines parallel to the current shoreline. They rise from the sea floor to a height of 5 m, 1 m of which is exposed at low tide. Exposure to sunlight and dry conditions during low tides may have killed the coral reefs. Most of the reef mass exposed in both of the exposures that were sampled is made up of dead organisms, so that original growth position may account for only a small part of the whole bank and represent the minimum low-water level at the time of their death. A detailed investigation of the coral-reef fauna of the study area was carried out by Kempf & Laborel (1967). They concluded that the most common species are Mussismilia sp., Siderastrea stellata, Millepora alcicornis, Dendropoma sp. and the sessile foraminifer Homotrea, which is very common in open spaces. Raised tidal flats are composed of shell-rich layers 10-30cm in width and are associated with sand and mud. These tidal flats are located to the south of the Galinhos beach and are partly covered by vegetation. The area is protected from direct wave action by a large spit to the north. These tidal flats represent mean sea level with a precision of + 1.0 m. Their fossil content is very similar to that described for the Holocene beachrocks. Peats of mangrove-swamp origin
286
F. H. R. BEZERRA E T AL.
Fig. 7. Beachrock cement phases: (a) acicular crusts of aragonite (plane-polarized light); (b) acicular crusts of (cross-polarized light); (e) micrite calcite cement (cross-polarized light); (d) micrite calcite cement (plane-polarized light); (e) cement dissolution and secondary porosity growth (plane-polarized light); (f) secondary porosity growth over crystalline calcite and a cicular aragonite (cross-polarized light). Q, quartz; Ac, acicular calcite; M, micrite calcite; Cc, crystalline calcite; Sp, secondary porosity.
display a minimum thickness of 1.5m and are composed mainly of wood fragments (tree branches) near the top and mud near the bottom of the deposit (Fig. 8). The upper part of the peats was deposited in the middle to upper foreshore, and the bottom part in the middle to lower foreshore (Martin et al. 1996b). Shells in living position are found mostly in the latter (Fig. 9), which can be used as sea-level indicator with a precision of + 1.0 m. A marine fauna in growth position occurs to the east of S~o Bento (Fig. 2). Its position indicates the minimum height of low-water level before its death.
Sample collection, pre-treatment and radiocarbon dating A group of 25 samples of fossils were selected for dating: 17 samples of shells in beachrock, mostly in facies (a); four samples of shells in peat and tidal flat deposits; three whole-rock samples in coral reef; and one sample of live shell (Table 1). Shells in living conditions were found in a few sites (Catanduba, Rio-do-Fogo, and Recuado). Other shell samples correspond to death assemblages but, as Richards (1982) has shown, wavedeposited fauna can be a better indicator of
COASTAL TECTONICS IN NE BRAZIL
287
Fig. 8. General view of the Rio do Fogo peat in the current intertidal zone. former waterlines than molluscs in growth position, especially if these can tolerate a wide range of depths. Whole-rock samples in beachrock were avoided because they are usually composed of fragments of other rocks, contain fossils of various ages, and may have passed through several stages of recrystallization, overgrowth, neomorphism, and dissolution. Altitude measurements of the various coastal features, including samples, were determined by levelling and were combined with horizontal positioning by global positioning system (GPS). Correction to local standard ports (Macau and Natal) followed procedures recommended by the Admiralty (1996) using tide-table predictions by the Brazilian Navy (1996). The zero level to which all measurements and corrections were made was the Brazilian 'Corrego Alegre' National datum. Radiocarbon dating by first-order assay was carried out at University College London according to the method developed by Vita-Finzi (1983,
Fig. 9. Bivalve shell in living position at Rio do Fogo peat (RF on Fig. 2).
Fig. 10. Electron scanning micrographs of analysed shells showing cross-lamellar structure in uncontaminated aragonite: (a) sample CH2; (b) sample BR2. 1991). Careful pre-treatment procedures were carried out to avoid or reduce contamination. Mechanical cleaning and acid leaching were the main operations used to remove contamination. These techniques were monitored by X-ray diffraction and inspection of acetate peels by light microscopy (Table 1). Samples showing signs of contamination were also analysed by scanning electron microscopy to identify primary and secondary calcite and aragonite (Fig. 10). Samples with clear signs of diagenetic alteration were rejected or subjected to further mechanical cleaning and acid leaching. The errors cited for the ages determined at University College London are based on sample activity, background and modern standard; the
288
F. H. R. B E Z E R R A E T AL.
Table 1. List of 14C ages used in this study Sample source
Height (m asl)
Lab number
Deposit
Species
X-ray
14C age (a BP)
Calibrated age (a Be at 2a)
PG/(1)
+0.60
UCL-423
b
*
A
2700 "+"80
92~O+330 .... -160
MC1/(1)
+1.80
UCL-354
b
*
A
1600+40
11~1+9o ~Jl_10 0
MC2/(1 )
+ 1.80
UCL-418
(live)
*
-
Modern
-
CDB1/(1)
+0.30
UCL-433
tf
*
A
3950 • 1 I0
~o~n+32o ~,/J,290
CDB2/(1)
+0.30
UCL-434
tf
*
C
45004- 130
47m+28o v_340
GA/(1)
+1,10
UCL-416
b
*
A
3550 -4- 100
"~A20+240 .... -220
FSA1/(1)
-0.50
UCL-410
b
*
A
6050 • 190
~aaa+430 .... -390
RECI/(1)
+3.90
UCL-397
b
*
A
5100 i 140
~a~a+290 .... -390
REC2/(1)
+5.40
UCL-393
mt
*
A/tr sC
4050 4- 110
ao7o+320 .... -280
GU/(1)
+0.60
UCL-431
b
*
A
3050 • 90
,~7~o+z30 .... -110
SAL/(1)
+1.20
UCL-417
b
*
A
3950 2 110
aoan+320 .... -290
PB/(1)
+0.10
UCL-420
b
*
A
4500 2 130
4710_350+28~
3750 2 110
~7n+29o -270
RF/(1)
- 0 . 20
UCL-409 .
pd
PJC/(1)
-0.40
UCL-424
cr
t
JC/(1)
+0.50
UCL-413
b
*
VC/(1)
-0.70
UCL-430
b
PR2/(1)
+0.10
UCL-425
cr
PR3/(1) BR1/(1)
-0.50 +2.20
UCL-361 UCL-403
cr b
~ *
.
*
.
A
.
1450 2 40
970_60+9~
A
4950 4- 150
5280_430+320
*
A
6300 2 200
t;'7Rfl+440 ,-,-Jv_450
t
-
1150 2 30
A
950 4- 30 4700 2 140
530_3o+4O AR90+420 .... -320
(;R13+40
. . . . 30
BR2/(1)
+1.80
UCL-404
b
*
A/tr sC
45002 120
A71o+260 .... - 320
GR1/(1)
+0.20
UCL-419
b
*
A
5600 + 170
r
GR2/(I)
0.00
UCL-421
b
*
A
6550 2 210
7n7o+36o .... -500
GR3/(1)
+0.70
UCL-405
b
*
A
5950 2 170
~q7n+370 v-,-,-,_390
CH2/(I)
+1.70
UCL-432
b
*
A/sC
5400 2 170
r- , - - , v
CH1/(1)
+1.50
UCL-414
b
*
A
65502210
70"70+360 .... -490
.
.
.
.
-350
330
P20/(2)
+ 1.0-2.0
--
b
*
--
6067 2 80
~ao0+2oo
P21/(2)
+1.0-2.0
-
b
*
-
6067 2 100
Jq'1'70+23~ v,-,-v_200
-
v ~ - - , , _
190
P14/(2)
+3.00
b
*
-
4737 2 130
aoTn+340 .... -330
P29/(2)
+3.00
b
*
-
4609 2 100
4830_280+2z0
BL/(3)
+1.0(?)
mfs
cement
-
6060 2 80
6480_18o+2OO
S-6/(3)
-3.0 to +4.0
rnfs
shell
-
2340 2 60
,~n+35o ~'~'~ 140
--
venerids Sample source: (1) this study; (2) Oliveira et al. (1990); (3) Silva (1991). Coastal deposit code: b, beachrock; cr, coral reef; pd, peat deposit; rot, marine terrace; td, tidal flat; mfs, marine and fluvial sediments. Species: * fossil assemblage on text; t coral-reef content on text. X-ray code: A, primary aragonite; C, primary calcite; sC, secondary calcite; (tr) mineral in trace quantity.
ages were calibrated using the curve o f Stuiver et al. (1986), and r o u n d e d up or d o w n to the nearest multiple of ten. T h e activity d e t e r m i n e d for a specimen collected live (MC2), 1 c p m above the p r e - b o m b b a c k g r o u n d level o f 7.7 cpm, shows that the residual effect o f the a t m o s p h e r i c b o m b tests swamps any a p p a r e n t reservoir effect with-
in the resolution o f the m e t h o d (C. Vita-Finzi, pers. comm.). F u r t h e r ages d e t e r m i n e d by conventional r a d i o m e t r i c dating in previous studies (Oliveira et al. 1990; Silva 1991), are also presented in Fig. 2 and Table 1 but were not included in the t i m e - d e p t h d i a g r a m because o f the lack o f i n f o r m a t i o n on sample quality.
COASTAL TECTONICS IN NE BRAZIL
Field and chronological correlation of beachrock Some of the coastal deposits which give overlapping ages can also be identified with a single palaeo-shorelines on the basis of field evidence. In the Cunhafi to Natal littoral zone, the Cunha6, Guaraira and Via-Costeira beachrock bodies range in height from 0.5 m below mean sea level to 2.0m above mean sea level and present sedimentary sections similar to the one depicted in Fig. 6b. There is good agreement between their ages (VC, G R 1 - G R 3 and CH1) and those obtained by Oliveira et al. (1990) for the Guaraira beachrock (samples P20 and P21). If the error bounds are extended to 2o., ages VC, GR1 and GR3 overlap at 6300-6350 cal. a BP, which is likely to be the real age of this group. Samples GR2 and CH 1 are 200 years older than the others and are perhaps redeposited death assemblages. It is possible to extend the unit to include the Farol de Sto. Alberto beachrock on the E-W-trending coast (Fig. 2) on the basis of age and field character. Chronological correlation is also possible between the Recuado, Barreta and Jacumfi beachrock bodies, whose ages overlap at 5060-5300 cal. a BP at 2o-. Another possible chronological association is found between the Galinhos and Lagoa-do-Sal beachrock bodies
289
and the Rio-do-Fogo peat, whose ages overlap at 3640-3670 cal. a BP. Age correlation between other samples is less evident. A few coastal deposits present discordant 14C ages, which might reflect several contributing factors. The age difference between CDB1 and CDB2 in the Catanduba tidal flat is likely to reflect the varied provenance of the detrital material. The age of the upper part of the Guaraira beachrock, where sample GR1 was collected, suggests that cementation processes continued for thousands of years. Despite these anomalies, it was possible to establish a local sea-level chronology and to compare it with predicted changes.
Comparisons between coastal chronology and glacio-hydroisostatic predictions: implications for tectonics The sea levels indicated by our findings, with the exception of beachrock facies (b) and tidal flats, do not represent the whole intertidal zone but reflect shells in the lower foreshore and upper shoreface transition zone during Holocene time. From Fig. 11 it can be seen that the great majority of Holocene coastal deposits in the study area are elevated relative to the present sea level. Sea-level
Fig. 11. Samples of the study area set against the Touros sea-level curve of W. R. Peltier (pers. comm., 1997).
290
F. H. R. BEZERRA E T A L .
behaviour, as indicated by beachrock data and other coastal deposits, points to two main phases of sea-level change. The first was a transgression which started at c. 7000cal. aBP and reached its maximum at c. 5500-5000cal. a BP. It was followed by a regressive phase which lasted until the present day. Samples of the same age and subenvironment but at different locations (e.g. BR1 and PB) plot at different heights, indicating sea-level oscillations. These sea-level oscillations are confirmed by the relationship between facies (a) and (b) and by different beachrock-cement phases. Our results also vindicate the data presented by Silva (1991) for the littoral zone near Macau, where a transgressive sequence starting at 7460 (+190/-160)cal. aBp and reaching a maximum at 5330 (+290/--310) cal. aBP is overlain by a regressive sequence of sediments. The data, which predominantly bear on low-water level, plot slightly below or on the Touros curve by W. R. Peltier (pers. comm., 1997), which represents mean sea level. The Touros curve correctly predicts a Holocene highstand in the study area at c. 5500-5,000 cal. a BP. Some sea-level oscillations of less than 2 m are also indicated by samples on and below the Touros curve (see Fig. 11). Some published Holocene coastal features are consistent with the relative sea-level fall predicted
by the Touros curve, which is possibly indicated by the absence of any age higher than 3 m after 4000cal. a BP. At least three generations of an extensive Holocene aeolian sand cover (Perrin & Costa 1982) occur in the current backshore zone of the study area. It is related to a relative sea-level fall in Holocene time which caused dune building by raising intertidal sandflats above the mean high tide level. In the Macau area, Silva (1991) presented an age of 2350 (+350/-140)cal. a BP (sample S-6) for deposits formed by the migration of sands from the shallow platform to the continent, presumably during the main regression phase. The values which do not fit the Touros curve by W. R. Peltier (pers. comm., 1997) may be the product of low-magnitude climatic oscillations, as proposed by Suguio et al. (1985) and Fairbridge (1992). Nevertheless, a recent study by Angulo & Lessa (1997) favours Holocene sealevel curves with few or no oscillations along the Brazilian coast. It may also be that minor changes in wind or currents can account for the anomalies (Damuth & Fairbridge 1970; Isla 1989; Gonzales & Weiler 1994). In some areas, however, a tectonic explanation is more appropriate. On the littoral zone near Silo Bento (Fig. 2), coastal emergence to the east of the Carnaubais fault system is matched
Fig. 12. Geological map of the Sgo Bento littoral zone showing the relationship between the Galinhos beachrock (GA), the Catanduba tidal flat (CDB1 and CBD2) and the Touros curve, and the location of the geoelectric sounding carried out by Caldas et al. (1997).
COASTAL TECTONICS IN NE BRAZIL
291
with geoelectric soundings across the Carnaubais fault system to the southwest of Silo Bento (Fig. 12) by Caldas et al. (1997), which showed a sequence extending from Neocampanian to Holocene time and 120m thick in the Umbuzeiro Graben. The base of the sequence displays a vertical offset of 60m along the Carnaubais fault system which gradually dies out upwards. Holocene activity is consistent with the coseismic faulting described by Takeya et al. (1989) in the area (Fig. 2).
Conclusion
Fig. 13. Detailed geological map of the S~o Bento littoral zone showing the relationship between the Farol de Sto. Alberto beachrock (FSA1), Recuado beachrock (REC 1), marine fauna in growth position (REC2), Guajiru beachrock (GU), and the Touros curve.
by submergence to the west. In the Macau High, the Galinhos beachrock (facies (a)) plots slightly below the Touros curve (mean sea level), indicating no uplift (Fig. 12). Similarly, near the Umbuzeiro Graben, which is directly affected by the Carnaubais fault system, the Farol de Sto. Alberto beachrock and the Catanduba tidal flat do not display any emergence (Fig. 13), in contrast to coastal deposits younger than 4000 cal. a BP east of S~o Bento, where sudden coastal emergence is indicated by a rich marine bivalve fauna in growth position (REC2) dated to 4240-3910 cal. a BP. This unit is continuous along more than 5 km of the littoral zone to the east of the Carnaubais fault system but not to the west of it (Fig. 13). It lies 5.5m above mean sea level, 1.5m above the Recuado beachrock dated to 5600-5290 cal a BP, and 5 m above the Guajiru beachrock dated to 2910-2740 cal a BP (Fig. 6a). From Fig. 11, it can be seen that local emergence by at least 5 m indicated by sample REC2 and GU, occurred during a short period of about 1320 years. Such local emergence points to tectonic movement and is in contrast to the much smoother sea-level regression that occurred after 5500-5000 cal a BP. Additional field evidence for tectonic movement is found at Silo Bento, which caused coastal emergence to the east and submergence to the west of the Carnaubais fault. Tectonic emergence and submergence is in agreement
In this attempt to analyse sea-level changes in the Holocene sedimentary record of NE Brazil, two beachrock facies associated with (a) the lower foreshore to upper shoreface and (b) the middle to lower foreshore were recognized, and beachrock was found to be a useful sea-level indicator in a mesotidal regime. Our sea-level data were found to fit the Touros curve proposed by W. R. Peltier (pers. comm., 1997) and confirmed the presence of the Holocene palaeo-shorelines predicted by him. In addition, however, sea-level oscillations superimposed on the Touros curve might be the result of minor climatic changes or wind-current changes in Holocene time. Tectonic influences can be detected locally, notably near the Carnaubais fault system, where rapid emergence of at least 5 m to the east of Silo Bento occurred at c. 4080-2780 cal. a BP. The significance of such intraplate activity for seismic hazard is clear and needs to be investigated further. We are greatly indebted to C. Vita-Finzi, I. Stewart, R. W. Fairbridge and K. Suguio for useful comments and suggestions. We also thank W. R. Peltier for providing the late Quaternary sea-level prediction for the study area. This research was funded mainly by University College London, and by the PADCT IIUFRN Project (No. 65.91.0366.00). We are very grateful to a CNPq Project (Vales Tect6nicos do Rio Grande do Norte) co-ordinated by Allaoua Saadi, which contributed with financial support for the last field trip.
References ADMIRALTY 1996. Admirahy Tide Tables." The Atlantic. Admiralty Hydrographic Department. The Hydrographer of the Navy, Southampton. ALEXANDERSSON, T. 1972. Mediterranean beachrock cementation: marine precipitation of Mg-calcite. In: STANLEY,D. J. (ed.) The Mediterranean Sea: A Natural Sedimentation Laboratory. Dowden, Hutchinson & Ross, Stroudsburg, PA, 203-223.
292
F. H. R. B E Z E R R A ET AL.
ALMEIDA, F. F. M., HASUI, Y., NEVES, B. B. B. & FUCK, R. A. 1981. Brazilian structural provinces: an introduction. Earth-Science Review, 17, 1-29. AMIEUX, P., BERMIER, P., DALONGEVILLE, R. & DE MEDWECKI, V. 1989. Cathodoluminescence of carbonate-cemented Holocene beachrock from the Togo coastline (West Africa): an approach to early diagenesis. Sedimentary Geology, 65, 261-272. ANGULO, R. J. & LESSA, G. C. 1997. The Brazilian sealevel curves: a critical review with emphasis on the curves from the Paranagu/t and Canan6ia regions. Marine Geology, 140, 141-166. BZIER, J. A. 1985. Diagenesis of Quaternary Bahamian beachrock: petrographic and isotopic evidence. Journal of Sedimentary Petrology, 55, 755-761. BIGARELLA, J. J., SALAMON, R. & MARQUES FLHO, P. L. 1961. M6todo para avaliaqio de construqio marinha. Boletim Paranaense de Geografia, 4-5, 11-115. BOURGEOIS, J. 1980. A transgressive shelf sequence exhibiting hummocky cross stratification: the Cape Sebastian Sandstone (Upper Cretaceous), southern Oregon. Journal of Sedimentary Petrology, 50, 681-702. BRAZILIAN NAVY 1996. Tide Tablesfor the Macau and Natal Ports. Divisio de Hidrografia e Navegagfio (DHN). CALDAS, L. H. O., MEDEIROS, W. E., DANTAS, E. P. & SA, E. F. J. 1997. Caracterizaqio geol6gica e geofisica de uma estrutura na borda leste do graben de Umbuzeiro, Bacia Potiguar/RN, e suas implicaq6es neotect6nicas. Proceedingsof the 5th Brazilian International Geophysical Congress, 2, 885-888. CAMPOS E SILVA, A., SILVA, D. D. & VASCONCELOS, M. D. T. 1964. Informa96es sobre a malacofauna dos beach rocks de Touros e Sio Bento do Norte, Rio Grande do Norte. Arquivos do Instituto de Antropologia, 1, 79-89. CLARK, J. A., FARRELL,W. E. & PELTIER, W. R. 1978. Global changes in post-glacial sea level: a numerical calculation. Quaternary Research, 9, 265-287. COOPER, J. A. G. 1991. Beachrock formation in low latitudes: implications for coastal evolutionary models. Marine Geology, 98, 145-154. DABRIO, C. J. 1982. Sedimentary structures generated on the foreshore by migrating ridge and runnel systems on microtidal and mesotidal coasts of S Spain. Sedimentary Geology, 32, 141-151. DAMUTH, J. E. & FAIRBRIDGE, R. W. 1970. Equatorial Atlantic deep-sea arkosic sands and Ice-Age aridity in tropical South America. Geological Society of America Bulletin, 81, 189-206. DARWIN, C. H. 1841. On a remarkable bar of sandstone on the Pernambuco coast of Brazil. Philosophical Magazine and Journal of Sciences, 19, 257-260. DELIBRIAS, C. & LABOREL, J. 1971. Recent variations of the sea level along the Brazilian coast. Quaternaria, 14, 45-49. DOMINGUEZ, J. M. L., BITTENCOURT, A. C. S. P., LEAO, Z. M. A. N. & AZEVEDO, A. E. G. 1990. Geologia do Quaternfirio Costeiro do estado de Pernambuco. Revista Brasileira de Geocidncias, 20, 208-215.
DUPRE, W. R. 1984. Reconstruction of paleo-wave conditions during the Late Pleistocene from marine terrace deposits, Monterey Bay, California. Marine Geology, 60, 435-454. FAIRBRIDGE, R. W. 1976. Shellfish-eating Preceramic Indians in coastal Brazil. Science, 191, 353-359. 1992. Holocene marine coastal evolution of the United States. In: FLETCHER III, C. H. & WEHMILLER, J. F. (eds) Quaternary Coasts of the United States." Marine and Lacustrine Systems. Society for Sedimentary Geology, Special Publication, 48, 9-19. FERREIRA, J. M., TAKEYA, M. K., COSTA, J. M., MOREIRA, J. A. M., ASSUMP~AO, M., VELOSO, J. A. & PEARCE, R. G. 1987. A continuing intraplate earthquake sequence near Jofio C~mara, Northeastern Brazil-preliminary results. Geophysical Research Letters, 14, 1042-1045. FLEXOR, J. M. & MARTIN, L. 1979. Sur l'utilization des gr6s coquilliers de la r6gion de Salvador (Br&il) dans la r6construction des lignes de rivages holoc6nes. In: SUGUIO, K., FAIRCHILD, T. R., MARTIN, L. & FLEXOR, J. M. (eds) Proceedings 1978 International Symposium on Coastal Evolution in the Quaternary. Silo Paulo, 343-355. GONZALES, M. A. & WE1LER, N. E. 1994. Argentinian Holocene transgression: sidereal ages. Journal of Coastal Research, 10, 621-627. HAYES, M. O. 1979. Barrier island morphology as a function of tidal and wave regime: ln: LEATHERMAN, S. P. (ed.) Barrier Islands. Academic Press, New York, 1 27. HOPLEY, D. 1986. Beachrock as sea-level indicator. In: VAN DE PLASSCHE, O. (ed.) Sea-level Research: A Manual for Collection and Evaluation of Data. Geobooks, Norwich, 15%173. INDEN, R. F. & MOORE, C. H. 1983. Beach environment. In: SCHOLLE, P. A., BEBOUT, D. G. & MOORE, C. H. (eds) Carbonate Depositionanl Environment. American Association of Petroleum Geologists, Memoir, 33, 211-265. ISLA, F. I. 1989. Holocene sea-level fluctuation in the southern hemisphere. Quaternary Science Reviews, 8, 359-368. KEMPF, M. & LABOREL, J. 1967. Formation de vermets et d'algues calcairessur dans les cdtes du Brazil. Revue des Travaux de la Station Maritime d'Endoume, 59. KINDLER, P. & BAIN, R. J. 1993. Submerged Upper Holocene beachrock on San Salvador Island, Bahamas: implications for recent sea-level history. Geologische Rundschau, 82, 241-247. MARTIN, L., FLEXOR, J. M., BITTENCOURT,A. C. S. P. & DOMINGUEZ, J. M. L. 1986a. Neotectonic movements on a passive continental margin, Salvador Region, Brazil. Neotectonics, 1, 87-105. --, MORNER, N. A., FLEXOR, J. M. & SUGUIO, K. 1986b. Fundamentos e reconstrugg.o de antigos niveis marinhos do Quaternfirio. Bulletin of the GeologicalInstitute, University of Sdo Paulo, 4, 1-161. MATOS, R. M. D. 1992. The northeast Brazilian rift system. Tectonics', 11, 766-791.
C O A S T A L T E C T O N I C S IN N E B R A Z I L
-
MAURY, C. J. 1934. Fossil invertebrata from northeastern Brazil. American Museum of Natural History, Bulletin, LXVII, 123-179. MENDON~A, M. I. 1966. O recife de Arenito de Tibau. Arquivos do Instituto de Antropologia, 2, 343-346. MEYERS, J. H. 1986. Marine vadose beachrock cementation by cryptocrystalline magnesian calc i t e - M a u i , Hawaii. Journal of Sedimentary Petrology, 57, 558-570. MOORE, C. H. JR 1971. Beachrock cements, Grand Cayman ISLAND, B. W. I. In: BRICKER, O. P. (ed.), Carbonate Cements. Johns Hopkins University Studies in Geology, 19, 9-12. NEWMAN, W. S., MARCUS, L. F., PARDI, R. R., PACCIONE, J. A. & TOMECEK, S. M. 1980. Eustasy and deformation of the geoid: 1000-6000 radiocarbon years BP. In: MORNER, N.-A. (ed.). Earth Rheology, Isostasy and Eustasy. John Wiley, New York, 555-567. NIMER, E. 1989. Climatologia do Brasil. Instituto Brasileiro de Geografia e Estatistica. Departamento de Recursos Naturals e Ambientais, Rio de Janeiro. OLIVEIRA, M. I. M., BAGNOLI, E., FARIAS, C. C., NOGUEIRA, a . M. B. & SANTIAGO, M. 1990. Consideraq6es sobre a geometria, petrografia, sedimentologia, diag~nese e idades dos beachrocks do Rio Grande do Norte. In: XXXVI Congresso Brasileiro de Geologia, Sociadade Brasileira de Geologia, 2, 621-634. PETTIER, W. R. 1982. Dynamics of the ice age earth. Advances" in Geophysics, 24, 1-146. - - 1 9 8 8 . Lithospheric thickness, Antarctic deglaciation history, and ocean basin discretization effects in a global model of postglacial sea level change: a summary of source of nonuniqueness. Quaternary Research, 29, 93-112. & J1ANG, X. 1996. Glacial isostatic adjustment and Earth rotation: refined constraints on the viscosity of the deepest mantle. Journal qf Geophysical Research, 101, 3269-3290. PERRIN, P. & COSTA,M. I. P. 1982. As dunas litorfineas da regifio de Natal, RN. In: SUGUIO, K., DE MEIS, M. R. M. & TESSLER,M. G. (eds) Proceedingsof the IVth Symposium on the Quaternary of Brazil, Brazilian Geological Society-PETROBRAS-University of Silo Paulo-IPT/CESP-Federal University of Rio de Janeiro, 291 304. RAMSAY, P. J. 1995. 9000 years of sea-level change along the Southern African coastline. Quaternary International, 31, 71-75. READING, H. G. & COLLISON,J. D. 1996. Clastic coast. In: READING, H. G. (ed.) Sedimentary Environmental Processes, Facies and Stratigraphy, 3rd edn. Blackwell Scientific, Oxford, 154-231. -
293
RICHARDS, G. W. 1982. Intertidal molluscs as sea-level indicators." a comparative study of modern and fossil Mediterranean assemblages. PhD thesis, University of London. SEMENUIK, V. & JOHNSON, D. P. 1982. Recent and Pleistocene beach/dune sequence, Western Australia. Sedimentary Geology, 32, 301-328. SHIPP, R. C. 1984. Bedforms and depositional sedimentary structures of a barred nearshore system, eastern Long Island, New York. Marine Geology, 60, 235-259. SILVA,C. G. 1991. Holocene stratigraphy and evolution of the A,cu River Delta, Rio Grande do Norte State, northeastern Brazil. PhD thesis, Duke University, Durham, NC. STODDART, D. R. & CANN, J. R. 1965. Nature and origin of beach rock. Journal of Sedimentary Petrology, 35, 243-273. STRASSER, A., DAVAUD, E. & JEDOUI, Y. 1989. Carbonate cement in Holocene beachrock: example from Bahiret el Bibau, southeastern Tunisia. Sedimentary Geology, 62, 89-100. STUIVER, M., PEARSON, G. W. & BRAZIUNAS,T. 1986. Radiocarbon calibration of marine samples back to 9000 yr BP. Radiocarbon, 28, 980-1021. SuGuIo, K., MARTIN, L., BITTENCOURT, A. C. S. P., DOMINGUEZ, J. M. L., FLEXOR, J. M. & AZEVEDO, A. E. G. 1985. Flutua96es do nivel relativo do mar durante o Quaternario Superior ao longo do litoral Brasileiro e suas impliea96es na sedimentag~o costeira. Revista Brasileira de Geocidncias, 15, 273-286. TAKEYA, M. K., FERREIRA, J. M., PEARCE, R. G., ASSUMP(~AO, M., COSTA, J. M. & SOPHIA, C. M. 1989. The 1986-1989 intraplate earthquake sequence near Jofio Cfimara, northeastern Brazil - evolution of seismicity. Tectonophysics, 167, 117-131. TIETZ, G. & MOLTER, G. 1971. High-magnesian calcite and aragonite cementataion in Recent beachrocks, Fuerteventura, Canary Islands, Spain. In: BRICKER, O. P. (ed.) Carbonate Cements. Johns Hopkins University Studies in Geology, 19, 4-5. VAN ANDEL, T. H. &LABOREL, J. 1964. Recent high relative sea-level stand near Recife, Brazil. Science, 145, 580-581. VITA-FINZI, C. 1983. First-order 14C dating of Holocene molluscs. Earth and Planetary Science Letters, 65, 389 392. - - 1 9 9 1 . First-order 14C dating Mark II. Quaternary Proceedings, 1, 11-17. WALCOTT, R. I. 1972. Past sea levels, eustasy and deformation of the Earth. Quaternary Research, 2, 1-14.
Slip parameters for the Rann of Kachehh, India, 16 June 1819, earthquake, quantified from contemporary accounts ROGER
BILHAM
CIRES, University of Colorado, Boulder, CO 80309-0216, USA (e-mail:
[email protected]) Abstract: The 16 June 1819 Rann of Kachchh earthquake was felt throughout much of
India. Although significant vertical movements of the ground caused flooding of regions near sea level, damming of a distributary of the Indus river, widespread liquefaction, and a local tsunami, the geometry of the fault plane has hitherto remained obscure. Dislocation models based on deformation data gathered 7 and 25 years after the earthquake suggest that a near-surface reverse fault slipped locally more than 11 m, and that rupture extended at least 80 km along strike. Estimates of maximum uplift and depression during the earthquake yield similar, but not identical, solutions to those based solely on a levelling profile across the zone of uplift, known as the Allah Bund. The inferred 50-70~ fault plane beneath the Allah Bund is unfavourably steep for reverse faulting, and its down-dip width (6-10 km) short for slip exceeding 10 m. Some forms oflistric fault geometry are also consistent with the observed surface deformation fields, with increased fault width (15-25km) and similar coseismic slip. A geometric moment magnitude of M = 7.7 + 0.2 is obtained from the inferred slip parameters, assuming uniform along-strike slip, consistent with a magnitude estimated empirically from the intensity distribution. Although a recurrence of the Kachchh earthquake is unlikely soon because of low inferred contractional strain rates in the region, the westward continuation of the Kachchh rift zone could host future ruptures contiguous with the 1819 event, with important consequences for the city of Karachi. The Rann of Kachchh (historically referred to as the Ran, Raun, or Runn of Cuch, Cutch, Kach or Kutch) is an arid region of western India devoid of vegetation, which lies close to sea level. Rann derives from the Persian word Eriyan meaning 'waste'. In the dry season the lowest part of the Rann is covered with a hard crust of evaporites, but during a vigorous monsoon the region is flooded to shallow depth and impassable on foot. The low hills of Kachchh separate the Rann of Kachchh from the Gulf of Kachchh to the south, and because of the Rann's low elevation above sea level, in some maps Kachchh is depicted as an island. The Rann is underlain by rift-like features, which have been mapped offshore beneath the continental shelf with a general east-west trend (Biswas 1989). The collision of India with Asia has initiated a compressional stress regime in a northwesterly direction, very different from that at the time of rifting, and the reactivation of these normal faults in a reverse sense is both anticipated (Khattri 1992) and demonstrated (Chung 1993; Chung & Gao 1995). Several researchers have speculated on substantial shifts in the coastline of N W India since the visit of Alexander the Great (see Burnes 1835; Haig 1894). In Medieval times the Rann is believed to have been connected to the Arabian Sea to a sufficient depth to permit marine access by large boats. The stranded remains of these boats are reported to have been excavated from
time to time and used as firewood by local villagers (Grindlay 1808; cited by Burnes (1834)). On the basis of the burial rate of nearsurface horizons, the sedimentation rate for the past 9000 years has been estimated as 2 mm/year by Gupta (1975). Uplift in the 1819 event created an 80 km long natural dam (the Allah Bund or Dam of God) across the Kori (Korree) branch of the Indus river (known as the Puran, Pharran, or Pooraun), which in 1826 was breached by a flood. The investigation of the 1826 flood resulted in surveys in 1827 and 1828 described by Burnes (1833, 1834), who estimated a 25 km width to the Allah Bund 7 years after the event. Lyell (1853) discussed the deformation in the earthquake including new materials elicited from Burnes and from the notebooks of early travellers. Suess (1904, pp. 43-47) and Wynne (1872) examined the evidence for uplift, but for want of unequivocal numerical data describing the deformation of 1819 concluded that it is unlikely to have occurred, a sentiment repeated as recently as 1976 (Glennie & Evans 1976). Although Baker (1846) measured a profile of the uplifted Bund using survey instruments, his data were omitted from his 1846 publication and were not generally known until rediscovered and printed by Oldham (1898). In 1926 Oldham collated and evaluated the available data, providing for the first time isoseismal maps and estimates of epicentral uplift and subsidence. Oldham (1926) opened
BILHAM, R. 1998. Slip parameters for the Rann of Kachchh, India, 16 June 1819, earthquake, quantified from contemporary accounts. In: STEWART,I. S. & VITA-FINZl,C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 295-319.
296
R. BILHAM
his account with a review of extant documents including a discussion of the provenance of Burnes' several publications. He shows the earliest of these may have been a quarto, lithographed edition, but was apparently unaware of the handwritten account from which subsequent accounts may have been printed, which Burnes presented to the Geological Society of London on 18 December 1833 with the preamble: 'The following pages contain a memoir which has been already lithographed in another shape, but which has been since recast to enable the author to state the facts as they occurred & afterwards draw his conclusions, instead of mixing both together.' The map that accompanied this was that prepared by W. Ballantine in 1831 entitled: 'A Sketch of the Runn and Countries adjacent to illustrate a Memoir on its formation, and the alterations of the Eastern Branch of the Indus by Lieut Burnes Ass. Qr. M r. Genl.' The deformation data are used in the present paper to provide constraints on the mechanism, magnitude and location of the 1819 earthquake. The credibility of the interpretation depends largely on the accuracy of the data, and it has been found necessary to examine some of the original sources cited by Wynne (1872) and by Oldham (1926). The paper first examines the characteristics of the Earth's surface near the epicentre and concludes that the plain was essentially featureless and that no former fault scarp was present. The reliability of the coseismic deformation data, and possible perturbation by postseismic effects and subsequent earthquakes in the period 1844-1846 is next evaluated. The numerical data are interpreted with the assumption that they were generated by a single event with northerly, or northwesterly, directed slip, with confidence intervals estimated from the original observations.
The surface of the Rann of Kachchh before the earthquake of 1819 In 1808 Captain R. M. Grindlay travelled through the Rann of Kachchh and provided qualitative information concerning the preseismic topography of the northern Rann of Kachchh. His path took him through Kori Creek to join the Puran river, which he identified as the easternmost distributary of the Indus, known northward as the East Nara river. In 1808 marine navigation in boats with moderate clearance was possible as far north as the Lallan Puttun dam, the first of several artificial dams encountered across the Puran river starting 24 km north of Sindri. A salt water anchorage existed south of this dam (bund),
north of which fresh water was impounded for irrigation. Grindlay noted in his journal: 'We passed Sindri, and observed several inferior branches leading through the Rann, among which we saw a few straggling men and women. About 20 miles beyond Sindri we reached Aly Bunder at night and came to anchor close to the mound that contains the fresh water' (cited by Burnes (1835)). No significant fault scarp or highland existed between Sindri and the next fort northward, before the earthquakes. Had there been, they would have been selected as sites for dwellings or defensive positions above the monsoon flood levels. In 1808, according to Grindlay (Burnes 1833), the surface of the Rann was covered with shrubs and bushes near the edge of the Rann that extended as far north as Allybunder, 32 km north of the Allah Bund. In a posthumous article, MacMurdo (1839) stated that: 'The Talpooras however erected two dams called the Morabund, which elevating the level draws the water into the Khattee and eastern districts, and another at Alibunder, with the same view. The waters do not know their way to the sea, which meeting no opposition is driven up to the dam of Alibunder. The mouths are gradually filling up with sand in the absence of the freshes [floods] that prevent its accumulation.'
16 June 1819 earthquake At approximately 7p.m., on 16 June 1819, Fort Sindri and masonry buildings in villages within a radius of c. 80 km were destroyed by a violent earthquake (see Appendix). A tsunami from the Arabian Sea surged across the Rann and sand vents in the region were active to a height of 2-3m for 3 days, venting water and gas (Bombay Public Consultations 1820-21; MacMurdo 1823). Reliable eyewitness reports near the epicentre are available only from Bhooj and Anjar, from where the British resident (Captain James MacMurdo) and an army officer (Lt. Colonel Colin Milnes) sent daily dispatches to the Government at Bombay. MacMurdo listed 1543 people killed in the event, mostly in Anjar and Bhooj where 1547 houses were completely destroyed and many more damaged. MacMurdo remarked that 'had the accident occurred in the night time, perhaps one third of the population of the province would have been buried in the ruins of their dwelling-houses'. He observed that damage to masonry structures was minor where these were constructed on rock, but catastrophic where constructed on soils. MacMurdo's meticulous records probably underestimate the total number of fatalities
THE 1819 RANN OF KACHCHH EARTHQUAKE because they fail to record damage in the northern Rann of Kachchh and southern Sind Province, which regions appear to have been close to the epicentre. The fort at Anjar was destroyed, and a similar fate befell the fort at Sindri in the Rann of Kachchh. Sketches of Fort Sindri before and after the earthquake collected by Lyell (1853) were reproduced by Johnston & Kanter (1992), and a view of the decaying ruins in 1869 was reproduced by Wynne (1872). According to testimony elicited by Burnes (1833, 1835), a tsunami flooded the Rann of Kachchh within minutes of the earthquake and survivors were forced to climb to the top of the ruin, from where they were rescued by boat the following morning. Writing from Jooria, Ballantyne, on the northern shore of Kachchh on 17 June related (MacMurdo 1823) that 'the whole town is a complete ruin' and on 18 June described ground fissures symptomatic of catastrophic lateral spreading; 'on examining the different rents, we found them to be of various extent from an inch to a foot in breadth; the depth however, being considerable, being 10, 15 and 20 feet. In some places a gravel-laden soil had been thrown out; in others a wet black earth.' Although the tsunami and subsequent sand venting caused transient flooding of the Rann, and numerous rivers were temporarily active in Kachchh, two permanent changes occurred in the Fort Sindri region: the foundations of the fort and the surrounding Rann for a radius of many kilometres subsided by more than 1 m, and a region 7 km north of the fort, including the bed of the Puran river, was elevated by 3-6m, preventing navigation northward into Sind province for several years. This natural dam became known as the Allah Bund. The details of the uplift and subsidence near Fort Sindri are discussed in following sections. Large public buildings were partly damaged 140 km to the east in Ahmedabad and Surat but damage to residential structures was insufficient to warrant claims for damage to the Government (Bombay Public Consultations 1820-21). At Ahmedabad a fifteenth-century mosque, famous for its shaking minarets (artificial stimulation of resonance in one caused sympathetic resonance of the other) was severely damaged, and its minarets were destroyed. Arguing that they were merely ornamental, the Government declined to rebuild the minarets. The earthquake was scarcely felt in Bombay, but reports were obtained of perceived motion at Kathmandu, Calcutta, the Baluchistan Hills and from Pondicherry south of Madras (Oldham 1926), a felt radius of 1600 km (see Fig. 1). On the basis of
297
Fig. 1. Rann of Kachchh earthquake 1819. Felt locations from Oldham (1926), 1897 isoseismal from Oldham (1899). the felt area, later compilations have assigned a magnitude of 8.3 to the event (see Dunbar et al. 1997), but many of the intensity reports are from secondary sources. Aftershocks occurred with decreasing frequency in the next few months. MacMurdo (1823) reported two to three events a day in June, one a day until August, one every 3 days in September, six in October and three in November.
Earthquakes 1845-1846 Damaging earthquakes in 1844, 1845 and 1846 are inferred from reports listed by Wynne (1872), but the information contained therein is sparse. Their closeness in time and the scant information available on damage led Oldham to suppose that the June 1845 event described by Nelson (1845) corresponded to the true date of the 1844 earthquake described by LeGrand Jacob (1860), who incorrectly ascribed Jeth Sumvut 1901 to 1844. Hindu Jeth Sumvut is June 1845. This interpretation is presumably correct in that a June 1844 event was not reported by Baker (1846), who was in the region from May to October 1844. However, some aspects of Nelson's 1845 letter are enigmatic. The identity of the author and the date of the letter to Nelson are not recorded in the publication, nor in the archives of the Geological Society of London for 15 December 1845, when it was entered into the minutes: 'One of Capt. McMurdo's guides was travelling on foot to him from Bhooj. The day he reached Lakhpat there were shocks of an earthquake, which shook down part of the walls of the fort, and some lives were lost. At the same time as the shock the sea rolled up the Koree (the eastern) mouth of the Indus, overflowing
298
R. BILHAM
the country as far westward as the Goongra river (a distance of 20 English miles), northward as far as a little north of Veyre (40 miles from the mouth of the Koree), and eastward to the Sindree lake. The guide was detained six days (from June 19th to 25th), during which time sixty-six shocks were counted. He then got across to Kotree, of which only a few small buildings on a bit of rising ground remain. Most of the habitants throughout the district must have been swept away, the best houses in Sind being built of sun-dried bricks, and whole villages consisting only of huts made of a few crooked poles and reed huts. The guide travelled 20 miles through water on a camel, the water up to the beast's body. Of Lak nothing was above water but a Fakeer's pole (the flagstaff always erected by the tomb of some holy man); and of Veyre and other villages only the remains of a few houses were to be seen. There are said to be generally two earthquakes every year at Lakhpat. The Sindree Lake of late years became a salt marsh' (letter to Nelson, 1846). Were it not for mention of Lake Sindri and the guide's route from Bhooj westward to Kotri via Lakhpat, the letter could be describing the 1819 event. Had this been written in 1819 the McMurdo referred to in the letter would have been MacMurdo, the British Resident, who was visiting Anjar on the night of the earthquake, and whose name was spelled in both ways in contemporary accounts. LeGrand Jacob (1860) described Lakhpat in November 1851 as an abandoned city of 20 inhabitants (he saw 12 only), 'the greater part of the houses deserted and many fallen down'. However, despite vivid memories of the earthquake of 1819, 35 years before his visit, the inhabitants did not ascribe the ruins of Lakhpat in 1951 to Nelson's alleged 1845 tsunami and earthquake, 6 years before his visit. Moreover, damage to the fort was not evident in 1851: 'The ramparts of the town completely encircling it, are lofty, nearly three miles round, with a parapet of about seven feet, banquette about six; numerous bastions, some boasting a cannon, all loop-holed for musketry and in good order'. Although Kotri in 1851 was a single 'upper roomed cottage' 6 miles north of Lakhpat approached by a circuitous 15 mile passage by camel 'owing to mud', consistent with the description of access to Kotri in Nelson's letter, this description also fits Lakhpat soon after the earthquake of 1819 (Burnes 1833). In contrast, Carless (1838) described the Lakhpat garrison as a thriving community in
1837, suggesting that damage to the fort in the 1840s did indeed occur: 'the walls are defended by numerous bastions, with guns mounted on them of all sorts and sizes. Most of them are so old as to be entirely useless . . . . It is now garrisoned by 50 Arabs and 150 Native soldiers, and contains a population of about 5000 persons, composed principally of merchants and Hindus, who have fled the tyranny of the Amirs.' Of relevance to the current study is that LeGrand Jacob (1860) described apparent uplift of the Allah Bund of 2-3 m in 1819 and widening of the Bund to 7.5km in 1844, with l m of additional uplift at Sunda, a navigational point on the river 5 km south of Sindri: 'Spent all the afternoon with my tent filled by the best informed men in town and port, assembled for me by the karbharee before alluded to, as having been with Burnes in Sind etc., an intelligent man, himself giving information, and helping to get it from me from others. One of the Rao's garrison in Sindree at the time of its destruction in AD 1819 was also present. The following is an abstract of the notes made after much examination and cross examination. The earthquake of Sumvut 1875 (AD 1819) that submerged Sindree, elevated the bed of the river to the height of 2 or 3 yards for the distance of 2 to 3 kos (5-7.5 miles) commencing about 2kos above Sindree: the spot is called Ullah Bund (God's embankment), but the monsoon has worn a water-way through it in an irregular narrow channel; the material being of clay, sand and gravel, this would soon be deepened and widened by any flow of water; the earth was also raised at a place called Sunda. The usual tide only reaches this Sunda, the spring tide now goes over it by a cubit, the earthquake of 1819 raised the ground as to leave the tide there waist high, but in Sumvut 1901 (AD 1844) [the Hindu lunar calendar Jeth Sumvut 1901 is actually June 1845 (see Byramjen 1845)] a series of shocks occurred, that raised the earth still more, so as to leave a cubit (foot and a half) as the greatest depth of water ever found there: these shocks also extended the breadth of the Allah Bund to the extent of 3 kos: before they occurred the usual tide went over the Sunda by about half a foot, but now not at all. At spring tides, however, a boat drawing a cubit of water can with some labour be taken over the Sunda. Pursuing the upward course of the river it is thus described: After passing the Sunda, a pool (Chuch) is reached called Muthar, where
THE 1819 RANN OF KACHCHH EARTHQUAKE the water is waist deep at all times; this is half a kos long; then comes the Ibraham Shah Peer flag-station, where there is nine feet or more, which continues past Sindri until the Ullah Bund is reached; through this Bund, as before explained, an irregular narrow channel continues the stream during the monsoon: at other seasons water terminates at the Ullah Bund so there is only the dry bed of the old river until we reach the 'chuch' called the Bundrejo Duryao, some three kos higher up, where the water is waist high and salt; it lasts for 4 kos, and is terminated by the Suyundwalla Bund above described. The earthquakes of 1844 here referred to I do not remember ever reading or hearing of, yet they are shown to have effected an important change in the earth's surface: the shocks are said to have lasted during a whole month (All Jeth Sumvut 1901), and were so threatening that whilst they lasted the inhabitants feared to sleep in their houses.' Oldham considered the changes in elevation described in LeGrand Jacob's account to be 'very improbable' based on the observed elevation of the foundations at Sindri Fort in later years. Also, in view of omission of mention of an earthquake from Baker's account, it is fairly certain that no second event occurred that may have distorted the deformation field recorded by him and modelled below, so that these later, somewhat uncertain events, are ignored.
Other events in the region The survival of the evidently precariously stable minarets at Ahmedabad for the 400 years preceding the 1819 event indicates that this was the largest earthquake to have occurred in the region in this time interval. Thus, an earthquake in May 1668 which caused 30000 houses to sink into the ground in the Indus Delta, and another of unknown severity near Surat in 1684 (Oldham 1883), were evidently too weak to cause the Ahmedabad minarets to collapse. Anjar, 60 km SE of the Allah Bund, was close to the epicentre of an M = 6.1 earthquake in 1956 in which 109 people were killed (Chung & Gao 1995).
Deformation in the 1819 event Two permanent changes occurred to the Rann of Kachchh in 1819: a basin was created with a total surface area of more than 1000 km 2 (Lake Sindri),
299
and just to its north a region of the Rann was uplifted with a steep scarp facing to the south (the Allah Bund). Permanent changes occurred elsewhere but these are poorly documented. Although Burnes (1833) evidently did not travel as far as the first of the artificial dams north of the Allah Bund, and did not undertake any precise measurements of the profile of the Allah Bund, he offered two consecutive glimpses of the channel cut by the Puran after the flood of 1826. On his first visit (28 March 1827) he was unable to define a precise northern limit to the Bund 'for it extends very far inland, perhaps 16 miles and by gradually sloping to the north, unites with the land, which renders it impossible to define its breadth with correctness ... the present channel through the Allah Bund ... is only one hundred and twenty feet wide, though from fifteen to eighteen feet deep'. His sketch map of the Bund reproduced by Oldham (1926) shows a 20kin long Bund with an acute angle just east of the Puran river (Fig. 2b). When he revisited this on 9 August 1828, he increased his estimate of the length of the Bund, but offered no details of its precise along-strike shape. His accompanying map shows the extended Bund in a very vague way (Fig. 2b). He remarked of the channel in 1828: 'The channel through the Allahbund I found to be wider, with more of the west side washed away, and changed from a sloping declevity, to a perpendicular bank like the eastern shore. I sailed two miles up the river or channel which the flood of 1826 cut through the Allah Bund, and found the water gradually to decrease from two and a half fathoms to as many feet, which I was informed was its depth as high up as Chatitar above the Allibund, and about 20 miles distant.' 'The banks of the channel which it cut through are of clay, and as they are perpendicular, and the river comes directly from the north, without any windings, I can compare it to nothing so correctly as a canal, nor does its breadth, when a little way up, destroy the resemblance, being only sixty six feet. The natural bund, so called, is certainly the most singular effect of the earthquake of 1819. To the eye it does not appear more elevated in one place than another, and being covered with a saline soil, has the appearance of the Runn on all parts'. In 1844 a survey of Sind irrigation by an engineer (Baker 1846) yielded a 61 km long profile along the bed and bank of the Puran and tributaries north of the Allah Bund. Baker described the surface soil of the Allah Bund as 'light and
Fig. 2. (a) The 1819 Kachchh earthquake dammed the Puran river north of a zone of uplift termed the Allah Bund, and submerged the region to its south surrounding the fort at Sindri. On the basis of morphological changes recorded by Survey of India maps, Oldham (1926) suggested that faulting may have extended a further 100 km to the east (dashed). Earthquake epicentres 1941-1996 3.5 < M < 5 from preliminary determination of epicentres (PDE). (b) Topography 1991 (TPC H-8D 1 : 500 000 Pilotage Chart) showing the location and original area of Lake Sindri, and Burnes' two sketches of the location of the Allah Bund. The 1827 sketch was reproduced by Oldham (1926) and the 1831 map by Burnes (1833). (e) Top and left: Satellite Views of the Rann of Kachchh in dry and wet seasons (1965 declassified Corona satellite images). The form of Lake Sindri in November approximates that mapped by Burnes. Right: Baker's 1844 sketch of the Puran River and Allah Bund superimposed on May 1965 image (inset area) reveals that his map of the course of the river north of the Bund was faithfull to approx. 100 m. The Allah Bund shows as a 90 km long, double arc facing the salt wastes of the Rann. Ancient drainages are present on its surface, and an anastomosing delta has formed north of the Bund mapped by Burnes. In places Lake Sindri has eroded the southern 1 km edge of the Bund presumably removing part of its 1819 crest. Lake Sindri beach berms are clearly expressed in the images.
THE 1819 RANN OF KACHCHH EARTHQUAKE
301
Figure 2. (continued) crumbling, and strongly impregnated with salt; at the depth of one and a quarter to two feet it has more consistency, and is mixed with shells such as are now found abundantly on the shores of the lake'. Baker did not extend his measurements south to Sindri. Baker's detailed levelling survey and map were accidentally omitted from his 1846 publication, and (despite an indication by the editor to the contrary) from subsequent volumes of the Bombay Geographical Society, and it was not until 1898 that the map of Baker's survey was discovered and published in India (Oldham 1898). Curiously, in his geological map of Kachchh, Wynne (1872) attributed features north of the Allah Bund to Baker's map which must have been available to him in some form, but his monograph was written apparently unaware of the levelling profile, as he cast doubt on the
authenticity of reported uplift. The section through the Bund is reproduced in Fig. 3 from the Mora Bund to Lake Sindri, projected in this figure as a function of north-south distance. From this figure it is clear that the Lallan Puttun dam separated fresh water from seawater before the 1819 event. Maps in the 100 years following the earthquake show the area of Lake Sindri to have shrunk. In recent topographical sheets (TPC H-8D 1:500 000 Pilotage Chart, Edition 4 1991) the salt flats of Lake Sindri are indistinguishable from the adjoining Rann, and although the Puran is not shown, a small ponded salt flat is apparent north of the Bund (Fig. 2b). It was still a basin in 1874 when a recurrence of the 1826 Indus flood occurred. Wynne (1872) found the fort to be a ruin surrounded by dry land, yet Major Smith (Royal Engineers) was quoted by
302
R. BILHAM
Col. Barton in 1877 (Bombay Gazetteer 1880) as reporting in January 1875 'the channel at the Allah Bund was 70-80feet across, the speed 2.5 feet to 4.5 feet per second', and Lake Sindri appeared as 'a stretch of clear blue water broken only by the ruined tower of Sindri fort.' Burnes' 1827 n-tap (reproduced by Oldham (1926)) showed the Allah Bund as two acute segments, but in his hand-written memoir (Burnes 1833) presented to the Geological Society of London on 15 December 1833, his printed map (inscribed, Ballantine 1831) showed it as a linear 50km zone striking N50W, and stated it to be perhaps 80km long (Fig. 2b). Oldham (1926, p. 99) in his summary map interpreted the Allah Bund as a sigmoidal zone (Fig. 2a) guided by topographical changes evident in the Survey of
India maps issued in 1880, and speculated also, from topographical evidence, that a fault strand striking N40E extends perhaps 150 km to the east (Fig. 2a). No field observations of surface faults that cut recent sediments were reported anywhere in the region, although many of the structural features in the Kachchh region conform to surface morphology reminiscent of 'Jura topography' (Wynne 1872), suggesting recent tectonic compression. Oldham (1926) summarized the accounts of Baker and Burnes as follows. The peak elevation of the Bund was c. 6.2m with a c. 600m wide scarp dipping at 0.65 ~ to the south and 0.052 ~ to the north. Estimates of the width of the northern scarp vary from 3-12kin (Baker) to 24kin (Burnes). Subsidence south of the scarp attained
Fig. 3. Baker's 1844 profile of bank and bed levels projected on a north-south section, and close-up view of section through the Allah Bund. Artificial dams are shown as vertical lines. Sea level is inferred from Grindlay's 1808 observation that the Lallan Puttun dam separated sea-level navigation from freshwater portage. Sindri lake level approximates high-tide level. The 1826 flood would have ponded only to the level indicated if the raised bed of the Puran continued 4 m below bank level through the Allah Bund.
THE 1819 RANN OF KACHCHH EARTHQUAKE depths of 3.5 m (although where Baker measured this depth is unclear), and extended with diminishing amplitude to 24 km south of the scarp. Fort Sindri is estimated to have subsided at least 1 m probably 1.6 m, and possibly >3.3 m. One of the uncertainties in these estimates concerns the undeformed surface level of the Bund before the earthquake, because it is to this level that estimates of coseismic deformation must be referred. As discussed above, there is little doubt that significant surface morphology was formed entirely by coseismic deformation associated with the 1819 sequence of earthquakes. Oldham (1926, p. 23) argued that if the observations are referred to a datum just north of the Allah Bund taken from Baker's map, absolute uplift may have been 1 m less, and absolute subsidence 1 m more, than the maximum values estimated from local datum levels, e.g. the bed of the Puran or the level of Lake Sindri, respectively. This, however, does not take into account the preearthquake seaward gradient of the land surface. The smooth surface of the bank of the Puran, mapped in Baker's profile, provides a surface whose extrapolation permits the absolute amplitude of uplift beneath the Allah Bund to be estimated. Baker's measurement datum was the level of water dammed behind the Mora Bund, but no precise estimate of sea-level elevation is provided. The absolute level of Baker's datum above sea level can be estimated to c. +0.3 m because it was possible to navigate to the base of the Lallan Puttun Dam before the earthquake, and because the level of Lake Sindri was replenished by high tides after the earthquake. With this assumption, two approximations to the pre-seismic land surface beneath the Allah Bund are possible: in one, a smooth curve is fitted to the bank of the Puran between 17 and 50 km north of the Allah Bund and extrapolated beneath the Bund, and in the other the curve is, in addition, constrained to fit to lowest estimated sea level at a distance of 50 km south of the Bund (Fig. 4). Using the first approximation the peak elevation c. 1 km north of the Bund increases from 6.2 to 6.6 m, and using the second it reduces to 6.1 m. The goodness of the fit to Baker's river bank data is superior (see residuals in Fig. 4) if the sea level constraint is ignored, and because it is by no means certain that the bank surface should asymptotically approach sea level, the higher estimate is thus considered more reliable. Data used in subsequent models are summarized in Table 1. The recorded region of maximum subsidence occurs along the Puran river and in the Sindri region along a well-travelled trade route. Numerical data for the Allah Bund are obtained only
303
where navigation was impeded and where a clear uplift profile was visibly manifest after the 1826 flood. Few roads exist to the east and yet fewer roads to the west that might have been explored after the event, and our knowledge of deformation is biased by this historical circumstance. Neither Baker nor Burnes travelled the length of the Bund to establish its lateral extent. In 1827 Burnes estimated its width as +16 miles (c. 50 km), but in 1828 revised it on the basis of travellers' accounts of newly necessary, circuitous routes around Lake Sindri, to 18 miles west to Ghari, and 24 miles east to Pacham Island (c. 80 km). Baker indicated its length would be too difficult to survey because of the absence of drinking water. The Survey of India maps later in the century were used by Oldham to confirm that the Bund was at least 80 km long, and that morphological features suggest faulting for more than 150 km. The southerly facing scarp width is likely to have been underestimated in 1826, as part of it was submerged, and in 1844, because it may then have been covered partly by sediments. Thus the south facing scarp could have been greater than the 600 m width estimated by Burnes, but because it determines only the closest approach of the subsurface fault to the surface, and has minor influence on deep slip parameters, its true width is of little consequence in the following analysis. The southern extent of subsidence is perhaps the most clearly defined because this formed a freshwater lake that eventually became saline and finally dried up. Unfortunately, because a deep channel existed through the lake, some of the depths in subsequent descriptions relate the channel depth and lake depth in ways that do not permit true bathymetry to be evaluated precisely. When Burnes visited this after the 1926 flood the main river channel was fresh, as was the surrounding water in Lake Sindri. In following years the Rann shallowed, and although much of this may have been due to sedimentation, it is possible that post-seismic deformation occurred (Oldham 1926). In 1827 the width of the channel though the Bund was 40 m wide but by 1828 the flow had ceased and the waters of the Rann were saline. Deformation tapers to low values near the northern and southern limits of rupture, but no deformation is apparent >24 km from the Allah Bund (Table 1). Slip on the fault is modelled as uniform slip in an elastic half-space using the formulation of Okada (1985). The procedure adopted is to compare the surface deformation arising from one or more subsurface dislocations with the observed surface deformation, and to reject those whose theoretical geometries do not result in satisfactory
304
R. B I L H A M
Fig. 4. Baker's 1844 profile, projected on a north-south line, with exponential approximations to the slope for the river bank (above), and residual elevations when these are subtracted from the observed data (below). A better fit to Baker's river-bank data is obtained if assumptions concerning morphological relations to inferred sea level are ignored. agreement. By assuming two-dimensional (2D) u n i f o r m slip in a northerly direction, for example, five u n k n o w n s remain to be determined: fault dip, slip, latitude, and the depth to the top and b o t t o m of the rupture. In practice, the latitude and depth of the rupture are determined to first order by the width and location of the section t h r o u g h the
Allah Bund. That is, the a p p r o x i m a t e ant• metry of the vertical d e f o r m a t i o n field requires an almost vertical fault whose surface extension must cut the Earth's surface at the s o u t h e r n m o s t expression of the Allah Bund. The half-width o f the s o u t h w a r d slope of the Allah Bund is approximately equal to the depth to the u p p e r
Table 1. Deformation estimates for north-south section through the Allah Bund Parameter
Maximum
Minimum
Model input
Northerly extent of uplift <0.1 m Maximum uplift Location of maximum uplift Ant• null Location of maximum subsidence Maximum subsidence Subsidence at Sindri Location of Fort Sindri Southerly extent of subsidence > -0.1 m
+24 km +6.6 m + 1 km 0 - 1 km -4.5 m -3.5m - 6 km -40 km
+6 km +6.1 m +0.6 km 0 -200 m -2.5 m -1.5m - 5 km -24 km
6• 6.34-0.3 m 14- 0.2 kill -3.5i0.3m -1.5• -6• -24•
Distances are measured relative to the inferred deformation-null separating uplift from subsidence. Uncertainties indicated as model input are used to estimate confidence levels of solutions.
THE 1819 RANN OF KACHCHH EARTHQUAKE
305
in Fig. 6. Assuming the values listed in Table 1 the model misfits in Fig. 6 are consistently lower for down-dip widths of less than 10 km than for larger fault widths. Acceptable combinations of dip and coseismic slip for a down-dip width of 10 km are 11.5 4- 1 m and 68 ~ + 7~ respectively (lo-, Fig. 7a).
6 5 .4 4 0
"" 3
Planar dislocations constrained by the Allah Bund profile
E 2
..~ 1 .o
~0
I
50
I
I
I
I
I
I
60 70 80 dip of fault in degrees
I
90
Fig. 5. The ratio of maximum uplift to maximum subsidence determines the dip of the fault in an elastic half-space. A 1.8 ratio of uplift to subsidence favours a dip close to 70~ largely independent of down-dip parameters (top and bottom of down-dip width indicated for two different solutions) surface of the dislocation. In addition, simple numerical tests (Fig. 5) show that a ratio of uplift to subsidence of 6:3.3 requires the dip on the causal rupture to be to the north at 65-70 ~ requiring reverse slip on the fault.
Planar dislocations constrained by maximum and minimum vertical deformation In principle, having assumed the location and depth of the dislocation by inspection, only three observations are required to constrain the dip, slip and down-dip width of the Kachchh rupture. Five data are available in Table 1 in addition to the continuous but incomplete spatial coverage from Baker's levelling data. Forward models were developed to estimate the sensitivity of the interpretation to each of the available data. The first suite of models (Fig. 6) emulates the subsidence at Sindri (+0.3m), maximum uplift north of the rupture (+0.3 m), maximum subsidence south of the rupture (+0.3 m), and far-field constraints of uplift and subsidence less than 2 0 + 10cm at distances + 2 4 k m from the Bund. These models ignore the location of maximum uplift and subsidence, and instead use only their amplitudes as constraints. Least-squares misfits between observations and model results for a range of possible slips and down-dip depths are estimated and contoured in terms of l - 3 a confidence intervals
In models illustrated in Figs 6 and 7 Baker's profile has not been used to constrain the 1819 rupture parameters. Baker's map view of the Bund (reproduced from a tracing of the original by Oldham (1898)) shows the section (lower expanded view in Fig. 3) to have been taken at N30E approximately normal to the strike of the Bund. Although Baker's numerical data are presumably more precise than the estimates for subsidence and uplift listed in Table 1 the profile has some puzzling characteristics. The transition between the almost linear northern slope and the undeformed surface of the desert (6 km from the southern edge of the Allah Bund) is too abrupt to be caused by elastic deformation. If the shallowing of the bed of the Puran is used to estimate the northern limit of deformation, the width of the uplifted Allah Bund might be estimated to extend perhaps 4 km further north. Baker suggested that the channel may have filled by bank collapse at the mouth of the incised cut through the Bund but not to the north. A possible reason for the abrupt transition between the Bund and the apparently undisturbed Rann north of the Bund may be erosion of the bank caused by drainage of the impounded waters in 1926. A further problem with the data concerns the surprisingly linear northerly dip to the deformation field, again atypical of elastic deformation. Notwithstanding these perceived problems with the levelling data, slip parameters were estimated from Baker's sectional profile of the Allah Bund by comparing the slope at 0.5km intervals with the theoretical slope estimated from a dislocation model. An observational uncertainty of 0.2 m per 0.5 km was assigned to these slope data, limited mainly by digitizing errors from Oldham (1898). Appropriate models require a down-dip width of 5 + 1 km, dipping at 504-5" with 1 2 4 - l m of slip (Fig. 7b). The shallower dips for the Bund profile result from the model's attempt to fit the steep northerly slope, in addition to the subsidence evident between 6 and 8 km north of the southern edge of the Bund (Fig. 4). A suite of models in which data were examined only from the southern
306
R. BILHAM
Fig. 6. Misfit contours for solutions for slip and down-dip width for four alternative northerly dips to the inferred rupture zone using constraints listed in Table 1. A down-dip width of c. 10km is favoured by the data, with a dip of at least 60~
5.5 km of the uplifted Bund favoured similar slip parameters. The models favoured by the Bund data evidently give shallower down-dip widths than the maximum-minimum deformation field used in Figs 6 and 7a, and are inconsistent with the reported subsidence in Lake Sindri. Subsidence at Fort Sindri is required to be less than 1 m, and maximum subsidence is preferred to be less than 2.2m values lower than those listed in Table 1.
Listric fault models The above models use planar dislocations with uniform slip. More realistic slip distributions and more complex geometries can be proposed that also fit the data, but their exploration would be conjectural in the absence of additional constraints. Because > 1 0 m of slip is large for a fault with < 1 0 k m down-dip width, a search for faults with longer down-dip width but similar surface deformation is of utility. A class of listric faults was examined with a 2D subsurface geometry of the form d = a + b e -Cx, where d is depth and x is distance from the southern edge of the Bund, and a, b and c are constants
chosen to best fit the observed surface deformation field. In the examples shown the listric rupture surface was approximated by ten short planar fault segments. These subsurface faults have steep near-surface dip and shallow dip at depth. Two geometries of several that approximate the data (but which predict a broader Bund profile) are shown in Fig. 7c. The total down-dip length to the faults increases by a factor of 2-3 in these models to 15-23 km. A detailed examination of listric fault models was considered unwarranted in the absence of constraints from horizontal deformation.
Location and Magnitude of the 1819 earthquake No accurate estimates for the epicentre for the 1819 earthquake have hitherto been proposed, although Chung & Gao (1995) attributed inappropriate locations south of the Allah Bund to Quittmeyer & Jacob (1979) and Chandra (1977), on the basis of those researchers' approximated coordinates. The above analytical solutions suggest that the 1819 rupture occurred 5-15 km north
THE 1819 RANN OF K A C H C H H EARTHQUAKE
307
Fig. 7. (a) Contours showing a range of slip and dip solutions that fit point data from Table 1, assuming a 10 km deep dipping planar dislocation. Best fitting solution shown left. (b) Best fitting solutions for a 5 km wide planar dislocation using only Baker's levelling profile (enlarged right) as a constraint. (c) Surface deformation arising from listric faulting. Listric faulting results in similar subsidence, much increased down-dip fault width, but broader uplift than that associated with planar faults.
or northeast of the Allah Bund. The longitude is not determined by the available data to better than 1~ The along-strike length of the Allah Bund was estimated by Oldham (1926) as 80-150 km, from which a geometric seismic moment can be estimated for the rupture. Because the eastward extension of the fault and its sense of slip is unconstrained by the historical record, it is ignored in the following estimates for magnitude. Using the relation M o = # x slip x L W, where M w = 2 / 3 ( l o g M o ) + 1 0 . 6 , the range of parameters
determined above corresponds to a local magnitude of MI, = 7.7 + 0.2 using typical values for the rigidity modulus #, and a range of down-dip widths of 10-20 km. An alternative method to estimate the magnitude is to use the intensity data reported for the event. An empirical relation between intensity and area of shaking constrained by six Indian events including the Anjar (1956) event was discussed by Johnston (1996), who offered a magnitude of M = 7.5-8 for the 1819 event (7.8 by Johnston & Kanter (1992)). A slightly lower
308
R. BILHAM
Table 2. Isoseismal areas and estimated moment magnitudes for the 1819 earthquake
Intensity
a b c Radius (km) Mo F94C (logl0(dyncm)) M
Felt
VIII
17.3 0.959 0.00126 1600 27.23 7.4
24.1 0.44 0.00586 140 27.66 7.7
Constants a, b, and c are preferred values from Johnston (1996). intensity magnitude can be estimated from the data of Fig. 1, where significant attenuation of intensity northward is evident, causing the isoseismals to be non-circular, especially for lower intensities. Moment magnitude, Mo, of an earthquake in the F94 model is related to the area, S enclosed within a specified isoseismal intensity contour, by an expression of the form logMo=a+blogS+cv/S
(1)
where the constants a, b and c are determined empirically for each isoseismal area. Intensity magnitudes are shown in Table 2, although the intensity data from which they are derived are sparse and of uncertain quality. A mean magnitude estimated from the intensity data is M = 7.5 4- 0.2 in reasonable agreement with the deformation data. Combining the intensity and deformation data, a preferred magnitude of M = 7.7 + 0 . 2 is assumed for the 1819 event. A more careful evaluation of authentic felt reports is needed to improve significantly on the mapped isoseismal estimation of magnitude. Discussion
Although several forms of subsurface slip geometry are admitted by the data, they share several features. Maximum and minimum vertical deformation values estimated 6 years after the earthquake yield preferred solutions for a 67 4-5 ~ dipping fault, with a down-dip width of 6-10 km, and the short profile across the Bund measured 25 years after the earthquake favours a dislocation with 50 + 5 ~ northerly dip, 12 + 1 m of slip and a down-dip width of 5-6 km. Listric faulting with dip shallowing to the NE is an alternative subsurface geometry that requires c. 11 m of slip on a steeply NE dipping nearsurface fault. As noted above, it is assumed that slip vector was at N30E, normal to the Bund, and
consistent with both the NE directed IndoAsian plate convergence vector (Paul et al. 1995) and the p-axes of regional earthquakes (Chung 1993; Chung & Gao 1995). A N45E slip vector would require steeper dips, and shorter downdip widths. Hence, the estimated dips are lower bounds, and a steep fault plane is a necessary, common feature of any interpretation of the data near the Allah Bund. At dips of 50-59 ~ 'Byerlee' friction causes a fault to 'lock' in response to horizontal compression, unless significant fluid overpressuring is available to reduce friction on the fault (Sibson & Xie 1998). The geometry of the Kachchh rupture is thus severely misoriented for reverse slip, and would have required fluid overpressuring to promote rupture. Fluid overpressuring is believed to be widespread in the lower crust (Sibson 1992). Reservoir-induced seismicity throughout India suggests that fluid pressures play an important role in triggering shallow seismicity, and it is possible that this may be a common slip mechanism for the Indian subcontinent. The thick sediments in the Indus fan and the Rann of Kachchh, moreover, have favourable conditions for overpressuring. A consequence of probable northeasterly directed slip is that Oldham's inferred fault east of the Allah Bund, if it indeed exists, could absorb a large left-lateral strike-slip component with insignificant convergence, and thus minor vertical deformation (Fig. 8). Oldham (1926) suggested that post-seismic surface changes occurred in the Sindri region that were not entirely the result of silting or precipitation of evaporites. The various sketches of Fort Sindri show it to have been initially surrounded by water close to the high-tide or monsoon-surge level, and a few decades later to have been surrounded by dry land. Oldham suggested that this change was caused by a relaxation of coseismic subsidence. The wavelength of the vertical changes is considered too short to result from viscoelastic adjustment of the elastic crust. However, J. N. Brune (pers. comm., 1997) has demonstrated in computer simulations and foam rubber models that dynamic effects associated with propagating wrinkles along the fault plane (Brune et al. 1993; Andrews & Ben Zion 1997) can cause over-shoot or undershoot during rupture, which may differ substantially from the static-frictionless deformation of models examined in this paper. Presumably, aftershocks, afterslip or post-seismic creep would bring surface deformation closer to the static deformation field. However, if this were the case in the Sindri region, and initial subsidence at Fort Sindri were an artefact of dynamic rupture, we would expect
THE 1819 RANN OF KACHCHH EARTHQUAKE
reverseslip/
local
I
50km Fig. 8. Simplified sketch of inferred NE-directed convergence during the 1819 earthquake. Steep reverse slip causes uplift NE of the Allah Bund (barbs). The existence and style of deformation east of the Bund is speculative, because contemporary field observations are absent.
that relaxation of the footwall would be evident also in relaxation of the hanging wall measured by Baker. The data of both Baker and Burnes were obtained from isolated samples of a feature whose along-strike length and surface geometry renders approximate any simple elastic deformation model. The amplitude of slip required in each solution is considerable for along-strike dimensions of 80 km, hence Oldham's somewhat weak evidence for 150 km of along-strike slip is consistent with this aspect of the data. In addition, a down-dip width of 6kin is unexpectedly small to permit > 10 m of surface rupture. Had this occurred, mean dilatational extension along each side of the fault plane would have exceeded 1000# strain, with a correspondingly high stress drop. The modelling presented here is insensitive to along-strike slip, and to variations of slip along strike. It is curious that the impressive dip-slip component resulted in no surface fault scarp, as it appears to have reached at least to within a few hundred metres of the surface. Presumably, for this to occur, the surface alluvium would have to have been draped over the rupture in the near surface. It is possible also that a fault scarp, or several scarps and fissures, did occur along parts of the Bund, the details of which were not related in second-hand accounts of the event. A curious feature of the region is the absence of a pronounced physiographical feature along the Bund (the Bund is a mound, not a mountain). This suggests that recurrence intervals are low, or that reverse slip is a relatively recent process for the Kachchh fault, which, like nearby faults associated with the Kachchh rift system, is currently being reactivated in a reverse sense. The recurrence of earthquakes in the Kachchh region would appear to be accessible to palaeoseismic techniques, and several issues associated with the 1819 event are worthy of field investigation. For example, the 1826 flood
309
will have deposited fresh-water sediments above the salt deposits on the floor of Lake Sindri, providing a measure of the current form of the subsidence basin. Investigations of ponding south of the Bund, along the Rann to the west and far east of the mapped expression of the 1819 Bund, would clarify the along-strike length of faulting, and its potential sinistral component. Investigations of ponding north of the Bund might also reveal the transient strand line of the 1826 flood. Surface studies of the eastern expression of the fault might reveal evidence for left lateral slip, although many of the drainages across the scarp would have been initiated only after the earthquake. The rate of secular strain contraction of India is not known in the Kachchh region but is immeasurably small (<0.1 # strain per century) on the Indian Peninsula (Paul et al. 1995). If similar rates prevail in the Kachchh region, the strain contraction released by the Kachchh event would require c. 100000 years for its renewal, an interval long compared with the time needed to erode the 1819 scarp, a problem common to Peninsular India earthquakes (Rajendran et al. 1996). It is therefore unlikely that future large events will recur soon near the epicentre of the 1819 Kachchh event. However, it is possible that the 1819 event may have broken a section of a larger fault system. The epicentres of large historical earthquakes are not well known. In 1534 a tsunami in the Gulf of Cambay (Logan 1887, p. 322) was reported by the crew of Vasco da Gama's fleet when it was anchored offshore at Dabul. Hobson-Jobson (Yule & Burnell 1903) located Dabul (Dabhul or Dabyl) between Goa and Bombay at 17~ a location shown on early European maps of navigation around the Indian coast. If this were near the epicentre, the event would be too far south to have been an Indus Delta event. A severe earthquake in the Indus Delta, to the east of Kachchh, occurred in 1668, in which 15 000 houses reportedly sank into the ground (Oldham 1883). Cities along the northern Arabian Sea are much larger now than when historical earthquakes visited nearby geographical regions. In particular, Karachi, less than 200km from Kachchh, now hosts a population of more than 12 million people. An M > 7 earthquake within 50 km of Karachi cannot be excluded.
Conclusions Solutions derived from subsets of data obtained 6 years and 25 years after the earthquake share coseismic reverse slip exceeding 11 m on a
310
R. B I L H A M
northeasterly dipping fault, terminating in the shallow subsurface. The a m o u n t of slip is large for inferred planar dislocations with d o w n dip widths of 5-10 km, a l t h o u g h longer d o w n - d i p widths are permitted by forms oflistric faulting. It is assumed that slip was oriented N E , n o r m a l to the strike of the Allah Bund, approximately parallel to the I n d o - A s i a n convergence direction, and to the p-axes of recent earthquakes in the region. The details of the d e f o r m a t i o n field across the uplifted Bund are inconsistent with simple elastic d e f o r m a t i o n , and it is considered possible that the d e f o r m a t i o n near Sindri was amplified by d y n a m i c processes during rupture. The geometric m o m e n t m a g n i t u d e of the event is estimated as M - - 7.7 + 0.2, similar to a magnitude of M = 7.5 -4- 0.4 obtained from the intensity distribution. Reverse slip on a fault dipping greater than 50 ~ requires fluid overpressuring of the rupture plane for slip to occur. A l t h o u g h this is not c o m m o n in the near surface, it m a y be ubiquitous at midcrustal depths in India, where artificial reservoirs frequently induce local earthquakes. Moreover, the R a n n of K a c h c h h and Indus Delta are clearly locations of sediment c o m p a c t i o n where fluid overpressuring w o u l d not be unexpected. Because e a r t h q u a k e recurrence intervals m a y exceed m a n y thousands of years it is likely that future large earthquakes near the A r a b i a n coastline o f India will occur in intervening regions, and not near the epicentral regions of recent m o d e r a t e events. That reverse slip can occur on a > 8 0 k m long rupture within 200 k m of K a r a c h i has i m p o r t a n t consequences for seismic hazards in that city, in view of the probable westward extension of the causal fault system. I am indebted to P. Bodin, A. Yin, and librarians at the India Office, Royal Geographical Society and Geologicalal Society of London, who assisted in the location of the references cited. I thank N. Ambraseys for additional materials and insightful comments, and R. Sibson, who has influenced my conclusions concerning the importance of fluid overpressuring The research was funded by the National Science Foundation.
References ANDREWS, D. J. & BEN ZION, Y. 1997. Wrinkle-like slip pulse on a fault between different materials. Journal of Geophysical Research, 102, 553-571. BAKER,W. E. 1846. Remarks on the Alla Bund and on the drainage of the Eastern part of the Sind Basin. Bombay Geographical Society Transactions, 7, 186-188. BlSWAS, S. K. 1987. Hydrocarbon exploration in western offshore basins of India. In: Recent
Geoscientific Studies in the Arabian Sea off India. Geological Survey of India Special Publication, 24, 185 194. Bombay Commercial Calendar and General Directory jor the year 1845. J. Byramjen, Bombay. Bombay Gazetteer 1880. Gazetter of the Bombay Presidency, 5. Cutch, Palanpur and Mahi Khanti, Bombay. BOMBAY PUBLIC CONSULTATIONS 1820-21. Exams Office Nov. 1820, From Boards Collections 15426-15427, Vol. 620 Record Department, India Office Library, F/4/620. BRUNE, J. N., BROWN, S. & JOHNSON, P. A. 1993. Rupture mechanism and interface separation in foam rubber model earthquakes: a possible solution to the heat flow paradox and the paradox of large overthrusts. Tectonophysics, 218, 59-67. BURNES, A. 1833. A memoir of the eastern Branch of the Indus giving an account of the alterrations pro-duced in it by the earthquake of 1819 and the bursting of the dams in 1826; also a theory of the Runns formation and some surmises in the route of Alexander the Great, Presented by the Honble the Governer in Council to the Lit. Soc. of Bombay, Camp at Lucput, 28 March 1827, 13 August 1828. Handwritten Manuscript presented by the Author to the Geological Society of London, 18 December 1833. With map inscribed W. Ballantine 1831. - - 1 8 3 4 . Travels into Bokhara. J. Murray, London, 3 vols. 1835. A memoir on the Eastern Branch of the River Indus giving an account of the alterations produced by it by an earthquake in 1819, also a theory of the Runn, and some conjectures on the Route of Alexander the Great, drawn up in the years 1827 28. Royal Asiatic Society, 3, 550-588 (dated 1833). CARLESS,T. G., 1838. On the Delta of the Indus. Royal Geographical Society Journal, 8, 361 366. CHANDRA, U. 1977. Earthquakes of Peninsula India a seismotectonic study. Bulletin of the Seismological Society of America, 67, 138%1413. CHUNG, W.-V. 1993. Source parameters of two riftassociated intraplate earthquakes in peninsular India: the Bhadrachalam earthquake of April 13 1969, and the Broach earthquake of March 23 1970. Tectonophysics, 225, 219-230. & GAO, H. 1995. Source mechanism of the Anjar, India, earthquake of 21 July 1956 and its seismotectonic implications for the Kutch rift basin. Tectonophysics, 242, 281-292. DUNBAR, P. K., LOCKRIDGE, P. A. & WHITESIDE, L. S. 1997. Catalog of Significant Earthquakes 2150 B.C.-1991 A.D. (with addendum through 1997) Including Quantitative Casualties and Damage. National Geophysical Data Center, Boulder, CO. GLENNIE, K. W. & EVANS, G. 1976. A reconnaissance of the Recent sediments of the Ranns of Kutch, India. Sedimentology, 23, 625-647. GUPTA, S. K. 1975. Silting of the Rann of Kutch during the Holocene. Indian Journal of Earth Science, 2, 163-175.
T H E 1819 R A N N OF K A C H C H H E A R T H Q U A K E HA1G, M. R. 1894. The Indus Delta Country, a Memoir
Chiefly on its Ancient Geography and History. Kegan Paul, Trench Trtibner & Co., London. JOHNSTON, A. C. 1996. Seismic moment assessment of earthquakes in stable continental r e g i o n s - 2 . Historical seismicity. Geophysical Journal Inter-
national. & KANTER, L. R. 1992. Stable continental earthquakes, Scientific American, 262, 68-75. KHATTRI, K. N. 1994. A hypothesis for the origin of Peninsular seismicity. Current Science, 67, 590-597. LEGRAND JACOB, G. 1860. Extracts from a journal I kept during a tour made in 1851 through Kutch... Transactions of the Bombay Geographical Society Society, 15, 56-66. LOGAN, W. 1887. Malabar. Reprint, Asian Educational Series, Madras, 1995. LYELL, C. 1853. Principles of Geology, l lth ed. Appleton & Co., London. MACMURDO, J. 1823. Papers relating to the earthquake which occurred in India in 1819. Literary Society of Bombay Transactions, 3, 90-116. MACMURDO, J. 1839. Observations of the Sindhoo, or River Indus, Bombay Geographical Society Proceedings, 5, 124. NELSON, R. E. 1845. Notice of an earthquake and probable subsidence of the land in the district of Kutch, near the mouth of the Koree, or eastern branch of the Indus, in June 1845. Geologieal Society, London, Quarterly Journal, ii, 103. OKADA, Y. 1985. Surface deformation due to shear and tensile faults in a half space. Bulletin of the Seismological Society of America, 75, 1135-1154. OLDHAM, R. D. 1898. A note on the Allah Bund in the north west of the Runn of Cuch. Geological Survey of India Memoirs, 28, 27-30. - - 1 8 9 9 . Report on the Great Earthquake of 12 June 1897. Geological Survey of India Memoirs, 29. - - 1 9 2 6 . The Cutch (Kachh) Earthquake of 16th June 1819 with a revision of the great earthquake of 12th June 1897. Geological Survey of India Memoirs, 46, 71-146. OLDHAM, T. 1883. Catalog of Indian earthquakes, Geological Survey of India Memoirs, 19, 163-215. PAUL, J. et al. 1995. Microstrain stability of Peninsula India 1864-1994. Indian Academy of Science (Earth and Planetary Science) Proceedings, 104(1 ). QUITTMEYER, R. C. & JACOB, K. H. 1979. Historical and modern seismicity of Pakistan, Afghanistan, northern India and southern Iran. Bulletin of the Seismological Society of America, 69, 773-823. RANJENDRAN, C. P., RAJENDRAN,K. & JOHN, B. 1996. The 1993 Killari (Latur) earthquake: an example of fault reactivation in the Precambrian crust. Geology, 24, 651-654. SmsoN, R. H. 1992. Implications of fault-valve behaviour for rupture nucleation and recurrence. Tectonophysics, 211, 283-293. & XIE, G. 1998. Dip range for intracontinental reverse fault ruptures: the truth not stranger than friction. Bulletin of the Seismological Society of America, 88, 1014-1022. -
-
311
SUESS, E. 1904. The face oJ the Earth. Translated by H. B. C. Sollas. Clarendon Press, Oxford. WYYNE, A. B. 1872. Memoir on the Geology of Kutch. Geological Survey of India Memoirs, 9. YULE, H. & BURNELL, A. C. 1903. Hobson-Jobson,
a Glossary of Colloquial Anglo-Indian Words and Phrases, and of Kindred Terms, Etymologieal, Historical, Geographical and Discursive. (CROOKE, W. (ed.), Reprint, Asian Educational Services, Madras, 1995.)
Appendix The following extracts are transcribed from local newspapers, and from approximately 50 pages of handwritten minutes and letters, sent to and from the Governor's Office in Bombay, concerning the 16 June 1819 Kachchh earthquake in the days following the event. Captain James MacMurdo was the British representative (Resident) in Kachchh, and in addition to these government reports, he subsequently submitted a collection of papers describing the earthquake to the Literary Society of Bombay. These were published posthumously in Volume 3 (1823). (MacMurdo died from cholera at the age of 33 and an obituary appears in the same volume.) Papers based on his early travels continued to surface for the next few decades and a brief history of his short career was summarized by Gosh (1977). Unlike Burnes and Baker, who described crustal deformation near Sindri 7 and 35 years, respectively, after the event, MacMurdo knew the area intimately before the earthquake and although he experienced the earthquake at first hand, he did so on the Kachchh mainland, and at the time these letters were written was unaware of deformation near and north of Sindri. The data describe severe damage in Bhooj and Anjar (Mercalli Intensity X), and hint at similar damage in many villages of Kachchh. Shaking at Ahmedabad and Surat is less severe (perhaps intensity 7) and was reported directly to Bombay. In Bombay the event was scarcely felt.
From Boards Collections 15426-15427, 1820-1821 Vol. 620 Record Department, India Office Library, F/4[620 Bombay Public Consultations, Exams Office Nov. 1820
Bombay, 7 July 1819 Extract Public Letters from Bombay to East India House, London. Dated 7 July 1819. 2. We are concerned to report to your Honorable Court that about twenty minutes past seven in the evening of the 16th last month, a slight shock of an earthquake was very perceptibly felt in various parts of the Island. The shock did not last above a minute and
312
R. B I L H A M
no injury appears to have been sustained from its effects, indeed the concussion was so slight that many persons did not notice it, and entertained doubts of it having taken place, but your Honorable Court will observe with regret from the enclosed documents that its consequences were very severely felt at the Northern Stations under this government, particularly at Ahmedabad and also in Cutch. At the former it has destroyed the beautiful shaking minarets of the Imra Musjed which were so long the ornament and admiration of the East, and done considerable damage to other Public and Private Buildings. 3. At Anjar the Fort wall with its towers and Guns have been levelled to the ground with three fourths of the houses in the Town, those which have been left standing have also sustained injury and the general destruction is emphatically stated by the resident to have reduced a flourishing population in one moment to wretchedness and misery. The lives lost on this occasion, as far as they could be correctly ascertained at the date of Captain McMurdo's dispatch amount to upwards of one hundred. 4. Similar damages have been done at Bhooj and we fear with an equal loss in human life. 5. We have still to learn the extent of this awful visitation, but private letters from all parts of Guzerat and Kattywar concur in stating it to have been felt with great severity through the country. 6. Since the preceding paragraphs were written we have been furnished with copies of private letters from Lt. Col. Colin Milnes of his Majesty's 65th Regiment and commanding the troops in Cutch which together with the accounts given in the Bombay Gazette of this date convey a most lamentable picture of the effects of the Earthquake.
Bombay, 8 July 1819 Public Department To Joseph Dart esq. Secretary at the East India House, London. Sir, In reference to the 5th paragraph of the letter from this government to the Honorable the Court of Directors of yesterday's date which has been closed and sent on the Lady Barringdon, I am directed by the Right Honourable the Governor in Council to transmit to you for information of the Honorable Court the accompanying copy of one from the Resident at Anjar dated the 19th Ultimo detailing the particulars of the injury sustained at that place and its vicinity from the earthquake with which they were visited on the 16th of last month and the three following days. I have the honor to be Sir, Your obedient Servant, signed H Newnham. Acting Chief Secretary., Bombay Castle 8 July 1819
Ahmedabad,
7 July 1819
Extract Bombay Public Consultations. Read the following letter .from the Acting Judge and Criminal Judge at Ahmedabad to Mr. Acting Secretary Newnham, dated the 17th Ultimo.
Sir, para.1 Yesterday afternoon at 7 o'clock precisely a very severe shock of an earthquake was felt in this city, the swell came from the direction of South West, there are various opinions as to the time the shock lasted. I think it continued about two minutes. 2. Happily no lives have been lost, but the damage which has been done has been considerable. The Minarets of the Juma Musjid, the highest and most beautiful in the place were thrown down, various other Minarets outside the walls have shared the same fate, and many of the Mosques have been otherwise shattered and much injured. One of the gates of the town has also fallen. Of the Government buildings the Adawlut, has alone been affected. None of the walls have actually fallen, but they have been cracked by the shock in several places, and it will be necessary therefore to request the engineer office at this place to inspect them. 3. Several private houses have I understand been destroyed but I am not yet acquainted with the particulars on this lead. I shall make enquiry concerning the damage this sustained. If it shall be found to be very heavy on a consideration of the circumstances in life of those by whom it has been sustained, the liberality of Government will not fail, I should hope, to afford some alleviation of their misfortune. 4. During the confusion occasioned by the shock, a prisoner who was in confinement for security escaped from the Gaol, the Sepoys on guard had all left the gate in the moment of alarm and he took this opportunity to pass out unobserved. 5. Several shocks have certainly been felt in the night and again very sensibly this morning. I have the honour to be etc. Chas. Norris Acting Judge and Criminal Judge, Ahmedabad Adawlut 1819.
Minutes 1 July: Ordered that Mr. Norris' be informed that the Governor in Council has great satisfaction in learning that no lives have been lost on the occasion of the earthquake felt at Ahmedabad on the 16th of last month, the effects of which were also felt at this place though in a trifling degree. Resolved that Mr. Norris be inJormed that the Governor in Council approves of his having called on the Engineers office to examine the damage which the Adawlut has sustained, and to make any temporary repairs that the circumstances may render necessary. Mr. Norris' will however be careful not to incur any considerable expense without previously submitting an estimate and obtaining the sanction of the government. On the receipt of the further report on the damage sustained by private houses, he will be furnished with the instructions of the government on any suggestions that he may deem worthy of notice. Read the .following letter Jrom the Superintendent of the Marine dated 24th. ultimo with enclosure. To The Right Honorable Sir Evan Nepean, President and Governor in Council Right Honorable Sir, Having received a communication from Captain J. Pruen the Commodore at Surat, under date 17th instant, reporting the circumstances which had come within his observation on the
THE
1819 R A N N
OF KACHCHH
awful occasion of an Earthquake with which Surat and its vicinity had been visited, and the matter recorded in Captain Pruen's letter may not have come to the knowledge of the Right Honorable the Governor in Council, I have deemed it expeditions to transmit copy thereof for the information of your honorable board. Henry Meriton, Superintendents Office Bombay 24 June 1819
EARTHQUAKE
313
Minutes: Ordered the Superintendent be informed that .from the violence of the shocks and their long continuance it is fortunate so little damage has been sustabledji'om the earthquake felt at Surat on the 16th Ultimo. Read the following letter from the Resident at Bhooj and Collector at Anjar to Mr. Acting Secretary Newnham dated the 17th Ultimo.
A n j a r , 17 J u n e 1 8 1 9 S u r a t , 17 J u n e 1 8 1 9 To Henry Meriton Esq., Superintendents Office Bombay 17 June 1819 I have the honor to inform you at 20 before 8 PM yesterday evening the City of Surat and for some miles round and the opposite bank of the Taply, were visited by that Phenomenon Earthquake in a very awful degree. Wen it first began I was lying down on my Couch, being still an Invalid. I found the whole house in serious agitation. The furniture all in motion and a small table close to me so much so as to keep striking the wall the lamp moving from east to west with the house about 6 or 8 inches each way. i got down stair as fast as possible about three minutes had transpired before I got out of the house, and I felt myself a little giddy. I found a number of people collected outside looking with astonishment at my house, which stands alone, and which was in such agitation I expected it to fall every second. The earth under our feet was by this time convulsed and seemed as if it was floating on a long ground swell trying to break it away through, and from its very great motion I expected to see the ground crack. The shock lasted about 5 or 6 minutes and appeared to me to run from East to West. The inhabitants were much alarmed. N o t a breath of wind was moving, with a clear sky, nor was there the least warning of its approach. On enquiry this morning I find several accidents have happened to houses, and at the village of Omer about 2 miles west, several Houses fell down. A Parsee Pagoda fell down in one side, and reports said one poor man was killed 10 minutes past 10 AM. We have just had another shock that lasted only one minute. I likewise felt two slight ones about 8:30 last night and at 10:10 am another shock. This shock stopped my watch, the glasses containing the oil in the lamp in two or three houses were upset. The well in the Jail whose water was about four feet below the earth was forced up to run over. The river water was likewise much agitated. A tank of water in the Bazar likewise threw its water out. Time alone will inform us whether it will be a partial convulsion of nature in the bowels of the Earth near this latitude caused by some great eruption at a far distant spot. I have felt it part of my duty to give you Sir, this slight information, but no doubt a more able pen, will give it Government. I have only related plain and simple facts for your information I have the honour to be etc. J. Pruen Surat 17 June 1819. P.S. I find that General Cook's watch stopped at 10:08 this morning.
Sir, It is with sincere regret that I have to inform you that this place was visited by an earthquake yesterday evening at 10 minutes before 7 O'clock. The effects of the shock, which lasted nearly 2 minutes, have been the leveling of the Fort Wall to the ground. Not a hundred yards of the wall remain in any one spot, and guns, towers etc. are all hurled in one mass of ruin. The destruction in the town has been distressing and awful. Not 1/4 of the houses are standing and those that do remain are all ruined. I cannot yet state the particulars of the losses, but I may in one word say that a flourishing population has been reduced in one moment to wretchedness and misery. I fear we shall have to lament the loss of upwards of one hundred people besides those hurt. Reports from the country state similar disasters in all the villages round about, and letters from Bhooj inform us that the Fort is much in the same condition as Anjar. Slight shocks continue to be felt and I shall (at) the first leisure moment, report such particulars as I may be able to collect. I have the honor to be etc. J. M c M u r d o
Ordered that our concern be expressed to Captain McMurdo at the effects produced jrom the earthquake felt at An/ar etc, with intimation that we are anxious to receive Jilrther accounts oJ" the extent of the damages sustained by the adjacent country. Recorded the following extracts of private letters from Colonel Milnes on the subject of the earthquake at Anjar.
B h o o j , 17 J u n e 1 8 1 9 We are at present in a shocking state of alarm here last evening between 6 and 7 o'clock we were visited by a dreadful e a r t h q u a k e - T h e wall that surrounded Bhooj is ahnost levelled with the ground and the few Towers which are left standing are merely broken remains; The houses generally unroofed; others in ruins, and most of the larger buildings including the Palace greatly injured. The wall of the Hill Fort down in many places and a complete breach near the gateway. The right of our camp rest a short distance on the left of the latter, fronting the tower, and extends along the bottom of the hill, to a little beyond the large tower on the south most point. I am happy to say that we have had no one materially hurt two Sepoys only bruised, who were on duty in the town, but I fear that a great many casualties have occurred there among the poor natives. Some hundreds are said to have lost their lives. There is at present so much confusion that the number cannot be ascertained.
314
R. B I L H A M
We had several shocks during the night and they have continued at intervals this day, the last one about two hours ago, when I could scarcely keep upon my legs, The sensation is horrible while it lasts. They have suffered we understand, in the same way at Anjar.
Bhooj, 18 June 1819 We still continue here in a state of doubtfulness and extreme anxiety. Every three hours we feel the Earth trembling under us but in a slighter degree. The inhabitants quitted the town yesterday, and slept out last night in the plains and about the neighbouring hills. The number of lives at present ascertained to be lost is almost 500. I have just received a letter from Captain MacMurdo (who went to Anjar before the late events) for Dooly Bearers to assist in clearing away the streets and gates there; which he says, if not done before the rain falls, the town will be swamped. They are in wretched state there it appears.
Bhooj, 19 June 1819 The last trembling we had was at 12 o'clock this day, rather severe, so that we are still kept in dread. Between 50 and 100 missing bodies have been discovered in the Town. Before this awful event took place we had not the least warning of its approach whatever. On the evening it occurred I took a short ride. The weather was delightful, a clear sky, a gentle breeze and perfectly cool, there having been a heavy fall of rain only a few days before. As I was returning home in a quick walk, some time after the Sun had set, when within about a quarter of a mile of the front of our camp I suddenly perceived something very unusual and extraordinary in the paces of my horse. His legs appeared to be in motion but he seemed to make no way whatever, at the same time, I felt a sort of dizziness in my head, and a sickness in my stomach, supposing this to proceed from the strange motion of the horse and that he was ill and might fall under me, I was thinking to dismount when my attention was distracted by an immense cloud of dust bursting out from the center of the hill Fort, which I took to be an explosion of gunpowder, and the first impression on me was that an accident had happened to the Magazine. But on casting my eyes to the left towards Bhooj, I observed the whole Town from one extreme to the other completely enveloped in a similar cloud, and on looking behind me I also observed the same appearance at no great distance in that quarter. I was then satisfied of its being dust and not Gunpowder, and concluded it to be some description of Typhoon or Hurricane, but still I was perplexed to account for its continuing so perfectly calm and serene about the spot where I was standing, and there not being the least symptom of a wind rising. I was just about moving quickly into Camp when I saw Captain Wilson, the assistant resident, who had been riding with me and from whom I had parted only a few minutes before, coming towards me from the town. He acquainted me
that he was entering by one of the gateways when there was a general crash and that the whole place had fallen down. Upon this I of course knew at once the cause, but until that moment had not the most distant idea of its being an earthquake. When I go to my tent I found that the Table which had been laid out for dinner was thrown over and everything on it smashed to pieces. The deposed Rao's mother and his Father's wives were among the sufferers in town. Some part of the Palace fell upon her, the body not yet found. Extract Bombay Public Consultations 14 July 1819 Read the following letter from the Resident at Anjar 19 June 1819.
Anjar, 19 June 1819 To William Newnham esq., Acting Chief secretary to Government of Bombay Sir 1. Since my address under the 17th instant accounts have been received from various quarters of the country. There is every reason to believe that the shock has destroyed in a greater or lesser degree, every fort and town from Arrisir to Luckput. many of the villages round about Anjar are reduced to heaps of rubbish, and I fear that those in Cutch and Wagir generally are little less injured. Bhooj has been a great sufferer. The wall so the town level with the ground. The palace in many parts in the same state, and the private dwelling houses in ruins. The loss of lives is not exactly ascertained but the lowest calculation makes it 500 people. The Rao's family has escaped, with the exception of the old lady, the widow of Rao Raidhan. Mandavee is stated to have suffered less than other places and is said to have lost only 125 people. Accounts from Coorbee state that town to be in ruins. 2. Our loss in Anjar has been greater that I had at first supposed. We have to lament the loss of 166 lives besides and double that number wounded, many of whom severely. Out of 4500 houses of which the town is composed, about 1500 are so completely destroyed as to not leave one stone upon another. They are overturned from the very foundation. About 1000 more are laid in ruins and so dreadful has been the shock, that of the standing some are injured and many uninhabitable. The fort cannot now bear that name, as there is no a third of it remaining in different parts, and even those are likely to fall with the first rains. 3. It is impossible to describe the misery of the unfortunate people. Their property buried in ruins, and exposed without the possibility of saving it [from] the weather. Their families, some among the ruins, and some in the open fields exposed in the same wretched condition, the calamity has been so general that not a labourer can be had for money. In the richer, and more respectable class of people, are seen sitting surrounded by their families, in the spot where their houses once stood in the most helpless and destitute situation. 4. I have not in my power to assist them materially, but what is in my power I have done. Free ingress and egress has been given to all property without taxes, and I venture to suggest to Government to continue this favour towards the people at least for some months to
T H E 1819 R A N N O F K A C H C H H come. It seems impossible to levy duties from a town in ruins. 5. I have set the labouring people about cleaning the streets and making a passage for the water to escape, for if the rains were to set in with violence the lower and greater part of the town would, in the present state, be 6 feet under water. 6. I applied to the commanding officer, for the assistance of a working party, but I am sorry to say that he did not think it proper to allow the men to be employed in assisting the inhabitants. Enclosed is the correspondence for the information of the government. 7. Since writing the forgoing, 150 Dooley Bearers have been kindly sent by Colonel Milnes to our assistance. I have the honour to be etc. J. McMurdo, Resident at Bhooj and Collector at Anjar. Anjar 19 June 1819. P.S. I have neglected to observe that the Public Buildings of every description including the judges dwelling house, offices etc. are rendered unsafe to inhabit.
Anjar, 17 June 1819; MacMurdo to Morgan requesting help To Captain James Morgan Commanding at Anjar. Sir, 1. It is of the utmost importance for the safety if the town of Anjar that the Warsamoree and Sortia gateways, and the water drains should be cleared of ruins, in order that the water may pass off, which would otherwise, in case of rain, swamp the better half of the town. 2. In consequence of the threatening appearance of the weather, and all the town people being at present too much employed in rescuing their families and small remains of property, I take the liberty to request, should you have no objection, that a proportion of the regular Sepoys be permitted to aid for a few days, for the public good by their labor, in working parties, to clear the passages for water and the gateways. I make this request with less hesitation as the dreadful misfortune has fallen with comparatively trifling weight upon the men of the Detachment. I have the honor to be etc., L McMurdo 17 June 1819
Anjar, 17 June 1819; Morgan to MacMurdo denying help To Capt, J. McMurdo Resident etc. etc. Sir Consistent with the military duties required of the Garrison under existing circumstances, I am concerned to say it is not in my power to comply with you request. Their duties, I am ready to allow should give way to necessities of greater magnitude, but until that is the case I conceive the employment of soldiers in occupations of the nature required, would be improper and inconsistent with the established use of the service.
EARTHQUAKE
315
I have the honor to be etc. Thos Morgan, Captain Commanding Anjar. 17 June 1819.
Ahmedabad, 28 June 1819; request to repair the mosque Sir, The Khazee of this city having the name of the Muhamedan inhabitants requested me to transmit the accompanying petition to Government to repair the Juma Muzjed. ! have the honour to forward it with a translation for the purposes of being laid before the honorable board. 2. Upon a subject of this kind ! can of course have little to say; there is certainly scarcely any thing that could afford so strong a contrast to the principles of the government to which we have succeeded, as an attention to the wishes of the people on this occasion, and that could excite in so lively a manner the gratitude of our subjects of all ranks and persuasions, for it must be observed that the memory of Sultan Ahmed is held in respect and veneration not only by Moosulmans but also by Hindoos, as that of their Protector and the Founder of their City. 3. There must at the same time be various considerations which must materially affect the decision of government on this question, and upon which I cannot presume to offer any observations. 4. With reference to my letter of the 17th I beg that you will inform the Hon. Board that private property has not suffered from the earthquake of the 16th in any way to make public assistance at all necessary. Charles Norris, 28 June 1819 Ahmedabad.
Translation of the petition 'In the evening of the 16 June 1819 great alarm and apprehension was occasioned throughout the city by the shock of an earthquake. It was so severe that all the people thought that the last day was come. All the buildings and houses old and new shook like rattans and were very nearby all falling and killing thousands of men. The time however of the people was not come and God was merciful so that the earthquake ceased, but great damage was done to the public buildings. The old stone buildings, which were erected four hundred years ago or more, the stone mosques, Razahs and domes both within and without the city, the stone minarets of the mosques, remarkable for their height and beauty, which need not have fallen before the last day, have all been broken and thrown down. But the greatest loss has been in the minarets of the Iuma Muzjed, which was built by Sultan Ahmed Badshah, the founder of the city. It stands in the middle of the Bazaar, and of the city, and its minarets were the greatest ornament to the city, and to the Mosque. Both the Mosque and the minarets were the largest in the place, and the Minarets were also in this particular, remarkable, in that a person going up one them and shaking it, could communicate the motion to the other, so that people came from all quarters to admire them. Both these minarets have been thrown down by the shock and the whole city deprived of its greatest ornament and the mosque is quite deformed.'
316
R. B I L H A M
[Another two pages follow asking the government to repair the Mosque etc. These are omitted for brevity]
Resolved that Mr. Norris' be informed that government deem it proper to decline entering into an)' engagement for the re-erection of the Minarets of the Iumma Musjed at Ahmedabad, as the expense attending it would be considerable and the buildings" themseh,es are more ornamental than useful.
B h o o j , 23 J u n e 1819 Read 21 July 1819 I have the honor to report for the information of government that since I returned to Bhooj I have ascertained the damage sustained in that town to be much greater than I had supposed. The loss in lives has not been correctly ascertained, as bodies continue to be dug out from the ruins. About a thousand have already been found. The fort is in a most ruinous state, but although there is little of it entire, there are few places so completely leveled to the ground as that of Anjar. As near as can be calculated seven thousand houses have been overturned, and few or none in the city left uninjured. The Palace which is an immense mass of building has been dreadfully shattered. All the upper parts overturned, and the pile, as low as the lower floor, rent and shook, so as to render the whole nearly uninhabitable. I am happy to say that the Rao's family has escaped without further loss than already reported. The Ex Rao is removed to tents near those of the residency, where he is guarded by a detachment of 100 Rank and File. The Rao Desul and all the females of the family, are also in tents outside of the town, where I hope to be able to persuade them to remain until some place can be made secure for their reception. I may observe that the whole of Cutch has suffered nearly equally in regard to the loss of houses, yet I am glad to say that the proportion of lives lost in different places bears no affinity. Perhaps Bhooj has lost as many as the whole of Cutch put together. In Mandavi 116 and in Luckput 150 are said to have suffered. The Iharyjas have in some instances lost members of their families. Koteree, Thera, Koira, Mothana and Nangercha are spoken of as having experienced the most dreadful effects from the shock, but perhaps there is little difference anywhere. A number of phenomenon are said to have occurred at the moment of the shock, but I shall only remark that which appears the most striking. The Runn and Bhunee on the north of Cutch, between that Province and the isolated district of Kouvra, which was quite dry was suddenly filled with a sheet of water, the extent of which east and west was not known, but its breadth was generally about 6 miles, and its depth gradually increased to upwards of 2.5 feet. after which, in a few hours, the waters subsided to about half that quantity. Horseman who crossed this tract on the day following the shock, describe a number of cones of soft sand elevated above the water, the tops of which were bubbling with air and water when they passed. As far as I have learned the sandy bed of every dry river in
Cutch was filled for a space of time with a flood of water. These waters have the colour and taste of the soil from whence they were ejected. The effects which this awful visitation may have on the prosperity of Cutch and consequently on our interests are very material, but I shall defer entering on the subject until the return of the Bhyant to Bhooj to their homes (wither the late event has called them) shall enable me to mature my sentiments. J. McMurdo, 23 June 1819 Anjar. [previous readers of the last sentence have attempted to elucidate its meaning through annotations to the syntax of the original without success]
Bhooj, 16 S e p t e m b e r 1819 I have the honour to submit for the information of government an estimate of the value of the houses destroyed by the late earthquake in the town of Anjar I have the honour etc. J. Macmurdo (Editor's Note: See Table at top of following page) J. McMurdo, 16 Sept, 1819 Camp Bhooj
B o m b a y , 15 O c t o b e r 1819 Captain Macmurdo is informed that we are concerned to find that the in/ury sustained by the late earthquake has been attended with so very heavy a loss to the inhabitants at Anjar as above reported.
B o m b a y , 1 N o v e m b e r 1819 Extract Public letters from Bombay Para 61. A subscription having been circulated for the relief of the Hon. Company's distressed subjects at Anjar who have suffered so severely from the effects of the late earthquake, as so fully described in the dispatches from Captain Macmurdo, transmitted to your Hon., Court under the dates the 7th and 8th July last and in the one recorded as per margin (1819 consultations 21 July Folio 1211), we have taken upon ourselves to subscribe the sum of Rupees four thousand on the part of the Honourable Company. Captain Macmurdo having been vested with the distribution of the amount. Para 62. The value of 1547 Houses destroyed on this occasion is estimated at Couries 4.73.790 which yielded an annual rent of Couries 26,778. [In margin: 1819 Consultations 20 October Folio.] The following entry is the first in the sequence in the India office collection, but its date is obviously related to the 1 November statement and may be a summary of it. It is written in a poor script.
B o m b a y , 7 J u l y 1819 The shock of an earthquake was severely felt on the 16th June 1819 at the northern Stations particularly at Ahmedabad and also in Kachchh. The sum of Rp.4000
THE No of casts
1819 R A N N
OF KACHCHH
EARTHQUAKE
317
Names of casts
Number of monthly rent value houses destroyed 1. Nunowana Brahmans 92 189 50,925 2. Bunyans 229 449 1 10,3975 3. Bhojiik 3 2 2 400 4. Shimalee Brahmins 95 147 40625 5. Bhattia 113 220 1 65700 6. Nagur Mettas 26 46 3 19,626 The list has altogether 46 lines of damage data sorted according to cast that are omitted for brevity. Monetary units are Couries (5000 couries ~ 1 rupee, Yule & Burnell 1903). 46. Bhunjia 2 1 2 150 total 1547 2231 2 4,73,790
has been subscribed on the part of the Co. for the relief of their distressed subjects at Anjar.
L o n d o n (to B o m b a y ) , 6 June 1821 Answer to public letter dated 7 July 1819 Public Department Despatches to Bombay 13 Sept 1820-31 Oct 1821, Vol. 45, E/4/1040 Record Department, India Office Library, London, p. 653. 33. We were greatly concerned on perusing the melancholy accounts of the devastation occasioned in various parts of the territories subject to your government by the awful visitation in question, which is stated to have been felt with such severity at Anjar in Cutch, and to have at once reduced a flourishing population to wretchedness and misery and which was attended with the loss of many hundred lives. 34. The relief extended to the surviving sufferers by means of a public subscription was highly creditable to the promoters of it, and we cordially approve of your having contributed 4000 Rs. on the part of the government. 35. We cannot close this subject without expressing our entire approbation of the conduct of Captain McMurdo. Footnote - see 8 July 1819 and also paragraph 61-62 1 N o v 1819.
Extracts from the Bombay Gazette 16 June-4 August 1819 The Bombay Gazette was published weekly. It reproduces verbatim letters, but does not disclose the writer's names. The following abstracts have been edited to suppress extraneous information (weather, speculation, etc.) from these accounts. The damage is concentrated along the coast of Cutch and Cambay and is less pronounced inland. Few details are given for the remoter sensed locations.
16 June 1819 No reports
23 June 1819 Earthquake: A slight shock being felt etc . . . . But it appears to have been so slight a convulsion that most people doubted the evidence. The west side of the island seems to have been most affected for we have been informed at Cambala, the undulations were distressing, and in the houses in the vicinity of the retreat, the lamps were shook violently. It was felt at Sion, and along the east side of the island but less distinctly and its duration was only of a few seconds. This is further corroborated by the following extract of a letter from Surat dated 17 June. 'At 20 minutes past seven yesterday evening I felt a strange trembling sensation . . . . It is strange my servants did not perceive it . . . . I felt as if undetermined whether I should stand or lie down.'
W e d n e s d a y 30 June 1819 The earthquake, We have received accounts from Ahmedabad, from Broach, Baroda and Khaira concerning this awful convulsion of nature. Extract of a letter from Ahmedabad, 18 June 1819 "It commenced gradually with a trembling of the earth attended with a rumbling noise: this increased every second, and was succeeded by a strong rushing noise, with a violent undulating motion, so that it was with difficulty we could keep on our legs. At this time, all the disagreeable sensation was experienced of being tossed in a ship at sea in a swell, and the rocking was so great that every moment we expected the earth to open beneath our feet.' ]Omitted here is an irrelevant paragraph on weather] 'The rocking motion affected persons in various ways in different situations. Those indoors experienced all the horrors attending the awful suspense in which they felt themselves, the walls of the houses shook and were rent in various places, the beams and the roofs cracked and appeared ready to fall on the inmates and crush them to atoms in an instant. People in wheeled carriages were nearly thrown out by the shaking of their vehicles, and animals of all kinds in a stationary position, felt confounded and became restless. Persons in motion, on horseback, were however unconscious of the shock.
318
R. B I L H A M
The proud minarets of the great Mosque, the Juma Muzjid, erected by the Sooltan Ahmud, the King of Guzerat and the founder of the city of Ahmedabad, which have stood for nearly 450 years, have tumbled to the ground. The Mosque itself has sustained less injury than could have been expected, and the handsome arch which separated the minarets has escaped unhurt. Another Muzjid, of elegant structure, which lies to the left of the road leading to the Shahee Bagh, denominated the Beebee's or Unchunt Koonkwee ke Muzjid, has shared the same fate. A gentleman while riding out saw the minars come down: the tops were thrown to a great distance, and immediately after the stones came tumbling down one after another. The only remaining shaking minarets, which are at all worthy of notice, and much inferior to the others have, I hear, been sadly fractured. They are situated west of the city outside the walls. The mausoleums (Rozas) and places of Moohamudan worship have suffered considerably, both in the city and in he surrounding country. Hindoo temples are few in number and of recent build in the city since its conquest from the true believers seventy years ago by the Murhuttus; consequently a very small number have been damaged. The walls of the Udalhut, an old building erected by the Murhuttus and the palace of the Peshwas viceroys in Guzerat have been much injured, and the walls rent in many places. The magnificent towers also forming the grand entrance to the citadel have been much shaken, and cracked in several places, especially the one in which the flag has been placed. Many private houses have been reduced to ruins, and 'tis most fortunate amidst all our disasters that not a single life has been lost, and but few accidents. Between the hours of 12 and 1 the same night we experienced two or three slight shocks and another at 6. At quarter before 10 we had one more severe, which shook the houses and caused the windows and doors to rattle violently. We were now on the alert and quitted our houses in haste but the shock did not continue above a few seconds, and was trifling compared with the one of the previous night. A 10:30 we were again visited slightly and at intervals the whole of the day. The last I felt occurred at 12:20 in the night and since then I cannot say I have experienced any more, although fancy has frequently led me to pause, and expect a return of this terrible visitation.' [Omitted here are several paragraphs of non-factual information] 'Reports from Kaira mention that the grand shock was experienced there twenty minutes after us and that it lasted only 37 seconds; two natives were killed by the falling of their homes & a good deal of damage has been done there. The Udulat has suffered and the walls rent, the Jain temple opposite to it has also received a terrible fracture. Kaira is distant from Ahmedabad 18 or 20 miles.'
Broach, 16 June 1819 'We had last night a very severe shock of an earthquake. The ground moved like the waves of the sea, and it was with the greatest difficulty I could keep on my legs. The walls of the house moved backwards
and forwards, and the lamps went with a very quick motion; the water in our well rose many feet with a great noise, and did not subside for an hour, after it was over. Europeans and natives ran into the streets; many native houses were thrown down, and several boats upset by the extraordinary motion of the river. It lasted above three minutes. I never in my life felt such an awful time, every moment expecting instant death. 17th June this morning at 10 o'clock we had another slight shock for a few seconds.' Our correspondent from Baroda. being a native expresses himself as follows. 'Last night I come from office, then we get Durtee Kup, ground so much shook, water jar is broken, Dinner is spoiled, all women and children, run away. No man understand this thing, only God, Lamp is cracked, Goat is gone away, all the persons is much frightened.'
14 July 1819 Earthquake 5 June 1819 in Mocha 'its effects were to overturn tables and chairs all over the adjacent country and in the factory at Mocha.'
Porebunder, 17 June 1819 'We yesterday experienced in this town and fort one of the most awful scenes of nature, that of a violent and destructive shock from an earthquake.' [I omit here a paragraph of personal narrative of the danger encountered of the collapse of walls of the fort leading to the narrow escape of the two officers' with their lives.] 'On reaching a spot of comparative safety, for then no place was safe, the attention was directed to a vast cloud of black dust, arising at about 300 yards distant, and from the sea face of the f o r t . . , approaching the cloud of dust, I found it to proceed from the fall of 9 Towers (20-30feet high), and large parts of the curtain, (22-25 feet high) leaving twenty one breaches of 40 and 60 yards wide. This devastation extended for 500 yards, and over a part of the fort which I had been walking on, not five minutes before.' [Irrelevant paragraph omitted.] 'I believe there are few houses throughout this large city which have not been more or less injured: some have fallen and blocked the streets in which they were situated. I am happy to say that not one life has been lost in this town, a circumstance which appears almost miraculous, from the danger which existed. The Earth opened, and water issued from the cavity, over an extensive piece of ground, in a plain, distant 14 miles hence. [Irrelevant paragraph omitted.] There has been several other shocks between 10 am and 2pm which brought some houses down and violently shook the seats of those who were seated within doors, which caused then to run out of their houses, but these inferior alarms are not to be compared with yesterday's awful phenomenon. [Irrelevant paragraph omitted.]
THE
1819 R A N N
OF KACHCHH
I am this moment informed that fifty men were killed by the fall of walls at Mangrole on this coast which is distant 60 miles in a SE direction.., half a Lac of Rupees damage estimated.
EARTHQUAKE
319
thought of. I observed that my tent was shaking and that my sword which was hanging against the wall flew backwards and forwards, in a way that I could not easily account for. This is the furthest north we have yet heard of, nor have we any accounts of it having extended farther south than Poonah.'
Porebunder, 18 June same correspondent 'they say the town of Kooteennahas suffered but not so much as Porebunder'.
Wednesday 4 August Camp Bhooj, 17 July
Bhooj, 23 June 1819 An extensive account of the earthquake as witnessed in Anjar and Bhooj containing identical information as found in the M a c M u r d o account of same date entered into the government report above. Presumably from MacMurdo. Editorial 'We also find that this awful phenomenon has also been witnessed at Calcutta, and singular to say on the same day that it was felt at this presidency.' An extract of a letter from Allahabad25 : 30~ 'we had a slight shock of an earthquake two or three evenings since. I never experienced so extraordinary a sensation, my first idea was that I must be extremely ill, for I felt the earth move in a way that I cannot describe and an earthquake was the last thing I
'About 1 o'clock on the 15th a severe shock was felt here, the tiles came off several houses. On the evening of the 15th another (7:30) slight and another in the morning and again this morning. 50 shocks since the 16th June. June 16 7:20 awful to a degree June 17 10:00 two slight ones June 18 7 am rather strong June 19 several slight ones June 21 at 9 rather strong June 23 at 2 a m strong, the house and furniture in great agitation 3/4 of an hour June 23/30 2 or three slight ones. July 8 at 11 pm slight July 11 at 5 slight July 21 at 10 pm strong. The house etc in agitation 3/4 of a minute.'
The effects of upper plate deformation on records of prehistoric Cascadia subduction zone earthquakes L I S A C. M C N E I L L 1'3, C H R I S
GOLDFINGER
& LAVERNE
2, R O B E R T
S. Y E A T S 1
D. K U L M 2
1Department of Geosciences, Wilkinson Hall, 104, Oregon State University, Corvallis, OR 97331, USA 2 College o f Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, OR 97331, USA 3 Present address." Department of Earth Sciences, UniversiO, of Leeds, Leeds' L S 2 9J J, UK Abstract: Geophysical data from the offshore Cascadia forearc reveal many Quaternary
upper-plate faults and folds. Most active structures are within the accretionary wedge, but significant deformation is also found on the continental shelf. Several faults and synclines project into adjacent coastal bays where deformation of Pleistocene marine terraces is reported. Rapidly buried marsh deposits and drowned forests in these coastal lowlands are interpreted to record coseismic deformation by prehistoric subduction zone earthquakes. The extent and amount of such coastal subsidence has been used to infer characteristic magnitudes and recurrence intervals. However, the record may incorporate both elastic strain release on the subduction zone and localized permanent upper-plate deformation. Movement on upper-plate structures may be triggered by a subduction zone earthquake, as observed in the Nankai and Alaskan forearcs. Alternatively, they may deform independently of subduction zone earthquakes. Regardless of which style of deformation predominates, the record of coseismic subsidence is likely to be affected. Crustal deformation may also contribute to the preservation of subsided marshes. Modelling of subduction zone earthquake characteristics based on coastal marsh stratigraphy is likely to be inaccurate in terms of: (a) total apparent rupture length and earthquake magnitude; (b) amount of subsidence and hence the position of the locked zone; (c) recurrence interval. Most of these shelf and coastal structures respond to N-S compression, in contrast to convergence-related northeasterly compression in the accretionary prism, but in agreement with the regional stress field. Despite low historical coastal and continental shelf seismicity, upper-plate faults may also pose an independent seismic hazard.
Geological and geophysical investigations of the Cascadia subduction zone during the last decade have increased public awareness of regional earthquake hazards from a subduction zone previously thought to be aseismic (Ando & Balazs 1979). Evidence for repeated abrupt subsidence in the last few thousand years is found in coastal bays along the active margin, in the form of buried marsh deposits and drowned forests (e.g. Atwater 1987, 1992; Darienzo & Peterson 1990; Atwater et al. 1995; Nelson et al. 1995; Yamaguchi et al. 1997). Previous workers have believed these deposits to be a result of coseismic subsidence and have attributed them to subduction zone earthquakes (e.g. Atwater 1987, 1992; Darienzo & Peterson 1990; Atwater et al. 1995; Nelson et al. 1995; Yamaguchi et al. 1997). However, the similarity of these deposits to the marsh stratigraphy of tectonically inactive coasts has led to the suggestion that many abrupt burials may be non-tectonic in origin and driven by local changes in intertidal environment (e.g. Long & Shennan 1994). A non-tectonic
origin cannot be eliminated except in certain cases where the event is found to be regionally widespread, is associated with other coseismic phenomenon such as tsunami deposits or liquefaction, or where the magnitude of subsidence is too large to be explained by nontectonic mechanisms. The coseismic v. nonseismic origin of the subsidence events will not be addressed by this paper, but remains a topic of debate. The chronology, distribution, and amount of subsidence for individual locations and events have been used to estimate recurrence intervals and magnitudes. A significant component of subsidence recorded at these sites could be attributed to localized permanent crustal deformation (Goldfinger et al. 1992b; Nelson 1992; Goldfinger 1994; McCaffrey & Goldfinger 1995; McCrory 1996; Nelson & Personius 1996; Clague 1997). This study shows that several bays lie within actively deforming synclines or on the downthrown side of faults mapped onshore and in the contiguous offshore inner shelf. Mapping
MCNEILL, L. C., GOLDFINGER,C., YEATS,R. S. & KULM,L. D. 1998. The effects of upper plate deformation on records of prehistoric Cascadia subduction zone earthquakes. In: STEWART,I. S. & VITA-FINZI,C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 321-342.
322
L . C . McNEILL E T A L .
the offshore region benefited from extensive geophysical datasets and the absence of thick vegetation, which often hinders coastal fieldwork. A question critical to the understanding of contributions by crustal structures to the subsidence record is: Do crustal and subduction zone earthquakes operate independently or together? Regardless of the answer to this question, localized upper-plate deformation calls into question calculations of recurrence interval and earthquake magnitude obtained from records of coastal subsidence. The objectives of this study were to consider the distribution of abruptly buried deposits in light of upper-plate Quaternary deformation in the offshore inner shelf and onshore coastal region, and to determine the effects on and limitations of this palaeoseismic record for determination of prehistoric subduction-zone earthquake characteristics. We show that such structures reflect the predominant structural regime and stress field of the shelf and coast, and suggest that these localized structures may pose an independent seismic hazard to coastal communities.
Methods Structures outlined in this paper were mapped from single-channel and multichannel seismic reflection profiles, sidescan sonar data, and submersible observations. Seismic data were collected by Oregon State University (OSU), University of Washington, the US Geological Survey, and the petroleum industry. Many of the seismic profiles extend to within a few kilometres of the coastline. One proprietary dataset consists of a closely spaced network (between 2 and 30 km apart) of high-quality, precision-navigated, migrated multichannel profiles. This dataset covers the shelf and uppermost slope of Oregon and Washington. Acquisition details of this proprietary dataset have been are discussed by McNeill et al. (1997). Many of the structures identified in this paper were mapped from this particular dataset, with the data shown here in the form of migrated time sections. Deep-towed sidescan sonar data were collected during several research cruises between 1992 and 1995, and were navigated using global positioning system (GPS). Sidescan sonar data on the shelf include the AMS150SIkHz and Klein systems. Details of sidescan sonar data, and sonar processing and imaging techniques have been given by Goldfinger et al. (1997b). The shallow-diving submersible, D E L T A , was used to dive on fault targets selected from sidescan sonar images and seismic reflection data.
Tectonic setting The Cascadia subduction zone is located off the coast of the northwestern United States and
southwestern Canada, a result of subduction of the Juan de Fuca and Gorda plates beneath the North American plate (Fig. 1 inset). Convergence is oblique to the northeast at a rate of 42 mm/year at latitude 47~ (NUVEL-1 plate motion model of De Mets et al. (1990)). No earthquakes have been recorded on the plate boundary during the period of recorded seismicity, with the possible exception of the 1992 Petrolia earthquake (Oppenheimer et al. 1993). Crustal seismicity in the North American plate and within the subducting Juan de Fuca plate is minimal but present, with greatest recorded seismicity in Washington (Puget Sound) and in the Gorda plate off northern California (e.g. Crosson & Owens 1987; Weaver & Baker 1988; Ludwin et al. 1991). The lack of seismicity on the megathrust led early workers to infer that subduction may have ceased or that subduction is aseismic (Ando & Balazs 1979). An alternative interpretation, that strain is currently accumulating and will be released in a future large-magnitude earthquake, was supported by evidence for abrupt potentially coseismic coastal subsidence (Atwater 1987) and regularly spaced turbidites, potentially earthquake induced, in submarine channels on the Juan de Fuca abyssal plain (Adams 1990). Both lines of evidence point to several events in the last few thousand years. Coseismic subsidence in the coastal Cascadia region is predicted from elastic dislocation models of the subduction zone cycle. Figure 2 illustrates this cycle, where the land surface a certain distance from the deformation front is expected to gradually uplift during the interseismic period, followed by sudden subsidence during the seismic event. This region is expected to coincide with the coast in Cascadia. A similar elastic response was recorded accompanying the 1960 Chilean and 1964 Alaskan subduction earthquakes (PlaNer 1969, 1972), with regions of coseismic uplift and subsidence identified. The earthquake potential of the Cascadia subduction zone based on coastal subsidence and abyssal plain turbidites was reinforced by analogies between Cascadia and Chilean-type subduction zones (Heaton & Kanamori 1984).
Regional stratigraphy and structure Cenozoic strata underlying the continental shelf consist of Eocene to Quaternary bathyal to neritic forearc basin and accretionary complex sequences resting in part on early Eocene basalt (Snavely 1987; Palmer & Lingley 1989; Snavely & Wells 1996). Several regional unconformities are prominent within the basinal sequence, including
UPPER PLATE DEFORMATION AND CASCADIA EARTHQUAKES
323
Fig. 2. Model showing the subduction earthquake cycle (after Darienzo & Peterson 1990; reproduced by kind permission of GSA). During the interseismic period (a period of hundreds of years), strain accumulates causing much of the Cascadia coastline to gradually uplift (between the two hinge lines). Coseismic strain release occurs during the subduction earthquake and causes this region to rapidly subside, resulting in drowning of the coastal area. The seaward hinge line represents the zero isobase, seaward of which coseismic uplift occurs. No vertical movement occurs at the hinge lines. Under purely elastic conditions, subsidence is completely recovered with no net subsidence or uplift through the earthquake cycle. NA, North American plate; JDF, Juan de Fuca plate.
Fig. 1. Neotectonic map of recent structures (late Pliocene to Holocene anticlines and faults only), including inner shelf and coastal structures outlined in this paper, and locations of coastal subsidence. Inset shows general tectonic setting with plate convergence vector (De Mets et al. 1990). Plates: P, Pacific; JDF, Juan de Fuca; NA, North American; G, Gorda. FZ, Fracture Zone; SNF, left-lateral South Nitinat Fault. Northern California offshore structures after Clarke (1990, 1992; reproduced by kind permission of AAPG and GSA).
one of late Miocene-early Pliocene age (MP) and one of late Pliocene-Pleistocene age. The Miocene basinal sequence of the Washington shelf is underlain by a thick sequence of Eocene to middle Miocene m61ange and broken formation, known as the Hoh Rock Assemblage and Ozette Formation onshore (Rau 1973; Orange et al. 1993; McNeill et al. 1997). Eocene Coast Range basaltic formations and the middle Miocene Columbia River Basalt Group locally interfinger with Eocene to Miocene strata in coastal Oregon and southern Washington, and extend onto the continental shelf. The lower continental slope is dominated by compressional tectonics within the active accretionary prism in response to plate convergence, with structural trends between N-S and N W - S E
324
L . C . McNEILL E T AL.
(Fig. 1; Goldfinger et al. 1992a, 1997a). A set of WNW-trending left-lateral strike-slip faults, a result of oblique convergence, was mapped on the continental slope and locally on the outermost shelf of Oregon and Washington (Fig. 1; Goldfinger 1994; Goldfinger et al. 1992b, 1996, 1997a). On the northern Oregon and Washington shelf and upper slope, E-W extension is common in the form of listric normal faulting related to the underlying mobile m61ange and broken formation (McNeill et al. 1997). Complex fold trends on the Washington upper slope may be partially controlled by mobilization of the m61ange and broken formation (McNeill et al. 1997). Structural styles are more varied on the continental shelf, where many of the Miocene to early Pleistocene structures are no longer active (Fig. 1; Goldfinger et al. 1992a). Active fold axes on much of the inner shelf trend perpendicular to the coastline and margin (Goldfinger et al. 1992b; Goldfinger 1994). These fold trends suggest N-S compression rather than dominant northeasterly compression within the active accretionary prism.
Active crustal structures and coastal subsidence Introduction
To date, coastal subsidence in the form of rapidly buried marshes has been identified at the following bay and river locations along the Cascadia subduction zone, from north to south: Vancouver Island sites (Ucluelet and Tofino), Neah Bay, Pysht River, Copalis River, Grays Harbor, Willapa Bay, Columbia River, Necanicum River, Nehalem Bay, Tillamook Bay, Netarts Bay, Nestucca Bay, Salmon River, Siletz Bay, Yaquina Bay, Alsea Bay, South Slough, Coquille River, Sixes River, and Humboldt Bay (Fig. 1; based on Atwater et al. 1995; Barnett 1997). Evidence of Quaternary structural downwarping is described or documented here at Grays Harbor, Willapa Bay, Nehalem Bay, Tillamook Bay, Netarts Bay, Siletz Bay, Yaquina Bay, Alsea Bay, South Slough, Coquille River, and Humboldt Bay (Table 1). At these locations, buried marshes are located within Quaternary synclinal folds or on the downthrown side of faults. Buried marsh locations where evidence of Quaternary deformation is inconclusive are also documented, along with two examples of buried marshes which may be located on the upthrown side of a Quaternary fault or on the crest of an anticline. Table 1 describes the Quaternary structures at each of these locations.
B u r i e d m a r s h e s in areas o f late Q u a t e r n a r y subsidence Grays Harbor. Structure contours of the late Pliocene-Pleistocene unconformity west of Grays Harbor indicate a narrow E-W to NE-SW trending depression (Fig. 3). This active syncline lies due west of Grays Harbor, which may be structurally controlled. Buried marshes have been found throughout the bay (Fig. 3) and may not all be associated with downwarping of this particular syncline. This depression is tectonically controlled and not the simple result of backfilling of a Pleistocene lowstand channel. Willapa Bay. Buried marshes and drowned forests provide evidence of rapid subsidence throughout Willapa Bay, southern Washington coast (Figs 1 and 3; e.g. Atwater 1987; Atwater & Hemphill-Haley 1996; Yamaguchi et al. 1997). Multichannel seismic reflection profiles 10-30 km west of the bay reveal a broad syncline 40km in length, extending from the northern to the southern end of the bay (Figs 3 and 4). The synclinal axis trends NW-SE to E-W and projects into the centre of Willapa Bay. To the north and south of the syncline, two bounding anticlines may be underlain by faults with reverse separation (strike-slip component unknown) (Fig. 3). The northern N-dipping reverse fault may be blind (A1, Fig. 4). Alternatively, this fault may be the southeastern projection of a left-lateral fault which deforms Holocene sediments on the continental slope (South Nitinat Fault of Goldfinger (1994) and Goldfinger et al. (1997a); SNF in Fig. 1). The Willapa Bay structures deform Eocene to late Miocene m61ange and broken formation and overlying late Miocene to Quaternary basinal sediments, and gently deform the seafloor. The northern fault was investigated during a submersible and sidescan sonar cruise in 1995. A submersible dive found no evidence of seafloor offset, possibly as a result of high-energy wave action in shallow water or because the fault is blind. However, evidence of carbonate cementation and algal mats was observed, suggesting fluid venting accompanying active faulting, as observed elsewhere on the margin (Kulm & Suess 1990; Goldfinger et al. 1997a). The southern anticline (A2) and underlying fault may be inactive in the Quaternary (minimal deformation of the Pliocene-Pleistocene unconformity in Fig. 4). Structure contours of the late PliocenePleistocene unconformity also outline the position of the Willapa syncline (Fig. 3). The different positions of the late Pliocene (dashed line), the early Pleistocene (unconformity
UPPER PLATE DEFORMATION AND CASCADIA EARTHQUAKES
325
Table 1. Quaternary deformation at marsh burial locations. Site
Local structures
Quaternary
Tofino Ucluelet Neah Bay Pysht River
unknown unknown inconclusive yes
unknown unknown inconclusive unproven
Copalis River
yes Langley anticline
yes
M c C r o O, 1996
Grays Harbor
yes
yes
Willapa Bay Columbia River (incl. Youngs Bay) Necanicum River Nehalem Bay
yes yes, Fern Hill fault, Youngs Bay syncline inconclusive yes, Cape Falcon fault
yes unproven
This study; Grim & Bennett 1969; Palmer & Lingley 1989 This study Niem & Niem 1985; Ryan & Stevenson 1995
Tillamook Bay
yes, Tillamook Bay fault unproven
Netarts Bay
yes, Nehalem Bank and Happy Camp faults
yes
Nestucca Bay Salmon River Siletz Bay
yes inconclusive yes
no inconclusive yes
Yaquina Bay Alsea Bay South Slough
yes yes yes
yes yes yes
Coquille River Sixes River Humboldt Bay
yes
yes
yes, anticline
yes
Kelsey 1990
yes, Little Salmon and Mad River faults, Freshwater syncline
yes
Clarke & Carver 1992
inconclusive unproven
References
Snavely 1987 (Calawah fault) Gower 1960; Tabor & Cady 1978; Wagner et al. 1987
This study; Niem & Niem 1985; Goldfinger et al. 1992 This study; Niem & Niem 1985; Wells et al. 1992; Goldfinger et al. 1992; Goldfinger 1994 This study; Parker 1990; Goldfinger et al. 1992; Wells et al. 1992, 1994; Goldfinger 1994 This study; Goldfinger et al. 1992 This study; Goldfinger et al. 1992; Goldfinger 1994 Kelsey et al. 1990; Goldfinger et al. 1992 Kelsey et al. 1990 McInelly & Kelsey 1990; Goldfinger et al. 1992; Goldfinger 1994 This study; McInelly & Kelsey 1990
Presence or absence of localized Quaternary deformation at marsh burial locations documented here and in published literature. Local structures are synclines or downthrown sides of faults indicating subsidence, except for those shown in italics, which show evidence of uplift. (See Fig. 1 for locations.)
structure contours), and the most recently active synclinal axis (continuous line) indicate that the fold axis has migrated north from Pliocene time to the present. Early Pliocene growth strata thicken towards the southern anticline (Fig. 4 A2), whereas overlying late Pliocene strata thin towards the anticline, indicating that growth of this anticline occurred during the late Pliocene. Before this time, the syncline was formerly a larger structure extending to the south. The position of synclinal folding with respect to Willapa Bay strongly suggests that the bay is structurally controlled. Exposures of Pleistocene marine terraces and underlying sediments are widespread and continuous throughout the central region of Willapa Bay. Amino acid racemization of shell deposits within the lowest altitude marine terrace sedi-
ments of the central bay suggest conflicting ages. Kvenvolden et al. (1979) identified two separate highstands, with absolute ages of 120 ka BP and 190 ka 4- 40 ka BP. Amino acid ages with supporting faunal correlations and palaeoecology from Kennedy (1978) and Kennedy et al. (1982) suggest a more probable and precise age of 8085kaBP (West 1986). Despite the unresolved terrace chronology, relative altitudes of continuous and correlative terraces have been documented. The lowest terrace surface is reported to be at a constant altitude of c. 13-15 m above sea level in the central region of the bay (West 1986). However, reconnaissance investigation of terrace elevations for this study found that the terrace actually reaches a minimum elevation of c. 9 - 1 0 m at Sandy Point (Sa, Fig. 3). There are no signs of landslides or slumping in this region,
326
L . C . McNEILL E T A L .
Fig. 3. Map of southern Washington coast and offshore region. Mapped shelf structures associated with the Willapa Bay syncline are active (late Quaternary, continuous lines) and Pliocene-early Pleistocene (dashed lines). Structure contours of a Pliocene-Pleistocene unconformity outline Quaternary synclinal deformation (contours in metres below sea level, 200 m interval, black hatching indicates the unconformity is truncated at the sea floor). Seismic tracklines used are represented by dotted lines. CR, Cedar River; N, Nemah; Sa, Sandy Point; St, Stony Point; BR, Bone River. Early Pleistocene palaeosols dip to the south between Stony Point and Bone River. Onshore structures from Walsh et al. (1987) and Atwater & Hemphill-Haley (1996); coastal subsidence locations after Atwater & Hemphill-Haley (1996). Bold line shows location of Fig. 4; A1, A2 and S are identified. Reproduced by kind permission of GSA.
and therefore terrace elevations are assumed to be undisturbed. In the northern bay, the lowest terrace, which West (1986) assumed to be the continuation of the central bay's lowest terrace (80-85kaBP?), increases in altitude to 24m (West 1986). South of Nemah (N, Fig. 3), limited exposures of terraces of unknown age range in altitude from 12 to 18 m. These altitude changes, although only loosely constrained by age control, are in agreement with a synclinal axis close to the centre of the bay. The stratigraphy of the estuarine sediments underlying the
lowest and youngest central bay terrace shows no clear evidence of deformation along a N-S transect (Kvenvolden et al. 1979; Clifton 1983, 1994; H. E. Clifton, pers. comm., 1997). However, a unit of older (early Pleistocene or late Pliocene?) well-indurated estuarine terrace sediments including a sequence of palaeosols, locally exposed within the youngest terrace, may be deformed (Fig. 5; Clifton 1994; H. E. Clifton, pers. comm., 1997). Between Stony Point and the Bone River (Fig. 3), these surfaces dip gently to the south (Fig. 5; Clifton 1994). If these surfaces were originally approximately horizontal, they are now deformed in agreement with the projection of the offshore projected synclinal axis. The W N W trend of Quaternary structures offshore contrasts slightly with NW- to NNWtrending structures which deform Miocene and older formations mapped east of Willapa Bay (Fig. 3; Walsh et al. 1987). This change in trend may reflect a change in compressive stress field from the early to late Tertiary, or non-linearity of mapped structures. One of two sub-parallel NW-trending synclinal axes southeast of the bay (Fig. 3; Walsh et al. 1987) could be the landward projection of the offshore structure. Atwater & Hemphill-Haley (1996) addressed the possibility of deformation by upper-plate structures in Willapa Bay by studying the marsh stratigraphy on and off the N-S trending South Bend antiform (Fig. 3). They found no difference in the amount of subsidence between these sites, although marshes are buried less deeply below the current marsh surface on, rather than off the antiform (Atwater & Hemphill-Haley 1996). However, this antiform deforms sediments no younger than Miocene and may not be currently active. An E - W trending structure, such as the Willapa Bay syncline, should show no measurable difference in amount of subsidence along an approximately E - W transect, and therefore the results of Atwater & Hemphill-Haley (1996) are not unexpected. The offshore northern bounding anticline and thrust fault project landward towards one location of marsh burial in the northernmost bay (Fig. 3). Evidence for subsidence at this site (Cedar River (CR in Fig. 3) of Atwater & Yamaguchi (1991)) is a western red cedar snag which died within a few months of AD 1700. This apparent contradiction can be explained if the crustal structure was not activated during the AD 1700 event, or if subsidence caused by the subduction event exceeded coseismic uplift on the crustal structure with resulting net subsidence, or if the fault and anticline change strike onshore to project to the north of this particular
UPPER PLATE D E F O R M A T I O N AND CASCADIA EARTHQUAKES
327
Fig. 4. N-S proprietary migrated multichannel seismic reflection profile 10 km west of Willapa Bay. The lower interpreted section shows late Miocene to Quaternary sediments and underlying m61ange are deformed by an active syncline and associated structures. To the north, Pleistocene sediments are deformed by an active N-dipping blind reverse fault. Regional angular unconformities: M, P, late Miocene-Pliocene; Plio-Pleist, Pliocene-Pleistocene.
subsidence site. In addition, the absolute vertical motion may differ from the observed relative vertical motion across this anticline. Seismic reflection profiles west of Nehalem Bay reveal synclinal deformation opposite the bay and uplift on the north side of the bay which may be equivalent to the Tertiary Cape Falcon fault of Niem & Niem (1985) (Fig. 6). The middle Miocene Columbia River Basalt (CRB in Fig. 7) is deformed by these structures and exposed at Cape Falcon (Niem & Niem 1985). The northern margin of Tillamook Bay is uplifted by the WNW-trending high-angle Tillamook Bay fault, which offsets Columbia River Basalt onshore (Wells et al. 1992). Offset on this N e h a l e m and Tillamook Bays.
Fig. 5. Early Pleistocene terrace sediments between Stony Point (St) and Bone River (BR) in Willapa Bay showing S-dipping palaeosols (after Clifton 1994; reproduced by kind permission of GSA), in agreement with the projection of the offshore synclinal axis to a position south of this location. (See Fig. 3 for locations.)
Fig. 6. Location map of northern Oregon coast and shelf, including active structures near Nehalem, Tillamook, and Netarts Bays. The offshore Nehalem Bank fault and associated syncline projects onshore as the Happy Camp fault (HCF) on the north side of Netarts Bay (Parker 1990; Wells et al. 1992, 1994). Tillamook and Nehalem Bay may also lie within active synclines or on the downthrown side of active faults (Tillamook Bay fault of Wells et al. (1992); Cape Falcon fault of Niem & Niem (1985)). Onshore structures after Parker (1990), Wells et al. (1992, 1994), and Niem & Niem (1985). Bold line represents N-S seismic reflection profile shown in Fig. 7.
E
G" ~
~ o Z
Z
..~
~
~-
0
N2Z~ ~Z
0.~
"
~-~ ~.% ,.~ ~ . ~ ~; ~ . ~ =
~'
~
~
~ ~
"-" ~ 0
~
o
c.~ " ~ .,,..~
"~
o ~ o ~o~
-~~.~ ~.~=o
9= Z
o'=~
o
Z ~~=, . ro = ~ 0 , . ~ ~ r-=~.>
UPPER PLATE DEFORMATION AND CASCADIA EARTHQUAKES fault is up to the north and may have a strikeslip component, but Quaternary deformation is unconfirmed. Deformation opposite Tillamook Bay is more complex than at Nehalem Bay, but a broad low between the Happy Camp fault at Cape Meares and an uplifted region at Twin Rocks can be seen in Fig. 7. This low is interrupted by smaller fold axes. The uplifted region north of the bay may be related to the Tillamook Bay fault (Figs 6 and 7). Owing to the thin Quaternary section on the innermost shelf, offshore Quaternary deformation cannot be confirmed, but the onshore Tillamook Bay fault suggests Quaternary activity. Rapid marsh burial has been identified at both bays (Atwater et al. 1995; Barnett 1997). Netarts Bay. Netarts Bay, on the northern Oregon coast (Fig. 1), is bounded to the north by WNW-trending, NE-dipping high-angle reverse faults which deform coastal sediments, similar in style to Tillamook Bay to the north (Wells et al. 1992, 1994). Faults thrust middle Miocene Columbia River Basalt Group (15 Ma) over late Pleistocene gravels, and may also offset the youngest Pleistocene marine terrace surface (Wells et al. 1994). The fault is known as the Happy Camp fault onshore (Parker 1990; Wells et al. 1992, 1994) and is the southernmost extension of the prominent Nehalem Bank fault zone, which deforms Miocene to Holocene sediments offshore (Fig. 6; Niem et al. 1990; Goldfinger 1994). This complex zone of deformation trends roughly N-S on the outer shelf and upper slope off the northern Oregon coast (25 km west of the coastline), but changes to the southeast at its southern end to project onshore just north of Netarts Bay (Fig. 6). No evidence of major strike-slip offset has been found along the northern segment of the fault, but its orientation (suggesting margin-parallel rightlateral offset where oblique convergence is partitioned into a compressional and strike-slip component), the presence of minor strike-slip faulting, and the linearity of the fault, which truncates bedding planes in AMS 150 kHz sidescan images, support a strike-slip component. However, the fault also shows significant vertical offset (both along its N-S and SE-trending segments) and we interpret it as a reverse fault system downthrown to the west and south. As the fault trends southeasterly, the zone of deformation becomes less complex, being characterized by a N-dipping (possibly blind) reverse fault, with Miocene sediments uplifted and exposed at the sea floor, and an asymmetrical syncline to the south (south end of Fig. 7), which lies immediately opposite Netarts Bay. Sidescan
329
images show bedrock within the hanging-wall anticline exposed on the sea floor and offset by minor N- to NNE-trending right-lateral faults. The vertical motion on the northern segment of the fault may be a flower structure or transpressional deformation. The nearshore Nehalem Bank fault clearly deforms and offsets the middle Miocene Columbia River Basalt Group (highly reflective in seismic reflection data, south end of Fig. 7) and overlying sediments. Absence of structural growth of strata within the syncline in Fig. 7 suggests this fault and associated fold post-date late Miocene sedimentation. Investigation of other seismic reflection data across both the southern and northern sections of the fault shows minimal thinning in late Miocene sediments and some thinning of Pliocene sediments across the fault-controlled anticline. This indicates that the fault was active as early as the late Miocene, but the bulk of deformation has taken place during the Pliocene and Quaternary. Vertical seafloor offset of 10-20m across the
Fig. 8. Location map of Siletz Bay on the central Oregon coast showing structures mapped offshore from single-channel and multichannel seismic reflection profiles and onshore fi'om Pleistocene marine terrace deformation (structures in Fig. 9 crosssection). Bold line indicates position of single-channel line in Fig. 10. Onshore and offshore deformation suggests that Siletz Bay is structurally downwarped within a syncline or on the downthrown side of a fault. Generalized locations of subsided marshes after Darienzo et al. (1994).
330
L . C . McNEILL ET AL.
Fig. 9. Cross-section of beach exposure of Pleistocene sediments underlying the youngest marine terrace at Siletz Bay. Marker horizons, including a clay horizon, gravel beds, and the wavecut platform were used to determine deformation of Pleistocene sediments and the location of Quaternary structures. The laterally continuous clay horizon is apparently offset across the bay (up to the north). Faults also offset the wavecut platform at Fishing Rock and Fogarty Creek. The trends of these structures are poorly defined.
fault zone is estimated from sidescan and seismic records, presumed to post-date late Pleistocene lowstand erosion on the shelf (Goldfinger 1994). Cooper (1981) and Parker (1990) also suggested that a west-plunging Miocene syncline is centred about Netarts Bay. Siletz Bay. Structures deforming the underlying sediments and wavecut platform of Pleistocene marine terraces (presumed 80 ka BP Whiskey Run terrace, West & McCrumb 1988; Kelsey 1990) have been identified in the Siletz Bay region of the central Oregon coast (Figs 1, 8 and 9). The wavecut platform and a locally continuous carbonaceous clay horizon, interpreted as a lagoonal deposit or Palaeosols and assumed to be initially sub-horizontal, were shown to be deformed. Variations in elevation of the wavecut platform and the clay horizon/ palaeosol may alternatively be controlled by existing topography at the time of formation, and not by deformation. Variations in altitude of these marker horizons indicate faulting, with vertical offsets of 5 - 3 0 m and broad folding,
with a wavelength of 8 - 1 2 k m (Figs 8 and 9), assuming the terrace is the same age throughout. Beach exposures alone indicate trends between N N W and SSW, but offshore data (see below) provide more precise trends. The clay horizon of the youngest terrace (80 ka BP) dips gently north between 5 km and 3 km south of the Siletz River mouth, where it is below beach level and projected below sea level (Fig. 9). This clay horizon is exposed again c. 10m above beach level just north of the river mouth, where the wavecut platform is at c. 2 m elevation (Fig. 9). The platform is presumed below sea level south of the river, inferred from the elevation and dip of the clay horizon. Projection of the clay horizon below beach level suggests maximum fault offset across the Siletz River mouth of c. 30m up to the north (offset could be a combination of folding and faulting). If this horizon is assumed to be the same age as the terrace (80 ka BP), this produces a late Quaternary vertical slip or subsidence rate relative to terrace levels across the river mouth of 0.4mm/ year. This is a maximum slip rate as sediments
Fig. 10. Line drawing of N-S trending OSU single-channel sparker profile, 6 km offshore Siletz Bay (bold line in Fig. 8). The MP (late Miocene-early Pliocene) unconformity is projected to the sea floor at the southern end of the profile. Profile shows synclinal deformation of presumed late Miocene strata off Siletz Bay, and sea floor offset by possible flexural-slip faults within an active synclinal fold west of Fishing Rock and Fogarty Creek. Fault dips are poorly constrained by seismic data.
UPPER PLATE DEFORMATION AND CASCADIA EARTHQUAKES underlying the terrace are somewhat older than 80kaBp. Siletz Bay may lie in a Quaternary syncline controlled by a fault at the northern end of the bay, similar to the structure observed at Netarts Bay. Fault offset (downthrown to the north) of the wavecut platform and marine terrace was also documented at Fishing Rock and Fogarty Creek (Fig. 8; Priest et al. 1994). Orientation of these two faults is poorly defined, but previously mapped faults onshore are oriented NW-SE and NE-SW. Poor exposure prevents the determination of any strike-slip component on onshore faults. The possible correlative of a syncline at Siletz Bay is traced on N-S trending seismic reflection profiles 4-17km offshore. Figure 10 is a line drawing of a N-S single-channel seismic profile 6kin west of Siletz Bay, which clearly shows synclinal deformation opposite the bay. The late Miocene-early Pliocene (MP) unconformity is truncated at the sea floor and therefore the age of the youngest strata is late Miocene. The sea floor indicates no synclinal deformation, therefore offshore Quaternary deformation at this scale is unconfirmed. The trend of the syncline across several profiles is between E-W and ESE-WNW. Deformed synclinaI sediments on middle- to outer-shelf profiles are truncated by the MP unconformity, indicating little or no activity since the late Miocene or early Pliocene time to the west; however, this unconformity is deformed by the southern bounding anticline which may project into the Gleneden Beach area. Onshore deformation of late Pleistocene marine terraces by the syncline points to the recent activity of these structures. Possible flexural-stip faults north of the river mouth (Fig. 8) were poorly imaged in the single-channel sparker profile and therefore have uncertain offset or dip. Faults with similar offset to those at Fishing Rock and Fogarty Creek (Fig. 9), identified in single-channel sparker lines between 2 and 6km offshore, may also be flexural-slip faults (Fig. 10). A flight of uplifted Pleistocene marine terraces is preserved north and south of Yaquina Bay (Figs 1 and 11) on the central Oregon coast. These terraces have been differentiated by age using amino acid enantiomeric (D:L) ratios in conjunction with the palaeoecology of fossil shells (Kennedy 1978; Kennedy et al. 1982) and a soil chronosequence (Ticknor et al. 1992; Ticknor 1993; Kelsey et al. 1996). These techniques indicate offset of marine terraces on the inferred Yaquina Bay fault of 75 m down to the south (Ticknor 1993; Kelsey et al. 1996). The fault juxtaposes the 80 ka BP terrace Yaquina and Alsea Bays.
331
(Qn) north of the bay against the 125 ka BP (Qy) terrace south of the bay (Fig. 11; Kelsey et al. 1996). This offset yields a slip rate of 0.6 ram/year. The continuation of the Yaquina Bay fault to the east was mapped by Snavely (1976), giving an ENE fault orientation (Kelsey et al. 1996). All core locations of buried marshes identified by Peterson & Priest (1995) are located on the south or downthrown side of the Yaquina Bay fault. Similar studies at Alsea Bay (Ticknor et al. 1992; Ticknor 1993; Kelsey et al. 1996) show that Quaternary faults strike generally N-S. The N-S striking Waldport fault zone vertically
Fig. 11. Central coastal Oregon between Yaquina Head and Alsea Bay, showing late Pleistocene marine terrace backedges and Quaternary faulting (after Kelsey et al. 1996). The onshore Yaquina Bay fault trends NE and downdrops marine terraces to the south. The Waldport fault zone (including the Lint Slough fault) downdrops terraces to the east, with greatest offset at Alsea Bay. Marine terraces in order of increasing age based on soil development chronology (Kelsey et al. 1996): Qn, Newport; Qw, Waconda; Qy, Yaquina; Qc, Crestview; Qfr, Fern Ridge; Qag, Alder Grove. Qh, Holocene beach and dune sands. 9 General subsided marsh locations (after Darienzo et al. 1994; Peterson & Priest 1995; Peterson & Darienzo 1996). Reproduced by kind permission of GSA.
332
L . C . McNEILL E T AL.
displaces terrace platforms down to the east (Fig. 11), with cumulative offset apparently greatest at Alsea Bay, suggesting a structural origin for this embayment (Kelsey et al. 1996). All rapid subsidence sites are on the downthrown side of the Waldport fault zone. Kelsey et al. (1996) concluded, from the evidence of Pleistocene terrace deformation, that both Yaquina and Alsea Bays are downwarped and structurally controlled by faults. Offshore data neither support nor refute the onshore terrace evidence. South Slough. Many active structures with N - S trends on the southern Oregon coast and shelf, where the deformation front is closer to the coastline, are interpreted to be part of accretionary prism-related deformation (Fig. 12). One example is the South Slough syncline which deforms Quaternary sediments southwest of Coos Bay on the southern Oregon coast (Fig. 12; Nelson 1987; Peterson & Darienzo 1989; Kelsey 1990; McInelly & Kelsey 1990) and may have produced multiple buried peats as an independent local structure (e.g. Nelson & Personius 1996). The syncline has been traced onto the shelf on seismic reflection profiles (Fig. 12; Goldfinger et al. 1992a; Goldfinger 1994). Both offshore and onshore deformation suggests that many faults are flexural-slip faults bounding active folds, such as the South Slough syncline, with fault slip parallel to bedding planes (McInelly & Kelsey 1990; Goldfinger 1994).
Coquille River. The Coquille fault (Fig. 12; Clarke et al. 1985; Goldfinger 1994) comes onshore just south of the Coquille River mouth, where it deforms Pleistocene marine terraces (McInelly & Kelsey 1990). The Whisky Run (80 ka BP) platform descends from an altitude of 35 m at Cape Arago to sea level just north of the Coquille River, deformed by the Pioneer anticline (Fig. 12; McInelly & Kelsey 1990). The terrace abruptly gains altitude to 18 m above sea level just south of the Coquille River at Coquille Point (Fig. 12). This altitude change is accompanied by a change in dip of platforms from southwest north of the river to west or seaward south of the Coquille River (Fig. 12; McInelly & Kelsey 1990). The Whisky Run platform is tilted slightly landward at Coquille Point, possibly resulting from deformation by the Coquille fault. The terrace elevation descends once again south of Coquille Point to reach sea level c. 10 km to the south (McInelly & Kelsey 1990). Fold trends in Tertiary and Mesozoic formations underlying the Whisky Run wavecut plat-
Fig. 12. Location map of the south-central Oregon coast and shelf showing Pliocene and Quaternary structures. Structures onshore based on deformed late Pleistocene terraces mapped by Kelsey (1990) and McInelly & Kelsey (1990). Offshore structures adapted from Goldfinger (1994), from seismic reflection and sidescan sonar data. Typical fractures and shears (with strike-slip offset) within Miocene and older sediments uplifted in the complex fault zone are also shown. Elevation contours of the Whisky Run wavecut platform (Qwr) shown in 10 m intervals (from McInelly & Kelsey 1990) indicate deformation by the Pioneer anticline and the Coquille fault. Coastal terrace altitudes are highest at Cape Arago (CA) and at Coquille Point (CP). Reproduced by kind permission of GSA. form also vary from north to south across the Coquille fault. North to south, fold trends are generally consistent with the variations in dip of the Pleistocene terraces and offshore Pleistocene fold and fault trends (Fig. 12). No strike-slip offset or recent deformation could be determined on onshore faults in Jurassic to Eocene strata. In seismic reflection profiles offshore, the fault zone appears as a ridge flanked by synclinal folding deforming the sea floor. Clarke
UPPER PLATE DEFORMATION AND CASCADIA EARTHQUAKES
et al. (1985) suggested the fault downdrops Pleistocene sediments to the northeast on the innermost shelf as observed in the deformed onshore marine terraces, but other seismic data across the fault indicate a fairly symmetrical ridge. Humboldt Bay. Two large thrust fault systems, the Little Salmon fault and the Mad River fault zone, deform Holocene sediments in the Humboldt Bay region of northern California (Fig. 1), and are interpreted to be part of the onshore expression of the southern Cascadia accretionary prism (Clarke & Carver 1992). The Freshwater syncline lies between these two fault systems within Humboldt Bay and Mad River Slough, and has produced Holocene subsidence resulting in stacked buried marsh and forest horizons (Clarke & Carver 1992). The most recent subsidence event is dated at 250-300 radiocarbon a BP, contemporaneous with the most recent event recorded throughout much of the subduction zone. Locations with inconclusive evidence for Quaternary subsidence Other abrupt subsidence sites are characterized by inconclusive evidence or no evidence of Quaternary deformation, in the form of crustal downwarping. Marsh burials have been found in Neah Bay (Waatch River) and the Pysht River on the northernmost Olympic Peninsula (Fig. 1; Atwater 1992). No active faults have been mapped and directly linked to evidence of rapid subsidence in Neah Bay. Buried marshes at Pysht River (Atwater 1992) lie on the downthrown side of a high-angle fault through Tertiary formations (Gower 1960; Tabor & Cady 1978). This fault is mapped parallel to the river (NE), downthrown to the west, and projects offshore to a similar fault in the Strait of Juan de Fuca which deforms an acoustic unit of Holocene age (Wagner & Tomson 1987). A second fault trends parallel to the coast (WNW) and projects into the bay. Buried marshes may lie on the upthrown side of this fault, but there is no evidence that it is recently active. High-resolution seismic profiles along the lower reaches of the Columbia River (Ryan & Stevenson 1995) indicate possible evidence of Quaternary faulting, including the NE-trending Fern Hill fault, which offsets Miocene Astoria Formation onshore (Niem
333
& Niem 1985). These faults have not yet been linked directly to locations of rapid marsh burial or liquefaction in the Columbia River (Atwater 1992, 1994; Obermeier 1995). Niem & Niem (1985) also mapped a WNW-trending syncline through Youngs Bay (Fig. 1), south of Astoria, where a drowned forest and rapidly buried marshes have been identified (Peterson et al. 1997, C. D. Peterson, pers. comm., 1997). Quaternary deformation across this structure remains unproven. An older syncline (deforming late Miocene and possibly Pliocene strata but with no conclusive evidence of Quaternary deformation) has been mapped on the outer shelf opposite Nestucca Bay (Fig. 1). No evidence of Pleistocene deformation has been reported onshore, but it has been suggested that Cape Kiwanda, the headland to the north of the bay which is composed of Miocene Astoria Formation and Smugglers Cove Formation, may be a structural high (Parker 1990). Evidence of Quaternary downwarping at Vancouver Island sites, the Necanicure River, and Salmon River also remains inconclusive, to judge from the available data.
Marsh burial located near structural uplifts Two possible exceptions to the hypothesis that buried marshes lie within tectonic downwarps are the Sixes River, southern Oregon, and the Copalis River, central Washington (Fig. 1). Kelsey (1990) mapped an E-W trending anticline, which deforms Pleistocene terraces, just north of Cape Blanco and coincident with the lower reaches of the Sixes River (Fig. 1). Buried marshes and tsunami sands have been identified on the southern limb of this anticline (Kelsey et al. 1993, 1998) in a cutoff meander of the Sixes River. Further examination of marshes in a N-S transect across the southern flank of the anticline may reveal differential subsidence (H. M. Kelsey, pets. comm., 1997). This apparent anomaly of local uplift and regional subsidence could be explained by subsidence being the net result of local uplift and regional subsidence, by the anticline not being triggered by every subduction zone earthquake, or by the fact that the anticline was active in Pleistocene but not during Holocene time. A second possible exception is an ENE-trending ridge (Langley ridge) located 5 km south of the Copalis River. Deformation is interpreted as anticlinal folding and diffuse faulting above a possibly N-dipping blind thrust fault (McCrory 1996), with buried marshes at the Copalis River on the upthrown or north side of this fault. In addition to buried marshes, Atwater (1992)
334
L . C . McNEILL ET AL.
identified liquefaction evidence from 900-1300 years ago at Copalis River with no indication of accompanying tectonic subsidence. This event could be attributed to movement on a local crustal thrust fault such as that underlying Langley Ridge. The dip of a blind fault is often difficult to determine from geomorphology and surface faulting, and this thrust fault may, in fact, dip to the south, with the Copalis River marshes lying on the downthrown side of a fault. Alternatively, a syncline to the north of this ridge may coincide with the Copalis River. This is supported by N-S single-channel and multichannel seismic profiles which indicate a series of closely spaced approximately E-W trending ridges and intervening synclines on the continental shelf. The Copalis River buried marshes may in fact lie within one of these synclines rather than being associated with the Langley Ridge structure 5 km to the south.
Discussion Implications f o r the Cascadia subduction zone earthquake record
The record of prehistoric subduction earthquakes on the Cascadia subduction zone, in the form of rapidly buried marshes, documents sudden submergence, inundation of coastal lowlands, and burial of the former land surface. Correlation of coseismic events between coastal bays, based on radiocarbon ages and dendrochronology, has allowed rupture lengths, magnitudes, and recurrence intervals of prehistoric Cascadia earthquakes to be proposed. In addition, estimates of amounts of coastal subsidence can be used to approximate the position of the rupture zone and earthquake magnitude using elastic dislocation modelling. The possible nontectonic origin of some submergence events should, however, be considered when assessing potential earthquake hazards. In this study, it has been demonstrated that many abruptly buried marsh locations can be linked to Quaternary structures (synclines and downdropped side of faults) which produce downwarping. The influence of upper-plate crustal deformation on the prehistoric earthquake record may lead to inaccuracies in calculations of magnitudes and recurrence intervals if based on the Holocene stratigraphy of coastal bays. Evidence of Quaternary deformation offshore is equivocal in some cases, but the use of offshore datasets and coastal exposures together has increased the probability of identifying recent activity. Only two possible examples of
rapid subsidence coincident with crustal uplift were identified. However, these apparent anomalies could be explained by active upper-plate structures not deforming during every subduction event. Localised upper-plate deformation at other subduction zones. Localized upper-plate deformation has been documented at subduction margins world-wide, with deformation both synchronous with and independent of subduction zone events. Examples include the Hikurangi margin of New Zealand (Berryman et al. 1989; Cashman & Kelsey 1990; Berryman 1993a, b), the Alaskan margin (Plafker 1972), the Nankai forearc of SW Japan (Maemoku 1988a, b; Maemoku & Tsubono 1990; Sugiyama 1994), and the Huon peninsula of Papua New Guinea (Pandolfi et al. 1994). Holocene terraces along the coastal Hikurangi margin off eastern North Island, New Zealand, are uplifted by movement on steep reverse faults of the onshore accretionary prism (Berryman et al. 1989), with clustering of terrace ages along the coast. Stratigraphic and ecological studies of Holocene terrace sediments on the Mahia Peninsula reveal that sedimentation was progradational between events, implying a lack of interseismic subsidence that would be expected with a subduction earthquake cycle and supporting the formation of Holocene coseismically uplifted terraces by local crustal structures (Berryman et al. 1997). Other earthquakes within the accretionary prism include the 1931 Hawkes Bay earthquake (Ms 7.8), caused by a fault cutting up from the megathrust (Hull 1990), and the 1855 Wairarapa earthquake, which may have originated on the megathrust and propagated into the upper plate along a blind thrust fault (Darby & Beanland 1992). In Alaska, significant deviations from the regional subsidence or uplift patterns during the 1964 earthquake (up to 12 m uplift across the Patton Bay fault on Montague Island, relative to a regional 2-4 m of uplift) were associated with movement on crustal faults contemporaneous with the subduction zone earthquake (Plafker 1969, 1972). Along the Nankai margin of Japan, two types of subduction earthquake have been inferred from coseismically uplifted terraces (Fig. 13; Sugiyama 1994). Subduction events where no permanent crustal deformation and therefore no uplifted terrace preservation occurred are known as Taisho type events (T, Fig. 13). Preserved uplifted terraces resulting from the triggering of crustal deformation are known as Genroku type events (G, Fig. 13).
UPPER PLATE D E F O R M A T I O N AND CASCADIA EARTHQUAKES
~
G I bench
levell T1
~,~
G 2 bench / s ....
T
~
'
"
~'
'
T2
G1
T3
T4
G2
T5
time
Fig. 13. Illustration of the resulting record of Taisho and Genroku subduction earthquakes on the Nankai margin (after Sugiyama 1994; reproduced by kind permission of Geofisica International). No permanent inelastic crustal deformation occurs during Taisho events (Tl-5); coseismic deformation is recovered in the interseismic period leaving no permanent record of the earthquake. Genroku events (G1, G2) involve local faulting or other inelastic crustal deformation leading to preservation of an uplifted bench. It should be noted that the coseismic and interseismic vertical motions are opposite to those expected on much of the Cascadia coastline.
Effects o f local crustal deformation on subsidence records. Recorded coseismic subsidence is the net result of regional elastic strain release from a subduction zone earthquake and local crustal deformation (permanent and/or elastic), assuming a tectonic origin for subsidence. For each event, subsidence could result from strain release on the plate boundary or on local structures, or a combination of the two. The contributions of each cannot be determined for prehistoric events, although the amount and pattern of subsidence at each location may suggest a particular mechanism. The apparent rupture length (and hence magnitude), amount of subsidence, and timing of coseismic events can be specifically affected in the following ways. (1) Triggering of local crustal faults beyond (along strike) the subduction rupture zone, as the release of the elastic load on the upper plate in one area causes loading in other areas, thereby increasing the apparent rupture length and magnitude (e.g. LS2 and SZE1 in Fig. 14). Similar patterns occurred during the Landers earthquake, where a sequence of delayed ruptures occurred within the fault zone (Sieh et al. 1993; Wald & Heaton 1994; Spotila & Sieh 1995), and more distant earthquakes were also triggered following the mainshock (Hill et al. 1993), although these events lacked surface rupture and geodetic change. (2) The amount of subsidence per event at each location is dependent not only on the magnitude of the subduction zone earthquake and position of the rupture zone, but also on the amount of localized upper-plate deformation
335
accompanying the earthquake. The amount of subsidence during a given earthquake cannot be used for elastic dislocation modelling of the locked zone. (3) Local crustal faults may move independently of subduction zone earthquakes and produce anomalous local coseismic subsidence and marsh burial (LS1 of Fig. 14). Correlation of subsidence events from site to site is dependent on age control with sufficient precision to distinguish such events. The majority of radiocarbon ages from marsh burials have large error bars, of the order of +50 years to hundreds of years (e.g. Atwater 1992; Atwater et al. 1995), with errors often larger than the suggested recurrence intervals. More recently, AMS and high-precision radiocarbon and dendrochronology ages in some locations have significantly reduced errors to 4-10-20 years or even to within a year or season (Nelson et al. 1995; Jacoby et al. 1997; Yamaguchi et al. 1997), but suitable material for such precise dating techniques is often unavailable (Nelson et al. 1996b). The most abundant high-precision data are available for the most recent subsidence event, which is dated within a few decades of AD 1700, and is consistent with evidence for a remote tsunami in Japan at that time (Satake et al. 1996). Older
Fig. 14. Coastal marsh stratigraphy in hypothetical cores showing subsided marsh and soil deposits produced by a variety of coseismic events. LS, localized subsidence; SZE, subduction zone earthquake. SZE1 represents a small subduction earthquake which triggers localized subsidence to the north, increasing the apparent rupture length. SZE2 (Taisho type) does not trigger inelastic deformation and does not preserve coseismic subsidence. SZE3 is a widespread subduction event with subsidence at all four locations. This earthquake may also trigger local structures which may contribute to subsidence and help preserve the buried soils (Genroku type).
336
L . C . McNEILL E T AL.
events are less accurately dated. Where error bars are large, it is impossible to distinguish between regional subsidence events and locally anomalous events which may have resulted from independent deformation on crustal structures. These anomalous events may be wrongly correlated with regional coseismic subsidence events, producing an inaccurate picture of the earthquake rupture zone. In addition, subduction earthquake recurrence intervals may be underestimated if independent local subsidence events contribute to the marsh stratigraphy. Synchronous v. independent movement of local structures and subduction earthquakes. It is unknown, in the absence of historic subduction zone earthquakes and little upper-plate seismicity, whether local crustal structures in Cascadia are triggered by subduction zone earthquakes or operate independently, or both. If movement on crustal structures is always synchronous with and triggered by subduction events, estimates of the subduction earthquake recurrence interval will be unaffected but magnitude calculations may be inaccurate. If these structures operate independently, deforming both during and between subduction events, both the magnitude and recurrence interval of subduction zone events will be affected. If patterns of strain release are similar to those of the Alaskan and Nankai subduction zones, we might expect crustal structures to be triggered by slip on the megathrust (PlaNer 1969, 1972; Sugiyama 1994). Minimal historic seismicity in the coastal and shelf region supports the hypothesis that these structures are predominantly triggered by subduction zone earthquakes, which also lack seismicity. Buried marshes similar in age to regional subsidence events have been attributed to upper-plate structures in South Slough, southern Oregon coast, and Humboldt Bay, northern California coast, with little visible evidence of significant rapid subsidence in other bays in this region, such as the Siuslaw River (Clarke & Carver 1992; Nelson 1992; Nelson & Personius 1996; Nelson et al. 1996a). If rapid subsidence is not regionally extensive, these are examples of crustal structures that were triggered by subduction zone events. Regional subsidence in this area may be small and only detectable by biostratigraphic investigations (compare with Mathewes & Clague (1994)), and pronounced sudden subsidence may only be recorded where local structures were triggered. Upper-plate structures are likely to have longer recurrence intervals than subduction zone earthquakes and may not be triggered by every subduction event.
Local structures may also produce tectonic subsidence independently of subduction events. Independent coseismic subsidence has been suggested as a likely cause of marsh burials on the southern Oregon and northern California coasts (Nelson 1992; Nelson & Personius 1996).
Preservation o f buried marshes. The sequence of Cascadia buried marshes indicates net submergence of the land or net relative sea level rise of 2-5m in the last 2000-4000 years. If the earthquake strain cycle were completely elastic and no other factors were involved, coseismic subsidence and interseismic uplift would cancel out and no buried marshes would be preserved. This argument is used for the Nankai subduction zone, where coseismically uplifted terraces are only preserved permanently when synchronous upper-plate uplift occurs (Sugiyama 1994). If similar patterns of deformation to those at Nankai occur along the Cascadia subduction zone, permanent deformation by local structures may help to preserve marsh burial. Not all subduction zone earthquakes would be recorded (e.g. SZE2 in Fig. 14) and subduction zone earthquakes would appear to be less frequent with longer recurrence intervals. One major difference between the Nankai and Cascadia coastlines is the sense of coseismic motion: Nankai experiences uplift whereas Cascadia experiences subsidence. Therefore, preservation of buried marshes in Cascadia may also be influenced by the following nontectonic factors, producing relative sea-level rise; (1) late Holocene eustatic sea-level rise; (2) isostatic forebulge collapse following the last glacial maximum; (3) compaction; or (4) changes in the geometry of coastal estuaries (this could cause relative sea-level rise or fall). These factors would not contribute to the preservation of uplifted terraces in Nankai, unless sea level was falling in late Holocene time, or this region experienced some form of isostatic uplift. Buried marsh preservation in Cascadia could also be attributed to regional tectonic subsidence resulting from, for example, subduction erosion. The rates and effects of these factors in late Holocene time are poorly known and therefore their contributions to the preservation of buried marshes can only be approximated. The rate of late Holocene global eustatic sea-level rise is hotly debated, with sorne estimates pointing to negligible rise during the last 5000 years (P. Clark, W. R. Peltier, pers. comm., 1997), very low rates (Clark & Lingle 1979; Bard et al. 1996), or a value which is currently very difficult to separate from local and regional factors,
UPPER PLATE DEFORMATION AND CASCADIA EARTHQUAKES including isostatic and tectonic factors, which dominate relative sea-level rise (Bloom & Yonekura 1990; Nelson et al. 1996b). Estimates of forebulge collapse on much of the Cascadia margin associated with isostatic re-equilibration following the last glacial maximum (LGM) are given by the models of Peltier (1996). Subsidence rates (or relative sea-level rise) for much of the Washington and Oregon coastline are estimated as 0-1 ram/year, with a maximum on the northern Oregon coast (M2 model of Peltier (1996)). However, the northern Olympic Peninsula and Vancouver Island, which underlay the Cordilleran ice sheet during the LGM, should be experiencing isostatic rebound. Tectonic subsidence may indeed dominate the preservation of buried marshes, but until other variables are better resolved, this hypothesis remains untested. Quantitative subsidence calculations. In general, subsidence patterns measured along the Cascadia subduction zone appear to be fairly consistent, with subsidence of 0.5-2 m for each burial event. Very large differences in the amount of subsidence per event, which might indicate localized subsidence contributions, have not been observed, as pointed out by Clague (1997). The large deviation in uplift magnitude recorded across the Patton Bay fault in Alaska, is not observed in subsidence on the Cascadia margin. However, the Patton Bay fault, within the coseismically uplifted zone, is within the accretionary prism and close to the deformation front, and therefore might be expected to experience more pronounced deformation. Measurements of Cascadia subsidence are invariably imprecise because biostratigraphic markers such as diatoms and plant assemblages have large vertical water depth ranges, but fairly large differences in subsidence can be detected (Atwater & Hemphill-Haley 1996). Small variations in measured subsidence through careful lithostratigraphic and biostratigraphic studies, such as those of Long &Shennan (1994), Nelson et al. (1996b), and Shennan et al. (1996), may eventually allow the contribution of local structures to the subsidence record to be determined. The estuarine stratigraphy at many Cascadia sites is strikingly similar to that observed at passive margin sites (Long & Sherman 1994), where a coseismic origin is unlikely. Some Cascadia subsidence events may therefore have non-seismic origins such as natural succession of intertidal environments from local changes in sea level, sedimentation rates, and ocean currents (Long &Shennan 1994; Nelson et al. 1996a, b).
N-S
337
compression
Most active structures on the inner shelf and coast in the Cascadia subduction zone are characterized by roughly E-W trends, in contrast to the predominantly N-S to NW-SE trends which result from plate convergence on the continental slope. We agree with Snavely (1987) and Goldfinger et al. (1992b) that this landward region of the forearc is under N-S compression, which is in agreement with regional N-S compression throughout the continental northwestern USA derived from late Tertiary upper-plate fault orientations, earthquake focal mechanisms, and borehole breakouts (Werner et al. 1990; Zoback & Zoback 1989). The regional N-S compressional stress field extends onto the middle to outer shelf in Washington and much of Oregon. In contrast, the southern Oregon and northern California shelf and coastal region are within the active accretionary prism, and deformation is in response to plate convergence leading to structures with N-S to NW-SE trends. Wang et al. (1995) suggested that the NE-directed strain accumulation caused by plate convergence can be considered a time-dependent local perturbation superimposed on the regional N-S compressive stress field, and thus the regional stress field and cyclic loading may coexist. The transition from regional N-S compression to predominantly plate convergence driven compression represents a significant structural domain boundary. This transition may act as a backstop and may be related to the long-term average position of the downdip end of the seismogenic locked zone. Despite the apparent independence of upper-plate structures from subduction zone deformation, it seems likely that these structures could be triggered by rupture of the subduction zone. Independent fault movement in response to regional compression is also possible and therefore these structures pose independent seismic hazards.
Conclusions Evidence of Quaternary deformation on the Cascadia coast and inner shelf is widespread, with crustal downwarping or fault offset coincident with many coastal lowlands. Rapidly buried marshes at these locations may be due to elastic strain release on the subduction megathrust, downwarping or fault displacement on upper-plate crustal structures, or both. Calculations of Cascadia subduction zone earthquake recurrence intervals, rupture zones, and magnitudes based on correlations of marsh burial
338
L. C. McNEILL E T AL.
events between sites may be complicated by the possibility of localized crustal fault movements and fold growth, in addition to non-seismic origins of the observed stratigraphy. The earthquake record is likely to be more difficult to resolve with the interaction of these multiple factors (Fig. 14). Prehistoric subduction zone earthquakes may have been of lower magnitude than previously estimated. Recurrence intervals for such earthquakes may be overestimated, if some events are not preserved as a result of little permanent deformation, or underestimated, if anomalous local subsidence events are wrongly linked to similar-age regional events along the margin. Higher-resolution records of marsh chronology and estimates of subsidence may eventually lead to the separation of regional and local factors. Meanwhile, the use of these records for modelling of the subduction earthquake cycle and prediction of prehistoric earthquake rupture zones should be undertaken with caution. The inner shelf and coastal structures are consistent with the regional N - S compressional stress field and inconsistent with subduction-driven compression. Despite low seismicity, these crustal faults may be seismic and pose significant shaking and ground deformation hazards to the coastal communities. We acknowledge the Minerals Management Service Pacific OCS Region at Camarillo, California, for supplying data used in this study. We thank the crews of the Jolly Roger and Cavalier, DELTA submersible pilots, Williamson and Associates of Seattle, Washington for sidescan sonar operations, and the scientific crews of the 1992-1995 research cruises. L.C.M. wishes to thank B. Atwater, P. Clark, E. Clifton, H. Kelsey, P. McCrory, A. Nien, S. Obemeier, R. Petier, C. Peterson, G. Priest and I. Shennan for helpful discussions. However, the interpretations and conclusions of the paper are entirely the responsibility of the author. We also thank two anonymous reviewers for their helpful comments and suggestions. This study was supported by NOAA Undersea Research Program at the West Coast National Undersea Research Center, University of Alaska Grants UAF-92-0061 and UAF-93-0035, National Science Foundation Grants OCE-8812731 and OCE-9216880, and US Geological Survey National Earthquake Hazards Reduction Program awards 14-08-0001-G1800, 1434-93-G-2319, 1434-93-G2489, and 1434-95-G-2635.
References ADAMS, J. 1990. Paleoseismicity of the Cascadia subduction zone: evidence from turbidites off the Oregon-Washington margin. Tectonics, 9, 569-584.
ANDO,M. & BALAZS,E. I. 1979. Geodetic evidence for aseismic subduction of the Juan de Fuca plate. Journal of Geophysical Research, 84, 3023-3027. ATWATER, B. F. 1987. Evidence for great Holocene earthquakes along the outer coast of Washington State. Science, 236, 942-944. - - 1 9 9 2 . Geologic evidence for earthquakes during the past 2000 years along the Copalis River, southern coastal Washington. Journal of Geophysical Research, 97, 1901-1919. - - (compiler) 1994. Geology of Liquefaction Features about 300 Years Old along the Lower Columbia River at Marsh, Brush, Price, Hunting, and Wallace Islands, Oregon and Washington. US Geological Survey Open-File Report, 94-209. & HEMPHILL-HALEY, E. 1996. Preliminary Estimates of Recurrence Intervals for Great Earthquakes of the Past 3500 Years at Northeastern Willapa Bay', Washington. US Geological Survey Open-File Report, 96-001. -& YAMAGUCHI,D. K. 1991. Sudden, probably coseismic submergence of Holocene trees and grass in coastal Washington state. Geology, 19, 706-709. , NELSON, A. R., CLAGUE, J. J. et al. 1995. Summary of coastal geologic evidence for past great earthquakes at the Cascadia subduction zone. Earthquake Spectra, 11, 1-18. BARD, E., HAMELIN, B., ARNOLD, M., MOTAGGIONI, L., CABIOCH, G., FAURE, G., & ROUGERIE, F. 1996. Deglacial sea-level record from Tahiti corals and the timing of global meltwater discharge. Nature, 382, 241-244. BARNETT,E. T. 1997. Potentialjor coastalflooding due to coseismic subsidence in the central Cascadia margin. MS thesis, Portland State University, Portland. BERRYMAN,K. R. 1993a. Age, height, and deformation of Holocene marine terraces at Mahia Peninsula, Hikurangi subduction margin, New Zealand. Tectonics, 12, 1347-1364. - - 1 9 9 3 b . Distribution, age, and deformation of late Pleistocene marine terraces at Mahia Peninsula, Hikurangi subduction margin, New Zealand. Tectonics, 12, 1365-1379. , BEU, A. & IRWIN, S. 1997. A high resolution model for the formation of a coseismic Holocene marine terrace sequence. INQUA Commission on Neotectonics, Late Quaternary Coastal Tectonics Abstracts. , OTA, Y., & HULL, A. G. 1989. Holocene paleoseismicity in the fold and thrust belt of the Hikurangi subduction zone, eastern North Island, New Zealand. Tectonophysics, 163, 185-195. BLOOM, A. L. & YONEKURA, N. 1990. Graphic analysis of dislocated shorelines. In: NATIONAL RESEARCH COUNCIL (ed.) Studies in Geophysics." Sea-Level Change. National Academy Press, Washington, DC, 104-115. CASHMAN, S. M., & KELSEY, H. M. 1990. Forearc uplift and extension, southern Hawke's Bay, New Zealand: Mid-Pleistocene to present. Tectonics, 9, 23-44.
UPPER PLATE DEFORMATION CLAGUE, J. J. 1997. Evidence for large earthquakes at the Cascadia subduction zone. Reviews of Geophysics, 35, 439-460. CLARK, J. A., & LINGLE, C. S. 1979. Predicted relative sea level changes (18,000 BP to present) caused by late-glacial retreat of the Antarctic ice sheet. Quaternary Research, 11, 279-298. CLARKE, S. H. JR 1990. Map Showing Geologic
Structures of the Northern CaliJbrnia Continental Margin. Scale 1:250000. US Geological Survey, Map MF-2130. - - 1 9 9 2 . Geology of the Eel River Basin and adjacent region: implications for late Cenozoic tectonics of the southern Cascadia subduction zone and Mendocino triple junction. American Association of Petroleum Geologists Bulletin, 76, 199-224. - & CARVER, G. A. 1992. Late Holocene tectonics and paleoseismicity, southern Cascadia subduction zone. Science, 255, 188-192. , FIELD, M. E. & HIROZAWA, C. A. 1985. Re-
connaissance Geology and Geologic Hazards of the Offshore Coos Bay Basin, Oregon. US Geological Survey Bulletin, 1645. CLIFTON, H. E. 1983. Discrimination between subtidal and intertidal facies in Pleistocene deposits, Willapa Bay, Washington. Journal of Sedimentary Petrology, 53, 353-369. - - 1 9 9 4 . Transgressive and Early Highstand System
Tracts: Pleistocene terrace deposits, Willapa Bay, Washington. Field guide prepared in conjunction with a Society of Economic and Petroleum Geologists Research Conference, Clastic Deposits of the Transgressive Systems Tract: Facies, Stratigraphy and Reservoir Character. COOPER, D. M. 1981. Sedimentation, stratigraphy and
facies variations of the lower to middle Miocene Astoria Formation in Oregon. PhD thesis, Oregon State University, Corvallis. CROSSON, R. S. & OWENS, T. J. 1987. Slab geometry of the Cascadia subduction zone beneath Washington from earthquake hypocenters and teleseismic converted waves. Geophysical Research Letters, 14, 824-827. DARBY, D. & BEANLAND, S. 1992. Possible source models for the 1855 Wairarapa earthquake, New Zealand. Journal of Geophysical Research, 97, 12375-12389. DARIENZO, M. E. & PETERSON, C. D. 1990. Episodic tectonic subsidence of late Holocene salt marshes, northern Oregon central Cascadia margin. Tectonics, 9, 1-22. & CLOUGH, C. 1994. Stratigraphic evidence for great subduction-zone earthquakes at four estuaries in northern Oregon, USA. Journal of Coastal Research, 10, 850-876. DEMETS, C., GORDON, R. G., ARGUS, D. F. & STEIN, S. 1990. Current plate motions. Geophysical Journal International, 101, 425-478. GOLDFINGER, C. 1994. Active deformation of the
Cascadia forearc: Implications for great earthquake potential in Oregon and Washington. PhD thesis, Oregon State University, Corvallis. , KULM, L. D. & YEATS, R. S. 1992a. Neotectonic
Map of the Oregon Continental Margin and
AND CASCADIA EARTHQUAKES
339
Adjacent Abyssal Plain. Scale 1:500 000. Oregon Department of Geology and Mineral Industries, Open-File Report, 0-92-4. , APPELGATE, B., MACKAY, M. & COCHRANE, G. R. 1996. Active strike-slip faulting and folding of the Cascadia plate boundary and forearc in central and northern Oregon. In: ROGERS, A. M., WALSH, T. J., KOCKELMAN, W. J. & PRIEST, G. (eds) Assessing and Reducing
Earthquake Hazards in the Pacific Northwest. US Geological Survey Professional Paper, 1560, 223-256. & MOORE, G. F. 1992b. Transverse structural trends along the Oregon convergent margin: implications for Cascadia earthquake potential. Geology, 20, 141-144. , MCNEILL, L. C., & HUMMON, C. ]997a.' Oblique strike-slip faulting of the central Cascadia submarine forearc. Journal of Geophysical Research, 102, 8217-8243. --, MCNEILL, L. C. & HUMMON, C. 1997b. Case study of GIS data integration and visualization in submarine tectonic investigations: Cascadia subduction zone. Marine Geodesy, 20, 267-289. GOWER, H. D. 1960. Geologic Map of the Pysht Quadrangle, Washington, scale 1:62 500. US Geological Survey Geologic Quadrangle Map, GQ-129.
GRIM, M. S. & BENNETT, L. R. JR 1969. Shallow seismic profiling of the continental shelf off Grays Harbor, Washington. University of Washington Department of Oceanography Special Reports, 41, 72-92. HEATON, T. H. & KANAMORI, H. 1984. Seismic potential associated with subduction in the northwestern United states. Bulletin of the Seismological Society of America, 74, 993-941. HILL, D. P. et al. 1993. Seismicity remotely triggered by the magnitude 7.3 Landers, California earthquake. Science, 260, 1617-1623. HULL, A. G. 1990. Tectonics of the 1931 Hawke's Bay earthquake. New Zealand Journal of Geology and Geophysics, 33, 309-330. JACOBY, G. C., BUNKER, D. E. & BENSON, B. E. 1997. Tree-ring evidence for an AD 1700 Cascadia earthquake in Washington and northern Oregon. Nature, 25, 999-1002. KELSEY, H. M. 1990. Late Quaternary deformation of marine terraces on the Cascadia subduction zone near Cape Blanco, Oregon. Tectonics, 9, 983-1014. , TICKNOR, R. L., BOCKHEIM, J. G. & MITCHELL, C. E. 1996. Quaternary upper plate deformation in coastal Oregon. Geological Society of America Bulletin, 108, 843-860. , WITTER, R. C. & POLENZ,*M. 1993. Cascadia paleoseismic record derived from late Holocene fluvial and lake sediments, Sixes River Valley, Cape Blanco, south coastal Oregon (abstract).
Eos Transactions, American Geophysical Union, 74, 199. , & HEMPHILL-HALEY, E. 1998. Response of a small Oregon estuary to coseismic subsidence and postseismic uplift in the past 300 years. Geology, 26, 231-234.
340
L.C.
M c N E I L L E T AL.
KENNEDY, G. L. 1978. Pleistocene paleoecology, zoogeography and geochronology of marine invertebrate faunas of the Pacific Northwest coast (San Francisco Bay to Puget Sound). PhD thesis, University of California, Davis. , LAJOIE, K. R. & WEHMILLER, J. F. 1982. Aminostratigraphy and faunal correlations of late Quaternary marine terraces, pacific Coast, USA. Nature, 299, 545-547. KULM, L. D. & SUESS, E. 1990. Relation of carbonate deposits and fluid venting: Oregon accretionary prism. Journal of Geophysical Research, 95, 8899-8915. KVENVOLDEN, K. A., BLUNT, D. J. & CLIFTON, H. E. 1979. Amino-acid racemization in Quaternary shell deposits at Willapa Bay, Washington. Geochimica et Cosmochimica Acta, 43, 1505-1520. LONG, A. J. & SHENNAN, I. 1994. Sea-level changes in Washington and Oregon and the 'earthquake deformation cycle'. Journal of Coastal Research, 10, 825-838. LUDWIN, R. S., WEAVER, C. S. & CROSSON, R. S. 1991. Seismicity of Washington and Oregon. In: SLEMMONS, D. B., ENGDAHL, E. R., BLACKWELL,D. & SCHWARTZ, D. (eds) Neotectonics of North America. Geological Society of America, DNAG CSMV-1, Boulder, CO, 77-98. MAEMOKU, H. 1988a. Holocene crustal movement in Muroto Peninsula, southwest Japan. Geographical Review of Japan (Series A), 61(10), 747-769. - - 1 9 8 8 b . Holocene crustal movement around Cape Ashizuri, southwest Japan. Geographical Sciences, 43, 231 240. - & TSUBONO, K. 1990. Holocene crustal movements in the southern part of Kii Peninsula, outer zone of southwest Japan. Journal of Geography, 99, 349 369. MATHEWES, R. W. & CLAGUE, J. J. 1994. Detection of large prehistoric earthquakes in the Pacific Northwest by microfossil analysis. Science, 264, 688-691. MCCAFFREY, R. & GOLDFINGER, C. 1995. Forearc deformation and great subduction earthquakes: Implications for Cascadia earthquake potential. Science, 267, 856-859. MCCRORY, P. A. 1996. Tectonic model explaining divergent contraction directions along the Cascadia subduction margin, Washington. Geology, 24, 929 932. MCINELLY, G. W. & KELSEY, H. M. 1990. Late Quaternary tectonic deformation in the Cape Arago-Bandon region of coastal Oregon as deduced from wave cut platforms. Journal of Geophysical Research, 95, 6699-6713. MCNEILL, L. C., PIPER, K. A., GOLDFINGER, C., KULM, L. D. & YEATS, R. S. 1997. Listric normal faulting on the Cascadia continental shelf. Journal of Geophysical Research, 102, 12 123-12 138. NELSON, A. R. 1987. Apparent gradual rise in relative sea level on the south-central Oregon coast during the late H o l o c e n e - implications for the great Cascadia earthquake hypothesis (abstract). Los Transactions, American Geophysical Union, 68, 1240.
-1992. Holocene tidal-marsh stratigraphy in SouthCentral O r e g o n - evidence for localized sudden submergence in the Cascadia subduction zone. In: FLETCHER, C. H. & WEHMILLER, J. F. (eds) Quaternary Coasts of the United States." Marine and Lacustrine Systems. Society for Sedimentary Geology Special Publication, 48, 287-301. --, & PERSONIUS, S. F. 1996. Great-earthquake potential in Oregon and Washington an overview of recent coastal geologic studies and their bearing on segmentation of Holocene ruptures, central Cascadia subduction zone. In: ROGERS, A. M., WALSH,T. J., KOCKELMAN,W. J. & PRIEST, G. (eds) Assessing and Reducing Earthquake hazards' in the Pacific Northwest. US Geological Survey Professional Paper, 1560, 91-114. --, ATWATER, B. F., BROBOWSKI, P. T. et al. 1995. Radiocarbon evidence for extensive plate-boundary rupture about 300 years ago at the Cascadia subduction zone. Nature, 378, 371-374. --, JENNINGS, A. E. & KASHIMA, K. 1996a. An earthquake history derived from stratigraphic and microfossil evidence of relative sea-level change at Coos Bay, southern coastal Oregon. Geological Society of America Bulletin, 108, 141-154. --, SHENNAN, I. • LONG, A. J. 1996b. Identifying coseismic subsidence in tidal-wetland stratigraphic sequences at the Cascadia subduction zone of western North America. Journal of Geophysical Research, 101, 6115-6135. NIEM, A. R. & NIEM, W. A. 1985. Oil and Gas Investigations of the Astoria Basin, Clatsop and Northernmost Tillamook Counties, Northwest Oregon. Scale 1:250000 Oregon Department of Geology and Mineral Industries Oil and Gas Investigation, 14. , SNAVELY, P. D., JR & NIEM, W. A. 1990. Onshore-Offshore Geologic Cross Section from the Mist Gas Field, Northern Oregon Coast Range, to the Northwest Oregon Continental Shelf Oregon Department of Geology and Mineral Industries Oil and Gas Investigation, 17. OBERMEIER, S. F. 1995. Preliminary Estimates of the Strength of Prehistoric Shaking in the Columbia River Valley and the Southern Hall" of Coastal Washington, with Emphasis for a Cascadia Subduction Zone Earthquake about 300 Years ago. US Geological Survey Open-File Report, 94-589. OPPENHEIMER, D., BEROZA, G., CARVER, G. et al. 1993. The Cape Mendocino, California, earthquakes of April 1992: subduction at the triple junction. Science, 261, 433-438. ORANGE, D. L., GEDDES, D. S., & MOORE, J. C. 1993. Structural and fluid evolution of a young accretionary complex: the Hoh rock assemblage of the western Olympic Peninsula, Washington. Geological Society' of America Bulletin, 105, 1053-1075. PALMER, S. P. & LINGLEY, W. S. 1989. An Assessment o/ the Oil and Gas Potential of the Washington Outer Continental Shelf. Washington State Division of Geology and Earth Resources Report, WSG 8 9 - 2 .
UPPER PLATE DEFORMATION AND CASCADIA EARTHQUAKES PANDOLFI, J. M., BEST, M. M. R. & MURRAY, S. P. 1994. Coseismic event of May 15, 1992, Huon Peninsula, Papua New Guinea: comparison with Quaternary tectonic history. Geology, 22, 239-242. PARKER, M. J. 1990. The Oligocene and Miocene geology of the Tillamook embayment, Tillamook County, northwest Oregon. MS thesis, Oregon State University, Corvallis. PELTIER, W. R. 1996. Mantle viscosity and ice-age topography. Science, 273, 1359-1364. PETERSON, C. D. & DARIENZO, M. E. 1989. Episodic, abrupt tectonic subsidence recorded in late Holocene deposits of the South Slough syncline: an onland expression of shelf fold belt deformation from the southern Cascadia margin. Geological Society of America Abstracts with Programs, 21, 129. & 1996. Discrimination of climatic, oceanic, and tectonic mechanisms of cyclic marsh burial, Alsea Bay, Oregon. In: ROGERS, A. M., WALSH, T. J., KOCKELMAN, W. J. & PRIEST, G. (eds) Assessing and Reducing Earthquake Hazards in the Pacific Northwest. US Geological Survey Professional Paper, 1560, 115-146. & PRIEST, G. R. 1995. Preliminary reconnaissance survey of Cascadia paleotsunami deposits in Yaquina Bay, Oregon. Oregon Geology, 57, 33 40. --, BARNETT, E., BRIGGS, G. G., CARVER, G. A., CLAGUE, J. J. & DARIENZO, M. E. 1997. Subsidence from Great Earthquakes in the Cascadia Subduction Zone, Vancouver Island, B.C., Washington, Oregon, and Northernmost California. Department of Oregon Geology and Mineral Industries, Open-File Report, 0-97-5. PLAFKER, G. 1969. Tectonics of the March 27 Alaskan earthquake. US Geological Survey Professional Paper, 543-I. - - 1 9 7 2 . Alaskan earthquake of 1964 and Chilean earthquake of 1960: implications for arc tectonics. Journal of Geophysical Research, 77, 901-925. PR1EST, G. R., SAUL, I. & DIEBENOW,J. 1994. Chronic Geologic Hazard Maps and Erosion Rate Database, Coastal Lincoln Count)', Oregon: Salmon River to Seal Rocks, 1:4800 scale maps. Oregon Department of Geology and Mineral Industries Open-File Report, O-94-11. RAU, W. W. 1973. Geology of the Washington Coast between Point Grenville and the Hoh River. Washington Division of Geology and Earth Resources Bulletin, 66. RYAN, H. F. & STEVENSON,A. J. 1995. Cruise Reportfor C1-94-0W: Reconnaissance High Resolution Geopulse Data Acquired for Seismic Hazard Studies along the Columbia River from July 18 22, 1994. US Geological Survey Open-File Report, 95-668. SATAKE, K., SHEMAZAKI,K., YOSHINOBU,T. & UEDA, K. 1996. Time and size of a giant earthquake in Cascadia inferred from Japanese tsunami records of January 1700. Nature, 379, 246-249. SHENNAN, I., LONG, A. J., RUTHERFORD, M. M e t al. 1996. Tidal marsh stratigraphy, sea-level change and large earthquakes, I: a 5000 year record in Washington, USA. Quaternary Science Reviews, 15, 1023-1059. -
-
-
-
341
SIEH, K., JONES, L., HAUKSSON, E. et al. 1993. Nearfield investigation of the Landers earthquake sequence, April to July 1992. Science, 260, 171-176. SNAVELY, P. D., JR 1976. Geologic map of the Yaquina and Toledo Quadrangles, Lincoln County, Oregon, scale 1:62 500. US Geological Survey Miscellaneous Investigations Series Map, 1-867. 1987. Tertiary geologic framework, neotectonics, and petroleum potential of the Oregon-Washington continental margin. In. SCHOLL, D. W., GRANTZ, A. & VEDDER, J. G. (eds) Geology and Resource Potential of the Continental Margin oJ Western North America and Adjacent Ocean Basins - Beau[ort Sea to Baja California. CircumPacific Council for Energy and Mineral Resources, Houston, TX, 305-335. -& WELLS, R. E. 1996. Cenozoic evolution of the continental margin of Oregon and Washington. In." ROGERS, A. M., WALSH, T. J.; KOCKELMAN, W. J., & PR1EST, G. (eds) Assessing and Reducing Earthquake Hazards in the Pacific Northwest., US Geological Survey Professional Paper, 1560, 161 182. SPOTILA, J. c~zSIEH, K. 1995. Geologic investigations of the 'slip gap' in the surficial ruptures of the 1992 Landers earthquake, southern California. Journal of Geophysical Research, 100, 543-559. SUGIYAMA, Y. 1994. Neotectonics of Southwest Japan due to the right-oblique subduction of the Philippine Sea plate. Geofisica lnternacional, 33, 53 76. TABOR, R. W. & CADY, W. M. 1978. The Structure of the Olympic Mountains, Washington-Analysis of a Subduction Zone. US Geological Survey Professional Paper, 1033. TICKNOR, R. L. 1993. Late Quaternary crustal deJormation on the central Oregon coast as deduced from uplifted wave-cut platforms. MS thesis, Western Washington University, Bellingham. --, KELSEV, H. M. & BOCKHEIM, J. G. 1992. Late Quaternary tectonic deformation along the central Oregon portion of the Cascadia margin as deduced from deformation of wave-cut platforms. Geological Society of America Abstracts with Programs, 24, 86. WAGNER, H. C. & TOMSON, J. H. 1987. Offshore Geolog)' of the Strait of Juan de Fuca, State of Washington and British Columbia, Canada. Washington Division of Geology and Earth Resources Open File Report, 87-1. WALD, D. & HEATON, T. 1994. Spatial and temporal distribution of slip for the 1992 Landers, California, earthquake. Bulletin of the Seismological Society of America, 84, 668-691. WALSH, T. J., KOROSEC, M. A., PHILLIPS, W. M., LOGAN, R. L. & SCHASSE, H. W. 1987. Geologic Map of Washington- Southwest Quadrant. Washington Division of Geology and Earth Resources, Geologic Map, GM-34. WANG, K., MULDER, T., ROGERS, G. C. & HYNDMAN, R. D. 1995. Case for very low coupling stress on the Cascadia subduction fault. Journal of Geophysical Research, 100, 12 907-12 918.
342
L.C.
M c N E I L L E T AL.
WEAVER, C. S. & BAKER, G. E. 1988. Geometry of the Juan de Fuca plate beneath Washington and northern Oregon from seismicity. Bulletin of the Seismological Society of America, 78, 264-275. WELLS, R. E., N~M, A. R. & SNAVELu P. D., JR 1994. Young deformation at Netarts Bay and elsewhere, northern Oregon coast (abstract). Proceedings of the Oregon Academy of Sciences, XXX, 39. , SNAVELY, P. D., JR & NIEM, A. R. 1992. Quaternary thrust faulting at Netarts Bay, northern Oregon coast. Geological Society of America Abstracts with Programs, 24, 89. WERNER, K. S., GRAVEN, E. P., BERKMAN, T. A. & PARKER, M. J. 1990. Direction of maximum horizontal compression in northwestern Oregon determined by borehole breakouts. Tectonics, 10, 948-958.
WEST, D. O. 1986. WNP-3 Geologic Support Services Coastal Terrace Study. Washington Public Power Supply System, Richland, 853-1010. -& MCCRUMB, D. R. 1988. Coastline uplift in Oregon and Washington and the nature of Cascadia subduction zone tectonics. Geology, 16, 169-172.
YAMAGUCHI, D. K., ATWATER, B. F., BUNKER, D. E., BENSON, B. E. & REID, M. S. 1997. Tree-ring dating the 1700 Cascadia earthquake. Nature, 389, 922-923. ZOBACK, M. L. & ZOBACK, M. D. 1989. Tectonic stress field of the continental United States. In. PAKISER, L. & MOONEY, W. (eds) Geophysical Framework of the Continental United States. Geological Society of America, Memoir, 172, 523-540
Late Holocene coastal tectonics at Falasarna, western Crete: a sedimentary study DALE
DOMINEY-HOWES
1"2, A L A S T A I R
DAWSON 2 & DAVID
SMITH 2
1 Coventry Centre Jor Disaster Management, School of the Built Environment, Coventry University, Coventry CV1 5FB, UK (e-mail:
[email protected]) 2 Centre for Quaternary Science, Division of Geography, Coventry University, Coventry CV1 5FB, UK Abstract: The late Holocene sedimentary record of Falasarna Harbour, western Crete, includes detailed evidence of tsunamis and serves as an independent dataset to evaluate the magnitude and timing of coastal tectonic movements in an area affected by contrasting tectonic regimes. Analysis of a foraminiferal assemblage makes it possible to identify suites of tsunami-deposited sediments within normal sedimentary sequences. The palaeo-environmental record is then complemented with a sequence of raised fossil marine notches. The transitional boundary between marine and terrestrial sedimentation indicates tectonic uplift at C.AD63-75+90 radiocarbon years BP, which is in conflict with previously published interpretations. No sedimentary evidence can be found for a tsunami believed to be associated with a large uplift event during AD365.
This paper presents the results of an investigation at Falasarna Harbour, western Crete, of sediments deposited by tsunamis reported to have occurred in the Aegean Sea region of Greece. Falasarna is located within an active extensional domain inboard of the compressional front associated with subduction of the Mediterranean plate, and, is therefore affected by different tectonic regimes. Any sedimentary evidence for tsunamis may help to shed light on the nature of coastal tectonic activity in this area. Biostratigraphic (Foraminifera) and lithostratigraphic evidence was used to determine the palaeoenvironmental history of Falasarna. We show that, following the construction of the harbout, sedimentation progressed until the site was affected by tectonic movement. The sedimentological data indicate that the harbour of Falasarna was raised above sea level, as reflected by a change from marine to terrestrial conditions. The data do not preserve evidence for a large vertical coseismic displacement reported to have occurred in AD 365 (Pirazzoli 1986).
Tectonic setting On a regional scale, the present form of the Aegean is the result of a set of complex interactions between phases of compressional and extensional tectonic normal faulting which result from the southward stretching and subsidence of the Aegean plate (Le Pichon & Angelier 1981; Mercier 1981). The Aegean region is composed of the Inner Hellenic Volcanic Arc and the Outer
Hellenic Arc (with subducting trench system) (Fig. 1). The Outer Hellenic Arc forms part of a rigid body thrusting over the Mediterranean basins and is associated with the development of the accretionary complex and a compressional front located south of Crete at the junction between the African and European plates (Lallemant et al. 1994). The Falasarna study site lies within an active extensional domain inboard of the compressional front and is found at the western end of the island of Crete, the most prominent feature of the Outer Hellenic Arc and which has been affected by uplift during the Holocene. The uplift is related to the development of the compressional front and the accretionary wedge by underplating of sediments from the downgoing plate beneath the continental upper plate (Le Pichon & Angelier 1981; Lallemant et al. 1994). Figure 1 provides a schematic representation of the main structural and tectonic components for the region in which Falasarna is located. It should be noted that the area to the south of western Crete is characterized by a series of E - W trending faults, whereas to the west of Crete, the main offshore faults trend N-S. To the NW of Crete, the orientation of submarine faults changes to NW-SE. Such radically different tectonic regimes in closely adjoining areas have prevailed since early Quaternary time (Angelier 1978); Jackson (1994) noted that the faults associated with the Outer Hellenic Arc are affected by normal, reverse and thrust movement. The area considered in this study is affected by both relatively shallow earthquakes associated with extensional faulting of the overriding
DOMINEY-HOWES, D., DAWSON, A. & SMITH, D. 1998. Late Holocene coastal tectonics at Falasarna, western Crete: a sedimentary study. In: STEWART, I. S. & VrrA-FINZI, C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 146, 343-352.
344
D. DOMINEY-HOWES E T AL.
SEA OF CRETE
%
I
km
0
t
1 O0
Fig. 1. Schematic representation of the main structural components of the Outer Hellenic Arc, subduction system and compressional front (Mediterranean Ridge and deformation front). Location of the study site is shown by the square.
upper-crustal plate and by deeper earthquakes associated with thrust movement of the subducting plate (Taymaz et al. 1990; Lallemant et al. 1994; Stewart, pers. comm. 1997).
Conceptual framework and methods It has been reported that the harbour of Falasarna contains a record of late Holocene sedimentation and that the harbour sediments include
deposits laid down by tsunamis which flooded the area in AD 66 and AD 365 (Pirazzoli et al. 1992). To elucidate the palaeoenvironmental history of this site, the lithostratigraphy of sedimentary sequences was investigated in five trenches excavated by the Greek Regional Archaeological Service at Falasarna. The elevation of all units was determined by instrumental levelling using a Zeiss Autoset Level. All levelling traverses were closed with no error greater than +0.01 m. In the
COASTAL TECTONICS IN CRETE absence of a Greek Datum, present sea level was assumed to be true mean sea level, as tidal variations rarely exceed 10 cm (P. A. Pirazzoli, pers. comm., 1994). It is recognized, however, that sea-level variations as a result of tidal cycles, atmospheric conditions, the nature of the geoid, storminess and seasonality may result in variations of mean sea level approaching 50cm (IAPSO 1985; Emery et al. 1988; Flemming & Woodworth 1988; Flemming 1992). Variations in the stratigraphy based upon changes in clast content, colour, lithology, matrix, shell content, stone content, clast roundness, structure and texture were recorded. Contiguous 0.05 m bulk samples (125 cm 3) of sediments were collected successively through the sequence from the base to the top of each trench for laboratory investigation including detailed biostratigraphic analysis based on Foraminifera. Foraminifera counting was carried out using a VMT 12 microscope at x 1 and x4 magnification. Reference was made to type collections at the Department of Micropalaeontology at the Natural History Museum, London, and the accounts of Sidebottom (1904-1909) and Cimerman & Langer (1991). Wherever possible, 300 individuals were counted in each sample to ensure statistical confidence. Particle size analysis was unsuccessful because detailed variations were obscured by the sample size adopted and since over 80% of the sediment is composed of marine biogenic CaCO3 (Pirazzoli et al. 1992).
Falasarna Harbour, site description Falasarna was a pirate port which operated between approximately the middle of the fourth century BC and the late first century BC (Hadjidaki 1988). The date at which the harbour became functional and the date at which it was finally abandoned are not known exactly (Hadjidaki 1988; Frost 1997), although the harbour was in existence by the time of Scylax of Caryanda in the middle of the fourth century BC (Hadjidaki 1988). The harbour is situated next to the Bay of Livadi in an enclosed position behind the eastern side of Cape Kutri at the southern end of the Grammvousa peninsula in western Crete. It is an quadrangular artificial harbour cut into the surrounding Mesozoic limestone and Scylax referred to Falasarna's status as a closed harbour (Pirazzoli et al. 1992, p. 375). The main harbour is 1 0 0 m • in size and is divided from a secondary basin 50m x 35m lying immediately to the east by a complex of walls and buildings. The main harbour is connected to the sea at its western side by a
345
channel which stretches 100m to the present shoreline (Fig. 2). The ground surface within the main harbour is at an elevation of 6.6 m above sea level (m a.s.1.). Hadjidaki (1988) suggested that Falasarna was probably destroyed by the Romans in 67 BC. She stated that the Romans sent Caecillius Metellus as a praetor to Crete to destroy a number of pirate strongholds. According to Pirazzoli et al. (1992), after its destruction the harbour rapidly filled with marine and terrestrial sediments and was inundated by tsunamis in AD66 and in AD 365, when it was uplifted to 6.6 m a.s.1, by an earthquake on 21 July. There is no evidence at Falasarna of continued occupation after the destruction of 67 BC.
Sedimentology In this paper only the results of the sedimentological analysis of Trench A are presented because this trench displays the clearest sedimentary record (Dominey-Howes 1996). Trench A is located within the main harbour basin (Fig. 2). The surface is at an elevation of 6.2m a.s.1, and the base is at 4.6 m a.s.1. (Fig. 3). Five lithostratigraphic units occur within the sedimentary sequence. The basal unit is a Foraminifera- and molluscrich, well-sorted fine to medium sand. The unit contains coarse rounded to angular grit. There are many small to medium-sized well-rounded to subangular limestone, sandstone and quartz clasts. The matrix contains many whole and comminuted marine molluscs. Many small subrounded pottery sherds are visible. Some crude horizontal bedding is apparent. The upper boundary of this unit is defined by an undulating unconformity. The basal unit is overlain by a mollusc- and Foraminifera-rich marine sand which contains rounded fine grit. The unit is further characterized by large numbers of whole and broken marine molluscs with no obvious orientation. Medium to large subrounded to subangular limestone and sandstone clasts and blocks (a-axis up to 10 cm) are present. There is no obvious bedding or structure, although clast a-axis tends towards the horizontal. The upper boundary of this unit is defined by an undulating unconformity. The middle unit is composed of a fine to medium-sized sand and grit and is further characterized by abundant well-rounded to subangular limestone and sandstone clasts. The smaller clasts are matrix supported whereas the larger blocks (a-axis from 20cm) are clast supported. The a-axis of the clasts tends towards the horizontal although there is no other
346
D. DOMINEY-HOWES E T A L .
Fig. 2. Detail of Falasarna Harbour and the position of the trenches. This paper presents the results of only Trench A. Cape Kutri is located immediately to the north of the harbour. The main channel connects with the Bay of Livadi.
obvious bedding or structures. Both whole and comminuted molluscs are present, as are small rounded pottery sherds. This unit is conformably overlain by a red silty clay sand which extends to the surface of the trench. There are many small to large subrounded to very angular limestone, sandstone and quartz clasts which are matrix supported. A few broken marine molluscs are found towards the base of this unit and there is no apparent bedding or structure. This unit is interrupted by numerous subrounded limestone blocks at 5.70-5.90ma.s.1. (Fig. 3). This unit represents the sediments reported to
have been deposited by a tsunami in AD365 (Pirazzoli et al. 1992). The a-axis of these blocks is up to 18 cm in length, and is approximately horizontal, and the blocks are matrix supported. More than 5200 Foraminifera were extracted and identified from the Trench A sediments and 28 species were identified (Table 1). The number of individuals (300+) per sample is high from the bottom of the trench upwards as far as sample 14 (Table 1); between samples 15 and 20 it declines rapidly and remains low (nine individuals in sample 21 to four in sample 23). None is recorded from sample 24 upwards to the surface
COASTAL TECTONICS IN CRETE
347
Fig. 3. Stratigraphy of Trench A according to the present authors and Pirazzoli et al. (1992). It can be clearly seen that the elevations of the base and top of Trench A vary between the two studies. (For a description of the lithostratigraphic units, refer to the text.) of the trench. The assemblage of Foraminifera present in samples 1-17 is dominated by Amm-
onia ber Ammonia parkinsoniana, Ammonia tepida, Elphidium advenum, Elphidium crispum, Globigerina ruber, Quinqueloculina aspera, Quinqueloculina bicornis and Quinqueloculina vulgaris (Table 1). These species make up 64% of the Foraminifera present within the Trench A samples. The rarest species of Foraminifera preserved are Quinqueloculina jugosa, Lachlanella variolata and Triloculina tricarinata, which constitute 0.4%, 1.6% and 1.8% of the total count, respectively. The number of individuals per sample of Cibicides advenum increases from an average of six specimens per sample in samples 1-14 to an average of 11 specimens per sample in samples 15-20. Similarly, Eponides repanda increases from an average of 2.2 specimens per sample in samples 1-14 to seven specimens per sample in samples 15-20. It is also worth noting that there is an increase in the number of broken tests for all species from an average of 21% of tests in samples 1-14 and 21-23 to an average of 55% of tests in samples 15-20. Pennate forms of Foraminifera total 555 specimens and the average percentage of broken tests is 26% in samples 1-14 and 21-23, compared with 83% in samples 15-20. The average percentage of broken centric forms in samples 1-14 and 21-
23 is 21%, and in samples 15-20 is 18%. Therefore, a higher percentage of pennate to centric Foraminifera are broken in samples 15-20. Table 2 gives the results of radiocarbon dating. The AMS (accelerator mass spectrometry) technique was used as no sample weighed more than 2.0 g. For the purpose of consistency all samples submitted for dating comprised specimens of the marine mollusc Hydrobia acuta. Calibration to calendar years was made by reference to the data of Stuiver & Braziunas (1993). The 13C/12C ratios are compared with those given as - 5 + 40%0 for the eastern Mediterranean by Stuiver & Braziunas (1993). As they lie well within the - 4 5 to +35%o range (Table 2) the shell ages reported here are taken to be reliable.
Interpretation of palaeoenvironmental history The most striking aspect of the lithostratigraphy of Trench A is that there is a clear change in the pattern of the sedimentation at c. 5.70 m a.s.l. and this change dates from AD 63 to AD 75 + 90 radiocarbon years BP. The abrupt change of sedimentation is from marine conditions characterized by the deposition of Foraminiferamollusc-rich sands to terrestrial silts and clay
348
D. D O M I N E Y - H O W E S
/-.,
'~
~
E T AL.
~
~
rT1
,.--..
~6
b
,~.~
~
~.'t.
:~ ~
~
~
~
~....~
.~
,..,,-.-.~
~
~
~
~
.-,.
~
,~
COASTAL TECTONICS IN CRETE
349
Table 2. 14C AMS dates of samplesfrom Trench A Sample
Measured
Conventional
13C/nC(%0)
Calibrated
Beat-81412 Beta-81413 Beta-81414 Beta-81415
2460 + 70 1900 + 60 1890 + 90 21104-60
2870 + 70 2270 + 60 2290 + 90 2530+60
0.0 -2.4 -0.8 +0.7
770-555 Bc AD25-I 50 35 B~AD160 335-155BC
The dates are plotted on Fig. 3.
(with some sand). This change is sudden, as indicated by the boundary unconformity, and is thought to represent relative sea- and land-level changes associated with vertical coseismic deformation. The conditions in which the deposition of large, well-rounded to subangular limestone and sandstone blocks between 5.20 and 5.60m a.s.1. were deposited are thought to be different from those that had operated during marine sedimentation. It is believed by the present authors that deposition of the larger blocks at this level probably relates to a high-energy low-frequency event such as a tsunami. The evidence to support this interpretation is described below. Samples 15-20 in Table 1 correspond lithostratigraphically to the sediments ascribed by Pirazzoli et al. (1992) to a tsunami in AD66 (Fig. 3). Sample 15 lies unconformably on top of the underlying sediments. The evidence provided by the Foraminifera shows clearly that the last episode of marine sedimentation within the harbour was probably during or just after the AD 66 tsunami event reported by Pirazzoli et al. (1992). There is no evidence for marine sedimentation taking place anywhere within the harbour after this event. Of the Foraminifera recovered from Trench A, 64% belong to the Ammonia, Elphidium and Quinqueloculina genera (Table 1). According to Murray (1991), these three genera are all found together in the shallow inner-shelf region. The Ammonia and Elphidium genera are also representative of brackish conditions, but the presence of Quinqueloculina is representative of open marine-lagoonal conditions. Parker (1958) identified a 'typical' shallow-depth assemblage between 0 and 25 m, which he referred to as the 'bay-open marine' assemblage. The presence of the high numbers of Ammonia, Elphidium, Miliolidae and Peneroplidae indicate a clear shallow, fully marine (lagoonal) assemblage at Falasarna similar to that identified by Parker (1958). Murray (1991) noted that Cibicides species range from the inner shelf (0-100 m) through the outer shelf (100-200 m) and in to the upper slope (200-2000m or upper bathyal). E. repanda,
however, was noted as characteristic of depth ranges from the outer shelf (100-200m) to the abyssal plain (4000 m). Its presence may be the result of post-mortem transport processes operating from the outer shelf-bathyal-abyssal depths into the shallow marine-lagoonal sediments of Falasarna. However, the processes that resulted in the deposition of the higher numbers of E. repanda and C. advenum in samples 15-20, were markedly different from those normally operating and are equated with a tsunami because tsunamis occur less frequently than other high-energy phenomenon such as storm surges. A tsunami may also explain the rise in the percentage of broken pennate Foraminiferal tests from 21% to 83% in samples 15-20. In laboratory experiments on the relative resistance of Foraminiferal tests to crushing, globular forms were more resistant than pennate forms (Wetmore 1987). Wetmore recognized that test strength is likely to be related to a complex set of factors such as general shape, character of partitions between chambers, the arrangement of chambers and test wall thickness. However, the strongest foraminiferal tests are those which belong to species that have biconvex to globular shapes and sutures which are only slightly depressed. Such a morphology, Wetmore believes, would allow a more uniform dispersion of compressive stresses associated with sediment impact in the marine environment. If pennate Foraminiferal forms are more susceptible to crushing (and breakage) associated with impact stresses, higher percentages of broken pennate forms in sediments thought to have been deposited during highenergy tsunami inundation would be expected. The data presented here support this hypothesis. Furthermore, such dramatic increases in the percentage of broken pennate forms of diatoms have been reported for tsunami-deposited sediments associated with the Storegga tsunami in Scotland (S. Dawson, pers. comm. 1995). Radiometric dating of marine shells of the species H. acuta from the base of the Trench A (Fig. 3) gave a conventional radiocarbon date of 2870+70 radiocarbon a BP (calibrated age 662 BC 4-70). This dates the onset of
350
D. DOMINEY-HOWES E T A L .
sedimentation within the harbour. A conventional radiocarbon date of 2270 + 60 radiocarbon abe (calibrated age AD75 4-60) has been obtained for sample 11 which comes from below the inferred AO 66 tsunami layer, and a conventional radiocarbon age of 2290 4- 90 radiocarbon a aP (calibrated age AD634-90) has been obtained for sample unit 16, which comes from within the proposed tsunami unit. These dates imply that the high-energy event which led to the deposition of the high percentages of broken (pennate) Foraminifera in samples 15-20 occurred between 35 Bc and AD 160 (but probably between AD 63 and 75 4- 90). The preceding interpretation strongly suggests that deposition of sediments associated with a tsunami of c. AD66 appears to be preserved within the Trench A stratigraphy and it is also noted that no Foraminifera are recorded above sample 23. Most significantly, no Foraminifera are recorded in samples 24-28 which correlate with the sedimentary unit that according to Pirazzoli et al. (1992), had been deposited by a tsunami during AD 365. The results of the present investigation broadly support the findings of a previous study by Pirazzoli et al. (1992) which sought to understand the palaeoenvironmental history of Falasarna. However, there are some discrepancies between the findings of the two investigations. Pirazzoli et al. (1992) reported that the base of Trench A is at an elevation of 5.0 m a.s.1, and the surface is at 7.0 m a.s.l., whereas the present study gave elevations of 4.6 m a.s.1, and 6.2 m a.s.l., respectively (Fig. 3). There are four possible explanations for these variations. First, tectonic activity resulting in relative sea- and land-level changes could have occurred between the two successive phases of investigations, but such tectonic activity is not known to the present authors. Second, the elevations of the present authors could be erroneous. However, the closing error for the Trench A traverse was 0.01 m. Third, the elevations of Pirazzoli et al. (1992) could be incorrect. P. A. Pirazzoli (pets. comm., 1997) stated that as the trench elevations reported by him and his coworkers were calculated by the Regional Archaeological Service, significant error may have been introduced. Fourth, errors associated with the assumption that present sea level is true mean sea level may result in variations between successive phases of investigation of up to 50 cm. The lithostratigraphy described in this paper is similar to that previously published, and the foraminiferal assemblages identified in the present investigation are similar to those with those identified by Pirazzoli et al. (1992), although
those workers reported that Foraminifera are only present within the sediments from the base of Trench A only as high as the layer they ascribed to the AD66 tsunami (sample 14 in Table 1). They stated that Foraminifera are not present in the AD66 tsunami layer, although they reappear above this unit (e.g. from sample 21 onwards). This conflicts with the findings of the present investigation, perhaps because the earlier interpretation was based on just 15 samples taken at regular intervals between 5.2 and 6.6ma.s.1. Finally, the calibrated age of 662 Bc 4-70 for the base of Trench A contrasts with a date of 522-340 BC at 4-1 SD for a sample 20cm farther up the stratigraphic column reported by Pirazzoli et al. (Fig. 3), and suggests that the harbour may actually have been in existence two centuries earlier than proposed by Pirazzoli et al. (1992). Discussion
In many parts of Crete, by mapping and dating sequences of uplifted, superimposed raised fossil marine notches, it is possible to identify those areas believed to have been affected by coseismic deformation associated with earthquakes (Spratt 1865; Flemming 1978; Pirazzoli 1986; Kelletat 1991). At Falasarna, a sequence of uplifted palaeo-shorelines suggests that a series of small, uniform episodes of subsidence occurred during the 2000-3000 years before c. 1530 4-40 radiocarbon aBl~ (Pirazzoli et al. 1981, 1982, 1992; Thommeret et al. 1981; Pirazzoli 1986) (Fig. 4). Radiometric dates on these raised shorelines show ages which decrease with increasing altitude ~"
Destru~ive
r
ea~hquake
11o
~kTsunami +7 267
66
.~.
"6
g
368
,_ . . . . . . . . . . . . . . . . . ,
,I .
. . . . .
255 t--
I
r--',
II I ....
,v ; .......
i-
I I I
I
__,
,,
$FA9
,,
'L,
I
i
I
I
'
7FA4 i
l
@ A
............
o
] III
, Iv'
,
365
__, ..... ---
,,
,, ,,
g
=
'v'
e~ []
Drop to present MSL
+5
1000
.
.
.
.
500
'
'BC'
'
0
'
'AD'
'
500
Year
Fig. 4. Band of relative sea-level changes at Falasarna between 1000 Bc and AD500 compared with historically destructive earthquakes and tsunamis in western Crete. Radiocarbon dates for the shoreline displacements are shown in Table 3. (After Pirazzoli et al. (1992).)
COASTAL TECTONICS IN CRETE
351
Table 3. Relative sea-level changes at Falasarna during historical times deduced from radiocarbon-dated shorelines. Shoreline
0 I II III IIIa IV IVl
Elevation (m a.s.1.)
+0 +6.6+0.1 +6.5+0.1 +6.35 4- 0.1 +6.25 + 0.1 +6.1 4-0.1 +5.9 :k:0.1
Displacement age
-,
aBP
Calibrated age range*
Inferred historical event
1530 + 40 i 600-1710 1780-1800 1880-1900 1950-1980 2250-2300 2500-2610
AD341-439 AD89-404 1613C~D 169 141 BC--AD69 235-18 Bc 728-378 Bc 991-759 Bc
AD 365 (?) AD 66 (?)
The shoreline numbers correspond to those shown in Fig. 4. (Adapted from Pirazzoli et al. (1992)). * Calibration according to Stuiver et al. (1986).
(Pirazzoli et al. 1992) (Table 3). The sedimentological analyses presented in this paper provide an independent record of late Holocene coastal tectonic movements. From Fig. 4 and Table 3 the palaeo-shoreline data ofPirazzoli et al. (1992) imply that an earthquake subsidence occurred c.AD 66, which displaced the contemporary shoreline from +6.35 4-0.1 to +6.5 +0.1 m. However, the sedimentology indicates a sudden change from marine to terrestrial conditions, which is believed to be associated with uplift of the harbour rather than subsidence. The sedimentology is thus in direct conflict with the geomorphological evidence. The last major tectonic displacement determined from the shoreline data in Fig. 4 relates to the uppermost of the emerged notches at c. 6.5 m a.s.1, radiometrically dated at 1530 4- 40 radiocarbon a ~31~(Table 3). The stratigraphic record preserves no evidence of this event in the form of tsunami-deposited sediments. It is difficult to understand why the stratigraphy at Falasarna records no evidence of such a large displacement. The implications of these findings are that, in general, it is difficult to correlate the very precise stratigraphic record with the raised palaeoshorelines at Falasarna and in particular with the +6.5 m shoreline associated with the inferred AD 365 tectonic uplift. Furthermore, the sedimentology does not reflect the suggested pattern of Holocene coseismic tectonic movements deduced from uplifted shoreline features.
Conclusions Data presented from Falasarna Harbour records a late Holocene palaeoenvironmental history which also contains detailed evidence of earthquake-related tsunamis. The results provide a useful independent dataset for evaluating the
magnitude and timing of coastal tectonic movements in an area affected by contrasting tectonic regimes. Sedimentation within the harbour began c. 662 Bc + 70, which is approximately two centuries earlier than previously reported. The results suggest that, on the basis of foraminiferal assemblage, assemblage variation and individual test preservation, it is possible to identify suites of tsunami-deposited sediments within normal sedimentary sequences. The evidence provided by the foraminifera shows clearly that the last episode of marine sedimentation within the harbour was during or just after the AD66 earthquake-tsunami event. Consequently, the sedimentology strongly suggests tectonic uplift of the harbour c. A D 66, rather than subsidence, as inferred from the raised palaeo-shoreline data. Significantly, there is no bio- or lithostratigraphic evidence to infer sedimentary deposition associated with a tsunami reported to have been generated by a large vertical tectonic displacement c. AD 365. It is not possible to interrelate distinct sedimentary horizons with raised shoreline features at Falasarna which have previously been used to describe late Holocene coastal tectonic movements in this area. Funding for this research was provided by the European Union under Contract EV5V-CT92-0175: Genesis and Impact of Tsunamis on the European Coasts (GITEC), administered by Directorate General XII (Science, Research and Development), Climatology and Natural Hazards Unit. We would also like to acknowledge the generous permission of E. Hadjidaki to undertake fieldwork at Falasarna Harbour.
References ANGELIER, J. 1978. Tectonic evolution of the Hellenic Arc since the Late Miocene. Tectonophysics, 49, 23-36.
352
D. D O M I N E Y - H O W E S E T AL.
CIMERMAN, F. & LANGER, M. R. 1991. Mediterranean Foraminifera. Ceneter Sazu Paleotoski Institut Ivana Takovca, Ljubljana, 2. DOM1NEY-HOWES, D. T. M. 1996. The geomorphology
MERCIER, J. L. 1981. Extensional-compressional tectonics associated with the Aegean Arc: comparison with the Andean Cordillera of south Peru-north Bolivia. Philosophical Transactions
and sedimentology of five tsunamis in the Aegean Sea region, Greece. Unpublished PhD thesis,
of the Royal Society of London, Series A, 300,
Coventry University. EMERY, K. O., AUBREY, D. G. & GOLDSMITH,V. 1988. Coastal neotectonics of the Mediterranean from tide-gauge records. Marine Geology, 81, 41-52. FLEMMING, N. C. 1978. Holocene eustatic changes and coastal tectonics in the northeast Mediterranean: implications for models of crustal consumption.
Philosophical Transactions of the Royal Society of London, Series A, 289(1362), 405-445. - - 1 9 9 2 . Predictions of relative coastal sea level change in the Mediterranean based on archaeological, historical and tide-gauge data. In: JEFTIC, L., MILLIMAN, J. D. & SESTtNI, G. (eds)
Climatic Change and the Level of the Seas. Vol. I The Mediterranean, 247-281. --
& WOODWORTH, P. L. 1988. Monthly mean sea levels in Greece during 1969-1983 compared to relative vertical land movements measured over different timescales. Tectonophysics, 148, 59-72. FROST, F. J. 1997. Tectonics and history at Phalasarna. In: SWINY, S., HOHLFELDER, R. L. & WYLDE SWlNY, H. (eds) Res Maritimae: Cyprus and the
Eastern Mediterranean from Prehistory to Late Antiquity. Cyprus American Archaeological Research Institute, Monograph Series, l, 107115. HADJIDAKI, E. 1988. Preliminary report of excavations at the harbour of Phalassarna in western Crete. American Journal of Archaeology, 92, 463-479. IAPSO 1985. Changes in relative mean sea level. Working Party Report of the International Association for the Physical Sciences of the Ocean Advisory Committee on Tides and Mean Sea Level. Eos Transactions, American Geophysical Union, 66, 754-756. JACKSON, J. 1994. Active tectonics of the Aegean region. Annual Review of Earth and Planetary Sciences, 22, 239-271. KELLETAT, D. 1991. The 1550BP tectonic event in the eastern Mediterranean as a basis for assessing the intensity of shore processes. Zeitschrift fu'r GeomotThologie , Supplementband, 81, 181-194. LALLEMANT,S., TRUFFERT, C., JOLIVET,L., HENRY, P., CHAMOT-ROOKE, N. & DE VOOGD, B. 1994. Spatial transition from compression to transition in the western Mediterranean Ridge accretionary complex. Tectonophysics, 234, 33-52. LE PICHON, X. & ANGELIER, J. 1981. The Aegean Sea.
Philosophical Transactions of the Royal Society' of London, Series A, 300, 357-372.
337-355. MURRAY, J. W. 1991. Ecology and Palaeoecology of Benthic Foraminifera. Longman, Harlow, UK. PARKER, F. L. 1958. Eastern Mediterranean Foraminifera. Swedish Deep Sea Expedition, 8, 219-279. PIRAZZOLI, P. A. 1986. The Early Byzantine tectonic paroxysm. Zeitschrift J~ir Geomorphologie, Supplementband, 62, 31-49. --, AUSSEIL-BADIE, J., GIRESSE, P., HADJIDAKI, E. & ARNOLD, M. 1992. Historical environmental changes at Phalasarna harbor, West Crete. Geoarchaeology, 7, 371-392. - - , THOMMERET, J., THOMMERET,Y., LABOREL,J. & MONTAGGIONI, L. F. 1981. Les rivages 6merg+s d'Antikithira (Cerigotto): corr61ations avec la Crete occidentale et implications cin6matiques et geodynamiques. Actes du Colloque Niveaux
Marins et Tectonique Quaternaires dans l'Aire Mdditerran~ene. University of Paris I, 49-65. & 1982. Crustal block movements from Holocene shorelines: Crete and Antikythira (Greece). Tectonophysics, 86, 2 2 4 3 . SIDEBOTTOM, H. (1904-1909). Report on the Recent Foraminifera from the coast of the Island of Delos (Grecian Archipelago). Manchester Literary and
Philosophical Society Memoirs and Proceedings. 48, 1-26; 49, 1-22; 50, 1-18; 51, 1-28; 52, 1-28; 53, 1-32. SPRATT, T. S. 1865. Travels and Researches in Crete. Vol. 2. J. van Voorst, London. STUIVER, M. & BRAZIUNAS, T. F. 1993. Modelling atmospheric 14C influences and ~4C ages of marine samples to 10,000B.C. Radiocarbon, 35, 137-189. --, PEARSON, G. W. & BRAZIUNAS,T. 1986. Radiocarbon age calibration of marine samples back to 9000 cal yr BP. Radiocarbon, 28, 980-1021. TAYMAZ, Y., JACKSON, J. & WESTAWAY, R. 1990. Earthquake Mechanisms in the Hellenic Trench near Crete. Geophysical Journal International, 102, 695-731. THOMMERET, Y., THOMMERET, J., LABOREL, J., MONTAGGIONI, L. F. & PIRAZZOLI, P. A. 1981. Late Holocene shoreline changes and seismotectonic displacements in western Crete (Greece). Zeitschrift jh'r Geomorphologie, Supplementband, 40, 127 149. WETMORE, K. L. 1987. Correlations between test strength, morphology and habitat in some benthic Foraminifera from the coast of Washington. Journal of Foraminiferal Research, 17, 1-13.
Possible tsunami deposits from the 1855 earthquake, North Island, New Zealand JAMES
R. G O F F 1'3, M I C H A E L URSULA
C R O Z I E R 1, V E N U S
COCHRAN 1 &
SUTHERLAND
l,
PHIL SHANE 2
1 School o f Earth Sciences, Victoria University of Wellington, PO Box 600, Wellington, New Zealand 2 Department of Geology, School o f Environmental and Marine Sciences, University of Auckland, Tamaki Campus, Private Bag 92019, Auckland, New Zealand 3 IGNS, PO Box 30-368, Lower Hutt, New Zealand (e-mail.
[email protected]) Abstract: A series of three fining-upward sequences from deposits in the Okourewa Stream
bank on the south coast of the North Island, New Zealand, investigated by grain-size, diatom, radiocarbon, geochemical and macrofaunal analyses have been tentatively interpreted as the products of a tsunami. The proposed event consisted of three separate waves (the second being the largest) generated by a surface rupture of a local fault. Changes in diatom assemblages and the presence of marine shells, pumice, and beach pebbles may represent a tsunami advancing inshore over beach, freshwater channel, and coastal wetland enviromnents. Deposition occurred between AD500 and 1890. The event in question may have currently the AI~1855 rupture of the West Wairarapa fault. Internationally, there has been a recent increase in the amount of literature concerning tsunami research on both contemporary and palaeotsunami events (e.g. Dawson et al. 1996). Previous research concentrated on palaeo-tsunami in two main regions, the Pacific (Australia (Bryant et al. 1992), Hawaii (e.g. Moore et al. 1994), Japan (e.g. Minoura et al. 1994), Pacific Northwest (e.g. Atwater 1987; Clague & Bobrowsky 1994)) and Northwest Europe (e.g. Dawson et al. 1988). This work has served to improve our understanding of tsunami signatures preserved in the sedimentary record, and a summary of diagnostic criteria used to identify palaeotsunami is given in Table 1. In New Zealand, tsunami research is generally limited to either modelling for coastal hazard planning purposes (e.g. Gilmour 1960; Victory et al. 1989; Van Dissen et al. 1994) or establishing a record of palaeo-tsunami events (Eiby 1982; DeLange & Healy 1986). DeLange & Healy (1986) have produced a valuable, yet limited, record of at least 32 documented tsunami dating back to AD 1840 which indicates that locally derived tsunami produce far larger waves than their exogenic counterparts. The lack of tsunami research in New Zealand is somewhat surprising bearing in mind that the country sits astride the boundary of the Pacific and Australian plates, and thus is subject to considerable locally generated tectonic activity. Moreover, the country is also exposed to several exogenic tsunami sources. The east coast of New
Zealand is exposed to tsunami generated by earthquakes occurring almost anywhere around the Pacific Ocean, but particularly South America (Heath 1976, 1977). The west coast is exposed to the Tasman Sea and Australia, where tsunami may be refracted around New Zealand or reflected from the continental shelf of Australia (Braddock 1969). There is evidence that the tectonically inactive coast of New South Wales, Australia, has been inundated by several large tsunami events (Bryant et al. 1992; Young et al. 1993, 1996), the most recent about 800 years ago. These records point to the existence of other tsunamigenic sources such as subaquatic landslides off the continental shelf or meteor impacts (Bryant et al. 1996). The Wellington region is located at the southern end of the North Island (Fig. l a) and is separated from the South Island by Cook Strait, a 22km wide, 1500m deep channel in which the edge of the continental shelf is within 0.5-2.0 km of the present-day coastline (Mitchell & Lewis 1980; Carter et al. 1988). There are four class-one active faults in the Wellington region (that is those that are known to have moved at least once in the last 5000 years): the Wairarapa, Wellington, Ohariu and Shepherds Gully faults (Van Dissen & Berryman 1996), all of which are believed to extend across Cook Strait (Carter et al. 1988). Berryman (1990) considered that movement on the Wellington, Wairarapa and Ohariu faults and the offshore extension of the Wairau fault (South Island)
GOFF, J. R., CROZmR, M., SUTHERLAND,V., COCHRAN,U. & SHANE, P. 1998. Possible tsunami deposits from 1855 in North Island, New Zealand. In: STEWART, I. S. & VITA-FINZI, C. (eds) Coastal Tectonics. Geological Society, London, Special Publications, 133, 353 374.
354
J. R. G O F F E T A L .
Table 1. S u m m a r y o f diagnostic criteria used to identify tsunami deposits
(1) (2) (3)
(4) (5) (6) (7)
(8)
(9) (10)
(11) (12) (13) (14)
Diagnostic characteristics
References (e.g.)
Generally fines inland and upwards Each wave can form a distinct sedimentary unit, although this is not often recognized in the sedimentary sequence Distinct upper and lower sub-units representing run-up and backwash can be identified, B U T investigation of recent tsunami deposits indicates that there is still considerable uncertainty about when most deposition occurs (during run-up or backwash) and so these sub-units may be related to other processes Lower contact is unconformable or erosional Can contain intraclasts of reworked material, but these are not often reported Often associated with loading structures at base of deposit Particle and grain sizes range from boulder layers (up to 750 m3), to coarse sand to fine mud. However, most deposits are usually recognized as anomalous sand units in peat sequences Generally associated with an increase in abundance of marine to brackish-water diatoms, but reworking of estuarine sediments may simply produce the same assemblage; preservation of frustules can be excellent, although many are often broken Marked changes in Foraminifera (and other microfossils) assemblages. Deeper-water species are introduced with catastrophic saltwater inundation Increases in the concentrations of sodium, sulphate, chlorine, calcium and magnesium occur in tsunami deposits relative to under- and overlying sediments; indicates saltwater inundation Individual shells and shell-rich units are often present Often associated with buried vascular plant material and/or buried soil Shell, wood and less dense debris often found 'rafted' near top of sequence Dating of tsunami sediments is problematic. Best results for dating are from units above and below to 'bracket' the event. Radiocarbon ages often equivocal because of older reworked carbon; dating of introduced marine macrobiota is preferred (and successful). Optical dating (OSL) is the best method available assuming the sediments were exposed to daylight during reworking by the tsunami
Foster et al. 1991; Dawson 1994 Ota et aI. 1985; Moore & Moore 1988; Clague & Bobrowsky 1994
could account for 80-100% of the strain rate measured in Wellington. The most recent movements of faults in the C o o k Strait region occurred on the southern portion of the Wellington fault c. AD 1350, c. AD 1250 and c. AD 1450 (Best 1923; Stirling 1992; Van Dissen et al. 1992; Ian R. Brown Associates Ltd 1995; Van Dissen & Berryman 1996); the AD1855 earthquake occurred on the West Wairarapa fault 2 0 k i n
Moore & Moore 1988; Dawson et al. 1996
Dawson et al. 1988; Moore & Moore 1988 Dawson 1994; Moore et al. 1994 Foster et al. 1991; Minoura & Nakaya 1991 Ota et al. 1985; Moore & Moore 1988; Minoura & Nakaya 1991; Dawson 1994; Minoura et al. 1994; Young et al. 1996 Dawson et al. 1988; Minoura et al. 1994; Hemphi11Haley 1996
Patterson et al. 1996; Sherman et al. 1996 Minoura et al. 1994; Goff & Chagu&Goff 1998
Moore & Moore 1988; Bryant et al. 1992 Clague & Bobrowsky 1994; Dawson 1994 Albertho & Martins 1996; Imamura et al. 1997 Dawson et al. 1988; Hansom & Briggs 1991; Clague & Bobrowsky 1994; Dawson 1994; Huntley & Clague 1996
east of Wellington (e.g. D a r b y & Beanland 1992; Stirling 1992). The AD 1855 earthquake represented a surface rupture of the West Wairarapa fault (Fig. l a) along a 200 k m segment from southern Hawke's Bay to the middle of C o o k Strait (Robinson 1986; Carter et al. 1988; Barnett et al. 1991). The offshore fault trace is sub-parallel to the coastline and marks the edge of the continental shelf
POSSIBLE TSUNAMI DEPOSITS (1855?), NEW ZEALAND
355
Fig. 1. Okourewa Stream: (a) regional location map; (b) study area (showing transect A-B), (e) Surveyed transect from storm ridge A (modern), to Okourewa valley floor, B. (Begg & Mazengarb 1996). Hull & McSaveney (1996) have produced a model to explain the complex pattern of surface faulting and ground deformation at the southern end of the West Wairarapa fault, which indicates that the centre of maximum uplift (>5 m) was near Turakirae Head as opposed to being several kilometres to the NE. The uplift rate is believed to have been about 2.9m per 1000 years over the last 7200 years (Hull & McSaveney 1996). Although the West Wairarapa fault is to the west of the study site (Fig. la), it gives an indication of the complex tectonic history of the region. The study site is in an area where the tectonic history is poorly understood, although the Okourewa valley is believed to follow the Pirinoa fault, c. 2 km west
of the Moanatahi Syncline (Ghani 1974, 1978). The projected average uplift rates for synclines in the region is between 0 and 2.2m per 1000 years for the last c. 125 000 years (Ghani 1974, 1978; Pillans 1986). There are few contemporary accounts of tsunami along the south coast of the Wellington region and all refer to the AD 1855 earthquake. This event generated a 9 - 1 0 m high tsunami in Cook Strait and a 3 - 4 m high wave at the entrance to Wellington Harbour, and caused severe seiching inside (Fox 1855; Barnett et al. 1991). In Palliser Bay, three waves inundated the coast, the first being larger than the following two (Mason 1855). Computer modelling of tsunami inside Wellington Harbour concurs with
356
J. R. GOFF ET AL.
contemporary accounts, but separates the signal of external (to the harbour) 'seismic forcing' from that of internal effects (Barnett et al. 1991). This effectively explains the unique response of Wellington Harbour to the AD 1855 earthquake, but places the events in Palliser Bay outside the parameters of the study by Barnett et al. However, it is believed that contemporary accounts underestimated tsunami height outside Wellington Harbour (A. Barnett, pers. comm., 1997).
Physical setting Okourewa valley (41~ 175~ is located in southern Wairarapa, c. 30 km east of Wellington, on the south coast of the North Island, New Zealand (Fig. l a). Okourewa Stream is oriented NE-SW, draining a catchment of about 10km 2 and emptying into a lagoon formed on the landward side of the AD 1855 storm ridge (Figs lb and c, and 2). The lagoon appears to drain by seepage through the storm ridge (about 5m above sea level (asl)), maintaining a maximum lagoon depth of about 1.0 m. There is little existing geological information relating specifically to Okourewa valley, although reconnaissance surveys of Tertiary- and Pleistocene-age geological features in the region were carried out by McKay (1878, 1879), King (1930, 1933) and Cotton (1942). The local geology comprises the Hautotara Formation of bluegrey mudstone (Pliocene) laid down in a marine inner-shelf environment (Begg & Mazengarb 1996), and unnamed Pleistocene marine terraces cut during several stages of the last interglacial (Palmer & Vucetich 1989). Pre-European settlement in the area is indicated by a Maori Pa (fort), which is situated on the west side of the valley, and the remnants of several shell middens occur on the valley slopes (Adkin 1959). The length of occupation is unknown, but Maori occupied the valley at the time of European settlement (c. AD 1840), providing the initial ferry service across the mouth of Lake Onoke (McIlraith, pers. comm., 1996). No field evidence exists for Maori occupation of the valley floor, although early European documents indicate the existence of a Pa adjacent to the east bank of the stream (Adkin 1959). At the time of European settlement, the lake drained into the sea east of the mouth of Okourewa valley. Subsequent river flooding in the early 1950s redirected the mouth to the west, to the position that it occupies today. The valley floor is currently under farmland, although access to a 1 km length of stream is restricted by forestry plantations.
Fig. 2. Schematic diagram of Okourewa valley showing sections described in the text: (a) Seaward end of valley; (b) landward end, 800 m north of Fig. (a).
Study methods Surveying A topographical profile of the beach, lagoon and Okourewa valley floor as far as site LF2 (Figs lb and c, and 2) was measured in relation to benchmark locations at Lake Ferry and the landward end of the valley (Stewart trig point). Mean sea level (msl) was estimated based upon the Cape Palliser tide levels (which are offset from the Wellington tide tables by +10min). The profile was repeated twice and averaged, with errors of 4-0.2 m caused mostly by natural variations in storm ridge height.
Sampling and sediment analyses Sections of exposed stream bank were cleaned by removing weathered sediment with a spade to expose a fresh surface and brushing loose material from the exposure. The sections were photographed and the exposed stratigraphy was described. Sediment samples were generally taken at O.1m intervals, and at smaller intervals where distinct units were not covered by the
POSSIBLE T S U N A M I D E P O S I T S (1855?), N E W Z E A L A N D standard sampling interval. Sediment samples were taken for grain-size and diatom analyses, wood and charcoal for radiocarbon analysis, and pumice and shells for identification. Grain-size analysis followed procedures described by Barrett & Brooker (1989). Organic material was removed with hydrogen peroxide treatment for 7 days. Salts and acids were removed with distilled water and then centrifuged. After the supernatant liquid was poured off, sediment samples were washed with sodium hexa-metaphosphate and wet sieved at 60 #m to separate the sample into coarse and fine fractions. Coarse fractions were dry sieved at 89 intervals (0.63-4.00mm) using a Frisch shaker. Fine fractions were dried for 24h at 100~ and a 1.5-2 g sub-sample was used in SediGraph analysis. Data were entered into a PC software package (SIZE) to produce grainsize distribution indices. Bulk density was determined by sampling a known volume of sediment (10cm3). Samples were dried for 24h at 100~ then reweighed, and dry bulk density was calculated. Loss on ignition (LOI) was determined by ashing at 550~ for 16h. Diatoms were concentrated by digesting samples in 27% hydrogen peroxide to remove organic matter, adding 32% hydrochloric acid to remove carbonate, and heating at 100~ for 30min. Sand was extracted by decanting and clay was removed by washing in sodium hexa-metaphosphate. Drops of the liquid were dried on a slide, mounted in naphrax and examined under x1600 magnification. Two hundred frustules were counted from each sample. Changes in relative concentrations of diatoms are based on the number of frustules counted along one transect of the cover slip (all samples were prepared with the same weight of sediment and diluted by the same amount, so that relative concentrations can be compared between samples). Fifteen pollen samples were taken from three sites (LF1, LF2, and LF5). Samples were taken at 5cm intervals down-section from the surface. Samples were collected from a 5 - 1 0 m m surface layer using a metal spatula and were prepared for palynomorph analysis using the technique outlined by Moore & Webb (1978), with at least 300 grains studied on each slide. As discussed below, the primary aim was to identify the first appearance of Pinus pollen, and detailed pollen spectra were not produced. Clast roundness and sphericity were measured at nine sites using the criteria laid down by Folk (1980). A total of 1150 greywacke clasts were measured, with the mean roundness and sphericity of 50 clasts being calculated for each sub-site, and a similar number for each of the additional sites. Fabric measurements were taken from prolate clasts at site LF2 to determine flow direction. Two subsamples of 51 clasts each were measured and Schmidt equal-area, lower hemisphere projections of stone a-axis fabrics were produced. In both cases, contours were drawn at every one, three, five, etc., points per (100/n) % of the projected area, where n is the sample size. The technique and presentation of data are well established for glacial and fluvial sediments (e.g. Mark 1973; Hicock et al. 1996; Maizels 1996).
357
Radiocarbon analysis Charcoal and wood fragments were taken from two sections for radiocarbon analysis. The largest available samples were analysed, to avoid possible contamination caused by the incorporation of organic material in sediments of a different age. Conventional ages are reported according to Stuiver & Polach (1977). Calibrated ages were obtained using the bidecadal curve developed by Stuiver & Pearson (1993), with 27 ~4C years subtracted for the Southern Hemisphere offset (T. Higham, pets. comm., 1997).
Identification of pumice and shells Glass shards from pumice samples and tephra were analysed by electron microprobe. Results were compared with the glass chemistry (CaO wt % v. FeO wt %) of sea-rafted pumice and tephra from four major volcanic events, the Kawakawa tephra event at 22 ka BP (Taupo), the Macauley Island rhyolite at 6kaBp, Loisels sea-rafted pumice at c. 0.5kaBP (P. Froggatt, pers. comm., 1997), and the Taupo pumice (tephra) at 1.85 ka BP. Shells were picked from stream bank exposures for subsequent identification. There was concern that reworking of old shell midden sites would produce considerable shell material of diverse origin and age. However, as all shells found were juvenile, with articulated and intact bivalves, it seems unlikely that they are derived from old shell midden sites.
Results Results o f stratigraphic, d i a t o m , r a d i o c a r b o n , geochemical a n d m a c r o f a u n a l analyses carried o u t at sites LF1, L F 2 , L F 5 a n d L F 6 are detailed in Figs 3 a n d 4. Sites L F 3 a n d L F 4 (Fig. 5) are discussed below.
Site LF1 (Figs 3a and 4a) Stratigraphically, three erosional c o n t a c t s can be recognized within the preserved sequence. Sedim e n t a r y units overlying each c o n t a c t represent f i n i n g - u p w a r d s sequences t e r m i n a t e d by a f u r t h e r erosional contact. T h e e x c e p t i o n is the u p p e r m o s t f i n i n g - u p w a r d s sequence, w h i c h a p p e a r s to t e r m i n a t e at a g r a d a t i o n a l c o n t a c t with an overlying s a n d y l o a m unit. T h e lowest erosional c o n t a c t overlies a basal unit (below 1.06 m) o f organic-rich silt. G r a i n sizes in each sequence generally fine t h r o u g h a series o f g r a d a t i o n a l c o n t a c t s f r o m pebbles to silt or sand. E a c h f i n i n g - u p w a r d s sequence has different characteristics. T h e lowest sequence has a distinct a n d u n i f o r m clasts u p p o r t e d unit (with a lens (intraclast?) o f
358
J. R. G O F F E T AL.
Fig. 3. Stratigraphy of sections (including, where applicable, mud and organic percentages, summary diatom assemblage, fabric data and chronological information): (a) section LF1, (b) section LF2, (c) section LF5, (d) section LF6.
POSSIBLE TSUNAMI DEPOSITS (1855?), NEW ZEALAND
359
Fig. 3. (continued)
deformed sand), overlain by a massive sand with organic content decreasing upwards. Decreases in both grain size and organic content continue through a gradational contact into a massive silt. An in situ marine shell, Dosinia lambata, was found at 0.73 m below ground level. The next sequence is a unit composed of a lower clast-supported section and an upper, matrix-supported section. This is overlain by the final sequence, which is predominantly a sand unit, with sand deformed around rare pebblesized clasts, fining upwards into a massive sand. Deformation appears to be associated with loading structures such as water escape features (e.g. flames). Changes in mud and organic content match the stratigraphic record, with the most significant changes occurring between 0.92 and 1.11 m. Both mud and organic contents decrease upsection across the lowest erosional contact. Diatom samples were analysed from 0.53 to 1.18m across two erosional contacts (0.60 and 1.08 m). Frustules for non-marine species were well preserved (marine ones were broken), and relative concentrations decreased above the lower erosional contact. In general, the diatom assemblages remain unchanged, although there are two exceptions. At 0.94 m above the lowest erosional contact, there is a slight increase in the
relative percentage of polyhalobous (marine) and a large increase in halophilous (brackishfresh) assemblages recorded in a sand lens sample. Corresponding decreases in mesohalobous (brackish) and oligohalobous (fresh) assemblages are evident. At 0.53-0.58 m above the next erosional contact, there is a marked increase in oligohalobous (fresh) assemblages, and a corresponding decrease in the others. Pollen samples taken in the upper section of the exposure record a distinct change in pollen assemblage at about 0.10m with the first appearance of Pinus pollen.
Site LF2 (Figs 3b and 4b) The stratigraphic record is similar to that at site LF1, with three erosional contacts. All are associated with overlying fining-upwards sequences. Underlying organic-rich silts are deformed by loading structures (e.g. water escape features or flames). The lowest sequence fines upwards from coarse to fine sand with one marine shell, D. lambata, found at the top (1.23 m) of the unit. The central sequence fines from clast-supported cobbles (a maximum a-axis of 0.20 m) and pebbles to a massive sand incorporating several marine shells near the upper contact (D. lambata). Measurements of 51 clasts each in the lower
Fig. 4. Diatom species, salinity and habitat: for section LF1 (0.53-1.118m) and section LF2 (1.27-1.82m).
(e)
Diatom Salinity, Section LF1
/
/
Depth~ om
/
/ -I
-1 / \ I
.J
J
I
-4 I
J
o
(d)
,,,,,,,,, 10 20 30 4 0
,
,,,,,,,,,, 10 20 30 40
,,,,,,,,, ,,,, ,,,,, 10 20 30 40 SO 60 70 0 10
_
,
Diatom Habitat, Section LF1
-f
J
cn,t
I
J
/
/ /
7
I
6O
7O
t
t-i I 80
"-'1 100
i
,n,n ,i, i, 10 20 30 40
Fig. 4. (continued)
i , i , i , i i , i , L , i , Li I r I , I , L , I , I , I , I , I , 0 10 20 30 O 10 20 30 40 0 10 20 30 40 GO 0 10 20 L~t
I
LiI 10
110
Fig. 4. (continued)
POSSIBLE TSUNAMI DEPOSITS (1855?), NEW ZEALAND
363
Fig. 4. (continued) (1.10-1.20 m) and upper (0.85-1.05 m) sections indicate that clasts possess a weak fabric, with flow directions from 241 ~ and 82 ~, respectively. The upper sequence fines upwards from clast-
supported pebbles and gravels to a massive sand unit, the latter containing wood with a 14C age of 1350+ 190aBP (Table 2). This is overlain by a unit of sandy loam.
Fig. 5. Section LF4: photograph showing stratigraphy of uppermost units.
364
J. R .
GOFF
E T AL.
t~
~ r,~ r.t3
,.._.., ..
,,~
<~
O
t.g5
.<
,<
t"q
r
rt---
I
I
Q O"
O
8 er
c'-,I
O o t"q tg%
,-..., r
r~
O
O
>
POSSIBLE TSUNAMI DEPOSITS (1855?), NEW ZEALAND Mud content changes most significantly across the lowest contact (1.55-1.72m), and generally matches the physical stratigraphy upsection, with higher percentages occurring in the finer units. Organic content is more variable than at LF1, with abrupt increases at the base of each fining-upwards sequence. Diatoms were analysed from 1.26 to 1.83m across the lowest erosional contact (1.60 m). Frustules for non-marine species were well preserved (marine ones were broken), and relative concentrations decreased above the contact. There were generally few polyhalobous (marine) diatoms, although there is a small increase immediately at and above the erosional contact. This section also preserves a minor increase in (mesohalobous) brackish diatoms. Pollen samples taken in the upper section of the exposure record a distinct change in pollen assemblage between 0.15 and 0.20m with the first appearance of Pinus pollen. A detailed analysis of the diatom assemblage is discussed below.
Site LF3 Stratigraphy is similar to that of previous sites. There are three erosional contacts (0.44, 0.88, and 1.11 m), each overlain by a fining-upwards sedimentary sequence. The lowest sequence decreases in grain size from a pebble-sand to fine sand, the central from cobbles to fine sand, and the upper from coarse to fine sand. Pumice (Table 2: Taupo pumice c.l.85kaBP) and marine shell (Table 2: D. lambata) material were found in the upper section of the central sequence near the upper contact.
Site LF4 (Fig. 5) Stratigraphy is similar to that at previous sites and the upper two erosional contacts are shown in Fig. 5. The basal organic-rich silts extend from 1.20 to 1.45 m below ground surface. An ash layer was identified at 1.42 m and subsequent geochemical analysis of glass shards indicates that it is the Kawakawa tephra dated at c. 22 ka BP. There are three erosional contacts (0.14, 1.07 and 1.22 m), each overlain by a fining-upwards sedimentary sequence. However, in the lowest sequence, the fining from pebbles-cobbles to medium sand occurs within 0.15m up-section. The central sequence fines from cobbles (maximum a-axis 0.36m) to fine sand, with the truncated upper sequence apparently grading from coarse to medium sand. The upper sequence appears to be incomplete and is overlain by humus. Pumice (Table 2: Taupo pumice c. 1.85 ka BP) and marine
365
shell (Table 2: D. lambata) material were found in the upper section of the central sequence near the upper erosional contact.
Site LF5 (Fig. 3c) Site LF5 is located c. 1.5 km inland, and is the most landward site reported in this paper. There is only one erosional contact (0.94 m), but two distinct gradational contacts up-section (0.40 and 0.67m) are marked by single layers of pebbles. Coarser pebbles were found in the lower of these two layers. Massive sand units overlie each pebble layer, the lower unit being coarser than the upper. Rare marine shells (D. lambata) were found at 0.06 m depth.
Site LF6 (Fig. 3d) This site was situated on a small tributary stream about 40m west of the main channel. Three erosional contacts were identified (0.10, 0.65 and 0.80 m below the surface). The contacts at 0.80 and 0.65 m had overlying fining-upwards sequences, but the uppermost contact was overlain by a thin, massive sand unit. A pumice-rich layer (Table 2: Taupo pumice), immediately over-lying a wood-rich layer, was situated immediately beneath the central erosional contact, and more pumice was found in the central finingupwards sequence. A sample from the wood-rich layer yielded a radiocarbon age of 1820 + 50 a BP (Table 2). Charcoal (0.10m) sampled from the uppermost massive sand unit produced a radiocarbon age of 2130 + 260 a BP (Table 2).
Clast roundness and sphericity (Fig. 6) The maximum variance for the indices of clast roundness and sphericity between storm ridge A and LF5 is 0.14. However, indices at storm ridge A and LFS, 1.5 km inland, are almost identical. There is some variability, mainly noticeable as a decrease in roundness and increase in sphericity between the two-end points. Samples taken from additional locations landward and to the west of the study sites (Fig. 2) are less spherical and less rounded, but are consistent within their own group.
Discussion
Stratigraphy Organic-rich silts were found at the base of each section. Furthermore, most sections contained a
366
J. R. GOFF E T AL.
Fig. 6. Clast roundness and sphericity indices ( I , roundness; U], sphericity). Location of storm ridge A is shown in Fig. 1; the locations of additional sites are shown in Fig. 2.
record of three fining-upwards sequences, the base of each being marked by an erosional contact (excluding the upper two contacts in LF5). With the exception of LF6 to the west of the main stream channel, there appears to be a general fining inland for each sedimentary sequence. This is visible on a gross scale from LF1 to LF5. The second fining-upwards sequence has coarser clasts and matrix (excluding LF6) and occupies a greater depth of profile in the central section of the study area from LF2 to LF4, which tends to suggest a higher-energy environment. At LF2, underlying sands are deformed by water escape structures, and fabric measurements are indicative of rapid, high-energy water deposition (Maizels 1996). Basal fabric shows flow in a landward direction, but in the finer, matrix-supported section flow directions are from the landward side. Fabric is stronger in the finer, matrix-supported section, which suggests a slightly lower-energy regime. Overlying sediments become finer, with rare shells present near the upper contact. It is possible that this sequence represents deposition from the incoming and outgoing segments of a catastrophic 'wave'. In the lowest part of the unit coarse, high-energy deposits may have been laid down so rapidly as to confer only a weak fabric. A subsequent decrease in energy
associated with the backwash of the wave laid down finer sediments, reoriented coarser clasts with a stronger 'landward' fabric, and laid down a layer of finer, more shell-rich material at the top of the sequence. The inferred depositional processes discussed above may be representative of the general depositional environment, although it appears to have varied considerably. For example, at LF2, the third fining-upwards sequence is finer than the underlying one and has no shells, although there are wood fragments. A possible interpretation is that a third 'wave' (out of a series of three) might have reworked much of the upper portion of the deposit laid down by the second 'wave', but being lower in energy introduced a generally finer particle size. Continued reworking could cause destruction of shell material and only rare wood fragments would be preserved. The stratigraphic interpretation above points to a rapid 'catastrophic' deposition of finingupwards sequences by a train of possibly three waves (Fig. 7). Fining-upwards and fininginland sediments, 'rafting' of less dense shell, wood and pumice material, sediment deformation, and the reworking of antecedent deposits laid down by the previous wave are supportive of this interpretation.
POSSIBLE TSUNAMI DEPOSITS (1855?), NEW ZEALAND
367
Fig. 7. Inferred depositional sequence in Okourewa valley caused by three waves during one tsunami event. gc, Gradational contact; ec, erosional contact.
Diatom assemblage (Figs 3a and b, and 4; Table 3) Sections LF1 and LF2 show relatively minor changes in marine and brackish diatoms in association with the lowest erosional contact. A second, mid-section, erosional contact shows distinct increases in fresh-water species. The study area is located in an exposed coastal environment and it is therefore highly likely that sediments eroded by incoming catastrophic waves included a high proportion of nearshore diatoms. Nearshore diatoms could have been introduced by both storm waves and wind. Erosion of sediments by catastrophic wave inundation would merely mix and redeposit a similar diatom assemblage (e.g. Hemphill-Haley 1996). This may explain the minor increase in polyhalobous (marine) diatoms above the lowest erosional contact, and the high numbers of mesohalobous (brackish) and oligohalobous (fresh) diatoms. An increase in oligohalobous (fresh) diatoms above the second erosional contact seems likely to have been 'inherited' from the antecedent conditions induced by the first wave. Landward, oligohalobous diatomrich sediments were drawn down by the backwash of the first wave, and subsequently reworked by the second wave.
The habitat assemblage is predominantly benthic, and is indicative of a shallow waterwetland area with deeper, open water nearby (planktonic species at LF 1). A peak in planktonic species at 0.94m (LF1) might indicate the seaward introduction of deeper water species, although as the peak consists of an individual species it may represent a specific source area. Up-section increases in tychoplanktic species (LF1) and epiphytic species (LF2: one species only) are interpreted as the introduction of material from the periphery of the wetland. Within the general context of site stratigraphy, up-section variations in planktonic, tychoplanktic and epiphytic species may therefore represent wave run-up and backwash. Although there is no significant influx of polyhalobous (marine) diatoms, the diatom species diagram for LF2 suggests that individual species fit a trend of saltwater inundation. The marine diatom Hyalodiscus cf. scoticus is not present in the basal silt sequence, but increases to 5% in the sediments of the inferred catastrophic wave. The brackish-marine diatom Nitzschia compressa lives in sand; it increases from 5% in the silt sequence to 15 % in the sandy layers above. The highest numbers of brackish species Campy-
lodiscus echeneis, Cocconeis scutellum, Nitzschia levidensis, Nitzschia circumsuta, Melosira
368
J. R. GOFF E T AL.
Table 3. List of diatom species identified in sections LF1 and LF2, their salinity and habitat preferences Species
Salinity*
Habitatt
Achnanthes clevei Grunow A. delicatula (Kutzing) Grunow A. delicatula spp. haukiana Lange-Bertalot A. lanceolata (Brebisson) Grunow A. ploenensis Hustedt Amphora coffeaeformis (Agardh) Kutzing A. libyca Ehrenberg A. cf. montana Krasske Aulacoseira sp. Campylodiscus echeneis Ehrenberg Cocconeis placentula Ehrenberg C. scutellum Ehrenberg Cyclostephanos dubius (Fricke) Round Cyclotella atomus Hustedt C. meneghiniana Kutzing C. stelligera Cleve & Grunow Cymbella sp. Diploneis elliptica (Kutzing) Cleve D. puella (Shumann) Cleve D. smithii (Brebisson) Cleve Epithemia adnata (Kutzing) Brebisson E. sorex Kutzing Fragilaria construens (Ehrenberg) Grunow F. pinnata Ehrenberg F. ulna (Nitzsch) Lange-Bertalot Gyrosigma acuminatum (Kutzing) Rabenhorst G. attenuatum (Kutzing) Rabenhorst Hyalodiscus lentiginosus John Hyalodiscus cf. scoticus (Kutzing) Grunow Melosira cf. dickeii (Thawaites) Kutzing M. cf. moniliformis (O. F. Muller) Agardh Navicula sp. N. capitata (Grunow) Ross N. cf. capitatoradiata Germain N. contenta Grunow N. cryptocephala Kutzing N. digitoradiata (Gregory) Ralfs N. elginensis (Gregory) Ralfs N. halophila (Grunow) Cleve N. placentula Ehrenberg (Grunow) N. pusilla W. Smith N. pygmaea Kutzing N. salinarum Grunow N. tenera Hustedt Nitzschia sp. N. acuminata (W. Smith) Grunow N. circumsuta (Bailey) Grunow N. compressa (Bailey) Boyer N. cf. frustulum (Kutzing) Grunow N. hungariea Grunow N. inconspicua Grunow N. levidensis (W. Smith) Grunow N. marginulata Grunow N. panduriformis Gregory N. cf. terrestris (Petersen) Hustedt N. tryblionella Hantzsch Pinnularia sp. Pinnularia subcapitata Gregory Pleurosigma angulatum (Quekett) W. Smith Rhopalodia acuminata Krammer R. brebissonii Krammer
Oligo Meso Oligo Oligo Oligo Oligo Oligo Oligo
Epiphytic Epiphytic Epiphytic Epiphytic Epiphytic Benthic Benthic Benthic
Meso Oligo Meso Halophil Halophil Halophil Oligo
Epipelic Epiphytic Epiphytic Planktonic Planktonic Planktonic Planktonic
Oligo Oligo Meso Oligo Oligo Oligo Oligo Oligo Otigo Oligo Poly Poly Oligo Meso
Benthic Benthic Benthic Epiphytic Epiphytic Tychoplank. Tychoplank. Epiphytic Epipelic Benthic Epiphytic Epiphytic Benthic Benthic
Oligo Oligo Oligo Oligo Meso Oligo Meso Oligo Halophil Halophil Meso Halophil
Epipelic Epipelic Aerophil. Epipelic Epipelic Epipelic Epipelic Epipelic Epipelic Benthic Epipelic Epipelic
Halophil Meso Meso Halophil Halophil Halophil Halophil Meso Meso Oligo Halophil Oligo Halophil Halophil Halophil
Benthic Benthic Epipelic Benthic Epipelic Benthic Epipelic Epipelic Epipelic Benthic Benthic Epipelic Benthic Benthic Benthic
POSSIBLE TSUNAMI DEPOSITS (1855?), NEW ZEALAND
369
Table 3. (continued) Species
Salinity*
Habitatt
R. gibba (Ehrenberg) O. Muller Surirella angusta Kutzing S capronii Brebisson S. cf. splendida (Ehrenberg) Kutzing S. subsalsa W. Smith Thalassiosira lacustris (Grunow) Hasle T. weisflogii (Grunow) Fryxell & Hasle
Oligo Oligo Oligo Oligo Halophil Oligo Halophil
Epiphytic Epipelic Epipelic Benthic Benthic Planktonic Planktonic
* Salinity: poly, polyhalobous, lives in marine conditions; meso, mesohalobous, brackish water; halophil, halophilous, fresh water, likes some salt; oligo, oligohalobous, fresh water, can tolerate some salt. t Habitat: benthic, bottom dwelling, or attached to a surface; epipelic, lives on silt; epiphytic, lives attached to plants; planktonic, lives in open water; tychoplank., tychoplanktic, lives in shallow water; aerophil, aerophilic, lives on damp soil.
cf. moniliformis and Nitzschia marginulata also occur in the inferred catastrophic wave deposit. Opposite trends occur in fresh-water species such as Cocconeis placentula and Thalassiosira lacustris, which decrease noticeably from the silts to the overlying sediments. The location of the stream mouth in AD 1855 and its association with the drainage channel of Lake Onoke and the barrier are known from the historic record (see above). However, before AD 1855 the nature of the nearshore geomorphology is unknown, although uplift would probably have isolated the stream behind successive storm ridges. It seems likely that the diatom assemblage recorded above the erosional contact reflects the sandy, brackish-marine environment of the Lake Onoke drainage channel. If this is the case, the deposit probably records the movement of a wave(s) over the storm ridges (cobbles and pebbles appear to be sourced from these) and through the drainage channel (Atwater 1992; Hemphill-Haley 1996). Abundant, wellpreserved brackish-marine taxa, and common tidal flat species (Campylodiscus echeneis, Cocconeis scutellum) suggest a locally generated catastrophic wave source (e.g. Clague & Bobrowsky 1994), although incorporation of the marine diatom Hyalodiscus cf. scoticus might indicate a more distantly sourced event (Hemphill-Haley 1996). However, the Hyalodiscus species generally live on seaweed or unattached on sediment of the inshore zone (Round et al. 1990). Therefore, in the absence of any other marine diatoms, it seems likely that the catastrophic wave(s) had a local source. Organic-rich silts beneath the lowest erosional contact pose something of a conundrum. They are relatively rich in brackish diatoms but are associated with the Kawakawa tephra (c. 22 ka BP). Therefore these sediments were laid
down when sea level was c. 120 m below the present level (e.g. Harris 1991). An estuarine interpretation would require uplift of about 120m over the last 22 000 years, which is contrary to the general tectonic history of the region (see above). Therefore, although the sediments were presumably fresh water, their brackish nature suggests a shallow ponded environment, with raised levels of salinity. The possible uptake of salts from the surrounding Hautotara Formation (see above) in conjunction with high winds blowing across off a lowered Cook Strait (c. 8-18 km offshore from the present sea level) and a narrow coastal platform may have generated conditions suitable for the survival of brackish diatoms (J. Begg, pers. comm., 1997). The introduction of windblown species is unlikely but cannot be discounted. A detailed analysis of at least the tectonic history and diatom assemblages of the region needs to be carried out before this conundrum can be fully resolved.
Pumice, shell, and wood material The geochemistry of glass shards from pumice samples indicates that all were from the Taupo eruption c. 1.85 kaBP. On the other hand, the geochemistry of glass shards from the tephra indicates that it is the Kawakawa tephra, dated at c. 22 ka BP. In most cases, pumice was situated in the upper part of fining-upwards sequences, and in all cases the clasts were well rounded. Similarly, most marine shells (and all intact ones) were deposited in the upper part of fining-upwards sequences. Several shells were preserved in life position, with dual valves, suggesting a lowenergy environment. All marine shells were identified as D. lambata, a sub-tidal species living in
370
J. R. GOFF E T AL.
10-30m of water. The bivalve is a suspension feeder, extracting food from organic detritus on the sea floor (Morton & Miller 1968; V. Anderlini, pers. comm., 1997). It is worth noting that before AD 1855 organic detritus on the sea floor would probably have been contaminated by older carbon and sediments introduced into the nearshore zone as a result of forest removal and land clearance carried out by European settlers (Peppercorne 1880; Pennington et al. 1976). Further more, before 1855 previous seismic activity would have had a similar effect. Although D. lambata shells would have the potential to date the catastrophic inundation event, they were not considered for radiocarbon dating because of the probable effects of contamination from old carbon (e.g. Goff 1997). Preferential preservation of shells and pumice suggests that they were rafted near the top of each wave, with passive deposition occurring during the waning energy regime of the backwash. This is consistent with the position of intact shells within a fining-upwards sequence, and the occurrence of shell fragments lower down in the sequence. It is assumed that there was an element of sediment reworking by each successive wave, in which case the incorporation of shell fragments lower down in the sequence is consistent with the preservation of intact ones higher up. Wood fragments and charcoal were rare, but were generally found near the upper part of fining-upwards sequences. Radiocarbon ages are not consistent with depth (i.e. older with depth), which is indicative of either reworking by catastrophic wave inundation or contamination by human disturbance of the area (e.g. Elliot et al. 1995; Goff 1997). Bearing in mind the Maori occupation of the valley in the recent past (Adkin 1959), it is surprising that no artefacts or reworked middens were evident in the exposed sections. Care was taken with the interpretation of shell fragments, for example, to ensure that they were not part of shell midden sites. However, although this source cannot be entirely discounted, it is argued that juvenile and articulated shells were probably introduced in association with some sea-floor sediments entrained by catastrophic waves.
Pollen
The introduction of exotic pollen species by European settlers (especially Pinus) has previously been used as a chronological marker (e.g. Dunbar et al. 1997). Pinus was first imported into Wellington in 1865, and by the mid-1880s it had been
widely distributed throughout the region (Shepherd & Cook 1988). The first appearance o f Pinus pollen in the sedimentary record is therefore thought to have occurred around 1890. In the three cases where the onset of Pinus pollen was recorded, the samples were taken across the uppermost gradational contact.
E v i d e n c e .for t s u n a m i inundation
If the sedimentary sequence discussed above was indeed laid down by catastrophic (tsunami?) waves, there should be a good correlation between the evidence discussed above and the diagnostic characteristics listed in Table 1. Table 1 lists a suite of key diagnostic characteristics for tsunamis reported by international researchers; the list is not exhaustive, and none of the individual tsunami reported show all of these characteristics. The evidence from Okourewa is recorded in the order Table 1. (1) Sequences fine upwards, and generally fine inland. (2) There are three distinct fining-upwards sequences; these are believed to be related to individual waves. (3) The sequences fine upwards, which suggests a waning energy environment associated with the run-up and backwash of a wave. Pebble fabric measurements in the upper and lower subunits of one fining-upwards sequence at site LF2 have been interpreted as representing deposition by run-up and backwash, respectively. (If this is the case, then most deposition seems to occur during the backwash.) (4) The lower contacts of all fining-upwards sequences are erosional. (5) One possible intraclast of reworked material was reported. (6) Loading structures were recorded at the base of LF2, and in the top sub-unit of the uppermost sequence in LF 1 (rapid dewatering?). (7) Particle and grain sizes vary from a maximum a-axis of 0.36m (LF4: large cobble) to a minimum of massive silt (LF1). Together, the three fining-upwards sequences can be recognized as an anomalous 'unit' between the organic-rich silts and sandy loam. (8) There are only minor changes between diatom assemblages above and below the lowest erosional contact at LF1 and LF2. However, this is consistent with some of the findings of Hemphill-Haley (1996). Frustules for nonmarine species were well preserved and marine ones were broken. (9) Foraminifera and other microfossils were not studied.
POSSIBLE TSUNAMI DEPOSITS (1855?), NEW ZEALAND (10) Geochemical analyses were not carried out. (11) Individual shells were found; some relatively 'shell-rich' sub-units were found but these appear to have been 'rafted' (point (13)). (12) No buried soil or vascular plant material was found; the lowest contact was erosional as opposed to unconformable. (13) One or more of shell, wood and pumice were found 'rafted' near the upper erosional or gradational contacts of the fining-upwards sequences. (14) Radiocarbon dates on wood fragments were equivocal (radiocarbon dating of shells was not carried out because of possible contamination). Dates from tephra and pollen 'bracket' the deposit. Alternative explanations for the origin of the deposits are not readily supportable by the evidence. The existence of pumice blocks so far from their volcanic source and the presence of marine bivalves within the sediment are good evidence of marine deposition. Roundness and sphericity data do not suggest fluvial deposition and show strong similarity to present-day beach deposits. The unusual nature of this deposit is the conjunction of large clasts with massive silt and intact juvenile marine shells, which suggests the juxtaposition of both active, high-energy and relatively passive, low-energy environments. This trend towards bimodality has been recorded in Hawaii (e.g. Moore & Moore 1988), and Australia (e.g. Bryant et al. 1992), and is considered to represent either individual waves of different energy or the run-up and backwash of individual waves. Similarly, the preservation of fragile marine shells is an indication of the complex flow regimes experienced when individual waves interact with the coastline (e.g. Bryant et al. 1992, 1996), although the sedimentological implications of these interactions are still poorly understood (Dawson 1994). We therefore attribute the series of finingupwards sequences to one tsunami inundation event composed of three waves. It is possible that the exposure represents three separate events, each successive event reworking some of the underlying material. However, tsunami inundations on exposed coastlines tend to destroy much of the evidence of prior events (e.g. Bryant et al. 1996) and records of multiple tsunami inundation are more likely to be preserved in sheltered coastal areas (e.g. Goff & Chagu~-Goff 1998). It is also possible that the deposits are evidence of a tectonically induced regression which abandoned beach deposits inland, but fining-upwards and fining-inland sequences are generally inconsistent with this suggestion. Unweathered,
371
abrupt erosional contacts are also indicative of rapid coastal processes, as opposed to gradual abandonment over a period of thousands of years. Furthermore, the nature of the diatom assemblage is indicative of deposition behind a fresh-water source. Abandoned beach deposits would be expected to have a far stronger marine diatom influence, and a greater number of fragmented frustules (marine and fresh-water).
Age o f tsunami inundation
The radiocarbon ages are equivocal but indicate that there has been substantial reworking of sediments. There are two known ages available at the location, the Kawakawa tephra (c. 22 ka BP) deposit found 0.20 m below the lowest erosional contact at site LF4, and the first appearance of Pinus pollen immediately above the uppermost fining-upwards sequence (at three sites). The unit in which Pinus pollen appears is thought to represent post-inundation conditions in Okourewa valley. Furthermore, the presence of'rafted' Taupo pumice (c. 1.85kaBP) and of a radiocarbon age of 13504-t90aBp suggests that the event occurred more recently than c. 1.5 ka BP. Goff & Chagu~-Goff (1998) reported at least three tsunami generated by fault ruptures in Cook Strait since c. 1.5 kaBP (Wellington fault: c. AD 1220 and c. AD 1440; West Wairarapa fault: AD 1855). Any one of these, or perhaps another fault rupture, could have been the source of the deposit in Okourewa valley. Reports indicate that the tsunami wave generated by the AD 1855 event was 'thirty to forty feet high in the straits, and Palliser Bay' (Fox 1855), and that 'immediately succeeding the first shock three large waves came up, the last higher than the first two' (Mason 1855). This contemporary evidence tends to support our interpretation of deposition by three waves. However, we suggest that the second wave was the largest and not the third as mentioned by Mason (1855). As New Zealand is in an active tectonic setting with much of the coastline subject to uplift, it is likely that a useful national and regional record of tsunami inundation can be acquired. Exposure to locally and distantly generated tsunami suggests that New Zealand may prove to be a natural laboratory for developing further tsunami signatures. Depending upon the nature of the coastline, 'exposed' or 'sheltered' environments are likely to yield different records, and therefore possibly different signatures.
372
J. R. G O F F ET AL.
Thanks are due to N. Preston, M. Sheppard, P. Hesp, Victoria University of Wellington (PHYG401 and 413) and Massey University (Geography) postgraduate students for field assistance and logistical support. We are grateful to M. Harper for identifying the first diatom assemblage from the site. This project was supported by a Victoria University of Wellington Internal Grants Committee award. References
ADKIN, G. L. 1959. The Great Harbour of Tara. Whitcombe and Tombs, New Zealand. ALBERTAO, G. A. & MARTINS, P. P., JR 1996. A possible tsunami deposit at the CretaceousTertiary boundary in Pernambuco, Northeastern Brazil. Sedimentary Geology, 104, 89-201. ATWATER, B. F. 1987. Evidence for great Holocene earthquakes along the outer coast of Washington State. Science, 236, 942-944. - - 1 9 9 2 . Geological evidence for earthquakes during the past 2000 years along the Copalis river, southern coastal Washington. Journal of Geophysical Research, 97 1901-1919. BARNETT, A. G., BEANLAND,S. & TAYLOR, R. G. 1991. Tsunami and Seiche Computation for Wellington Harbour. Report for the Museum Project Development Board. BARRETT, P. J. & BROOKER, M. R. 1989. Grain Size Analysis at Victoria University of Wellington. Laboratory manual, Victoria University of Wellington. BEGG, J. G. & MAZENGARB, C. (eds) 1996. The Geology of the Wellington area, Geological Map 22, scale 1.'50000. Institute of Geological & Nuclear Sciences, New Zealand. BERRYMAN, K. R. 1990. Late Quaternary movement on the Wellington Fault in the Upper Hutt area, New Zealand. New Zealand Journal of Geology and Geophysics, 33, 257-270. BEST, E. 1923. Miramar and its history: how Motukairangi was discovered and settled by Polynesians and how in times long past it became Miramar Peninsula. Transactions and Proceedings of the New Zealand Institute, 54, 779-791. BRADDOCK, R. D. 1969. Tsunami propagation over large distances. In: ADAMS, W. M. (ed.) Tsunamis in the Pacific Ocean. East-West Centre Press, Honolulu, 285-304. BRYANT, E. A., YOUNG, R. W. & PRICE, D. M. 1992. Evidence of tsunami sedimentation on the southeastern coast of Australia. Journal of Geology, 100, 753-765. --, -- & 1996. Tsunamis as a major control on coastal evolution, southeastern Australia. Journal of Coastal Research, 12, 831-840. CARTER, L., LEWIS, K. B. & DAVEY, F. 1988. Faults in Cook Strait and their bearing on the structure of central New Zealand. New Zealand Journal of Geology and Geophysics, 31, 431-446. CLAGUE, J. J. & BOBROWSKY,P. T. 1994. Evidence for a large earthquake and tsunami 100 400 years ago on western Vancouver Island, British Columbia. Quaternary Research, 41, 176-184.
COTTON, C. A. 1942. Shorelines of transverse deformation. Journal of Geomorphology, 3, 27-37. DARBY, D. J. & BEANLAND, S. 1992. Possible source models for the 1855 Wairarapa Earthquake, New Zealand. Journal of Geophysical Research, 97, 12 375-12 389. DAWSON, A. 1994. Geomorphological effects of tsunami run-up and backwash. Geomorphology, 10, 83-94. --, LONG, D. & SMITH, D. E. 1988. The Storegga slides: evidence from eastern Scotland for a possible tsunami. Marine Geology', 82, 271-276. --, SHI, S., DAWSON, S., TAKAHASHI, T. & SHUTO, N. 1996. Coastal sedimentation associated with the June 2nd and 3rd tsunami in Rajegwesi, Java. Quaternary' Science Reviews, 15, 901-912. DELANGE, W. P. & HEALY, T. R. 1986. New Zealand tsunamis 1840-1982. New Zealand Journal of Geology and Geophysics, 29, 115-134. DUNBAR, G. B., MCLEA, W. & GOFF, J. R. 1997. Holocene pollen stratigraphy and sedimentation, Wellington, New Zealand. New Zealand Journal of Geology' and Geophysics, 40, 325-333. EIBY, G. A. 1982. Two New Zealand tsunamis. Journal of the Royal Society of New Zealand, 12, 337-351. ELLIOT, M. B., STRIEWSKI, B., FLENLEY, J. R. & SUTTON, D. G. 1995. Palynological and sedimentotogical evidence for a radiocarbon chronology of environmental change and Polynesian deforestation from Lake Taumatawhana, Northland, New Zealand. Radiocarbon, 37, 899-916. FOLK, R. L. 1980. Petrology' of Sedimentary Rocks. Hemphill, Tulsa, OK. FOSTER, I. D. L., ALBON, A. J., BARDELL, K. M. et al. 1991. High energy coastal sedimentary deposits: an evaluation of depositional processes in southwest England. Earth Surface Processes and Landjbrms, 16, 341-356. Fox, W. 1855. Letter from W. Fox dated February 6 1855. In: GODLEY, J. R. (ed.) Papers, Typescripts, Vol. H. Canterbury Museum, 678-679. GHANI, M. A. 1974. Late Cenozoic vertical crustal movement in the central part of New Zealand. PhD thesis, Victoria University of Wellington. 1978. Late Cenozoic vertical crustal movements in the southern North Island, New Zealand. New Zealand Journal of Geology' and Geophysics, 21, 117-125. GILMOUR, A. E. 1960. Tsunami warning chart. New Zealand Journal of Geology and Geophysics, 4, 132-135. GOFF, J. R. 1997. A chronology of natural and anthropogenic influences on coastal sedimentation, New Zealand. Marine Geology, 138, 105-117. - & CHAGUI~-GOFF, C. 1998. A Late Holocene record of environmental changes from coastal wetlands: Abel Tasman National Park, New Zealand. Quaternary International, in press. HANSOM, J. D. & BRIGGS, D. J. 1991. Sea-level change in Vestfirdir, north west Iceland. In: MAIZELS,J. K. & CASELDINE, C. (eds) Environmental Change in Iceland. Past, Present and Future. Kluwer, Dordrecht, 79-91. HARRIS, T. F. W. 1991. Greater Cook Strait." Form and Flow. DSIR Publications, Wellington.
POSSIBLE T S U N A M I D E P O S I T S (1855?), N E W Z E A L A N D HEATH, R. A. 1976. The response of several New Zealand harbours to the 1960 Chilean tsunami. Tsunami Research Symposium 1974. Royal Society of New Zealand Bulletin, 15, 71-82. - - 1 9 7 7 . Circulation and Hydrology of Wellington Harbour. New Zealand Oceanographic Institute Summary 12. HEMPH1LL-HALEY, E. 1996. Diatoms as an aid in identifying late-Holocene tsunami deposits. Holocene, 6, 439-448. HICOCK, S. R., GOFF, J. R., ELAN,O. B. & LITTLE,E. C. 1996. Subglacial till fabric: on the interpretation of subglacial till fabric. Journal of Sedimentary Research (Series A), 66, 928-934. HULL, A. G. & MCSAVENEY, M. J. 1996. A 7000-year record of great earthquakes at Turakirae Head, Wellington, New Zealand. IGNS Client Report 1996/33493B.10. Prepared for the Earthquake Commission. HUNTLEY, D. J. & CLAGUE, J. J. 1996. Optical dating of tsunami-laid sands. Quaternary Research, 46, 127-140. IAN R. BROWNASSOCIATESLTD. 1995. An investigation of latest displacement events on the Wellington Fault in the Lower Hutt area. Earthquake Commission Research Project 93/128. IMAMURA, S., SUBANDONO,G., WATSON, G., MOORE, A., TAKAHASHI, T., MATSUTOMI, H. & HIDAYAT, R. 1997. Irian Jaya earthquake and tsunami cause serious damage. Los Transactions, American Geophysical Union, 78, 200-201. KING, L. C. 1930. Raised beaches and other features of the south-east coast of the North Island of New Zealand. Transactions of the New Zealand Institute, 61,498-523. - - 1 9 3 3 . Tertiary molluscan faunas from the southern Wairarapa. Transactions of the New Zealand Institute, 63, 334-354. MAIZELS, J. K. 1996. Sediments and landforms of modern proglacial terrestrial environments. In: MENZlES, J. (ed.) Modern Glacial Environments. Processes, Dynamics and Sediments. Glacial environments: Vol. 1. Butterworth-Heinemann, Oxford, 365-416. MARK, D. M. 1973. Analysis of axial orientation data, including till fabrics. Geological Society of America Bulletin, 84, 1369 1374. MASON, T. 1855. Letter from Thomas Mason to his uncle, Papers M S Paper 54.'2. ATE. National Library of New Zealand. MCKAY, A. 1878. Report on the East Wairarapa District. Reports of Geological Explorations, New Zealand Geological Survey, 11, 14-24. - - 1 8 7 9 . Report on the southern part of the East Wairarapa District. Report of Geological Explorations, New Zealand Geological Survey, 12, 75-85. MINOURA, K. & NAKAYA, S. 1991. Traces of tsunami preserved in intertidal lacustrine and marine deposits. Journal of Geology, 99, 265-287, , --, UCHIDA, M. 1994. Tsunami deposits in a lacustrine sequence of the Sanriku coast, Northeast Japan. Sedimentary Geology, 89, 25-31.
373
MITCHELL, J. S. & LEWIS, K. B. 1980. Cook Strait bathymetry, 2nd edn. New Zealand Oceanographic Institute chart coastal series 1:200 000. MOORE, G. W. & MOORE, J. G. 1988. Large-scale bed forms in boulder gravel produced by giant waves in Hawaii. Geological Society of America Special Paper, 229, 101-110. MOORE, J. G., BRYAN, W. B. & LUDWIG, K. R. 1994. Chaotic deposition by a giant wave, Molokai, Hawaii. Geological Society of America Bulletin, 106, 962-967. MOORE, P. D. & WEBB, J. A. 1978. An Illustrated Guide to Pollen Analysis. Hodder and Stoughton. London. MORTON, J. & MILLER, M. 1968. The New Zealand Sea Shore. Collins, London. OTA, Y., PIRAZZOLI,P. A., KAWANA, T. & MORIWAKI, H. 1985. Late Holocene coastal morphology and sea-level records on three small islands, the South Ryukyus, Japan. Geographical Review of Japan, 5 8 B , 185-194. PALMER, A. S. & VUCETICH, C. G. 1989. Last Glacial loess and early Last Glacial vegetation history of Wairarapa Valley, New Zealand. New Zealand Journal of Geology' and Geophysics, 32, 499-513. PATTERSON, T. R., GUILBAULT,J.-P. & CLAGUE, J. J. 1996. Marsh foraminiferal infaunal habit and taphonomic processes: significance for high-resolution sea-level studies. (abstract). In: MURRAYWALLACE, C. V., RoY, P. S. & LENHAM, K. H. (eds) Abstracts of 1996 Annual Meeting IGCP Project 367: Late Quaternary Coastal Records of Rapid Change. Application to Present and Future Conditions. Sydney, N.S.W., 55. PENNINGTON, W., CAMBRAY, R. S., EAKINS, J. D. & HARKNESS, D. D. 1976. Radionuclide dating of the recent sediments of Blelham Tarn. Freshwater Biology, 6, 317-331. PEPPERCORNE, F. 1880. Influence of forests on climate and rainfall, Transactions and Proceedings of the New Zealand Institute, 12, 24-32. PILLANS, B. J. 1986. A late Quaternary uplift map of North island, New Zealand. Royal Society of New Zealand Bulletin, 24, 409-417. ROBINSON, R. 1986. Seismicity, structure and tectonics of the Wellington region, New Zealand. Geophysical Journal of the Royal Astronomical Society, 87, 379-409. ROUND, F. E., CRAWFORD,R. M. & MANN, D. G. 1990. The Diatoms. Biology and Morphology of the Genera. Cambridge University Press, Cambridge. SHENNAN, I., LONG, A. J., RUTHERFORD, M. M. et al. 1996. Tidal marsh stratigraphy, sea-level change and large earthquakes, I: a 5000 year record in Washington, USA. Quaternary Science Reviews, 15, 1023-1059. SHEPHERD, W. & COOK, W. 1988. The botanic garden W'ellington." a New Zealand history 1840-1987. Millwood Press, New Zealand. STIRLING, M. 1992. Late Holocene beach ridges displaced by the Wellington Fault in the Lower Hutt area, New Zealand. New Zealand Journal of Geology and Geophysics, 35, 447-453.
374
J. R. G O F F E T AL.
STUIVER, M. & PEARSON, G. W. 1993. High-precision bidecadal calibration of the radiocarbon time scale, AD 1950-500BC and 2500-6000BC. Radiocarbon, 35, 1-23. & POLACH, H. A. 1977. Discussion: reporting of 14C data. Radiocarbon, 19, 355-363. VAN DlSSEN, R. J. & BERRYMAN, K. R. 1996. Surface rupture earthquakes over the last ~1000 years in the Wellington region, New Zealand, and implications for ground shaking hazard. Journal of Geophysical Research, 101, 599%6019. ---, , PETTINGA, J. R. & HILL, N. L. 1992. Paleoseismicity of the Wellington-Hutt Valley segment of the Wellington Fault, North Island, New Zealand. New Zealand Journal of Geology and Geophysics, 35, 165-176.
--,
LESTER, R. & BARNETT, A. G. 1994. Tsunami Hazard Study for the Hawke's Bay Region.
Hawke's Bay Regional Council. VICTORY, S. J., BUNDGAARD,H. I., BARNETT,A. G. & WARREN, I. R. 1989. Prediction of Seiching and Tsunamis' in Wellington Harbour. Barnett Consultants Ltd, prepared for Museum of New Zealand. YOUNG, R. W., BRYANT, E. A. & PRICE, D. M. 1996. Catastrophic wave (tsunami?) transport of boulders in southern New South Wales, Australia. Zeitschrft fu'r Geomorphologie, 40, 191-207. , W I R T H , L. M. & PEASE, M. 1993. Theoretical constraints and chronological evidence of Holocene coastal development in central and southern New South Wales, Australia. Geomorphology, 7, 317-329.
Index Note: tables and graphs are indicated in italics Agadir region 1731 and 1960 earthquakes 247-52 active folding, coastal morphology 241-4 Ait Lamine-Kasbah fault-related fold 247-52 Landsat image 241 uplift rates 245 implications 251-2 U-Th dating of fossils 244, 245 Ait Lamine-Kasbah fault-related fold 1960 earthquake 240, 247 Morocco 239-53 recent surface faulting and displacements 247-51 aminostratigraphy 57-8, 186, 260 Anadara trapezia, racemization of amino acids 260 Apennines, uplift patterns 71-110 Aquapora palmata, U-Th dating 21, 23-5, 244, 245 Arca sp. 79, 87 archaeological sites evidence for vertical movement 147-63 see also submarine human occupation sites (SILOS) Arctica islandica 116 Argentina, relative sea-level change 10 asthenosphere post-event stress relaxation 165 viscosity 174 Atlas Mts, Morocco, South Atlas Thrust Front, marine terraces 239-53 Australia Coorong to Mt Gambier coastal plain 255-67 geomorphological context 258 Holocene sedimentation 261-2 Quaternary volcanism 262-3 thermoluminesce 259-60, 263 volcanic edifices 265 Miocene Gambier Limestone 265 Portland Volcanic Province 262 relative sea-level change 10 radiocarbon dating 23 Australian plate, collision zone, E Indonesia, surface uplift 213-24 back-arc basins, Lau Basin 269 Banda Arc-Australian plate collision zone, E Indonesia Quaternary surface uplift 213-24 vertical strain conclusions 222-3 Barbados predicted eustatic curve 35 relative sea-level change 10, 22-5 beachrock facies, Brazil, NE 279-93 Bering Strait 12 Brazil, NE, Holocene beachrock facies 279-93 Barreiras Formation 285-6 chronology vs glacio-hydro-isostatic predictions 289-91 glacio-hydro-isostatic models 281-2 radiocarbon dating 287-9 British Isles, relative sea-level change 19-22 Bronze Age sites, evidence for vertical movement 147-63
Brunhes-Matuyama geomagnetic polarity reversal 255, 259 Calabria-Peloritan Arc, Pleistocene uplift 116 calcretes, Australia 255 California late Quaternary tectonism 179-97 deformation of 125 ka and later shorelines 184-91
Palos Verdes Peninsula 189-90 Peninsular Ranges 189-94 Southern Coast Ranges 186-7 Transverse Ranges 187-9, 191-3 Mesa Hills, Santa Barbara dating and correlation of marine terraces 57-61 oxygen-isotope correlation of marine terraces and uplift 57-69 data 62-4 Santa Barbara fold belt (SBFB) 58-61 uplift rates 65-8 Quaternary marine limit 182-4 San Clemente and san Nicolas Island terrace sequences 184 tectonic implications 191-4 carbon 14 see radiocarbon dating Cascadia Subduction Zone 199-211,321-42 active crustal structures and coastal subsidence 324-33 coastal deformation 200, 321-42 earthquakes 321 42 marsh burial locations 325, 333, 335-6 regional stratigraphy and structure 322-4 Cerastoderma sp. 75-89 Chandler wobble frequency 4 Cladocora coespitosa 77-81, 85-7 coastal indicators of sea-level change rock-cut installations 147-55 other indicators 155-6 Conus spp. 75, 83, 87 convergent-margin deformation, Washington, Olympic Coast 199-211 coral dating hermatypic corals 185 Iran 229-236 Quaternary surface uplift, Banda Arc-Australian plate collision zone, E Indonesia 213-24 solitary corals 65, 185 U-Th dated coral sequences 21, 23-5, 57, 65, 219, 220, 231, 244, 245 coral reefs dominant taxa 34 palaeo water depths 36-39 Quaternary uplift 31-9 see also Papua New Guinea, Huon Peninsula Cordilleran ice-sheet, Juan de Fuca lobe 208 Coseismic uplift 32 Crete, Falasarna harbour relative sea-level change and historical earthquakes 350 radiocarbon dating 351 tsunami signature 343-52 deglaciation, ICE-4G model 7-11
376
INDEX
earthquake-generated tsunami 343-52 Cook Strait, New Zealand (1855) 355-6 earthquakes Agadir region (1731, 1960) 247-52 Ait Lamine-Kasbah (1960) 240, 247 Cascadia Subduction Zone 321-42 contemporary accounts, Rann of Kachchh, India (1819) 295-320 Crete, Falasarna harbour (AD 65) 350 Greece, Gulf of Corinth, Helike Delta (373BC, 1861, 1995) 41-56 Rann of Kachchh, India (1819 and 1844-46) 295-320 subduction earthquake cycle 323 and tsunami signatures 343-52 Elephas sp. 81 ETOPOS model 5 Fennoscandian rebound 13-18 Fiji see Lau-Colville Ridge Florida, submarine human occupation sites (SHOS) 137 foraminiferal associations radiocarbon dating 48 tsunami signatures, Crete 346-50 fossils, U-Th dating 244, 245 France, submarine human occupation sites (SHOS) 135 Gibralter, submarine human occupation sites (SHOS) 135 glacial isostatic adjustment (GIA), global 1-29 relative sea-level change, model-data comparisons 18-27
theory of relative sea-level change 2-8 tuning model parameters 8-18 glacial rebound 5-8, 11-18 glacio-hydro-isostatic models, Brazil, beachrock facies 281-2 Global Positioning System (GPS) small-scale device 228 surveying, Iceland, vertical motion 165-78 Globorotalia inflata 116 Glycymeris sp. 75-89 Gorda plate 200, 322 Greece, Gulf of Corinth, Helike Delta 41, 42, 47, 49, 51 Holocene uplift and subsidence 41-56 borehole data 48 modelling absolute sea level and coastal uplift 43-5 shoreline data 47 stratigraphy 49-53 halokinetic deformation 233 Heaviside step function 5 Holocene Brazil, NE 279-93 Crete, Falasarna harbour 343-52 glacial isostatic adjustment (GIA), global 1-29 Greece, Gulf of Corinth, Helike Delta 41-56 Iran 225-37 Papua New Guinea, Huon Peninsula 33 Huon Peninsula see Papua New Guinea ICE-4G model of deglaciation 7-11
Iceland, vertical motion, GPS surveying 165-78 Askja magma chamber 171-2 Krafla spreading segment 171-4 Northern Volcanic Zone (NVZ) 166-72 Tj6rnes Fracture Zone (TFZ) 166, 173 Vatnaj6kull 168-170, 174 India, Rann of Kachchh 295-320 1819 and 1844-46 earthquakes contemporary accounts 296-305 appendix, correspondence 311-19 listric fault models 306-7 planar dislocations 305-6 secular strain contraction rate 309-10 Indonesia first seafaring 130 Quaternary surface uplift, Banda Arc-Australian plate collision zone 213-24 Semau and Rote island surveys (West Timor) 217-222 Indo-Australian Plate island-arc volcanism 270 surface uplift 213-24 interseismic deformation 209 Iran, South Coast marine terraces and uplift rates 225-37 Qeshm Island 226-9, 232-6 island-arc volcanism, Pacific and Indo-Australian Plates 270 isostatic adjustment, global glacial 1-29 Israel coast Acre peninsula 150-1, 158-60 rock-cut installations 147-55 other coastal indicators 155-6 Italy deformation of 125 ka marine terrace 71-110 distribution of uplift 99-101 uplift rates 100 Ionian coast 94-106 Palinuro, submarine human occupation sites (SHOS) 135 Southern Apennines 102-5 Southern Calabria, lineament-coplanarity analysis 111-127 Calabro-Peloritan Arc 112-17 fault scarps 116-17 fracture planes vs existing faults 123-6 marine terraces 116-17 synthesis of marine terrace data 114-15 Tyrrhenian coast 72-94 Japan, Tokyo Bay, relative sea-level change 10 Java Trench 214 Juan de Fuca plate 200, 322 Laurentide rebound 14 Lau-Colville Ridge, Pacific Ocean 269-78 geotectonic context 269 late Cenozoic evolution 270 model for pre-Holocene uplift 272-7 terraces, suggested chronology 274 lineament-coplanarity analysis 117-27 listric faults, models 306-8 Lithophaga lithophaga 46 lithosphere-asthenosphere, post-event stress relaxation 165
INDEX Makran geological province 226 marine terraces 199 marsh burial locations, Cascadia Subduction Zone 325, 333, 335-6 Mediterranean coast rock-cut installations Israel 147-54 other parts 154-5 submarine human occupation sites (SHOS) 129-46, 147-63 terrace data on RSL 114-15 Milankovitch hypothesis 257, 259 molluscs, oxygen-isotope correlations 57-69 Morocco, South Atlas Thrust Front, marine terraces 239-53 Neolithic sites, evidence for vertical movement 147-63 New Zealand, tsunami signatures 353-74 Cook Strait, 1855 earthquake-generated tsunami 355-6 Okourewa Valley 355-369 Wellington fault 353-6 West Wairarapa fault, 1855 earthquakes 354-5 notches, emerged 44, 272, 274 'ocean function', defined 2 Olivella biplicata, oxygen-isotope correlation of marine terraces and uplift 63 Olympic Coast, Washington, Pleistocene convergentmargin deformation 199-211 Oregon-Washington Coast, Pleistocene uplift 321-42 Ostraea sp. 77 oxygen-isotope stratigraphy 58 correlations marine terraces and uplift, Santa Barbara, California 57-69 thermoluminescence studies, Australia 259-60 Pacific Ocean, islands with relative sea-level history 24, 25 Pacific Plate Lau-Cotville Ridge, late Cenozoic evolution 270 tsunami sources 355 Palaeolithic sites, evidence for vertical movement 147-63 Papua New Guinea, Huon Peninsula 24-6, 31-9 phosphate, avian, chronology 272 Pinus pollen 371 Pleistocene deformation, Cascadia Subduction Zone 321-42 Pleistocene uplift Apennines 71-110 California, Mesa Hills, Santa Barbara 57-69 convergent-margin deformation 199-211 glacial isostatic adjustment (GIA) 1-29 Washington, Olympic Coast 199-211 Washington-Oregon Coast 321-42 pollen, radiocarbon dating analysis 370, 371 Quaternary coastal deformation, Cascadia Subduction Zone 200, 321-42 Quaternary surface uplift Australia, SE, Coorong to Mt. Gambier coastal plain 255-67
377
Banda Arc-Australian plate collision zone, E Indonesia 213-24 California 179-97 Iran 225-37 Mediterranean coast, terrace data on RSL 114-15 Papua New Guinea, Huon Peninsula 31-9 Washington, Olympic Coast 199-211 Quaternary volcanism, Australia, SE 262-3 racemization, amino acids 260 radiocarbon dating Brazil, NE, Holocene beachrock facies 287-9 calibration, Southern Hemisphere 357 glacial rebound 18-19 Greece, Gulf of Corinth, Helike Delta 46-53 hard water effect 46, 50 Hydrobia acuta, AMS 347, 349 limitations 58 Qeshm, Iran 230 relative sea-level change, Crete, Falasarna 351 reservoir effect, Greece 45-6 shells, in tsunami signature analyses 369-70 wood and charcoal 50-1,262-3, 369-70 Rann of Kachchh, India, contemporary accounts of earthquakes 295-320 relative sea-level change based on U-Th dated coral sequences 21, 23-5, 219, 244, 245 Brazil, NE, Holocene beachrock facies 279-93 British Isles 19-22 Papua New Guinea, Huon Peninsula 24-6 predictions, VM2 and ICE-4G models 9 submarine human occupation sites (SHOS) 139-41 VM1 and VM2 vicosity models 20 rock-cut coastal installations Iran 234 Israel 147-63 salt, dome, halokinetic processes 233 sea-level markers rock-cut installations 149 see also relative sea-level change shell middens 130 shells, in tsunami signature analyses 369-70 Sicily, deformation of the 125 ka marine terrace, Tyrrhenian coast 98-9 slip parameters, Rann of Kachchh, India (1819 and 1844-46) 295-320 Strombus bubonius 71, 75-89 submarine human occupation sites (SHOS) 129-46, 147-63 age distribution 135-7 depth distribution 137-8 Florida, Warm Mineral Springs 137 France, Fermanville, nr Cherbourg 135 geographical distribution 138-9 Gibraltar, Gorham's Cave 135 Israeli coast 147-63 Italy, Palinuro 135 number and rate of discovery 130-4 relative sea-level change derivation 139-41 taphonomy 134-5 see also rock-cut coastal installations
378
INDEX
thermal infrared multispectral scanning (TIMS) 32 thermoluminescence dating 259-60, 263 Timor Trough, location maps 214, 216 Timor, West 217-222 Tonga-Kermadec ridge 270 tsunami signatures 370-2 Crete, Falasarna harbour 343-52 Helike, Greece 41 New Zealand 353-74 U-Th dating, coral sequences 21, 23-5, 219, 244, 245 correlation methods 57-8 principles 220 uplift rates 205-207 viscosity, asthenosphere 174 viscosity models VM1, VM2, VM3 6-27
volcanic provinces Australia 262, 265 Iceland 166-72 Washington, Olympic Coast convergent-margin deformation 199-211 Kalaloch syncline 199-200, 204-7 Pleistocene strata deformation, Whale Creek 202-3 Washington-Oregon Coast, Pleistocene uplift 321-42 wave-cut surfaces 200-201 X-ray diffraction analysis 63-4, 218-19 Yellowstone hotspot, compared with Iceland 177 Zagros geological province 226
Coastal Tectonics edited by
I. Stewart (Brunel University, UK) and C. Vita-Finzi (University College, London) This volume concerns the application of high-resolution coastal records to developing and testing tectonics models. The case studies are at scales ranging from global to regional and local, and deal with glacio-isostasy, relative sea-level changes and seismic and aseismic crustal deformation at a variety of time scales. Data are drawn from a wide variety of disciplines, including geology, seismology, geodesy, geomorphology, geochronology and archaeology. The contributions within the book are from leading practitioners of field, laboratory and theoretical aspects of coastal tectonics. They provide a multidisciplinary, international review at a time when satellite monitoring, advanced modelling and sophisticated dating are transforming the subject from a descriptive historical branch of Earth science into a predictive tool addressing real societal needs. The book will be of interest to a wide audience including academics with an interest in tectonics and/or Quaternary science and professionals involved in seismic hazard or coastal development/protection.
•
394 pages
•
21 chapters
•
188 illustrations
•
index
ll!Hl!!!!i ISBN 1 - 8 6 2 3 9 0 - 2 4 - X
