Late Paleozoic Glacial Events and Postglacial Transgressions in Gondwana (GSA Special Paper 468)

Special Paper468

B -

THE GEOLOGICAL SOCIETY OF AMERICA®

Late Paleozoic GRaciall Event§ a n dl Po§tgRadiaR Tiran§gire§§fton §

lin G ondlwan a

India~

Africa

Antarctica

Late Paleozoic Glacial Events and Postglacial Transgressions in Gondwana

edited by Oscar R. López-Gamundí Hess Corporation One Allen Center 500 Dallas Street Houston, Texas 77002 USA and Luis A. Buatois Department of Geological Sciences University of Saskatchewan 114 Science Place Saskatoon, SK S7N 5E2 Canada

Special Paper 468 3300 Penrose Place, P.O. Box 9140

Boulder, Colorado 80301-9140, USA

2010

Copyright © 2010, The Geological Society of America (GSA), Inc. All rights reserved. GSA grants permission to individual scientists to make unlimited photocopies of one or more items from this volume for noncommercial purposes advancing science or education, including classroom use. For permission to make photocopies of any item in this volume for other noncommercial, nonprofit purposes, contact The Geological Society of America. Written permission is required from GSA for all other forms of capture or reproduction of any item in the volume including, but not limited to, all types of electronic or digital scanning or other digital or manual transformation of articles or any portion thereof, such as abstracts, into computer-readable and/or transmittable form for personal or corporate use, either noncommercial or commercial, for-profit or otherwise. Send permission requests to GSA Copyright Permissions, 3300 Penrose Place, P.O. Box 9140, Boulder, Colorado 80301-9140, USA. GSA provides this and other forums for the presentation of diverse opinions and positions by scientists worldwide, regardless of their race, citizenship, gender, religion, or political viewpoint. Opinions presented in this publication do not reflect official positions of the Society. Copyright is not claimed on any material prepared wholly by government employees within the scope of their employment. Published by The Geological Society of America, Inc. 3300 Penrose Place, P.O. Box 9140, Boulder, Colorado 80301-9140, USA www.geosociety.org Printed in U.S.A. GSA Books Science Editors: Marion E. Bickford and Donald I. Siegel Library of Congress Cataloging-in-Publication Data Late Paleozoic glacial events and postglacial transgressions in Gondwana / edited by Oscar R. LópezGamundí and Luis A. Buatois. p. cm. — (Special paper ; 468) Includes bibliographical references. ISBN 978-0-8137-2468-3 (pbk.) 1. Geology, Stratigraphic—Paleozoic. 2. Glacial landforms—Gondwana (Continent) 3. Periglacial processes—Gondwana (Continent) 4. Gondwana (Continent) I. López-Gamundí, Oscar R. II. Buatois, Luis A. QE654.L375 2010 551.7′2—dc22 2010019884 Cover, front: Map showing the distribution of glacial basins in Gondwana. From Isbell, J.L., “Environmental and paleogeographic implications of glaciotectonic deformation of glaciomarine deposits within Permian strata of the Metschel Tillite, southern Victoria Land, Antarctica” (Chapter 3, this volume). Back: Postglacial transgressive scenarios. (Upper left) Fjord environment influenced by extreme freshwater discharge from retreating glaciers. A freshwater ichnofauna occurs in the transgressive deposits (Guandacol Formation). (Lower right) Coastal environment without direct influence of freshwater influx from melting ice masses. A freshwater ichnofauna is present in coastal lakes and temporally inundated flood plains (Trace-Fossil Assemblage 1). A brackish-water ichnofauna (Trace-Fossil Assemblages 2 and 3) occurs in distal-bay facies (Tupe Formation). From Desjardins, P.R., Buatois, L.A., Mángano, M.G., and Limarino, C.O., “Ichnology of the latest Carboniferous–earliest Permian transgression in the Paganzo Basin of western Argentina: The interplay of ecology, sea-level rise, and paleogeography during postglacial times in Gondwana” (Chapter 8, this volume).

10 9 8 7 6 5 4 3 2 1

Contents

Introduction: Late Paleozoic glacial events and postglacial transgressions in Gondwana . . . . . . . . . . v Oscar R. López-Gamundí and Luis A. Buatois 1. Transgressions related to the demise of the Late Paleozoic Ice Age: Their sequence stratigraphic context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Oscar R. López-Gamundí 2. From bergs to ergs: The late Paleozoic Gondwanan glaciation and its aftermath in Saudi Arabia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 John Melvin, Ronald A. Sprague, and Christian J. Heine 3. Environmental and paleogeographic implications of glaciotectonic deformation of glaciomarine deposits within Permian strata of the Metschel Tillite, southern Victoria Land, Antarctica . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 John L. Isbell 4. Formation of euxinic lakes during the deglaciation phase in the Early Permian of East Africa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Thomas Kreuser and Gebretinsae Woldu 5. Stratigraphic and paleofloristic record of the Lower Permian postglacial succession in the southern Brazilian Paraná Basin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Roberto Iannuzzi, Paulo A. Souza, and Michael Holz 6. “Levipustula Fauna” in central-western Argentina and its relationships with the Carboniferous glacial event in the southwestern Gondwanan margin . . . . . . . . . . . . . . . . . . . . 133 Gabriela A. Cisterna and Andrea F. Sterren 7. Ichnology of late Paleozoic postglacial transgressive deposits in Gondwana: Reconstructing salinity conditions in coastal ecosystems affected by strong meltwater discharge . . . . . . . . . . . . 149 Luis A. Buatois, Renata G. Netto, and M. Gabriela Mángano 8. Ichnology of the latest Carboniferous–earliest Permian transgression in the Paganzo Basin of western Argentina: The interplay of ecology, sea-level rise, and paleogeography during postglacial times in Gondwana . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 Patricio R. Desjardins, Luis A. Buatois, M. Gabriela Mángano, and Carlos O. Limarino 9. Reconstruction of a high-latitude, postglacial lake: Mackellar Formation (Permian), Transantarctic Mountains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 Molly F. Miller and John L. Isbell

iii

The Geological Society of America Special Paper 468 2010

Introduction: Late Paleozoic glacial events and postglacial transgressions in Gondwana Oscar R. López-Gamundí Hess Corporation, 500 Dallas Street, Houston, Texas 77002, USA Luis A. Buatois Department of Geological Sciences, University of Saskatchewan, 114 Science Place, Saskatoon, SK S7N 5E2, Canada

The stratigraphic record suggests that glaciations have occurred episodically at different time intervals in Earth’s history (Crowell, 1982, 1999). One of those glaciations affected the Gondwanan Supercontinent during the late Paleozoic and constituted the longest period of continuous glaciation in the Phanerozoic (Eyles, 1993). Carboniferous to Early Permian glaciogenic successions have been known on all the subcontinents of Gondwana, most notably South America, Africa, India, and Australia, and later work expanded to Antarctica and the Middle East (Fig. 1). This glacial age can be subdivided into three distinct episodes (López-Gamundí, 1997). Glacial episodes II and III occurred during the early Late Carboniferous and the Late Carboniferous–Early Permian, respectively. An earlier, shortlived glacial episode in the Late Devonian–earliest Carboniferous (glacial episode I) identified in central and northern South America (Fig. 1) extended even further the duration of this ice age (Veevers and Powell, 1987). In general, the locus of ice cover, and its stratigraphic record, progressively moved across Gondwana from South America to Australia (Crowell, 1999), tracking the transpolar trajectory across Gondwana. This polar wander across the Gondwanan Supercontinent controlled paleolatitudes and accounts for the diachroneity of glacial episodes I, II, and III; however, the exact timing of waxing and waning of ice centers during each glacial episode (particularly for the longest-lived episode III) seemed to have been influenced by basin dynamics, topographic barriers, glaciation styles, and other local factors. The recognition of ancient glacial deposits of similar late Paleozoic age in South Africa and South America (Du Toit, 1927) helped, in conjunction with other lines of evidence, to argue in favor of the principles of seafloor spreading and indirectly to build the theory of plate tectonics (Wegener, 1915). The pioneering work during the first half of the last century was solidi-

fied subsequently by the seminal work led by John Crowell and Lawrence Frakes (Frakes and Crowell, 1967, 1969; Frakes et al., 1969; Frakes and Crowell, 1970; Crowell and Frakes, 1971a, 1971b, 1972; Frakes et al., 1971; Crowell, 1978). With the additional help of later contributions, the body of evidence about the duration, areal extent, and influence of the Late Paleozoic Ice Age (LPIA) on the biota has significantly grown. However, uncertainty remains over the exact timing of onset and demise of each glacial episode of the LPIA, particularly when attempts are made to link these glacial episodes in Gondwana with cyclothems in the Northern Hemisphere (particularly the United States and Europe), following Wanless and Shepard’s (1936) hypothesis. The picture becomes particularly blurred if, as exemplified by Wright and Vanstone (2001) for the Viséan carbonate successions in the Northern Hemisphere (UK), glacioeustatic sea-level oscillations invoked to account for high-frequency cyclicity had an approximate 100 ka periodicity, which may correspond to Milankovitch eccentricity. Thus, far field studies can sometimes be based on shaky grounds, particularly owing to the difficulty of estimating the magnitude and hierarchy of the near field glacioeustatic fluctuations (Rygel et al., 2008) and the less than optimal chronostratigraphic resolution of the Gondwanan faunas and floras associated with the LPIA. Maximum expansion of Gondwanan continental ice sheets occurred during earliest Permian time (glacial episode IV) under paleoatmospheric CO2 levels as low as the present ones to values of up to 12 times higher by the late Early Permian (Montañez et al., 2007). Widespread Early Permian (Sakmarian) collapse of ice sheets coincided with the onset of rising atmospheric CO2 levels, after which time surface temperatures and atmospheric partial pressure of carbon dioxide (pCO2) rose. The more detailed and deeper our knowledge about the LPIA gets, the better positioned

López-Gamundí, O.R., and Buatois, L.A., 2010, Introduction: Late Paleozoic glacial events and postglacial transgressions in Gondwana, in López-Gamundí, O.R., and Buatois, L.A., eds., Late Paleozoic Glacial Events and Postglacial Transgressions in Gondwana: Geological Society of America Special Paper 468, p. v–viii, doi: 10.1130/2010.2468(00). For permission to copy, contact [email protected]. ©2010 The Geological Society of America. All rights reserved.

v

vi

López-Gamundí and Buatois

Glacial Episode III

Arabia Glacial Episode II

2

Glacial Episode I

A F R I C A

340 Ma Polar Path 360 Ma 340 Ma

INDIA

1 360 Ma

4 S O U T H

AUSTRALIA 320 Ma

5 A M E R I C A

7

A N T A R C T I C A

1 7

8

1

7

260 Ma 265 Ma

1

6

250 Ma

280 Ma

7 3

7

1

7

9 0

2000 km

Figure 1. Gondwana Supercontinent and the Late Paleozoic Ice Age (LPIA) basins with glacial evidence in their stratigraphic record highlighted. Glacial episodes after López-Gamundí (1997); polar path after Powell and Li (1994). Locations of contributions in this volume indicated by numbers corresponding to chapters in the volume.

we will be to understand the relationship between shifts in pCO2, temperature, and ice volume and greenhouse gas forcing of past and future climates. Additionally, a better understanding of the mechanisms of postglacial transgressions along basin margins will allow us to refine our search for fossil fuels through the identification of potential marine source rocks and coals. We hope this volume contributes to both ends. This volume’s contributions constitute a wide range of topics related to this extreme paleoclimatic episode in Earth’s history. Original presentations were part of the IGCP 471 project (“Evolution of the western Gondwana during the late Paleozoic: Tectonosedimentary record, paleoclimate and biological changes”)–sponsored session Late Paleozoic glacial events and postglacial transgressions in Gondwana during the 32nd International Geological Congress (Florence, August 2004). It was evident at that time that consensus had been reached on some basic problems about the LPIA, but some new challenges had emerged. Some of the unanswered questions that this volume attempts to address revolve around (1) relatively less known glacial deposits in some regions of Gondwana (Central Africa, Kreuser and Woldu, Chapter 4; Arabia, Melvin et al., Chapter 2); (2) the controversy between a single massive ice sheet versus numerous glacial centers and alpine glaciers (Isbell, Chapter 3);

(3) the chronostratigraphic resolution of paleofaunas (Cisterna and Sterren, Chapter 6) and paleofloras (Ianuzzi et al., Chapter 5) that coexisted with extreme glacial and relatively milder early postglacial conditions, and the presence of freshwater and brackish water ichnofaunas related to postglacial marine transgressions (Buatois et al., Chapter 7; Desjardins et al., Chapter 8); (4) the characterization of high-latitude, postglacial lakes (Miller and Isbell, Chapter 9); and (5) the search for a unifying sequencestratigraphic model for the glacial-postglacial transition (LópezGamundí, Chapter 1). The contributions included in this volume cover a broad geography across Gondwana, but they do not have the objective of giving a state-of-the-art review of the LPIA, a theme that has been periodically dealt with since Hambrey and Harland’s (1981) volume. A recent update can be found in Fielding et al. (2008). Rather, the present volume is focused on key specific topics related to the LPIA that, in our opinion, required further study. These topics deal with the two main episodes identified during the LPIA: the early Late Carboniferous (Namurian–Westphalian) glacial episode II, mostly confined to the Paleopacific margin of southern South America, and a much more widespread early Permian glacial episode III, which affected the rest of the supercontinent (López-Gamundí, 1997; Isbell et al., 2003). Instead of providing

Introduction brief summaries of specific areas, the authors were encouraged to expand their views, providing full documentation. The book consists of two main parts. The first half deals with sedimentologic, paleoenvironmental, and paleoclimatic aspects of the glacial event. The second half explores paleobiologic aspects of glacial and glacially influenced ecosystems. The first contribution, which is by López-Gamundí, focuses on the sequence stratigraphy of the late Paleozoic glacial event and the subsequent postglacial phase, setting the stage for the rest of the volume. He notes that, irrespective of the age, there is a common stratigraphic stacking pattern in each of the transgressive events. As a result of the combined effect of fast glacioeustatic sea-level rise and subsidence along basin margins, a drastic landward facies shift took place in the transition from glacially dominated to glacially influenced early postglacial environments. Available information allows recognition of two basic types of transgressive systems tracts (TSTs): (1) a complete TST, with a basal diamictite unit, followed by shale with ice-rafted debris (IRD) and IRD-free shales, culminating in a maximum flooding surface; and (2) a base-cut TST in which the TST is dominated by open-marine shales, generally devoid of IRD. The other three contributions are case studies based on specific areas of Gondwana but bearing implications at a more global scale. Melvin et al. (Chapter 2) provide a detailed characterization of the Upper Carboniferous–Lower Permian Unayzah Formation of subsurface Saudi Arabia. This unit (subdivided into four members) is particularly relevant, because it provides a full picture of the paleoclimatic evolution in this region of Gondwana, from glacial through postglacial to semiarid and arid conditions. The wide spectrum of facies documented includes tillite, reworked diamictite, glaciolacustrine fines and turbidites, fluvial deposits, paleosoils, and eolianites. These authors differentiate between climatically and tectonically controlled transgressions and provide correlations across the Arabian Peninsula. In Chapter 3, Isbell explores the paleoenvironmental and paleoclimatic implications of glaciomarine deposits in the Permian Metschel Tillite of the Transantarctic Mountains. In contrast to previous interpretations, he proposes that ice entered the area from at least two different ice centers on opposite sides of the basin. Abundant evidence of glaciotectonic structures is documented, including sheared diamictites and thrust sheets. The global significance of this study resides in that multiple glaciers contain less ice volume than a single massive ice sheet, impacting on global climate and eustatic sea level in a different way than would have a single massive ice sheet. Kreuser and Woldu (Chapter 4) provide a detailed characterization of Permian euxinic lake deposits preserved in the Idusi Formation of the Ruhuhu Basin in Tanzania. The succession reflects a transition from glacial to postglacial deposits, culminating in the development of distal alluvial fans during climatic amelioration. These extensive anaerobic stratified lakes provided appropriate conditions for deposition of black shales with abundant organic matter (up to 11% total organic carbon [TOC]). These authors underscore the importance of these deposits as source rocks and

vii

outline the regional extent of these anoxic lakes in various late Paleozoic basins of eastern and southern Africa. The nature of continental and shallow-marine ecosystems during glacial and postglacial times is still poorly understood. The last five contributions of this book focus on this topic, touching also on biostratigraphic implications. In Chapter 5, Iannuzzi et al. provide for the first time a sequence-stratigraphic and paleoenvironmental framework for palynozones and plant zones in the Rio Grande do Sul portion of the Paraná Basin. These authors find that the boundaries of palynozones lie near the maximum flooding surfaces. In addition, they note that the plant zones previously defined correspond to distinct ecofacies and are better regarded as ecozones rather than as biozones. Based on this analysis, a link is suggested between the increase in floral diversity of the Glossopteris-Rhodeopteridium Zone and the appearance of more complex coastal ecosystems as recorded in the Rio Bonito Formation. Cisterna and Sterren (Chapter 6) evaluate taphonomic and paleoecologic aspects of the Levipustula Fauna. This fauna is typical of lower Upper Carboniferous glacially related deposits in the Andean basins of Argentina. Based on studies in different stratigraphic units of the Calingasta-Uspallata Basin, these authors are able to distinguish two associations: intraglacial and postglacial. They also note that the postglacial association is more diverse and displays more abundance than the intraglacial association. This increase in diversity and abundance is explained as a result of less stressful conditions resulting from climatic amelioration. The last three contributions deal with ichnology. In particular, trace fossils are ideally suited for ecosystem studies because they provide in situ evidence of organism-substrate interactions. In Chapter 7, Buatois et al. summarize ichnologic data from eight different Gondwanan basins (Paganzo, San Rafael, Tarija, Paraná, Karoo, Falkland, Transantarctic, and Sydney) and note the presence of fresh-water ichnofaunas in direct association with glacially influenced coasts affected by strong discharges of meltwater as a recurrent theme. These authors suggest that freshwater conditions prevailed in coastal areas during most of the postglacial times because of a strong discharge of fresh water from melting of the ice masses. They conclude that the classic marine-nonmarine dichotomy used in ichnologic studies may be misleading in this type of setting. Desjardins et al. (Chapter 8) document the ichnofauna present in transgressive deposits of the uppermost Carboniferous and lowermost Permian Tupe Formation of the Paganzo Basin in western Argentina. These authors discuss ichnologic aspects of the transition from postglacial fluvial deposits to bay deposits formed during a rise in sea level. The recognized trace-fossil assemblages reflect the changing environmental conditions that result from a base-level rise. This study underscores the interplay of ecology, sea-level rise, and paleogeography as controlling factors for trace-fossil distribution. In Chapter 9, Miller and Isbell focus on the paleoecologic implications of an ichnofauna formed in a large and deep turbiditic Permian lake, recorded in the Mackellar Formation of the Transantarctic Mountains. The Mackellar ichnofauna is of low

viii

López-Gamundí and Buatois

diversity, and the degree of bioturbation is generally low, suggesting oxic conditions and restriction of the benthos to areas with low rates of sedimentation. They compare this Permian lake with modern Lake Agassiz, and conclude that in spite of its high paleolatitude (~80°S), the lake was dynamic and biologically active. ACKNOWLEDGMENTS Finally the editors want to thank the following colleagues for dedicating their time to reviewing this volume’s contributions: Lucia Angiolini (Università degli Studi di Milano, Italy), Christoph Breitkreuz (Institut für Geologie, Freiberg, Germany), Roberto d’Avila (Petrobras, Brazil), Jim Collinson (USA), Almerio Barros França (Petrobras, Brazil), Dirk Knaust (StatoilHydro, Norway), Ricardo Melchor (Universidad de La Pampa, Argentina), Marcello Guimarães Simões (Sao Paulo State University at Botucatu, Brazil), Lynn Soreghan (University of Oklahoma, USA), Luis A. Spalletti (Universidad de La Plata, Argentina), Antonio Rocha-Campos (University of Sao Paulo, Brazil), Alfred Uchman (Jagiellonian University, Poland), John Veevers (MacQuarie University, Australia), and Joonas Virtasalo (University of Turku, Finland). REFERENCES CITED Crowell, J.C., 1978, Gondwana glaciation, cyclothems, continental positioning and climate change: American Journal of Science, v. 278, p. 1345–1372. Crowell, J.C., 1982, Continental glaciation through geologic time, in Climate in Earth History: Studies in Geophysics, Washington, D.C., National Academy Press, p. 77–82. Crowell, J.C., 1999, Pre-Mesozoic Ice Ages: Their Bearing on Understanding the Climate System: Geological Society of America Memoir 192, 106 p. Crowell, J.C., and Frakes, L.A., 1971a, Late Palaeozoic glaciation of Australia: Journal of the Geological Society of Australia, v. 17, p. 115–155. Crowell, J.C., and Frakes, L.A., 1971b, Late Paleozoic glaciation: Part IV, Australia: Geological Society of America Bulletin, v. 82, p. 2515–2540, doi: 10.1130/0016-7606(1971)82[2515:LPGPIA]2.0.CO;2. Crowell, J.C., and Frakes, L.A., 1972, Late Paleozoic glaciation: Part V, Karoo Basin, South Africa: Geological Society of America Bulletin, v. 83, p. 2887–2912, doi: 10.1130/0016-7606(1972)83[2887:LPGPVK ]2.0.CO;2. Du Toit, A.L., 1927, A Geological Comparison of South America with South Africa: Washington, D.C., Carnegie Institute, 157 p. Eyles, N., 1993, Earth’s glacial records and its tectonic setting: Earth-Science Reviews, v. 35, p. 1–248, doi: 10.1016/0012-8252(93)90002-O. Fielding, C.R., Frank, T.D., and Isbell, J.L., eds., 2008, Resolving the Late Paleozoic Ice Age in Time and Space: Geological Society of America Special Paper 441, 354 p.

Frakes, L.A., and Crowell, J.C., 1967, Facies and paleogeography of late Paleozoic Lafonian diamictite, Falkland Islands: Geological Society of America Bulletin, v. 78, p. 37–58. Frakes, L.A., and Crowell, J.C., 1969, Late Paleozoic glaciation: I, South America: Geological Society of America Bulletin, v. 80, p. 1007–1042, doi: 10 .1130/0016-7606(1969)80[1007:LPGISA]2.0.CO;2. Frakes, L.A., and Crowell, J.C., 1970, Late Paleozoic glaciation: II, Africa exclusive of the Karroo basin: Geological Society of America Bulletin, v. 81, p. 2261–2286, doi: 10.1130/0016-7606(1970)81[2261:LPGIAE ]2.0.CO;2. Frakes, L.A., Amos, A.J., and Crowell, J.C., 1969, Origin and stratigraphy of Late Paleozoic diamictites in Argentina and Bolivia, in Amos, A.J., ed., Gondwana Stratigraphy, IUGS Symposium (Buenos Aires, 1967): Earth Sciences, v. 2, p. 821–843. Frakes, L.A., Matthews, J.L., and Crowell, J.C., 1971, Late Paleozoic glaciation: Part III: Antarctica: Geological Society of America Bulletin, v. 82, p. 1581–1604, doi: 10.1130/0016-7606(1971)82[1581:LPGPIA ]2.0.CO;2. Hambrey, M., and Harland, W., 1981, Earth’s Pre-Pleistocene Glacial Record: Cambridge, UK, Cambridge University Press, 1044 p. Isbell, J.L., Miller, M.L., Wolfe, K.L., and Lenaker, P.A., 2003, Timing of late Paleozoic glaciation in Gondwana: Was glaciation responsible for the development of northern hemisphere cyclothems?, in Chan, M.A., and Archer, A.W., eds., Extreme Depositional Environments: Mega End Members in Geologic Time: Geological Society of America Special Paper 370, p. 5–24. López-Gamundí, O.R., 1997, Glacial-postglacial transition in the late Paleozoic basins of southern South America, in Martini, I.P., ed., Late Glacial and Postglacial Environmental Changes, Quaternary, Carboniferous–Permian and Proterozoic: Oxford, UK, Oxford University Press, p. 147–168. Montañez, I., Tabor, N.J., Niemeier, D., DiMichele, W.A., Frank, T.D., Fielding, C.R., Isbell, J.L., Birgenheier, L.P., and Rygel, M.C., 2007, CO2-forced climate and vegetation instability during late Paleozoic deglaciation: Science, v. 315, p. 87–91, doi: 10.1126/science.1134207. Powell, C.McA., and Li, Z.X., 1994, Reconstruction of the Panthalassan margin of Gondwanaland, in Veevers, J.J., and Powell, C.McA., eds., Permian– Triassic Pangean Basins and Foldbelts along the Panthalassan Margin of Gondwanaland: Geological Society of America Memoir 184, p. 5–9. Rygel, M.C., Fielding, C.R., Frank, T., and Birgeinheier, L.R., 2008, The magnitude of late Paleozoic glacioeustatic fluctuations: A synthesis: Journal of Sedimentary Research, v. 78, p. 500–511, doi: 10.2110/jsr.2008.058. Veevers, J.J., and Powell, C.M., 1987, Late Paleozoic glacial episodes in Gondwanaland reflected in transgressive-regressive depositional sequences in Euramerica: Geological Society of America Bulletin, v. 98, p. 475–487, doi: 10.1130/0016-7606(1987)982.0.CO;2. Wanless, H.R., and Shepard, F.P., 1936, Sea level and climatic changes related to late Paleozoic cycles: Geological Society of America Bulletin, v. 47, p. 1177–1206. Wegener, A., 1915, Die Entsehung der Kontinente und Ozeane: Braunchweig, Germany, Vieweg, 367 p. Wright, V.P., and Vanstone, S.D., 2001, Onset of Late Paleozoic glacio-eustasy and the evolving climates of low latitude areas: A synthesis of current understanding: Journal of the Geological Society [London], v. 158, p. 579–582.

MANUSCRIPT ACCEPTED BY THE SOCIETY 21 DECEMBER 2009

Printed in the USA

The Geological Society of America Special Paper 468 2010

Transgressions related to the demise of the Late Paleozoic Ice Age: Their sequence stratigraphic context Oscar R. López-Gamundí Hess Corporation, 500 Dallas Street, Houston, Texas 77002, USA ABSTRACT The Gondwanan Icehouse Period spanned between the mid-Carboniferous and Early Permian waning by the early Late Permian. Early postglacial sea-level rise related to the final stage of the Late Paleozoic Ice Age favored creation of accommodation space with preservation potential for productive anoxia events in the newly inundated shelves and peat-forming conditions favored by rapid water table rise in updip positions in the basin. The combined effect of fast glacioeustatic sea-level rise and subsidence along basin margins led to a drastic landward facies shift; the newly created space was sufficient to accommodate a transgressive systems tract (TST) that, irrespective of the age of the glacial episode, exhibits common characteristics across Gondwana. High fresh-water discharges related to the retreat of glaciers resulted in associated reduction in coastal salinity. Therefore, fjord-like settings as part of early postglacial inland seas seem to be a valid analogue for many of these TSTs. The examples of glacial-postglacial transitions analyzed in this contribution are present in a variety of basin types, namely, those ranging from backarc foreland basins to rifts. In all of them a clear retrogradational stacking pattern is detectable in the transition from glacially dominated settings to glacially influenced early postglacial environments. Examples from South America (Calingasta-Uspallata and Paganzo Basins), South Africa (Karoo Basin), Peninsular India (several Gondwana basins), and eastern Australia (Tasmania Basin) help define two basic types of TSTs: (1) complete TSTs, with a basal part of clast-poor, massive to poorly stratified diamictites, thinly bedded diamictites, shales with ice-rafted debris (IRD) and IRD-free shales, and an upper part dominated by open-marine shales representing the maximum flooding of the shelf; and (2) base-cut TSTs in which the basal transgressive portion is mostly omitted, and the TST is exclusively represented by open-marine shales generally devoid of IRD. Whereas the complete TSTs are common in cases in which high sediment supply rates via rain-out, ice rafting, and settling of fines prevail during the early phase of deglaciation, the base-cut TSTs, on the other end, reflect the dominance of drastic sealevel rises related to fast glacier retreats.

INTRODUCTION

the Earth (Fig. 1). Evidence of this glaciation is widespread in the Gondwana Supercontinent (Hambrey and Harland, 1981; Crowell, 1983, 1999; Eyles, 1993; Fielding et al., 2008a). This glacial age coincides with a high paleolatitude for Gondwana and the

The Late Paleozoic Ice Age (LPIA) is one of the best recorded episodes of extreme climatic conditions that affected

López-Gamundí, O.R., 2010, Transgressions related to the demise of the Late Paleozoic Ice Age: Their sequence stratigraphic context, in López-Gamundí, O.R., and Buatois, L.A., eds., Late Paleozoic Glacial Events and Postglacial Transgressions in Gondwana: Geological Society of America Special Paper 468, p. 1–35, doi: 10.1130/2010.2468(01). For permission to copy, contact [email protected]. ©2010 The Geological Society of America. All rights reserved.

1

2

López-Gamundí

Arabia

20 19

T e

A F R I C A 18

t h

INDIA 17

13

y

s

15 14

31 32

16

S O U T H 12

A M E R I C A

7

34 30

AUSTRALIA

8 9

2

29

A N T A R C T I C A 10

6

11 24

5

1

21 3

P a

n

23

h

a

l

a

s

s

27

28 26

22

4

t

33

a

25 0

2000 km

Figure 1. Reconstruction of Gondwana supercontinent with simplified outlines of principal basins during the Late Carboniferous–Early Permian (glacial episodes II and III). Positions of Arabia and Madagascar after De Wit et al. (1988); South American basins after López-Gamundí et al. (1994); African basins from Veevers et al. (1994) and Visser (1997a); Falkand-Malvinas Islands in their pre-breakup position east of the coast of South Africa as originally proposed by Adie (1952); Indian basins after Wopfner and Casshyap (1997); Australian basins modified from Struckmeyer and Totterdell (1992) and Lindsay (1997). 1—Calingasta-Uspallata and Paganzo; 2—Tarija; 3—San Rafael; 4—Tepuel-Genoa; 5—Sauce Grande; 6—Chaco-Paraná; 7—Paraná; 8—Huab; 9—Kalahari; 10—Karoo; 11—Falkland-Malvinas Islands; 12—Zambesi; 13—Congo; 14—Tanzania; 15—Malagasy; 16—Peninsular India; 17—Extra-Peninsular India (Himalayan); 18—Salt Range; 19—Yemen–South Arabia; 20—Oman; 21—Pensacola Mountains; 22—Transantarctic Mountains; 23—Tasmania; 24—Murray; 25—Sydney; 26—Bowen; 27—Galilee; 28—Cooper; 29—Pedirka-Arckaringa; 30—Perth; 31—Carnarvon; 32—Canning; 33—Browse; 34—Bonaparte.

growth of high standing topography when Gondwana collided with Laurasia to create Pangea (Eyles, 2008). The stratigraphic record derived from this glaciation is abundant and diverse: different types of glacioterrestrial and glaciomarine sediments associated with glacially abraded surfaces are present in a wide variety of basin types and tectonic settings. Despite this variability of facies, basin types, and tectonic regimes, most of these glaciated margins were affected during the early deglaciation phase by a sea-level rise that imparted a series of conspicuously diagnostic stacking patterns that can be interpreted in sequence stratigraphic terms. Examples of such glacial-postglacial transition have been documented not only for the LPIA but also for the Pleistocene glaciation. Owing to its relation with large hydrocarbon reservoirs, evidence of such transition from surface and subsurface studies has been abundantly documented for the Late Ordovician glaciation that affected most of the North African and Arabian platforms (Beuf et al., 1971; Deynoux et al., 1985; Vaslet, 1990;

Ghienne et al., 2007). The objective of this contribution is to illustrate the sequence stratigraphic context of the transgressions in relation to the demise of the LPIA. The emphasis of this study is on the section temporally constrained to the initial deglaciation when glacioisostatic rebound lags behind a rapid eustatic rise in sea level. CHRONOSTRATIGRAPHIC FRAMEWORK The record of the LPIA seems to cluster around three distinct episodes (Veevers and Powell, 1987; López-Gamundí, 1997; Isbell et al., 2003). The Late Devonian is the oldest of these episodes (episode I of López-Gamundí, 1997); it was originally described by Caputo (1985) and is apparently confined to a broad SW-NE belt across the central and northern parts of South America (Caputo el al., 2008; López-Gamundí and Buatois, this volume) from Peru (Carlotto et al., 2004; Cerpa et al., 2004) and

Transgressions related to the demise of the Late Paleozoic Ice Age Bolivia (Díaz Martínez and Isaacson, 1994) to Brazil (Solimões Basin, Eiras et al., 1994; Amazonas basin, Cunha et al., 1994; Parnaiba, Goes and Feijo, 1994) and recently extended to the Central African Republic and Niger in Africa (Isaacson et al., 2008). A late Fammenian age has been suggested for this glacial event on the basis of palynological evidence mostly from Bolivia (Isaacson et al., 1999). A northward extension of this short-lived, areally confined glaciation has been proposed for North America (Cecil et al., 2004; Isaacson and Díaz Martínez, 2005; Brezinski et al., 2008), but its evidence is tenuous so far and requires further work. The postglacial transgressive sections analyzed herein correspond to glacial episodes II and III as defined by López-Gamundí (1997). The former is associated with the mid-Carboniferous Levispustula fauna, and the latter is characterized by the Early Permian (Sakmarian) Eurydesma fauna. These two cold-water faunas have been identified in the marine fine-grained sediments that rest directly on, or in similar facies interbedded with, glaciogenic deposits across the Gondwana Supercontinent. The age of the Levipustula fauna has been traditionally considered Namurian–Westphalian (Roberts et al., 1976; see Cisterna and Sterren, this volume, for a recent review on the age of this fauna) but more recently has been confined to the early Namurian on the basis of absolute ages from SHRIMP (sensitive high-resolution ion microprobe) zircon-based dating in interbedded tuffs in Australia (Roberts et al., 1995; Fielding et al., 2008b). Glacial episode II is restricted to the westernmost part of the Gondwanan Supercontinent in southern South America (Amos and López-Gamundí, 1981a; González, 1990; López-Gamundí, 1984, 1989, 1997). Glacial episode III is widespread across the Gondwanan basins from South America (Frakes and Crowell, 1969; Rocha-Campos and dos Santos, 1981; López-Gamundí, 1997; Rocha-Campos et al., 2008) across Africa (Crowell and Frakes, 1972; Theron and Blignault, 1975; Visser, 1983, 1987a, 1987b, 1989, 1997a, 1997b; Wopfner and Kreuser, 1986; Von Brunn, 1994, 1996), Arabia (Helal, 1964; Kruck and Thiele, 1983; McClure, 1980; Braakman et al., 1982; Levell et al., 1988; Melvin and Sprague, 2006; Melvin et al., this volume) to India (Niyogi, 1961; Smith, 1963; Casshyap and Qidwai, 1974; Ghosh and Mitra, 1975; Casshyap and Srivastava, 1987), and Antarctica (Lindsay, 1970; Miller, 1989; Collinson et al., 1994; Isbell et al., 2008) to Australia (Crowell and Frakes, 1971; Struckmeyer and Totterdell, 1992; Lindsay, 1997). SHRIMP ages from the lower Dwyka Formation (Bangert et al., 1999) in the Karoo Basin (Fig. 1) and the Itararé Group (Rocha-Campos, 2006, in Rocha-Campos et al., 2008) in the Paraná Basin (Fig. 1) indicate a Late Carboniferous (Stephanian) age for the onset of glacial episode III in South Africa and South America. There seems to be consensus on a Sakmarian age for the widespread Early Permian collapse of ice sheets (Dickins, 1996; Isbell et al., 2003). Tuff zones at the base of the postglacial Prince Albert Formation in the southwest Karoo Basin provided SHRIMP ages of ca. 290 Ma (Bangert et al., 1999), a mid-Sakmarian age based on the numerical time scale of Gradstein et al. (2004). Exceptionally, local glacial conditions per-

3

sisted after the waning of the Gondwanan Ice Sheet Complex (Eyles, 1993) until the end of the Early Permian (Artinskian) along some margins and upland regions of western Gondwanan basins (northern margins of the Karoo Basin and the Kalahari Basin, Fig. 1) and particularly in eastern Gondwana, where small alpine ice caps provided glacial debris during three more minor glacial episodes in eastern Australia (Jones and Fielding, 2004). The proof of these waning glacial conditions elsewhere in Australia is mostly derived from the presence of ice-rafted debris (IRD) in the form of isolated dropstones in the Tasmania Basin (Fig. 1, Clarke and Banks, 1975) and the Sydney Basin (Fig. 1, Eyles et al., 1998). Indirect evidence for the presence of these waning, post-Sakmarian cold periods in eastern Australia is provided by studies that show lowered atmospheric pCO2 before the permanent transition to an ice-free Earth ca. 260 Ma (Montañez et al., 2007). The glacial-postglacial transition exhibits a similar retrogradational stacking of facies irrespective of their ages; therefore the analysis attempted in this contribution will be focused on the sedimentological characteristics of these transgressive deposits and their sequence stratigraphic context rather than on their chronostratigraphic significance. Analogies for vertical stacking patterns, for example, will be highlighted irrespective of the ages of the successions. Furthermore, probably the best documented pre-Pleistocene glacial-postglacial transition is the one that corresponds to the Late Ordovician (Hirnantian, late Ashgill) glaciation in North Africa and the Arabian Platform (Beuf et al., 1971; McClure, 1978; Hambrey, 1985; Vaslet, 1990; Le Heron et al., 2007; Ghienne et al., 2007), which shares sedimentological and sequence stratigraphic similarities with the late Paleozoic examples analyzed herein. Both glacial ages share a similar pattern of deglaciation characterized by a drastic landward facies shift during a fast transgression. However, unlike the short-lived Late Ordovician glaciation restricted between ~1 m.y. (Brenchley et al., 1994) and 10 m.y. (Ghienne, 2003), the LPIA is the longest lived glaciation in Earth history, perhaps one to two orders of magnitude longer than the Late Ordovician event (Eyles, 1993). Thus, eustatic fluctuations caused by numerous ice advances and retreats are commonly recorded across Gondwana for the late Paleozoic. The emphasis in this contribution is on the sedimentological and sequence stratigraphic characteristics of the wellpreserved glacial-postglacial transition rather than on the shorter lived, often cannibalized, glacial-interglacial fluctuations. Four examples have been selected; they cover a wide spatial and temporal range from west to east: Calingasta-Uspallata and Paganzo Basins (glacial episode II), Karoo and associated basins (glacial epsiode III), basins in Peninsular India (glacial episode III), and the Tasmania Basin in Australia (glacial episode III) (Fig. 1). SEQUENCE STRATIGRAPHIC CONTEXT FOR POSTGLACIAL TRANSGRESSIONS The term glacioeustasy is referred in this contribution to the process that generates changes in sea level that can be related to

4

López-Gamundí

changes in ice volume. A first approximation to the magnitude of eustatic sea-level rises related to the rather abrupt demise of any glaciation is provided by estimates of volumes of ice released. Table 1 is a compilation of the estimated present-day area and volume of glaciers and the maximum sea-level rise potential. It is worth noting that minor melting episodes related to small, local ice caps have negligible effects on sea-level fluctuations, as opposed to major waxing of large areas (i.e., East Antarctica). In the latter scenario, significant (~65 m) maximum sea-level rises are expected. This effect could have been attenuated if several small ice sheets, rather than a large single one, were present as suggested for at least part of the LPIA by Isbell et al. (2003). A similar approach has been used by Le Heron and Dowdeswell (2009) for their calculation of the postglacial sea-level rise related to the Late Ordovician glaciation. These authors estimated a postglacial eustatic sea-level rise of ~75 m on the basis of a small ice-sheet hypothesis and concluded that this approach adequately accounts for the estimated magnitude of postglacial transgression (~45–80 m) associated with ice-mass decay. Their estimate of 50 m of postglacial sea-level rise only from the contribution of an ice sheet in the North African Platform (the largest of the three ice sheets modeled) is in agreement with independent estimates derived from studies on depth-related benthic fossil communities (Ross and Ross, 1996). The effects of isostasy and its lateral variations along a continental margin (i.e., differential cross-shelf isostatic response of a thin crust) may potentially mask the otherwise dominant effect of glacioeustatically induced sea-level rises on the stratal architecture and sequence development of the early postglacial basin fill. This was frequent in tectonically active regions affected by glaciation, in which the individual factors related to the glaciation (eustasy and isostasy) and those of local origin (i.e., tectonically induced subsidence) are difficult to discern. Studies of Pleistocene glaciation along the western Canadian continental shelf show this extreme variability (Clague et al., 1982). Whereas a marine transgression owing primarily to eustatic rise occurred in the southern regions of the Strait of Georgia adjacent to the central and northern Strait of Georgia, it appears that 100 m of isostatic adjustment associated with local tectonic changes were

offset by eustatic effects resulting in minimal sea-level fluctuations (Barrie and Conway, 2002). Nevertheless, the Pleistocene and pre-Pleistocene glacial-postglacial stratigraphic record provides abundant examples of deepening-upward sections that indicate a drastic increase in accommodation space during the early stages of deglaciation. This record is confined to a portion of the basin where the postglacial sea-level rise and the isostatic subsidence caused by water loading might have exceeded the combined effect of glacial erosion and glacio-isostatic rebound (Bjorlykke, 1985; Nystuen, 1985), resulting in a subsequent glacioeustatic transgression (Crowell, 1978). Areas close to the ice margin are commonly dominated by isostatic depression and rebound, whereas more distal areas are dominated by eustatic changes like submergence during deglaciation (Miller, 1996). The main controls on stratal architecture in any given sequence can be narrowed down to (1) accommodation space (the summation of subsidence rates and sea-level fluctuations), and (2) sediment supply. Figure 2 illustrates some of the possible conditions under which accommodation space is created during a stage of sea-level rise in a subsiding basin (Jervey, 1988). Case A illustrates a classic example in sequence stratigraphy where the summation of a constant (in this case slow to moderate) subsidence rate and a symmetrical sinusoidal eustatic sea-level curve creates a relative sea-level curve (accommodation space). Cases B, C, and D illustrate three alternative outcomes of the most likely scenario for postglacial transgressions when the eustatic sea-level cycle shows a clear asymmetry from a faster (glacioeustatic) sea-level rise (expressed by the steep part of the eustatic sea-level curve). The subsidence rates vary and the eustatic sealevel cycle remains constant in these three last cases (Figs. 2B, 2C, and 2D). Case C differs from case B only in that case C includes a faster subsidence rate; in this situation the postglacial sea-level rise could be partially masked because very fast subsidence rates exceed the sea-level rate with accommodation space not significantly destroyed, and consequently, even when eustatic fall is taking place, deep-water conditions similar to those dominant during the transgressive stage may persist in the highsea-level stage. Finally, case D illustrates an extreme case with variable subsidence rates (from initially moderate to fast) and

TABLE 1. ESTIMATED PRESENT-DAY AREA AND VOLUME OF GLACIERS AND MAXIMUM SEA-LEVEL-RISE POTENTIAL Percent Volume Percent Maximum sea-level-rise Geographic region Area 2 3 (%) (km ) (%) potential (km ) (m) Ice caps, ice fields, valley glaciers, etc. 680,000 4.24 180,000 0.55 0.45 Greenland (inland ice) 1,736,095 10 .82 2,600,000 7 . 90 6. 5 0 Local ice caps and other glaciers 48,599 0.30 20,000 0.06 0.05 Antarctica 13,586,400 84.64 30,109,800 91.49 73.44 East Antarctica 10,153,170 26,039,200 64.80 West Antarctica 1 ,9 1 8 , 1 7 0 3,262,000 8.06 Antarctic Peninsula 4 46,690 227, 100 0. 46 Ross Ice Shelf 53 6 , 0 70 2 2 9, 6 0 0 0 .01 Ronne-Filchner ice shelves 532 ,20 0 351 ,90 0 0.11 Totals 16 , 0 5 1 , 0 9 4 100 .0 0 32,909,800 1 00 .0 0 8 0. 4 4 Notes: From Williams and Ferrigno (2008). Values in italicized rows are a subset of the “Antarctica” row values.

Transgressions related to the demise of the Late Paleozoic Ice Age

5

B RS

L

L ES

nce

side

Sub

Depth/Thickness

Depth/Thickness

A

RSL

Time

Depth/Thickness

RSL

ce

den

i ubs

ES L

Time

ES L

D Depth/Thickness

C

ce

iden

s Sub

RSL

S

ES L

S

Time

creation of deep-water conditions during transgression and even deeper water conditions during the high-sea-level stage. In cases C, and particularly D, subsidence rates are so high that the basin undergoes no decrease in accommodation even though eustatic fall may be occurring (Emery and Myers, 1996). Cases C and D correspond to the concept of forced transgression of Chough and Hwang (1997). The classic definition of sequence is adopted in this contribution. Sequences are composed of depositional packages or systems tracts (Brown and Fisher, 1977) deposited during specific phases of the relative sea-level cycle (Posamentier and Allen, 1999). A systems tract is a linkage of contemporaneous depositional systems and is defined by the nature of its boundaries (see below). Three principal tracts are used in this contribution: lowstand, transgressive, and highstand systems tracts (LST, TST, and HST, respectively). Boundaries between system tracts are defined by key sequence-stratigraphic surfaces: 1. Sequence boundary (SB): an unconformity and its correlative conformity corresponding to the most regressive configuration of stratigraphic architecture. A sequence boundary is expressed as a facies dislocation, a surface where rocks of a shallower facies rest directly on rocks of a significantly deeper facies (Emery and Myers, 1996). This facies dislocation (basinward facies shift) implies a fall in relative sea level. 2. Flooding surface (FS): a surface that separates abruptly deeper deposits from underlying shallower water sediments (Van Wagoner et al., 1990). This surface implies a rise in relative sea level and can be generated by eustasy,

e

nc

ide

s ub

Figure 2. Accommodation space as a function of subsidence rates and sealevel changes. Relative sea level (RSL) is equivalent in this case to accommodation because both curves begin at zero-depth water. Modified from Jervey (1988) and Emery and Myers (1996). (A) Accommodation space curve as the summation of a sinusoidal eustatic sealevel curve (ESL) and a slow to moderate subsidence rate. B, C, D: Resulting accommodation space from an asymmetric sinusoidal sea-level curve (ESL) and variable subsidence rates. (B) Slow to moderate subsidence (same as in A). (C) Moderate to high subsidence rate; note higher resulting accommodation space than in B; note that with faster subsidence, maximum accommodation is progressively later. (D) Increasing subsidence rate, which creates the most accommodation space. See discussion in text.

Time

tectonics, and/or delta-lobe switching; when recognizable globally, a flooding surface can be assigned to a eustatic sea-level rise. 3. Maximum flooding surface (MFS): the surface that corresponds to the most transgressive stratigraphic architecture. Maximum flooding surfaces can be identified by the presence of condensed sections that reflect distant sediment sources at the peak of transgression (Loutit et al., 1988; Posamentier and Allen, 1999). In shelfal and basinal settings, maximum flooding surfaces can be identified by evidence of condensation (i.e., firmgrounds), fine-grained (silt-clay) deposits, the presence of high organic content (expressed as high gamma-ray values in well logs), phosphate levels, fossiliferous beds, and outer shelf carbonates. In paralic successions the MFS is coeval with the most landward position of the shoreline (Emery and Myers, 1996). In areas where it is difficult to identify and trace laterally a maximum flooding interval (MFI) or zone is recognized for such a surface. The MFS should lie within the MFI, but no sedimentological evidence could be found to define such a surface. 4. Transgressive ravinement surface (TRS): the erosional surface cut by wave action during transgression. This transgressive surface is the first significant marine flooding surface across the shelf within a sequence (Van Wagoner et al., 1988). It defines the base of the TST. As pointed out by Posamentier and Allen (1999), some potential confusion arises when terms like transgressive surface and flooding surface are used because they are commonly

6

López-Gamundí used interchangeably. The former is reserved in this contribution for a surface “marking the onset of significant and extended period of transgression within a succession” (Posamentier and Allen, 1999, p. 5), restricting flooding only to the process of subsequent inundation.

FACIES ASSOCIATIONS RELATED TO POSTGLACIAL TRANSGRESSIONS With few exceptions in which terrestrial subglacial tills and associated facies are the dominant deposits, the stratigraphic record of the LPIA and its subsequent final deglaciation record across Gondwana have been preserved mostly in subaqueous settings. The main reason for the skewed record is possibly the significantly different preservation potential of terrestrial and subaqueous (marine or lacustrine) glacial environments. Preservation potential depends on a delicate balance of long- and shortterm components. The long-term component is largely controlled by the geotectonic setting (i.e., forearc basin, cratonic basin, etc.) and the related tectonic activity of the basin after deposition (i.e., postdepositional erosion); the short-term component is connected to the erosion coeval with sedimentation (Nystuen, 1985). This short-term component of the preservation potential is significantly low for terrestrial settings. Two main facies associations were connected with glacier retreat by López-Gamundí (1997) based on either nonmarine or marine conditions that prevailed in any specific part of a basin. Those are, respectively, a valley-glacier-retreat facies association and a submarine-retreat facies association. The Eyles et al. (1983) facies code for diamictite sequences is used in this contribution. Valley-Glacier-Retreat Facies Association This facies association developed along basin margins under predominantly subaerial or shallow subaqueous conditions during the retreat of a glacier front. Owing to the relatively low initial subsidence along the basin margins, the fill of these valleys is modest in thickness when compared with the submarine-retreat facies association developed basinward. Any sea-level rise has significant effects inland by raising the water table. This influence of a sea-level rise on mires was summarized by Bohacs and Sutter (1997), who estimated that a 5 m rise in sea level would propagate inland 50 km from the shore across a low-topography coastal area, producing a 3.5 m rise in the groundwater table. Terrigenous organic matter can be preserved to form coals only when and where the overall increase in accommodation approximately equals the production rate of peat (Bohacs and Sutter, 1997). For mires, base level is the groundwater table. Subsidence varies much more slowly than does the groundwater table and thus is commonly the major long-term control of accommodation in nonmarine settings. The postglacial facies in the valley-glacier-retreat setting rests on basement or, more commonly, on glaciogenic facies represented by thin (5–10 m), laterally discontinuous, massive

to crudely stratified diamictite (interpreted as subglacial till) beds partly associated with lacustrine shales with IRD. The glacial beds grade upward to fluvial deposits, coals, and carbonaceous shales, which represent the postglacial stage. The typical fluvial facies are predominantly coarse-grained sediments made up of pebble to cobble conglomerates, pebbly sandstones with trough cross-bedding, and medium- to fine-grained rippled sandstones deposited in a gravelly braided channel complex or as part of a subaqueous proximal outwash fan during a phase of glacier retreat. In coastal environments, marine influence can be expressed as marine fossiliferous shales, considered as maximum flooding surfaces, whereas the underlying coals and carbonaceous rocks are interpreted as initial flooding surfaces within a TST. In exclusively nonmarine settings the coals may represent the correlatives of maximum flooding surfaces developed basinward. Coals and carbonaceous shales present in this facies association have several characteristics that can be related to the accumulation of peat on initial rapid transgressions and increasing accommodation space. They are relatively thin and areally restricted as a result of increasing accommodation (cf. Bohacs and Sutter, 1997). Also, they are characterized by high sulfur (S) content since the original mires near the coast may have been submerged in brackish or salt water. In peat forming environments with brackish and marine influence (i.e., rheothropic mires), groundwater and seawater are important sources of primary S. In these brackish- to marine-influenced peats, the S content can be high (>10% in extreme cases), whereas in fresh-water peats tend to be low (1 m), laterally discontinuous beds of compacted massive diamictites (Dmm), interpreted as lodgment tills, rest on the striated surfaces. This basal facies is followed by tabular beds of massive diamictites (Dmm), locally bedded (Dms) and interpreted as subglacial ablation tills associated with cohesive debris flow beds represented by massive to bedded diamictites (Dms[r]) with soft sediment deformation. This facies association is grouped in an LST by Canuto et al. (2001). The LST is followed by sandstones, siltstones, and claystones with dropstones (Fld) of a TST, which, in turn, is overlain by deltaic sandstones stacked in a progradational pattern (highstand and regressive glacio-isostatic systems tracts). The uppermost of these sequences (S7) represents the period of final deglaciation in the Paraná Basin. This glacial-postglacial transition lies stratigraphically between the uppermost part of the Itararé Group (Rio do Sul Formation), containing the LST and TST, and the basal Triunfo Member of the superjacent Rio Bonito

15

Formation (HST). A similar sequence stratigraphic framework was proposed by Vesely and Assine (2004) in their study that incorporated both outcrop and well-log data. They identified five depositional sequences bounded by disconformities traceable over 400 km across the basin in an E-W depositional strike section. Three successive facies associations were defined within each sequence, but the lower and upper ones are absent in places. The lower facies association occurs only in the two lowermost sequences and is made up of subglacial facies (mostly Dmm); it corresponds to a glacial-maximum systems tract as defined by Vesely and Assine (2004). This tract grades upward into conglomerates and sandstones, which in turn are overlain by diamictites, turbidites, and shales with dropstones, labeled by the authors as a deglaciation systems tract (equivalent to the transgressive systems tract of Canuto et al., 2001). The deglaciation facies rests directly on the bounding erosional surfaces where the subglacial facies of the glacial maximum system is absent basinward. The fine-grained laminated facies of the upper part of the deglaciation systems tract represents the record of maximum glacial retreat during interglacial periods. Highest gamma-ray readings were used to determine the stratigraphic position of the MFSs within the deglaciation systems tracts. Falkland-Malvinas Islands Similarities between the late Paleozoic glacial-postglacial sections in the Karoo Basin and the Falkland Islands (Fig. 9) were first suggested by Adie (1952). Later paleomagnetic (Mitchell et al., 1986; Taylor and Shaw, 1989), radiometric (Mussett and Taylor, 1994), and paleocurrent (Storey et al., 1999; Trewin et al., 2002) data further confirmed Adie’s original correlation. The pre-breakup reconstructed position of the Falkland microplate east of the coast of South Africa (Figs. 1 and 6) is the result of an eastward translation and a 180° rotation from its present location (Marshall, 1994). The glaciogenic deposits are grouped under the Lafonia Formation (Frakes and Crowell, 1967; Scasso and Mendía, 1985; Bellosi and Jalfin, 1987), and the overlying postglacial sediments are part of the Port Sussex Formation (Frakes and Crowell, 1967). The contact between the uppermost Lafonia diamictites and the lowermost member of the Port Sussex Formation (Hells Kitchen Member) was characterized as an “abrupt undulating surface with a local relief of 0.25 m” by Trewin et al. (2002), who also note that the member thickness ranges from 3.5 to a maximum of 10 m and is made up of a coarsening-up cycle grading from “black fissile mudstone to fine- and medium-grained sandstones with granules and pebbles deposited as dropstones” (p. 9). Trewin et al. (2002) interpret the contact between the Lafonian diamictite and the base of the Hells Kitchen Member as an erosional event that probably developed under shallow-water conditions during a transgression related to deglaciation. This contact is considered herein as a transgressive ravinement surface (TRS) that initiated the postglacial TST. Trewin et al. (2002, p. 9) suggest that the Hells Kitchen Member sediments were deposited “probably during a transgression associated with deglaciation” and compared the section with a similar

16

López-Gamundí Si

SG Si

Bonete Fm.

SG

Ripon Si S

G

Fm.

Piedra Azul Fm.

Collingham Fm. Whitehill Fm.

Prince Albert Fm.

Sauce Grande Fm.

Dwyka Fm.

Highstand systems tract Transgressive systems tract

Port Sussex Fm.

Lowstand

Lafonia Fm.

systems tract

100 m 0 Black shales (Fl)

Massive diamictite (Dmm)

Shales (Fl) and fine-grained sandstones (Sr)

Coarse sandstone

Dropstone shale (Fld), thinly bedded diamictite (Dms) and shales (Fl)

Fine to medium sandstone

Stratified diamictite (Dms)

Basement

Figure 9. Correlation of the postglacial mudstones (glacial epsiode III) across the Ventana fold belt (V), Karoo Basin (K), and Falkland Islands (FI). Grain-size scale in logs: Si—silt; S—sand; G—gravel.

facies arrangement described by Visser (1993) for the base of the Prince Albert Formation in the Karoo Basin. The Hells Kitchen Member is overlain by interbedded black, pyritic, partly carbonaceous shales and siliceous mudstones (Black Rock Member, with a minimum thickness of 120 m); TOC content in the carbonaceous shales ranges ~2%–3% in the east to exceptionally up to 40% toward the west (Trewin et al., 2002). TOC versus S plots for two samples of the Black Rock Member indicate fresh-water salinities (Fig. 8B). Sauce Grande Basin–Ventana Fold Belt Keidel (1916) first recognized the similarities between the Cape fold belt and the adjacent Karoo Basin in South Africa, and the Ventana fold belt and contiguous Sauce Grande Basin in Argentina (Fig. 9). The similarities are particularly striking for late Paleozoic times, when the overall basin evolution and subsidence histories in both regions are considered (Du Toit, 1927; López-Gamundí and Rossello, 1998). Sedimentation in the Sauce Grande Basin adjacent to the Ventana fold belt began with

glaciogenic deposits by the Late Carboniferous (Keidel, 1916; Harrington, 1947; Coates, 1969; Amos and López-Gamundí, 1981b). This is approximately the same time interval when sedimentation, dominated by glacial deposits as well, began also in the Chaco-Paraná and Paraná Basins (Fig. 1). The basal fill of the Gondwana cycle in both basins is characterized by glacial deposits, mostly nonmarine in the Chaco-Paraná Basin (Russo et al., 1987; Fernández Garrasino, 1996; Winn and Steinmetz, 1998) and predominantly marine in the Paraná Basin (Rocha-Campos and dos Santos, 1981; Zalán et al., 1990; França and Potter, 1991; França, 1994; Rocha-Campos et al., 2008). Late Paleozoic glacial deposits exposed in the Ventana fold belt are grouped under the Sauce Grande Formation. Equivalent diamictites and overlying finer grained units (Piedra Azul and Bonete Formations) also have been reported from offshore wells (Lesta et al., 1980; Amos and López-Gamundí, 1981b; Fryklund et al., 1996; Juan et al., 1996). The age of the Sauce Grande Formation is poorly constrained as pre–Early Permian by the presence of Eurydesma fauna and Glossopteris flora in the Bonete

Transgressions related to the demise of the Late Paleozoic Ice Age Formation and post–mid-Carboniferous on the basis of palynological studies in equivalent sediments found in offshore wells (Archangelsky, 1996). The base of the Sauce Grande Formation lies with regional unconformity on Devonian metasedimentary basement. The unit is mostly made up of massive to crudely stratified diamictites with subordinate rhythmites, sandstones with ripples, and scarce conglomerates toward the top (Andreis et al., 1987). The massive (Dmm) and stratified (Dms) diamictites have been interpreted as glaciomarine (rain-out tills) partially remobilized downslope by gravity flows (Dms[r]) (Coates, 1969; Andreis, 1984; Andreis and Torres Ribeiro, 2003). The marine setting is confirmed by the presence of a solitary marine bivalve in the diamictites (Harrington, 1955). Detailed sedimentological studies by Andreis and Torres Ribeiro (2003) allowed subdivision of the Sauce Grande Formation in three megacycles. The lower megacycle (maximum thickness, 700 m) is composed of abundant diamictites, sandstones, and scarce conglomerates. The middle megacycle (~50 m thick) contains only sandstones and conglomerates. The upper cycle (~350 m thick) is made up of abundant fine- to coarse-grained sandstones; thick-bedded, massive, clast-poor, muddy diamictites (Dmm); thin-bedded stratified diamictites (Dms[r]); and shales, with scattered dropstones. The Dmm facies is interpreted as rainout tills, whereas the Dms(r) facies is considered glacial material remobilized downslope as gravity (mostly debris) flows. Andreis and Torres Ribeiro (2003) interpret the finer grained nature (evidenced by increasing participation of shales and sandstones to the detriment of diamictites, particularly of the clast-rich subtype) of the uppermost megacycle as a response to the transgression caused by glacier retreat that continued during Piedra Azul times. Dickins (1984) also relates the Piedra Azul transgression to the early Sakmarian glacioeustatic sea-level rise. The facies types and their vertical stacking in the Sauce Grande upper megacycle show similarities to the deglaciation sequences described by Visser (1997b) for the Karoo Basin (cf. Figs. 7 and 9). The Sauce Grande sediments grade upward to shales, bioturbated mudstones with gastropods, and subordinate fine-grained sandstones with wave and current ripples (Piedra Azul Formation), passing upward into bioturbated mudstones with fossils of the Eurydesma fauna (Harrington, 1955; Rocha-Campos and Carvalho, 1975; Amos, 1980), and fine-grained sandstones with wave ripples and cross-bedding (Andreis et al., 1979). Coarsening upward parasequences (up to 40 m thick each) are common in the Piedra Azul Formation; they consist, from bottom to top, of mudstones, fine-grained sandstones with wave ripples, and channelized medium- to coarse-grained sandstones with trough crossstratification (Andreis and Japas, 1996). Gondwana Basins of Peninsular India The Gondwana basins of Peninsular India (Fig. 1) have been considered fault bounded troughs developed along preexisting zones of weakness imparted by Precambrian structural fabrics (Naqvi et al., 1974). The basin fill is commonly asymmetric, with an overall increase in thickness toward one of the boundary faults,

17

a feature typical of extensional rift basins. A significant strikeslip component has been proposed for these basins (Chakraborty et al., 2003; Chakraborty and Ghosh, 2005). The Talchir Formation in the Indian peninsular basins (Fig. 10) shows evidence of sedimentation under glacial influence (Ghosh and Mitra, 1975; Casshyap and Srivastava, 1987; Veevers and Tewari, 1995). Key evidence of glacial influence during Talchir sedimentation is provided by striated pavements on basement rocks and boulder pavements, striated and bulletshaped boulders in diamictites, and dropstone shales (Banerjee, 1966; Ahmad, 1975) Marine faunas have been reported at or near the top of the glaciogenic Talchir Formation in several Gondwanan basins of India (see Veevers and Tewari, 1995, for a review). These findings may be related to a marine transgression, interpreted by several authors as having been caused by a eustatic rise in sea level owing to deglaciation. Previous interpretations, however, originally considered that the glacial sediments and finer grained postglacial sediments were deposited at the margin of lake basins, despite the rather gradational passage to marine deposits with abundant invertebrate fauna dominated by bivalves in the Rewa Basin (Ghosh, 1954), Daltonganj-Rajhara Basin (Dutt, 1965, in Goswami, 2008), and Bokaro Basin (Sengupta et al., 1999) (Fig. 10A). The most distinctive element of these marine faunas is Eurydesma, a cold-water bivalve genus widespread in the Early Permian of the Gondwana Supercontinent (Harrington, 1955; Dickins, 1961; Dickins and Shah, 1977; Runnegar, 1979). An inspection of the section described by Gosh (2003) in the Satpura Basin (Fig. 10B) suggests that the black shales and subordinate, decimeter-scale limestone beds that rest on the glaciogenic Talchir diamictites can be interpreted as part of the postglacial marine transgression rather than as the result of sedimentation of predominately fine-grained (clay and silt) material and carbonate muds in a basinal lacustrine environment with a sporadic marine incursion, containing a marine fauna characterized by Eurydesma and presence of other bivalves, ostracodes, and foraminifers. Marine acritachs have been identified in several basins and, in a few cases, associated with Eurydesma (Venkatachala and Tiwari, 1988). Associated plant remains also have been reported in several localities (Gosh, 2003). The postglacial section in the Satpura Basin is a marine, deep-water facies association; isolated ripple trains can be interpreted as starved ripples produced by weak bi- or unidirectional bottom currents. This finding extends also this marine embayment of the Indian Peninsula farther west from the areas where marine faunas were originally described (Rewa, Daltonganj-Rajhara, Bokaro Sub-basins of the Damodar Basin, Fig. 10A). The alternative interpretation should invoke a setting in which deep lacustrine facies are systematically interrupted by open-marine shales with Eurydesma fauna deposited below wave base without any intervening facies. Bose at al. (1992) studied the sedimentary features of the Talchir sandstone facies in the West Bokaro Sub-basin (Fig. 10A) and interpreted them as the product of coarse-clastic sedimentation “in fjord-like glacier-fed coastal troughs” (p. 95). Supporting evidence for a fjord-like setting with low salinity from a fresh-water

18

López-Gamundí Micritic limestones

78°E

Son

N

86°E Daltonganj-Rajhara

Rewa Basin

24°N

Satpura Basin

B

∗

Fine-grained sandstones

Raipur

Starved ripples Burrows

Paleocurrents Marine fauna (Eurydesma)

Talchir Basin

Irai

Trough cross-stratification

Massive diamictites (Dmm)

Raniganj

Mahanadi Basin Bedasar

20°N

Jahria

B

Rahmajal

Damodar Basin

Son Basin

∗

∗ ∗

∗

West Bokaro

Hummocky cross-stratification

Shales / mudstones

Barakar Fm.

Bay

TST

of

Godavari Basin

A

100 km

∗

Localities with marine fauna

20 m

Bengal

Talchir Fm. Si S

G

Si S

G

Figure 10. (A) Gondwana deposits in Peninsular India. Note presence of marine fossils (Eurydesma fauna) in postglacial mudstones (glacial episode III). (B) Section of the glacial-postglacial transition in the Satpura Basin. Modified from Ghosh (2003). Grain-size scale in logs: Si—silt; S—sand; G—gravel.

contribution from retreating glaciers comes also from ichnological studies. Bhattacharya and Bhattacharya (2007) studied the ichnofacies associated with these postglacial intervals in the Raniganj Sub-basin (Fig. 10A) and concluded, on the basis of low ichnodiversity, sporadic distribution of the traces, small burrow dimensions, absence of any body fossils, and dominance of worms and annelids as trace-makers, that stressed environmental conditions (cold climate and low marine salinity) prevailed. These conditions are assigned to the influx of glacier meltout fresh water in an ice-marginal sea during climatic amelioration and deglaciation. The postglacial TST exhibits a deepening-upward facies trend initiated with (1) mudstones with dropstones; (2) interbedded laminated to massive siltstones, mudstones, and fine-grained sandstones with current and combined-flow ripples and hummocky cross-stratification; and (3) overlying IRD-free, massive, locally bioturbated mudstones. Nereites and Zoophycos ichnoassemblages are common in the two latter facies (Bhattacharya and Bhattacharya, 2007). These facies with abundant trace fossils are conspicuously IRD-free and indicative of a retreat of the icegrounding line with a concomitant influx of glacier meltwater into the basin and subsequent lowering of the marine salinity. The facies of siltstones, mudstones, and fine-grained sandstones is interpreted as the product of sedimentation in an open shelf influenced by waves, whereas the overlying mudstone facies represents offshore background sedimentation from suspension of clay-silt material. It is here proposed that the early TST is represented by the facies of mudstones with dropstones; the MFI

in the late TST is represented by the mudstone facies. A similar interpretation has been proposed for the upper Talchir interval in the adjacent West Bokaro Sub-basin (Fig. 10A) by Bhattacharya et al. (2005). The basal Talchir glaciogenic section is made up of breccias, matrix- and clast-supported conglomerates, and coarse-grained sandstones (their conglomerate-sandstone, TCS, facies association) and is followed upward by fine-grained sandstones with hummocky cross-stratification, sandy siltstones with dropstones, and subordinate clast-supported conglomerates (sandstone-siltstone, TSS, facies association). This fining-upward, retrogradational section culminates with alternating thinly bedded fine-grained sandstones and mudstones with marine fossils (mollusks) and thick, multistoried, fine-grained sandstones with common hummocky and swaley cross-stratification (fine sandstone–mudstone, TSM, facies association). Bhattacharya et al. (2005) interpret this retrogradational stacking pattern as the result of progressive deglaciation during a eustatic sea-level rise and deposition of shelf sediments under a transgressive phase for the TSS and TSM facies associations. Geochemical evidence is also supportive of an environment with significant fresh-water contribution. Bhattacharya et al. (2002) studied the geochemistry of calcareous nodules in finegrained sediments (siltstones, rhythmites) that rest on the coarse glaciogenic basal part of the Talchir Formation in the Damodar, Mahanadi, and Godavari Basins (Fig. 10A). Four nodules of similar composition (micritic) from the uppermost Dwyka tillite of the Karroo Basin were also analyzed for comparative purposes.

Transgressions related to the demise of the Late Paleozoic Ice Age The oxygen and carbon isotopic ratio (δ18O and δ13C) values from both Talchir and Dwyka samples were similar and indicate a fresh-water environment of formation. Tasmania Basin During the late Paleozoic, Tasmania (Fig. 1) was positioned close to polar latitudes ~70° (Late Carboniferous) and ~80° (Early Permian) (Scotese and Langford, 1995; Li and Powell, 2001). The Tasmania Basin (Fig. 1) was glaciated throughout the Late Pennsylvanian (Clarke and Forsyth, 1989; Dickins, 1996). Ice-flow indicators suggest ice centers on the west and northwest and depocenters on the east. Local topographic highs created a fragmented shelf where glacially influenced sedimentation took place (Clarke and Forsyth, 1989; Hand, 1993). There is consensus (Banks, 1981; Clarke and Forsyth, 1989; Hand, 1993) that the glaciomarine diamictite and associated rhythmites of the Tasmania Basin were deposited in a fjord-like seaway characterized by rain-out and settling of fines with minor coarse debris deposited by rafting and turbidity currents originating from the glacier grounding line (cf. Bartek and Anderson, 1991). The glaciomarine sediments pass upward to marine siltstones and mudstones (Woody Inglis Siltstone, Quamby Mudstone, and equivalent units) with abundant glendonite concretions and elements of the Eurydesma fauna (Clarke and Banks, 1975). This interval has been assigned to a widespread marine transgression that covered most of Tasmania as glaciers retreated (Hand, 1993) during the early–middle Sakmarian (Brakel and Totterdell, 1993). These fine-grained deposits are overlain by the Bundella Formation (late Sakmarian), which includes the Darlington Limestone. Dropstones and glendonite concretions are abundant within the limestone (Rogala et al., 2007); a high-abundance, low-diversity Eurydesma fauna of calcareous invertebrates (mostly brachiopods, bryozoans, and Eurydesma bivalves) has been identified (Clarke and Forsyth, 1989; Rogala et al., 2007). The limestones consist of bioclastic floatstones, rudstones, and grainstones deposited in neritic shelfal environments during sea-level highstands (Rogala et al., 2007). Cold conditions persisted until the Late Permian, as evidenced by dropstones and cold-water Eurydesma fauna in Kungurian to Capitanian beds (Clarke and Forsyth, 1989). The glacial-postglacial transition in the Tasmania Basin has been studied in outcrops and in subsurface (Figs. 11–13). Its contact is in general sharp (mudstones resting on mostly massive diamictites), but in a few localities a distinctive facies between the glaciogenic diamictites and the postglacial mudstones has been identified. This facies consists of thin (0.5–1 m of bed thickness), stratified diamictites (Dms) with dropstone clusters and pebble nests (Domack et al., 1993). These thin diamictite beds have flat, nonerosive lower contacts, and alternate with the isolated pebbly mudstones and dark gray to black mudstones; eventually the mudstones predominate. The postglacial black mudstones contain scattered dropstones; framboidal pyritic and large (up to 10 cm) calcareous concretions (glendonites) also have been identified (Domack et al., 1993). Glendonites are calcite pseudomorphs after ikaite, a hydrated type of calcium car-

19

bonate, CaCO3 · 6H2O (Suess et al., 1982), and seem to occur in those mudstone zones rich in organic matter derived from the unicellular alga Tasmanites. This mineral is suggested to be an authigenic precipitate, forming at low temperatures from interstitial waters of organic-rich sediments, undergoing microbial degradation, accumulating rapidly in cold bottom waters (Suess et al., 1982; Shearman and Smith, 1985). High alkalinity from decomposing organic matter also enhanced precipitation of ikaite (Bischoff et al., 1993; DeLurio and Frakes, 1999). Glendonites in the Sydney Basin (Fig. 1) are also commonly associated with organic-rich shelf facies with dropstones (Thomas et al., 2005; Selleck et al., 2007). The layers rich in Tasmanites (named tasmanite beds) are laminated and occupy a consistent stratigraphic position ~10– 30 m above the contact between the mudstone-prone postglacial interval and the underlying diamictic section (Figs. 11 and 12, Domack et al., 1993). Pyrite nodules and dropstones are abundant in the tasmanite beds. TOC values for the tasmanite levels in two boreholes (Douglas River and Ross 1) reach exceptionally ~20%, with high hydrogen indexes (HI = 868 mg HC/g TOC); even higher HI values >900 mg HC/g TOC have been reported from other localities (Fig. 14, Revill et al., 1994). These high HI values indicate that the kerogen contains hydrogen-rich type I organic matter (Tissot and Welte, 1984). Sulfur values obtained rarely exceed 2%; when plotted against TOC content they fall on both fields (marine sediments and fresh-water sediments), indicating conditions of variable salinity (Fig. 8B). As indicated by Domack et al. (1993), the association of glendonite with the TOC-rich tasmanite levels suggests a marine origin and also low temperatures (close to freezing), a condition required for the formation of the precursor of glendonite (ikaite) (Shearman and Smith, 1985). Rao and Green (1982) estimated independently a sea surface temperature for Early Permian Tasmania of –1.8 °C, close to the present average near the Antarctic ice shelf of –1.9 °C. Cold water deposition is consistent with the low diversity of the associated Eurydesma fauna (Clarke and Banks, 1975). The organic-rich tasmanite beds in the Woody Island Formation and equivalent units represent condensed sections within an MFI; a similar view was proposed by Rogala et al. (2007). DISCUSSION Framework of the Postglacial TSTs Punctuation of long-term transgressions by repeated shortterm regressions is caused by the tendency of sediment supply rates to outmatch, for short periods, accommodation increase rates (Cattaneo and Steel, 2003). This is a particularly common pattern in most transgressions during periods devoid of glaciations. This punctuated shoreline movement has a complex zigzag shoreline trajectory (Helland-Hansen and Gjelberg, 1994) that results in a retrogradational (or backstepping) stacking pattern (retrogradational parasequence set, Van Wagoner et al., 1990). As part of the Pleistocene Ice Age, the late Quaternary transgressive

20

López-Gamundí

Trunbridge 0

G Si S

Ross-2 Si

Douglas River

Ross-1

SG

Si

SG

G

60

G 120

G

Depth (meters)

180

G

G

240

G 300

G 360

Ross-2 Tunbridge

420 Shales (black mudstones)

Diamictites

Silty mudstones

Basement

480

Douglas River

Ross-1

Eaglehawk Neck

Dropstones (IRD)

Tasmanite beds

G

Glendonites

100 km

Figure 11. Cross section showing stratigraphic position of tasmanite levels within the postglacial mudstone section in the Tasmania Basin. From Domack et al. (1993).

deposits may represent an unusual record of high-frequency and high-amplitude sea-level oscillations driven by glacioeustasy, and therefore better analogues for the sections previously discussed. Rather, the postglacial transgressions analyzed herein seem to be related to abrupt upward-deepening of facies and the scarcity of surfaces of ravinement (erosion by wave action), culminating in a level of deepest facies, commonly termed the maximum flooding interval (MFI) or surface (MFS). Direct juxtaposition of offshore mudstones over glacial deposits (most commonly rain-out tills or less abundant subglacial tills) is not unusual. In other cases a diagnostic intervening facies association dominated by thin-bedded debris flow deposits with IRD and dropstone shales is present. This facies association is the result of the complex interaction between glacier-derived sedimentation, gravity (debris flow) currents, and rain-out deposition. Although

variations within this theme are possible along a continental margin subjected to deglaciation, as sediment supply may be laterally variable in a basin, the rate of sea-level rise is significantly higher than any other process. In that sense the postglacial TST identified in the late Paleozoic Gondwanan basins can be equated to the late Quaternary transgressive deposits resulting from an unusual record of high-frequency, high-amplitude sea-level rise driven by glacioeustasy. The response to this type of postglacial retreat is the creation of a continuous transgression triggered by steady (and fast) sea-level rise (Curray, 1964; Cattaneo and Steel, 2003). Experimental sequence stratigraphy provides insights into some of the salient features of fast transgressions. Heller et al. (2001) showed that a rapid base-level-rise rate leads to an abrupt shoreline retreat with maximum flooding at the end of the maximum base-level rise. They also noticed that the magnitude of

Si S

G

TOTAL ORGANIC CARBON (%)

0

2

4

6

8

0

2

4

6

8

290 290

300

300

310

310

320

330

320

G 330

G

Shales (Fl) with Tasmanites

Mudstones with granules and pebbles

Mudstones

Stratified diamictites (Dms), thinly bedded

Dropstone shales (Fld)

Stratified diamictites (Dms)

Fine-grained sandstones

Basement

Glendonites

Figure 12. Douglas River borehole section (modified from Domack et al., 1993), with TOC (total organic carbon) values for the Tasmanites-rich interval. Grain-size scale in log: Si—silt; S—sand; G—gravel. See location in Figure 11.

22

López-Gamundí

A

C

transgression is far greater for the rapid base-level rise than during the slow rise, and that the time of maximum flooding is nearly synchronous with the end of the rapid base-level rise. In cases with slow rises, maximum flooding takes place far earlier than at the end of the rise.

B

Figure 13. Cores from the glacial-postglacial transition in the Tasmania Basin; see Figure 11 for location. (A) Massive diamictite (Dmm), probably of rain-out origin at base of Ross-1 well. (B) IRD-dominated section, Eaglehawk. (C) Stratified diamictites (Dms) with soft sediment deformation, Eaglehawk. (D) Organic-rich shale, Tasmanites-rich interval, Douglas River. Bar scale: 10 cm. IRD—ice-rafted debris.

D

Even in the cases where postglacial offshore mudstones are directly resting over the glacially derived, diamictitic section, the interval with the highest TOC values is invariably slightly higher in the interval dominated by offshore mudstones; thus most maximum flooding shales analyzed herein do not seem to be basal

Transgressions related to the demise of the Late Paleozoic Ice Age 1,000

23

I

900 Oonah Latrobe

800

Hydrogen Index

Douglas River

II

700 600 500

Figure 14. Geochemical characteristics of the tasmanite beds. From Domack et al. (1993).

400 300

100 km

200 100

Oonah

Latrobe

III

Douglas River

Upper seam Lower seam

0 380

400

420

440

460

480

500

T max °C

transgressive black shales sensu stricto (BT model of Wignall, 1991, 1994). The organic-rich, condensed sections in the Falkland Islands and Tasmania, a few meters above the contact with the diamictitic succession, are examples of this type of maximumflooding black shales. In other, leaner postglacial mudstone sections the maximum flooding interval is inferred by the abundance of marine fossils and dominance of fine-grained sediments (cf. most sections in the Calingasta-Uspallata and Paganzo Basins, the Piedra Azul interval in the Ventana fold belt, and the Lower Barakar Formation in several basins of Peninsular India). Postglacial TSTs through Space and Time As noted by Andrews (1997) the stratigraphic record may not be able to furnish the nuances of an unstable ice sheet at the time of its disintegration, especially without precise chronological control such as in the case of the LPIA. Thus, although appealing, it is somewhat incorrect to infer direct correlations between glacial histories and ice volume record. Despite this potential impossibility, the stratigraphic record of the late Paleozoic glacial-postglacial transition shows at hierarchies equivalent to third-order cycles a remarkable consistency in terms of facies associations, facies stacking patterns, and sequence-stratigraphic framework evolution. Other Occurrences of Postglacial TSTs Related to the LPIA The common characteristics of the glacial-postglacial transition highlighted in this contribution for the Calingasta-Uspallata and Paganzo Basins in western Argentina, the Karoo Basin in South Africa, the Gondwana basins of India, and the Tasmania

Basin in Australia seem to be present in other basins of the Gondwana Supercontinent affected by the LPIA. In Oman (Fig. 1) the glaciogenic sequence (Lower Permian Al Khlata Formation) is developed in a rift setting. It rests disconformably over Proterozoic basement and consists of diamictites, conglomerates, sandstones, and mudrocks ranging from glaciofluvial, glaciolacustrine, and alluvial to paralic environments (Levell et al., 1988; Al-Belushi et al., 1996; Martin et al., 2008). Evidence of glacial abrasion is provided in outcrops by striated surfaces on the dolomitic basement of late Proterozoic age (Braakman et al., 1982). A paleo– ice flow from northeast to southwest was inferred by Al-Belushi et al. (1996). Southward in the subsurface of the South Oman Salt Basin the Al Khlata diamictites were entirely deposited by rain-out and debris flow with no evidence for the preservation of true tillites (Aitken et al., 2004). The upper part of the Al Khlata Formation consists of a sheetlike diamictite abruptly overlain by the Sakmarian Rahab shale, which includes varve-like laminated mudrocks (rhythmites) with dropstones deposited in a large fresh-water to brackish-water body, according to Hughes Clarke (1988) and Levell et al. (1988). The dropstones progressively disappear upward, leading to Levell et al. (1988) to interpret this uppermost section of the Rahab shale as a deglaciation sequence. This unit is followed by the Saiwan Formation. In subsurface the Saiwan Formation unconformably overlies fine-grained sandstones and siltstones of the Rahab shale or rests directly on diamictites of the Al Khlata Formation. Its lowest part (lower member of the Gharif Formation) includes a restricted marine interval with the acritarchs, termed the maximum flooding shale by Guit et al. (1995) and interpreted as a possible postglacial eustatic flooding event (Stephenson and Osterloff, 2002). The base

24

López-Gamundí

of the Saiwan Formation, dated biostratigraphically as late Sakmarian (Angiolini et al., 2003), includes bioclastic sandstones with assemblages of filter-feeding brachiopods and bryozoans, which compare closely with the postglacial depositional sediments described from low-energy species of Antarctic shelves and indicate, according to Angiolini et al. (2003), the final stage of the Gondwanan deglaciation. TSTs Associated with the Late Ordovician Glaciation The proposed correspondence between lowstand incision of paleovalleys filled with glacial sediments during a time of ice buildup and subsequent glacioeustatically induced transgression immediately after glacier retreat is not exclusive to the LPIA. The Late Ordovician glaciation, present mainly in intracratonic basins of Mauritania, Mali, Morocco, Algeria, Libya, Tunisia, Jordan, and Saudi Arabia (Fig. 15), is particularly abundant in examples of these mechanisms, particularly for the postglacial transgression and its well-known product in North African and Arabian Platforms: the world-class source rock informally known as “the hot black shales” (Keeley and Massoud, 1998; Jones and Stump, 1999; Lüning et al., 2000; Carr, 2002). The hot shales are defined by an arbitrary cutoff (>200 API units) in the gamma-ray curves of well logs (Lüning et al., 2000). This value correlates approximately with TOC values of 3% for maturities around the oil window. Initial marine transgression is marked by an erosional ravinement surface locally overlain by thin, residual shallow-marine sands (Boote et al., 1998). Recent paleoglaciological reconstructions of the Late Ordovician Saharan ice sheet (Le Heron and Craig, 2008) suggest a stepwise, southward recession from the ice maximum, followed by a postglacial transgression. The hot shales (basal Tanezzuft Shale) in the North African Platform were deposited during the initial transgression in paleodepressions (formed by

0

Basins with glacial deposits

km 500

N

Shield areas

6

4

W es te rn

Sa ha ra

Atlantic

Jordan

Ghadames Basin

3 Mauritania Algeria Taoudeni Basin

Outcrops of glacial deposits

Mediterranean Sea

Tunisia

Morocco

Ocean

previous glacial outwash valleys and structural depressions) that enhanced restricted circulation; stratigraphically they correspond to the early TST (Lüning et al., 2000). The postglacial shale unit receives different lithostratigraphic denominations (Aïn Deliouine Formation in Morocco, Argiles à Graptolites in Algeria, basal Tanezuft Shale in Lybia, Qusaiba Member in Saudi Arabia, Sahmah Formation in Oman, Batra Mudstone and Mudawwara in Jordan, Abba Formation in Syria, Akkas in Iraq, Dadas Formation in southeastern Turkey, and Ghakum Formation in Iran), but it seems to correspond to a single, fast contemporaneous episode of postglacial inundation across much of the North African and Arabian Platforms. Detailed studies of the associated graptolite faunas indicate that the deposition of the organic-rich hot shales was a synchronous event across North Africa through Arabia of Llandovery (Rhuddanian) age during the Early Silurian (Lüning et al., 2000; Miller and Melvin, 2005). In Arabia (1 in Fig. 15) the hot shales are highly fossiliferous and pyritic. They correspond to the basal part of the Qusaiba Member (Qalibah Formation), the principal source rock for Paleozoic hydrocarbons in Saudi Arabia (Mahmoud et al., 1992), and were deposited as a condensed sequence on a sedimentstarved shelf; the high TOC values have been interpreted as the result of high productivity in high-latitude water masses (Jones and Stump, 1999). Sequence stratigraphic studies (Lüning et al., 2000; Dardour et al., 2004) frame the record of this glacial event and its associated subsequent transgression in the context of an LST of glacial sediments covered by transgressive marine shales (TST). Dardour et al. (2004) proposed a Late Ordovician–Silurian (second-order) super-sequence comprising a periglacial lowstand, the Tanezzuft transgressive to early highstand shales, and the Acacus Formation

Illizi Basin

Iraq

5 2

Murzuq Basin

Lybia

Saudi Arabia

Egypt

1 Mali

Oman

d

Chad

Re

Sudan Niger

UAE

Se a

Yemen

Figure 15. Map of North Africa and Middle East regions, with record of Late Ordovician glacial deposits and associated Early Silurian postglacial transgressive deposits. Based on compilation from several sources by Le Heron et al. (2009). Numbers refer to outcrop and subsurface records of postglacial transgressions discussed in the text.

Transgressions related to the demise of the Late Paleozoic Ice Age late HSTs. The lower boundary for this super-sequence is well preserved in the northern part of the Murzuk Basin (SW Lybia, 2 in Fig. 15) and in the contiguous Illizi Basin (SE Algeria, 3 in Fig. 15) as deep, incised valleys. This topography was gradually infilled by a heterogeneous glacial (mostly glaciofluvial) lowstand clastic facies that grades laterally into a more uniform, distal facies in the Ghadames Basin (eastern Algeria, southern Tunisia, and NW Libya) farther north (4 in Fig. 15, Dardour et al., 2004). Several higher frequency glacial sequences are recognized that may reflect several cycles of glacial advance and retreat. The North African Platform was subsequently flooded by a transgressive facies, the Tanezzuft Shale, which contains at its base a short-lived Rhuddanian anoxic event represented by thin but regionally extensive organic-rich hot shales (Lüning et al., 2000), interpreted here as the MFI of the sequence. The aggregate thickness of the hot shales rarely exceeds 20 m in a thin (~40 m), hot-shale–bearing basal section of the Tanezzuft Shale in the Ghadames and Illizi Basins (Lüning et al., 2000). The postglacial TST was followed by regressive highstand sedimentation commencing in the late Llandoverian; the HST continued to the end of the Silurian with sedimentation of a northerly prograding shelfal to fluviodeltaic wedge. Similar scenarios have been described elsewhere for the same Late Ordovician–earliest Silurian glacial-postglacial transition. Turner et al. (2005) described lowstand channel incisions and fill of those channels with glaciofluvial and shoreface sandstones in southern Jordan (5 in Fig. 15) related to a fall of relative sea level during ice buildup in southern Arabia. Each incision has been correlated with the first glacial advance during a stage of reduced accommodation space. This stage was followed by glacial melting and marine transgressive filling of the incised valleys as accommodation space increased. This process was repeated four times with glacier re-advances evidenced by glacially scoured, striated surfaces. The final transgressive filling (equivalent to the postglacial TST here) is characterized by the hot blackshale interval locally known as the lower Batra mudstone. The base of the black shale is coincident with the MFS (Armstrong et al., 2005). Recently Armstrong et al. (2009) concluded that the Batra Formation black shale was deposited in a short-lived, permanently stratified marine basin where an influx of fresh water from melting ice and nutrients resulted in enhanced photic-zone primary productivity and organic matter sedimentation. They proposed the stratified basins and fjords of east Antarctica as possible modern analogues. Similar valleys incised to depths exceeding 600 m (McClure, 1978; Vaslet, 1990), and subsequently filled by glacial diamictites and pro-glacial sandstones, have been traced into the subsurface of northern Saudi Arabia with seismic data (McGillivray and Husseini, 1992). After the retreat of glaciers the Arabian Platform was flooded by a rapid, glacioeustatic sea-level rise (Konert et al., 2001) that set anoxic conditions for the deposition and preservation of organic-rich hot shales (Qusaiba shale). The postglacial transgression (TST) was followed by a thick (>1000 m) coarsening-upward sequence of shales and sandstones (HST) of

25

Llandovery to Pridoli age that prograded basinward (Mahmoud et al., 1992; Konert et al., 2001). A similar sequence is described for the glacial-postglacial transition by Le Heron et al. (2007) in the Anti-Atlas of Morocco (6 in Fig. 15) where stratified, clast-poor, sandy diamictites, lying directly above a striated surface, pass vertically into transgressive tidal deposits (sigmoidally cross-bedded sandstones), which in turn pass upward into offshore mudstones (Aïn Deliouine Formation). Rare outsize boulders in clast-poor diamictites deform underlying laminae, indicative of dropstones from icebergs. The overall fining-upward section has been interpreted as having been deposited during ice sheet retreat in the earliest Silurian. Asymmetry in Sea-Level Rates: Its Impact on the Stratigraphic Record The asymmetry (fast rates of sea level rise, cf. Fig. 2) in the eustatic sea-level curve defines a clear stratigraphic signature, especially in cases where the sediment supply rate is relatively moderate to low with respect to the sea-level rise. The extent of erosion and creation of a TRS depends on the rate of rise of relative sea level. In cases such as the postglacial TST with a relatively rapid sea-level rise, erosion is minimized. Thus TRSs are infrequent, and transgressive sediments are preserved. The postglacial signature is characterized by retrogradational stacking patterns. Two subtypes of TSTs, which should be considered end members within a continuum for the submarine-retreat facies association (Fig. 16), are identified: 1. Complete TSTs: In cases where the sediment supply rate is fast enough to keep up with the sea-level rise rate, the lower part of the TST is dominated by rain-out processes, gravity flows (mostly proximal and distal debris flows), and background sedimentation of fines by settling from suspension. The resulting association, called here early TST, is made up of three main facies stacked in a retrogradational pattern: (i) clast-poor, massive to poorly stratified diamictites; (ii) thinly bedded diamictites; and (iii) shales with IRD. The upper part of this type of TSTs (complete TSTs) is made up of open-marine, IRD-free shales, unusually associated with fine-grained sandstones with bi- or unidirectional ripples indicative of tenuous wave reworking in shallow-marine environments. Clastic dilution owing to high sediment contributions would prevent accumulation of significant volumes of organic matter in the shales. 2. Base-cut TSTs: Where the basal transgressive portion is mostly omitted, the early TST is represented, from base to top, by distinctive distal debris-flow deposits (in some cases associated with low-density Bouma-type turbidites) and shales with IRD. The upper part of the TST is represented by basinal (below wave base) IRD-free shales. Alternatively, in extreme settings with a very high sea-level rise rate, the entire TST is made up exclusively of IRD-free shales (the basal shale model of Wignall,

26

López-Gamundí

Sediment supply rate

Sea-level rise rate

Si S

G

Si S

TST

G

Si

SG

TST

TST LST LST LST

Base-cut TST

Complete TST Black shales (Fl)

Thin-bedded diamictite (Dms)

Shales (Fl) and fine-grained sandstones (Sr)

Thick-bedded diamictite (Dms)

Dropstone shale (Fld)

Diamictite (Dmm) with boulder beds

Figure 16. Postglacial transgressive systems tract (TST) spectrum, based on relative influence of sediment supply and sea-level-rise rates. See text for further discussion. Grain-size scale in logs: Si—silt; S— sand; G—gravel. LST—lowstand systems tract.

1994). Wave reworking is unlikely in these fast transgressions. These base-cut TSTs are common in many glacial-postglacial transitions and basically reflect the drastic sea-level rise related to ice melting. Owing to a relatively low sediment supply rate with respect to the sea-level-rise rate, organic-rich shales could be deposited and preserved. Water Salinity during Deglaciation Widespread evidence indicates that fresh-water contributions into the inland seas developed in early postglacial times. This fresh-water incursion, as melting icebergs, is interpreted to have had the combined effect of creating a layer of brackish water and a relatively high suspended-sediment load, as described for fjords in Greenland by Syvitski et al. (1987, 1996). Three independent lines of evidence suggest drastic fluctuations of salinity from normal marine through brackish to fresh waters for these postglacial seas: (1) ichnofacies associations, (2) associated fauna, and (3) geochemical characteristics. An inland sea paleogeographic model has been proposed for the Transantarctic Basin (Barrett et al., 1986; Collinson and Miller, 1991; Miller and Collinson, 1994; Collinson et al., 1994). Early Permian palynoflora (fossil spores and pollen), recorrelated

with Australian palynomorph zones, is present in sediments resting on glaciogenic diamictites (Pagoda Formation). This palynoflora indicates that the glaciation in the Transantarctic Mountains was restricted to the Asselian–Sakmarian (Isbell et al., 2005). These deposits above the glaciogenic diamictites are grouped under the Mackellar Formation of the Central Transantarctic Mountains (Fig. 1). This unit is mostly made up of black, finely laminated shales and subordinate fine-grained sandstones with ripples, resting on the glaciogenic deposits of the Pagoda Formation (Lindsay, 1970; Miller, 1989); the upper Pagoda Formation is characterized by massive shales with IRD (Miller et al., 1991). C/S ratios in these sediments are extremely high, indicative of fresh to slightly brackish waters (Miller et al., 1991); TOC content is low (20 ft (6 m). Examples are illustrated in Figures 3A, 3B, and 4. The shear zones display a complex variety of features ranging from brittle shear to ductile shear to simple softsediment deformation. These zones are characteristically overlain and underlain by much thicker intervals of sandstone that are devoid of any such deformation (Figs. 3A and 4B). The shear zones are highly distinctive wherever they are seen in core, and all these zones recorded to date clearly correlate with intervals that display a spiky, high gamma profile on wireline logs (e.g., Fig. 4B). Commonly, they are also associated with a strongly chaotic signature on image logs (M.H. Prudden, 2005, personal commun.). This has enabled tentative identification of the shear zones in uncored wells: in such cases the chaotic zones are separated by considerable thicknesses (several tens of feet) of apparently normally stratified sediment, as has been observed in the cored examples. Origin and Evolution of the Unayzah C Member The character of the multistory bedsets of the Unayzah C member suggests that it was deposited in an environment that was dominated by processes of repeated erosion and deposition. The multiple scoured contacts imply a strongly channelized environment, and the grain size and primary sedimentary structures are suggestive of high energy conditions, as indeed are the thick beds with abundant dispersed mud clasts. The diffuse fabric seen in many of these sandstones is ascribed to dewatering, and the associated argillaceous crinkly laminations probably represent elutriated fines associated with that dewatering. The depositional system of the Unayzah C member was thus dominated by an abundance of channels characterized by rapid, high energy, sandy bedload sedimentation—i.e., an extensive alluvial braided plain. The thin siltstone beds represent rare low-stage suspension fallout deposits of fine-grained sediment upon the upper surfaces of the channel braided bars. As has been discussed above, limited biostratigraphic evidence suggests that the Unayzah C member is Late Carboniferous (Stephanian = Moscovian-Gzhelian) in age (Stephenson et al., 2003). This indicates that it was laid down relatively early following formation of the Hercynian unconformity, and, further, that deposition commenced during an early stage of the late Paleozoic Gondwanan glaciation. It is clear that these fluvial sediments have a widespread distribution throughout the subsurface across the study area (Fig. 5). Similar extensive fluvial sands and gravels are identified within postglacial outwash deposits associated with the several phases of retreat of the Pleistocene ice sheets in northern Europe. It is inferred that the Unayzah C coarse-grained

45

sandstones and conglomerates analogously represent extensive glaciofluvial outwash deposits associated with retreat phases of the late Paleozoic Gondwanan ice sheets in Saudi Arabia. The occurrence of an extensive line of end-moraine complexes known as the Rehburg Line (Fig. 6) that is associated with the north European Pleistocene glacial outwash deposits is of special significance with regard to the Unayzah C member. This line extends >500 km from the North Sea in the west to Hanover, Germany, in the east (Bennett, 2001) and marks the approximate extent of glacier ice during the Rehburg Phase (Drenthe advance) of the Saalian Glaciation (Van der Wateren, 1995). Several push moraine complexes were documented along the Rehburg Line (Van der Wateren, 1985, 1987, 1994), wherein their architecture consists of a number of subhorizontal nappes that have been displaced horizontally by the ice, in some places as much as 6 km. These nappes are bounded above and below by a number of shear zones (Bennett, 2001). The description of the Unayzah C member presented above has highlighted the recognition of well-developed zones that show a range of deformation styles from brittle to ductile shear and between which the rock is essentially undeformed. Furthermore, recent investigations by the senior author of the Carboniferous-Permian lower Juwayl Formation (Unayzah C equivalent) at the outcrop in the Wajid region of southwest Saudi Arabia also revealed the presence of discrete zones of low-angle dislocation (including shear and overfolding) within sandstones that are interpreted as glacial outwash within glacial paleovalleys. (That outcrop work will be presented in detail in a future publication.) All of this evidence, considered in conjunction with the earlier conclusion that the top of the Unayzah C member represents an unconformity of some magnitude, suggests the following model for the evolution of the Unayzah C member. Following the mid-Carboniferous tectonic event and the near-coincident inception of the late Paleozoic Gondwanan South Polar glaciation, the ice ultimately extended across the southern part of the Arabian plate to the vicinity of wells 13 and 14 in this study. That is to say, it spread to paleolatitudes at least as far north as the present-day location of the Al Batin Arch (see Fig. 1A) and therefore significantly farther north than had been previously documented (e.g., Stampfli and Borel, 2004). This glacial event was long-lived and possibly persisted for some 30–45 m.y., from the mid-Carboniferous (late Visean to early Namurian, or Serpukhovian) to Early Permian (Sakmarian) times (Al-Husseini, 2004). Considerable uncertainty exists, however, regarding the specific timing of the onset of this glaciation in Arabia, not least because of the possibility of a number of separate phases of glacial advance and retreat within its overall duration. Melvin and Sprague (2006) and Osterloff et al. (2004a) discussed the likelihood that older glaciogenic deposits may very well have been successively removed as a result of extensive cannibalization during later glacial advances in Saudi Arabia and Oman, respectively. In the Unayzah C member in the subsurface of eastern-central Saudi Arabia, each of the inferred glacial retreat phases resulted in substantial volumes of glaciofluvial outwash sands and gravels

Figure 5. Stratigraphic section across the study area, showing thickness variations and lateral extent of the Unayzah C member, in relation to the Unayzah B member and the unnamed middle Unayzah member (“UmUm”). Datum is the top of the unnamed middle Unayzah member. Lower bounding surface is the Hercynian unconformity (HU). Note the extent of downcutting into the Unayzah B member by the pre-Khuff unconformity (PKU) in wells 2 and 3. Map shows line of section.

Figure 6. Map of the northwest European Plain, showing the Rehburg Line of push moraines. These moraines are associated with phases of glacial advance of the Pleistocene Fenno-Scandian ice sheet. Modified from Bennett (2001).

Late Paleozoic Gondwanan glaciation in Saudi Arabia being deposited across the area upon laterally extensive alluvial braided plains. During subsequent readvances the ice sheets progressed across the glaciofluvial outwash of each preceding retreat phase. In the process they formed a number of glacially induced, push moraine nappes, similar to those documented by Van der Wateren (1985, 1987, 1994) along the Rehburg Line from the Pleistocene of the north European Plain (Fig. 6). Ultimately, this process created a thick pile of subhorizontal, superimposed units of glaciofluvial sands and gravels, separated by distinct shear zones (Melvin and Sprague, 2006). It is likely that these repeated processes of glacial thrusting associated with readvances of the ice sheet were also responsible for the remobilization of pore fluids within the outwash sands, thus giving rise to the widespread evidence of dewatering that is observed within these rocks. The top Unayzah C unconformity is considered to represent the subglacial surface at the time of the final advance of the late Paleozoic Gondwanan ice sheet in Saudi Arabia. UNAYZAH B MEMBER The preceding discussion has shown how the Unayzah C member is characterized across the study area by a limited number of depositional facies, namely multistory quartzose sandstones and rare conglomerates that were deposited upon an extensive glaciofluvial outwash braided plain. Consequently, that member displays a fairly distinctive, monotonous wireline log signature. This is in stark contrast to the rocks that make up the Unayzah B member. Melvin and Sprague (2006) described in detail >1355 ft (406.5 m) of core from the Unayzah B member in 13 wells, and identified 8 distinctive depositional facies from within that stratigraphic unit. The characteristics of those various facies are presented below. In well 7, continuous core was recovered, not only from the entire Unayzah B member as it is represented in that well but also from the complete section through the overlying “un-named middle Unayzah member” (Melvin and Sprague, 2006), as well as the uppermost parts of the underlying Unayzah C member. This cored interval thus represents a crucial record of the stratigraphic relationships in the Unayzah in that well, and is reproduced as a reference section in Figure 7. Sediment Fracture-Fill Features (?Periglacial Deformation) In well 4, recovered cores show that the Unayzah B member rests directly upon the Hercynian unconformity (i.e., the Unayzah C member is absent at that location). A laminated sandstone 3 ft (0.9 m) thick is present ~2 ft (0.6 m) above the unconformity. It displays a sharp, subvertical fracture in core that is filled with a variety of coarse-grained detritus including mud clasts and granule-sized quartz grains. This feature has been discussed and tentatively interpreted by Melvin and Sprague (2006) as a frost contraction wedge formed in a periglacial setting. Similar features have been documented in modern periglacial settings (French,

47

1996; Ruegg, 1983), as well as from Proterozoic glaciogenic sediments from Greenland by Moncrieff and Hambrey (1990). Stratabound, Internally Deformed Deposits (Push Moraines) This facies is well displayed in two zones in well 7 (Fig. 7, 26.5–42.0 ft and 55–68 ft). At each location the rock displays severe structural deformation in the form of high- to low-angle reverse faulting (listric thrusting) (Fig. 8A) and overfolding. The rocks comprise sandstones and silty mudrocks that contain dispersed pebbles and small cobbles of siltstone and very fine– grained sandstone (diamictite) (Fig. 8B). Palynological analyses of these fine-grained deposits have yielded a heavily reworked assemblage of palynomorphs of Silurian, Devonian, and Early Carboniferous age, as well as significant quantities of monosaccate pollen that cannot be older than Namurian (Serpukhovian) in age (the “Cm palynoflora assemblage”: J. Filatoff, 2004, written commun.; Fig. 7). These intervals of deformed rock are stratabound by undeformed sediments laid down in a conformable succession (Fig. 7) and are thus interpreted as having been disrupted at, or very shortly after, their time of deposition. Furthermore, it can be shown (see below) that they are interstratified with sediments that are demonstrably glaciogenic in nature. They have thus been interpreted as the product of glacial deformation and specifically identified as the preserved remnants of glacial push moraines (Melvin and Sprague, 2006). Bennett (2001) described how push moraines display a wide range of morphologies at a range of scales, from a few meters to features that extend for several kilometers. The overthrust outwash sands and gravels described earlier from the Unayzah C member fall into the latter category. The stratabound, internally deformed deposits in the Unayzah B member are much smaller features, with an average thickness of ~14 ft (4.2 m). Boulton et al. (1999) identified four broad categories of push moraines, wherein the style of deformation involves either fans of imbricate thrusts or superimposed subhorizontal overthrusted nappes. Regarding the former, the smallest are generally no more than 16.4 ft (5 m) high and can be found in both terrestrial and subaqueous environments (Boulton, 1986; Boulton et al., 1999; Bennett, 2001). These small push moraines are more representative of the stratabound, internally deformed deposits seen in the Unayzah B member in well 7. In the western part of the study area the Unayzah B member in wells 2 and 3 (see Fig. 1A) directly overlies rocks of early Paleozoic age, i.e., it rests directly upon the Hercynian unconformity. It is 40 ft (12 m) thick in well 2, and 65 ft (19.5 m) thick in well 3, and in each well it is characterized throughout by poorly sorted sediment and by an abundance of small-scale faults and low-angle shear planes (Melvin and Sprague, 2006). This pervasive intraformational deformation of the Unayzah B member in these wells is also attributed to direct sustained contact with ice.

Figure 7. Core log from well 7, showing the complete section through the unnamed middle Unayzah member (“UmUm”) and the Unayzah B member in that well, and their respective contacts with the Unayzah A member above and the Unayzah C member beneath. Note: Cm indicates the locations of the proven occurrence of the heavily reworked “Cm palynofloral assemblage” in the Unayzah B member (J. Filatoff, 2004, written commun.).

Late Paleozoic Gondwanan glaciation in Saudi Arabia

49

Figure 8. Core photographs from well 7, showing features observed in the glacially tectonized facies of the Unayzah B member. (A) Interval of sandstone exhibiting high-angle, reverse faulted (thrusted) dislocations (see Fig. 7, 58–61 ft). (B) Poorly sorted pebbly sandstone (diamictite), displaying a strongly sheared fabric (see Fig. 7, 48–51 ft).

A

B

Massive, Very Poorly Sorted Pebbly Siltstones (Diamictite) Diamictites are poorly sorted sediments containing a wide range of particle sizes in a relatively fine matrix (Bates and Jackson, 1980). The massive diamictites described herein are commonly observed within the Unayzah B member in the Saudi Arabian subsurface, being recorded in this study from wells 1, 2, 3, 4, 5, 7, 10, and 12. In some cases they are severely deformed (e.g., wells 2, 3, and 7: see previous discussion) as a result of being incorporated in push moraines. Where they have been identified in core they range in thickness from ~6 ft (1.8 m) in well 10 to 14 ft (4.2 m) in well 12, to >200 ft (60 m) in well 5 (Fig. 9). In general the diamictites display almost no stratal fabric and are extremely poorly sorted. Thus they comprise a

host rock of gray-green (and locally red-brown) siltstone to very fine–grained sandstone within which occurs an abundance of well-rounded to subangular, dispersed grains of medium to very coarse quartz sand as well as up to 5% coarser material including granules and pebbles of granite, quartz, feldspar, black chert, gray siltstone, sandstone, and mudstone (Fig. 10A). Cobblesized clasts of fine-grained sandstone have been observed in places (Fig. 10B). Several workers investigating diamictitic deposits in glacial settings have identified a number of subfacies, including lodgment tillites, proximal and distal water-lain tillites, and debris flows (e.g., Visser, 1982; Levell et al., 1988; Moncrieff and Hambrey, 1990). The latter authors conceded that “there are numerous cases in which the interpretation remains open to question.” We

Figure 9. Core logs from wells 10, 12, and 5, showing associations of glaciolacustrine depositional facies in the Unayzah B member. (A) Sediment gravity flows in well 10 (0–25 ft; note the dropstone at 17 ft), overlain by multiple laminasets of ripple cross-laminated sandstones (25–44 ft) and a thin diamictite (44–50 ft), followed by mudrock in the uppermost 5 ft. (B) Laminated mudrock (0–7.5 ft) in well 12, overlain by massive diamictite (7.5–20.0 ft). This is followed by stratified diamictite (20–28 ft) and sediment gravity flows (28–41 ft) in the upper part. (C) Extremely thick interval of massive diamictite in well 5 on top of a unit of sediment gravity flows (0–10 ft). Wireline logs in this well suggest a total >200 ft (60 m) of diamictite.

Late Paleozoic Gondwanan glaciation in Saudi Arabia

A

C

51

B

D

concur with Melvin and Sprague (2006), who discussed in detail the massive diamictites of the Unayzah B member and concluded that they most likely represent resedimented glacial debris laid down at the bottom of glacial lakes as debris flow deposits. Interstratified Pebbly Siltstones and Laminated Mudstones: Stratified Diamictite This distinctive facies within the Unayzah B member has been recorded from wells 4, 7 (Fig. 7, 68–85 ft), and 12. It is heterolithic in character and comprises a host rock of very fine– grained, laminated mudstone (with associated rare, very thin beds of rippled, very fine–grained sandstone in some places), within which occur thin beds (millimeter to decimeter scale) of poorly

Figure 10. Core photographs showing features associated with massive and stratified diamictites in the Unayzah B member. (A) Typical example of massive diamictite, showing the very poorly sorted nature of the sediment, and the apparently random orientation of clasts (well 12). (B) Cobble-sized clast of fine-grained sandstone in massive diamictite (well 5). (C) Stratified diamictite in well 7, showing finegrained glaciolacustrine laminite with a thin (2 cm) interval of (type 1) very poorly sorted diamictite. Note the extreme fissility in the lower part of the laminite subfacies. (D) Cobble-sized clast of diamictitic material (till pellet) in type 1 stratified diamictite. Note the irregular, diffuse edges to the till pellet (well 12).

sorted pebbly siltstone material (diamictite) (Fig. 10C). Two types of this diamictite are observed. Type 1 beds are 0.2–4.0 in. (0.5–10.0 cm) thick and have indistinct upper and lower boundaries. They consist of silty mudstone that contains abundant dispersed grains of fine- to very coarse–grained sand, as well as rare granules of quartz, feldspar, and granite. Significantly, they also commonly contain distinctive pelletoid clasts of material that is itself very poorly sorted (diamictitic) in nature. These pelletoid clasts range in size from a few millimeters to 15 cm (Fig. 10D), are commonly flattened parallel to bedding, and show diffuse or “ragged” edges. Type 2 beds are 2–12 in. (5–30 cm) thick, with sharp, commonly loaded lower bed boundaries, and are similarly very poorly sorted. The most extensive development of this depositional facies recorded to date occurs in well 7 (Fig. 7, 68–85 ft).

52

Melvin et al.

There, the type 1 beds become increasingly rarer toward the top of the 17 ft interval, where the laminated mudrocks predominate: this interval of stratified diamictite effectively fines upward in this well. The laminated mudrocks represent the background sedimentation of this depositional facies. To date, they have proved to be palynologically barren (N.P. Hooker, 2008, personal commun.) and are interpreted as lake-bottom deposits (as opposed to marine). In the type 1 diamictite beds the diffuse nature of the bed boundaries is suggestive of rain-out through the water column of debris derived from icebergs floating on the surface of the water body, probably during warm season meltout. The identification of the poorly sorted pelletoid clasts is significant in this regard, as they represent “till pellets” sensu Ovenshine (1970). That author considered such deposits to identify uniquely the existence of glacier ice in very close proximity to the environment of deposition. The type 2 diamictites, with their sharp bed contacts, more likely represent debris flow deposition on the bottom of glacial lakes. In well 7 the progressive diminution up-section of type 1 diamictite beds in favor of the laminated mudrocks (described above) suggests increasingly ice-distal conditions through time and/or progressive deepening of the lake. Nonstratified Pebbly Siltstone: Rain-Out Diamictite These rocks comprise a subfacies of the stratified diamictites, described above. In well 8, directly overlying the Unayzah C member, and constituting the entire Unayzah B member in this well, is a 2 ft (0.67 m) thick bed of diamictite. It is mud supported and very poorly sorted, and it displays no internal stratification. It contains an abundance of dispersed grains of fine- to very coarse–grained sand as well as rarer dispersed granules and small pebbles of quartz and sedimentary rock fragments. Melvin and Sprague (2006) illustrated how, at the top of this bed, there occurs a small (1 cm) pebble with its long axis oriented perpendicular to bedding. In the lower part of the bed, indistinct mottling is observed that is suggestive of bioturbation (Melvin and Sprague, 2006). Similar mottled diamictites are seen in well 7, within the stratified diamictite facies. The noted occurrence of the pebble with its long axis oriented perpendicular to bedding supports an origin for these rocks related to melt-out from floating ice in a glacial lake. Similar rain-out diamictites of glacial origin were described and discussed from the Karoo in South Africa by Visser (1982), albeit on a larger scale. The possible bioturbation in these Unayzah B sediments is suggestive of a seasonal aspect to their deposition and supported by its recognition in the stratified diamictites in well 7. Thus, in relatively warm periods, coarse-grained detritus was released from melting ice floes in a glaciolacustrine setting; having settled upon the lake bottom, the sediment was colonized by organisms that were suited to the relatively mild conditions. When harsher conditions prevailed, sediment rain-out was minimized, faunal activity dwindled, and sedimentation was reduced to suspension settling of very fine–grained silt and clay.

Multistory Graded Sandstones: Glaciolacustrine Gravity Flow Deposits This facies is recognized in core from wells 7, 9, 10, 12, 13, and 17. It is most fully developed in well 9, where it is >230 ft (69 m) thick. There it comprises sandstone beds that are 1–5 ft (0.3–1.5 m) thick, displaying sharp, commonly erosional basal contacts, and in many cases they are clearly graded (Fig. 11). They display an abundance of fluid escape structures, dominated by elutriation pillars and dish structures (cf. Lowe and LoPiccolo, 1974). These sandstones are organized into (at least) two upwardthinning and -fining bedset packages (Fig. 11). A similar package of sharp based, graded sandstones occurs in well 10 (Fig. 9A, 0–25 ft), and in that well a cobble sized, angular sandstone clast was recognized, “floating” on the top of one of the graded beds (Fig. 9A, 17 ft). Other occurrences of sharp-based graded sandstones in the Unayzah B member have been described in detail from other wells by Melvin and Sprague (2006). The graded sandstones described above from well 9 (Fig. 11) display all the characteristics of rapidly deposited, highly fluidized sediment gravity flows such as have been described by Lowe and LoPiccolo (1974). The organization of the beds into thick, upward-thinning and upward-fining packages is significant. It suggests sustained deposition in an environment within which, for each package of sediment, the deposits became increasingly distal with respect to their source. It is proposed that these gravity flow deposits represent sublacustrine outpouring of sediment from melting glacier ice. The two distinct upward-fining packages of sandstones may represent either the successive retreat of two different glacial sources (e.g., valley glaciers?), or they may reflect an element of glacial readvance and retreat in the location of this well. That these sublacustrine gravity flow sandstones in the Unayzah B member have a glacial origin is supported by the noted occurrence of the anomalously large clast described from well 10 (Fig. 9A). This is interpreted as a dropstone, which melted out of ice floating on the lake surface, and fell upon the lakebottom gravity flow beneath. In well 13, dropstones have been recognized in fine-grained lake deposits (Melvin and Sprague, 2006) that were subsequently overlain by upward-thinning and -fining packages of gravity flow deposits, similar to the deposits described herein at well 9. Furthermore, the overall trend of those deposits in well 13 was to become thinner and finer upward, suggesting that the sediment supply was becoming more distant (i.e., the rocks became more ice-distal in character), and/or the lake was becoming deeper. Ripple-Drift Cross-Laminated Sandstone: Sublacustrine Glacial Outwash This depositional facies has been observed only in core from well 10, where it occurs in a unit that is ~20 ft (6 m) thick (Fig. 9A, 25–45 ft), resting directly above the interval of sublacustrine gravity flow sandstones mentioned above. It comprises three bedsets, 3–11 ft (0.9–3.3 m) thick, each separated from the other

Late Paleozoic Gondwanan glaciation in Saudi Arabia

53

Figure 11. Wireline gamma-ray (GR) log, showing cored interval and associated core log from well 9, showing thick development of multiple beds of highly fluidized sediment gravity flows (as indicated by abundance of dish structure and elutriation pipes). Note how the beds are arranged in thick packages that show a general upward thinning and fining trend (see text for full discussion). These gravity flow sandstones sit directly upon an interval (uncored) with a high gammaray-log signature, which has yielded samples of the heavily reworked Cm palynoflora (J. Filatoff, 2004, written commun.).

by a thin interval of siltstone. Each bedset consists of fine- to very fine–grained sandstone characterized by multiple, thin (1–3 cm) laminasets of climbing ripple-drift cross-lamination. Type A ripple-drift cross-lamination of Jopling and Walker (1968) dominates the succession, although it is seen to pass upward into type B ripple-drift cross-lamination in the uppermost 3 ft (0.9 m). The sandstones that display type A ripple-drift crosslamination have been interpreted to represent sustained bedload transport of sediment that originated as subaqueous outwash material in a glacial lake (Melvin and Sprague, 2006). Analogous sublacustrine glacial outwash deposits, dominated by laminasets of ripple-drift cross-laminated sandstones, have been described from the Pleistocene of Canada, where they are associated with lake-bottom kame deltas (Gustavson et al., 1975) as well as subaqueous esker fan deposits (Rust and Romanelli, 1975). The uppermost sets of type B ripple-drift cross-lamination in well 10 in the present study suggest decreasing energy, whereby increased fallout from suspended load prevailed over bedload transport. This inferred loss of energy is taken to indicate that the original source of sediment (melting ice) had become somewhat farther removed from the location of well 10, i.e., this succession of ripple-drift cross-laminated sandstones represents an increasingly ice-distal facies association. Mudrock Facies: Distal Glaciolacustrine Deposits Fine-grained sediments of this depositional facies have been observed in a number of wells (e.g., wells 7, 10, 12, 13, and 17) (Melvin and Sprague, 2006). In well 10 (Fig. 9A, 50.5–69.5 ft), 19 ft (5.7 m) of dark gray mudrock overlies a 6 ft (1.8 m) thick diamictite. The lowermost 10 ft (3 m) of this mudrock deposit comprises a series of stacked, thin (centimeter scale) muddy siltstone beds that are sharp based and are distinguished from each other only by very thin (millimeter scale) claystone partings. This lower unit fines upward gradationally into increasingly argillaceous, homogeneous dark gray mudrock; in the uppermost 0.13 ft (4 cm) it becomes a pale gray claystone. The muddy siltstones of the lower part of this mud-prone interval display the characteristics of very fine–grained gravity flow deposits (“distal turbidites”) and pass upward into essentially featureless mudstone, interpreted as suspension fallout deposits. They have been proven to be palynologically barren (J. Filatoff, 2004, written commun.) and are considered to have been laid

54

Melvin et al.

down on the bottom of a lake. The overall upward-fining of these muddy sediments from siltstone to claystone is considered to represent a gradual deepening of the lake through time. Origin and Evolution of the Unayzah B Member The foregoing discussion has highlighted the large number of depositional facies that characterize the Unayzah B member. Except for the inferred periglacial frost wedging described earlier, and the case of wells 2 and 3 in the western part of the study area (to be discussed below), each of those facies is interpreted to have been deposited in a lacustrine environment (palynological evidence for a marine setting being consistently lacking). In most cases there is evidence, direct or indirect, to suggest a glacial influence upon that environment. Thus, the stratabound, internally deformed deposits in well 7 have been interpreted above to represent the remains of glacial push moraines and as such provide strong evidence of an ice-contact setting. The remaining depositional facies are indicative of a spectrum of glacially related environments, ranging from ice-proximal to ice-distal. Considering the volumes of sediment they represent, the >200-ftthick, massive, very poorly sorted pebbly siltstones (diamictites) at well 5 (Fig. 9C), and the >230-ft-thick sequence of multistory graded sandstones (glaciolacustrine gravity flow deposits) in well 9 (Fig. 11) most likely represent deposition in an ice-proximal sublacustrine environment. In the latter case the organization of the gravity flow deposits may suggest either multiple glacial sources or possibly some degree of glacial readvance, as has been discussed earlier. Ice-distal settings are suggested by the very fine–grained sediment (mudrock facies) seen in well 10 as well as by the interstratified pebbly siltstones and laminated mudstones (stratified diamictite) that are developed in wells 7 and 12. Sediments indicative of deposition in settings intermediate in the “proximal-distal spectrum” are represented by nonstratified pebbly siltstone (rain-out diamictite) and ripple-drift crosslaminated sandstone (sublacustrine glacial outwash), as well as the thinner bedded occurrences of massive, very poorly sorted pebbly siltstones (diamictite) and multistory graded sandstones (glaciolacustrine gravity flow deposits). The assignment of these various depositional facies to a position in the “ice proximal-distal spectrum” is substantiated, not only by their grain size, thickness, and general sedimentological characteristics, but also by their associations relative to each other from well to well. That is to say, the vertical association of depositional facies commonly suggests an overall passage in any given location from an ice-proximal setting to one that was ice-distal. This is well displayed in well 10 (Fig. 9A). There, a lowermost interval of at least 25 ft (7.5 m) of medium- to coarsegrained multistory graded sandstones (glaciolacustrine gravity flow deposits) (with dropstones) represents ice-proximal sublacustrine fan deposition. This is overlain by ~20 ft (6 m) of multiple laminasets of ripple-drift cross-laminated sandstone (Fig. 9A, 25–45 ft). These sediments were interpreted earlier also to be sublacustrine glacial outwash detritus, within which the upward pas-

sage from type A to type B ripple-drift cross-lamination (sensu Jopling and Walker, 1968) suggests increasingly ice-distal conditions. They are directly overlain by a massive diamictite deposit that is only ~6 ft (1.8 m) thick. Its limited thickness suggests an intermediate setting in terms of its proximity to any glacial source (especially when compared with the very great thickness of the same facies at well 5). The uppermost 19 ft (5.7 m) of the Unayzah B succession in well 10 comprises ~10 ft (3 m) of mudprone distal turbidites that pass up into 9 ft (2.7 m) of featureless mudrock (Fig. 9A, 50.5–69.5 ft). These fine-grained sediments were laid down in an extremely ice-distal setting. This sedimentary succession at well 10 thus represents a passage through time from a relatively ice-proximal setting to one that was distinctly ice-distal. The clear inference from this evidence is that, relative to the location of well 10, either the ice was becoming increasingly distant (i.e., retreating) or the lake was becoming deeper, or both. It was noted above how Melvin and Sprague (2006) observed similarly that in well 13, a series of upward-thinning and fining packages of gravity flow deposits in the Unayzah B member displayed an overall trend toward being thinner and finer upward, again suggesting that the sediment supply (from the ice) was becoming more distant or the lake was becoming deeper. In well 7 the sediments of the Unayzah B member that directly overlie the top Unayzah C unconformity consist of ~23 ft (6.9 m) of sublacustrine gravity flow deposits (Fig. 7, 3–26.5 ft). These are interpreted as relatively ice-proximal, sublacustrine fan deposits. They are overlain (Fig. 7, 26.5–42.0 ft) by the lower of two units of stratabound, internally deformed deposits (push moraines), inferred to represent ice-contact conditions. The succession continues upward with ~13 ft (3.9 m) of massive, very poorly sorted pebbly siltstones (diamictite) (Fig. 7, 42.0–55.0 ft), and that is again interpreted to be a relatively ice-proximal deposit. The return of an ice-contact setting is seen in the second of the stratabound, internally deformed deposits (push moraines) (Fig. 7, 55–68 ft). The highest part of the Unayzah B member in well 7 (Fig. 7, 68–85 ft) is characterized by well-developed and sustained deposition of interstratified pebbly siltstones and laminated mudstones (stratified diamictite). Those mud-prone sediments represent a significantly more ice-distal setting than any of the underlying sediments. Furthermore, the upward diminution of the amounts of glacially derived sand grains in this facies in this well (described earlier) does itself strongly imply continued retreat of the ice from this location. Thus, in well 7 the Unayzah B member displays evidence for an overall gradual retreat of the ice whereby ice-proximal conditions gave way to ice-distal conditions, but during which at least two distinct icecontact deformational events occurred. That is to say, the overall retreat of the ice was interrupted by at least two minor readvances of the ice sheet in this location. The possibility of minor glacial readvance within the Unayzah B member was also inferred earlier from the architecture of the gravity flow sandstones described from well 9. It is clear from the foregoing examples from individual wells that across most of the study area the vertical association of

Late Paleozoic Gondwanan glaciation in Saudi Arabia depositional facies in the Unayzah B member generally displays an increasingly ice-distal aspect up-section. In all cases this is interpreted as representing the steady retreat of the Gondwanan ice sheet in early Permian times. Locally, there is evidence for some minor readvance of the ice (e.g., at wells 7 and 9). Support for these sedimentological conclusions can be found in the palynology. The Late Carboniferous–Early Permian assemblages that characterize the OSPZ2 palynozone of Stephenson et al. (2003) (which in turn characterizes the Unayzah B member in Saudi Arabia) comprise a variety of spores, pollen, and fresh-water algae (N.P. Hooker, 2008, written commun.). These assemblages can be interpreted as showing a progression from a cold, dry climate in the Late Carboniferous to a slightly warmer and wetter climate in the Early Permian. The latter represents the transition from a glacial to a deglacial phase, and is reflected in the palynoflora by the change from low diversity spores and monosaccate pollen-dominated assemblages (glacial phase) to high diversity spores with fresh-water algae (deglacial phase) (N.P. Hooker, 2008, written commun.). The palynofloral assemblages of the deglacial phase are commonly derived from the massive diamictites described above as being relatively common within the Unayzah B member. Figure 12 is a stratigraphic section that is hung using the top of the “un-named middle Unayzah member” (Melvin and Sprague, 2006) as a datum. It incorporates both the Unayzah B member and the “un-named middle Unayzah member.” It is clear that the Unayzah B member exhibits considerable thickness variation across the study area. Specifically the thickest sections of this stratigraphic unit are seen in wells 5 and 9: it has been discussed above how these two wells contain sediments that are the most ice-proximal in their character. The thinnest occurrence of the Unayzah B member is at well 8 (Fig. 12), where it is represented by only 2 ft (0.6 m) of nonstratified pebbly siltstone (rain-out diamictite) resting directly upon the top of the Unayzah C unconformity. Considered in the context of the preceding discussion, these stratigraphic relationships suggest the following model for the evolution of the Unayzah B member. Following the various retreats and inferred readvances of the Gondwanan ice sheets that led to the deposition and deformation, respectively, of the Unayzah C member (see earlier discussion), the subglacial surface of the final readvance was effectively preserved as the top of the Unayzah C unconformity. Thereafter, the final melting and ultimate withdrawal of the ice took place. As the ice melted, the subglacial topography was exposed, and the lows began to be filled with the meltwaters from the retreating ice sheets, with the consequent formation of abundant glacial lakes. Initially the deepest of those lakes were infilled with large volumes of resedimented glacial outwash material in the form of massive, very poorly sorted ice-proximal debris flows (diamictites) (e.g., well 5, Fig. 9C) and highly fluidized gravity flow sands that developed significant lake bottom turbidite deposits (e.g., well 9, Fig. 12). As the ice continued to retreat, more meltwater and more sediment were supplied, and the numerous glacial hollows were filled and spilled over with rising floodwaters.

55

Locally minor readvances of the ice occurred, with the creation of minor push moraine deposits. In general, however, the terminal retreat of the ice was reflected and preserved in the sedimentary record in the form of increasingly ice-distal deposits. The landscape changed from one of numerous isolated and deep glacial lakes to one that was awash with glacial meltwaters as the lakes overspilled their boundaries and became connected. This terminal (maximum) flood scenario is preserved at well 8, where the thin rain-out diamictite directly overlies the top of the Unayzah C unconformity and appears to link the much deeper glacial lake deposits that characterize wells 7 and 9 (Fig. 12). In the western part of the study area, around wells 2 and 3 (Fig. 12) a somewhat different situation prevailed. It was noted earlier how in each of these two wells the Unayzah B member is represented in its entirety by relatively thick intervals of glacially deformed sediment, which are attributed to sustained, direct contact with the ice. Furthermore, it is clear from the lack of relevant depositional facies that when the ice did eventually melt in this area, the meltwaters were not ponded in topographic lows to form significant lakes such as has been described widely from across the region. The possible corollary exists therefore that the western part of the study area may have been topographically high during Unayzah B times. This is a not altogether unlikely scenario, given the proximity of these westerly locations to the Al Batin Arch of Faqira et al. (2009) (see Fig. 1A). Indeed, given the suggestion by some workers (e.g., Eyles, 1993) that ice coverage may have extended over to present-day east Africa, it is interesting to speculate upon the possibility of a center of high altitude (alpine) glaciation in western central Saudi Arabia at this time. UNNAMED MIDDLE UNAYZAH MEMBER The unnamed middle Unayzah member was recognized and defined in the extensive core-based study of the lower Unayzah Formation carried out by Melvin and Sprague (2006). Its facies characteristics show that it is very different from the glaciogenic Unayzah B member, and yet it displays a distinctive character that sets it aside also from the Unayzah A member. It is a stratal unit that represents a transition from the Unayzah B member to the Unayzah A member, even though its boundaries are sharp and readily identified in core. Among the wells examined in this study, the unnamed middle Unayzah member occurs in a number of situations. Thus it may be identified directly overlying the Unayzah C member, as at well 6 (Fig. 12), although more commonly it sits upon the glaciogenic facies of the Unayzah B member. In places it is absent, as at well 11 (Fig. 12), and elsewhere it may have been removed by erosion at the pre-Khuff unconformity, as seen in wells 2 and 3 (Fig. 12). In well 7 the unnamed middle Unayzah member appears to have been completely cored, and its relationships to the Unayzah B and A members are clearly exposed (Fig. 7). This well thus serves as a valuable reference well. The nature of the individual depositional facies and the stratigraphic boundaries of the unnamed middle Unayzah member are discussed below.

56

Melvin et al.

Figure 12. Stratigraphic section across the study area, showing facies architecture within the Unayzah B member and the overlying unnamed middle Unayzah member. Datum is the top of the unnamed middle Unayzah member. In wells where this sediment package is thick (e.g., wells 7, 9), the lowermost parts of the section are dominated by glaciogenic deposits (Unayzah B member). These deposits are overlain in those wells by nonglacial, red lacustrine siltstones and associated sandstones of the floodplain facies of the unnamed middle Unayzah member. Where the overall package is thin (e.g., well 6), it comprises only the latter member. Note the effects of the pre-Khuff unconformity (PKU) in wells 2 and 3. Map shows line of section. Modified from Melvin and Sprague (2006). GR—gamma-ray.

Fluvial Sandstones The best example of fluvial sandstones in the unnamed middle Unayzah member is seen in core from well 8. There, Melvin and Sprague (2006) identified a sandstone interval that is 30.5 ft (9.15 m) thick (Fig. 13A, 29.5–60.0 ft). It directly overlies some 10 ft (3 m) of red sandy siltstones (alluvial floodplain deposits: see below) and occurs ~18 ft (5.4 m) above the top of the very thin Unayzah B member in this well. It is also overlain by a thick interval of 45 ft (13.5 m) of red sandy siltstones of the alluvial floodplain facies (see below). Fine- to mediumgrained, moderately sorted sandstones occur in stacked beds that are 2–4 ft (0.6–1.2 m) thick. They display sharp, erosional contacts, with local mud-clast concentrations, and pass upward into low-angle trough cross-lamination overlain in places by finer–

grained, ripple cross-laminated sandstones (Fig. 13A). These characteristics of the sandstones, and their occurrence embedded within red very fine–grained (silty) deposits, suggest deposition in a channelized, heavily oxidized (i.e., terrestrial) setting. They are thus interpreted to be fluvial in origin, and their architecture of stacked shallow channel deposits encased and isolated within finer–grained floodplain deposits further suggests deposition in relatively high-sinuosity or even anastomosing rivers. Fluvial channel sandstone deposits, isolated within fine-grained floodplain siltstones, were described from the Cutler Formation (Permian to Pennsylvanian) of New Mexico by Eberth and Miall (1991), and were similarly interpreted to be anastomosed river deposits. In well 7 there occurs a similar interval of sandstones that shows cross-bedding and an upward-fining profile; these are also considered to be of fluvial origin (Fig. 13B, 41–50 ft).

Late Paleozoic Gondwanan glaciation in Saudi Arabia

57

Figure 13. Core logs from wells 8, 7, and 6, illustrating the nature of the unnamed middle Unayzah member (“UmUm”). Note in particular the distinct sandstone units overlying siltstone intervals, and the similarities and differences displayed among them. See text for detailed discussion.

Eolian Sandstones This facies is well displayed in the unnamed middle Unayzah member in wells 7 and 6 (Fig. 13B and 13C). In well 7, there is an interval of ~17 ft (5.1 m) of low-angle to flat-laminated, well-sorted fine-grained sandstones wherein the grains are well rounded and appear frosted (Fig. 13B, 53–70 ft). In well 6, texturally similar flat-laminated sandstones are interbedded with more poorly sorted sandstones, displaying well-developed adhesion ripples (Fig. 13C, 23–40 ft). Thin (centimeter scale) beds of

dark gray siltstone are also present. Overlying these sandstones in well 6 is an interval 5 ft (1.5 m) thick comprising well-sorted, well-rounded, and frosted grains that occur in high-angle, grainsize-segregated cross-laminations (Fig. 13C, 40–45 ft). The cross-laminations are disrupted by small-scale synsedimentary faults (Fig. 13C, 42–43 ft). Synsedimentary deformation in similar facies in the unnamed middle Unayzah member has been observed in several other places. It is particularly well developed in western parts of the study area in central Saudi Arabia, in the vicinity of well 24 (see Fig. 1A).

58

Melvin et al.

Melvin and Sprague (2006) described how the base of the unnamed middle Unayzah member in well 7 features a thin (2 ft; 0.6 m) red sandstone that rests abruptly upon the dark gray stratified diamictite at the top of the Unayzah B member in that well (Fig. 13B, 4–6 ft). That sandstone has high-angle crosslaminations that become quite diffuse in the lower levels of the bed. It also displays local nested concentrations of subrounded to subangular, granule-sized grains. It has a sharp, granule- and pebble-strewn basal contact with the underlying glaciolacustrine stratified diamictites of the Unayzah B member. The latter rocks display considerable cracking in their uppermost part, and the red sandstone is seen to penetrate deeply into those dislocations (Fig. 14A). The flat-laminated sandstones described above from well 6 and well 7 have the character of eolian sand sheet deposits, such as were well documented by Fryberger et al. (1979). The

A

beds showing adhesion ripples represent damper, sandy interdune conditions, and the thin siltstone interbeds suggest shortlived, shallow bodies of water in an otherwise eolian dominated environment. The overlying cross-laminated sandstone interval in well 6 is interpreted to be a residual eolian dune deposit. The synsedimentary dislocations may represent gravitational collapse on the dune slip face. Alternatively, Melvin and Sprague (2006) suggested the possibility that these dislocations may relate to sediment readjustment in response to melting of bodies of ice or snow that may have been trapped during dune migration. Analogous features were described from modern cold-climate eolianites from the Rocky Mountains of Colorado by Ahlbrandt and Andrews (1978) and from coeval Permian deposits of the Gondwanan sequence of Australia (Williams et al., 1985). Ostensibly, the thin red sandstone that occurs at the base of the unnamed middle Unayzah member in well 7 (Fig. 13B,

B

Figure 14. Core photographs showing the nature of the lower and upper contacts of the unnamed middle Unayzah member in well 7. (A) Gray, silty glaciolacustrine mudstone at the top of the Unayzah B member, abruptly overlain by red eolian sandstones at the base of the unnamed middle Unayzah member (“UmUm”) (see Fig. 13B, 5 ft). At the contact, note how the glaciolacustrine deposits are cracked, and the coarse basal sediment of the red sandstones has deeply invaded the cracks. (B) Red siltstone and silty, very fine–grained sandstone. The mottled and rooted lower interval is the uppermost part of the unnamed middle Unayzah member (“UmUm”) in this well, and represents a paleosol (Fig. 13B, 72.5–77 ft). The abruptly supradjacent siltstones are assigned to the Unayzah A member.

Late Paleozoic Gondwanan glaciation in Saudi Arabia 4–6 ft) displays the character of a residual, basal eolian dune sandstone deposit. The juxtaposition of such terrestrial sediment relative to the underlying glaciolacustrine stratified diamictites has significant stratigraphic implications that are discussed further below. Red Sandy Siltstones: Alluvial Floodplain Deposits The unnamed middle Unayzah member is commonly characterized by extensive deposits of sandy siltstones and silty, very fine–grained sandstones that are generally red, ranging from reddish brown to reddish purple, and locally red-green variegated. They are well displayed in core from well 6 (Fig. 13C, 10.5– 22.0 ft) and well 7 (Fig. 13B, 7–33 ft). In well 8, 10 ft (3 m) of red siltstones are seen below the fluvial sandstones described above (Fig. 15, 19.5–29.5 ft), and a further 47 ft (14.1 m) of red sandy siltstones are present above those sandstones (Fig. 15, 60–107 ft). In that well the highest parts of the unnamed middle Unayzah member show a distinct, albeit subtle upward-coarsening grain size trend (Fig. 15, 107–129 ft). Although generally these red sandy siltstones display very diffuse bedding characteristics, in places they are interbedded with red, silty very fine–grained sandstones that can display irregular lamination or occur as very thin (centimeter scale), current-rippled beds. Rarely, these beds have sharp upper contacts, showing evidence of sand-filled desiccation cracks. These fine-grained red beds were laid down under highly oxidizing conditions, as suggested by the widespread red coloration. They are interpreted to have been deposited in a terrestrial environment dominated by low energy conditions. The generally diffuse bedding characteristics are indicative of heavily waterlogged sediment, and the local development of ripple lamination suggests the presence of minor current activity. Thus this facies is considered to represent subaqueous deposition within shallow floodplain lakes. Locally, these lakes were at times infilled to the point of exposure and desiccation. In that context, the upwardcoarsening interval at the top of the unnamed middle Unayzah member in well 8 (Fig. 15, 107–129 ft) is tentatively interpreted as a small floodplain lacustrine delta. Paleosols In well 7, the uppermost part of the unnamed middle Unayzah member is represented by ~4 ft (1.2 m) of red-purple-gray variegated argillaceous and silty very fine–grained sandstones (Fig. 13B, 72–76 ft; Fig. 14B). These sandstones are tightly silica cemented and poorly sorted, containing an abundance of dispersed (floating) grains of medium- to coarse-grained sand. They are characterized by well-developed root traces up to 1 ft (30 cm) long, cutans, and other indicators of pedogenic development (Fig. 14B). In well 6 the eolian sandstone described earlier from the unnamed middle Unayzah member is overlain by ~4 ft (1.2 m) of poorly sorted sediment (Fig. 13C, 46–50 ft) that contains abundant floating grains of medium to coarse sand. In many

59

places these sand grains appear to be concentrated within subvertical cracks within the sediment. This interval in well 6 displays characteristic subspherical fractures throughout. Similar features have also been noted at the top of the unnamed middle Unayzah member in well 9. These variegated, texturally immature rocks are interpreted as paleosols. In most places where they are recognized, they represent the stratigraphically highest occurring facies within the unnamed middle Unayzah member. Origin and Evolution of the Unnamed Middle Unayzah Member The contact that separates the Unayzah B member from the unnamed middle Unayzah member is extremely sharp (Fig. 14A) and is recognized across the study area (Fig. 12). Melvin and Sprague (2006) interpreted the contact to represent a regionally disconformable surface. The various facies that characterize the uppermost part of the underlying Unayzah B member represent widespread lake-dominated sedimentation at the time of maximum flooding of glacial meltwaters across the landscape in Early Permian times, as discussed above. The (possible) cold-climate eolian, fluvial, and fine-grained alluvial red-bed deposits of the unnamed middle Unayzah member clearly represent a very different, terrestrially dominated depositional setting and suggest that a highly significant drainage event marked the end of Unayzah B deposition. Martini and Brookfield (1995) described how, in the Pleistocene Bowmanville Bluffs in Ontario, Canada, a very sharp demarcation occurs between clay-rich glaciolacustrine rhythmites and overlying sand-prone sediments. They suggested that the widespread distribution of this contact and the nature of the overlying successions were the result not of erosion but of a sudden drop in lake level caused by the retreat of a glacier. Teller et al. (2002) noted that during the last deglaciation of the Quaternary period, melting of the Laurentide Ice Sheet in North America led to the release of very large volumes of stored precipitation to the oceans. There were a number of lakes along the margin of the Laurentide Ice Sheet, whose confining ice or sediment dams failed during the deglaciation. Of these, proglacial Lake Agassiz was by far the largest, covering a total of more than one million square kilometers over its 4000 yr history (Teller et al., 2002). Frequent changes in lake levels of Lake Agassiz have been documented. These changes were abrupt and often involved the release of several thousand cubic kilometers of water. The exact time for each lake drawdown (outburst) can only be estimated, but it is believed that lake levels for most phases could have been drawn down in a matter of months to a few years (Teller et al., 2002). Similarly, in northern Russia during the last glaciation of the Quaternary Period, the North Polar ice sheets expanded and blocked north-flowing rivers such as the Yenissei, Ob, and Pechora. As a result, south of the ice sheets a number of large ice-dammed lakes formed that were considerably larger than any lake on Earth today (Mangerud et al., 2004). The final

Figure 15. Core log through ~240 ft (72 m) of core in well 8, extending from the uppermost part of the Unayzah C member to the top of the Unayzah A member. Note the stratigraphic boundaries and the overall upward-coarsening profile from 132 to 238 ft. Modified from Melvin et al. (2005).

Late Paleozoic Gondwanan glaciation in Saudi Arabia drainage of the best mapped lake (namely Lake Komi in the Pechora Lowlands of northern Siberia) was modeled, and it was concluded that it probably emptied within a few months (Mangerud et al., 2004). These results dramatically demonstrate the likelihood of geologically instantaneous drainage events related to glaciolacustrine settings at times of terminal glacial retreat. Following the probably dramatic drainage event that is inferred to have terminated glaciolacustrine deposition of the Unayzah B member, the Permian landscape was dominated for the most part by very low lying alluvial floodplains that accommodated deposition of the large volumes of fine-grained (silt-sized) material that now constitutes the greater part of the unnamed middle Unayzah member. A number of river systems traversed these flood basins, but they were probably relatively isolated in occurrence, as is indicated from the limited evidence in the apparent relative isolation of their resulting channel deposits (Figs. 13A and 15). Toward the end of the period of deposition of the unnamed middle Unayzah member, relatively drier conditions became established, accompanied by a reduction in sediment supply, and the fluvial sands were reworked into a number of isolated eolian deposits (Figs. 13B and 13C). The possibility that these sandstones may represent cold-climate eolianites has considerable implications for the Permian paleogeography of the study area, particularly in the western part. There, around well 24 (Fig. 1A), these postulated cold-climate eolian deposits are particularly well developed. It was mooted earlier that the possibility exists that a high-altitude Permian glaciation was active and centered on the Al Batin Arch (see Fig. 1A). If such was the case the likelihood that these deformed eolian sandstones of the unnamed middle Unayzah member are indeed cold-climate deposits, laid down at a relatively high altitude, is rendered significantly more credible. Significantly, ongoing palynological studies from the same general area appear also to suggest assemblages having “montane” affinities (N.P. Hooker, 2008, personal commun.). Across the study area as a whole, active deposition of the unnamed middle Unayzah member ultimately ceased, and soilforming processes prevailed. These resulted in the formation of the paleosols that in many places characterize the top of the unnamed middle Unayzah member. Those soils represent a depositional hiatus and consequently mark a disconformable surface that separates the unnamed middle Unayzah member from the overlying Unayzah A member. UNAYZAH A MEMBER The Unayzah A member has a widespread distribution across the study area. Generally its thickness ranges from 150 to 300 ft (45–90 m) but does not display the extreme variation recognized in the Unayzah C or Unayzah B members. Although identified in the subsurface across the area of investigation, the Unayzah A member nonetheless is difficult to correlate in detail from wireline logs. This is because of the intrinsic variability in depositional facies distribution, confirmed from examination of >3150 ft (945 m) of Unayzah A core in the present study. This

61

work has identified several different depositional facies (summarized below), and these and their resultant facies associations have permitted the recognition of three major depositional environments within the Unayzah A member. All of these environments are strongly indicative of a highly arid continental setting, and hence are suggestive of a significant and sustained climate change that became manifest at the beginning of deposition of the Unayzah A member. Semiarid Ephemeral Lake Deposits These rocks characterize the lower intervals of the Unayzah A member almost everywhere it occurs. They comprise silty, very fine–grained sandstones with minor sandy siltstones that range in color from brick red and red brown to buff yellow and pale to dark gray. They display variably lenticular to indistinct crinkly lamination, and locally small-scale ripple cross-lamination is seen. Rarely, over-steepened and folded sediment is observed. Elsewhere, small-scale vertical features have been noted and interpreted as syneresis cracks. These very fine–grained rocks are specifically characterized by the common occurrence of thin, flat-lying (horizontal) laminae, commonly only one or two grains thick (Fig. 15, 136–186 ft). The laminae comprise grains of wellrounded, medium to coarse sand. Commonly, bedding and lamination within these rocks are disrupted by subvertical, sand-filled cracks that have a downward penetration of 2–5 cm. These fine-grained silty sediments dominate the lower parts of the Unayzah A member, as was noted above. As such, they commonly overlie the similar red siltstones that were described earlier from the unnamed middle Unayzah member, separated from them in most places only by the paleosols that mark the end of deposition of the latter unit. The subtle but critical difference between these two siltstone facies lies in the abundance of coarser-grained sand laminae, as well as a greater number of sand-filled vertical cracks, that are observed within the lower zones of the Unayzah A member. The siltstones within the unnamed middle Unayzah member have been interpreted above as the depositional infill of alluvial floodplain lakes. Sedimentological evidence for those lakes being relatively long-lived bodies of standing water is much greater than is seen in the lower Unayzah A member. In the latter case, the sediment is similarly interpreted to be of a shallow lacustrine origin, wherein waterlain deposits are indeed preserved (as indicated by the ripple forms and evidence of slumping). However, there is much more abundant evidence for fluctuation of the lake levels, and consequent repeated and widespread exposure and desiccation within these lower Unayzah A silty, very fine– grained sandstones. That evidence includes the common occurrence of subvertical sand-filled cracks (desiccation cracks) and the abundant horizontal laminae of coarser-grained sand. These features are interpreted as being related to quasi-planar adhesion laminae (sensu Hunter, 1980) or adhesion laminations (Kocurek and Fielder, 1982) that were laid down upon exposed damp surfaces. The lowermost interval of the Unayzah A member is thus

62

Melvin et al.

interpreted to represent deposition within an ephemeral lake basin setting (playa lakes) and as such indicates much drier conditions than those that prevailed during deposition of the unnamed middle Unayzah member. Semiarid Ephemeral Stream Deposits These deposits constitute the greater part of the upper Unayzah A member in a number of wells across the study area. Thus, for example in well 8, the upper part of the Unayzah A member displays a number of interbedded sandstones and siltstones (Fig. 15, 186–238 ft). The sandstones are thin-bedded units, rarely more than ~1 ft (30 cm) thick, that make up the lower parts of upwardfining couplets. They are fine to medium grained, moderately to poorly sorted, and generally either massive or flat laminated with sharp, erosional lower contacts (locally with numerous small mud clasts). These sandstones can be single units, or in places they comprise thin amalgamated beds (Fig. 16A). They are commonly overlain by finer-grained beds of similar thickness that form the upper parts of the depositional couplets referred to above. These beds comprise very fine–grained silty sandstone that displays well-developed ripple cross-lamination (Fig. 16B). In places these silty, rippled sandstones contain an abundance of mud drapes that display well-developed desiccation features (Fig. 15, 194–196 ft; Fig. 16C). The lower, sharp-based and coarser sandstones of these upward-fining couplets are interpreted as having been deposited rapidly by flash floods, and the overlying, finer-grained, and ripplelaminated sediment represents the waning flow deposits of those

A

B

flood events. The repeated occurrence of desiccation in the finer, mud-draped sediment emphasizes the episodic nature of deposition of these beds in a semiarid environment; i.e., these are the sporadic deposits of ephemeral streams. The overall depositional profile of the Unayzah A member at well 8 (Fig. 15, 132–238 ft) displays a clear upward-coarsening trend, passing gradationally from a siltstone-dominated regime upward into medium- and coarse-grained sandstones. The siltstone-dominated interval represents the semiarid ephemeral lake (playa lake) deposits of the lower Unayzah A member, discussed above. The upward-coarsening depositional trend that overlies these playa lake sediments in well 8 represents the progradation into the lake of a terminal alluvial fan, or a lake-marginal bajada, that was traversed by a number of shallow, ephemeral stream channels. Eolian Erg Center to Erg Margin Deposits Although the semiarid ephemeral stream deposits described above are encountered in the upper Unayzah A member in several places in the subsurface of eastern and central Saudi Arabia, they do not everywhere represent the dominant depositional facies association of that stratigraphic unit. In many places the uppermost part of the Unayzah A member consists of a complex of sandstone facies that can be ascribed in general to an eolian setting (Fig. 17), and which are summarized below. Eolian dune cross-bedded sandstones are common throughout the upper Unayzah A member. They comprise fineto medium-grained, well to very well sorted sandstones, with very well rounded and frosted grains of quartz sand that occur

C

Figure 16. Core photographs illustrating aspects of the semiarid ephemeral stream facies association of the upper Unayzah A member (see text for discussion). (A) Amalgamated medium-grained sandstones representing flash flood deposition. Note the sharp, scoured, and loaded contacts (arrows), suggestive of rapid, high energy sedimentation. (B) Argillaceous, ripple laminated fine- to very fine–grained sandstones indicative of waning flow conditions following flash flood events. (C) Abundance of desiccated clay drapes in the ripple laminated facies (arrows).

Figure 17. Core log through ~210 ft (63 m) of core in well 21, representing the various facies identified in the eolian facies association of the upper Unayzah A member. Note in particular the abrupt horizontal upper terminations of the cross-bedded facies in many places. The facies are identified as follows: (i) eolian dune cross-bedded sandstones; (ii) eolian sand sheet sandstones; (iii) paleosols; (iv) sandy interdune deposits (damp interdunes); (v) silty interdune deposits (wet interdunes). Modified from Melvin et al. (2005).

64

Melvin et al.

in pronounced high angle (>30°), grain size–segregated crosslaminations. Those laminations may be very closely spaced (“pin-striped,” sensu Fryberger and Schenk, 1988) (Fig. 18A) and inversely graded, where they are interpreted as wind-ripple laminations (subcritically climbing translatent strata: Kocurek and Dott, 1981) that formed on the slip faces of eolian dunes. Alternatively, the cross-laminations are a few centimeters thick, in which case they are better interpreted as grain flow cross-strata (sensu Kocurek and Dott, 1981) that formed by gravity sliding of sand down the dune slip faces. These eolian dune cross-bedded deposits are readily identified on downhole image logs, which, crucially, allows this facies to be recognized in uncored wells. Eolian sand sheet deposits are also widely recognized within the upper Unayzah A member. They comprise fine- to

medium-grained sandstones, with very well rounded and frosted grains of sand that occur in low angle to flat, grain size–segregated laminations (Fig. 17, 90–98 ft; Fig. 18B). Many sets of these laminations have a pin-striped appearance, suggesting an origin from wind-ripple migration. Flat-based, lenticular (convex-up) accumulations of coarse-grained, well rounded sand are also observed within these deposits. It is possible in places to distinguish very low angle truncations separating different sets of the low-angle laminated sand. Some examples of this facies display disruption of the laminae, attributed to plant roots, as well as to other unknown forms of faunal bioturbation (insects? spiders?) (Fig. 18B). All of these characteristics compare favorably with features seen in eolian sand sheets, as they were described by Fryberger et al. (1979).

C A

B

D

Figure 18. Core photographs illustrating facies representative of the eolian erg center to erg margin facies association in the upper Unayzah A member. (A) Pronounced high-angle cross-lamination of the eolian dune facies. These laminations include very closely spaced (pin-striped) wind ripple laminations (wr) and thicker grain flow laminations (gf). (B) Very low-angle wind ripple lamination (wr) of the eolian sand sheet depositional facies. Note: low-angle truncations (arrowed) and subtle high-angle disruptions (insect burrows?) (arrowed, ib). These laminated sands overlie an interval at the bottom of the photograph that shows indistinct adhesion ripple (ar) structures. (C) Irregular crinkly laminations that are associated with damp interdune settings. White patches are irregular occurrences of anhydrite cement. (D) Disrupted and brecciated sandstone typical of many paleosols found within the arid eolian facies association of the upper Unayzah A member.

Late Paleozoic Gondwanan glaciation in Saudi Arabia Sandy interdune deposits (damp interdunes) are commonly recognized in the upper Unayzah A member. They consist of moderately sorted, fine- to medium-grained sandstones that contain a heterogeneous assemblage of depositional structures, including irregular to crinkly, variably continuous laminations and very thin (centimeter scale) beds with well-developed adhesion ripple laminations (Fig. 17, 84–90 ft; Fig. 18C). The occurrence of adhesion ripple laminations suggests that these sediments were laid down upon a damp substrate (cf. Kocurek and Fielder, 1982). The irregular nature of the laminations is reminiscent of textures observed in damp interdune, or sandy sabkha, environments (e.g., see Fryberger et al., 1983). Given the association of these sediments with facies (described above) that were clearly deposited in an arid, eolian-dominated setting, these irregularly laminated sandstones of the upper Unayzah A member are similarly interpreted to represent a damp interdune environment in close proximity to the paleo–water table. Silty interdune deposits (wet interdunes) comprise siltstones and silty, very fine–grained sandstones showing variably continuous to lenticular and crinkly lamination. In places small-scale ripple cross-lamination is seen, suggesting subaqueous deposition. These sediments occur in intervals that are rarely >1 ft (30 cm) to 2 ft (60 cm) thick. They show many depositional similarities to the semiarid lake deposits described above. The significant difference lies in the thickness of their occurrence. They are thus interpreted not as ephemeral (playa) lake sediments but rather as the deposits of temporary, very shallow interdune ponds that formed at times when the water table rose above the depositional surface within an otherwise arid (desert) environment. Sandy paleosol horizons have also been recognized at a number of localities within the eolian-dominated upper Unayzah A member. They are 0.3–4.0 ft (0.09–1.2 m) thick (Fig. 17, 53–57 ft, 126 ft, 151 ft), and comprise poorly sorted, very fine to medium-grained sandstones, with patchy carbonate cementation and showing varying degrees of disrupted texture (Fig. 18D). These paleosol horizons suggest that the paleo–water table was relatively high at the time of their formation. Individually, the above-described eolian and eolian-related facies occur commonly throughout the upper part of the Unayzah A member. The mutual associations of these facies, as well as their abundances relative to each other, vary markedly from well to well, and indeed from field to field across the study area. Thus in some places the upper part of the Unayzah A member is dominated (almost to the exclusion of any other facies) by stacked eolian dune cross-bedded sandstones that are clearly seen in both core and image-log data. These dune-dominated occurrences are interpreted as representative of eolian erg-center settings. Elsewhere a much greater mix of the various eolian facies (described above) is seen: those mixed facies associations are interpreted as eolian erg-margin deposits. Similar eolian-related facies associations have been described from the Permian Cedar Mesa Sandstone in Utah by Mountney and Jagger (2004). If well 21 (Fig. 17) is taken as a typical example of a well within which the facies of the upper Unayzah A member are eolian (as opposed to

65

ephemeral fluvial) dominated, the repeatability of facies is clear, and the facies boundaries are seen to be sharp (Fig. 17). In particular, the eolian dune cross-bedded sandstones are in many cases abruptly truncated by a horizontal contact with the overlying facies (Fig. 19A; Fig. 17, 24 ft, 42 ft, 125 ft, 151 ft, 163 ft). These abrupt horizontal truncations of eolian cross-strata are identified as Stokes surfaces (after Stokes, 1968) and are attributed to a rising water table in the Permian eolian depositional environment (see Fryberger et al., 1988). The stratigraphic significance of the identification in the core of such rises in the paleo–water table was emphasized by Melvin et al. (2005). Subsidiary to the current work, a field-scale study was undertaken of the Unayzah A member at the southern end of the giant Ghawar structure in eastern Saudi Arabia that included wells 15 through 20 (see Fig. 1A). Details of that study are discussed by Melvin et al. (2010). There, the lower part of the Unayzah A member (described above as dominated by ephemeral lake facies) was seen in cores from many of the wells to terminate upward in a thin (decimeter scale), subtle upward-fining trend, with a passage from silty, very fine–grained sandstone to siltstone sensu stricto (Fig. 19B). That upward-fining trend can be interpreted as a deepening of the lower Unayzah A lake. The surface representing the top of this vertical, upward-fining trend among the wells has been interpreted further to represent in a lateral dimension the maximum extent of the lake (MEL) (Melvin et al., 2010). Above this MEL horizon, each well within the subsidiary study area displays its own assemblage of the various upper Unayzah A member eolian depositional facies (described above). Within these assemblages, respectively, a number of cyclical rises in the paleo–water table can be recognized based on the identification of Stokes surfaces in core, as well as on image logs, and the general associations of the various facies. Crucially, when the data from each well are hung stratigraphically using the MEL horizon as a datum, many of the interpreted rises in the paleo–water table that are recognized above the MEL in the various wells correlate exactly, irrespective of the depositional facies within which they are found (Fig. 20) (Melvin et al., 2010). This correlation extends along a distance of ~65 km. Thus, a correlatable layering scheme is crucially established that is related to paleo–water table fluctuations within the eolian-dominated Unayzah A member. The correlation demonstrably carries through different depositional facies tracts and is intrinsically established on sequence stratigraphic principles. Terminal Facies of the Uppermost Unayzah A Member The preceding discussion has described how the Unayzah A member is characterized by a wide variety of different depositional facies that generally fall into three distinctive facies associations, all of which are strongly indicative of deposition within a semiarid to arid setting. Thus this member almost universally in its lower part comprises playa lake deposits that pass upward into either ephemeral stream facies or eolian erg to erg margin facies. The ultimate termination of these conditions is represented in

Figure 19. Core photographs illustrating features of stratigraphic significance within the eolian facies association of the upper Unayzah A member. (A) Abrupt upper horizontal termination of pronounced eolian dune cross-bedding. This is a “Stokes surface” and is overlain in this example by sand sheet and interdune facies. It represents a rise through the dune deposits of the paleo–water table. Note: Large white spots are postdepositional nodules of anhydrite. (B) Upwardfining of very fine–grained sandstones into siltstone at the top of the lower Unayzah A member ephemeral lake deposits in well 18. The abrupt contact with overlying coarser (eolian) sandstones defines the maximum extent of the lake (MEL) in this well. See text for full discussion.

A

B

Figure 20. Stratigraphic cross section through the Unayzah A member in selected wells at the southern end of the Ghawar structure (see Fig. 1A). The section has the MEL (maximum extent of the lake) horizon (see text for discussion) as its datum. Note how a number of maximum wetting horizons above the MEL can be identified (from facies relationships seen in core and image logs), and that these appear to be correlatable along the length of the section, irrespective of the facies within which they are found. These wetting cycles are interpreted to be related to fluctuations in the paleo–water table. PKU—pre-Khuff unconformity.

Late Paleozoic Gondwanan glaciation in Saudi Arabia many wells across the study area by very different and distinctive facies at the top of the Unayzah A member (Fig. 21). Thus in well 24 the stratigraphically highest eolian deposits (Fig. 21A, 1–9 ft) are overlain by 21 ft (6.3 m) of finer-grained “interdune” deposits that differ from those normally seen in the upper Unayzah A member in that they display minor evidence of burrowing. Those sandstones are overlain by a highly distinctive interval 7 ft (2.1 m) thick, showing an abundance of burrows (Fig. 21A, 32–39 ft; Fig. 22A). Above this intensely bioturbated zone there occurs in well 24 a thick development of paleosols (Fig. 21A, 39–83 ft). Those paleosols appear to comprise at least five individual paleosol deposits each of which displays a number of probable soil “zones.” The facies characterizing these zones vary from very argillaceous, poorly sorted sandstones with an abundance of cutans and possible rooting features to extremely well developed, massive to nodular silcretes (e.g., Fig. 21A, 68–70 ft; Fig. 22B). This 44 ft (13.2 m)–thick interval of paleosols is superseded in this well by a heavily rubbleized interval of red silty mudstone that displays in its lower part an abundance of very thin (millimeter scale) laminations of fine- to mediumgrained sandstone. This sequence in well 24 can be compared favorably with the highest part of the upper Unayzah A member in well 21. In the latter well, some 300 km distant from well 24 (see Fig. 1A), the highest eolian dune sand is overlain by ~20 ft (6.0 m) of paleosol deposits (Fig. 21B, 10–30 ft). Those rocks subsequently pass upward through some upward-fining sandstones into an interval of low-angle laminated sandstone wherein the laminations are severely disrupted by intense burrowing (Fig. 21B, 39–45 ft; Fig. 22C). Overlying these burrowed sediments is another interval of thickly developed paleosols (Fig. 21B, 45–76 ft). Again, these paleosols appear to comprise a number (five or six) of stacked, individual paleosol deposits, within which various facies representing soil zonation occur. Those zones are similar to those identified in well 24 (see above) in that they vary from very argillaceous, poorly sorted sandstones with an abundance of cutans and apparent root-related structures to well-developed massive to nodular silcretes (e.g., Fig. 22D). It appears that, notwithstanding the great distance that separates wells 24 and 21, there are great similarities between the two wells in terms of the facies associations in the uppermost parts of the Unayzah A member. Specifically there appears to have been a thin, probable marine event, identified by intensive burrowing, that was followed in both cases by the development of thick, stacked paleosol deposits. In well 24 these paleosols are overlain by fine-grained red beds; in well 21 there is a 5 ft gap of non-recovery of core that is superseded by sandstones of shallow-marine origin and ascribed to the Basal Khuff Clastics member of the Khuff Formation (see later discussion).

67

a large number of depositional facies are readily identified. It is also evident that most of those facies reflect deposition within a significantly arid environment. Thus, depending upon paleogeographical location within the study area, depositional facies associations reflect either (1) arid to semiarid conditions that are characteristic of terminal alluvial fans and bajadas or (2) widespread eolian erg systems, both of which encroached upon ephemeral lakes (playas) whose dimensions fluctuated throughout time. Stratigraphic analysis of one of these erg systems by Melvin et al. (2010) revealed a high degree of cyclicity among facies that results in a layering scheme that is correlatable both within and between depositional facies tracts (see earlier discussion). This internal stratigraphy within the eolian-dominated Unayzah A member has been interpreted to be a direct reflection of fluctuations in the level of the paleo–water table throughout the time of its deposition. These correlatable rises in the paleo–water table are intrinsically allostratigraphic in nature, and it is tempting to consider their cyclicity as possibly having its basis in the Earth’s orbital fluctuations. If the Gondwanan Permo-Carboniferous ice sheet was still significantly (albeit very distally) in existence during deposition of the Unayzah A member, and if such orbital fluctuations had an impact on the melting of the ice and the consequent postglacial global transgression, then it may be that even within the terrestrial rocks of the Unayzah A member there is a record of the pulsatory nature of that transgression. It appears to be highly significant in this regard that the stratigraphically highest of these terrestrial deposits in locations so widely separated as well 24 and 21 (Fig. 1A) are superseded by a thin zone of sandstone that displays a distinctly marine signature in the form of intense burrowing activity (Figs. 21 and 22). This would suggest that the episodic rises in base level inferred from the cyclicity identified from rises in the paleo–water table did indeed ultimately manifest themselves with a breakthrough of marine waters at the very end of the time of deposition of the Unayzah A member. It is clear that the end of deposition of the Unayzah A member in most areas was marked by a (probably prolonged) period of minimal deposition to nondeposition. This was reflected initially by the intensely bioturbated sandstone (indicating low rates of sedimentation) and was manifest ultimately by the development of the very thick and pervasive paleosols seen in many places at the top of this stratal unit. The widespread extent of those paleosols testifies to the development of a sustained, high water table that may indicate a change in climate (decreasing aridity). The significance of this inferred prolonged period of minimal deposition to nondeposition in terms of the overall tectono-stratigraphic evolution of Saudi Arabia is discussed in the final section of this paper. BASAL KHUFF CLASTICS MEMBER OF THE KHUFF FORMATION

Origin and Evolution of the Unayzah A Member It is clear from the foregoing discussion that the Unayzah A member is a highly complex stratigraphic unit, within which

Senalp and Al-Duaiji (1995) observed that the Unayzah Formation at outcrop is truncated by an unconformity known as the pre-Khuff unconformity. In the subsurface of the study

Figure 21. Core logs through the uppermost Unayzah A member at (A) well 24 and (B) well 21. These wells are ~300 km apart. Nonetheless, they display a remarkable similarity in their stratigraphic organization. Note in particular the occurrence in each well of a pronounced bioturbated sandstone zone (bs) ~25–30 ft (7.5–9.0 m) above the highest occurrence of eolian dune cross-bedded sandstone (dxb). In each case the bioturbated zone is overlain by a similar thickness (30–35 ft, 9.0–10.5 m) of paleosols (ps) that display similar internal stratigraphic organization. See text for full discussion.

Late Paleozoic Gondwanan glaciation in Saudi Arabia

A

B

69

C

D

Figure 22. Core photographs showing aspects of the uppermost Unayzah A member in wells 24 and 21. (A) Intensely bioturbated sandstone in well 24 (Fig. 21A, 33–38 ft). (B) Pervasive development of silcrete in well 24 (Fig. 21A, 68–70 ft). (C) Intensely bioturbated sandstone in well 21 (Fig. 21B, 38–43 ft). (D) Pervasive development of silcrete in well 21 (Fig. 21B, 53 ft).

area (Fig. 1A), this unconformity separates the sandstones of the Unayzah from the overlying Khuff Formation. Although the Khuff Formation is widely recognized as a carbonate-dominated stratigraphic unit (Sharland et al., 2001; Vaslet et al., 2005), its lowermost deposits in many places are characterized by a series of alternating sandstones, shales, and thin carbonates that sit directly upon the pre-Khuff unconformity. Those deposits are described herein as the Basal Khuff Clastics member of the Khuff Formation. They include the unit identified by Senalp and Al-Duaiji (1995) and Vaslet et al. (2005) as the Ash Shiqqah member of the Khuff Formation. The depositional facies of the Basal Khuff Clastics member change in their character increasingly in a westward direction, as will be discussed below. Furthermore, the higher carbonate members of the Khuff Formation eventually overstep the Basal Khuff Clastics member, as well as the Unayzah and ultimately all older Paleozoic deposits, in the same direction. In extreme westerly locations, these upper Khuff carbonates rest directly upon Proterozoic rocks of the Arabian Shield.

Upward-Fining Units of the Basal Khuff Clastics In many places in the western part of the study area, and exemplified at wells 24 and 23 (Fig. 23A and Fig. 23B), the preKhuff unconformity is abruptly overlain by pebble conglomerates and very coarse–grained, pebbly sandstones that fine upward over several feet into argillaceous fine-grained sandstones and (ultimately) dark gray mudstones. The upward-fining sandstones and conglomerates have clasts that are variably subrounded to angular in nature and polymict in composition. Well 24, for example, displays an assemblage of angular detritus, including quartz, feldspar, jasper, granite, sandstone, siltstone, and mudstone. The mudstones that cap these upward-fining units in places display rooted textures in their upper part and commonly pass up into thin bioturbated sandstone and associated bioturbated mudrocks with calcareous concretions (e.g., well 23, Fig. 23B). The upward-fining character of these coarse-grained sediments in the western parts of the study area, and the lack of marine indicators except higher in the section, suggest that the

Figure 23. Selected core logs through the Basal Khuff Clastics member across the study area. (A) In well 24 the pre-Khuff unconformity (PKU) is overlain by a coarse breccia that fines upward into a highly carbonaceous paleosol. That in turn is overlain by ~5 ft (1.5 m) of thin-bedded marine bioturbated sandstones that pass up into gray mudstones (well 24). (B) In well 23 the pre-Khuff unconformity is abruptly overlain by upward-fining pebbly sandstone that passes upward into gray pedogenically altered mudrock. That soil is overlain by very thin, bioturbated fine-grained sandstone. (C) At well 15 the pre-Khuff unconformity is overlain by a mudrock-dominated interval that contains a few significant bioturbated sandstones. The mudrock facies show a cyclicity wherein fissile shales pass upward into more blocky and highly carbonaceous mudstones. (D) At well 17 the section directly above the preKhuff unconformity is characterized by a number of upward-fining cycles each of which passes into well-developed argillaceous and carbonaceous soils. These are superseded by a mudrock-dominated succession that contains bioturbated sandstones in its lower part, and within which can be seen a similar subtle cyclicity as at well 15. (E) The Basal Khuff Clastics member at well 21, in the eastern part of the study area, solely comprises shallow-marine depositional facies, including bioturbated sandstones and mudstones, and cross-bedded sandstones. See text for full discussion of these wells and their relationships.

Late Paleozoic Gondwanan glaciation in Saudi Arabia lowermost parts of the Basal Khuff Clastics member in those areas are characterized by fluvial deposition. The coarseness of the sandstones is indicative of a significant drop in base level (and/or uplift in the source area) pursuant to the creation of the pre-Khuff unconformity. Paleosols of the Basal Khuff Clastics In a number of the wells that have cored the Basal Khuff Clastics member, distinctive paleosols are recognized that are, significantly, quite different in character from those soils identified and earlier described from the top of the Unayzah A member. Thus in wells 24 (Fig. 23A, 35–40 ft) and 17 (Fig. 23D, 11–28 ft) fine- to very fine–grained, poorly sorted and very argillaceous and carbonaceous sandstones are identified low in the Basal Khuff Clastics section, overlying upward-fining coarser-grained sediment described above. The example from well 17 comprises four small-scale, upward-fining beds that display a high degree of vertical rooted texture throughout (Fig. 23D, 11–28 ft). These paleosols of the Basal Khuff Clastics member generally occur low in that stratal unit, directly subjacent to rocks displaying evidence of a marine influence in their deposition. This association leads to their interpretation as representing coastal marshlands in low-lying (estuarine) areas. As such they represent the first signs of the onset of the Khuff marine transgression. Marine Sandstone Deposits of the Basal Khuff Clastics In well 21 (Fig. 1A) the thick paleosols described above from the uppermost Unayzah A member (see Fig. 21B) are separated by ~6 ft (1.8 m) of non-recovery of core from a unit that is at least 21 ft (6.3 m) thick, and comprises sandstones and mudstones that are entirely different in character (Fig. 23E, 12–33 ft). Those sediments are assigned to the Basal Khuff Clastics member in this well. They comprise (1) interbedded dark gray, silty mudstones and thin (centimeter scale), heavily bioturbated fine-grained sandstones (Fig. 24A); (2) massive, intensively bioturbated sandstones up to 3 ft (0.9 m) thick; and (3) cross-bedded, fine-grained sandstones that occur in bedsets up to 5 ft (1.5 m) thick, with individual sets up to 0.5 ft (0.15 m) thick (Fig. 24B), and which only rarely display any bioturbation. In well 17 (Fig. 23D), the paleosols described above from the lowermost parts of the Basal Khuff Clastics member are overlain by a heavily bioturbated sandstone unit 6 ft (1.8 m) thick that contains large (cobble-sized) clasts of very fine–grained silica rock (Fig. 24C). Those clasts are probably derived from nearby nodular silcrete horizons within the uppermost Unayzah A paleosol (see earlier discussion). In well 15 the pre-Khuff unconformity is seen clearly to truncate eolian sandstones of the Unayzah A member (Fig. 23C). This unconformity is overlain by a very thin (1–2 cm) sandstone bed, above which the sequence is dominated by dark gray carbonaceous mudstones. Within those mudstones a very thin (0.3 ft, 0.09 m) bioturbated sandstone (Fig. 24D) occurs ~5 ft above the pre-Khuff unconformity in this well (Fig. 23C, 13–14 ft). In well

71

24 (Fig. 23A) the paleosol described above from the Basal Khuff Clastics is abruptly overlain by ~5 ft of thin-bedded (decimeter scale) coarse-grained sandstones that are burrowed throughout. The depositional facies described above from well 21 were laid down in a shallow-marine environment wherein the thin sets of cross-bedded sandstone represent offshore marine bars, and the various bioturbated facies are indicative of slightly deeper (or otherwise protected) conditions that allowed a proliferation of infaunal activity. The bioturbated sandstone in well 17 was likewise deposited in a marine setting. It is probable that the large cobble-sized clasts of siliceous rock were eroded by storms from silcretes in the uppermost Unayzah paleosols in nearby locations. The coarse-grained (i.e., sand-sized) marine detritus observed in wells 21 and 17 is not recognized in well 15. There, the first definitive indications of marine deposition are seen in the thin bioturbated sandstone that occurs ~5 ft above the pre-Khuff unconformity (Fig. 23C, 13–14 ft; Fig. 24D). Significantly, palynological data from that zone show an influx of abundant scolecodont debris that can be interpreted as indicating a significant marine influx at that point (J. Filatoff, 2004, personal commun.). It is clear that a similar marine depositional signature is also evident in well 24 in the western part of the study area (Fig. 23A, 41–47 ft). Mudrocks of the Basal Khuff Clastics In wells 17 (Fig. 23D) and 15 (Fig. 23C) the marine sands described above are overlain by up to 40 ft (12 m) of dark gray mudrocks. They locally contain thin-walled bivalve shells and display a number of subtle cycles, each of which is ~5 ft (1.5 m) thick. Within each cycle, there is a passage upward from fissile, gray, silty mudstone to blocky and crumbly, highly carbonaceous and rooted mudstone (Fig. 23C, 15–47 ft; Fig. 23D, 45–77 ft). These fine-grained depositional cycles are interpreted to represent successive periods of lagoonal infill in a marginal marine setting, each one terminating in a coastal swamp. Origin and Evolution of the Basal Khuff Clastics Member of the Khuff Formation It is clear from the foregoing not only that the Basal Khuff Clastics member represents an extremely heterogeneous assemblage of depositional facies but also that that variation is geographically constrained. Thus in the southeast of the study area around well 21 the lowermost Basal Khuff Clastics member is specifically characterized by significant deposits of shallowmarine sandstones. Somewhat to the north and west, around well 17, the lowermost Basal Khuff Clastics consist of very different sediments, comprising relatively thin, upward-fining sandstones that have been interpreted above to be highly argillaceous and carbonaceous paleosols. Those paleosols are believed to have formed in a low-lying coastal plain or estuarine marsh environment. Above these paleosols in well 17, bioturbated marine sandstones are recognized, although they are not as fully developed as

72

Melvin et al.

A

B D

C

Figure 24. Core photographs showing some characteristics of the marine sandstone facies of the Basal Khuff Clastics member. (A) Interbedded mudstones and fine-grained sandstones, displaying intense burrowing activity (well 21). (B) Cross-bedded sandstones of the offshore shallow-marine-bar environment (well 21). (See also Fig. 23E.) (C) Intensely bioturbated marine sandstone, containing a large clast of chert rock (Ch). The latter is considered to have been ripped out of nearby paleosol deposits by storms (well 17). (D) Thin, sharp-based bioturbated sandstone overlying gray mudrock (m). The sandstone is interpreted to be a relict shallow-marine storm deposit laid down below storm wave base (well 15). See text for further discussion.

those that occur in well 21; at well 15, shallow-marine sandstone is also recognized, but there it occurs only as a very thin bed (Fig. 24D) within an otherwise mudstone-dominated sequence. Those mudstones show a subtle cyclicity suggestive of repeated infill of lagoons in a coastal setting. In other words, the Basal Khuff Clastics in this location display a subtly less marine signature than equivalent deposits somewhat to the southeast at well 21. Even farther to the west in the study area, at well 23 the Basal Khuff Clastics member is represented in its lowermost part by

a fluvial deposit that is ultimately transgressed by a thin marine sandstone (Fig. 23B). At well 24 a basal fluvial breccia passes up into a paleosol that is in turn capped by ~5 ft (1.5 m) of thinbedded bioturbated sandstones (Fig. 23A, 32–47 ft). The clear distinction within the lowest parts of the Basal Khuff Clastics between shallow-marine depositional facies in the southeast and continental (fluvial) facies in the west demonstrates the gradual encroachment of the marine transgression from the southeast to the west in earliest Khuff times. The transgression

Late Paleozoic Gondwanan glaciation in Saudi Arabia reached its maximum expression with deposition of the various carbonate-evaporite cycles of the Khuff Formation, most recently described by Vaslet et al. (2005).

73

tiate the much closer association of the Unayzah B member with the unnamed middle Unayzah member in mid- to high southerly latitudes, from the near tropical setting of the Unayzah A member. These observations are discussed further below.

PALEOMAGNETIC STUDIES DISCUSSION In the course of conducting these stratigraphic studies of the Unayzah and lowermost Khuff Formations, an opportunity arose to carry out a paleomagnetic pilot study on core samples from two of the wells, wells 7 and 22. A total of 38 samples were collected under appropriately controlled, nonmagnetic conditions, and analyzed with a view to determining whether or not the paleolatitude at the time of their deposition could be identified. The samples were taken from rocks that were independently assigned to the Unayzah B member, the un-named middle Unayzah member, and the Unayzah A member. The results of this pilot study are presented in Figure 25. Although the samples from each stratigraphic unit show a range of values, it is clear that each unit has its own signature with respect to its interpreted paleolatitude and that there is no overlap in the respective data sets. Thus the samples from the Unayzah B member were deposited in paleolatitudes represented by a mean value ~75° S. The unnamed middle Unayzah member appears to have been laid down when the study area was at a (mean) paleolatitude of ~55° S. Deposition of the Unayzah A member did not occur until the area of investigation lay in paleolatitudes of ~27° S. These data clearly reflect the northward drift of the Arabian plate throughout the Early to Middle Permian, as was independently documented by previous authors (Beydoun, 1991; Al-Fares et al., 1998). They also appear to clearly differen-

Unayzah A

Younger

Unnamed middle Unayzah member

Unayzah B

0

30 60 Paleolatitude (degrees S)

90

Figure 25. Diagram based on paleomagnetic data, showing relative position with respect to paleolatitude of the Unayzah B member, the unnamed middle Unayzah member, and the Unayzah A member. These data sets individually display varying degrees of spread but nonetheless appear to show no overlap with each other. They illustrate clearly the northward drift of the Arabian plate from Asselian–earliest Sakmarian time (Unayzah B member) to Artinskian–Kungurian(?) times (Unayzah A member). The results were obtained from core plugs collected under controlled (nonmagnetic) conditions and were analyzed by E. Hailwood (CoreMagnetics).

Sharland et al. (2001) discussed how Tectonostratigraphic Megasequence TMS AP5 is the last megasequence of the Arabian plate to have been dominated by siliciclastic sediments (Unayzah Formation). Only the Basal Khuff Clastics member, overlying the pre-Khuff unconformity at the base of TMS AP6, represents any further siliciclastic deposition in Saudi Arabia prior to the northward drift of the Arabian plate into subtropical latitudes where carbonate and evaporite deposition predominated (cf. Beydoun, 1991; Al-Fares et al., 1998). TMS AP5 is marked at its base in Saudi Arabia by the Hercynian unconformity, which represents erosion associated with the mid-Carboniferous “Hercynian” tectonic inversion event (Al-Husseini, 2004). Those erosional processes were intensified with the almost coincident inception of the late Paleozoic Gondwanan glaciation in Arabia. The upper boundary of TMS AP5 is the pre-Khuff unconformity, dated as mid-Tatarian in Saudi Arabia by Stephenson and Filatoff (2000b). It represents erosion following the early Late Permian rifting associated with the creation of the Neotethys Ocean (cf. Bishop, 1995; Loosveld et al., 1996). It can thus be considered to be the “Break-up Unconformity” associated with the continental rifting and spreading of the Sanandaj-Sirjan and central Iran terranes from the Arabian plate (Sharland et al., 2001). Angiolini et al. (2003), however, believe that in the Sakmarian of Oman there are two tectono-eustatic transgressive events that are related to the end of the Gondwanan glaciation and the concomitant tectonic evolution during rifting and initial opening of the Neotethys Ocean. Those authors thus consider the “Break-up Unconformity” to be represented by an unconformity that is seen within TMS AP5, and which formed much earlier than the pre-Khuff unconformity (Angiolini et al., 2003). Clearly, the evolution of Tectonostratigraphic Megasequences TMS AP5 and TMS AP6 on the Arabian plate represents a complex interplay of tectonic, climatic, and eustatic relationships. An attempt is made below to discuss that evolution, in the context of the depositional and stratigraphic development of the various units described above within the Unayzah, as well as the Basal Khuff Clastics member of the Khuff Formation of Saudi Arabia. Inevitably, that discussion necessitates consideration of the equivalent stratal units elsewhere on the Arabian plate, specifically Oman. As has been discussed above, the late Paleozoic South Polar glaciation that affected so much of the Gondwanan continent is considered by Al-Husseini (2004) possibly to have commenced in Arabia as early as middle Carboniferous (late Visean–early Namurian or Serpukhovian) times, and to have persisted for 30–45 m.y. until the Early Permian (Sakmarian). This provided ample time for an unknown number of glacial advances, retreats,

74

Melvin et al.

and readvances to have taken place, both in Saudi Arabia and Oman. Such multiple events are implicitly recognized in Oman from palynological evidence within the lower Al Khlata (Osterloff et al., 2004a), which presumably reflects different interstadial depositional periods. In Saudi Arabia the palynology is not so forthcoming. It is tempting to think that the glacial readvances that have been postulated herein in relation to the shear zones within the Unayzah C member may be correlative by some means with the Oman section. This is clearly an important area for future study. The final retreat phase of the Gondwanan glaciation in Arabia (and by implication the onset of climatic amelioration) is manifest in the great variety of predominantly glaciolacustrine depositional facies recognized and described above in the Unayzah B member in Saudi Arabia as well as in the upper Al Khlata Formation (including the Rahab Shale) of Oman (Osterloff et al., 2004a). Paleomagnetic evidence from the current study appears to suggest that at this time Saudi Arabia lay in high southerly latitudes ~75° S. Notwithstanding minor readvances such as have been recognized in the local occurrence of stratabound push moraine deposits (Fig. 7), the overall character of the Unayzah B member across most of the study area illustrates ongoing glacial retreat, evident in the continuing filling and spilling over of an environment that was dominated by glacial lakes. This wholesale flooding of the postglacial landscape in Saudi Arabia is reflected in the clearly “transgressive” (deepening) nature of the glaciolacustrine depositional facies associations. In the western part of the study area the Unayzah B is characterized predominantly by glacially induced deformation features. These are considered to reflect the longer term presence of significant bodies of ice in these locations, and possibly even an element of long-lived alpine glaciation associated with highlands of the Al Batin Arch. The uppermost deposits of the Unayzah B member across most of the subsurface of eastern Saudi Arabia (and of the Rahab Shale in Oman, Osterloff et al., 2004a) provide clear sedimentological evidence of the maximum melt-out of the Gondwanan ice sheet. It remains to be seen whether or not these rocks represent the maximum flooding event of the entire Tectonostratigraphic Megasequence TMS AP5. That in turn demands a consideration of the extent to which the postglacial transgression is reflected by purely climatic (i.e., warming) controls, or by overriding plate tectonic factors. The only maximum flooding surface (MFS) that has hitherto been formally recognized in TMS AP5 was identified in Oman by Sharland et al. (2001) as MFS P10, which is represented by a bioturbated shale directly below the Haushi Limestone. This is the “maximum flooding shale” of Guit et al. (1995) and occurs within the postglacial lower Gharif member (see Fig. 2). Stephenson and Osterloff (2002) showed that MFS P10 is marked by the occurrence of the acritarch Ulanisphaeridium omanensis, and Stephenson et al. (2003) subsequently assigned it to the OSPZ3b Subbiozone. Angiolini et al. (2006) showed that the marine deposits of the overlying Haushi Limestone are Early Permian (Sakmarian) in age and correlative with the OSPZ3c Subbiozone

of Stephenson et al. (2003). Melvin and Sprague (2006) proposed that the unnamed middle Unayzah member red beds in Saudi Arabia can be correlated with the lower Gharif member in Oman and therefore may be considered generally correlative with the OSPZ3 Zone. This implied that in Saudi Arabia MFS P10 would somehow have equivalence within the wholly terrestrial deposits of the unnamed middle Unayzah member of Melvin and Sprague (2006). Ongoing palynological studies are providing encouraging results that support this original proposition of Melvin and Sprague (2006) that there is at least some stratigraphic equivalence between the unnamed middle Unayzah member and the lower Gharif member (N.P. Hooker, 2008, personal commun.). In Oman, Angiolini et al. (2003) considered the paleoclimatic and paleotectonic characteristics of the Saiwan Formation at outcrop (equivalent in the subsurface to the Haushi Limestone at the top of the lower Gharif member). They concluded from paleontological, paleoecological, and petrographic evidence that the opening of the Neotethys Ocean commenced north of Oman in about midSakmarian times (Angiolini et al., 2003). From the foregoing discussion of the sedimentary history of the Unayzah Formation, it would appear nonetheless that these proposed early stages of the Neotethys marine transgression did not find recognizable expression in the terrestrial unnamed middle Unayzah member in Saudi Arabia. The significant drainage event postulated earlier as marking the end of Unayzah B member times, and, most significantly, the end of the Gondwanan glaciation in Saudi Arabia, may well have been initiated by distant tectonic events related to an incipient breach as a first step toward opening of the Neotethys Ocean. In Oman, Guit et al. (1995) suggested the presence of a regional intra-Gharif unconformity related to a drop in relative sea level at the top of the lower Gharif member, and Osterloff et al. (2004b) also described how the base of the middle Gharif marks an overall regression. In Saudi Arabia the contact between the unnamed middle Unayzah member and the overlying Unayzah A member is considered to be a sequence boundary that is at least disconformable on the subcropping formation (Melvin and Heine, 2004). It heralds the onset of a major shift in paleoclimatic conditions, under which deposition of the Unayzah A member became dominated by arid to semiarid eolian and ephemeral stream sedimentation. That climatic change was probably brought about by the northward drift of the Arabian plate into lower southerly latitudes (see Beydoun, 1991; Al-Fares et al., 1998). Northward drift is clearly reflected in the paleomagnetic data presented in the current paper (Fig. 25). Those data demonstrate a large amount of northward migration toward tropical latitudes and can be inferred to represent a significantly long period of time between deposition of the unnamed middle Unayzah member and the Unayzah A member. This in turn would suggest that the stratigraphic boundary between these two members represents a hiatus of considerable significance. The earlier discussion of the Unayzah A member has highlighted the identification in places of a number of zones within this stratigraphic unit, each of which is related to a demonstrable rise in the paleo–water table and is correlatable over tens of

Late Paleozoic Gondwanan glaciation in Saudi Arabia kilometers, irrespective of the depositional facies within which it occurs. The possibility has already been mooted that this cyclical rise in the paleo–water table through the sediments of the Unayzah A member may be related to a pulsed rise in distant sea level, and as such may be a phantom expression of ongoing marine transgression. The recognition of an apparently widespread bioturbated (and therefore probably marine) sandstone zone close to the top of the Unayzah A member is highly relevant in this regard. The middle Gharif member in Oman is considered the probable stratigraphic equivalent of the Unayzah A member in Saudi Arabia (Stephenson et al., 2003). It has been described by Osterloff et al. (2004b) as consisting “largely of non-marine clastics, indicated by red continental paleosols and non-marine facies.” Significantly, the latter authors also note that a “rapid vertical interdigitating nature of marine to non-marine clastics” has been recorded in the middle Gharif member from cores from central Oman. It is possible that such alternations of facies in Oman reflect a more marineward expression of the cyclicity that has been seen and described herein from the predominantly continental sediments of the Unayzah A. That cyclicity culminates with the prominent bioturbated (and therefore probably marine) zone that occurs in some wells in the uppermost parts of the Unayzah A member (cf. Fig. 21), suggesting final breakthrough in latest Unayzah A time by marine waters into the study area of eastern and central Saudi Arabia. Thereafter the uppermost Unayzah A member displays widespread development of generally thick paleosols, suggesting a prolonged period of nondeposition, with no marine indicators. This is interpreted to be a reflection of the up-doming that has been postulated by Sharland et al. (2001) in relation to the thermal uplift that preceded continental rifting and spreading, and which culminated in the regional pre-Khuff unconformity. It was noted earlier that the palynological changeover between OSPZ4 and OSPZ5 is “probably the greatest recorded in the Permian palynological succession in Oman and Saudi Arabia” (Stephenson et al., 2003). This palynological change finds expression in Oman also in changes in depositional (environmental), chemostratigraphic, and mineralogical (heavy minerals) signatures between the middle Gharif and upper Gharif members (Stephenson et al., 2003; Osterloff et al., 2004b). It is thus strongly implicit that a significant depositional hiatus exists between these two members. Osterloff et al. (2004b) noted how the upper Gharif member consists of four cycles (namely cycles 5–8), the uppermost of which displays evidence of “tidal/ estuarine” environments. The preceding discussion has demonstrated a similar change in Saudi Arabia from the widespread arid continental setting in the Unayzah A member to the mixed fluvial to shallow-marine setting of the Basal Khuff Clastics member. Furthermore, the Basal Khuff Clastics member exhibits a distinctive and different heavy mineral assemblage relative to the underlying sandstones of the Unayzah A (R.W.O’B. Knox, 2003, written commun.). It is thus evident that a significant depositional changeover exists between the Unayzah A member and the Basal

75

Khuff Clastics member: the boundary for that changeover is manifest as the pre-Khuff unconformity. Osterloff et al. (2004b, p. 115) postulated some equivalence in the sequence development of the uppermost cycle of the upper Gharif member in Oman compared with the upper Basal Khuff Clastics member, while stressing that no age equivalence is implied. Indeed, Stephenson (2006) recently clarified how in Oman the upper Gharif member is assigned to palynozone OSPZ5 of Stephenson et al. (2003), whereas the Basal Khuff Clastics member in Saudi Arabia is assigned definitively to the younger palynozone OSPZ6 (see Fig. 2). It has been shown earlier in this paper how in Saudi Arabia the lowermost part of the Basal Khuff Clastics member displays very different depositional facies from place to place, ranging from shallow marine in the southeastern part of the study area to fluvial deposits in more westerly locations (Fig. 23). This geographical demarcation of facies clearly reflects the onset of fully marine transgression to the south and east of the study area. That transgression progressed steadily westward, as is reflected in (1) the diachroneity of deposits from the upper Gharif member to the Basal Khuff Clastics, (2) the heterogeneity of the facies tracts within the Basal Khuff Clastics, and, ultimately, (3) by the superposition of Khuff carbonates over Proterozoic basement rocks in western central Saudi Arabia, as discussed previously. The pre-Khuff unconformity marks the top of the Tectonostratigraphic Megasequence TMS AP5 of Sharland et al. (2001). Those authors describe how “progressive thermal uplift (the precursor to continental rifting and spreading) is interpreted to have occurred … culminating in the regional ‘pre-Khuff unconformity’ at the top of this megasequence” (Sharland et al., 2001, p. 85). The change in Oman from continental or marginal-marine deposits of the Gharif Formation to fully marine Khuff Formation carbonates in the Wordian (Broutin et al., 1995) (equivalent to early Kazanian) is interpreted “to reflect marine flooding following thermal collapse of the new Arabian plate passive margin with Neotethys” (Sharland et al., 2001, p. 89). Those authors observe that “in north Arabia the transition probably occurs slightly later, at the base of the Tatarian (equivalent to middle Capitanian) . . . possibly indicating a very rapid ‘unzipping’ effect from southeast to northwest.” The present paper presents data that fully support these concepts. Furthermore, it seems logical that in lithostratigraphic terms the upper Gharif in Oman is indeed an older equivalent of the Basal Khuff Clastics member in Saudi Arabia and marks the onset in Oman of the major transgression that ultimately took place across Arabia, pursuant to the creation of the Neotethys Ocean. CONCLUSIONS 1. In middle Carboniferous times the Arabian plate was subjected to an episode of major tectonic activity that was followed by a period of erosion and the creation of the Hercynian unconformity. These events were closely followed by the inception of the late Paleozoic Gondwanan glaciation in Arabia, which

76

Melvin et al.

resulted in deposition of the Unayzah Formation upon the Hercynian unconformity. 2. The Unayzah C member, where present, sits directly upon the Hercynian unconformity. It varies greatly in thickness and comprises quartzose sandstones that were laid down during multiple retreat phases of the ice sheets upon glaciofluvial outwash braided plains. During subsequent intervening readvances of the ice sheets, these outwash deposits were overridden, cannibalized, and pushed into stacked piles of glacially tectonized (push moraine) deposits, manifest as relatively undeformed sandstones separated by distinct shear zones. The top of the Unayzah C member is an unconformity and represents the final subglacial surface of the Gondwanan ice sheet in Saudi Arabia. 3. The terminal melt-out phase of the Gondwanan glaciation is identified in the various facies of the Unayzah B member, which was deposited in high southerly latitudes, ~75° S, based on paleomagnetic data. These Unayzah B sediments comprise a number of different ice-proximal to ice-distal glaciolacustrine depositional facies. Facies associations consistently demonstrate progressive retreat of the ice and a concomitant sustained increase in flooding by meltwaters of an environment that was characterized by an abundance of glacial lakes. The top of the Unayzah B member can be taken to represent the climatic maximum flooding surface related to the melting of the Gondwanan ice sheets in Saudi Arabia and is considered equivalent to the top of the Rahab Shale in Oman. There is evidence to suggest a possible center of relatively prolonged upland glaciation in the western part of the study area, associated with the Al Batin Arch. 4. The unnamed middle Unayzah member is separated from the underlying Unayzah B member by a sharp stratigraphic boundary, considered to be a regional disconformity, which is interpreted to represent a significant drainage event that marked the end of the glaciation in Saudi Arabia. Paleomagnetic data suggest that the unnamed middle Unayzah member was laid down in paleolatitudes ~55° S. This member comprises red-brown, alluvial floodplain sandy siltstones, within which occur various isolated sandstone facies representing fluvial and possibly coldclimate eolian deposits. It is commonly capped by a paleosol horizon that represents a terrestrial highstand deposit, believed to be equivalent to the marine Haushi Limestone highstand deposit at the top of the lower Gharif member in Oman. That those paleosols represent a prolonged period of limited deposition or nondeposition is supported by the paleomagnetic evidence. The paleosols are therefore considered to mark a significant disconformity between the unnamed middle Unayzah member and the overlying Unayzah A member. 5. The lower Unayzah A member was deposited in widespread ephemeral (playa) lakes, whereas the upper Unayzah A sediments were deposited predominantly within arid to semiarid eolian dune fields and associated interdune deposits as well as within ephemeral stream systems and minor playa lakes. In places the eolian deposits display an internal stratigraphy that is related to cyclical fluctuations in the paleo–water table, and which is correlatable within and between significant facies tracts in this

stratal unit. Those rises in the paleo–water table possibly reflect ongoing pulsed phases of a distant marine transgression, an idea that is given support by the recognition of a heavily bioturbated sandstone zone at the very top of the Unayzah A member in two wells at either end of the study area. That zone is thought to represent final breakthrough of the transgressing marine waters. In most places the very highest parts of the Unayzah A display thick, well-developed paleosol horizons. These represent an extended period of minimal deposition and are considered to reflect the long period of uplift that was related to thermal doming prior to rifting and collapse that led to the creation of the pre-Khuff unconformity. The Unayzah A member is considered equivalent to the middle Gharif member of Oman. 6. The Unayzah A member is truncated by the pre-Khuff unconformity, which marks the upper boundary of Tectonostratigraphic Megasequence TMS AP5 in Arabia. It is overlain by sandstones and shales of the Basal Khuff Clastics member of the lowermost Khuff Formation. The depositional facies of the Basal Khuff Clastics are characterized in southeastern areas by shallowmarine sandstones and shales that pass westward into more terrestrial facies (fluvial), reflecting the onset of a major westwarddirected marine transgression. That transgression reached its fullest development with the widespread marine carbonates of the Khuff Formation. It is an expression of the successful opening of the Neotethys Ocean, and as such represents a tectonically related major flooding event. 7. The late Paleozoic Gondwanan glaciation per se in Saudi Arabia is represented in part by the Hercynian unconformity, as well as by the deposition and deformation of the Unayzah C member. Its final retreat from Saudi Arabia is seen in the rocks of the Unayzah B member, the top of which represents the maximum glacial melt-out. This is the maximum expression of postglacial transgression that can be ascribed solely to climatic (warming) factors. The end of the glaciation in Saudi Arabia is marked by an inferred dramatic drainage event, manifest in the disconformity that separates the glaciolacustrine Unayzah B member from the terrestrial deposits of the unnamed middle Unayzah member. Thereafter, the rocks are continental in nature, and their relationships with distant marine transgression (for which there is tantalizing evidence) must necessarily incorporate the likelihood of significant tectonic influences. These tectonic influences are ultimately manifest in the pre-Khuff unconformity. Above that unconformity the rocks of the Basal Khuff Clastics member of the Khuff Formation reflect the marine transgressive flooding associated with the opening of the Neotethys Ocean. ACKNOWLEDGMENTS We acknowledge the Saudi Arabian Ministry of Petroleum and Mineral Resources and the Saudi Arabian Oil Company (Saudi Aramco) for granting permission to publish this paper. The evolution of our thoughts on the Unayzah Formation has benefited from many discussions with many colleagues at Saudi Aramco. In particular we wish to acknowledge Mark Prudden,

Appendix: Core Log Legend Planar lamination

Mud cracks

Cross lamination Burrows Trough crosslamination Rooted horizons Low-angle lamination Current ripples

Pedogenic “slickoliths” Floating sand grains

Climbing ripples

Mud clasts

Coarse sand lamination

Pebbles, cobbles

Wispy lamination

Chert

Adhesion ripple

Carbonaceous layer

Graded bedding

Upward-shoaling interval

Dish structure with elutriation pillars Soft sediment deformation

Stylolites

Flame structures

Vertical fractures

Soft sediment loading

Low-angle shear

High-angle stylolites

Grain size G S Z M Gravel Sand Silt Mud/clay Figure A1. This figure provides a legend to which reference should be made for all figures throughout the paper comprising sedimentological core logs.

78

Melvin et al.

Kent Norton, and Roger Price of Saudi Aramco’s Exploration Organization for much stimulating debate. John Filatoff, Nigel Hooker, and Merrell Miller provided biostratigraphic information, and Stephen Franks shared petrographic data regarding these enigmatic rocks. Peter Sharland also provided valuable insights on the occurrence of late-glacial drainage events during the Pleistocene glaciation. Ernie Hailwood of CoreMagnetics provided the analyses of the samples selected for the paleomagnetic (paleolatitude) studies that are referenced herein. We nonetheless accept sole responsibility for the ideas presented in this work. We thank George Grover and the Saudi Aramco Publications Review Committee for their time and effort, and for insightful comments in reviewing this paper. Kathleen Haughney of the Saudi Aramco Exploration and Producing Information Center was her usual cheerful self in chasing down some of the more elusive reference material. Gene Cousart of Aramco’s Cartography Department demonstrated outstanding commitment and professionalism in his approach to producing the figures that are presented in this paper. Ali Al-Zahrani and Hadi Al-Uraij are thanked for their tireless efforts in arranging for the layout of many thousands of feet of Unayzah core in the Saudi Aramco Core Storage facility. REFERENCES CITED Ahlbrandt, T.S., and Andrews, S., 1978, Distinctive sedimentary features of cold-climate eolian deposits, North Park, Colorado: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 25, p. 327–351, doi: 10.1016/0031 -0182(78)90048-2. Aitken, J.F., Clark, N.D., Osterloff, P.L., Penney, R.A., and Mohiuddin, U., 2004, Regional core-based sedimentological review of the glacially influenced Permo-Carboniferous Al-Khlata Formation, South Oman Salt Basin, Oman: GeoArabia, Abstract, v. 9, no. 1, p. 16. Al-Belushi, J.D., Glennie, K.W., and Williams, B.P.J., 1996, Permo-Carboniferous glaciogenic Al Khlata Formation, Oman: A new hypothesis for origin of its glaciation: GeoArabia, v. 1, p. 389–404. Al-Fares, A.A., Bouman, M., and Jeans, P., 1998, A new look at the middle to Lower Cretaceous stratigraphy, offshore Kuwait: GeoArabia, v. 3, p. 543–560. Al-Hajri, S.A., and Owens, B., eds., 2000, Stratigraphic Palynology of the Palaeozoic of Saudi Arabia: GeoArabia Special Publication 1, Gulf PetroLink, Bahrain, 231 p. Al-Husseini, M.I., 2004, Pre-Unayzah unconformity, Saudi Arabia, in AlHusseini, M.I., ed., Carboniferous, Permian and Early Triassic Arabian Stratigraphy: GeoArabia Special Publication 3, Gulf PetroLink, Bahrain, p. 15–59. Angiolini, L., Balini, M.E., Garzanti, E., Nicora, A., and Tintori, A., 2003, Gondwanan deglaciation and opening of Neotethys: The Al-Khlata and Saiwan Formations of Interior Oman: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 196, p. 99–123, doi: 10.1016/S0031-0182(03)00315-8. Angiolini, L., Stephenson, M.H., and Leven, E.J., 2006, Correlation of the Lower Permian surface Saiwan Formation and subsurface Haushi limestone, Central Oman: GeoArabia, v. 11, p. 17–38. Bates, R.L., and Jackson, J.A., 1980, Glossary of Geology (2nd edition): Falls Church, Virginia, American Geological Institute, 751 p. Bennett, M.R., 2001, The morphology, structural evolution and significance of push moraines: Earth-Science Reviews, v. 53, p. 197–236, doi: 10.1016/ S0012-8252(00)00039-8. Beydoun, Z.R., 1991, Arabian Plate Hydrocarbon Geology and Potential—A Plate Tectonic Approach: American Association of Petroleum Geologists Studies in Geology 33, 77 p. Bishop, R.S., 1995, Maturation history of the Lower Palaeozoic of the Eastern Arabian Platform, in Al-Husseini, M.I., ed., Middle East Petroleum Geosciences GEO94: Gulf PetroLink, Bahrain, v. 1, p. 180–189.

Boulton, G.S., 1986, Push moraines and glacier-contact fans in marine and terrestrial environments: Sedimentology, v. 33, p. 677–698, doi: 10.1111/j.1365 -3091.1986.tb01969.x. Boulton, G.S., van der Meer, J.J.M., Beets, D.J., Hart, J.K., and Ruegg, G.H.J., 1999, The sedimentary and structural evolution of a recent push moraine complex: Holmstrombreen, Spitsbergen: Quaternary Science Reviews, v. 18, p. 339–371, doi: 10.1016/S0277-3791(98)00068-7. Braakman, J.H., Levell, B.K., Martin, J.H., Potter, T.L., and van Vliet, A., 1982, Late Palaeozoic Gondwana glaciation in Oman: Nature, v. 299, p. 48–50, doi: 10.1038/299048a0. Broutin, J., Roger, J., Platel, J.-P., Angiolini, L., Baud, A., Bucher, H., Marcoux, J., and Al-Hashmi, H., 1995, The Permian Pangea. Phytographic implications of new palaeontological discoveries in Oman (Arabian Peninsula): Comptes Rendues Academie Scientifique Paris, t. 321, Serie IIa, p. 1069–1086. Eberth, D.A., and Miall, A.D., 1991, Stratigraphy, sedimentology and evolution of a vertebrate-bearing, braided to anastomosed fluvial system, Cutler Formation (Permian-Pennsylvanian), north-central New Mexico: Sedimentary Geology, v. 72, p. 225–252, doi: 10.1016/0037-0738(91)90013-4. Evans, D.S., Bahabri, B.H., and Al-Otaibi, A.M., 1997, Stratigraphic trap in the Permian Unayzah Formation, central Saudi Arabia: GeoArabia, v. 2, p. 259–278. Eyles, N., 1993, Earth’s Glacial Record and Its Tectonic Setting: Earth-Science Reviews, v. 35, 248 p., doi: 10.1016/0012-8252(93)90002-O. Faqira, M., Rademakers, M., and Afifi, A.M., 2009, New insights into the Hercynian Orogeny, and their implications for the Paleozoic hydrocarbon system in the Arabian Plate: GeoArabia, v. 14, p. 199–228. Ferguson, G.S., and Chambers, T.M., 1991, Subsurface stratigraphy, depositional history, and reservoir development of the Early-to-Late Permian Unayzah Formation in central Saudi Arabia: Bahrain, Proceedings of the Society of Petroleum Engineers (SPE) Middle East Oil Show, 7th, SPE Paper 21394, p. 487–496. French, H.M., 1996, The Periglacial Environment (2nd edition): Harlow, UK, Addison Wesley, Longman, 341 p. Fryberger, S.G., and Schenk, C.J., 1988, Pin stripe lamination: A distinctive feature of modern and ancient eolian sediments: Sedimentary Geology, v. 55, p. 1–15, doi: 10.1016/0037-0738(88)90087-5. Fryberger, S.G., Ahlbrandt, T.S., and Andrews, S., 1979, Origin, sedimentary features, and significance of low-angle eolian “sand sheet” deposits, Great Sand Dunes National Monument and vicinity, Colorado: Journal of Sedimentary Petrology, v. 49, p. 733–746. Fryberger, S.G., Al-Sari, A.M., and Clisham, T.J., 1983, Eolian dune, interdune, sand sheet, and siliciclastic sabkha sediments of an offshore prograding sand sea, Dhahran area, Saudi Arabia: American Association of Petroleum Geologists Bulletin, v. 67, p. 280–312. Fryberger, S.G., Schenk, C.J., and Krystinik, L.F., 1988, Stokes surfaces and the effects of near-surface groundwater-table on aeolian deposition: Sedimentology, v. 35, p. 21–41, doi: 10.1111/j.1365-3091.1988.tb00903.x. Guit, F.A., Al-Lawati, M.H., and Nederlof, P.J.R., 1995, Seeking new potential in the Early–Late Permian Gharif play, west central Oman, in Al-Husseini, M.I., ed., Middle East Petroleum Geosciences GEO94: Gulf PetroLink, Bahrain, v. 2, p. 447–462. Gustavson, T.C., Ashley, G.M., and Boothroyd, J.C., 1975, Depositional sequences in glaciolacustrine deltas, in Jopling, A.V., and McDonald, B.C., eds., Glaciofluvial and Glaciolacustrine Sedimentation: Society of Economic Paleontologists and Mineralogists Special Publication 23, p. 264–280. Helal, A.H., 1964, On the occurrence and stratigraphic position of PermoCarboniferous tillites in Saudi-Arabia: Geologische Rundschau, v. 54, p. 193–207, doi: 10.1007/BF01821178. Hughes Clarke, M.W., 1988, Stratigraphy and rock unit nomenclature in the oil producing area of interior Oman: Journal of Petroleum Geology, v. 11, p. 5–60, doi: 10.1111/j.1747-5457.1988.tb00800.x. Hunter, R.E., 1980, Quasi-planar adhesion stratification—An eolian structure formed in wet sand: Journal of Sedimentary Petrology, v. 50, p. 263–266. Husseini, M.I., 1992, Upper Palaeozoic tectono-sedimentary evolution of the Arabian and adjoining plates: Journal of the Geological Society [London], v. 149, p. 419–429, doi: 10.1144/gsjgs.149.3.0419. Jopling, A.V., and Walker, R.G., 1968, Morphology and origin of ripple-drift cross-lamination with examples from the Pleistocene of Massachusetts: Journal of Sedimentary Petrology, v. 38, p. 971–984.

Late Paleozoic Gondwanan glaciation in Saudi Arabia Kellogg, K.S., Janjou, D., Minoux, L., and Fourniguet, J., 1986, Explanatory notes to the geologic map of the Wadi Tathlith Quadrangle, Kingdom of Saudi Arabia: Deputy Minister for Mineral Resources, Ministry of Petroleum and Mineral Resources, Kingdom of Saudi Arabia, 27 p., Geoscience Map GM-103C, scale 1:250,000, sheet 20G. Kocurek, G., and Dott, R.H., Jr., 1981, Distinctions and uses of stratification types in the interpretation of eolian sand: Journal of Sedimentary Petrology, v. 51, p. 579–595. Kocurek, G., and Fielder, G., 1982, Adhesion structures: Journal of Sedimentary Petrology, v. 52, p. 1229–1241. Konert, G., Al-Afifi, A.M., Al-Hajri, S.A., and Droste, H.J., 2001, Paleozoic stratigraphy and hydrocarbon habitat of the Arabian Plate: GeoArabia, v. 6, p. 407–442. Kruck, W., and Thiele, J., 1983, Late Palaeozoic glacial deposits in the Yemen Arab Republic: Geologisches Jahrbuch, Reihe B, Regionale Geologie Ausland, v. 46, p. 3–29. Levell, B.K., Braakman, J.H., and Rutten, K.W., 1988, Oil-bearing sediments of Gondwana glaciation in Oman: American Association of Petroleum Geologists Bulletin, v. 72, p. 775–796. Loosveld, R.J.H., Bell, A., and Terken, J.J.M., 1996, The tectonic evolution of interior Oman: GeoArabia, v. 1, p. 28–51. Lowe, D.R., and LoPiccolo, R.D., 1974, The characteristics and origins of dish and pillar structures: Journal of Sedimentary Petrology, v. 44, p. 484–501. Mangerud, J., Jakobsson, M., Alexanderson, H., Astakhov, V., Clarke, G.K.C., Henriksen, M., Hjort, C., Krinner, G., Lunkka, J.-P., Moller, P., Murray, A., Nikolskaya, O., Saarnisto, M., and Svendsen, J.I., 2004, Ice-dammed lakes and rerouting of the drainage of northern Eurasia during the Last Glaciation: Quaternary Science Reviews, v. 23, p. 1313–1332, doi: 10.1016/ j.quascirev.2003.12.009. Martini, I.P., and Brookfield, M.E., 1995, Sequence analysis of Upper Pleistocene (Wisconsinan) glaciolacustrine deposits of the north-shore bluffs of Lake Ontario, Canada: Journal of Sedimentary Research, v. 65, p. 388–400. McClure, H.A., 1980, Permian–Carboniferous glaciation in the Arabian peninsula: Geological Society of America Bulletin, v. 91, p. 707–712, doi: 10 .1130/0016-7606(1980)912.0.CO;2. McClure, H.A., and Young, G.M., 1981, Late Paleozoic glaciation in the Arabian peninsula, in Hambrey, M.J., and Harland, W.B., eds., Earth’s Pre-Pleistocene Glacial Record: Cambridge, UK, Cambridge University Press, p. 275–277. McClure, H.A., Hussey, E.M., and Kaill, I.J., 1988, Permian-Carboniferous glacial deposits in southern Saudi Arabia: Geologisches Jahrbuch, Reihe B, Regionale Geologie Ausland, v. 68, p. 3–31. McGillivray, J.G., and Husseini, M.I., 1992, The Palaeozoic petroleum geology of central Arabia: American Association of Petroleum Geologists Bulletin, v. 76, p. 1473–1490. Melvin, J., and Heine, C.J., 2004, Sequence stratigraphy of an eolian gas sand: Layering in the Permian Unayzah-A reservoir at south Ghawar, Eastern Saudi Arabia: GeoArabia, Abstract, v. 9, p. 103. Melvin, J., and Sprague, R.A., 2006, Advances in Arabian stratigraphy: Origin and stratigraphic architecture of glaciogenic sediments in PermianCarboniferous lower Unayzah sandstones, eastern central Saudi Arabia: GeoArabia, v. 11, p. 105–152. Melvin, J., Sprague, R.A., and Heine, C.J., 2005, Diamictites to eolianites: Carboniferous–Permian climate change seen in subsurface cores from the Unayzah Formation, east-central Saudi Arabia, in Reinson, G.E., Hills, D., and Eliuk, L., eds., 2005 CSPG Core Conference Papers and Extended Abstracts CD: Calgary, Canadian Society of Petroleum Geologists, p. 237–282. Melvin, J., Wallick, B.P., and Heine, C.J., 2010, Advances in Arabian stratigraphy: Allostratigraphic layering related to paleo–water table fluctuations in eolian sandstones of the Permian Unayzah A reservoir, South Haradh, Saudi Arabia: GeoArabia, v. 15, p. 55–86. Moncrieff, A.C.M., and Hambrey, M.J., 1990, Marginal-marine glacial sedimentation in the late Precambrian succession of East Greenland, in Dowdeswell, J.A., and Scourse, J.D., eds., Glacimarine Environments: Processes and Sediments: Geological Society [London] Special Publication 53, p. 387–410. Mountney, N.P., and Jagger, A., 2004, Stratigraphic evolution of an aeolian erg margin system: The Permian Cedar Mesa Sandstone, SE Utah, USA: Sedimentology, v. 51, p. 713–743, doi: 10.1111/j.1365-3091.2004.00646.x.

79

Osterloff, P., Penney, R., Aitken, J., Clark, N., and Husseini, M.I., 2004a, Depositional sequences of the Al Khlata Formation, subsurface Interior Oman, in Al-Husseini, M.I., ed., Carboniferous, Permian and Early Triassic Arabian Stratigraphy: GeoArabia Special Publication 3, Gulf PetroLink, Bahrain, p. 61–81. Osterloff, P., Al-Harthy, A., Penney, R., Spaak, P., Williams, G., Al-Zadjali, F., Jones, N., Knox, R., Stephenson, M., Oliver, G., and Al-Husseini, M.I., 2004b, Depositional sequences of the Gharif and Khuff Formations, subsurface Interior Oman, in Al-Husseini, M.I., ed., Carboniferous, Permian and Early Triassic Arabian Stratigraphy: GeoArabia Special Publication 3, Gulf PetroLink, Bahrain, p. 83–147. Ovenshine, A.T., 1970, Observations of iceberg rafting in Glacier Bay, Alaska, and the identification of ancient ice-rafted deposits: Geological Society of America Bulletin, v. 81, p. 891–894, doi: 10.1130/0016-7606(1970)81 [891:OOIRIG]2.0.CO;2. Roland, N.W., 1979, Geological Map of the Arab Yemen Republic, Sheet Sadah, 1:250,000: Hanover, Germany, Federal Institute of Geoscience and Natural Resources. Ruegg, G.H.J., 1983, Periglacial eolian evenly laminated sandy deposits in the Late Pleistocene of NW Europe, a facies unrecorded in modern sedimentological handbooks, in Brookfield, M.E., and Ahlbrandt, T.S., eds., Eolian Sediments and Processes: New York, Elsevier, Developments in Sedimentology 38, p. 455–482. Rust, B.R., and Romanelli, R., 1975, Late Quaternary subaqueous outwash deposits near Ottawa, Canada, in Jopling, A.V., and McDonald, B.C., eds., Glaciofluvial and Glaciolacustrine Sedimentation: Society of Economic Paleontologists and Mineralogists Special Publication 23, p. 177–192. Senalp, M., and Al-Duaiji, A., 1995, Stratigraphy and sedimentation of the Unayzah reservoir, central Saudi Arabia, in Al-Husseini, M.I., ed., Middle East Petroleum Geosciences Conference, GEO’94: Gulf PetroLink, Bahrain, v. 2, p. 837–847. Sharland, P.R., Archer, R., Casey, D.M., Davies, R.B., Hall, S.H., Heward, A.P., Horbury, A.D., and Simmons, M.D., 2001, Arabian Plate Sequence Stratigraphy: GeoArabia Special Publication 2, Gulf PetroLink, Bahrain, 371 p., with 3 charts. Stampfli, G.M., and Borel, G.D., 2004, The TRANSMED transects in space and time: Constraints on the paleotectonic evolution of the Mediterranean domain, in Cavazza, W., Roure, F., Spakman, W., Stampfli, G.M., and Ziegler, P.A., eds., The TRANSMED Atlas—The Mediterranean Region from Crust to Mantle: Berlin, Heidelberg, Springer, p. 53–80. Stephenson, M.H., 1998, Preliminary correlation of palynological assemblages from Oman with the Granulatisporites confluens Oppel Zone of the Grant Formation (Lower Permian), Canning Basin, Western Australia: Journal of African Earth Sciences, v. 26, p. 521–526, doi: 10.1016/S0899 -5362(98)00030-X. Stephenson, M.H., 2004, Early Permian spores from Oman and Saudi Arabia, in Al-Husseini, M.I., ed., Carboniferous, Permian and Early Triassic Arabian Stratigraphy: GeoArabia Special Publication 3, Gulf PetroLink, Bahrain, p. 185–215. Stephenson, M.H., 2006, Stratigraphic note: Update of the standard Arabian Permian palynological biozonation; definition and description of OSPZ5 and 6: GeoArabia, v. 11, p. 173–178. Stephenson, M.H., and Filatoff, J., 2000a, Correlation of CarboniferousPermian palynological assemblages from Oman and Saudi Arabia, in AlHajri, S., and Owens, B., eds., Stratigraphic Palynology of the Palaeozoic of Saudi Arabia: GeoArabia Special Publication 1, Gulf PetroLink, Bahrain, p. 168–191. Stephenson, M.H., and Filatoff, J., 2000b, Description and correlation of Late Permian palynological assemblages from the Khuff Formation, Saudi Arabia and evidence for the duration of the pre-Khuff hiatus, in Al-Hajri, S., and Owens, B., eds., Stratigraphic Palynology of the Palaeozoic of Saudi Arabia: GeoArabia Special Publication 1, Gulf PetroLink, Bahrain, p. 192–215. Stephenson, M.H., and Osterloff, P.L., 2002, Palynology of the deglaciation sequence represented by the Lower Permian Rahab and Lower Gharif members, Oman: American Association of Stratigraphic Palynologists Contribution Series, v. 40, p. 1–32. Stephenson, M.H., Osterloff, P.L., and Filatoff, J., 2003, Palynological biozonation of the Permian of Oman and Saudi Arabia: Progress and challenges: GeoArabia, v. 8, p. 467–496. Stokes, W.L., 1968, Multiple parallel-truncation bedding planes—A feature of wind-deposited sandstone formations: Journal of Sedimentary Petrology, v. 38, p. 510–515.

80

Melvin et al.

Teller, J.T., Leverington, D.W., and Mann, J.D., 2002, Freshwater outbursts to the oceans from glacial Lake Agassiz and their role in climate change during the last deglaciation: Quaternary Science Reviews, v. 21, p. 879–887, doi: 10.1016/S0277-3791(01)00145-7. Van der Wateren, D.F.M., 1985, A model of glacial tectonics, applied to the icepushed ridges in the Central Netherlands: Geological Society of Denmark Bulletin, v. 34, p. 55–74. Van der Wateren, D.F.M., 1987, Structural geology and sedimentology of the Dammer Berg push moraine, FRG, in Meer, J.J.M., ed., Tills and Glaciotectonics: Rotterdam, A.A. Balkema, p. 157–182. Van der Wateren, D.F.M., 1994, Proglacial subaquatic outwash fan and delta sediments in push moraines—Indicators of subglacial meltwater activity: Sedimentary Geology, v. 91, p. 145–172, doi: 10.1016/0037-0738(94 )90127-9. Van der Wateren, D.F.M., 1995, Structural geology and sedimentology of push moraines: Processes of soft sediment deformation in a glacial environment and the distribution of glaciotectonic styles: Mededelingen Rijks Geologische, v. 54, 167 p.

Vaslet, D., Le Nindre, Y.-M., Vachard, D., Broutin, J., Crasquin-Soleau, S., Berthelin, M., Gaillot, J., Halawani, M., and Al-Husseini, M.I., 2005, The Permian-Triassic Khuff Formation of central Saudi Arabia: GeoArabia, v. 10, p. 77–134. Visser, J.N.J., 1982, Upper Carboniferous glacial sedimentation in the Karoo basin near Prieska, South Africa: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 38, p. 63–92, doi: 10.1016/0031-0182(82)90065-7. Wender, L.E., Bryant, J.W., Dickens, M.F., Neville, A.S., and Al-Moqbel, A.M., 1998, Paleozoic (pre-Khuff) hydrocarbon geology of the Ghawar area, eastern Saudi Arabia: GeoArabia, v. 3, p. 273–302. Williams, B.P.J., Wild, E.K., and Suttill, R.J., 1985, Paraglacial aeolianites: Potential new hydrocarbon reservoirs, Gidgealpa Group, southern Cooper Basin: Australian Petroleum and Exploration Association Journal, v. 25, p. 291–310.

MANUSCRIPT ACCEPTED BY THE SOCIETY 21 DECEMBER 2009

Printed in the USA

The Geological Society of America Special Paper 468 2010

Environmental and paleogeographic implications of glaciotectonic deformation of glaciomarine deposits within Permian strata of the Metschel Tillite, southern Victoria Land, Antarctica John L. Isbell* Department of Geosciences, University of Wisconsin–Milwaukee, Milwaukee, Wisconsin 53021, USA

ABSTRACT Popular reconstructions of late Paleozoic glaciation depict a single massive ice sheet centered over Victoria Land and extending outward over much of Gondwana. This view is untenable, as interpretations presented here indicate that glaciogenic strata in southern Victoria Land were deposited in a glaciomarine setting, and that ice entered the area from at least two different ice centers on opposite sides of the depositional basin. Reports from other areas also reveal that multiple ice sheets, ice caps, and alpine glaciers diachronously waxed and waned across Gondwana during the Carboniferous and Permian. Glaciogenic rocks of the Lower Permian Metschel Tillite contain glaciotectonic structures and glaciogenic deposits that include (1) sheared diamictites, (2) thrust sheets, (3) massive and stratified diamictites, and (4) sheet sandstones. These features formed as subglacial deforming beds, thrust moraines at glacial termini, and as glaciomarine deposits associated with temperate glaciers. A glaciomarine setting, rather than a glaciolacustrine setting, is suggested, owing to the abundance of meltwater plume deposits. A wedge-shaped sandstone body at the base of the overlying Weller Coal Measures was deposited as a grounding-line fan. Results of this study imply deposition in ice-marginal glaciomarine settings from ice radiating out of multiple glacial centers. These findings are significant because multiple glaciers, covering a given area, contain considerably less ice volume than a single massive ice sheet. Therefore, the waxing and waning of multiple ice masses during the late Paleozoic would have influenced global climate and eustatic sea level much differently than would have a single massive Gondwanan ice sheet. INTRODUCTION

Permian (Fig. 1; Veevers, 1994, 2001; Zeigler et al., 1997; Hyde et al., 1999; Scotese et al., 1999). In these models, ice flowed radially outward from a glacial spreading center over Victoria Land, Antarctica, and extended, in one direction, to glaciomarine margins in the Ellsworth Mountains and South Africa (Lindsay, 1970; Barrett, 1991; Veevers, 2001). In the opposite direction, models show ice flowing out of northern Victoria Land to terrestrial ice

Many paleogeographic reconstructions of the late Paleozoic Gondwanan Ice Age (LPGIA) depict an immense ice sheet covering Gondwana during the Mississippian, Pennsylvanian, and *[email protected]

Isbell, J.L., 2010, Environmental and paleogeographic implications of glaciotectonic deformation of glaciomarine deposits within Permian strata of the Metschel Tillite, southern Victoria Land, Antarctica, in López-Gamundí, O.R., and Buatois, L.A., eds., Late Paleozoic Glacial Events and Postglacial Transgressions in Gondwana: Geological Society of America Special Paper 468, p. 81–100, doi: 10.1130/2010.2468(03). For permission to copy, contact [email protected]. ©2010 The Geological Society of America. All rights reserved.

81

82

Isbell

Africa

India Au

S. America

A.

str ali a

Antarctica VL

CTM EM

Figure 1. (A) Map showing the distribution of glacial basins in Gondwana. (B) Gondwana reconstruction after Powell and Li (1994), showing the location of a hypothetical ice sheet covering Gondwana (modified from Ziegler et al., 1997) with an ice spreading center over Victoria Land, Antarctica (Lindsay, 1970; Barrett, 1991). Hypothetical flow directions are from models proposed by Lindsay (1970), Crowell and Frakes (1971), Barrett (1991), and Veevers (2001). CTM—central Transantarctic Mountains, EM—Ellsworth Mountains, NZ—New Zealand, VL—Victoria Land.

NZ

B. 0

km 1500

Antarctica

CTM

VL

EM

Gondwana Glacial Basins

NZ Gondwanan Ice Sheet

margins in southern Australia and glaciomarine margins in Tasmania (Fig. 1; Crowell and Frakes, 1971; Barrett, 1991; Veevers, 2001). Throughout Antarctica, early workers identified upper Paleozoic diamictites as tillites deposited subglacially from the terrestrial portion of the ice sheet as it waxed and waned across Gondwana (e.g., Lindsay, 1970; Barrett and Kyle, 1975; Barrett et al., 1986). The occurrence of an ice spreading center in Victoria Land supplying ice to the massive Gondwanan Ice Sheet is central to many paleogeographic and climatic models for the late Paleozoic. These models are based on work conducted prior to the advent of modern glacial facies models, and therefore their validity has not been rigorously tested in Victoria Land, the proposed terrestrial center of the ice sheet. In Antarctica (Fig. 2), upper Paleozoic glacial deposits thin from the central Transantarctic Mountains into southern Victoria Land. There, glaciogenic deposits are discontinuous (Barrett and Kyle, 1975; Collinson et al., 1994). In southern Victoria Land, thick glaciogenic deposits are reported as glacial terrestrial valley fills (Barrett, 1972; Barrett and Kyle, 1975; McKelvey et al., 1977; Barrett and McKelvey, 1981). Barrett and McKelvey (1981) suggested that the valleys preferentially preserved the deposits from erosion during postglacial rebound, whereas, outside the valleys, partial if not complete erosion of the glacial

Victoria Land Ice Spreading Center 0

km 1500

deposits occurred prior to deposition of the overlying Lower Permian Weller Coal Measures. The occurrence of the glacial valleys and regional thinning of the glacial strata toward Victoria Land led Lindsay (1970) and Barrett (1991) to conclude that southern Victoria Land lay adjacent to a late Paleozoic glacial spreading center and that deposition occurred from the terrestrial portion of the ice sheet (Fig. 1B). However, data presented herein suggest that some of the relief on the preglacial contact separating upper Paleozoic glacial rocks above from Devonian strata below resulted from glaciotectonic thrusting along concave upward shear planes rather than from erosion and development of glacial valleys, and that deposition occurred within a glaciomarine setting. Even though glaciotectonic structures and their significance have been recognized in Gondwanan rocks in South Africa and South America (e.g., Visser and Loock, 1982; Rocha-Campos et al., 2000), descriptions of pre-Pleistocene glaciotectonic features are scarce. Deformed upper Paleozoic glaciogenic rocks also occur in Antarctica. The identification and interpretation of these strata are enigmatic, as deformation has been attributed to overriding of the sediments by Late Carboniferous or Permian ice, slumping along glacial valley walls, and/or owing to tectonic upheaval (Barrett, 1972; Barrett and Kyle, 1975; McKelvey

Permian deposits, southern Victoria Land, Antarctica

A

83

B

Figure 2. (A) Map of Antarctic sites showing the location of southern Victoria Land. (B) Location map of southern Victoria Land showing Kennar Valley, Mount Metschel, and Mount Ritchie.

et al., 1977; Barrett and McKelvey, 1981; McElroy and Rose, 1987; Spörli, 1992). This paper presents sedimentologic data collected from Mount Metschel, Mount Ritchie, and Kennar Valley in southern Victoria Land (Fig. 2) during the 2000–2001 and 2001–2002 austral field seasons. Here, strata of the Devonian Aztec Siltstone, the Permian Metschel Tillite, and the basal few meters of the Weller Coal Measures (Fig. 3) are described and interpreted (1) to determine whether a major late Paleozoic ice sheet spreading center was located in Victoria Land, (2) to determine the setting in which glaciogenic sediments were deposited (i.e., subglacial, periglacial, glaciomarine), and (3) to document the occurrence of glaciotectonic structures within upper Paleozoic strata in Antarctica. Understanding upper Paleozoic glaciogenic strata in southern Victoria Land will help to resolve the nature, timing, and extent of the LPGIA, which is of importance, as this glacial interval is one of the best analogues for predicting what will happen to Earth systems during the transition out of the current Cenozoic Ice Age. STRATIGRAPHY Glaciotectonic structures and glaciogenic deposits described in this paper are found at the top of the Devonian Aztec Siltstone, in the Lower Permian Metschel Tillite, and at the base of the Lower Permian Weller Coal Measures in southern Victoria Land (Fig. 3). The rocks are part of the Taylor (Devonian) and Victoria (Permian and Triassic) Groups of the Beacon Supergroup. Barrett (1972), Barrett and Kyle (1975), McKelvey et al. (1977), Barrett and McKelvey (1981), and McElroy and Rose (1987) provide

Figure 3. Generalized stratigraphic section of Devonian and Permian rocks exposed in southern Victoria Land, Antarctica.

an overview of the occurrence and distribution of Devonian and upper Paleozoic rocks in this area. The Aztec Siltstone is up to 217 m thick and consists of interbedded shale, siltstone, and cross-bedded sandstone (Fig. 3). Siltstones and shales are a few centimeters to several meters thick, whereas sandstones vary from 0.1 to 15 m in thickness (McKelvey et al., 1977; McElroy and Rose, 1987). The occurrence of fining-upward cycles, red beds, rootlet horizons, mud cracks, and conchostracan fossils all suggest that deposition occurred in an alluvial setting (McKelvey et al., 1977; McPherson, 1978, 1979). A microflora containing Geminospora lemurata and fossil fish near the top of the unit, including Bothriolepis, Groenlandaspis, and Turinia gondwana, indicate a Middle to Late Devonian age (Helby and McElroy, 1969; McKelvey et al., 1972; Ritchie, 1975; Young, 1988, 1989, 1991; Turner and Young, 1992).

84

Isbell

Glaciogenic strata of the Metschel Tillite unconformably overlie rocks of the Aztec Siltstone (Fig. 3). The Metschel Tillite is 0–85 m thick and consists of diamictite, conglomerate, sandstone, and shale, which are locally intraformationally deformed (Barrett and Kyle, 1975; McKelvey et al., 1977; Barrett and McKelvey, 1981; McElroy and Rose, 1987). Although the age of the Metschel Tillite is unknown, it is here considered to be Early Permian on the basis of its position below rocks of the Weller Coal Measures, which contain Lower Permian palynomorphs, and also on the basis of correlation of the Metschel Tillite with glaciogenic rocks of the Darwin Tillite exposed 130 km to the south. The Darwin Tillite contains Asselian-Sakmarian palynomorphs, as do all other palynomorph-bearing upper Paleozoic glaciogenic rocks in Antarctica (cf. Barrett and Kyle, 1975; Kyle, 1977; Kyle and Schopf, 1982; Lindström, 1995; Askin, 1998). Strata of the Weller Coal Measures rest both disconformably and conformably on diamictites of the Metschel Tillite (Fig. 3; McKelvey et al., 1977). Conglomerate and cross-bedded, coarsegrained sandstone containing quartz pebbles occur at the base of the formation, whereas upward within the unit, fine to coarsegrained sandstones are interstratified with siltstones, shales, and coals. Strata near the base of the formation contain fossil leaf impressions of Gangamopteris, Glossopteris, and Cordaites (Kyle, 1976; Pyne, 1984; Cúneo et al., 1993). Kyle (1977) correlated the Weller Coal Measures with Lower Permian (Artinskian) rocks in Australia (Australian Stage 4 Palynomorph Zone of Evans, 1969) based on microfloras from the middle and upper parts of the formation. GLACIOGENIC DEPOSITS AND GLACIOTECTONIC STRUCTURES Two types of deformed and two types of undeformed facies associations within Devonian and upper Paleozoic strata are reported in this paper. The associations are (1) sheared and homogenized sandstone and diamictite, (2) large-scale thrust sheets, (3) diamictite, and (4) sheet sandstone facies associations.

gently dipping (12°–24°) sandstone, mudstone, and paleosols in the Aztec Siltstone that are internally folded (folding may be the result of numerous microscopic thrust faults) and cut by thrust faults with centimeter-scale displacements; Fig. 4A); (3) 2 m of pervasively thrust faulted Aztec Siltstone where individual sandstone beds are internally folded and thrust faulted (Fig. 4B); (4) a sharp contact that truncates underlying folded and faulted rocks of the Aztec Siltstone below from diamictites of the Metschel Tillite above; (5) 1.5 m of thrust faulted diamictite consisting of a chaotic mixture of sandstone pods and boudins, including material from the Aztec Siltstone, and granite and quartz clasts up to 0.3 m in diameter; (6) 2–3 m of pervasively foliated (dips of 8°–30°), chaotic to homogenized diamictite containing granite and quartz clasts up to 0.2 m in diameter (Fig. 4C); and (7) locally up to 2 m of almost completely homogenized foliated diamictite containing granite, and quartz clasts up to 0.2 m in diameter (Fig. 4D). Foliation, thrust faults, and axial planes of folds within the deformed zone dip at up to 44° toward 306°. At the top of the deformed zone a sharp contact separates highly foliated and homogenized diamictites below from undeformed, weakly stratified to stratified diamictite above (Fig. 4E). A similar but much thinner deformed zone occurs on the east side of Mount Ritchie near the top of the Metschel Tillite. There, weakly stratified and stratified diamictite are overlain by a 1-m-thick sandy interval containing numerous small-scale thrust faults, and an overlying 1-m-thick foliated diamictite. The small-scale thrust faults dip at up to 45° toward 243°. This foliated diamictite is overlain by a sharp contact with weakly stratified diamictite or by an erosional surface that separates strata of the Metschel Tillite below from conglomerates and cross-bedded sandstones of the Weller Coal Measures above. Interpretation The presence of small-scale thrust faults, chaotic bedding– boudinage structures, and foliated homogenized materials at Mount Metschel suggest that sandstone and shale at the top of the Aztec Siltstone and diamictite at the base of the Metschel Tillite were deformed by shear. In general, deformation within

Sheared and Homogenized Sandstone and Diamictite Facies Association Description On Mount Metschel (Fig. 2) an asymmetric deformation zone separates the 50(+)-m-thick Devonian Aztec Siltstone from the overlying 21.5-m-thick Permian Metschel Tillite (Fig. 4). At this site the Aztec Siltstone consists of 0.1–0.5-m-thick interbeds of cross-laminated and horizontally laminated, fine- to mediumgrained sandstone, gray to red mudstone, and calcrete-bearing paleosols. Near the top of the unit, strata pass upward through a deformed zone, which shows increasing strain upward, into foliated diamictite of the Metschel Tillite (Fig. 4). The following zones occur from the base to the top of this sequence, which straddles the Aztec-Metschel contact: (1) undeformed interbeds of sandstone and mudstone of the Aztec Siltstone; (2) 2–2.5 m of

Figure 4. Stratigraphic column of deformed strata of the Devonian Aztec Siltstone and Permian Metschel Tillite exposed on the eastfacing slopes of Mount Metschel, interpreted to be the deposits of a subglacial deformation bed (glaciotectonite and deformation till). Orientations of structural features and interpreted ice-flow directions are also shown. (A) Thrust faults with centimeter-scale displacements cut strata of the Aztec Siltstone. Brunton compass for scale. (B) Pervasively thrust faulted sandstone at the top of the Aztec Siltstone. Rock hammer for scale. (C) Pervasively foliated diamictite and sandstone near the base of the Metschel Tillite. Ice axe for scale. (D) Pervasively foliated and partially homogenized diamictite near the base of the Metschel Tillite. Ice axe for scale. (E) Stratified diamictite with lobeshaped, boulder-bearing sandy diamictite near the top of the Metschel Tillite at Mount Metschel. Ice axe for scale. vm—vector mean, mvl— mean vector length, n—number of measurements.

E

D

Stratified diamictite

Dipping foliation

Lobe-shaped sandy diamictite

(c)

Ice flow direction

n=11

Sand f mc

vm=126 mvl=0.95 n=11

Diamictite

Clay Silt

Dip of shear planes

C

14m

Metschel Tillite

Undeformed stratified diamictite with granite and quartz clasts up to 1 m in diameter

(d)

B

Homogenized diamictite with granite and quartz clasts up to 0.2 m in diameter Pervasively foliated, diamictite with granite and quartz clasts up to 0.2 m in diameter Thrust faults

Thrust faulted diamictite with sand boudins and granite and quartz clasts up to 0.3 m in diameter

Sharp Contact Aztec Siltstone

Pervasively thrust faulted (dm to m displacements) sandstone that is internally folded and thrust faulted

A Small-scale thrust faults

Internally folded and thrust faulted with cm displacements Small-scale thrust faults

0

Undeformed Aztec Siltstone

86

Isbell

individual tectonic and subglacial shear zones displays a symmetrical increase in strain away from upper and lower boundaries toward a homogenized central region of high finite shear strain (Van der Wateren, 1987, 2002). The distribution of strain in the shear zone at Mount Metschel is asymmetrical, with changes in structures indicating that strain increased upward from undeformed rock in the Aztec Siltstone into the overlying homogenized diamictite at the base of the Metschel Tillite. However, structures indicating decreasing strain do not occur as the homogenized diamictite passes abruptly upward into undeformed stratified diamictite. Rocks at the contact between the Aztec Siltstone and the Metschel Tillite are similar to sediment deformed and deposited subglacially. Under certain conditions, shear between glacial ice and its unconsolidated or weakly consolidated substrate results in formation of a deforming bed beneath the ice-sediment interface (Alley, 1991; Benn and Evans, 1998). For ice streams and surging glaciers, much of the forward motion of the glacier may be accounted for within such a unit. Formation of a deforming bed is favored by the occurrence of (1) subglacial waters, (2) unfrozen unconsolidated or weakly consolidated (sedimentary) substrates, and (3) substrate materials that inhibit water from draining away into the underlying sediments (Alley, 1991; Boulton, 1996). Under these conditions, glacial ice forms the “mirrored top” of the idealized symmetrical shear zone with strain increasing upward within the substrate toward the ice-sediment interface and then decreasing away from the interface into the overlying glacier (Boulton, 1979; Boulton and Jones, 1979; Boulton and Hindmarsh, 1987; Van der Wateren, 2002). A vertical strain profile within a subglacial deforming bed consists of (1) undeformed sediment at depth, (2) materials deformed by brittle failure and faulting, (3) pervasive ductile shearing of materials, and (4) plowing and sliding of ice and debris along the ice-sediment contact (Banham, 1977; Boulton, 1987; Alley, 1991; Hart and Boulton, 1991; Benn, 1995; Benn and Evans, 1996; Van der Wateren, 2002). A décollement may separate structural zones displaying different degrees of strain (Banham, 1977; Boulton, 1987). Subglacially sheared deposits are classified by Benn and Evans (1996) as glaciotectonites if the deposits contain structural characteristics of the original parent material, or as deformation tills if the material has been homogenized. Recently, van der Meer et al. (2003) reported that all subglacial tills form from a combination of both deformation and lodgment. At Mount Metschel, deformed Aztec Siltstone has all of the characteristics of a glaciotectonite, whereas the overlying foliated diamictites at the base of the Metschel Tillite display characteristics of a deformation till. The two formations are separated by a décollement, which would have served as the structural boundary between brittle and ductile deformation. The décollement also occurs at the position of the regional unconformity that now separates the two formations. The occurrence of this type of deformation indicates that rocks of the Upper Devonian Aztec Siltstone were weakly consolidated during glaciation. Such an interpretation is tenable, as Cambrian and Ordovician sandstones

of the upper Midwestern United States are weakly consolidated and were locally deformed by glaciogenic activity during the Pleistocene. The interstratification of sands and muds in the Aztec Siltstone likely inhibited subglacial drainage into underlying aquifers, possibly resulting in high pore-water pressures and the formation of a subglacial deforming bed during overriding of the area by a Permian glacier. The orientation of the thrust faults and shear planes at Mount Metschel indicate local ice movement toward 126°. Barrett and Kyle (1975) reported striations at Mount Metschel oriented toward 120°. However, these striations were likely slickensides contained within the shear zone. The deformed zone at the top of the Metschel Tillite on the east-facing slope of Mount Ritchie is also interpreted to have formed subglacially as a glaciotectonite and deformation till. The occurrence of weakly stratified and stratified diamictite (see section below on Diamictite Facies Association) directly below this unit, coupled with the orientation of the structural fabric within the deformed zone, suggests that grounded ice advanced northeastward (063°) into a glaciomarine setting. Large-Scale Thrust Sheet Facies Association Description of Strata at Mount Richie At Mount Ritchie (Fig. 2), thrust sheets in the Metschel Tillite, up to 15 m thick, occur at and just above the contact with undeformed strata of the Devonian Aztec Siltstone (Fig. 5). The sheets, which are imbricately stacked, make up a 600+-m-long duplex along a north-northwest– to south-southeast–trending ridge near the summit of the mountain. These strata are exposed on the west-facing slope of Mount Richie. Along much of the ridge the contact between the thrust sheets and underlying strata is covered by snow and scree. Where exposed, the substratum includes (1) diamictite, (2) deformed interstratified diamictite and sandstone, (3) breccia, and (4) strata of the Aztec Siltstone (Figs. 5A, 5B, 5C). At its northern end the duplex ramps up onto a 4–6-m-thick, fine- to medium-grained massive sandstone (see the section below on Sandstone Sheet Facies Association) in the Metschel Tillite. There, individual thrust sheets are bounded by

Figure 5. (A) Photo of imbricate thrust sheets exposed on the westsouthwest face of Mount Ritchie, interpreted to be part of a subaquatic thrust moraine complex. For scale, the distance from the base of the Weller Coal Measures to the base of the sill is 25 m. (B) Interpretive map of the thrust sheets on the west-southwest face of Mount Ritchie. Stereonet shows the orientation of the thrust sheets, and the rose diagram shows paleo ice-flow directions interpreted from the orientations of the thrust sheets. (C) Brecciated zone marking the location of a shear plane at the base of a thrust sheet on Mount Ritchie; 15-cm-long ruler for scale. (D) Thrust sheets of the Metschel Tillite, composed of strata of the Aztec Siltstone at Mount Ritchie. The sheets are bounded by major thrust faults (MTF), whereas recumbent folds and small-scale thrust faults (STF) with meter-scale displacements occur within the sheets. Person and Jacob’s staff (1.6 m) for scale. vm—vector mean, mvl—mean vector length, n—number of measurements.

A

B

C

D

88

Isbell

listric to sigmoidal-shaped shear zones that consist of either a 0.05–0.3-m-thick, intensely foliated mudstone containing rare granite and gneiss granules, pebbles, and cobbles or locally by a 0.5–3-m-thick brecciated and/or chaotically folded zone. The thrust sheets are sigmoidal in shape and consist of massive sandstone, which contains an internal mosaic of annealed fractures and dewatering structures. The sheets within this complex dip at up to 45° toward 127°. Near the back of the thrust complex, several thrust sheets consist of an interbedded succession of crossstratified and horizontally laminated sandstones and mudstones of the Aztec Siltstone (Figs. 5A, 5B, 5D). These sheets rest on the backs of the thrust sheets described above. Internally, the Aztec thrust sheets contain internal deformational structures that include (1) high-angle normal faults, (2) small-scale reverse faults with displacements of centimeters to a few meters, (3) overturned folds, and (4) recumbent folds (Fig. 5D). The folds and smallscale thrust faults indicate compression toward the WNW. The thrust sheets are overlain by weakly stratified and stratified diamictite (see Diamictite Facies Association, below) that lap onto and drape over the thrust complex. Description of Strata at Kennar Valley At Kennar Valley (Fig. 2), rocks of the Metschel Tillite are exposed primarily on a ridge that extends northeastward into the center of the valley. Here, highly deformed rocks of the Metschel Tillite occur between undeformed strata of the Devonian Aztec Siltstone below, and undeformed strata of the Weller Coal Measures above (Figs. 6 and 7). Within the Metschel Tillite the style of deformation changes progressively along the ridge toward the southwest from ductile, to brittle, to undeformed (Fig. 6). A chaotic zone of ductile deformation is exposed on the eastern valley wall and on the northeastern end of the central ridge. The folding consists of large-scale recumbent and isoclinal folds and/or nappe-like structures, with structural displacement toward the west-southwest. The strata consist of interstratified, mediumto coarse-grained, massive to internally deformed sandstones and stratified diamictites (Figs. 6A–6D). Some of the sandstone units contain dewatering features (pipes, sheets, and dish structures). Low-angle thrust-fault–bounded packages up to 10 m thick occur throughout the middle and southern parts of the central ridge (Figs. 6 and 7A–7C). These thrust sheets occur in front, and on top of, the chaotic ductile zone described above. The faults, marked by 0.01–1-m-thick, boudin-bearing, foliated and homogenized mudstones and siltstones (Figs. 7D and 7E), dip at up to 51° toward the east-northeast (toward 72°), indicating structural transport of the sheets toward the west-southwest (252°). These faults have high-angle dips along the northeast part of the ridge but become subhorizontal toward the southwest. Internally within the thrust-fault–bounded packages, beds are slightly folded and contain abundant high-angle normal and listric-shaped reverse faults (Figs. 7C, 7F, 7G). The normal faults are common at the bases of the thrust sheets and occur within the concave-up troughs of small, open synclines (Fig. 7C). On the west-northwest side of the central ridge (Fig. 7A and 7B), stacked thrust sheets

1–7 m thick consist of (1) weakly stratified diamictite; (2) interbedded conglomerate and faulted, cross-bedded to massive sandstone containing dish structures and dewatering pipes; and (3) coarsening-upward siltstone to sandstone successions with siltstones containing isolated ripples (lenticular bedding) and sandstones containing cross-laminae, horizontal laminations, wave ripple laminae, swaley cross-stratification, load and flame structures, and scattered dropstones. The contact with the underlying Aztec Siltstone is covered by scree. However, some thrust sheets cut below the level of the contact (Figs. 7A and 7B). Undeformed conglomerates and sandstones of the Weller Coal Measures rest on an erosion surface with a relief of several meters cut into the underlying Metschel Tillite. Interpretation Deformational structures in the Metschel Tillite at Mount Ritchie and Kennar Valley indicate formation caused by horizontal compression. These structures include listric-shaped thrust faults, sigmoidal-shaped thrust sheets, fold and thrust nappes, and recumbent folds. Shear zones at the base of individual thrust sheets are marked by boudin-bearing, foliated mudstones and decimeter- to meter-thick breccia zones. Substantial evidence indicates that deformation occurred primarily during deposition of the Metschel Tillite. This evidence includes (1) occurrence of the thrust sheets between undeformed strata of the Aztec Siltstone and glaciogenic deposits of the Metschel Tillite below, and undeformed glaciogenic and fluvial strata of the Metschel Tillite and Weller Coal Measures above; (2) truncation and overriding of glaciogenic deposits of the Metschel Tillite by the thrust sheets; (3) composition of the thrust sheets, which consist primarily of glaciogenic deposits; (4) occurrence of both soft sediment deformational (i.e., massive sandstones, dewatering structures, and chaotically folded sandstones and diamictites) and brittle (normal and reverse faults) structures within the thrust sheets, indicating that both ductile and brittle deformation and water expulsion from the sediments occurred, which is characteristic of deformation of unconsolidated and weakly consolidated deposits; and (5) onlapping and overlapping of the thrust complexes by diamictites, which indicate that continued glaciogenic deposition occurred across relief generated by the compressional structures. Thrust sheets truncating (Kennar Valley), and composed of strata (Mount Ritchie) derived from the underlying Aztec Siltstone, indicate that some excavation of bedrock also occurred. Because the zone of deformation appears to have occurred in unconsolidated glaciogenic sediment, and because it is confined between undeformed strata of the Aztec Siltstone below and the Permian Weller Coal Measures above, the compressional structures are interpreted to be intraformational features associated with glacial activity rather than being due to tectonic stresses. In glaciogenic settings, compressional features result from either sliding-slumping or glaciotectonic deformation, both of which produce similar structures. Slides occur on slopes where gravitational failure of the substrate results in rotational

A

Weller Coal Measures

B Brittle zone Unconformities Surfaces (bedding and faults)

Chaotic zone (Fig. 6C)

Metschel Tillite

C

D

E Brittle

Ductile

Glacier

Figure 6. Folded and thrust-faulted strata of the Metschel Tillite at Kennar Valley that are interpreted to have formed as a subaquatic thrust moraine. (A, B) Highly folded strata (right side of photo), giving way to strata deformed by thrust faults (left side of photo). The strata are exposed on the east-northeast end of the central ridge in Kennar Valley. The photo shows a slope and cliff face ~50 m high. (C) Closeup of the highly folded portion of the cliff face shown in Figures 6A and 6B. (D) Highly deformed strata contained within the chaotic zone exposed on the westnorthwest side of the central ridge in Kennar Valley. Jacob’s staff (1.6 m) for scale. (E) Schematic diagram showing formation and orientation of deformation in a modern thrust moraine (diagram modified from Croot, 1988).

A

B

C

D Thrust sheet

Major thrust fault Normal faults

Thrust fault

E

F

Major thrust fault

G

Major thrust fault

Normal faults

Internal thrust fault

Figure 7. Strata exposed on the west-northwest–facing slope of the central ridge in Kennar Valley, interpreted to be part of a thrust moraine complex. (A, B) Thrust faults and thrust sheets exposed along the northern end of the central ridge. The Weller Coal Measures are ~28 m thick, for scale. Stereonet shows the orientation of thrust faults, and the rose diagram shows the ice-flow directions interpreted from fault orientations. (C) Thrust sheet showing basal thrust fault and internal accommodation faults. Jacob’s staff (1.6 m) for scale (white arrow). (D, E) Foliated and homogenized mudstones mark the position of major thrust faults. (F) Small-scale thrust fault at the base of a thrust sheet. Jacob’s staff (1.6 m) for scale (white arrow). (G) Normal faults near the base of a thrust sheet; 15-cm-long ruler for scale. VM—vector mean, MVL—mean vector length, N—number of measurements.

Permian deposits, southern Victoria Land, Antarctica extension of a sediment mass along listric glide planes followed by downslope sliding of a coherent block away from a headwall. Deposition occurs when the block comes to rest at the toe of the slide. Slumps form in a similar fashion. However, slumps display internal folding and faulting. The resulting slide-slump structures consist of either a single block or stacked blocks or sheets (multiple slide-slump events), each bounded below by a sheared glide plane (Allen, 1985; Collinson and Thompson, 1989; Ricci Lucchi, 1995; Benn and Evans, 1998). Slump and/or slide blocks are identified by the following criteria: (1) truncation of underlying strata from rotational extension in areas adjacent to the slumpslide scarp; (2) deposits, which at the head of the slump-slide dip away from the headwall scarp; (3) deposits, which at the toe of the slump-slide typically rest on slopes that dip in the direction of sliding; (4) deposits typically composed of only one to several sheets with younger sheets resting on the backs of older sheets; (5) the occurrence of slump fold noses; (6) an increase in the degree of deformation in the direction of transport owing to compression at the toe of the slump-slide; and (7) deposits that typically do not contain excavated bedrock blocks (cf. Tucker, 1996; Collinson and Thompson, 1989; Ricci Lucchi, 1995; Stow, 2005). Slides and slumps are commonly associated with debris flows, which are often triggered during formation and movement of the slump and slide blocks. Structures formed by glaciotectonic compression develop at ice margins during glacial advance and involve displacement of subglacial and proglacial sediment, and weak rock by both ductile and brittle deformation (Aber et al., 1989; Hart and Boulton, 1991). Deformation is facilitated by (1) horizontal stresses resulting from high ice overburden pressures up-glacier, (2) weak subglacial and periglacial substrates that produce décollements during failure, (3) high pore-water pressures that reduce the cohesive and yield strength of ice marginal and subglacial sediments and coherent substratum, and (4) by the presence of reverse slopes along the glacial margin that increase horizontal stress (Bluemle and Clayton, 1984; Aber et al., 1989; Boulton and Caban, 1995). Ductile deformation includes formation of open to overturned folds, whereas brittle failure results in the development of lowangle overthrusts, listric thrust faults, imbricately stacked thrust sheets, and nappes (Hart, 1990; Van der Wateren, 2002). “Piggyback” thrusting is common where early formed proximal sheets are carried on the backs of later formed distal blocks. In this scenario, intense folding and high-angle reverse faulting occur along the ice margin, decreasing to lower angle emplacement of thrust sheets outward (Eybergen, 1987; Croot, 1988; Van der Wateren, 2002). Thrust moraines are the surface expression of this deformation (Aber et al., 1989). Several criteria can be used to identify thrust sheet complexes produced by glaciotectonic deformation. These include (1) truncation of underlying strata throughout the zone of deformation, especially along frontal thrusts; (2) occurrence of compressional features throughout; (3) décollement–thrust fault surfaces that ramp up and over truncated proglacial sediment near the leading edge of the frontal thrust; (4) thrust complexes where older

91

thrust sheets rest on top of younger sheets or where early formed sheets display near-vertical orientations owing to rotation and piggyback transport on younger sheets that formed along frontal thrust faults during continued compression; (5) a decrease in the degree of deformation in the direction of transport owing to the formation of frontal thrust faults; and (6) thrust complexes that typically contain excavated blocks of bedrock (Croot, 1988; Aber et al., 1989). Small thrust sheets composed of periglacial material also occur along both terrestrial and subaqueous ice margins. These thrust blocks or push moraines, which are typically no more than 5 m high, form during small-scale seasonal advance and retreat cycles of the ice front. During a seasonal advance, periglacial material is pushed or “bulldozed” into a morainal ridge. Although these features are commonly composed of supraglacial debris dumped at the ice margin, some also include proglacial sediment (Bennett and Glasser, 1996). Compressional deformation in the Metschel Tillite is consistent with formation by glaciotectonic deformation of periglacial deposits. Supraglacial material does not occur within the thrust sheets. At Mount Ritchie, shear zones truncate glaciogenic and/or Devonian strata throughout the thrust complex. Near the front of the complex the basal thrust sheet ramps up and over preexisting deposits, with the orientation of the truncation surface dipping in a direction opposite to that of the direction of transport for the sheets. Because thrust sheets of the Aztec Siltstone rest on these basal imbricated sheets, the “Aztec” sheets likely formed early, only to be later transported piggyback on younger structures that developed along frontal thrust faults. Dip directions of the thrust sheets at Mount Ritchie suggest glaciotectonic transport, and therefore glacial advance toward 307°. At Kennar Valley the lateral change from highly contorted beds, recumbent folds, and thrust nappes to thrust sheets indicates a decrease in the intensity of deformation from ductile to brittle in the direction of structural transport, which was toward 252°. The high-angle orientation of the proximal deformation zone suggests that these may have been the first sheets to have formed and that they were later transported piggyback on younger thrust sheets and rotated into high angles during continued propagation of the thrust front. This progression of structures is similar to proximal to distal changes in deformation style seen within modern ice-marginal thrust moraine complexes (Fig. 6E; cf. Croot, 1988; Van der Wateren, 2002). Therefore, structures exposed at Kennar Valley are interpreted to have formed as glaciotectonic features along an ice margin. During emplacement, flexure of the deforming mass would have produced both extensional and compressional stresses within individual thrust sheets where the thrust sheets were flexed during transport into broad synclines and anticlines. Such stresses would have resulted in the formation of internal accommodation features (i.e., normal and listric-shaped reverse faults), which would have facilitated flexing of the thrust sheets. In modern glacial settings, glaciotectonic deformation occurs along both terrestrial and glaciomarine ice margins (Bennett

92

Isbell

et al., 1999; Van der Wateren, 2002). Massive, weakly stratified, and stratified diamictite facies associated with the thrust sheets at Mount Ritchie suggest that deposition occurred in a basinal setting (see next section). The lithologies and sedimentary structures of strata contained within the thrust sheets at Kennar Valley, which include stratified to massive diamictite (see next section), coarsening upward siltstone to sandstone successions containing wave ripple laminae, and swaley cross-stratification, also suggest that ice advanced into a glaciomarine environment. Diamictite Facies Association Description Massive to weakly stratified diamictite occurs at Kennar Valley and at Mount Ritchie (Figs. 2 and 8). At Kennar Valley several of the thrust sheets are composed entirely of this type of diamictite (see large-scale thrust-sheet-facies association, Figs. 8A and 8B). At Mount Ritchie, massive and weakly stratified diamictites typically have gradational upper and lower contacts with stratified diamictite. However, weakly stratified diamictite is also truncated by conglomerates and sandstones in the overlying Weller Coal Measures. The diamictite consists of a clay to fine-grained sand matrix containing scattered to abundant clasts. Laminae and thin wisps of silt and sand, which are better sorted than the surrounding matrix, define faint stratification. Clasts of granite, quartzite, and gneiss up to 1 m in diameter, pierce stratification, and the long axes of some clasts are oriented at high angles to bedding. Within massive to weakly stratified diamictite units, isolated centimeter- to meter-thick and decimeter- to meter-wide pods of contorted and massive sandstone and conglomerate with dewatering structures occur (Fig. 8C and 8D). Injection structures in the form of diamictite diapirs intrude these pods. Stratified diamictites are common at Mount Metschel, Mount Ritchie, and within the thrust sheets at Kennar Valley (Figs. 8E–8G). These diamictites have a similar matrix and clast (lithology and size) composition as those of the massive and weakly stratified diamictites. Stratification is distinct, however, and consists of centimeter- to decimeter-thick semi-continuous layers of silty sand–rich layers alternating with mud-rich layers. In a few places, intercalations of centimeter-scale units of normally graded to cross-laminated, medium-grained sandstones delineate stratification (Fig. 8G). The bases of the sandstones commonly display load structures. Within the stratified diamictite, clasts up to 30 cm in diameter are common and cut stratification. Within the stratified diamictites, decimeter-thick and tens-of-meters-long lenses, or lobe-shaped bodies of massive sandstone and conglomerate, occur (Fig. 4E). These bodies characteristically contain sharp bases with abundant centimeter- to decimeter-scale load structures. The bodies are also intruded by diapirs of stratified diamictite. Boulders up to 0.7 m in diameter protrude from the top of the lobe-shaped bodies. On the north-northwest and west-southwest sides of Mount Ritchie, just below the contact with the overlying Weller Coal Measures, locally abundant accumulations of clasts up to 1 m

in diameter occur (Fig. 9A). Many of these clasts cut stratification and have their long axes oriented at high angles to bedding. Faceted and striated clasts also occur. At the base of the overlying Weller Coal Measures, gravel clasts (pebbles, cobbles, and boulders) protrude downward into the underlying diamictites, as do load structures at the base of coarse-grained sandstones and pebble to cobble conglomerate units. Intrusion of the Weller by diamictite diapirs also occurs, as does interfingering of Metschel diamictites and Weller sandstones and conglomerates (Fig. 9B). These coarse-grained Weller sediments are part of a 6- to 10-m-thick wedge-shaped coarse- to very coarse grained trough-cross-bedded sandstone body (Figs. 10A–10D). When traced over several hundred meters, this body displays an internal geometry characterized by low-angle downlapping beds and surfaces. When viewed on exposures perpendicular to paleoflow orientations, the downlapping units are cut by numerous smallscale (meters to tens of meters wide and up to a few meters deep) cross-bedded, sand-filled channels. Paleocurrent directions are oriented toward 234° (Figs. 10C and 10D). Interpretation Massive and stratified diamictites are commonly interpreted to have formed in both polar and temperate glaciomarine systems by a number of different processes. In polar settings the temperature and strength of the ice allow for the development of floating glacial tongues and ice shelves. Under these conditions, ice decouples from the bed and begins to float just seaward of a grounding line. Here, in this proximal setting, melt-out of basal debris from the underside of the glacier allows both fine- and coarse-grained particles to settle through the water column and to produce massive diamictites. Sedimentation rates are relatively high near the grounding line, but they decrease distally owing to the loss of debris-rich basal ice (Evans and Pudsey, 2002). Stratified diamictites occur where sedimentation rates are lower, and where bottom currents (e.g., tidal pumping) winnow out finer grained particles (Domack et al., 1999). Dropstones are produced by either iceberg rafting in open-marine settings or as rain-out from the debris-poor distal portions of the ice shelf or glacial tongue (Evans and Pudsey, 2002). Under temperate conditions, warmer, weaker ice typically results in tidewater glaciers entering the sea as an actively calving ice cliff. Under these conditions the glacial terminus is grounded or partially floating. Here, glacial deposition is dominated by meltwater and, to an equal or lesser extent, by rafting of debris by icebergs. Grounding-line fans form where subglacial meltwaters, emerging from conduits along the ice front, deposit sand and gravel. As velocity in the effluent flow drops, buoyant forces become dominant, and the inflowing fresh water rises to the surface to form an overflow plume that transports fine sand, silt, and clay seaward. Sedimentation occurs as particles released from the plume settle through the water column (Cowan and Powell, 1990). Plume sedimentation, in conjunction with dumping of coarse debris during calving at the ice front, and by the release of clasts from the melting of icebergs, produces stratified

A

B Weakly stratified diamictite

Massive diamictite Thrust faults

Massive diamictite

C

D

Sandstone lens

Sandstone lens

E

F

G

Figure 8. (A) Massive diamictite contained within two different thrust sheets exposed on the west-northwest–facing slope of the central ridge in Kennar Valley. Person for scale. (B) Massive diamictite grading into weakly stratified diamictite on the west-northwest–facing slope of the central ridge in Kennar Valley. The diamictite unit is ~4 m thick at its thickest point in the photo. (C) Deformed sandstone lens contained within weakly stratified diamictite exposed on the eastern side of Mount Ritchie. Person for scale. (D) Deformed pebbly sandstone lens containing diamictite diapirs near the summit of Mount Ritchie. Trekking pole for scale. (E) Stratified diamictite containing a lens of massive diamictite at Mount Metschel. Ice axe for scale. (F) Stratified diamictite, Mount Ritchie; 15-cm-long ruler for scale. (G) Interstratified diamictite and beds of medium- to coarse-grained sandstone, Mount Ritchie. Persons for scale.

A

A

B

B C

Figure 9. (A)Weakly stratified diamictite at the top of the Metschel Tillite at Mount Ritchie. Note the abundance of clasts and the high dip angle of the long axes of many of the clasts. Clasts in the overlying Weller Coal Measures also project down into the underlying diamictite. Hammer for scale. (B) Interfingering of weakly stratified diamictite of the Metschel Tillite, and sandstone and conglomerate of the Weller Coal Measures. Trekking pole for scale.

D

E

Figure 10. (A, B) Sandstone sheet containing diamictite lenses and a channel body in the Metschel Tillite. A wedge-shaped sandstone at the base of the Weller Coal Measures containing downlapping surfaces is also shown. For scale, the distance from the base of the Weller Coal Measures to the base of the dolerite sill (Jurassic) is 25 m. (C, D) Wedge-shaped sandstone body at the base of the Weller Coal Measures, containing broad, channel-like scours. For scale, the distance from the base of the Weller Coal Measures to the base of the dolerite sill is 25 m. (E) Dewatering pipes within massive sandstone of the sandstone sheet facies association. Mechanical pencil for scale.

Permian deposits, southern Victoria Land, Antarctica diamictites in proximal glaciomarine positions (Cowan and Powell, 1991; Smith and Andrews, 2000). Massive diamictites occur in ice distal settings owing to iceberg rafting and scouring (Dowdeswell et al., 1994). However, massive and stratified deposits and/or mudstones lacking clasts can occur in either proximal or distal glaciomarine positions because of changes in sea ice and/or oceanographic conditions (Smith and Andrews, 2000; Dowdeswell et al., 2000). In both polar and temperate settings, stratification is also produced by periodic remobilization of the deposits by sediment gravity flows (Evans and Pudsey, 2002; Powell and Domack, 2002). In the Metschel Tillite, the occurrence of massive, weakly stratified, and stratified diamictites with gradational upper and lower contacts is suggestive of deposition in proximal to distal glacial basinal settings during fluctuations in the location of the ice front. Although it is unknown whether strata in southern Victoria Land were deposited under glaciolacustrine or glaciomarine conditions, a glaciomarine setting seems likely owing to an abundance of stratified diamictites, which suggests that sedimentation occurred from settling of particles from buoyant low-density meltwater plumes. Clasts in these deposits were likely introduced from the melting out of debris from ice fronts or from the release of debris rafted by icebergs. The abundance of massive, normally graded, and cross-bedded sandstone layers and pods within the strata is indicative of deposition from subaqueous meltwater as tractive flows and as sediment gravity flows, suggesting temperate glacial thermal conditions (Mackiewicz et al., 1984; Powell and Domack, 2002). Sandstones were deposited subaqueously on water-saturated substrates, which, when loaded, failed, producing load structures, dewatering structures, and disruption of the sandstone bodies by intrusion of diamictite diapirs. Such unstable substrates are the result of high sedimentation rates in ice proximal zones (Boulton, 1990). The occurrence of lens- and lobe-shaped sandy diamictite beds with protruding boulders is suggestive of deposition from debris flows with the boulders introduced as ice-rafted debris. However, small lenses of sand and conglomerate could have formed as iceberg dump structures (Thomas and Connell, 1985). Ice rafting of debris was likely an important component of sedimentation in both proximal and distal locations as indicated by the occurrence of clasts penetrating stratification (Thomas and Connell, 1985). Locally abundant boulders at the top of the Metschel Tillite at Mount Ritchie may indicate either iceberg dump structures or dumping of clasts at the ice front during calving events (Thomas and Connell, 1985; Powell and Domack, 2002). Further evidence of a glacial origin for these strata is provided by the occurrence of striated and faceted dropstones. Strata at the base of the overlying Weller Coal Measures show evidence of deposition contemporaneous with or shortly following deposition of the Metschel Tillite. Evidence includes (1) interfingering of Metschel diamictites with Weller sandstones and conglomerates, (2) boulders and cobbles in the Weller conglomerates protruding downward into the underlying diamictite,

95

and (3) Metschel diamictites intruding Weller strata as diapirs. The occurrence of load structures and diapirs suggests failure of a water-saturated Metschel substrate owing to deposition of the overlying Weller sandstones and conglomerates. Downlapping surfaces within the basal Weller strata indicate progradation of a wedge-shaped body across proximal Metschel glaciomarine deposits (massive diamictite and abundant dropstones). The basal sandstone and conglomerate body has a geometry and internal features that are similar to grounding-line fans described by Powell (1990) and Powell and Alley (1997). Sandstone Sheet Facies Association Description On the northwest side of Mount Ritchie, strata near the middle of the Metschel Tillite consist of a 16.5-m-thick succession of fine- to medium-grained cross-bedded sandstone, fine- to medium-grained massive sandstone containing dewatering pipes (Fig. 10E), shale, and massive to weakly stratified diamictite. These units are contained within a sheetlike sediment body that is in erosional contact with underlying shales and diamictites (Figs. 10A–10D). Sandstone within the sheet is laterally continuous. However, diamictite and shale are discontinuous within the sheet and either drape underlying beds or form lens- to podshaped bodies. On the eastern end of the west-northwest face of Mount Ritchie a large channel is incised to a depth of 10 m into the sheet (Figs. 10A and 10B). Laterally, channel margins are concordant with beds in the underlying sandstone. The channel is filled with clast-supported conglomerate, cross-bedded sandstone, massive sandstone containing dewatering pipes, and by massive to weakly stratified diamictite. The sheet and channel are overlain by, and interfinger with, massive, weakly stratified, and stratified diamictite (see section on Diamictite Facies Association, above). On the west-southwest side of Mount Ritchie the sandstone sheet is overridden by thrust sheets (see section on Large-Scale Thrust Sheet Facies Association, above; Figs. 10C and 10D). Interpretation The interfingering of sandstone and massive to stratified diamictite (see section on Diamictite Facies Association, above) suggest that the sandstone sheet was deposited in a glaciomarine setting. Features within the sandstone sheet, which include channeling, conglomerates, and cross-bedded sandstone, indicate deposition from high-energy tractive currents. Massive sandstone with dewatering pipes suggests rapid sedimentation rates possibly from either suspension or sediment gravity flows. These conditions are common near the grounding lines of temperate tidewater glaciers (Powell, 1990). In this setting, fresh-water effluent flow, emanating from the base of the glacier, deposits wedgeshaped bodies of sand and gravel known as grounding-line fans. These bodies are laterally continuous for hundreds of meters and are commonly cut by channels as the effluent flow cuts into early

96

Isbell

formed deposits (Powell, 1990). Away from the glacial front, current velocity drops, allowing the low-density fresh-water flow to detach from the bottom and rise to the surface to form a buoyant overflow plume. Rapid sedimentation from the rain-out of sand, silt, and clay from the plume results in deposits with high initial water contents (Cowan and Powell, 1990). These ice proximal deposits are unstable and are highly susceptible to dewatering and/or to remobilization as sediment gravity flows. Addition of coarse debris from ice rafting results in a complex interfingering of sands, conglomerates, muds, and diamictites (Powell, 1990; Powell and Domack, 2002). The sandstone sheet facies association is here interpreted as a grounding-line fan. This interpretation is consistent with the observed sedimentary features and with the overriding of the sandstone sheet on the west-southwest side of Mount Ritchie by ice proximal thrust sheets, which also are common near the grounding line of some tidewater glaciers (Bennett et al., 1999). OVERALL DEPOSITIONAL SETTING Deformational and depositional lithofacies associations in the Metschel Tillite suggest that sedimentation occurred in ice marginal and glaciomarine settings. Deformational features are interpreted as glaciotectonites, deformation tills, and thrust duplexes. Glaciotectonites and deformation tills resulted from subglacial deforming beds, which, in modern settings, typically form beneath ice streams, outlet glaciers, tidewater glaciers, and surging glacial lobes where high pore-water pressures facilitate deformation of unconsolidated or weakly consolidated substrates. Thrust duplexes formed as periglacial thrust moraines along ice margins. The occurrence of these strata and their glaciotectonic structures suggests that deposition in southern Victoria Land occurred at or near glacial termini. The occurrence of massive and stratified diamictites, lonestone-bearing deposits, sheet sandstones, and sandstones with swaley cross-stratification and wave ripple laminations suggests that Permian glaciers in southern Victoria Land advanced into, and retreated from glaciomarine settings. Owing to the occurrence of meltwater plume deposits (stratified diamictites), the depositional setting was most likely glaciomarine. Evidence in the form of glaciotectonite, deformation till, and cross-bedded sandstone indicates that abundant meltwater was present at the time of deposition, and therefore that temperate thermal conditions characterized the depositing ice. Such thermal conditions are characteristic of tidewater glaciers where deposition is dominated by subglacial deforming bed conditions, grounding line processes, meltwater outflow, meltwater plumes, and iceberg rafting of debris. Only a few ice-flow directions have been reported from upper Paleozoic glaciogenic rocks in southern Victoria Land (Barrett and Kohn, 1975; McKelvey et al., 1977). Flow directions from these data are ambiguous. Although deformation structures reported in this paper represent local glaciotectonic displacement, these features also provide a record of paleo-ice-flow directions. Transport directions were derived from displacement directions

of thrust sheets, orientation of foliation, and analyses of folds. Interpretation of these data suggests that glacial ice converged on southern Victoria Land off the East Antarctic craton to the west (e.g., glaciotectonite and deformation tills at Mount Metschel and Mount Ritchie) and off an area in the direction of the present Ross Sea to the east (Fig. 11; e.g., thrust sheets at Kennar Valley and Mount Ritchie). If these directions are correct, then ice advanced from glacial centers on opposite sides of the depositional basin. Expansion of ice from these centers then allowed advance of ice margins into a glaciomarine setting in southern Victoria Land. DISCUSSION Models for the Late Paleozoic Ice Age place Victoria Land beneath the center of an immense Gondwanan Ice Sheet that waxed and waned throughout the Mississippian, Pennsylvanian, and Permian (e.g., Scotese et al., 1999). These models predict that terrestrial ice flowed southward across southern Victoria Land out of a major glacial spreading center (Lindsay, 1970; Barrett, 1991; Veevers, 2001). However, lithofacies and paleocurrent data presented here do not support such a conclusion. Instead, the results of this study suggest that expansion of temperate glaciers, flowing out of smaller glacial centers, converged on southern Victoria Land and extended into and retreated out of a glaciomarine setting during the Early Permian. No evidence for Carboniferous glaciation occurs. Therefore, glaciogenic strata of the Metschel Tillite are inconsistent with deposition in this area from a single, massive, long-duration Gondwanan Ice Sheet. Recent work in other areas of Antarctica and Gondwana are also challenging the prevailing view of the extreme size and duration of late Paleozoic glaciation. In the central Transantarctic Mountains, ongoing facies and paleocurrent analyses do not support the traditional view of terrestrial ice flowing radially out of Victoria Land (Isbell et al., 1997, 2005, 2008; Isbell, 1999). Instead, results show that ice converged on an elongate basin whose long axis was oriented parallel to the present trend of the mountain range (Figs. 1 and 11). In the central Transantarctic Mountains, glaciers, grounded along basin margins in the direction of the present polar plateau and in the direction of the Ross Sea–Marie Byrd Land, flowed transversely off the margins into a glaciomarine setting. This scenario is similar to that proposed for southern Victoria Land, thus suggesting the occurrence of multiple Permian glacial centers in Antarctica (Isbell et al., 1997, 2005, 2008; Isbell, 1999). Elsewhere in Antarctica, interpretations of earlier work suggest that a glacial center on the Ellsworth Mountains crustal block supplied ice to glaciomarine settings in the Ellsworth and Pensacola Mountains (Fig. 11C; Frakes et al., 1971; Matsch and Ojakangas, 1991; Collinson et al., 1994). In Queensland and New South Wales, Australia, Jones and Fielding (2004), Birgenheier et al. (2005, 2009), and Fielding et al. (2005, 2008) reported the occurrence of short, discrete intervals of mountain-valley glaciations, ice caps, and/or small

Permian deposits, southern Victoria Land, Antarctica

C

A b

B

97

°

Figure 11. Tectonic transport direction for rocks of the Metschel Tillite as indicated by the orientation of (A) thrust faults and foliation in glaciotectonite and deformation till (Mount Metschel and Mount Ritchie) and the orientation of (B) thrust faults associated with thrust sheets (Kennar Valley, Mount Ritchie). (C) Map showing interpreted flow directions in southern Victoria Land, plotted with data from glaciogenic strata elsewhere in the Transantarctic and Ellsworth Mountains (EM), and in South Africa. Data from Frakes et al. (1971), Barrett (1981), Collinson et al. (1994), Visser (1997), Isbell (1999), Lenaker (2002). v.m.—vector mean.

ice sheets. Their findings are in marked contrast with earlier reports that suggested that much of Australia was covered by a continental-scale polar ice sheet. The hypothesis that numerous ice centers occurred in Gondwana during the late Paleozoic is not a new concept. Work by Crowell and Frakes (1970), Caputo and Crowell (1985), Eyles (1993), López Gamundí (1997), Limarino et al. (2002), Isbell et al. (2003), Henry et al. (2008), and Isbell et al. (2008) showed that multiple ice sheets, ice caps, and alpine glaciers diachronously waxed and waned as Gondwana drifted across the late Paleozoic South Pole. The glacial record in southern Victoria Land is consistent with the concept of multiple ice centers within Gondwana, and interpretation of the record disproves that Antarctica was covered by a single, massive ice sheet during the late Paleozoic. Identification of the size and duration of Gondwana glaciation is of great importance in developing an understanding of

Earth systems during the late Paleozoic. Data obtained from the Metschel Tillite clearly show that multiple glaciers were active in southern Victoria Land during the Permian. These results, coupled with recently reported data in Australia and South America, strongly suggest that Gondwana glaciation was characterized by numerous glacial centers and alpine glaciers rather than by a single massive ice sheet. Because the geographic area of ice cover, ice volume, and the number of ice sheets are directly related, multiple glaciers, for a given land area, contain considerably less ice volume than a single massive ice sheet (Crowley and Baum, 1991; Isbell et al., 2003). Therefore, multiple Gondwana glaciers would have had a completely different impact on Earth’s natural systems than that of a massive ice sheet. For example, the waxing and waning of multiple ice sheets would have resulted in considerably smaller changes in eustatic sea level than those produced by growth and decay of a single glacier covering the same geographical area.

98

Isbell

CONCLUSIONS Glaciotectonic deformation and glaciomarine deposits within the Metschel Tillite indicate that late Paleozoic glaciation was characterized by temperate glacial conditions along Permian ice margins in southern Victoria Land rather than by polar glacial conditions associated with a glacial spreading center as previously hypothesized. Deposition occurred beneath glaciers undergoing deforming bed conditions and in a glaciomarine setting in front of either alpine glaciers, outlet glaciers, ice streams, or surging glacial lobes. The direction of glaciotectonic transport indicates that ice converged on southern Victoria Land from two directions. One source of ice was from the direction of the present East Antarctic craton. The other source of ice was from the opposite direction, that of the present Ross Sea. These findings imply that multiple glaciers were active in this region. Because multiple ice sheets covering a given area contain less ice than a single ice sheet, multiple glaciers would have influenced Earth systems (e.g., eustatic sea level, climate) differently than the massive-ice-sheet models predict. ACKNOWLEDGMENTS Discussions with Rosemary Askin, Jim Collinson, Ellen Cowan, Dyana Czeck, Pete Flaig, Tom Hooyer, Mark Johnson, Carlos Oscar Limarino, and Molly Miller are greatly appreciated. I also thank Paul Lenaker, Rosemary Askin, Tim Cully, Molly Miller, and Keri Wolfe for their help in the field. Peter Barrett, Luis Buatois, Jim Collinson, Chris Fielding, and Antonio Carlos Rocha-Campos provided valuable comments on earlier drafts of this paper. The U.S. National Science Foundation, Raytheon Polar Services, the New York Air National Guard, Kenn Borek Air LTD, Trans World Logistics, and Petroleum Helicopters Incorporated provided logistic support for fieldwork in Antarctica. National Science Foundation grants OPP-9909637, ANT-0440919, ANT0635537, and OISE-0825617 supported this work. REFERENCES CITED Aber, J.S., Croot, D.G., and Fenton, M.M., 1989, Glaciotectonic Landforms and Structures: Dordrecht, Kluwer Academic Publishers, 200 p. Allen, J.R.L., 1985, Principles of Physical Sedimentology: London, George Allen & Unwin, 272 p. Alley, R.B., 1991, Deforming-bed origin for southern Laurentide till sheets?: Journal of Glaciology, v. 37, p. 67–76. Askin, R.A., 1998, Floral trends in the Gondwana high latitudes: Palynological evidence from the Transantarctic Mountains: Journal of African Earth Sciences, v. 27, p. 12–13. Banham, P.H., 1977, Glacitectonics in till stratigraphy: Boreas, v. 6, p. 101– 105. Barrett, P.J., 1972, Late Paleozoic glacial valley at Alligator Peak, southern Victoria Land, Antarctica: New Zealand Journal of Geology and Geophysics, v. 15, p. 262–268. Barrett, P.J., 1981, History of the Ross Sea region during the deposition of the Beacon Supergroup 400–180 million years ago: Journal of the Royal Society of New Zealand, v. 11, p. 447–458. Barrett, P.J., 1991, The Devonian to Jurassic Beacon Supergroup of the Transantarctic Mountains and correlatives in other parts of Antarctica, in

Tingey, R.J., ed., The Geology of Antarctica: Oxford, UK, Oxford University Press, p. 120–152. Barrett, P.J., and Kohn, B.P., 1975, Changing sediment transport directions from Devonian to Triassic in the Beacon Supergroup of South Victoria Land, Antarctica, in Campbell, K.S.W., ed., Gondwana Geology: Canberra, Australian National University Press, p. 15–35. Barrett, P.J., and Kyle, R.A., 1975, The early Permian glacial beds of southern Victoria Land and the Darwin Mountains, Antarctica, in Campbell, K.S.W., ed., Gondwana Geology: Canberra, Australian National University Press, p. 333–346. Barrett, P.J., and McKelvey, B.C., 1981, Permian tillites of southern Victoria Land, Antarctica, in Hambrey, M.J., and Harland, W.B., eds., Earth’s Pre-Pleistocene Glacial Record: Cambridge, UK, Cambridge University Press, p. 233–236. Barrett, P.J., Elliot, D.H., and Lindsay, J.F., 1986, The Beacon Supergroup (Devonian-Triassic) and Ferrar Group (Jurassic) in the Beardmore Glacier area, Antarctica, in Turner, M.D., and Splettstoesser, J.F., eds., Geology of the Central Transantarctic Mountains: Washington, D.C., American Geophysical Union, Antarctic Research Series, p. 339–428. Benn, D.I., 1995, Fabric signature of subglacial till deformation, Breidamerkurjökull, Iceland: Sedimentology, v. 42, p. 735–747, doi: 10.1111/ j.1365-3091.1995.tb00406.x. Benn, D.I., and Evans, D.J.A., 1996, The interpretation and classification of subglacially deformed materials: Quaternary Science Reviews, v. 15, p. 23–52, doi: 10.1016/0277-3791(95)00082-8. Benn, D.I., and Evans, D.J.A., 1998, Glaciers and Glaciation: London, Edward Arnold, 734 p. Bennett, M.R., and Glasser, N.F., 1996, Glacial Geology: Ice Sheets and Landforms: New York, Wiley & Sons, 364 p. Bennett, M.R., Glasser, N.F., Crawford, K., Hambrey, M.J., and Huddart, D., 1999, The landform and sediment assemblage produced by a tidewater glacier surge in Kongsfjorden, Svalbard: Quaternary Science Reviews, v. 18, p. 1213–1246. Birgenheier, L.P., Fielding, C.R., Frank, T.D., and Roberts, J., 2005, Stratigraphic record of late Paleozoic Gondwanan ice age in New South Wales, Australia: A review and revision of the Carboniferous System: Geological Society of America Abstracts with Programs, v. 37, no. 7, p. 256. Birgenheier, L.P., Fielding, C.R., Rygel, M.C., Frank, T.J., and Roberts, J., 2009, Evidence for dynamic climate change on sub-106-year scales from the late Paleozoic Glacial record, Tamworth Belt, New South Wales, Australia: Journal of Sedimentary Research, v. 79, p. 56–82. Bluemle, J.P., and Clayton, L., 1984, Large-scale glacial thrusting and related processes in North Dakota: Boreas, v. 13, p. 279–299. Boulton, G.S., 1979, Processes of glacier erosion on different substrata: Journal of Glaciology, v. 23, p. 15–38. Boulton, G.S., 1987, A theory of drumlin formation by subglacial sediment deformation, in Menzies, J., and Rose, J., eds., Drumlin Symposium: Rotterdam, A.A. Balkema, p. 25–80. Boulton, G.S., 1990, Sedimentary and sea level changes during glacial cycles and their control on glacimarine facies architecture, in Dowdeswell, J.A., and Scourse, J.D., eds., Glacimarine Environments: Processes and Sediments: Geological Society of London Special Publication 53, p. 15–52. Boulton, G.S., 1996, Theory of glacial erosion, transport and deposition as a consequence of subglacial sediment deformation: Journal of Glaciology, v. 42, p. 43–62. Boulton, G.S., and Caban, P., 1995, Groundwater flow beneath ice sheets: Part II—Its impact on glacier tectonic structures and moraine formation: Quaternary Science Reviews, v. 14, p. 563–587, doi: 10.1016/0277 -3791(95)00058-W. Boulton, G.S., and Hindmarsh, R.C.A., 1987, Sediment deformation beneath glaciers: Rheology and sedimentological consequences: Journal of Geophysical Research, v. 92, p. 9059–9082, doi: 10.1029/JB092iB09p09059. Boulton, G.S., and Jones, A.S., 1979, Stability of temperate ice caps and ice sheets resting on beds of deformable sediment: Journal of Glaciology, v. 24, p. 29–43. Collinson, J.D., and Thompson, D.B., 1989, Sedimentary Structures: London, Chapman & Hall, 207 p. Collinson, J.W., Isbell, J.L., Elliot, D.H., Miller, M.F., and Miller, J.M.G., 1994, Permian-Triassic Transantarctic basin, in Veevers, J.J., and Powell, C.M., eds., Permian-Triassic Pangean Basins and Foldbelts along the Panthalassan Margin of Gondwanaland: Geological Society of America Memoir 184, p. 173–222.

Permian deposits, southern Victoria Land, Antarctica Cowan, E.A., and Powell, R.D., 1990, Suspended sediment transport and deposition of cyclically interlaminated sediment in a temperate glacial fjord, Alaska, U.S.A., in Dowdeswell, J.A., and Scourse, J.D., eds., Glacimarine Environments: Processes and Sediments: Geological Society of London Special Publication 53, p. 75–89. Cowan, E.A., and Powell, R.D., 1991, Ice-proximal sediment accumulation rates in a temperate glacial fjord, south-eastern Alaska, in Anderson, J.B., and Ashley, G.M., eds., Glacial Marine Sedimentation: Paleoclimatic Significance: Geological Society of America Special Paper 261, p. 61–73. Croot, D.G., 1988, Morphological, structural and mechanical analysis of neoglacial ice-pushed ridges in Iceland, in Croot, D.G., ed., Glaciotectonics: Forms and Processes: Rotterdam, A.A. Balkema, p. 33–47. Crowell, J.C., and Frakes, L.A., 1970, Ancient Gondwana glaciations, in Haughton, S.H., ed., Proceedings and Papers of the Second Gondwana Symposium, South Africa: Pretoria, CSIR, p. 469–476. Crowell, J.C., and Frakes, L.A., 1971, Late Paleozoic glaciation: Part IV, Australia: Geological Society of America Bulletin, v. 82, p. 2515–2540, doi: 10.1130/0016-7606(1971)82[2515:LPGPIA]2.0.CO;2. Crowley, T.J., and Baum, S.K., 1991, Estimating Carboniferous sea-level fluctuations from Gondwana ice extent: Geology, v. 19, p. 975–977, doi: 10 .1130/0091-7613(1991)0192.3.CO;2. Cúneo, N.R., Isbell, J.L., Taylor, T.N., and Taylor, E.L., 1993, The Glossopteris Flora in Antarctica: Taphonomy and paleoecology, C.R.: Buenos Aires, International Congress of Carboniferous and Permian Stratigraphic Geology, 12th, p. 13–40. Domack, E.W., Jacobson, E.A., Shipp, S., and Anderson, J.B., 1999, Late Pleistocene–Holocene retreat of the West Antarctic Ice-Sheet system in the Ross Sea: Part 2—Sedimentologic and stratigraphic signature: Geological Society of America Bulletin, v. 111, p. 1517–1536, doi: 10.1130/ 0016-7606(1999)1112.3.CO;2. Dowdeswell, J.A., Whittington, R.J., and Marienfeld, P., 1994, The origin of massive diamicton facies by iceberg rafting and scouring, Scoresby Sund, East Greenland: Sedimentology, v. 41, p. 21–35, doi: 10.1111/j.1365-3091 .1994.tb01390.x. Dowdeswell, J.A., Mackensen, A., Marienfeld, P., Whittington, R.J., Jennings, A.E., and Andrews, J.T., 2000, An origin for laminated glacimarine sediments through sea-ice build-up and suppressed iceberg rafting: Sedimentology, v. 47, p. 557–576, doi: 10.1046/j.1365-3091.2000.00306.x. Evans, J., and Pudsey, C.J., 2002, Sedimentation associated with Antarctic Peninsula ice shelves; implications for palaeoenvironmental reconstructions of glacimarine sediments: Journal of the Geological Society [London], v. 159, p. 233–237, doi: 10.1144/0016-764901-125. Evans, P.R., 1969, Upper Carboniferous and Permian palynological stages and their distribution in eastern Australia, in Amos, A.J., ed., Gondwana Stratigraphy: Paris, UNESCO, p. 41–54. Eybergen, F.A., 1987, Glacier snout dynamics and contemporary push moraine formation at the Turtmannglacier, Wallis, Switzerland, in Van Der Meer, J.J.M., ed., Tills and Glaciotectonics: Proceedings of the International Union for Quaternary Research (INQUA) Symposium, Amsterdam, 1986: Rotterdam, A.A. Balkema, p. 217–231. Eyles, N., 1993, Earth’s Glacial Record and Its Tectonic Setting: Earth-Science Reviews, v. 35, 248 p., doi: 10.1016/0012-8252(93)90002-O. Fielding, C., Frank, T., Birgenheier, L., Thomas, S., Rygel, M., and Jones, A., 2005, Revised Permian glacial record of eastern Australia: Geological Society of America Abstracts with Programs, v. 37, no. 7, p. 256. Fielding, C.R., Frank, T.D., Birgenheier, L.P., Rygel, M.C., Jones, A.T., and Roberts, J., 2008, Stratigraphic imprint of the late Palaeozoic ice age in eastern Australia: A record of alternating glacial and nonglacial climate regime: Journal of the Geological Society, London, v. 165, p. 129–140. Frakes, L.A., Matthews, J.L., and Crowell, J.C., 1971, Late Paleozoic glaciation: Part III, Antarctica: Geological Society of America Bulletin, v. 82, p. 1581–1604, doi: 10.1130/0016-7606(1971)82[1581:LPGPIA ]2.0.CO;2. Hart, J.K., 1990, Proglacial glaciotectonic deformation and the origin of the Cromer Ridge push moraine complex, North Norfolk, England: Boreas, v. 19, p. 165–180. Hart, J.K., and Boulton, G.S., 1991, The interrelation of glaciotectonic and glaciodepositional processes within the glacial environment: Quaternary Science Reviews, v. 10, p. 335–350, doi: 10.1016/0277-3791(91)90035-S. Helby, R.J., and McElroy, C.T., 1969, Microfloras from the Devonian and Triassic of the Beacon Supergroup, Antarctica: New Zealand Journal of Geology and Geophysics, v. 12, p. 376–383.

99

Henry, L.C., Isbell, J.L., and Limarino, C.O., 2008, Carboniferous glacigenic deposits of the proto-Precordillera of west-central Argentina, in Fielding, C.R., Frank, T.D., and Isbell, J.L., eds., Resolving the Late Paleozoic Ice Age in Time and Space: Geological Society of America Special Paper 441, p. 131–142. Hyde, W.T., Crowley, T.J., Tarasov, L., and Paltier, W.R., 1999, The Pangean ice age: Studies with a coupled climate–ice sheet model: Climate Dynamics, v. 15, p. 619–629, doi: 10.1007/s003820050305. Isbell, J.L., 1999, The Kukri Erosion Surface; a reassessment of its relationship to rocks of the Beacon Supergroup in the central Transantarctic Mountains, Antarctica: Antarctic Science, v. 11, p. 228–238, doi: 10.1017/ S0954102099000292. Isbell, J.L., Gelhar, G.A., and Seegers, G.M., 1997, Reconstruction of preglacial topography using a post-glacial flooding surface: Upper Paleozoic deposits, central Transantarctic Mountains, Antarctica: Journal of Sedimentary Research, v. 67, p. 264–272. Isbell, J.L., Miller, M.F., Wolfe, K.L., and Lenaker, P.A., 2003, Timing of late Paleozoic glaciation in Gondwana: Was glaciation responsible for the development of northern hemisphere cyclothems?, in Chan, M.A., and Archer, A.W., eds., Extreme Depositional Environments: Mega End Members in Geologic Time: Geological Society of America Special Paper 370, p. 5–24. Isbell, J.L., Miller, M.F., Askin, R.A., Lenaker, P.A., and Koch, Z.J., 2005, Late Paleozoic glaciation in Antarctica: Are models depicting an immense ice sheet correct?: Geological Society of America Abstracts with Programs, v. 37, no. 7, p. 257. Isbell, J.L., Koch, Z.J., Szablewski, G.M., and Lenaker, P.A., 2008, Permian glacigenic deposits in the Transantarctic Mountains, Antarctica, in Fielding, C.R., Frank, T.D., and Isbell, J.L., eds., Resolving the Late Paleozoic Ice Age in Time and Space: Geological Society of America Special Paper 441, p. 59–70. Jones, A.T., and Fielding, C.R., 2004, Sedimentological record of the late Paleozoic glaciation in Queensland, Australia: Geology, v. 32, p. 153–156, doi: 10.1130/G20112.1. Kyle, R.A., 1976, Palaeobotanical studies of the Permian and Triassic Victoria Group (Beacon Supergroup) of south Victoria Land, Antarctica [Ph.D. thesis]: Wellington, New Zealand, Victoria University of Wellington, 306 p. Kyle, R.A., 1977, Palynostratigraphy of the Victoria Group of south Victoria Land, Antarctica: New Zealand Journal of Geology and Geophysics, v. 20, p. 1081–1102. Kyle, R.A., and Schopf, J.M., 1982, Permian and Triassic palynostratigraphy of the Victoria Group, Transantarctic Mountains, in Craddock, C., ed., Antarctic Geosciences: Madison, University of Wisconsin Press, International Union of Geological Sciences, p. 649–659. Lenaker, P.A., 2002, Sedimentology of Permian glacial deposits in the Darwin Glacier region, Antarctica [M.S. thesis]: Milwaukee, University of Wisconsin–Milwaukee, 173 p. Limarino, C.O., Césari, S.N., Net, L.I., Marenssi, S.A., Gutierrez, P.R., and Tripaldi, A., 2002, The Upper Carboniferous postglacial transgression in the Paganzo and Río Blanco basins (northwestern Argentina): Facies and stratigraphic significance: Journal of South American Earth Sciences, v. 15, p. 445–460, doi: 10.1016/S0895-9811(02)00048-2. Lindsay, J.F., 1970, Depositional environment of Paleozoic glacial rocks in the central Transantarctic Mountains: Geological Society of America Bulletin, v. 81, p. 1149–1172, doi: 10.1130/0016-7606(1970)81[1149:DEOPGR ]2.0.CO;2. Lindström, S., 1995, Early Permian palynostratigraphy of the northern Heimefrontfjella mountain-range, Dronning Maud Land, Antarctica: Review of Palaeobotany and Palynology, v. 89, p. 359–415, doi: 10.1016/0034 -6667(95)00058-3. López-Gamundí, O.R., 1997, Glacial-postglacial transition in the Late Paleozoic basins of southern South America, in Martini, I.P., ed., Late Glacial and Postglacial Environmental Changes: Quaternary, Carboniferous–Permian, and Proterozoic: Oxford, UK, Oxford University Press, p. 147–168. Mackiewicz, N.E., Powell, R.D., Carlson, P.R., and Molina, B.F., 1984, Interlaminated ice-proximal glacimarine sediments in Muir Inlet, Alaska: Marine Geology, v. 57, p. 113–147, doi: 10.1016/0025-3227(84)90197-X. Matsch, C.L., and Ojakangas, R.W., 1991, Comparison in depositional style of “polar” and “temperate” glacial ice; late Paleozoic Whiteout Conglomerate (West Antarctica) and late Proterozoic Mineral Fork Formation (Utah), in Anderson, J.B., and Ashley, G.M., eds., Glacial Marine Sedimentation;

100

Isbell

Paleoclimatic Significance: Geological Society of America Special Paper 261, p. 191–206. McElroy, C.T., and Rose, G., 1987, Geology of the Beacon Heights area, southern Victoria Land, Antarctica: New Zealand Geological Survey Miscellaneous Series Map 15 and Notes, 47 p., scale 1:50,000. McKelvey, B.C., Webb, P.N., and Kohn, B.P., 1972, Stratigraphy of the Beacon Supergroup between the Olympus and Boomerang Ranges, Victoria Land, in Adie, R.J., ed., Antarctic Geology and Geophysics: Oslo, Universitetsforlaget, p. 345–352. McKelvey, B.C., Webb, P.N., and Kohn, B.P., 1977, Stratigraphy of the Taylor and lower Victoria Groups (Beacon Supergroup) between the Mackay Glacier and Boomerang Range, Antarctica: New Zealand Journal of Geology and Geophysics, v. 20, p. 813–863. McPherson, J.G., 1978, Stratigraphy and sedimentology of the Upper Devonian Aztec Siltstone, southern Victoria Land, Antarctica: New Zealand Journal of Geology and Geophysics, v. 21, p. 667–683. McPherson, J.G., 1979, Calcrete (caliche) paleosols in fluvial red-beds of the Aztec Siltstone (Upper Devonian), southern Victoria Land, Antarctica: Sedimentary Geology, v. 22, p. 267–285, doi: 10.1016/0037-0738(79 )90056-3. Powell, C.M., and Li, Z.X., 1994, Reconstruction of the Panthalassan margin of Gondwanaland, in Veevers, J.J., and Powell, C.M., eds., Permian-Triassic Pangean Basins and Foldbelts along the Panthalassan Margin of Gondwanaland: Geological Society of America Memoir 184, p. 5–9. Powell, R.D., 1990, Glacimarine processes at grounding-line fans and their growth to ice contact deltas, in Dowdeswell, J.A., and Scourse, J.D., eds., Glacimarine Environments: Processes and Sediments: Geological Society [London] Special Publication 53, p. 53–73. Powell, R.D., and Alley, R.B., 1997, Grounding-line systems: Processes, glaciological inferences and the stratigraphic record, in Barker, P.F., and Cooper, A.C., eds., Geology and Seismic Stratigraphy of the Antarctic Margin, 2: Washington, D.C., American Geophysical Union, Antarctic Research Series, v. 71, p. 169–187. Powell, R., and Domack, E., 2002, Modern glaciomarine environments, in Menzies, J., ed., Modern and Past Glacial Environments: Oxford, UK, Butterworth-Heinemann, p. 361–389. Pyne, A.R., 1984, Geology of the Mt. Fleming area, South Victoria Land, Antarctica: New Zealand Journal of Geology and Geophysics, v. 27, p. 505–512. Ricci Lucchi, F., 1995, Sedimentographica: A Photographic Atlas of Sedimentary Structures: New York, Columbia University Press, 255 p. Ritchie, A.R., 1975, Groenlandaspis in Antarctica, Australia and Europe: Nature, v. 254, p. 569–573, doi: 10.1038/254569a0. Rocha-Campos, A.C., Canuto, J.R., and dos Santos, P.R., 2000, Late Paleozoic glaciotectonic structures in northern Paraná Basin, Brazil: Sedimentary Geology, v. 130, p. 131–143, doi: 10.1016/S0037-0738(99)00110-4. Scotese, C.R., Boucot, A.J., and McKerrow, W.S., 1999, Gondwanan palaeogeography and palaeoclimatology: Journal of African Earth Sciences, v. 28, p. 99–114, doi: 10.1016/S0899-5362(98)00084-0. Smith, L.M., and Andrews, J.T., 2000, Sediment characteristics in iceberg dominated fjords, Kangerlussuaq region, East Greenland: Sedimentary Geology, v. 130, p. 11–25, doi: 10.1016/S0037-0738(99)00088-3.

Spörli, K.B., 1992, Stratigraphy of the Crashsite Group, Ellsworth Mountains, West Antarctica, in Webers, G.F., Craddock, G.F., and Splettstoesser, J.F., eds., Geology of the Ellsworth Mountains, Antarctica: Geological Society of America Memoir 170, p. 21–35. Stow, D.A.V., 2005, Sedimentary Rocks in the Field: A Colour Guide: London, Elsevier Academic Press, 320 p. Thomas, G.S.P., and Connell, R.J., 1985, Iceberg drop, dump, and grounding structures from Pleistocene glacio-lacustrine sediments, Scotland: Journal of Sedimentary Petrology, v. 55, p. 243–249. Tucker, M.E., 1996, Sedimentary Rocks in the Field: New York, Wiley & Sons, 133 p. Turner, S., and Young, G.C., 1992, Thelodont scales from the middle–late Devonian Aztec Siltstone, southern Victoria Land, Antarctica: Antarctic Science, v. 4, p. 89–105, doi: 10.1017/S0954102092000142. van der Meer, J.J.M., Menzies, J., and Rose, J., 2003, Subglacial till; the deforming glacier bed: Quaternary Science Reviews, v. 22, p. 1659–1685, doi: 10.1016/S0277-3791(03)00141-0. Van der Wateren, F.M., 1987, Structural geology and sedimentology of the Dammer Berge push moraine, FGR, in van Der Meer, J.J.M., ed., Tills and Glaciotectonics: Rotterdam, A.A. Balkema, p. 157–182. Van der Wateren, F.M., 2002, Processes of glaciotectonism, in Menzies, J., ed., Modern and Past Glacial Environments: Oxford, UK, ButterworthHeinemann, p. 417–443. Veevers, J.J., 1994, Case for the Gamburtsev Subglacial Mountains of East Antarctica originating by mid-Carboniferous shortening of an intracratonic basin: Geology, v. 22, p. 593–596, doi: 10.1130/0091-7613(1994)022 2.3.CO;2. Veevers, J.J., 2001, Atlas of Billion-Year Earth History of Australia and Neighbours in Gondwanaland: Sydney, GEMOC Press, 76 p. Visser, J.N.J., 1997, A review of the Permo-Carboniferous glaciation in Africa, in Martini, I.P., ed., Late Glacial and Postglacial Environmental Changes: Quaternary, Carboniferous–Permian, and Proterozoic: Oxford, UK, Oxford University Press, p. 169–191. Visser, J.N.J., and Loock, J.C., 1982, An investigation of the basal Dwyka Tillite in the southern part of the Karoo Basin, South Africa: Transactions—Geological Society of South Africa, v. 85, p. 179–187. Young, G.C., 1988, Antiarchs (placoderm fishes) from the Devonian Aztec Siltstone, southern Victoria Land, Antarctica: Palaeontographica, v. 202, 125 p. Young, G.C., 1989, The Aztec fish fauna (Devonian) of southern Victoria Land: Evolutionary and biogeographic significance, in Crame, J.A., ed., Origins and Evolution of the Antarctic Biota: Geological Society [London] Special Publication 47, p. 43–63. Young, G.C., 1991, Fossil fishes from Antarctica, in Tingey, R.J., ed., The Geology of Antarctica: Oxford, UK, Oxford University Press, p. 538–567. Ziegler, A.M., Hulver, M.L., and Rowley, D.B., 1997, Permian world topography and climate, in Martini, I.P., ed., Late Glacial and Postglacial Environmental Changes: Quaternary, Carboniferous-Permian, and Proterozoic: Oxford, UK, Oxford University Press, p. 111–146. MANUSCRIPT ACCEPTED BY THE SOCIETY 21 DECEMBER 2009

Printed in the USA

The Geological Society of America Special Paper 468 2010

Formation of euxinic lakes during the deglaciation phase in the Early Permian of East Africa Thomas Kreuser Gebretinsae Woldu Geology Department, University of Asmara, P.O. Box 1220, Eritrea

ABSTRACT The continental glaciation of Gondwanaland in the Late Carboniferous–Early Permian left traces in many places in southern and eastern Africa. This paper focuses on the last glacial advance and consecutive deglaciation leading to the formation of large euxinic lakes with high concentrations of organic matter. The Idusi Formation in the Tanzanian Ruhuhu Basin (initiating the Karoo cycle, which extends into the Triassic) provides the type section for this depositional sequence. It is subdivided into a lower Lisimba Member, the basal unit of glacial origin, and an upper Lilangu Member, characterized by postglacial black shale and rhythmites as evidence of a climatic amelioration on a large regional scale in Africa. Thickness and facies variations are attributed to a pronounced paleotopography as the result of scouring glaciers and local tectonic events. There is a gradual change between the members, reflecting a continuous climatic amelioration and change of sediment supply. The lacustrine environment was terminated by the onset of braided stream deposition (Mpera Sandstone Member); an erosional unconformity between the units marks the start of initial rifting in the Early Permian. This is followed by the development of extensive coal swamps in a temperate climate, where organic matter predominated over clastic supply. Periglacial deposits with tillites and rhythmites, containing dropstones, are overlain by glaciolacustrine laminites intercalated with glaciofluvial marginal deltaic sediments. Deglaciation provided water and accommodation space for the evolution of extensive anaerobic stratified lakes, which were the focus of prolific deposition of organic matter. This black shale may contain up to 11% TOC (total organic carbon) content. Eventually, the lake became shallower and was succeeded by alluvial fan deposition. The duration of the glaciation and deglaciation was ~20–25 m.y., and the lacustrine phase lasted ~4–5 m.y. These ages have been verified by palynology (Granulatisporites confluens Oppel zone). The hydrocarbon potential of the black shale was estimated by Rock-Eval pyrolyses. Hydrogen index, maximum temperature (Tmax), and vitrinite reflection were used to determine kerogen type, maturity stage, and subsidence history. A promising potential with respect to gaseous hydrocarbon generation was detected from both the euxinic black shale and the overlying coals. A comparison with other Tanzanian

Kreuser, T., and Woldu, G., 2010, Formation of euxinic lakes during the deglaciation phase in the Early Permian of East Africa, in López-Gamundí, O.R., and Buatois, L.A., eds., Late Paleozoic Glacial Events and Postglacial Transgressions in Gondwana: Geological Society of America Special Paper 468, p. 101–112, doi: 10.1130/2010.2468(04). For permission to copy, contact [email protected]. ©2010 The Geological Society of America. All rights reserved.

101

102

Kreuser and Woldu Karoo basins reveals similar conditions in TOC values and temperature history. The wide regional extent of the anaerobic lacustrine black shale of the deglaciation event in several eastern and southern African basins evinces a similar climatic and regional tectonic framework in the pre-breakup phase of Gondwanaland during the Early Permian. This period of time may be of some importance in the future when the economic potential with respect to hydrocarbon generation of the Permian basins is scrutinized in more detail.

INTRODUCTION The present work is a general review of the results of a working group from the University of Cologne, Germany, and the University of Dar es Salaam, Tanzania, during a research period of more than 15 yr. Therefore, most of the findings and interpretations given in the following chapters have already been published elsewhere. The focus of this paper was mainly to collect data from different working groups and combine them into a general description of the late Paleozoic glacial deposition, the onset of Permian rifting, and the associated history of lacustrine deposition which initiated the onset of the Karoo Supercycle in southern and eastern Africa. Continental glaciation in Late Carboniferous to Early Permian times occurred at numerous locations on the African continent and other former Gondwanan continents. In Africa, glacial and postglacial deposition represents the onset of the Karoo depositional cycle, which continued until the Early Triassic. In South Africa the glacial deposits are referred to as the Dwyka Formation, which has been studied intensively by numerous authors (Rust 1975; Martin, 1981; Visser, 1989); in Tanzania the succession was described by Wopfner and Kreuser (1986), Kreuser (1987), and Wopfner and Diekmann (1996). In East Africa the glaciogene sediments are best exposed in southwestern Tanzania, where the early Karoo deposits are exposed along the uplifted rift shoulders of Lake Malawi and Lake Rukwa. This paper summarizes the lithological and depositional history of a number of Tanzanian basins: the Ruhuhu, Songwe Kiwira, and Galula Basins, with respect to their glacial and postglacial depositional history. Special emphasis is placed on the accumulation of organic matter in large periglacial and postglacial lakes and the thermal history of these organic source rocks. Additionally, a stratigraphic approach is presented to compare the Tanzanian sections with other southern and eastern African localities of comparable age and focus on the ubiquitous onset of glacial and postglacial deposits in the region. The glacial nature of these deposits was first established by Spence (1957). Later, Wopfner and Kreuser (1986), Kreuser (1987), Diekmann (1993), Wopfner and Diekmann (1992), and Wopfner and Diekmann (1996) described these sequences in detail and established a modern nomenclature and a depositional model. The most detailed lithologic investigation with respect to facies distribution and regional differentiation was performed by Diekmann (1993), when the formal lithostratigraphic terminology was established. Geochemical analysis of the nature and

thermal history of Karoo source rocks was performed by Kreuser et al. (1988), Dypvik et al. (1990), Diekmann (1993), and Kreuser (1995b). The present paper summarizes the most important features and incorporates a model of the depositional environment and thermal history of these potential source rocks for hydrocarbon generation. Within the Karoo succession, several phases of lacustrine development occurred in which organic matter accumulated (Kreuser, 1995b); however, only the periglacial to postglacial development is highlighted in this paper. Similar sedimentary successions were recorded from neighboring countries in Africa: South Africa (Rust, 1975), Madagascar (Besairie, 1972), Congo (Boutakoff, 1948), Ethiopia (Worku and Astin, 1992), and Oman (Qidwai, 1988), which are mentioned here for comparative purposes. The Tanzanian succession serves as a reference section, which is comparatively well described and analyzed. STRATIGRAPHY The southwestern Tanzanian Karoo basins comprise diverse facies successions of glacial and postglacial deposits (Fig. 1): 1. Ruhuhu Basin (subdivided into a western Mchuchuma sub-basin and an eastern Ngaka sub-basin) along the eastern fault scarp of the Malawi Rift. Post-Karoo movements have dissected the Ruhuhu Basin into several half grabens, i.e., the Lumecha sub-basin. Owing to the southeastern tilt of the half grabens, outcrops of the Idusi Formation lie mainly on the northern and western margins of these blocks. 2. Songwe-Kiwira Basin, situated northwest of the Ruhuhu Basin. 3. Galula Basin, along the southern boundary of the Rukwa Rift (Dypvik et al., 1990; Mliga, 1994). The type section of the glaciogenic sequence in Tanzania is in the Ruhuhu Basin in the gorge of the Idusi River in the northwestern Mchuchuma sub-basin and is known as the Idusi Formation (Kreuser et al., 1990; Fig. 2). Comparable stratotypes were detected along the Ketewaka River in the northeastern Mchuchuma sub-basin and along the Nyamangami River of the western Ngaka sub-basin (Fig. 2). Additional sections of the Idusi Formation were obtained from logs of exploration boreholes drilled by CDC (Colonial Development Corp.) and MADINI (Geological Survey of Tanzania) and were described by Wopfner and Diekmann (1996).

Formation of euxinic lakes, East Africa 30˚

103

34˚

38˚E

AFRICA

O

UGANDA

CO

NG

Tanzania Lake Victoria

RWANDA

Nairobi 2˚S

Kenya basin

DEM. REP

BURUNDI

Tanga basin

INDIAN OCEAN

TANZANIA Lake Tanganyika

ZANZIBAR

Mikumi basin

Namwele-Mkomolo

6˚

Dares Salaam Ruvu Valley

Rukwa basin Muse Lake Rukwa

Galula Songwe-Kiwira

Karoo basins

Ruhuhu basin

0

ZAMB

Border

250

500 km

Lake Nyasa

IA

Faults

WI MALA

Post-Karoo basins

Lu w tro egu ug h

LEGEND

Rufiji valley

Ruvuma basin

10˚

Njuga Mhukuru

Mbamba Bay

Maniamba basin

MOZAMBIQUE

Figure 1. Regional distribution of continental Karoo (Permian-Triassic) rocks in Tanzania and neighboring countries.

At the type locality in the Idusi River gorge the Idusi Formation reaches a thickness of 240 m. Because of facies variations, Diekmann (1993) subdivided the Idusi Formation into two lithostratigraphic units, a lower Lisimba Member and an upper Lilangu Member. Lisimba Member This member is named after a tributary of the Idusi River (thickness, 170 m, Fig. 3). The lower succession consists of different lithotypes of diamictite, conglomerate, and sandstone as well as silty mudstone, exhibiting large dropstones and lonestones (Diekmann, 1996). Higher in the succession, laminites are dominant in a mudstone matrix with abundant lonestones. The Lisimba Member exhibits a typical olive green color caused by abundant chlorite and has an arkosic composition. The thickness ranges up to 420 m owing to a strong paleorelief of the Pre-

cambrian basement (Kreuser, 1987). No palynological evidence has been found in the lower Lisimba Member, and owing to the stratigraphic position of the younger Lilangu Member, the lower member may have been deposited in Late Carboniferous times (Diekmann, 1996). The basal unit of the Lisimba Member consists of massive diamictite, grading into faintly bedded diamictite. Higher in the section, siltstone with dropstones up to 60 cm in diameter is present. The middle portion of the section consists of siltstone and mudstone with slumping structures. The upper part consists of olive green laminites with a few dropstones and sandstones. Lilangu Member Named after a tributary of the Idusi River (thickness, 70 m, Fig. 3), the Lilangu Member developed gradually from the glaciogenic succession beneath and is separated locally by an

104

Kreuser and Woldu

Figure 2. Distribution of sub-basins and geology of glacial and postglacial deposits in the Ruhuhu Basin. Note the fault-controlled basin margins along the northeast trending blocks.

unconformity from the overlying Mchuchuma Formation (Semkiwa, 1992). The boundary between the Lisimba and Lilangu Members is characterized by the first black fissile shale that is highly pyritic, containing calcareous concretions (Wopfner and Diekmann, 1996). Above this are black organic-rich siltstones, locally bituminous, intercalated with sandstone lobes and breccias. The amount of organic matter is high (kerogen types III and IV, from 3% to 6%). Palynological analyses indicate assem-

blages assigned to the Granulatisporites confluens Oppel zone, indicating a late Asselian to early Sakmarian age (Wopfner and Kreuser, 1986). Regional Lithological Variations The reference section along the Ketewaka River of the Idusi Formation measures ~100 m in thickness and exhibits faceted

Formation of euxinic lakes, East Africa

105

Figure 3. Composite log of the Idusi Formation at the type locality in the Idusi gorge and the northwestern Mchuchuma sub-basin. Legend is attached (for location, see Fig. 2). Based on Wopfner and Diekmann (1996).

and striated diamictites at the base and continues in a similar way along the Idusi River into laminites farther upward (Fig. 4). The Lilangu Member, however, exhibits a different facies development. Although the dominant black, organic-rich shale is present, intercalations with sandy lobes, several of which attain 10 m in thickness, are present, with an overall increased thickness of 120 m. Diekmann (1993) termed these sediments the Muhimbi facies, named after a tributary of the Ketewaka River. In the Ngaka sub-basin (Fig. 1) the “long section” of the Idusi Formation reaches a thickness of >700 m. In contrast to correlative lithologic successions, the lower 140 m is characterized

by sandstone with cross-bedding, plus siltstone and conglomerate that contain cobble-size lonestones. Diekmann (1993) termed this section the Ndongosi facies after the escarpment it forms, and interpreted it as glaciofluvial in origin. This is overlain by 160 m of massive deltaic sandstones, referred to as the Nanderuka facies (Diekmann, 1993). The Lilangu Member reaches a maximum thickness of ~300 m; common intercalations of mudstone are present, and a generally higher sand to shale ratio is observed. Both south and north of the “long section” the Idusi Formation is characterized by a considerably condensed succession with a thickness of up to 40 m. For that reason, other authors have used the term “short section” (McKinlay, 1954). The threefold lithologic subdivision and most of the described lithofacies are recognizable (Fig. 5). However, a section north of the Mkapa River exhibits a massive basal diamictite with faceted and striated clasts of various diameters resting on basement, interpreted as a lodgment till (Diekmann, 1996). It is followed by reworked diamictite and green bedded sandstones. The Lilangu Member has the typical black shales at the base, but the upper part is usually absent, probably having been removed by erosion prior to deposition of the Mchuchuma Formation (Diekmann, 1996). Songwe Kiwira Basin Approximately 50 km west of the northern end of Lake Malawi (Fig. 1), another small basin appears, named after the

106

Kreuser and Woldu paleovalley filled with basal diamictites interfingering with sandstone and conglomerate, capped by a few meters of laminites. Galula Basin This small basin, along the southwestern flank of the Rukwa Rift (Fig. 1), is an erosional remnant of Karoo strata. Here, Karoo strata are preserved along the depressed side of tilted fault blocks. Mbede (1993) noted seismic surveys indicating that in the deepest part of the Rukwa graben, some 7 km below ground level, 5 km of Karoo strata is present. There is only one outcrop of the Idusi Formation along the Chizi River, of ~13 m thickness (Wopfner and Diekmann, 1996). The succession begins with a basal diamictite, overlain by laminites of the Kipololo facies, topped by laminated mudstones with microclasts. The Lilangu Member is not developed here. Mhukuru Basin Located 65 km south-southwest of Songea near the Mozambique border (Fig. 1), this basin contains a well-developed succession of the Idusi Formation, with a maximum thickness of ~120 m. The Lisimba Member exhibits basal diamictites with dropstones and sandstones and green laminites. The overlying Lilangu Member consists of 20 m of black shales (Diekmann, 1993). DEPOSITIONAL ENVIRONMENT Lisimba Member

Figure 4. Log of the Ketewaka section in the northwestern Mchuchuma sub-basin. See Figure 3 for legend. Based on Wopfner and Diekmann (1996).

main rivers Kiwira and Songwe, which cross it. The southern part of the basin continues into Malawi, exhibiting an originally larger areal extent of Karoo deposits. The northern part of the basin is covered by Quaternary basalts of the Rungwe volcanics. The Idusi Formation is restricted to parts of the Lisimba Member, starting with a basal diamictite with cobbles that locally reach 30 cm in diameter. This unit is interpreted as a lodgment tillite (Dypvik et al., 1990). Facies variations can be studied in borehole sections drilled by the Coalfield Exploration Team of the Peoples Republic of China (Anonymous, 1979). Wopfner and Diekmann (1996) presented a depositional model that shows a 60-m-deep

The basal diamictites consist of massive to faintly bedded, poorly sorted mixtures of clasts resting in a sandy-silty matrix. Clasts, derived from local basement rocks, are faceted and locally striated, which is characteristic for glacial transport (Wopfner and Kreuser, 1986). In this paper, diamictites are interpreted as subaqueous melt-out and flow tillites, intercalated with meltwater deposits. Lodgment tills rarely occur. Depositional evidence for this interpretation is given in detail in Wopfner and Kreuser (1986). In the long section they are replaced by an upwardly fining succession, starting with conglomerate and ending with mudstone. This is interpreted as a transition from braided-river to subaquatic conditions in a proglacial environment (Diekmann, 1993; Fig. 6). The dominant lithology of the Lisimba Member is characterized by silty and sandy mudstones containing rare lonestones with diameters up to 80 cm. The lonestones are interpreted as ice-rafted dropstones. The fine clastics were settling from meltwater plumes as suspended particles. A few ripple laminations and sandy layers are current indicators that were formed during subaqueous flows (Diekmann, 1996). Laminites consist of alternating silt-clay couplets (8–15 cm) with common deformation structures, and some microclasts and rare dropstones are also present, which were deposited as periglacial varves during seasonal changes from a thermally stratified

Formation of euxinic lakes, East Africa

107

Figure 5. Correlation chart of three sections of the Idusi Formation in the Ruhuhu Basin, displaying a typical threefold subdivision. In addition to the depositional environments, the successive climatic changes are seen at the right. Modified from Diekmann (1993).

water column (Diekmann, 1993). Diekmann interpreted the laminites as sandy layers from proximal positions that developed during meltwater discharge during times of episodically slightly higher temperatures (Fig. 6). Lilangu Member No sharp boundary could be discerned between the Lisimba and Lilangu Members. A transitional zone exists where the silt gradually disappears and is replaced by black to olive gray shale with a high TOC (total organic carbon) content of up to 0.5%. Early diagenetic concretions are commonly observed wherein pyrite replaces plant remains. These basal shales represent a deepwater lacustrine environment during a postglacial stage (Diekmann, 1993). Overlying the basal shales are black, organic-rich siltstones, reflecting a high amount of terrestrial organic debris that consists of kerogen types III and IV. TOC values of up to 12% were

detected. These siltstones contain fossil-wood remains and remnants of Gangamopteris. Rhythmites are intercalated with the massive siltstones and consist of couplets of silt and medium sand laminae. This succession is interpreted as sapropel-rich deposits of a deep-lake environment with restricted circulation in a thermally stratified water column (Diekmann, 1993). Rare dropstones probably signal seasonal ice rafts on the lake, with common graded bedding. Some lenticular sandstone bodies are intercalated into the black siltstones, which exhibit sole marks (flute and groove casts) and current lineations. These sandstones are interpreted as turbidites accumulating at the apex of subaquatic channels on delta fronts (Diekmann, 1993; Fig. 6). Toward the top of the member an increase in sand-size grains was detected, which developed parallel with a decrease in TOC. Wopfner and Diekmann (1996) interpreted these silty sandstones as part of a shallow lake environment, which alternated with upper-flow-regime sandstone deposits accompanied by lag deposits with common mud clasts.

108

Kreuser and Woldu

Figure 6. Paleogeographic model of the depositional environment during deglaciation time. (A) Periglacial setting with ice rafts (1), valley glaciers (2), glacial outwash (3), reworked glaciofluvial residue (4). (B) Postglacial lake model with laminites, rhythmites, lakeshore flora, and associated marginal deltaic lobes.

PALEOGEOGRAPHY Records of the Permo-Carboniferous glaciation in Africa are in most cases incomplete, and commonly only the last glacial advance was spared from erosion. Deeply incised paleovalleys around the centers of glaciation are known from several localities, i.e., from South Africa (Visser, 1989, 1997), Namibia (Martin, 1981), and Congo (Boutakoff, 1948). From other Gondwanan continents, comparable features have been recorded—e.g., Campos (1994) from Brazil, and Woolfe (1994) from Antarctica. In Tanzania a number of such tectonically controlled paleovalleys have preserved much of the sedimentary record, the most impressive of these being the Ruhuhu Basin, with almost 900 m of sediments. This is the only region where such a thick succession was preserved in Tanzania, attributed probably to thermal subsidence during the early phase of rifting in the Early Permian of East Africa. The Idusi Formation appears to record the last glacial advance and the succeeding stages of glacial retreat and degla-

ciation. This climatic amelioration is evinced in the lithologic development of the Idusi Formation irrespective of local facies or thickness variations. The correlation between lithology and depositional environment is shown in Figure 5. The Lisimba Member, characterized by massive diamictite, records the last glacial advance and subsequent retreat. These are end-glacial or periglacial deposits that are overlain by a proximal glaciolacustrine succession intercalated with glaciofluvial and marginal deltaic deposits. The middle part of the Lisimba Member is characterized by mudstones with dropstones that record a distal facies of a proglacial lake environment. The final deglaciation is characterized by the deposition of fine-grained, organic-rich sediments of the Lilangu Member, which took place in a large, shallow lake that developed as an expansion of formerly proglacial to periglacial lake systems. Plant remains and sporadic desiccation cracks exhibit times of exposure when the lake was reduced in extent and depth. This large lake not only filled the paleorelief but overstepped onto higher areas where it was directly in contact with Precambrian

Formation of euxinic lakes, East Africa basement. During successive deglaciation and exposure of the lowland areas, which were increasingly vegetated, deposition switched to plant debris and silty-clayey sediments that culminated in the accumulation of organic carbon–rich black shale. These deposits took place in an anaerobic environment at the lake bottom, interchanging at marginal positions with deltaic lobes and subaqueous fans (Fig. 5). The top of the Lilangu Member records the filling of the lake and subsequent alluvial fan deposition. Wopfner (1996) interpreted the angular unconformity between the Idusi Formation and the overlying Mchuchuma Formation (Mpera Sandstone Member) as a basinwide tectonic event, which in the Galula and Songwe-Kiwira Basins led to erosion and removal of the sediments of the Lilangu Member. REGIONAL TECTONIC CONTROL AND APPROXIMATE DATING The deposition of black shales during the euxinic stages of lake development of the Ruhuhu Basin and other early Karoo basins was a significant event that can be traced regionally across East Africa. The abundance of organic matter led to anaerobic conditions in those regionally extensive deposits. Alternatively, anaerobic conditions may have led to extensive preservation of organic matter. Postglacial conditions were characterized by a climatic amelioration and represented a transition from cold and arid to temperate and humid conditions. Shortly thereafter, coalrich cyclothems prevailed, which recorded seasonal variations during deposition of the Mchuchuma Formation (Kreuser, 1991). Coal deposition commenced as an almost simultaneous event throughout entire Gondwanaland, which likely contributed to a lowering of atmospheric CO2 (Wopfner and Diekmann, 1996). The unconformity between the underlying Lilangu Member and the overlying Mpera Sandstone is evidence for a basinwide tectonic event that probably led to the complete erosion of the black shale in the Galula and Songwe-Kiwira Basins. The Mpera Sandstone is interpreted as braided stream deposits that initiated the second depositional sequence of the Karoo succession (Kreuser et al., 1990). This second unit, the Mchuchuma Formation (Fig. 5), was dominated by coal swamps and is also important for the source rock assessment of the Karoo succession in Tanzania (Kreuser et al., 1988). On a regional scale in Africa south of the Sahara, glaciogenic and postglacial deposits are known from South Africa, Zimbabwe, Zambia, Malawi, Kenya, Madagascar, Congo, and Mozambique. The basal diamictite is present in Zimbabwe (Mid-Zambezi and Sabi-Lunde Basins; Kreuser et al., 1990). In Zambia glacial deposits are known from the Gwembe, Luano, Luangwa, and Barotse Basins, respectively (Kreuser et al., 1990). In Malawi there is some evidence for the glacial character of the basal beds. In Kenya some indicators point to a glacial origin of the basal beds (Martin, 1981), and a similar record was noted for Mozambique (Kreuser, 1995b). There is no doubt that a glacial character exists in Madagascar in the Morondava Basin, whereas

109

an erosional hiatus separates Precambrian to Permian strata in the Majunga Basin (Besairie, 1972). For the Congo, basal diamictites are recorded and interpreted as being of glacial origin (Boutakoff, 1948). The black shale above the diamictite is not of such a ubiquitous distribution. In Zambia the Mukumba Member shows similarities to the Lilangu Member as being part of the Luwumba Formation in the Luangwa Valley, both separated by an unconformity (Kreuser et al., 1990). In Malawi the black shales are present in the Nkana and Livingstonia Basins, separated from the underlying tillites by an unconformity (Kreuser et al., 1990). Also in Madagascar, black shales follow the tillites in the Morondava Basin (Besairie, 1972). In the Congo the tillites are overlain by two separate black shale zones that locally interfinger with fluvial sandstones, both separated by local unconformities (Kreuser, 1995a). The Lilangu Member is partly developed in those basins which survived later tectonic exhumation. Locally it was not deposited at all, but deposition evolved from a glacial to a peat swamp environment, marked by either a rapid transition or a regional unconformity between the glacial and postglacial units. However, the viability of much of the data presented in the literature is not convincing, as much of the evidence has not been verified by detailed lithological descriptions; thus some of the stratigraphic nomenclature is not up to date with lithostratigraphic correlations. Additionally, dating by palynology or paleontology is poor for these continental deposits, and much more work is needed in order to establish a reliable litho-chronological framework. Apparently the Karoo basins farther east in the vicinity of the coastal area were never reached by the Paleozoic glaciation, or at least are not exposed at the surface (Kreuser et al., 1990). No basal diamictites and succeeding euxinic black shales are known from the coastal basins of Tanzania, Kenya, and Somalia (Mbede, 1997). Wopfner and Diekmann (1996) suggested a duration for the deposition of the Lilangu Member of ~5–6 m.y. In comparison with the approximate duration of the glaciation, which lasted ~20–25 m.y., deglaciation would have taken one quarter of the overall time. The revised estimate is based on palynological analyses that indicate a late Asselian to early Sakmarian age (Granulatisporites confluens Oppel zone; Wopfner and Kreuser, 1986) for the Lilangu Member of the Idusi Formation. ORGANIC GEOCHEMISTRY OF EUXINIC LACUSTRINE BLACK SHALES OF THE LILANGU MEMBER Several authors analyzed samples with high organic contents from Lower Permian lacustrine strata from East Africa in order to estimate their hydrocarbon potential (Kreuser et al., 1988; Dypvik et al., 1990; Kagya, 1991; Diekmann, 1993). Only a few samples were collected from the postglacial deposits, however; most of the samples came from coal seams in the Mchuchuma Formation or from Upper Permian deposits in rift lake environments (Mpanju,

110

Kreuser and Woldu

1999). For the sake of comparison, some of these results are mentioned here as well, but our focus is on the black shale of the Lilangu Member. The hydrocarbon potential of the Lilangu Member black shale was evaluated by Rock-Eval pyrolyses, and the hydrogen index (HI in mg HC/g TOC) was measured. We also analyzed the Tmax (°C) values during successive heating of samples, and this was matched with vitrinite reflection data. Temperature ranges (Rock-Eval pyrolyses) for most of the samples were from 425 °C to 460 °C, with the black shale from the Lilangu Member showing the highest temperatures. These samples exhibit a fair hydrocarbon generation potential, ranging from 65 to 150 mg HC/g TOC. Some TOC values reach 11%, but most are