Diss. ETH No. 11763

The dating of surfaces using in situ produced 10ae, 26Al and 36CI, with examples from Antarctica and the Swiss


presented by Susan Denise Ivy Ochs M.S. M.S. Civil Engineering Oregon State University born May 8, 1958 citizen of U.S.A.

accepted on the recommendation of Prof. Dr. Christian Schluchter, examiner Dr. Peter W. Kubik, co-examiner Dr. j(jrg Beer, co-examiner PD Dr. Rainer Wieler, co-examiner

Zurich 1996 To Michael

Cover illustration: 'Gletscherstein.' The dates 1660 and 1767 are carved into the surface of a boulder found at the foot of a moraine at Goschener Alp, . They are thought to indicate the dates of pilgrimages undertaken to halt the advancing Damma Glacier (modified from Holzhauser 1982, original reference Schaller-Donauer 1925). Acknowledgments

This project was funded by ETH research grant 0-20-624-92 to Christian Schluchter. The fieldwork was supported by Swi:;s National Science Foundation grant 21-28971.90 to Christian Schluchter and US National Science Foundation Polar Programs 9020975.

The following people provided rne with much needed support and encouragement for which I can never thank them enough:

Christian SchlOchter: for his boundless and contagious enthusiasm for all things Quaternary, being there when you needed him but otherwise giving me a long line, for the most careful sampling-really the most important step, cigars during sampling, and pleasant moments in a 'Beiz-thing' after sampling; Peter Kubik: for explaining many things about physics to me (luckily only what I really needed to know), those long inte·grations, and always putting me in the first cassette of the run; Jurg Beer: for letting me use the modern lab space at EAWAG, and making me feel like a member of UP; Rainer Wieler: for being a last minute coreferent, and careful reviews; Michael Ochs: for making me comfortable with the whole idea of wet chemistry, figuring out what the ICP-AES was actually doing, making it stop, and then getting consistent and reproducible numbers for Al, chiseling out many samples, carrying many kilos of rock, and taking beautiful pictures of the sampling sites; Robert Ivy: for smashing many rocks during his 'vacations' to Switzerland, making most of the maps for the Geology paper, and showing me how to make maps with a computer; lrka Hajdas (and Wojtek): for an infinite number of fruitful discussions, for cutting me slack in the 14C lab; Laura Bruno: for her persistence especially when we were writing the first proposal; 36 Hans Arno Synal: for CI measurements and enthusiastic support of investigating 36CI in rock surfaces; Martin Suter: for giving all of the members of the AMS group relatively free rein in their research interests; Peter Signer: for his relentless support of the geology-AMS-noble gas collaboration; Willi Wolfli: for inception of a varied and robust AMS research group; 14 14 Georges Bonani: for being a nice C boss, for measuring a few C when I needed them, and for keeping THE MACHINE happy; Alan Thompson: for supporting us in the ETH grant evaluation process; 14 Wally Broecker: for hiring me to do C and always encouraging the next step; 14 Sue Trumbore: for getting me the job in the C lab and many interesting discussions; Ed Brook: for answering any and all questions through email, and conscientious reviews; Hanns Kerschner: for his enthusiasm for glacial geomorphology, fixing up the introduction to the Julier Pass paper, and the tip about the boulders at the toe of Koefel's; George Denton and David Sugden: for the excitement they incite for glaciologic and geomorphologic problems, reading the first (very) rough version of the Sirius paper on the train to Jungfraujoch and patiently explaining the real meaning of my Antarctic exposure dates; George Denton: for collecting great samples from New Zealand, and of course again his enthusiasm; David Sugden and Mike Bentley: for fun days sampling in Iceland; Jeff Klein: for explaining the banana just one more and encouraging me to write the Geology paper; Fred Phillips: for his helpfulnes~: in all matters 36CI and/or exposure dating; Marek Zreda: for showing me how to make 36CI samples; Markus Knauer: for showing me his extraction of 10Be and 26AI from meteorites.

All the members of the Teilchenphysik group, every single person who did a shift to measure these samples (especially a night shift) deserves a million thank yous: Wolfgang Gruber: for taking care of many little things and right when I needed it done (ordering stuff--even worse returning stuff), making neat stands for the columns, covers for the samples etc. etc.; Peter Kagi: for babysitting THE MACHINE and keeping it happy; Nora Steiner: for pressing some of my targets and fun in the 14C lab.

The whole Umweltphysik group at EAWAG for making me feel like I belong to them: 36 Silvia Boll ha Ider: for helping me through some of the CI sample preparation and many discussions in DS7; Alfred Luck: for some of the data reduction and his cheerfulness; Gerhard Wagner: for interesting discussions and for some of the data reduction; Yvo Weidmann: for the most beautiful SEM pictures and ordering lots of stuff; Caroline Stengel: for letting me into her labs without a fight; and many other friends at EAWAG who got me through lunches and other stuff: especially Meg (and David), Yael (and Tony), Therese, Bouziane and Raoul (for the Ottawa font). David Kistler: for keeping the ICP-AES running and letting me use it at any time. Thanks also go to the whole lngenieurgeologie group at ETH, especially: Conrad Schindler: for accepting me into the group; Rita Gysin: for help in innumerable matters (Fremdenpolizei etc. etc.).

The whole Happy Ice Age group (now of Uni Bern) provided (and still do) a big friendly Quaternary family: especially Beni MOiier (and Catherine): for shedding light on the Swiss Quaternary paradigm; Sylvio Tshudi and Joerg Schafer: for some fresh ideas about exposure dating; Michael Helfer (and Otti Rohner) for sampling in Antarctica and insights into Antarctic geology. Eve SchlOchter deserves a big hug for not minding when Christian literally ran-off to sample whenever and wherever, and for yummy Happy Ice Age fests.

And a few, more personal thanks: lrka: for always being my best friend, no matter what; My family: especially Mama and Baba, Robert and the mom for unflagging encouragement; Michael: he is clearly the shadow coauthor of this dissertation. From the sampling through the chemcial extraction on to the final writing of each paper, his input can be felt. He constantly patiently explained in great detail what was actually happening during the Be, Al, and Cl chemical extraction procedures. Without this insight, this whole thing might not have been possible. But all this is not meant to diminish the value of the affection, emotional support and belief in me which he gave endlessly. Abstract

A primary concern today in ice age research is to elucidate the worldwide timing of glacial fluctuations. This information is needed to define the mechanisms by which climatic signals are initiated, and then transferred throughout the various global systems (i.e. oceans, atmosphere, and terrestrial environments). Because of the importance of integrating the terrestrial records of the former extent of both ice sheets and alpine glaciers, the direct dating of rock surfaces with in situ produced cosmogenic isotopes is 10 26 36 an essential tool. These isotopes (e.g. Be, AI, and CI) are produced in the surfaces of rocks due to interactions with cosmic rays. In order to be able to reliably determine exposure ages from rock surfaces with both long and short exposure , ocr goal at the onset of this project was to set-up the extraction procedure for Be, Al, Cl from rock samples. 73 separate rock or mineral dissolutions, involving 52 different rock samples, were performed to 1) make sure that 0 meteoric ' Be was being removed, 2) verify the reproducibility of the procedure, and 3) eventually, address several questions on the timing of glacier fluctuations in two specific geographic areas: Antarctica and the . Exposure dates we have determined from erratic boulders in the Sirius Group at Mount proved conclusively that this outcrop of the Sirius Group is more than 5.8 million years old. Minimum ages for Sirius Group deposits at Table Mountain and Mount Feather are 2.9 Ma and 2.3 Ma, respectively. The respective concentrations of cosmogenic isotopes are controlled and the ages are true minima. At Table Mountain, the plateau surface carved into the Beacon Sandstone 10 bedrock yielded minimum Be exposure ages of 5.5 Ma. No erosion (in situ weathering) rate higher than 70 cm/Ma could be calculated from our data. Indeed many surfaces have been affected by weathering rates of less than 5 cm/Ma (based on the accepted 10 26 10 production rates for Be and Al). The very high Be concentrations we have measured at all three sites completely rule out significant tectonic uplift of these surfaces during the last 3 Ma. Our resu Its have provided support for the idea that the East Antarctic Ice Sheet, a crucial variable in projections of greenhouse scenarios, is a stable feature of Antarctica. They also show that much of the landscape in Antarctica has not been significantly modified geomorphologically sino~ at least the early Pliocene and is indeed a relict landscape. We have investigated three key sites in Switzerland related to the timing of glaciations of the Alps. At the first site, we sampled boulders located along the crest of an 10 26 Egesen moraine at Julier Pass. The overall age we obtained, determined with Be, AI 36 and CI, for the double-walled moraine (11, 100 years) confirms that it was formed during the Younger Dryas cold snap. At Grimsel Pass, we determined exposure ages from ii polished surfaces that were formed during the Last Glaciation. Our 10Be and '"Al exposure ages indicate that the pass area was ice-free by no later than 12,000 years ago. Erratic blocks mark the farthest extent of the Solothurn arm of the Rhone Glacier in the 10 26 region of Wangen a. d. Aare. The exposure ages ( Be and Al) we measured from some of these blocks are the first direct dating of the farthest extent of the Rhone Glacier. According to our dates, the glacier withdrew from its maximum position at or slightly before 20,000 years ago. This provides evidence that the Solothurn lobe of the Rhone Glacier reached its maximum extent late in the last glacial cycle, roughly contemporaneous with the maximum extent of the Rhine Glacier and the worldwide Last Glacial Maximum. iii


Ein primares Ziel der heutigen Eiszeitforschung ist die Aufklarung der zeitlichen Abfolge weltweiter Gletscherfluktuationen. Diese Information ist notig um die Mechanismen zu definieren, durch welche klimatische Signale zunachst initiiert und dann auf vielfaltige globale Systeme (Ozeane, Atmosphare, terrestrische Systeme) Ubertragen werden. Da es sehr wichtig ist die terrestrischen Aufzeichnungen der frUheren Ausdehnung von sowohl kontinentalen als auch alpinen Gletschern zu integrieren, ist direkte Datieren von Felsoberflachen mit Hilfe in situ-produzierter kosmogener Isotope 0 26 36 ein unersetzliches Werkzeug. Diese Isotope (z. B. ' Be, AI und (1) werden in den Oberflachen von Gesteinen aufgrund von Wechselwirkungen mit der kosmischen Strahlung erzeugt. Unser Ziel am Ausgangspunkt dieses Projekts war es die Extraktion von Be, Al und Claus Gesteinsproben zu etablieren, um Expositionsalter fUr Gesteinsoberflachen sowohl mit langen als auch kurzen Expositionszeiten zuverlassig bestimmen zu konnen. Es wurden 73 Probenaufbereitungen, basierend auf 52 verschiedenen Gesteinsproben, durchgefUhrt, um 1) sicherzustellen, dass meteorisches ' 0 Be entfernt wird, 2) die Reproduzierbarkeit des Aufbereitungsverfahrens zu verifizieren und 3) schliesslich mehrere Fragen bzgl. der zeitlichen Abfolge von Gletscherfluktuationen in zwei spezifischen geographischen Gebieten anzugehen: der Antarktis und den Schweizer Al pen. Expositionsalter, die wir ain erratischen Blacken in Sedimenten der Sirius Group am Mount Fleming bestimmt hat1en, beweisen schlUssig, dass dieser Aufschluss der Sirius Group mehr als 5.8 Millionen Jahre alt ist. Minimalalter fUr Ablagerungen der Sirius Group an Table Mountain und Mount Feather betragen 2.9 und 2.3 Millionen Jahre. Die jeweiligen Konzentrationen kosrnogener Isotope werden durch Erosion kontrolliert und die Alter sind deshalb echte Minima. Die Plateauoberflache, die am Table Mountain in den anstehenden Beacon Sandstone geschnitten ist, liefer! minimale ' 0 Be-Expositionsalter von 5.5 Millionen Jahren. Aus unseren Dalen konnten keine Erosionraten (in situ- Verwitterung) grosser als 70 cm pro Millionen Jahre berechnet werden. Tatsachlich wurden viele Oberflachen durch Verwitterungsraten kleiner als 5 cm pro Millionen Jahre 0 beeinflusst (basierend auf akzeptierten Produktionsraten fUr ' Be und 26AI). Die sehr 0 hohen ' Be Konzentrationen, die wir an alien drei Lokalitaten bestimmten, verneinen die Moglichkeit einer signifikanten tektonischen Aufhebung dieser Oberflachen wahrend der letzten 3 Millionen Jahre. Unsere Resultate unterstutzen das Konzept, dass das Ostantarktische lnlandeis, eine entscheidende Variable in Vorhersagen von Treibhausszenarien, ein stabiler Bestandteil der Antarktis ist. Sie zeigen ferner, dass ein Grossteil der antarktischen Landschaft seit mindestens dem Jungpleistozan keine iv signifikanten geomorphologischen Veri!nderungen erfuhr und in der Tat eine relikte Landschaft darstellt. Wir haben drei Schlussellokalitaten in der Schweiz untersucht, die einen Bezug zur Abfolge der Vergletscherungen in den Alpen herstellen. Ander ersten Lokaliti!t beprobten wir Blocke entlang des Kamms einer Egesenmorane am Julierpass. Das 10 26 36 durchschnittliche Alter (11 '100 Jahre), dass wir mit Be, AI und CI fur die Doppelwallmorane bestimmten, bestatigt, dass sie wahrend des Kalteeinbruchs der Jungeren Dryas gebildet wurde. Am Grimselpass bestimmten wir Expositionsalter von Gletscherschliffen, die wahrend der letzten Vergletscherung entstanden. Unsere 10Be- 26 und Al-Exposition salter zeigen, dass das Passgebiet nicht spater als vor 12'000 Jahren eisfrei wurde. Erratische Blocke markieren die Maximalausdehnung des Solothurner Arms 10 des Rhonegletschers im Gebiet von Wangen a. d. Aare. Die Expositionsalter ( Be und 26 Al), die wir an einigen dieser Blocke ermittelten, stellen die erste direkte Datierung der Maximalausdehnung des Rhonegletschers dar. Gemass unserer Oaten zog sich der Gletscher von seiner Maximalposition vor 20'000 Jahren, oder wenig fruher, zuruck. Dies ist ein lndiz dafur, dass der Solothurner Arms des Rhonegletschers seine Maximalausdehnung spat irn letzten Glazialzyklus erreichte, ungefi!hr zeitgleich mit der Maximalausdehnung des Rheingletschers und dem weltweiten glazialen Maximum. ACKNOWLEDGMENTS ABSTRACT ZUSAMMENFASSUNC iii


8 EXPOSURE DATING 7 B 1 THEORETICAL ASPECTS 7 61.1 INTRODUCTION 7 61 .2 DEVELOPMENT OF THE FIELD OF EXPOSURE DATING 7 61 .3 PRODUCTION OF COSMOGENIC ISOTOPES 9 Bl .3.1 complicating factors - known 12 B 1.3.1. i' Sample thickness 12 B Shielding of the rock surface 15 61.3.2 complicating factors - unknown 15 B Magnetic field effects 16 B Unknown factors specific to the site or rock surface 16

B2 EXPERIMENTAL ASPECTS 23 62.1 INTRODUCTION 23 62.2 SAMPLING 23 62.3 SAMPLE PREPARATION 24 10 62.3.1 6e and 26AI extraction from 24 82.3.1. 1 Selective chemical dissolution - the making of a pure quartz separate 25' Dissolution of pure quartz and chemical separation of Be and Al 28 ICP-AES measurement of total Al 35 62.3.2 Extraction of Cl from whole rock 37 62.4 AMS MEASUREMENT 38

10 B3 USING FOf BE EXPOSURE DATING STUDIES: A NOTE OF CAUTION 41 63.1 ABSTRACT 42 63.2 INTRODUCTION 42 63.3 SAMPLES AND ANALYTICAL PROCEDURES 44 B3.3.1 Sample information 44 10 63.3.2 Be sam~le preparation and measurement 44 3 1 63.3.3 He and Ne measurement in pyroxene 45 B3.4 PRODUCTION RATES 46 63.5 RESULTS AND DISCUSSION 46 63.5.1 Quartz 48 63.S.2 Pyroxene 50 63.6 CONCLUSIONS 52 C ANTARCTICA 55 C1 OVERVIEW OF THE CHRONOLOGY OF ICE SHEET VARIATIONS 55 C1.1 INTRODUCTION 55 C1 .2 OF THE LANDSCAPE AND CLIMATE OF ANTARCTICA 55 C1.2.1 The ocean record 55 C1 .2.2 The terrestrial record 60 C 1 .3 A DIFFERENT POINT OF VIEW: THE DYNAMIC ICE SHEET MODEL 61 C1 .4 GEOLOGIC SETTING 62 Cl .5 SIRIUS GROUP 63 Cl .5. J Sirius Group outcrop at Table Mountain 64 C1 .5.2 Sirius Group outcrop at Mount Feather 64 C1 .5.3 Sirius Group outcrop at Mount Fleming 65 Cl .6 SUMMARY OF PART C 65


26 C3 lOBE AND AL EXPOSURE AGES FOR THE PLATEAU SURFACE AT TABLE MOUNTAIN AND THE SIRIUS GROUP AT MOUNT FLEMING, 3 MOUNT FEATHER AND TABLE MOUNTAIN 76 C3.1 ABSTRACT 77 C3.2 INTRODUCTION 77 C3.3 METHODOLOGICAL DETAILS 78 C3.4 RESULTS AND DISCUSSION 81 C3.4.1 Beacon Sandstone Plateau Surface at Table Mountain 81 C3.4.2 Sirius Group at Table Mountain 82 C3.4.3 Sirius Group at Mount Fleming and Mount Feather 85 C3.4.4 Uplift of the Transantarctic Mountains 85 C3.5 SUMMARY AND CONCLUSIONS 87









1 submitted to Earth and L~tters. 2 Geology, 1995, v. 23, p. 1007-1010. 3 Terra Antarctica, 1996 (in press). 4 Eclogae Geologicae Helvetiae, 1996, v. 89, p. 1049-1063. A INTRODUCTION ANI[) OVERVIEW

A 1 INTRODUCTION Detailed changes in amplitude and frequency of the ice ages have been estimated 18 by tracking 0 0 changes in both deep sea cores (Shackleton and Opdyke 1973, Hays et al. 1976, lmbrie et al. 1984, Martinson et al. 1987) and in the Greenland Oohnsen et al. 1992, Alley et al. 1993) and Antarctic ice cores (Jouzel et al. 1987, 1993, 1995, Lori us et al. 1985). In contrast to these two records is the more patchy terrestrial sedimentary record. However, the importance of the terrestrial record cannot be ignored; it contains the only direct evidence of actual extents and movements of glaciers (i.e. till, SchlOchter 1992). The crucial task is to correlate the terrestrial record with the deep sea sediment and records. This is the only way to really see if cold events around the globe have been synchronous or not. For such a task, the dating of terrestrial deposits is of utmost importance. Therefore, the goals of this work are simple and straightforward. The use of rock surface dating with in situ produced cosmogenic isotopes (exposure dating) to date features that were never datable before. Such features which are produced by a glacier directly include a glacially polished surface, a boulder on the crest of a moraine, an erratic boulder. The direct evidence of glacier movements on the continents can now be fitted into the emerging picture of global paleoclimate. The timing of glacial fluctuations can be elucidated in detail and world-wide, regional, or local synchroneity or nonsynchroneity can be revealed. The concept of exposure dating is, in principle, quite simple. Cosmic rays impinging upon an exposed rock surface induce nuclear reactions within the mineral lattices yielding the cosmogenic i~;otopes. The concentration of the cosmogenic isotope, which is measured by mass spectrometric methods (accelerator mass spectrometry for long-lived radioisotopes and high resolution mass spectrometry for stable isotopes), is a direct measure of the length of time the geologic surface has been exposed. By combining 10 6 26 the radionuclides with different half-lives Be .5 x 10 yr) and AI 6 x 105 36 5 21 yr) and CI (t 112=3.0l x 10 yr) and the stable cosmogenic isotopes (3He and Ne), complicated exposure , in a variety of rock types and with exposure ages ranging from ca. 5000 to ca. 5,000,000 years can be investigated.

A2 OVERVIEW OF CHRONOl.OGY OF THE ICE AGES 18 The 0 0 variations seen in ocean cores, as recorded in foraminifera tests, reflect both the temperature and the isotopic makeup of the water at the time. The latter reflects world-wide ice volume, as ice locks up more of the lighter oxygen isotope. Estimations about ice volume can be made after one makes certain assumptions about the 2 temperature effect. Oxygen isotope ratios recorded in ice provide a direct gauge of air 18 temperature. In both cases, a shift towards heavier 8 0 (more negative values) indicates either directly cooler temperatures (ice record) or inferred cooling based on increased ice volume (ocean record). In ice cores, the timescale is constrained by annual layer counting. Annual layers 18 are defined by dust and/or 8 0 changes (among other parameters) back to ca. 14,000 BP. Before that date, layer counting is combined with ice sheet models, which account for the flow of ice at greater depth (e.g. Johnsen et al. 1992, Alley et al. 1993). Using radiocarbon locked in foraminifera, the deep sea record has been dated back to 40,000 14 C years BP. Prior to that time, paleomagnetic reversals, dated by K/Ar in volcanic rocks provide anchor points. For the most part the time scale has been constructed by tuning 18 the 3 0 record to orbital variations as described by Milankovitch (e.g. lmbrie et al. 1984, Martinson et al. 1987). The deep sea record goes back much further than either the Greenland or the Antarctic ice core records, i.e. millions of years as compared to hundreds of thousands of years. The most complete long terrestrial records are the loss deposits of China, which reach back to the Pliocene/Pleistocene transition. Matching of paleomagnetic reversals of the loss record (Heller and Liu 1982, Kukla et al. 1988, Kukla and An 1989) has allowed direct correlation of the highs of magnetic susceptibility (associated with the paleosols) with the deep sea record (Kukla et al. 1988). The paleosols within the loss deposits are coeval with the warmer intervals (interglacials and interstadials) recorded in oceanic 3180, and appear to reflect changes in the paleomonsoon track over China (Banerjee 1995, Bloemendahl et al. 1995). Other terrestrial records are less continuous than the loss deposits. Such records include glacigenic sediments (tills) and glacially-related sediments (glaciolacustrine and fluvioglacial). Past climatic events are recorded not only in the sediment which has been preserved, but in the time intervals where sediment is missing as well (not deposited or later eroded off). Deposits related to a previous cold event may have been stripped by later glacier movements. Additionally, the evidence left by a glacier is not only depositional but is erosional as well, e.g. glacially polished surfaces. The only way to fit the terrestrial evidence into the time frame of climate change built with ice and deep sea cores is by relative means (correlation) and/or by . For the Holocene and the Last Glacial Maximum, the application of on organic material has provided a somewhat complete picture of the timing of events -where material is available. For mountainous areas, or especially dry areas where vegetation is sparse, the radiocarbon method cannot be used, due to a lack of plant 14 remains. Because of its high precision when appropriate material can be found C 3

should be the first method to choose. But often, organic material related to the event of interest cannot be found. The presence of an ash layer in several sedimentary records provides a marker 14 40 9 which can be dated (by C of entrained wood, or by K/Ar or Ar/3 Artechniques) and provides a useful tool for correlation. In central Europe, the Laacher See Tephra (LST) has provided just such an invaluable marker bed for the Allerad (van den Bogaard and Schminke 1985, Hajdas et al. 1995). Unfortunately, there is no evidence found as yet of ash layers older than the LST deposited in Switzerland. The method of exposure dating can fill these methodological time gaps and/or problems due to a lack of material. Exposure dating can be used beyond the time range of 14 radiocarbon, i.e. older than 40,000 C yr BP. Exposure dating can also be used for the 40 9 time scales younger than 100,000 years where K/Ar and Ar/3 Ar have yet to prove their accuracy. Neither the organic material entrained in a moraine nor an ash layer mantling a moraine are the direct product of a glacier's advance. Both of the aforementioned dating methods provide only a relative date. In both methods the age of the material is determined, while exposure dating dates a process, i.e. the uncovering of the surface due to glacial retreat. The method of exposure dating offers the unique advantage that the record of the event itself can be directly dated, e.g. a boulder on a moraine. Indeed, a glacier can transport a previously exposed boulder. But the attainment of the boulders final position with its top surface facing upward is reached only when the glacier retreated from that position and the moraine stabilized. In that sense exposure dating dates the event directly. Additionally, exposure dating is the only means by which the age of erosional surfaces (glacially polished surfaces) can be directly determined, so that finally, such information can be correlated and flt into the overall climate picture. A variety of geological events can be dated with exposure dating (Chapter Bl .2). Here we are primarily concerned with features of glacial origin. Our goal is to address the question of the world-wide synchroneity of cold events by obtaining more accurate dates 18 for glacial advances and retreats. The shape of the S 0 curves from various records indicate apparently synchronous climatic behavior (e.g. Bond et al. 1993), but the terrestrial sedimentary records have yet to be fit into the puzzle. Interpretations about synchroneity or the lack thereof of cold periods in different regions cannot be made until a precise chronology is established in every case. For example, in the last glacial cycle, the maximum of global ice volume occurred 14 18 at ca. 20,000-18,000 C yr BP, based on both 3 0 from ocean cores and sea level interpretations ( and Shackleton 1986, Shackleton 1987). On the other hand, the 18 temperature record reflected by ice core 3 0 does not show a marked temperature low at 20,000 years ago. Conversely, the lowest temperature was reached several times periodically during the whole of the last glacial cycle (since Se) (e.g. Sowers et al. 4

1993). Direct evidence of the farthest extent of the Fennoscandian ice sheet indicates this 14 occurred at ca. 28,000 C yr BP (e.g. Fronval et al. 1995). For the Laurentide Ice Sheet, the furthest extent on Baffin Island occurred at roughly 100,00-80,000 years ago ( et al. 1993 and references therein). While along the southern margin of the ice sheet, the 14 furthest extent was reached at ca. 20,000 C yr BP, with the maximum in the mid-west being reached slightly earlier than in either the east or west (e.g. Fullerton and Richmond 1986, Richmond and Fullerton 1986). Layers containing high concentrations of ice-rafted debris recorded in sediment cores taken from the North Atlantic which are known as Heinrich events (Bond et al. 1992, 1993, Bond and Lotti 1995) seem to have been synchronous with glacial advances in Chile (Lowell et al. 1995) and perhaps as well with several marked glacial advances in western North America (Clark and Bartlein 1995). The Younger Dryas abrupt climatic deterioration appears to have been felt almost instantaneously world-wide. Evidence from both lake and ocean sediments has come from all regions of the globe (Peteet et al. 1993, Peteet 1995, Troelstra et al. 1995, and references therein). For example, outside of the traditional European Younger Dryas realm, well-dated moraines associated with this event are found in western North America (Gosse et al. 1995a) and in the Southern Alps of New Zealand (Denton and Hendy 1992). If we hope to answer the question of regional or world-wide synchroneity of glacier expansions, the question whether glaciers should behave synchronously must be addressed. Glaciers respond to a decrease in temperature and/or increase in accumulation with a change in mass balance (accumulation minus ablation), which may eventually be reflected in a length change. Because there is a broad spectrum of factors which affect how and how fast a given glacier responds to a change in input, in detail, glacier responses may not appear synchronous (Gillespie and Molnar 1995). Response time is a function of glacier size; cirque and glaciers respond after tens to hundreds of years, while ice caps may take many thousands of years (e.g. Sugden and John 1976). Additional factors which play an important role in whether a given climate signal will be amplified or dampened by a glacier include 1) the slope of the glacier and how much of its volume is near the equilibrium line position, 2) , topography, and gradient of the bed, and 3) the thermal characteristics of the ice itself (e.g. Meier 1965, Sugden and John 1976, Greuel I 1989, Bull 1991 ). The glaciers of the Alps may be especially responsive to climate change because of their mid-latitude location. The mid-latitudes are sensitive to changes in insolation and storm track position. Regional to local insolation and precipitation variations can act to either amplify or dampen a glacier's response to a change in temperature. When the various response times and modes, and site 5 specific/local factors for each ice body are considered, a degree of asynchroneity in detail should not be surprising.

A3 A BRIEF HISTORY OF EXf'OSURE DATING IN SWITZERLAND Even though at the present time (summer 1996) the method of exposure dating is seen more and more as an invaluable tool for dating Quaternary (as well as older) deposits, when we initiated this project in 1991, the method was anything but accepted by earth scientists. At that time, our main concerns were not which moraine or which erratic shall we date, but whether we can actually date anything, whether the resulting date is meaningful within the geological and geomorphologic constraints, whether we can only date rocks with high concentrations of cosmogenic isotopes. The most basic concern was whether we can get the same age (concentration) from the same rock more than once. With the completion of these first two dissertations (Bruno 1995 and this work), one can readily see how far we have really come. Indeed, the field of exposure dating has matured as well. This is due in no small part to the integration of experienced (local) / geomorphologists from the beginning within a given dating project (for example see Gosse et al. 1995a, I 995b). A coincidental sequence of events led to the initiation of the method of exposure dating in Switzerland. On a one-day excursion to look at the glacial landscapes around the ZUrich area, Wally Broecker introduced me to George Denton and Christian SchlUchter. From there, Christian and I discussed the use of AMS (I was at the time doing AMS radiocarbon dating for Wally Broecker) to date rock surfaces. A vital turning point 36 was the addition of Peter Kubik, a coauthor on the landmark CI exposure dating papers of Fred Phillips and Marek Zreda, to the AMS group. JOrg Beer (EAWAG), another enthusiastic exposure dating supporter, provided the sorely lacking robust laboratory space. At precisely the same moment, Professor Peter Signer, Rainer Wieler, and Laura Bruno were contemplating the use of their talented noble gas mass spectrometers (Tom Dooley for one) to measure non-extraterrestrial samples. The group effort involved in the writing of the ETH proposal proved to be worthwhile. The final touch was added when Professor Alan Thompson became our representative in the ETH proposal evaluation process.

A4 ORGANIZATION OF THIS DISSERTATION Th is dissertation is made up of several publications, supported by all the theoretical and experimental details relating to the measurement and interpretation of exposure ages. In addition, several chapters present data on exposure dating in Switzerland which have not yet been published. This dissertation is divided into five parts. The first part (A) is the 6 introduction and synopsis. It includes a brief introduction which serves primarily as an overview of the present state of ice age chronology and shows the distinct need for a way to date glacial landscapes especially where other methods cannot be applied. Equations are numbered sequentially throughout the whole dissertation and to avoid duplication, references for each chapter are contained in the single reference section at the end of this dissertation. Part B contains details on the method of exposure dating itself, both theoretical (Bl) and methodological (B2). These details are important for understanding both the range of possible applications as well as the limitations of this method. The second chapter of part B provides detailed descriptions of the chemical preparation steps we have used to extract Be, Al and Cl from rocks and to make a target measurable with AMS. Chapter B3 we present two sequential dissolution experiments performed to evaluate the usefulness 10 of pyroxene for Be exposure dating. In part C, the data and interpretation of exposure ages determined from rocks from the Dry Valleys region of Antarctica are presented. An introduction chapter (Cl) provides an overview and the geologic setting in light of the question of a stable or unstable East Antarctic Ice Sheet. The data are presented in the form of two chapters the first of which was published in the November 1995 issue of Geology (C2) and the second (C3) which is in press (1996) in the conference volume (Terra Antarctica) on Antarctic Earth Sciences VII meeting in Siena (1995). A brief synopsis of the landscape development of the Dry Valleys which includes our exposure dating results is also given with the conclusions (C4). Part D, is made up of chapters addressing individual temporal problems of the glaciations of the Swiss Alps, preceded by an overview of ice age chronology (Dl ). Chapter D2 presents a paper published in Eclogae Geologicae Helvetiae (1996). This paper concerns the exposure ages of boulders from an Egesen moraine at Ju lier Pass, Graubunden. In Chapter D3, exposure ages of glacially polished surfaces at Grimsel Pass (Bern/Wallis) which provides more information about the deglaciation of the high mountain passes in the Swiss Alps are presented. The results of exposure dating the tops of huge erratic boulders which demarcate the maximum extent of the Solothurn lobe of the Rhone Glacier are presented in Chapter D4. Chapter D5 is an attempt to fit the exposure dating results into the general picture of the evolving landscape of Switzerland at the end of the Last Glaciation and to address once more the question of synchroneity of glacier movements in the northern Swiss foreland. In part E, we use the data we have generated from Antarctica and from Julier Pass to 26 calculate production rates for rnBe and Al, and compare our calculated production rates to published ones. Part F contains a listing of the conclusions as well as an overview of the work that still needs to be done. 7



81.1 INTRODUCTION This chapter deals with the theoretical aspects of the method of rock surface exposure dating, we begin with a brief overview of the history of this field. It shows the breadth of the temporal geological problems that have been attacked thus far using this 10 method. Emphasis is placed on Be, 26AI, and 36CI as those are the isotopes studied here. The section which follows contains a summary of the production mechanisms of the three isotopes of interest. Such a discussion necessarily includes coverage of factors that can affect production. These factors may be well known, such as the thickness of the sample processed, or may be unknown, such as the amount of snow that covered a surface during exposure.

81 .2 DEVELOPMENT OF THE FIELD OF EXPOSURE DATING As described in more detail below, exposure dating is based on the concept that in a rock surface that has been either newly created (e.g. by scouring at the bottom of a glacier) or is suddenly uncovered (when sediment or soil cover is eroded off), costnogenic isotopes are produced within the mineral lattices as a result of exposure to cosmic rays (Table B 1.1 ). If prior to exposure the production of cosmogenic isotopes was negligible then, following uncovering, the production rate has been increased so that in a few thousand years the cosmogenic isotopes will be present in measurable quantities.

Ta bl e B 1.1. Ch aracteristics o f t h e cosmogenic isotooes 10Be, 26A I an d 36C. I

Isotope Production Rate T112 Limit of Sample Example of (at sea level, Applicability Material Sample Size high latitude)

toBe 6 atoms/I! SiO')·vr• 1.5 Ma to -5 Ma Quartz 10-50 g 26AI 37 atoms/I! SiO.,·vr• 716 ka to-2.5 Ma Ouartz 10-50 I! 3•c1 Rock Chemistry 301 ka to -1 Ma Whole 100 g dependent: e.g. Rock Granite 9 atoms/g·yr

Basalt 15 atoms/g·vr # *Nishiizumi et al. 1989 #Zreda et al. 1991, Phillips et al. 1996a 8

Recent summaries in the field of exposure dating include Cerling and Craig (1994b) 10 and Kurz and Brook (1994), applications and details pertaining specifically to Be and 36 26AI are given in Nishiizumi et al. (1993), and details with respect to CI are given in Dorn and Phillips (1991 ), Zreda (1994) and Zreda and Phillips (1994). Davis and 36 Schaeffer (1955) completed the first exposure dating of rock surfaces by measuring CI in a high elevation (3300m), high-Cl phonolite. They used traditional counting methods. 3 10 21 26 36 Although the suite of isotopes, He, Be, Ne, AI, and CI had been measured in lunar samples and meteorites (e.g. Nishiizumi et al. 1991 ), it was not until the development of highly sensitive mass spectrometers in the early 1980's, that the measurement of the small concentrations found in terrestrial materials became possible. 3 In 1986, successful exposure age determinations using He (Craig and Poreda 1986, Kurz 21 1986), 10Be and 26 Al (Klein et al. 1986, Nishiizumi et al. 1986), Ne (Marti and Craig 36 1987) and CI (Phillips et al. 1986) were reported. The dating of rock surfaces using cosmogenic isotopes has been applied to a broad spectrum of geological events. Of particular relevance here is the application of this method to the dating of features of glacial origin, e.g. glacial polish (Nishiizumi et al. 1989, Cerling and Craig 1994b). The precise dating of a glacier's maximum stand has been accomplished by exposure dating boulders at the crest of moraines (Phillips et al. 1990, Zreda 1994, Gosse et al. 1995a, 1995b, Zreda et al. 1995, Phillips et al. 1996b). In Antarctica, where very low erosion rates make exposure dating especially fruitful (Nishiizumi et al. 1991 ), the timing of events related to a sequence of moraines in Arena Valley was elucidated (Brown et al. 1991, Brook et al. 1993, 1995b) and the spatial and temporal relationships of deposits of the Ross Sea drift were examined (Brook et al. 1995a). As a consequence of the difference in half-lives, erosion rates in Antarctica have 10 26 been examined typically with the pair Be and AI (Nishiizumi et al. 1991, Nishiizumi et al. 1993, Brook et al. 1995b). Deciphering the timing of eruption of a flow has also been shown to be 21 possible using the cosmogenic isotopes 3He and Ne (e.g. Anthony and Poths 1992, 36 36 10 Cerling and Craig 1994a), CI (Zreda et al. 1993) and CI and Be (Shepard et al. 1995). The approximate date of caldera collapse at Reunion has been determined using 21 Ne (Staudacher and Allegre 1993). Catastrophic flooding events, which were common on many of the continents towards the end of the last glaciation, can be dated (such as the Big Lost River Flood in Idaho, Cerling et al. 1994). 10 Finally, because the maximum amount of Be that can be produced at a given elevation is known, constraints can be placed on the maximum possible amount of uplift in a given period of time, for e>

81 .3 PRODUCTION Of COSMOGENIC ISOTOPES 10 21 26 The cosmogenic isotopes 3He, Be, Ne, AI, and 36CI are produced in the upper surface of a rock by nuclear reactions induced by cosmic rays. The primary cosmic ray spectrum consists of galactic cosmic rays (energies 1-100 GeV), which originate outside our solar system, and of solar cosmic rays (energies 10 -100 MeV) (e.g. Lal and Peters 1967, Reedy 1987). The latter are unimportant for the production of cosmogenic isotopes in rocks. The galactic cosmic ray spectrum is made up of ca. 85% protons, ca. 14% Cl- particles, and ca. 1% heavier particles (Lal and Peters 1967, Lal 1988). As cosmic rays enter the earth's atmosphere, a cascade of secondary particles is produced, including protons, neutrons and muons of which neutrons are the dominating component These particles produce cosmogenic isotopes in a rock surface by spallation reactions, muon- induced reactions and neutron capture reactions. Spallation is a nuclear reaction where several lighter particles are emitted from a target nucleus hit by cosmic ray particles. For example, within a quartz grain located at the surface of an exposed rock, an oxygen atom hit by a cosmic ray particle will spall losing several nucleons. One of the atoms that can be thus produced within the.quartz crystal lattice is a 10Be atom.

10 26 Ta ble B 1.2. Reactions or t h e prod uction of Be and Al in quartz ioBe 26AI 16 Spallation reactions 0(n,4p3n"l 1°Be 28 10 28 Si(n,x•) Be Si(n,p2n)26 Al Negative muon capture 160(µ',3o3n) 10Be 28Si(µ·,x)10Be 28Si(µ-,2n)26Al 10 23 a reactions \i(a.,o) Be Na(a,n)26AI 9 10 n reactions Be(n;y) Be 10B(n,p)10Be

nC(n,a.) 1oBe the total of protons and neutrons that are emitted in the reaction, not meant to imply the actual make-up of the emitted particles. *the x stands for several protons and/or neutrons that are emitted in the reaction.

10 26 The production reactions for Be and AI are shown in Table Bl .2. Radiogenic production (Sharma and Middleton 1989) of the isotopes as a result of 1) decay of U and Th and 2) spontaneous fission of U yielding a-particles and neutrons is also possible by the reactions shown in Table Bl .2. For the radionuclides this contribution is usually 3 21 small, but it may be significant for the production of the stable noble gases ( He and Ne) 10 in rocks with old crystallization ages and high U and Th contents (Poreda and Cerling 1992, Ceding and Craig 1994a). We can calculate the production of the isotopes with time and thus determine the age of the exposed rock surface with Eq. 1.

N=-1p ( e-(A.+pe/A)T)+Noe-AT (1) A.+ PE A where 10 26 N is the number of atoms/gram, (of Si02 for Be and AI),

N0 is the number of atoms/gram of the cosmogenic isotope present at the beginning of exposure, Pis the local production rate in atoms/gram•yr, Tis the length of time the surface has been exposed in yrs, 1 is the decay constant of the radionuclide in 1/yr, 3 pis the rock density in g/cm , e the erosion rate in cm/yr, and 2 A is the cosmic ray attenuation length in the rock surface in g/cm •

The second term reduces to N0 when the concentration present at the beginning of exposure is attributable to radiogenic production. 10 26 While Be and Al are produced primarily by spallation, 36CI is produced by several mechanisms (Table Bl.3), that is from spallation and thermal neutron capture. For 36 that reason, the production rate for CI is composition dependent and must be calculated individually for each rock analyzed (Phillips et al. 1986, Zreda et al. 1991 ). For 36CI, P in Eq. 1 is therefore calculated by:


where P is the local production rate in atoms 36CI /kg rock.yr, 36 'I' is the production rate of CI from K or Ca, 36 in atoms CI /kg rock.yr.unit of concentration used for CK and Cea, C are the concentrations of Ca or K, +n is the thermal neutron capture rate in units of neutrons/kg rock.yr, 35 2 cr35 is the thermal neutron absorption cross section for CI in cm , 35 N3s is the number of CI atoms per kg of rock, 2 cr; is the thermal neutron absorption cross section for element i in cm , and N; is the number of atoms of element i /kg of rock. 11

36 The production rate for CI must then be calculated using the measured concentrations of the major elements, the concentrations of the trace elements with high neutron capture cross sections (mainly Band Gd), and the total Cl concentration for that 36 35 rock. Non-cosmogenic CI can be produced by neutron capture on CI (Table B1 .3), with the neutrons originating from the U and Th decay chains and secondary (o;,n) 36 14 reactions. The resulting Cl/CI ratios can be as high as as 10· , but they are usually more 15 on the order of 3 x 10· (Feige et al. 1968, Andrews et al. 1986, 1989, Fabryka-Martin 1988). The contribution from this production mechanism can be calculated by using measured U and Th concentrations, the concentrations of the targets for the (o;,n) reactions and the Cl concentrations. This must then be subtracted from the measured 36 Cl/CI ratio for that rock.

Tab Ie Bl .3. Reactions or·36 C I prod uction (Fa brv1 k a-Martin 1988, Zreda 1994). 36CI

Spallation of K and Ca 39K(n,2n2p) 36CI 4 °Ca(n,2n3o) 36CI Thermal neutron activation of Cl 3sCl(n,y) 36CI Thermal neutron activation of K 39K(n,o;) 36CI 39 36 Negative muon capture by K K(µ-,p2n) CI Negative muon capture by Ca 40Ca(µ',a.) 36CI

Figure B 1 .1 shows the increase of the concentration of several cosmogenic isotopes in the surface of a rock at sea level with time. This diagram was made assuming no erosion, that is E = O in Eq. 1 . When there is no erosion, the noble gas concentration increases linearly with time. The concentration of the various radionuclides, on the other hand, will eventually become constant. This is called saturation or secular equilibrium, and occurs after 3 to 4 half-lives. Saturation is the concentration where production and decay are balanced. Reaching of saturation limits the application of the method with 10 26 36 radionuclides to ca. 5 Ma, 2.5 Ma, and 1 Ma for Be, Al, and CI, respectively (Table B1.3). Several assumptions must be satisfied in order for the relationships of Eq. 1 and 2 to have simple solutions. The first assumption is that the concentration of the cosmogenic isotope at the beginning of exposure is zero or is known (N0 in Eq. 1). That means that there should have been no recent prior exposure of the rock surface. During exposure, the system must have been closed with respect to either gain or loss of the respective isotope. Loss can occur by diffusion or leaching from within the mineral grain (i.e. 12

3 diffusion of He from quartz, Brook and Kurz 1993). Gain would most likely occur by 10 contamination with meteoric water in the case of Be, which contains several orders of 10 magnitude more Be than is produced in situ in the rock (Chapter 63). The last assumption is that the correct production rates are known and have not varied with time.

1.00E+ 10 ...----....--....----.-___,,.., 3He

1.00E+04 +-·--+--+---+--...... of I(.) co «> 0 0 0 0 ~ + + ""' + + w w w w w 0 0 §l $ C! C! $

Time in years

3 10 26 36 Figure B 1.1. Increase in the concentrations of He, Be, Al and CI in a rock surface as a function of exposure time (at sea level).

81 .3.1 Complicating factors-known The factors affecting isotope build-up that are known or at least can be described mathematically, so that the production rate can be corrected for, include shielding by surrounding hillsides, sample thickness (depth in the rock), latitude, and altitude. Production rates vary on the surface of the earth due to atmospheric thickness and magnetic field spatial variations, so that the production rate for each rock must be adjusted based on the rock's location. Detailed description of these variations by Yokoyama (1977) and Lal and Peters (1967) have been supplanted by Lal (1991 ), where the variations have been set out in polynomial form. The uncertainty of this relationship is estimated to be ±10% (Lal 1991).

81 .3.1.1 Sample thickness Because secondary particles are attenuated as they travel through rock, the production rate of cosmogenic nuclides decreases with depth. For isotopes produced through spallation reactions, this depth dependence can be expressed as: 13

(3) where

P, is the production rate at depth x in atoms/g Si02.yr,

P0 is the production rate at the surface in atoms/g Si02.yr, 3 p is the rock density in g/crn , x is the depth in cm, and 2 A is the cosmic ray attenuation length in g/cm .

P/ 0 0.2 0.4 0.6 0.8

Depth (cm) 30•-·~--i----+-·-•t----+~~

Figure B 1.2. The decrease of the production rate with depth from the surface of the rock expressed as a fraction of the su1iace value. The dashed line is the actual relative production rate at that depth (Eq. 3), while the bold line is the production rate integrated and thus applicable for the whole interval (Eq. 4).

According to Eq. 3, at a depth of, for example, 10 cm, the production rate is 83% of 3 the surface production rate (using the density of granite, 2.7 g/cm , and an attenuation 2 length of ·1 so g/cm , Brown et al. 1992). This relationship is shown by the dashed line in Fig. B1.2. The integrated production rate over a certain depth interval is shown by the bold line in Fig. B1.2 and is described by the following equation which is obtained by integration of Eq. 3:

P(Otox)--Po- A (1 -e-px/A) (4) px

The variables are the same as in Eq. 3, Pio 10 xi is the average production rate for the interval from the surface downward to the given depth. This equation applies to a slab of rock of thickness x. For example, for a slab 10 cm thick the overall production rate is 92 % of what it would be right at the surface. 14

Because 36CI is produced not only by spallation, but by thermal neutron capture as well, the depth correction for 36CI is not so straightforward as for 10Be and 26 Al. The depth profile of thermal neutrons is different from that of fast neutrons. Thermal neutrons are able to escape back out of a rock surface (and the sides as well) of a boulder or knob. The variation of the neutron flux with depth (Fabryka-Martin 1988, Zreda 1994) is expressed as: (5)

where c?x is the thermal neutron capture rate at depth x in neutrons/kg rock.yr,

cp0 is the thermal neutron capture rate at the rock surface in neutrons/kg rock.yr, 3 pis the rock density in g/cm , x is the depth in cm, a,c,b and d are constants.

P/Po 0 0.5 1 1.5 2 0


40 Depth (cm) 60



Figure Bl .3. The applicable relative production rate integrated over a given thickness interval of rock. The thin line is for spallation products while the bold line is for thermal neutron capture reactions.

For the whole thickness interval the following equation applies:


where a=-19.676; b= 0.0383; c= 2.9519; d=0.0064 15

These constants are from Zreda (1994) and were fitted to measured thermal neutron intensities of Fabryka-Martin (1988). Eq. 6 is plotted as the bold line in Fig. Bl .3. Notably, the neutron flux reaches a maximum at a depth of within the rock. For comparison, the thin line is the depth correction for spallation reactions. Both lines are the integrated production rate at depth for the entire interval. Shielding of the rock surface Shielding of the rock surface from surrounding hillsides and the dip angle of the rock surface sampled must also be taken into consideration because these factors reduce the cosmic ray flux that the surface 'sees'. The angular dependence of the cosmic ray flux relative to the vertical flux (Nishiizumi et al. 1989) is:

F(e) = sin 23 e (7)

where F(0) is the relative angular dependence of the cosmic ray nucleonic component, 0 is angle from vertical.

This correction must be calculated including both the dip of the rock surface and the surrounding skyline. As an example, for a rock surface dipping at an angle of 25°, the reduction of the production rate will be less than 2 %.

Bl .3.2 Complicating factors - unknown Several other factors may seriously affect the production rate of the isotope of interest in the rock surface (and therefore the exposure age). But because they pertain to events in the past that cannot be known precisely, the net effect of these factors is difficult to constrain. Such factors can be divided into factors relating to the cosmic ray flux, and those that are site specific. Factors affecting the cosmic ray flux, and thus the production of the cosmogenic isotopes, include variation of the primary cosmic ray flux with time (Klein et al. 1990, Reedy et al. 1990, Fink et al. 1990), changes in solar activity (e.g. Beer et al. 1991) and variations in the earth's magnetic field (e.g. Castagnoli and Lal 1980, Lal 1988). Increases in either solar activity or the strength of the earth's magnetic field act to reduce the flux of galactic cosmic rays reaching the earth's surface and therefore decrease production of cosmogenic isotopes (Lal and Peters 1967). 16

81 .3.2.1 Magnetic field effects The magnetic field deflects cosmic rays, thus the effect is weaker at the poles where the magnetic field lines point inward. The effect of the earth's magnetic field can be attributed to variations in the intensity of the magnetic dipole moment (field paleointensity) and changes in the location of the earth's dipole axis with respect to its geographic axis in time (dipole wobble). Lesser effects may be due to short term variations of non-dipole fields. Paleointensity measurements, performed on archaeological materials, lava flows, and sediments have been used to track the paleogeomagnetic field (McElhinny and Senanayake 1982, Mazaud et al. 1991, Meynadier et al. 1992, Tric et al. 1992, Thouveny and Creer 1992, Merrill and McFadden 1994). Measurements for the same time period but at different locations on the earth do not yet provide an unequivocal picture (e.g. Thouveny et al. 1993, Lehmann et al. 1996). Agreement has been reached over the 10 intensity low at ca. 40,000 yr BP. This low is apparently reflected in an increase of Be production as recorded by unusually high concentrations measured in ice cores (Raisbeck et al. 1987, Beer et al. 1988, 1991) and in deep sea sediments (Castagnoli et al. 1995, McHargue et al. 1995, Robinson et al. 1995). Although the earth's present magnetic dipole is ca. 11° off from the geographic pole (near 270 °W and 79 °N), during the last 10,000 years the geomagnetic pole has on average roughly approximated the geographic pole (Ohno and Hamano 1992, 1993, Sternberg 1992). That means for samples with exposure ages of more than 10,000 years, the geographic rather than geomagnetic pole should be used in the latitude and altitude corrections. We have used the geographic latitude for all our calculations.

81 .3.2.2 Unknown factors specific to the site or rock surface Site specific factors which are for the most part unknown (for the period of exposure) and therefore are difficult to quantify include: • shielding by ice, snow or sediment; erosion of the surface, spalling of large flakes from the rock surface. • burial following exposure; • shifting, rolling over or exhuming of boulders, • alteration of the minerals due to weathering of the rock, prior exposure (grains in sandstone, for example) All of these processes except the last one, yield exposure ages which are too young. Indeed they are more likely than those yielding too old ages (prior exposure) (Phillips et al. 1986, Kurz and Brook 1994, Zreda 1994, Zreda et al. 1995, Phillips et al. 1996b). 17

Similar to the discussion above concerning the thickness of the rock slab, burial of the surface by snow, ice, or sediment reduces the production rate as well. The surface which is sampled today may actually have been covered, for example by snow or till, during the assumed period of exposure. But today little evidence of that coverage remains. An example of just such a scenario is discussed in detail in Chapter 03, which is about samples from the Grimsel Pass region. To a lesser extent, trees block cosmic rays. Indeed, present day vegetation may have little to do with that in the past, making corrections for vegetation shielding difficult. For example, many glacial erratics of the Swiss Alpine foreland are located in fields, while during much of the Holocene those areas were covered with forests. Cerling and Craig (1994b) estimated the reduction of cosmogenic isotope production due to trees and leaves to be on the order of 4 % for samples taken in forests in the Puy de Dome area of France. Exposure ages which are too young can also occur due to post-depositional movement or settling of the boulders on the moraine. Such processes include turning over of blocks by rolling down the sides of the moraine or constant overturning of blocks due to heaving action, etc. Large boulders located on a moraine crest may roll off as the finer material erodes quickly away in the first 50-100 years as the moraine takes on a more subdued shape (see also Hallet and Putkonen 1994, Zreda et al. 1994). There are several methods to get a handle on possible erosion or weathering of the rock surface itself, none of which are unequivocal. Rock surface lowering may be the result of active erosion and transport by wind or water (temperate regions) or may be just the result of in situ weathering (as in Antarctica). The multiple production mechanisms of 36CI make it possible to reveal erosion and erosion rates with a single isotope, by 36 measuring a CI depth profile and/or by measuring different mineral separates from the same depth (e.g. Dep et al. 1994). Ideally, the concentrations of 36CI with depth will reveal the maximum of the neutron absorption depth (Fig. Bl .3). In certain situations where more exact knowledge of the erosion rate is needed especially when the exposure time is less than several hundred thousand years, the 36CI depth profile or mineral separates method can be applied. Erosion or weathering of the rock surface may also be revealed by using a combination of isotopes, e.g. 10Be and 26Al. An advantage of using this isotope pair is that they are measured in one sample (only one chemical preparation). This has its most important application for surfaces with exposure ages in the range of several hundred thousand to several million years. 10 Figure B"l .4 shows the effect of erosion on the accumulation of Be with exposure 10 time. The upper bold line follows the increase of the Be concentration when there is no 10 erosion (e=O in Eq. 1). Each of the lines below follows the increase of the Be concentration when erosion is included. Erosion or weathering always removes the upper 18 few millimeters or centimeters of the rock surface. This is the area where the production rate of the cosmogenic isotope is the highest. If this part is constantly being removed, an effectively lower production rate for the whole time of exposure results. For the cosmogenic radionuclides, the end effect is that saturation is reached earlier at a lower concentration (compare the upper line with the lines below it in Fig. Bl .4). The magnitude of erosion for each individual rock can be roughly revealed by 26 plotting an isotopic ratio vs. the concentration of one of the isotopes, e.g. Al/1°Be vs. 10 26 10 Be (Fig. B 1.5). In the left hand part of the plot, the AV Be ratio is controlled by the 26 10 production ratio, i.e. ca. 6. Since Al has a shorter half-life than Be, it reaches 26 saturation earlier and the Al/1°Be ratio decreases with longer exposures. The banana- shaped area is known as the erosion island (Klein et al. 1986, Lal 1991, Nishiizumi et al. 26 10 1991, 1993). In the case of no erosion, the Al/ Be ratio follows the upper boundary with time. The lower right hand end of the erosion island is where both isotopes are in 26 10 saturation ( Al/ Be ratio is 2.88) and the ratio does not change with longer exposure times. The lines inside show the evolution of the ratio with time for different erosion rates. The lower dashed line connec..15 the final ratios for each erosion rate and is not itself an evolution line. Note that neither of the axes portrays the passage of time. Significant coverage of the surface following exposure can be revealed by the 26 Al!1°Be ratio. If a rock surface had been covered by enough ice or sediment, after a 26 10 period of exposure, that cosmic rays are blocked, both Al and Be begin to decay. 26 10 26 10 Since Al has a shorter half-life than Be, the Al/ Be ratio decreases. The plotted results would then fall below and to the left of the erosion island in Fig. B1 .5. In general, it is difficult to distinguish simple exposure histories (with erosion only) from complicated exposure histories involving periods of erosion and periods of burial. This is because each scenario does not result in a unique location on the erosion island plot. Complicated exposure histories cannot be revealed with a noble gas measurement alone. Since the noble gases do not decay, they retain their isotope inventory from every exposure, even after coverage by ice or sediment. The cosmogenic noble gases record a composite exposure history made up of all episodes of exposure. 19

L....i.....-...,_---1 NO EROSION 5 cm/Ma 9 i8: 10 cm/Ma 0 "' 8 106 00 7 106 E 6 106 30 cm/Ma f: 6 -E' 5 10 50 cm/Ma 6 E"' 4 10 ~ 3 106 Ill ~ 2 106 Ill 106 :& 107 Time in years LI)

Fig. B1 .4. The effect of erosion or weathering on the increase of 10Be concentration as a function of time (sea level). The upper bold line is plotted (Eq. 1J withe =o. The four lines beneath it are plotted with the stated values for€.


2Ma Ill 0 ~-- 5Ma <( 4 "'


10 Be atoms/gram

26 Figure B1 .5. Erosion island plot of Al/10Be vs. 10Be. 20

It is a common misconception that cosmogenically produced 21 Ne (or 3 He) do not reach saturation because they do not decay O.. =OJ. Indeed, the rate of growth of the concentrations of the cosmogenic noble gases are also decreased by erosion or weathering. The concentration of the noble gas reaches an apparent or pseudo saturation in the presence of erosion. Such an effect is governed by Eq. 1, where A.= O, and the following equation results:

N=i>x p (1 e-(pe/A)T ) (8)

The effect of erosion on 21 Ne concentration with exposure time is shown in Fig. B1 .6. The straight bold line shows the increase of the 21 Ne concentration with no erosion (e = 0), while the lines below show how the 21 Ne levels off when there is erosion. Interestingly, an erosion rate of only 23 cm/Ma results in pseudo-saturation of 21 Ne after ca. 5 Ma, meaning that the effective half-life is ca. 1.5 Ma or the same as 10Be.


ON 2.50E+08 v; E 2.00E+08 2 -....OJ E l.50E+08 -§ l.OOE+08 (J) z « 5.00E+07

O.OOE+OO 8 "' 0 + ~ Cf"' '; "' w Li.I .t § 8 8 ci iri s ~ N Time (years)

Figure B1.6. Effect of erosion on the increase of the 21 Ne concentration. The bold line is with no erosion, the thin line with 5 cm/Ma, the long dashed line with 10 cm/Ma and the short dashed line with 20 cm/Ma.

Erosion rates cannot be elucidated with one stable isotope alone. But, erosion rates may be revealed by plotting a stable cosmogenic isotopes vs. a radionuclide (Graf et al. 1991 ), e.g. 21 Ne;1°Be vs. 10Be. Such a plot is shown in Fig. Bl .7. Results which plot 21 within the area of the curving lines have been affected by erosion with respect to the 21 Ne and 10Be concentrations. The plot of 21 Ne/0 Be vs. 10Be is especially useful for revealing prior exposure. Such results would plot above the 21 Ne/1°Be vs. 10Be erosion island in Fig. Bl .7. They would plot above because there is more 21 Ne than expected for the 10Be concentration. Prior exposure is the case where a rock surface has been exposed to cosmic rays before the exposure that we are interested in. The resulting exposure age would be 'too old'. There are probably many scenarios whereby a rock can have been exposed before it arrives in the setting in which we find it. These include exposure along a cliff surrounding a glacier, exposure in an older moraine, and where sub-glacial erosion has been insufficient to remove all of the previously exposed rock. The coarseness and angularity of much of the material on moraines indicates it is supra-glacial in origin (e.g. Summerfield 1991 ). That leaves open the possibility that a boulder may first have been exposed in a cliff above the glacier. Later this cliff could have collapsed and the blocks landed on the surface of the glacier. If the exact same face came to rest facing upwards, then it would also be exposed as it is on the glacier surface. Eventually the large boulder could be carried and left along a moraine crest. For this boulder to have acquired several periods of exposure, the exact same rock face must always have been exposed upward to cosmic rays.

~ 10+----l---+-·---+----+- "';;i l 8 +-----!---+-----;------+-

~ ~ ~ ~ ~ + w w w ~,._ ~ ;? lil ~ 10Be atoms/gram

Figure B1 .7. Plot of 21 Ne/ 0Be vs. 10Be. The bold line is for no erosion, the thin line for erosion of 10 cm/Ma, and the dashed line for 20 cm/Ma (after Graf et al. 1991 ). 22

Perhaps more likely is a boulder from an older moraine getting reworked by a glacier into a younger moraine. This is especially possible in Antarctica, where the cold- based glaciers are not capable of extensive sub-glacial erosion. A glacier terminating at roughly the same position could leave boulder trains very nearly on top of earlier deposits. Such peaking through of older boulders into younger moraines has been offered as an explanation for some of the spread in data in the Arena Valley moraines (Brown et al. 1991, Brook et al. 1993, 1995b). Still, the same rock face would have to have been exposed in both episodes. As we often date several boulders from a single moraine, we postulate that those with prior exposure would stand out as older outliers. A final example of prior exposure involves a glacially polished surface that may have a non-zero initial concentration. That may occur if a glacier did not erode more than three meters of rock in its passing. At a depth of three meters the production of cosmic rays is 1 % of its surface value. The magnitude of the effect of residual concentrations from prior exposure depends not only on the depth in the rock, but the length of time the rock was exposed before the glacier scoured it. For example, if a rock had been exposed 100,000 years prior to glacial polishing. Then 3 m of rock are removed by sub-glacial erosion, the contribution to the concentration accumulated after glacial retreat (ca. 1 0,000 years) would be less than 5 %. This can also be checked by measuring the concentration of a cosmogenic noble gas. Because the noble gases are stable and do not decay, they record all previous episodes of exposure. The noble gas age would thus be older than the radionuclide age (which had begun to decay during coverage of the surface by ice), but as well depends on the actual duration of exposure and time of burial. Weathering and alteration of individual mineral grains of the rocks surface may lead to open system behavior for the isotope and/or contamination with, for example, meteoric 10Be. In the case of quartz, which is very weathering resistant, it does not appear to be a problem. But for other minerals (especially pyroxene, or plagioclase), contamination may be significant (see also Chapter B3). The hydrophilic nature of chloride means that meteoric 36CI is often washed through rocks, i.e. is not retained through adsorption. In contrast, 10Be is much more surface reactive. Extreme weathering may result in changes in the major element composition which would affect the calculated 36CI production rate (Zreda 1994) and a spurious exposure age may result. It is impossible to correct for these unknown factors although one can make justified estimates. Additionally, as discussed above, multiple isotope investigations can help to unveil complicated exposure histories. 23


82.1 INTRODUCTION This chapter is devoted to a detailed description of the methods used to extract Be, Al and Cl for AMS measurement from rock samples. It is broken down into three parts: 1) information on the field methods and sampling, 2) sample preparation chemistry for 10Be and 26AI from quartz and 36CI from whole rock, and 3) AMS measurement. As listed in Appendix H1, 73 individual rock or mineral separates from 52 different rock samples were dissolved during the course of this dissertation. This does not include quartz or beryl dissolved for blanks. Many of these were performed 1} as a check on the reproducibility of the extraction procedures which we set-up, and 2) to check that all meteoric 10Be contamination had been removed, so that reliable ages could be determined.

82.2 SAMPLING Perhaps the single most important factor for obtaining exposure dating results which make sense and provide an as accurate as possible date for the feature under investigation, is the sampling. The degree to which a chosen rock surface or boulder is representative of the geomorphologic feature or event to be dated is of course the first consideration. Whether or not a boulder has remained stable and completely exposed since it was deposited must also be assessed. In this light, the presence of an experienced, local or geomorphologist in every sampling expedition is essential. Samples were taken with the help of a chisel from the center of the rock surface (usually more than 20 cm from the edge), whether it be a boulder or a glacially polished knob. Quartz veins were removed by chiseling along the vein margins, while for bare rock faces, small cracks were taken advantage of. Average thickness of the slabs removed for sampling was 5-10 cm. We generally tried to sample high-standing surfaces, such as the largest boulders, because for such surfaces, it is most likely that snow would be blown off. Sketches were made and photographs were taken of every sample and site. The strike and dip of the rock surface were recorded before chiseling. Approximate slab thickness as well as were also noted in the field. Cosmic ray shielding by the surrounding countryside was estimated or measured with a clinometer in the field, and determined from topographic maps. For a rock of granitic composition with an exposure age in the range of 10,000 years, ca. 50 g of very pure quartz are needed for the making of BeO and A!i03 AMS targets. This is extracted from 0.5-1 kg of 0.5-2 mm sized crushed rock. For the 36CI 24 procedure, ca. 100 g of rinsed whole rock are needed(< 0.4 or 0.5 mm grain size). Therefore, samples on the order of 1-2 kg were taken from each rock surface.

82.3 SAMPLE PREPARATION Prior to sample preparation, thin sections were prepared for many rock samples. A brief inventory of petrographic characteristics of the rocks investigated are tabulated in Appendix H2. In thin section one can check the size, percent and nature of the quartz in each rock, and the accompanying mineralogy. This information was used to pick the correct grain size to use to gain monomineralic quartz grains. Additionally, one can head off complications related to the presence of certain minerals which may come up during sample preparation. As discussed in later chapters (D3 and 04), the success of the Be and Al extraction procedure is strongly dependent on the size and nature of the quartz. The upper surface of the rock samples was cleaned of lichen etc. with a wire brush. Rock samples were then smashed to small chunks with a sledge hammer, and afterwards put into a steel jaw crusher set to 1-2 mm grain size. The crushed material was sieved through nylon sieves, and the larger chunks (> 2 mm) were put through the jaw crusher again. Further size reduction was accomplished by using a steel mortar, followed again by sieving. Generally, the grain size groups 0.5-1 and 1-2 mm were separated, but this depended on the actual mineral grain size within each rock. These two size fractions were used for the Be and Al sample preparation. The < 0.4 or< 0.5 mm size fraction of the crushed whole rock was retained for 36CI measurement. For the extraction procedures, general cleaning of all labware (including all disposable pipettes, centrifuge tubes, and plastic bottles) entailed rinsing with ultrapure (18 MO) water and dilute HCI (not used in the case of 36CI samples). Teflon beakers (or bottles in the case of Cl) were cooked for several hours in dilute HN03, then rinsed at least 10 times with ultrapure water.

82.3.1 Be and Al extraction from quartz Apart from a few test samples made from meteorites, the extraction of ' 0Be and 26AI from silicate rock had never been performed at EAWAG or at the lnstitut fur Teilchenphysik, thus it was necessary to set up the complete extraction procedure from beginning to end. With respect to 10Be and 26AI sample preparation, the following procedures were realized at EAWAG by modifying procedures described in the literature (Nishiizumi et al. 1989, Brown et al. 1991, Nishiizumi et al. 1991, Brown 1992, Kohl and Nishiizumi 1992, Brook 1993). These included the gaining of pure quartz grains from granites, sandstones, quartzites, schists, and gneisses; complete 25 quantitative dissolution of the quartz; separation of Be and Al by cation exchange; and final preparation of BeO and Al 20 3 AMS targets. The sample preparation chemistry for Be and Al can be divided into two areas; 1) the gaining of a pure quartz separate from the crushed whole rock and 2) complete digestion of the quartz to extract the Be and Al to be made into AMS targets. An essential part of the former is that during the procedure meteoric 10Be is also removed.

82.J.1.1 Selective chemical dissolution - the making of a pure quartz separate The coarsely crushed whole rock was rinsed thoroughly with ultrapure water before beginning the selective dissolution procedure (Kohl and Nishiizumi 1992). In the rinsing, the fine grained material was discarded. About 50-70 g of the crushed rock were put in a 1 L bottle, which was then filled to 0.5 L with ultrapure water and 50 ml of 40% HF. Such a bottle can be left on the shaker table for several days. Between each HF step, the rock material was rinsed three or more times with ultrapure water. For the next HF step the bottle was filled again with water and HF as described above. After HF3 or 4, the sample material was transferred to a glass beaker then examined for purity under a binocular microscope. Large grains were removed with a pair of tweezers. For an average granite or granodiorite (e.g. Ju lier Pass samples), after HF3, a few grains of feldspar were all that was remaining. At this point, if the grain size did not seem small enough that the quartz was present as single grains, this material was ground gently with an agate mortar and pestle. After the final HF step (5 or 6), the quartz was transferred to a glass beaker, rinsed at least 10 times with ultrapure water, then dried on a hot plate. The quartz yield is rock type dependent, but for granites, 8-10 g of clean pure quartz was usually won from 50 g of whole rock. In addition to examination under the binocular microscope, we measured the total Al concentration by ICP-AES in aliquots from several etching steps for samples from each rock type (results for granite 217b are shown in Fig. B2.1 ). When only pure quartz is present, the Al content of the material generally should be less than a few hundred ppm. We noted that the total Al content does not flatten out completely until the 4th or 5th HF etching step (d. Kohl and Nishiizumi 1992). Figure B2.1 shows the weight loss during each etching step, which was rapid at first, as the were dissolved, then leveled off as just the outer portions of the quartz grains were dissolved. In general, the first one or two HF steps eliminate the possible meteoric 10Be (see also discussion below), while the last few HF steps wear-off the remaining feldspar grains, that were adhering to the quartz grains. This last process is important so that the total Al concentration is low enough that an AMS-measurable 26Al/17 Al ratio is attained. 26

100 90 ic 80 ------~ 70 -----

Figure B2.1, Sample weight remaining (in percent of the starting sample weight) after the indicated selective dissolution step (i.e. HCI leach, HF1, HF2, etc.) for granite sample 217b. The total Al content in ppm (as measured by ICP-AES) from an aliquot of the indicated step are also shown (note that we did not check the Al concentration in every step).

Even after 5 or 6 HF steps, there are very recalcitrant minerals, such as zircon and epidote, that are not attacked by weak HF. For example, the honey-colored zircon noted to be present after 5 HF steps in samples made from the granitic erratics (Chapter 04), was especially resistant to HF. As zircon will still not dissolve even in the final digestion with concentrated HF it does not present a problem. On the other hand epidote (and sometimes muscovite) are not removed during the purification steps with weak HF, but are dissolved with concentrated HF. Because we need large amounts of quartz to measure the very young exposure ages of the samples from the Alps, the presence of even 1 % accessory minerals dilutes the Be (and Al) in the final sample material. This drastically reduces the efficiency of the AMS measurement. For example, in recent measurements we noted that the Fe, Mg, and Ca contributed by the ubiquitous epidote found in the Aare Granites of our Grimsel Pass samples (even in the quartz veins) diluted the Be causing low AMS source currents (see also chapter 03). Thus, we had to add a heavy-liquid mineral separation step after HF4 then followed by HFS. We used sodium metatungstate to separate the heavier epidote from the quartz, and therefore make a purer quartz separate. 27


add Be spike

HF digest

heat to dryness 3X with Hl'.03 ,o, ,

redissolve with HO heat to near dryness take up In 1 M HCI Al

load onto cation exchange column elute Al wtth 4.5M HCI Be heat to near drynes lake Up In 1M HO elute Be with 1M HQ I

ppt with NH,o.CH heat to near dryness pH8 lake up in 1 M HO centrifuge, decant pp! with NH,,oti heat to dryness pH 8-9 In quartz cruclbl

centrifuge, decant Bake in furnace heattodryness at 850 'C in quartz crucible

Bake in furnace mix with copper at 850 'C press Into target

mix wlt h copper press into target

Figure B2.1. Flow chart showing the basic sample preparation steps for 1 Ose and 26AI extraction from quartz (ppt=precipitate). 28 Dissolution of pure quartz and chemical separation of Be and Al The flow diagram (Fig. B2.2) shows the sequence of steps involved in sample preparation from dissolving the pure quartz (obtained by selective chemical dissolution) to final pressing. The dried teflon beaker was weighed and the quartz was 9 poured in, beaker and quartz were weighed again. Then, 0.5 ml of a Be(N03Ji solution (Merck ICP standard 1 mg/ml) was added as carrier using an Eppendorf autopipette. The sample was weighed again following the carrier addition. HF (48%) was added to just cover the quartz. The beaker with the sample and HF was heated on a hot plate at 80 °C, until all of the quartz had dissolved. Approximately 5 g of quartz are dissolved in 24 hr., that means that a 50 g quartz sample may take up to 1O days to dissolve.

After the quartz had dissolved completely, concentrated HN03 was added and the last remaining SiF4 was fumed off. After at least three fumings with HN03, concentrated HCI was added, the solution was fumed again several times. The sample was then heated until it was nearly dry to homogenize the carrier and sample Be. During this fuming, it is important that the temperature be kept low (80 °C), because

AICl3 can be lost to volatilization at 183 °C. The loss of Al at this point (i.e. before the aliquot for ICP-AES has been taken) can lead to serious errors in the total Al measurement and thus in the calculated ages. This fuming step is also crucial because fluoride which is not fumed off can interfere with the column separation. AIF 3 complexes will not adsorb to the exchange resin as strongly as Al +, resulting in some of the Al coming off the column too early (i.e. being found in the Be eluant). Additionally, any Ca found in the sample solution can be separated by the first precipitation (before the column). But in the presence of fluoride, fluorite precipitates with the Be(OH}i and Al(OHh and is therefore carried through onto the column where it can cause column overload. These topics are discussed in detail in Ochs and Ivy-Ochs (1996). Note that fuming with perchloric acid can better guarantee removal of fluoride but can must be done in an appropriate fume hood. The sample solution was then transferred to a 100 ml flask using a disposable pipette. The beaker was rinsed several times with 18 MO water which was also added to the flask. After filling up to the mark, the flask was shaken thoroughly turning it upside down several times. For the ICP-AES aliquot we took 5 ml. The aliquot was stored in a sealed tube. After the aliquot had been taken the solution was heated to reduce the volume to ca. 2 ml. The solution was transferred to a disposable centrifuge tube and all hydroxides, including Be(OH)2 and Al(OHh, were precipitated with NH40H. In general, we added NH40H until a precipitate began to form, then checked that the pH was at 8. 29

In Figures and B2.3b, the speciation of both Be and Al over the complete pH range are shown. Speciation has been calculated using MINEQL+ (Environmental Systems Research), an upgraded version of the well-known speciation code MINEQL (Westall et al. 1976). A complete list of stability constants used can be found in Ochs and Ivy-Ochs (1996). For these calculations, we have used actual measured sample concentrations for Be and Al. The speciation modeling indicates that both Be and Al will precipitate at pH 8. The influence and interference of an excess of Ca, Mg, and/or Fe on the efficiency of the Be and Al extraction chemistry based on speciation calculations are presented in detail in Ochs and Ivy-Ochs (1996). The solution was centrifuged and the liquid was decanted. The precipitate was redissolved with concentrated HCI. This solution was transferred back to the teflon beaker and heated to near dryness. Some of the sample solutions had a fair amount of iron present, which was shown by the solutions bright yellow color. It may have come from very fine magnetite inclusions in the quartz grains. The iron should be removed prior to the cation exchange column step because it may overload the column and interfere with separation efficiency. It can be removed with an anion exchange column or with a MIBK (methyl isobutyl ketone) extraction (Knauer 1994).

I I I I BO I I 70 (a) I I ·-·-- Al+3 I 60 I 50 ---- - AIOH+2 1 40 - - --- Al(OH)4· 1 I 30 --- Al(OH)3 (am) 1 1 I 20 I I I 10 I I

0 2 3 4 5 6 7 8 9 10 11 12 13 14


Figure B2.3a. Speciation of Al plotted as a function of pH. 30

100 I 90 I I (b) I 80 I I ----- Be+2 70 I I - • - • - • BeOH+ I 60 I I - • · - • · · Be(OH)3- 50 I I · - • · · · - • Be(OH)4-2 40 I I ---Be20H+3 I 30 I --- Be3(0H)3+3 t 20 i:" " 10 - Be(OH)2(s) I;"

0 0 2 3 4 5 6 7 8 9 10 11 12 13 14 pH

Figure B2.3b. Speciation of Be plotted as a function of pH (similar diagrams for Ca and Mg are presented in Ochs and Ivy-Ochs 1996).

For this extraction, the sample solution in 20 ml 4.5 M HCI was transferred to a separatory funnel and 20 ml of MIBK were added. The solution was shaken vigorously and let stand. The sample was then found in the lower clear solution which was transferred back to the teflon beaker. The upper yellow part contained the MIBK and

the FeCl 3. With this extraction, no back extraction was necessary, because the Be and Al are found in the aqueous phase not the organic. The solution was heated to near dryness and taken up in 1-2 ml of 1 M HCI for loading onto the cation exchange column. Teflon columns of 1 cm diameter and 20 cm length were used. The water equilibrated volume of resin (Biorad analytical grade Ag50W X 8 cation exchange resin, H+ form) was 10 ml for samples of 5-10 grams quartz and 20-25 ml for samples of> 40 g quartz. The resin was pretreated with 5 column volumes water, 5 volumes 9 M HCI, 5 volumes 4.5 M HCI, then 5 volumes of 1 M HCI. The Be fraction was found in column volumes 4-7. Column volumes 3-8 were eluted into the teflon beaker. Several column volumes of l M HCI were allowed to run through before the Al fraction was taken. The Al fraction was eluted with 2-3 column volumes of 4.5 M HCI. 31

The sample solution was again heated to dryness and taken up in 2 ml of 1 M

HCI and transferred to a disposable centrifuge tube. Be(OHh (respectively, Al(OH)3 ) was precipitated at pH 8 with NH4 0H. The precipitate was centrifuged for 1/2 hour and the solution decanted. Ca. 5 ml of ultrapure water and one drop NH40H were added to rinse the precipitate. The precipitate was rinsed twice and the supernatant decanted. The hydroxide was transferred to a new quartz crucible (1 cm diameter 1.5 cm long) using a 0.5 ml disposable pipette, then dried under infra-red heat lamps for ca. 2 hr. Using tweezers, the crucibles were moved to a quartz sled. Tthen each crucible was covered with a quartz lid. To convert Be(OH)z and Al(OH)J to the oxides the sled was placed in a muffle furnace with the following program: 2 hr. at 150 °C for final drying, a ramp to 850 °C, then 2 hr. at 850 °C. The BeO (A'203) was mixed with pure copper powder (ratio of roughly 1 :4), then the mixture was pressed into the 1.5 mm hole of the copper disk target holders used at the PSI/ETH AMS facility. All samples were pressed inside of a laminar-flow dust cabinet (inflow).

Ta bl e B2.1. Examp!es o f c hem1stry bl an k va ues or various materia s. 10 Material origin Be/lBe 14 beryl ETH mineral collection 2 x 10- beryl ETH mineral collection 1 x 10·14 15 beryl ETH mineral collection 9 x 10· quartz crystal industrial 3 x 10·14 quartz crystal surface sample, 3 x 10·14 but brief exposure quartz from granite drill core, 11 .5 m deoth 2 x 10·14 1 auartz from granite drill core, 125.2 m depth 2 x 10·14 1 Merck carrier 3 x 10·14 Merck carrier 2 x 10·14 Merck carrier 1x10·14 data from Niedermayer (1995), prepared with the author at EAWAG ..

Chemistry blanks for 10Be Procedural blanks for 10Be were made from quartz, beryl crystals, and from commercial carrier material. The beryl samples were made from crushed single beryl crystals obtained from the ETH mineral collection. Blanks from both natural and industrial quartz crystals were also made. Additionally, quartz separated from whole rock by our standard techniques was used to make procedural 32 blanks. Such samples included surface samples assumed to have very short exposures as well as samples taken from deep drill cores (Niedermayer 1995). The measured 10BefBe ratios for the various blanks are shown in Table B2.1. All of the procedural blank ratios were, within the errors,indistinguishable from the Zurich AMS facilty's house blank (solid BeO from Merck) measured during the respective run 1 (generally 1-4 x 10- \ In that sense the above values should also be considered to be maximum values. Therefore, we could not detect any 10Be that might have been introduced during sample preparation from water, labware, etc. On the other hand, comparison of resu Its for quartz with carrier added, carrier alone and beryl seemed to confirm the idea that commercial carriers contain some 10Be (Middleton et al. 1984).

Boron 10B is an interfering isotope in AMS measurement of 10Be, and has to be kept as low as possible. Although various authors have emphasized fuming with perchloric acid right before the final precipitation to remove boron, we did not perform this procedure. Brown (1992) noted that boron is predominantly a contaminant in dust and tapwater, so rinsing all labware copiously (even disposable plasticware) with ultrapure water (18 MO) seemed to reduce the boron suffiecently. This was especially true with respect to the final Be(OH)2 precipitate, which must be rinsed at least once with 18 Mn water. We also noted that samples which had been precipitated but left standing in the test tube (uncentrifuged) for a long period of time (weeks to months) had very low B levels. Solubility constants for various borate/boric acid species (Smith and Martell

1976) show H3B03 is the least soluble B species. That means that traces of B in the solution will not lead to any B-containing solids. There is no boron solid that would precipitate at pH 8. Thus, any B in the sample precipitate must have just been carried down with the precipitating Be(OH)i. Such B could be eliminated by letting the solution stand, so the B can go back into the bulk of the solution, or with much rinsing and stirring or shaking. Finkel and Suter (1993) also recommended rinsing (with stirring) the precipitate several times. Boron is volatile at the high baking temperatures (850 °Cl which were used to convert the Be(OH)i to BeO. This was shown when samples we had produced (normally low B levels) were baked with samples of different origin, which had high B levels. At that point in time none of the sample crucibles was equipped with a quartz lid. As a result, our samples and even the blank carrier material (which contains no boron initially) had high B levels. Thereafter, we always covered the sample crucibles with lids during baking.

Sample reproducibility and meteoric 10Be A detailed discussion of the origin and processes by which meteoric 10Be can contaminate rock samples is found in Chapter 33

10 B3. It was very important that we could prove that all traces of meteoric Be were completely, consistently and routinely removed by our extraction protocol. This is 10 because atmospheric, and thus rainwater, concentrations of Be can be at least three 10 orders of magnitude greater than the in-situ produced Be. For samples with short exposure times, even as little as 1% contamination would yield spurious results. 10 To check that the meteoric Be had been removed Be targets were measured from several of the etching steps. These tests were done for several different rock types (granite, sandstone and quartzite). Results for two sandstones and a granite are shown 10 in Figure B2.4. We found that the meteoric Be is already completely removed by the first two HF steps (similar to Brown et al. 1991, 1995, Nishiizumi et al. 1991). But this can also be dependent on the grain size, shape and individual characteristics of the quartz grains. The series of HF etchings remove between 30-50 % of the outer portions 10 of the quartz grains. As another type of check for removal of meteoric Be, we processed a surface sample of quartz (from granite) collected as a loose clast from the Grimsel Pass region. This rock sample is believed to have no or only a very brief exposure. The results are shown in Table B2.1, the measured ratio for this sample was in the same range as other blanks processed. Although this rock had been exposed to meteoric 10Be it was removed by our cleaning procedure. 0 Meteoric ' Be may be difficult to remove from weathered non-quartz mineral grains. SEM (scanning electron microscope) photographs (Chapter B3) of unleached and leached quartz and pyroxene show that a surface cleaning of the latter may be especially difficult because of its cleavage. The cleavage surfaces open up pathways to 10 the interior of the grain. As the mineral grain weathers, meteoric Be can become locked inside the grain as secondary minerals form. Quartz lacking cleavage and being very resistant to weathering remains a closed system, not allowing contamination by meteoric 10Be. 10 The many tests we performed to check for the removal of meteoric Be also afforded us a check on sample preparation reproducibility. As shown in Figure B2.4, results from leach steps HF2 and HF3 for sandstone 218 agree remarkably well. For sandstone 213a, we prepared aliquots from HF2 and HF4 which also agree within the stated errors. For granite sample 217b, aliquots from dissolution steps HF5, HF6, HF7, and HFB were prepared. The standard deviation from the mean of these four samples was 5 %. Additionally, a piece of 217b was completely remade from the rock crushing 10 stage onward (the sample was made after 5 HF steps). The Be results for this sample (shown as the small shaded band in Fig. B2.4) also agree very well with those of the other four targets. 34

0 20 40 60 80 100 % dissolved

Figure B2.4. For sandstone 213a, the plotted results show the removal 10 of meteoric Be. The agreement of the results from different dissolution steps for each rock also provides a check on sample reproducibility. 35

Another form of reproducibility check is the comparison of the various isotopes 26 measured from a single rock sample. Because of the agreement of Al concentrations and 10Be concentrations measured in the same rock (see Table C2.1 ), we also have 36 confidence that 26AI can be extracted reproducibly. Finally, a comparison of CI and 10Be measured in the same rock sample from the Julier Pass moraine indicates that 36 36 although we have less experience with C! sample preparation, the CI results do agree well with the 10Be results (see Chapter 02). ICP-AES measurement of total Al Quartz contains on the order of 50-500 ppm Al, depending on the nature of the quartz (inclusions, etc.). For large quartz sample dissolutions (ca. 50 g) no Al carrier was added. Because the final exposure age is derived from the measured 26Al/27 Al ratios, the inherent Al concentration must be determined not only very precisely but very accurately. We measured the Al concentration in an aliquot of the solution of the dissolved quartz (see section B2.3.1 ). This avoids serious problems inherent with measuring on merely an aliquot of the mineral separate (e.g. sample inhomogeneity). We measured the Al concentration using ICP-AES (inductively-coupled plasma atomic- emission spectrometry) at EAWAG. Problems with apparent instability of the ICP-AES were later found to be attributable to the background corrections made by the instrument's data reduction program. Dependence on background correction was avoided by using three standard additions for each sample aliquot, which enabled us to use the raw counts from the machine. The problems with background correction arose because even though the standard solutions can be matrix matched, e.g. by adding the appropriate concentrations of acids, every sample has a different matrix of additional cations. The principle of standard additions is to add known amounts of Al to several aliquots of the sample material. We performed the standard additions as follows: 0.5 ml of the aliquot of sample solution were pipetted into four test tubes; an additional 2, 4, and 6 ppm Ai were pipetted into three of the tubes, respectively. The tubes were filled up to 5 ml. The standard solution is then truly matrix-matched because the standard curve is measured in the sample solution. Thus a calibration curve is calculated individually for every single sample. As an additional step to increase precision, we used two separate wavelengths for the Al determination. The outcome of this was that as the 396.15 line is very close to a Ca peak (e.g. Lichte et al. 1980), the 308.22 line was generally the one we used. But this depended on the Ca content of the sample solution. 36

< 0.4 mm crushed, rinsed whole rock

add AgN03 to ppt AgCI let stand > 24 hrs.

centrifuge, ecant HF

redissolve AgCI with NH40H

add BaN03 to ppt BaS04 I I let stand> 24 hrs.


ppt AgCI with HN03

let stand > 24 hrs.

centrifuge, ecant

dry at 80 °C PRESS'-AMS

Figure B2.5. Steps for the extraction of 36Ci from a crushed whole rock sample. 37

82.3.2 Extraction of Cl from whole rock Sample preparation for 36CI followed the method of Zreda (1994) very closely. In Figure 82.5, the various steps of the sample preparation are shown. From the same 10 26 batch of crushed rock as was used for the Be and AI extractions, we took the< 0.4 (or < 0.5) mm size fraction for 36CI sample preparation. It is important to note that exactly the same crushed rock sample was used for all three isotopes. The first step was to rinse the crushed whole rock to remove any possible 36 meteoric CI. The crushed whole rock was left overnight in a solution of weak HN03, then rinsed at least five times with 18 Mil water. Some of the day-size fraction was washed away and discarded in this step. The rinsed crushed rock was dried in a glass beaker in an oven at ca. 80 °C, then stored in plastic bags. From this rinsed material 10-15 grams were split and sent for elemental analysis (XRAl Ontario, Canada). Samples (50-1 00 g) of dry whole rock were placed in 2 l teflon bottles with

concentrated HN03 and HF. As per Zreda (1994), for each gram of rock, 2.5 ml of HF

and 1 ml of HN03 were added. No Cl carrier was added. The bottles were kept in a 90 °C water bath until the rock had dissolved completely, which took several days. Usually some material was present at the bottom of the bottles. This material was white and thought to be a new precipitate (reprecipitated silica gel ?). There was never any dark material which might have meant there was undissolved rock remaining. X- ray diffraction analysis of similar material indicated it was made up predominantly of precipitated fluoride salts (F. Phillips, personal communication 1995). The contents of the 2 l bottle were then transferred to several 250 ml teflon centrifuge bottles. These bottles were centrifuged cold (l 0 °C) in order to separate the precipitated material from the HF solution. The solution was decanted to 50 ml teflon centrifuge tubes, to which 2 ml of 0.1 M AgN03 solution was added to precipitate AgCI. The solutions were allowed to stand at least 24 hours to allow nucleation of the AgCI. Then, the tubes were centrifuged (30 minutes at 7500 rpm). The supernatant was decanted, the last drops were removed with a rinsed disposable plastic pipette, so that no Agel would be lost. Water (18 MQ) was added to each tube and they were shaken to rinse the AgCI. The tubes were once again centrifuged, the liquid was decanted in the same manner, and all the AgCI was transferred to a single disposable 10 ml tube, which then was filled with water. This tube was centrifuged and the supernatant decanted.

In the next step, the AgCI was redissolved with NH 40H and unwanted hydroxides were precipitated. Usually, the Agel at the bottom of the tube had to be mechanically broken up with a plastic stirrer to get it to dissolve. The pH was checked to make sure that the Agel was in solution and Ca or Mg hydroxides had precipitated. If there appeared to be a lot of hydroxide precipitate, this was separated by 38 centrifuging. At this point, the sample is in solution, so the precipitate is discarded but the solution is retained. Because 36S interferes with the AMS measurement of 36CI, S must be separated from the sample material during the chemical extraction. In order to do this, we added

2 ml of saturated Ba(N03)z solution and the sample solution was left to stand for at least 24 hours. The BaS04 precipitated as white, sugary crystals along the walls of the test tube. The contents of the tubes were centrifuged and then the BaS04 was removed by filtering (45 µ). Phase separation by filtering seems to be a good way to separate the

BaS04, since the S levels of our samples were generally low. Separation of BaS04 by centrifuging and decanting may allow tiny BaS04 particles to remain with the sample material during decanting. The pH of the filtered solution is then lowered by adding several ml of concentrated HN03 until a milky cloud of AgCI appears. After centrifuging and rinsing two times, the AgCI is dried in the tube in an oven at 80 °C bevore pressing. The sample material must be kept out of light to avoid photolytic decomposition of AgCI. After drying, the AgCI was broken into small pieces with a small stainless steel spoon and worked into the hole of a tantalum target holder with a needle so that the hole is completely filled. The target was then pressed at 5 bars. Another piece of AgCI was put into the hole, pressed again, and so forth, until the AgCI was used up. As listed in Appendix Hl, several more 36CI extractions were completed than yielded results. The most common problem was there not being enough AgCI to be pressed into a target holder. In such cases, the total Cl content had not been measured prior to sample preparation. That means that the weight of rock needed to get enough AgCI had just been estimated (in these cases underestimated). In general there were no problems noted with the extraction procedure described by Zreda (1994).

Elemental Analysis As discussed in chapter B1, elemental analyses are needed for each rock analyzed for 36CI. The production rate for each rock must be individually calculated. Major elements, Gd, B, and total Cl were measured using XRF, prompt gamma neutron activation and an ion selective electrode, respectively, by XRAl (Ontario, Canada). These elements were all measured on the rinsed crushed rock.

82.4 AMS MEASUREMENT The concentrations of 10Be, 26AI and 36CI were determined using AMS at the ETH/PSI tandem facility at ETH-Zurich, Honggerberg (Finkel and Suter 1993 and references therein). The following discussion is partially specific to only this facility. Some details on the Zurich AMS facility are listed in Table B2.2. 39

10 26 36 Table B2 ..2 AMS characteristics for measurement of Be AI and CI ' ioBe 26AI 36CI 10 26 36 ratio Be/'Be Al/27 Al Cl/CI 1 stable isotope 1-2 ppm in granite 20-500 ppm in 50-500 ppm in concentration quartz granite or basalt

back2round ratio <10·14 <10-14 <10"15 tar2et material BeO Al203 AgCJ extracted BeO- Al- Cl- ion source 0.5-1 0.05-0.2 10-20 current (uA) orecision 3-5% 3-10% 1-5% interferin2 isobar JOB 36s 1 9 for quartz Be carrier is added.

A typical BeO target from 50 grams of quartz exposed 10,000 years contains 6 7 10 only 10 - 10 atoms of Be. Because of the extreme sensitivity and low background of AMS such a small sample can be used and measured within about one hour. The half- 10 life of Be is so long that it would take decades to measure such a sample by decay counting. If one would count however only for months, then several tens of kilograms of quartz would be needed. On the other hand, a conventional mass spectrometer would have to be run at very high mass resolution to separate isobaric ions and molecules, therefore reducing the efficiency to such a degree that it could not be used for the low isotopic ratios we are dealing with. The advantage of an accelerator mass spectrometer over a conventional mass spectrometer is that the ions can be accelerated to such high energies that even while attaining high transmissions all ions can unambiguously be identified through nuclear detection techniques. The overall 3 4 efficiency of the AMS system is on the order of 10" to 10" . In brief, the tandem accelerator mass spectrometer is made up of three sections, the low energy section including the ion source, where negative ions are created, the tandem accelerator itself, where the ions are accelerated to high energies (MeV instead of keV in the case of a conventional mass spectrometer), and the high energy section, where the positive ions are analyzed and led to the detection system. In the ion source the target material is sputtered off the sample surface (e.g. BeO) by a stream of cs• ions. Negatively-charged ions are extracted from the sample material and mass analyzed. Upon entering the tandem, the negative ions are accelerated to the positive potential in the center of the tandem. There they pass through a thin foil or some gas where they 40 are stripped of several of their electrons and become positively charged ions. At this point, highly positive-charged molecules are unstable and break up into their constituents. All positive ions are then accelerated a second time and injected into the mass spectrometer on the high energy side of the accelerator. There the unwanted molecular breakups can easily be filtered out. Isobaric atomic interferences such as 10s in the case of 10Be, and 36S in the case of 36CI, can be unambiguously removed and/or identified in the detection system, which is based on the nuclear properties of high energy ions. The ETH/PSI AMS system does not measure absolute isotopic ratios. Targets made of standard material are run frequently together with the samples (unknowns). The measured ratios of the unkowns are normalized to the standard. For 10Be, the 12 standard S555 is used (1°Be J9Be S555=95.5 x 10- ), which is a secondary standard to the materials used to determine the half-life of 10Be (Hoffmann et al. 1987). The 26 26 12 standard material for AI is ZAL94 with Al;27 Al 526 x10- , a secondary standard based on the reference material of Sarafin (1985). For 36CI, the standard material K380/4 (PSI, Villigen, Switzerland} is used with 36Cl/CI 15.24 x 10·12 (Synal et al. 1994). The values of the appropriate blanks have been subtracted. The final quoted error (1 cr) is a combination of counting statistics, reproducibility derived from the standard measurements and repeated measurements of the samples, the chemistry blanks. This is comibined with a 5 % sample reproducibility (see section B2. 3.1.2), which is considered to be an upper limit. 41


S. Ivy-Ochs lnstitut fur Teilchenphysik, lngenieurgeologie, ETH-Honggerberg, CH-8093 Zurich, Switzerland

P.W. Kubik Paul Scherrer lnstitut c/o lnstitut f(.ir Teilchenphysik, ETH Honggerberg, CH-8093 Zurich, Switzerland

J. Masarik Department of Nuclear Physics, Komensky University, SK-84215 Bratislava, Slovakia l. Bruno lnstitut fur lsotopengeochemie, ETH-Zurich, CH-8092 Zurich, Switzerland

J. Beer EAWAG, CH-8600 DObendorf, Switzerland

C. Schliichter Geologisches lnstitut, Universitat Bern, CH-3012 Bern, Switzerland

1 submitted to Earth and Planetary Science Letters. 42

83.1 ABSTRACT We report results from two separate sequential dissolution series of separated from dolerite clasts collected from the Sirius Group sediments at Table Mountain, Antarctica. Although quartz is a ubiquitous mineral additional geological 10 problems could be addressed with a broader spectrum of utilizable minerals for Be 26 3 21 and AI exposure dating studies. In the case of pyroxene, He and Ne are quantitatively retained and exposure ages can be calculated. Therefore, measurement of 10 Be as well in the same mineral would complement the noble gas results by revealing possible episodes of prior exposure. 10 In the first dissolution series, the final Be concentration was 26 % greater than 10 3 the expected Be concentration, which was calculated using the measured He and 21 Ne exposure age (2.4 Ma}. In the next dissolution series, on a different pyroxene separate, we completed seven different dissolution steps. Even after etching away 49 %, 10 10 whereby the Be concentration appeared to have leveled off. The final Be concentration was greater than the saturation value (for that elevation) and was therefore significantly greater than the concentration expected after only 1.8 Ma of exposure 3 21 (based on He and Ne concentrations}. We compare the results from the pyroxene 10 tests with sequential dissolution series on quartz separates. In quartz, the Be concentration levels off to a pure in-situ component after only one HF leaching; this fact 26 10 can also be verified by Al measurements. A possible explanation is that meteoric Be is found within the secondary minerals inside of the pyroxene grains. Thus it cannot easily be removed, even with a rigorous cleaning procedure. We do not wish to imply that quartz is the only mineral which can be safely used 10 for Be exposure dating studies. Unaltered pyroxene, especially from younger rocks, 10 may be suitable for exposure studies. On the other hand Be exposure ages measured on minerals other than quartz, especially from old or weathered rocks, must be viewed with caution until they have been verified by measurement of another cosmogenic isotope.

83.2 INTRODUCTION Cosmogenic isotopes are produced in measurable amounts in situ in the upper decimeters of rocks exposed to cosmic rays. Measurement of the concentration of these isotopes allows the calculation of the amount of time elapsed since formation or uncovering of the rock surface (Lal 1991 }. This is related by:

N =-p-(1-e-(A.+pE/A)T) (9) A.+ pe A 43

1 1 where N is the number of atoms·g· , Pis the production rate (atoms·g- .yr·\ Tis the length of time the surface has been exposed (yr), A, is the decay constant (yr"\ pis the rock density (g•cm·\ e is the erosion rate (cm.yr·\ and A is the cosmic ray 2 attenuation length (g•cm- ). 10 26 The mineral most often used for studies with Be and Al is quartz (Lal and 10 26 Arnold 1985). Successful results using Be and Al in quartz have shown the versatility and broad applicability to geomorphological studies that are possible using this mineral (e.g. Nishiizumi et al. 1993). We report here the results from attempts to use the mineral pyroxene for exposure dating studies. Interest in measuring exposure ages in pyroxenes has been sparked by the controversy over the age of the Sirius Group sediments in Antarctica (Kennett and Hodell 1995, and references therein). Quartz-bearing clasts are rare in Sirius Group sediments while dolerite clasts (plagioclase and pyroxene) are common. This has forced the use of 10 minerals other than quartz for Be exposure dating, for example plagioclase from a dolerite clast {Graham and Harwood 1995). Determining the age of these high elevation glacial deposits is the key to solving the controversy over a stable (Denton et al. 1993) or dynamic East Antarctic Ice Sheet (Webb and Harwood 1991 ). Results from Cerling (1990), Poreda and Cerling (1992) (summarized in Cerling 3 21 and Craig 1994a) and Bruno (1995) indicate that He and Ne are quantitatively retained in pyroxene. Stable cosmogenic isotopes accumulate over all earlier episodes of 10 3 21 exposure. As Be (t112=1.5 Mal decays, while He and Ne do not, periods of earlier 21 exposure followed by periods of burial would show up as an older apparent Ne, 3 10 respectively He, exposure age in relation to the se age. The use of stable and radionuclides in concert can often provide the most information. 10 We present Be measurements from sequential leaching tests of two different pyroxene mineral separates. For comparison, we show leaching results from three quartz samples, two from sandstones and one from granite. The objective of these tests was to determine what percent of the pyroxene (by weight) must be etched away to remove all 10 10 contaminating meteoric Be. The atmospheric Be production rate is around three orders of magnitude per unit mass higher than in a rock at sea level (Lal 1991) meaning very small amounts of contamination can yield spurious results.

Interpretation of the exposure age data and the implications for both the age of the Sirius Group sediments and the stability of the East Antarctic Ice Sheet are discussed in detail in Ivy-Ochs et al. (1995) and Bruno (1995). 44

83.3 SAMPLES AND ANALYTICAL PROCEDURES 83.3.1 Sample information The samples investigated are part of a project to determine the age of the Sirius Group sediments and related landscape elements in the Dry Valleys region of Antarctica, 3 10 21 26 using in situ produced cosmogenic He, Be, Ne and AI (Bruno 1995, Ivy-Ochs et al. 1995). All of the samples discussed below are from Table Mountain, details of sample 213a, 218 and 217b locations are shown in Ivy-Ochs et al. (1995) and in Bruno {1995) for samples 4 and Sa. The latter were chipped from the top surface of two dolerite boulders found embedded in the Sirius Group sediments.

10 83.3.2 8e sample preparation and measurement 10 The procedure we used for Be sample preparation generally follows that of Brown et al. (1991) and Kohl and Nishiizumi (1992), which are for quartz. In the case of the pyroxenes (which contain at least Al, Fe, Mg and Ca, in addition to Sil, minor modifications were necessary because it is more difficult to isolate Be in a solution from dissolved pyroxene than quartz (Ochs and Ivy-Ochs 1996). From sandstone (samples 213a and 218) or granite (217b) selective chemical dissolution with a weak HF solution was used to obtain a pure quartz fraction (cf. Kohl and Nishiizumi 1992). This involved an HCI leach step, followed by several dilute HF steps. Thereafter, the sample preparation procedure generally followed that as described below for the pyroxenes.

Table B3 .1 . Major element concentrations in we1m. h t oercent oxt'd e. Element 4 Sa

Si02 51.84 50.38 Al,01 4.54 5.99 Fe201 16.4 17.87 MnO 0.29 0.3 MgO 14.14 13.53 Cao 10.99 10.27 Na20 0.48 0.59 K10 0.18 0.27 Ti02 0.63 0.78 P,o, 0.13 0.12

In order to obtain monomineralic grains, the 0.2-0.5 mm fraction was chosen for the dolerites, based on the grain size observed in thin section. For sample 4, the 4S pyroxene grains were hand-picked under a binocular microscope. For Sa, we used a heavy liquid (methylene iodide) and a Frantz magnetic separator to separate the pyroxene from the plagioclase. This purified pyroxene fraction was then examined under the binocular microscope and remaining grains with adhering fragments of other minerals than pyroxene were removed. The pyroxenes we used were augitic, major element compositions (Table B3. 1) were measured by ICP-AES (Bruno 199S). The weighed mineral separates were placed in 60 ml closed teflon jars and put on a shaker table with approximately 40 ml 18 Mn water, S ml HN03 and S ml 48% HF. After several hours, the HF solution was decanted into a teflon beaker and the remaining mineral grains were rinsed several times with 18 Mn water, the rinse water being added to the teflon beaker containing the decanted sample solution. The material left in the teflon jar was put on a shaker table overnight in 18 Mn water. In the morning, this water was also added to the beaker containing the sample solution, and the sample was rinsed again, again adding the rinse water to the sample solution. The teflon jar with the undissolved portion was dried and weighed to record the amount dissolved in that step. 9 Be carrier was immediately added to the decanted solution. After fuming off SiF4, Be and Al were separated on a cation exchange column. Excess Fe was separated from the Be fraction by using a methylisobutylketone extraction (Knauer 1994) or an anion exchange column. When Al was present in the Be fraction, the complete column procedure was repeated. Finally, the hydroxides were precipitated and then converted to 10 the oxides at 8SO °C Be/Be ratios were measured at the PSI/ETH tandem AMS facility. The errors listed in Table B3.2 include the statistical (counting) error, the error due to the normalization to the standards and blanks, and the reproducibility of several measurements of an individual sample (S %).

83.3.3 3He and 21 Ne measurement in pyroxene 3 21 Details of the He and Ne measurement procedure and data analysis can be found in Bruno (199S). The noble gas concentrations in Table B3.3 are corrected for 21 atmospheric Ne and nucleogenic He and Ne. Atmospheric Ne was determined by 20 assuming that all measured Ne is atmospheric. The respective correction is less than 3 S.0% for sample 4 and 13.4% for sample Sa. Nucleogenic He was determined via the 4 measured He concentration, which was assumed to be completely radiogenic (and quantitatively retained). In all cases, nucleogenic contributions were less than 1.1 %. The 3 samples should not contain any inherited mantle He since the trapped Ne is atmospheric, i.e. no traces of a MORB Ne component are visible. The concentrations of the cosmogenic noble gases are accurate to within S%. This number includes the reproducibility of the measurements, the uncertainty of the standard gas amounts and the corrections for non-cosmogenic contributions. 46

83.4 PRODUCTION RATES 10 The production rate of Be in pyroxene irradiated at Earth's surface by galactic- cosmic-ray (GCR) particles is calculated using LCS, the Los Alamos LAHET Code System (Prael and Lichtenstein 1989), which is a general-purpose, Monte Carlo computer code that treats the relevant physical processes of particle production and transport. LCS, its tests, the basics of its physical model, and its adaptation to the production of cosmic-ray- produced (cosmogenic) nuclides in meteorites (Masarik and Reedy 1994) and Earth's surface and atmosphere (Masarik and Reedy 1995) are described in detail elsewhere.

Using calculated fluxes of particles in the Earth's surface and measured and evaluated cross sections, we calculated production rates of cosmogenic nuclides. In pyroxene, 10Be is produced from 0, Mg, Al, Si, and Fe. For these elements, absolute 10 elemental production rates for Be were calculated:

P( 10Be) = 10.87[0] + 0.39[Si] + 0.45[AI] + 0.52[Mg] + 0.16[Fe] (10) where the weight percent of the target element is contained in brackets (Masarik and 10 Reedy 1996). These Be production rates have uncertainties estimated at -20%. The production rates calculated for our two pyroxene samples are shown in Table B3.3. In 10 comparison, empirically determined production rates for quartz are 6 atoms Be per gram Si02 per year (Nishiizumi et al. 1989). In order to calculate the elemental contributions from Mg and Al to the production of 21 Ne, we used the production ratios of Masarik and Reedy (1996) combined with the 21 measured production rate for Ne from Si02 (Niedermann et al. 1993), i.e. 21 21 21 21 P NeMg:P Nes; =3.32 and P NeA1:P Nes; =1.64. The resulting contributions are 149.4 3 3 for Mg, 73.9 for Al and 45 for Si. The He production rate we used is 115 atoms He per gram pyroxene per year (Cerling and Craig 1994a). The production rates were then scaled to the altitudes of our samples (Lal 1991 ).

83.5 RESULTS AND DISCUSSIOI~ 10 Table B3.2 contains the measured Be atoms per gram of quartz or pyroxene for each step in the five different dissolution series we performed. Four of these data points were previously reported in Ivy-Ochs et al. (1995) (as indicated in Table B3.2). Exposure ages (Table B3.3) have been calculated by solving Eq. 9 for T and assuming no erosion. In Figure B3.1, the ' 0 Be concentrations are plotted after normalization to sea level. All concentrations are for pure mineral separates (either pyroxene or quartz). Table B3.2. Sample weig ht ht at was d isso Ive d in each step an d t h e respective AMS-measured 10 Be concentrations. 10 1 Sample Mineral Altitude (m) Dissolution Weight (grams) Dissolved (%) Cumulative (% Be atoms/gram Error (%) No. Step 4 pyroxene 2030 HCI 0.0478 1.0 1.0 1.54 x 109 6.0 HF1 1.9248 82.2 83.2 1.67 x 108 7.8 HF2 0.3697 16.8 100 5.30 x 107 10.9 Sa pyroxene 2030 HF2 0.6905 14.2 25.2 1.48 x 108 5.5 HF3 0.2691 5.5 30.7 1.29 x 108 7.1 HF4 0.4376 9.4 40.1 1.50 x i08 7.6 HFS 0.4301 8.9 49.0 0.85 x 108 7.8 HF6 2.4826 51.0 100 1.18 x 108 6.1 4 217b ouartz 1820 HF5 1.3971 8.6 31.1 3.52 x 107 9.2 HF6 2.9749 18.4 49.5 3.77 x 107 5.6 HF 7 2.9935 18.4 67.9 4.03 x 107 5.6 HF8 5.1831 32.1 100 3.61 x 107 6.1 5 213a ouartz 2080 HCI 3 1.0 1.0 8.00x 108 HF 2 4.7063 14.2 57.6 5.57 x 107 6.5 HF4 11.1623 33.0 100 5.70 x 107 l 5.9 5 218 Quartz 2170 HF2 1.5461 5.8 15.6 3.37 x 107 2 10.3 HF 3 3.1343 11.6 27.2 3.39 x 107 2 7.8 1 Errors are at the 1cr level, mcludmg the stat1St1cal (counting) error, the error due to the normalization to the standards and blanks, and the reproduc1b1hty.. of several measurements of the same sample. 2 Data from Ivy-Ochs et al. (1995). 3 This 10Se fraction was assumed to have been leached from the surfaces of all the grains, see text. 4 Quartz from granite. 5 Quartz from sandstone. 48

83.5.1 Quartz 10 8 A Be concentration of 8.0 x 10 atoms/gram was obtained for the HCI leach of quartz sample 213a (calculated using the total initial sample weight). This is at least an 10 order of magnitude greater than the Be measured in HF2 and HF4. In sandstone 218, 26 10 we measured AI as well, and the age (1.3 Ma) agrees well with the Be age (1.4 Ma), 10 verifying that the Be concentration reached is the in situ component with no remaining 10 meteoric contamination. The Be concentrations for steps HFS, HF6, HF7 and HF8 for quartz from granite sample 21 ?bare the same (within the stated errors) and an exposure 26 age of 2.6 Ma was calculated. The agreement with the AI results (2.7 Ma) shows that 10 these procedures have indeed cleaned away the meteoric Be. Such dissolution series 10 experiments support the idea that the majority of the meteoric Be is weakly adsorbed on the surfaces of the quartz grains (Brown et al. 1991, 1995, Kohl and Nishiizumi 1992). The resistance of quart2 to weathering (i.e. no secondary mineral formation 10 within the grains) prevents the uptake of meteoric Be into the quartz crystal.

10 3 21 Table B3.3. Measured Be, He and Ne concentrations and calculated ex oosure ages. Sa mole 4 Sa 10 6 Be (10 atoms/g) 53±5.8 118±8.3 10 1 Be P.R. (atoms/g yr) 5.2 5.1 Age (Ma) (1°Be) 4.1 2 3 6 3 3 He (1 0 atoms/1~) 1522±75 1185±45 3 4 He P.R. (atoms/g yr) 115 115 21 6 3 3 Ne (10 atoms/g) 357±18 273±11 21 5 Ne P.R. (atoms/g yr) 25.4 25.3 21 Age (Ma) (3He, Ne) 2.4 1.8

10 Be production rate (P.R.) calculated using the formula for elemental contributions gi1ven in the text. The measured concentration was higher than the secular equilibrium concentration thus an exposure age could not be calculated. Data from Bruno (1995). 4 Cerling and Craig (1994a). 21 Ne P.R. calculated with the following elemental production rates Mg: 149, Al: 74, Si: 45 (see text). 49

o 213a D 217b /:; 218

7 • 4 3 10 • Sa

. . .'. ff .·.

2 107

E f!! Ol (;s + E 0 (ii Q) al ;: 1 07 9 106 8 106 7 106 6 106

5 106

4 106 0 20 40 60 80 100

% Dissolved

Fig. B3. l. 10Be concentrations normalized to sea level. Filled symbols are for pyroxene, open symbols are for quartz. The shaded band indicates the secular equilibrium concentration of 10Be for pyroxene. so

83.5.2 Pyroxene We performed two separate dissolution series on pyroxenes, from two dolerite boulders (4 and Sa) with different exposure ages. The eight measured data points are 10 9 listed in Table B3.2. For sample 4, the Be concentration in the HCI leach was 1.S x 10 10 atoms Be/gram pyroxene. This is two orders of magnitude greater than the 7 10 concentration measured in the last dissolved fraction (S.3 x 10 atoms Be/gram 10 pyroxene). From the Be concentration in the final step, an exposure age of 4.1 Ma was 3 21 calculated. The exposure age determined by He and Ne is 2.4 Ma. The measured 10 Be concentration differs from the expected (using an exposure age of 2.4 Ma) 7 10 concentration (4.2 x 10 atoms Be/gram pyroxene) by 26 %. 10 For pyroxene sample Sa, we measured the Be concentrations in HF2, HF3, HF4, HFS and HF6. Prior to the final step almost half of the original sample weight had 10 already been dissolved away (49 %). At this point, the Be concentration appears to have for the most part leveled off. There was no point in performing more dissolution steps after HF6 because the remaining pyroxene grains were completely bleached and 10 could easily be mechanically destroyed. The final concentration of Be in pyroxene Sa 7 10 is roughly twice the saturation concentration for this elevation (6.7 x 10 atom Be/gram pyroxene). Therefore, an exposure age could not be calculated. Clearly, the measured concentration is significantly greater than the expected concentration based on the 1.8 3 21 Ma He and Ne exposure age. 10 The disagreement of the noble gas and Be results may have three causes: 1) the production rates are not correct, 2) the noble gases may have been lost through diffusion 10 and 3) there is excess Be. The fact that there is less of the noble gases with respect to 10 Be means that prior exposure can be ruled out as the cause of the disagreement of the 3 21 ages. The consistent internal agreement of He and Ne ages from many pyroxene 10 26 21 samples (Bruno 199S) and Be, AI and Ne ages from several quartz samples (Bruno 199S, Ivy-Ochs et al. 199S) indicates that the production rates we have used are 3 21 appropriate. We also interpret the consistent agreement of He and Ne ages in a variety of pyroxene samples (Bruno 199S) as ruling out significant diffusional loss of these isotopes from the pyroxenes. 10 The only remaining explanation is the presence of excess Be in the sample Sa 10 pyroxenes, and probably in sample 4 as well. Radiogenic production of Be (Sharma and Middleton 1989) cannot be called upon as these rocks do not have abnormally high U, Th or Li concentration. Therefore, we suggest that we were unable to remove all of 10 the meteoric Be from the pyroxene grains. This may also be an explanation for the 10 scatter in the data, indicating that more or less of the meteoric Be was removed inconsistently in different cleaning steps. 51

10 We suggest that meteoric Be derived from atmospheric production has probably been introduced into the pyroxene grains during the processes of weathering and alteration. Weathering and dissolution studies of pyroxenes in both the laboratory and in naturally weathered pyroxenes (Berner et al 1980, Schott and Berner 1982, 1985, Eggleton 1986, Colin et al. 1993) indicate that pyroxene does not weather uniformly just on the mineral surface. Conversely, etch pits concentrate along crystallographically weak zones and reach into the grain cores. As the pyroxene dissolves, clays and Fe hydroxides form along these zones, taking up structural elements from the dissolved pyroxene and from the surrounding aqueous medium (Eggleton 1986, Casey et al. 1993, Banfield et al. 1995). In the clay structure, Be may substitute for Al and for some divalent cations (Kabata-Pendias and Pendias 1992). of Be within the structure of secondary minerals begins as it is adsorbed on the highly reactive surfaces of clays and hydroxides or their precursors, and is subsequently incorporated as the minerals grow (Kusakabe and Ku 1984). 10 The amount of meteoric Be actually available to the forming clays may be inferred from the concentrations we have measured in the surface HCI leach fractions 8 9 10 10 (ca. 10 -10 atoms Be/gram) which are in the same range as Be concentrations 8 10 10 measured in secondary minerals (10 to 10 atoms Be/gram bulk sediment) (Monaghan et al. 1983, Pavich et al. 1986, Valette-Silver et al. 1987, Bourles et al. 1992). This is at least an order of magnitude greater than the expected in situ component 7 10 in our pyroxenes (ca. 10 atoms Be/gram). Although Antarctica is a polar desert, thin films of water are present and are likely enough to weather the primary ninerals over the course of millions of years (Ugolini and Anderson 1973, Ugolini 1986, and Claridge 1987). For example, Campbell and Claridge (1987) noted Mg-rich smectites as weathering products of dolerites in 10 Antarctica. Notably, the concentration of Be in precipitation in Antarctica is especially high because of the infrequency of precipitation events (Lal and Peters 1967). The two rocks we have sampled have been subject to weathering at least as long as they have been on the surface of the Sirius Group deposits at Table Mountain, i.e. more than 3 million years (Bruno 1995, Ivy-Ochs et al. 1995). Both of the dolerite boulders, 4 and Sa, exhibit a thin rust-colored rind clearly indicating oxidation of the primary minerals. In thin section, individual pyroxene grains exhibit patchy alteration and replacement by clays and iron hydroxides along cleavage fractures as well as in the grain cores. Figures BJ.2a and B3.2b are scanning electron micrographs of etched quartz and etched pyroxene, both after HF2. Individual quartz grains easily survive the many HF washings. As shown in Figure 2a, it is difficult say that the grain has even been acid leached. In stark contrast, pyro;

After several steps individual grains were not being cleaned on the outside but were being destroyed and broken into smaller grains. In summary, the problem is twofold. First, weathering of the pyroxenes may result 10 in precipitation of meteoric Be-containing clay and/or hydroxide minerals deep inside of the pyroxene grains. In each grain, the weathered zones are difficult to separate from primary unaltered pyroxene. Second, the pyroxene grains are not robust enough to make it through the numerous HF steps during which the contamination might be removed. 10 No Be exposure dating studies have been attempted as yet with pyroxene. 10 26 Nishiizumi et al. (1990) reported Be and Al exposure ages and calculated erosion rates from olivine separated from a SOO ka lava flow from Hawaii. In their study they removed 11 % of the outer surface of the grains by HF etching. That their samples were 10 3 21 free of meteoric Be was confirmed by agreement with He and Ne measurements (Marti and Craig 1987, Cerling and Craig 1994b). Shepard et al. (1995) determined the 10 36 36 age of a lava flow in Nevada (ca. 4S ka) using both Be and CI, where CI was 10 measured on whole rock and Be on the olivine separates. In both of these examples, 10 the Be results were confirmed by measurement of at least one other cosmogenic 10 isotope. Difficulty in removing meteoric Be from olivine was encountered by Seidl (1993) in her multi-isotope study of a 5.2 Ma lava flow (Hawaii). In all three of the aforementioned studies the HF leaching technique developed for quartz and applied 36 here was used. Zreda (1994) noted contamination by meteoric CI in olivine grains altered to iddingsite. Olivine yields consistent results when the rock is not very old and the grains themselves are relatively fresh. It is probable that unweathered pyroxene in 10 younger rocks may actually be suitable for Be studies. Because of the importance of enlarging the cast of minerals that can be used for 10 Be exposure dating studies, continued tests are encouraged, despite the negative results we present here as an undeniable warning. In that light we are continuing such tests using different leaching techniques.

83.6 CONCLUSIONS The two dissolution series we performed indicate that meteoric 1OBe may be 10 difficult to separate in even weakly altered mineral grains. Clearly if meteoric Be were only adsorbed on the surface of the pyroxene grains, it should have been removed by our sequence of seven etching steps. One possible explanation is that crystallographically controlled dissolution during weathering of the pyroxene 10 accompanied by clay formation allows meteoric Be to penetrate and to be locked in 10 the interior of the grains. This is in contrast to quartz, where all of the meteoric Be is 10 only adsorbed on the surface. It may be that, similar to olivine, Be exposure ages can be measured in pyroxenes from young rocks where the grains are not very altered. First it 53 must be shown that they can be cleaned, which can only be proven by measurement of another isotope. Indeed leveling off in a dissolution series does not prove the absence of 10 meteoric Be. We emphasize again the multi-nuclide approach to check for gain or loss of one of the isotopes or to check for prior exposure. Very careful interisotope 10 comparisons are necessary before drawing conclusions from Be measured in minerals that are less resistant to weathering than quartz, such as pyroxene or plagioclase. 54

Figure B3.2a. SEM photographs of unleached and leached quartz grains.

Figure B3.2b. SEM photographs of unleached and leached pyroxene grains. 55



Cl.1 INTRODUCTION The history, and thus, in a sense, the past and present 'health' of the Antarctic Ice Sheets (Fig. Cl .1) are important parameters in predicting possible greenhouse scenarios. The West Antarctic Ice Sheet is clearly the more precarious of the two ice sheets because it is marine based (Rott et al. 1996), but the East Antarctic Ice Sheet locks up the greatest volume of water on earth (Thiel 1962, Drewry 1983) and thus has a critical position with a view towards possible global warming and rising sea level. The Neogene basal tills of the Sirius Group sediments are the only remaining evidence and terrestrial record of the pre-Quaternary character of the ice of East Antarctica. Although we are beginning to understand where a certain glacier was and the nature of the glacier, the key question remains, and that is when. The Sirius Group has so far eluded direct dating. For example, key ash layers that could provide at least minimum ages have not been reported overlying Sirius Group sediments. The measurement of the exposure age of surface clasts on the Sirius Group deposits, the goal of this investigation, is such an attempt to directly determine the age of the Sirius.

Cl .2 EVOLUTION OF THE LANDSCAPE AND CLIMATE Of ANTARCTICA The making of the Dry Valleys landscape can be attributed to the action of streams, glaciers, and wind, with the climate as the dominant ruling force. Investigation into the evolution of the Antarctic climate, the reconstruction of which process dominated when, brings together such diverse research areas as paleoceanography, in viewing the record contained in ocean sediment, and geomorphologic mapping, analyzing the mark left on the landscape by the different processes. In the paragraphs below, we try to follow the various lines of evidence from which a climatic history and a picture of the landscape evolution can be built (Fig. Cl .2, and its references). Since a controversy exists over just how warm Antarctica might have become during the Pliocene, our overview focuses necessarily on that time period.

Cl .2.1 The Ocean Record The onset of glaciation in Antarctica has been causally linked to its thermal isolation following Australian break-up and the subsequent establishment of the circum- polar current around 40 Ma (Kennett 1977). The build-up of ice on Antarctica has been Atlantic


Ocean Indian

Ocean 18(f

Figure C1 .1. Map of Antarctica. Sirius Group glacial deposits are indicated by dots (after Denton et al. 1991 ). 57

18 inferred from the shift to heavier 0 0 during the Eocene (Matthews and Poore 1980, Prentice and Matthews 1988). Similarly, significant increases in polar ice volume are 18 likely reflected in the o 0 shifts to heavier ratios at ca. 36 Ma and 14 Ma (Kennett and Shackleton 1976, Shackleton and Kennett 1975). Ice-rafted debris (IRD) found in Oligocene sediments in southern ocean cores (e.g. Hayes and Frakes 1975, Barron et al. 1989) provide evidence for the presence of ice on Antarctica. With time, IRD is found in sediments farther and farther to the north indicating the increasi.ng importance of glaciers with time on Antarctica (Wise et al. 1991). A notable cooling of ocean temperatures, as well as the increased importance of upwelling, is revealed by calcarE!OUS giving way to diatomaceous. This transition was found in sediment at the Oligocene/Miocene boundary (Kennett and Barker 1990, Burckle and Pokra!; 1991 ). Another important mineralogical change noted in southern ocean cores is seen in the clay mineralogy. Prior to roughly the Oligocene/ Miocene boundary, ii lite dominated the clay mineralogy of sediments while afterwards smectite did (Robert and Maillot 1990). When chemical weathering dominates as the process breaking down rocks then illite is the dominant clay mineral in offshore sediments. Pollen and wood of Nothofagus (southern beech) in Oligocene sediments {e.g. Kemp and Barrett 1975, Truswell 1986) give additional evidence that prior to the Miocene the climate in Antarctica was more temperate. More recent evidence (Francis 1995) shows that the trees that were found were environmentally stressed and probably lived close to the ground rather than upright. The interpretation of sea level curves for ice volume on Antarctica is equivocal, especially with respect to the Pliocene. The 60 m global Pliocene sea level rise postulated by Haq et al. (1987) would require a reduction in volume of the East Antarctic Ice Sheet (assuming no Greenland (-7.4 m) nor West Antarctic Ice Sheets (-5 rn) equivalent) by half to two-thirds. Evidence that early Pliocene sea level rise may have been more on the order of 30 m (Kennett and Hodell 1995) comes from (1) studies at the U.S. Atlantic coastal plain, namely 25-35 m sea level rise (Krantz 1991) and a 35 ± 18 m sea level rise between 3.5-3.0 Ma (Dowsett and Cronin 1990); and (2) the Enewatak Atoll, where a sea level increase of 29-36 m has been estimated (Wardlaw and Quinn 18 1991 ). Neither sea level estimates nor 0 0 curves provide evidence that a significantly smaller East Antarctic Ice Sheet has existed since the middle Miocene (Kennett and Hodell I 993). Fig. Cl .2. Summary of the landscape evolution in Antarctica (references in Table C1 .1 ). 59

I able£ l.L5.elected.references furfig~CL2. li'.'3.a Hodell and Venz 1992 Kennett and Shackleton 1975 Kennett and Hodell 1993 Matthews and Poore 1980 Miller et al 1987 Prentice and Matthews 1988 Savin et al. 1975 Shackleton and Kennett 1975 Woodruff et al. 1981

E11idence ior~glaciationsfoce.Ollgocene including. IRQandglaciomarine. sediments Barrett et al. 1989 Barron et al. 1989 ODP Leg 119 Breza et al. 1992 ODP Leg 120 Grobe et al. 1990 ODP site 693 Hayes and Frakes 1975 DSDP Wise et al. 1985

DiaJ:am/carbonata.transition Noltmfagus Kennett and Barker 1990 Askin and Markgraf 1986 Burckle and Pokras 1991 Hill 1989 Hill and Truswell 1993 Uplift and denudation Kemp and Barrett 1975 Gleadow and Fitzgerald 1987 Mildenhall 1989 Fitzgerald 199 2 Truswell 1986 Fitzgerald et al. 1986 Sugden et al. 1995 Wilch et al. 1993

AsQages EjordsJn_Wrlgbt'lLallay Marchant et al. 1993a, 1993b Hall et al. 1993

Oligocene/Miocene.age.fortbe.Sirius Barrett and Powell 1982 Brady and McKelvey 1979 Mercer 1972 Sugden 1992. 60

Cl.2.2 The Terrestrial Record On land, the history of the landscape has been deciphered by detailed mapping and Ar/Ar dating of ash layers mantling the landscape. Noel Potter and other workers (Marchant et al. 1993a, 1993b) have found numerous undisturbed in-situ ash deposits in arid landscape features of the Dry Valleys, e.g. in relict thermal cracks and on desert pavement. Clearly, the underlying feature predates the deposition of the ash, when evidence shows the ash is in situ (i.e. not reworked). The ages of the ashes determined by single crystal laser fusion 40ArJ3 9 Ar are up to 15 Ma (Marchant et al. 1993a, 1993b). It is also important to note that the volcanic glasses have not been chemically altered to clays, indicating a very dry climate since their eruption (Marchant et al. 1993a, Denton et al. 1993). Denton et al. (1993), Marchant et al. (1993, 1993b) and Sugden et al. (1995a) concluded that for at least the last 15 Ma the Dry Valleys have had a cold desert, hyper arid climate. Conversely, sometime before 15 Ma (though the oldest ash only provides a minimum age limit) the climate was neither as cold nor as dry since processes such as fluvial down-cutting, scarp retreat and planation dominated as shown by the morphology of the landscape. It was during this period that almost all of the landscape was sculpted in a semi arid environment (Denton et al. 1993, Sugden et al. 1995a). The amount of material removed by denudation to give the present Dry Valleys their present character corresponds to a wedge-shaped slice of rock, ca. 4 km thick at the thickest point near the coast (Sugden et al. 1995a). This estimate was based on several parameters including apatite fission track data (e.g. Fitzgerald 1992). Apatite fission track studies were used to constrain the magnitude and timing of uplift and exhumation of the Transantarctic Mountains. Crustal uplift has been episodic, with most occurring in the early Cenozoic (Gleadow and Fitzgerald 1987, Fitzgerald et at. 1986, Fitzgerald 1992). Minor tectonic adjustments since that time include ca. 400 m of subsidence in the Miocene, during which the Dry Valleys were filled with fjords (Hall et al. 1993), and a subsequent 300 m of uplift constrained by 3 Ma old, sub-aerial eruptive rocks (Wilch et al. 1993). The work of glaciers did play a part in carving the features of the Dry Valleys. But this was probably subordinate to the work from fluvial processes (Sugden et al. 1995a). The Transantarctic Mountains appear to have been overridden by ice from the East Antarctic Ice sheet at least once and this occurred before 13.6 Ma (Denton et al. 1991, 1993). During the Pliocene, the climate and extent of glaciation seem to have been little different from the conditions both before and after (Denton et al. 1993). The Quatermain I Till has a minimum 10Be exposure age of 4.4 Ma, indicating that the existed in a state similar to the present one since that time (Brown et al. 1991, Brook et 61 al. 1993). Pliocene glaciation was of a similar nature (cold-based) and of only slightly greater extent than Quaternary glaciation (Denton et al. 1993). The preservation of glacial ice lying undisturbed beneath an in-situ 8 Ma old ash layer implies as well the presence of a stable climate since the Miocene (Sugden et al. 1995b). These ideas may be roughly drawn together as the 'stabilist' point of view. The main tenets are: the climate in Antarctica has been a cold desert climate similar to today's since at least 15 Ma (Denton et al. 1993; Marchant et al. 1993a, 1993b), uplift and denudation were complete by that time (Sugden et al. 1995a), and glaciation in the Pliocene differed little from the maximum Quaternary glaciation (Denton et al. 1993).

Cl .3 A DIFFERENT POINT OF VIEW: THE DYNAMIC ICE SHEET MODEL The climatic development of Antarctica portrayed in Fig Cl .2 leaves out one critical piece of information, marine diatoms (age 3.1-2.5 Ma) that have been reported from Sirius Group sediments, in both the Beardmore region and in the Dry Valleys (Harwood 1986). The dynamic model is based on the idea that several times since the East Antarctic Ice Sheets inception it has melted down to ca. 1/3 its present size. This allowed a seaway to form in the area of the Wilkes and Pensacola basins (the last time was interpreted to have been between 4.8 and 3.0-2.5 Ma). Following each collapse episode, ice once again built up and diatoms from these seaways were transported with ice moving from inland Antarctica towards the Transantarctic Mountains. The diatoms were then deposited with the sub-glacial sediments of the Sirius Group. The age ranges of the reworked diatoms have been confirmed by 40 Ar/39 Ar dating of ash in the CIROS 2 core (2.8 Ma, Barrett et al. 1992). A!; it may be hard to explain an ice sheet so huge that it transported the diatoms to 2750 m, it has been put forth that the Transantarctic Mountains were subsequently uplifted at a rate of ca. 1 km/Ma since the Late Pliocene (Behrendt and 1991 ). It was also suggested that during the Pliocene and Pleistocene, this uplift was accompanied by intense erosion, both fluvial and wet-based glacial (Behrendt and Cooper 1991 ). If the Sirius sediments are less than ca. 3 Ma, old then the contained fossil wood in the Beardmore region implies that trees had existed in refuges through colder intervals, then later expanded in the warmer Pliocene (Webb and Harwood 1991 ). Temperatures on the order of 20-25 °C warmer than present (McKelvey et al. 1991, Hambrey and Barrett 1993) and rainfall in the range of 60-110 cm/yr have been suggested for that time interval (Mercer 1986). 62

Table Cl .2. Formations of the Beacon Supergroup' in South Victoria Land (Barrett 1992).

Age Formation

Jurassic Ferrar Kirkpatrick Basalt Mawson formation Lashly Formation

Triassic Feather Conglomerate

Victoria Weller Coal Measures Group disconformitv E. - Metschel Tillite L Carboniferous

disconformitv ...... Aztec siltstone

Beacon Heights Orthoquartzite

Arena Sandstone

Devonian Taylor Group Altar Mountain Fm.


Terra Cotta Si Its tone

Windy Gully Fm.

E. Paleozoic- Igneous and Metamorphic basement L Precambrian 1 Horizontal sills of Ferrar Dolerite intrude into the Beacon Supergroup sediments but are not shown here.

C1 .4 GEOLOGIC SETTING The bedrock geology of East Antarctica can be characterized by the features seen along the Transantarctic Mountains (Tingey 1991 ). It consists of a Precambrian to Early Paleozoic continental shield which was intruded by rocks of granitic composition (Granite Harbour Intrusives, Gunn and Warren 1962) during the Late Paleozoic (ca. 400- 550 Ma) (laird 1991 ). These rocks are cut by an erosional surface known as the Kukri Peneplain (Gunn and Warren 1962). Deposited on this peneplain are flat-lying, continental and nearshore sediments of the Beacon Supergroup (Table C1 .2) (Barrett et a1. 1972, 1992). The Devonian to strata comprise quartzose sandstones of the 63

Taylor Group which underlie interbedded sandstones, conglomerates, coal measures, as well as volcaniclastics (in the upper part) of the Victoria Group (Barrett et al. 1992). In the Early to middle Jurassic, horizontal sills of tholeiitic Ferrar Dolerite (Harrington 1958) were intruded along the contact of the Beacon with the basement rock, as well as within both the basement rocks and Beacon beds. The Ferrar 5upergroup includes the Kirkpatric Basalts which are the eruptive equivalents of the Ferrar Dolerite, as well as the volcanogenic Carapace Sandstone and the Mawson Tillite (Tingey 1991 ). The Table Mountain cross section (Fig. Cl .3) provides an example of the Transantarctic Mountains as seen in the Dry Valleys region. Located on the eroded surfaces of this bedrock terrain are deposits of glacial origin and the Miocene to Recent McMurdo Volcanics (Harrington 1958). Tile volcanic rocks, with compositions of phonolite to basalt and are found all along the Transantarctic Mountains. Antarctic glacial deposits, specifically those located in the Dry Valleys, can be generally divided into two groups: 1) the compact diamictites found at high elevation roughly grouped together as Sirius Group and 2) the more recent (cold-based) glacial sediments deposited within the present-day glacial drainage system.

Cl .5 SIRIUS GROUP The age of the glacigenic sediments of the Sirius Formation (Mercer 1968, 1972), now Sirius Group (McKelvey et al. 1991 ), lies at the heart of the stable vs. dynamic ice sheet controversy. The Sirius Group includes a diverse package of fades, including both lodgement till, glaciolacustrin•e and fluvioglacial sediments, as well as colluvium (Denton et al. 1991 ). In situ wood stumps have been found within Sirius Group sediments at Oliver Bluffs (e.g. Hill and Truswell 1993). Several authors have come to the conclusion that some of the Sirius Group sediments were deposited by glaciers near the pressure melting point at their base, i.e. by either more temperate or very thick glaciers or both (e.g. Mercer 1968, 1985). Outcrops of sediments classified as Sirius Group are found as remnant patches scattered along the length of the Transantarctic Mountains (Fig. Cl .1 ), generally at elevations greater than 1600 m. For details on the locations, and references for each of the Transantarctic Mountains Sirius Group outcrops, we refer to Denton et al. (1991). The rocks we have analyzed in this study come from three of the Sirius Group outcrops located in the Dry Valleys: • Table Mountain (ca. 1800 m) (Barrett and Powell 1982, Helfer and SchlOchter 1995), • Mount Feather (ca. 2750 m) (Brady and McKelvey 1979, 1983), and • Mount Fleming (ca. 21 OOm) (Denton et al. 1991, Stroeven et al. 1992, Helfer and Schluchter 1995). 64

Other Dry Valleys locations of high elevation glacial deposits include Shapeless Mountain (ca. 2400 m) (Mayewski 1975), Carapace Nunatak (ca. 2000 ml (Mayewski and Goldthwaite, 1975), Allan Hills (ca. 1750 m) (Borns in Denton et al. 1991 ), and Coombs Hills (ca. 1900m) (Brady and McKelvey 1983). The Sirius Group outcrops in the Dry Valleys are the subject of detailed field and laboratory investigations which are being carried out by Prentice, ~ichlOchter, and co-workers (Helfer, Stroeven). Although the relation in time of all the Sirius deposits is unclear, they have been classified together (McKelvey et al. 1991) based on the following general characteristics: • its lithology composed of massive diamictite; 'mud-rich till' (Denton et al. 1991 ), which are apparently 'wet-based', • the general compactness or semi-lithifled character of the sediments, • the presence of interbedded stratified layers to lenses, which are either more sand- rich and/or -rich, • its deposition on pre-Tertiary rocks, which are often grooved and striated, • its location at relatively high elevation. Denton et al. (1991) noted that the Sirius deposits can be separated into two distinct groups: basal tills found at high elevation, and till interbedded with stratified layers found at lower elevation. The latter occurs mostly as wedges along valley walls and can be classified as ice marginal fades.

Cl .5.1 Sirius Group outcrop at Table Mountain Barrett and Powell (1982) described in detail the Sirius Group beds found at Table Mountain. More recently, an investigation into the lithology and texture of this deposit is in the process of being completed by M. Helfer. The Sirius Group outcrop is found on a gently sloping bench on the NW side ofTable Mountain. This bench is cut into Terra Cotta Siltstone along its lower contact to a Ferrar Dolerite (Fig. Cl .3). This deposit covers an area of ca. 2 km 2 and is up to 15 m thick. It is made up of diamictite (basal till), interbedded with sand-rich and gravel-rich layers. Barrett and Powell (1982) have interpreted the latter to have been formed in a proglacial environment. Clasts, which range up to boulder size, consist almost solely of dolerite (some are striated). Barrett and Powell suggested a late Oligocene to early Miocene age. No diatoms have been described from the Table Mountain Sirius Group outcrop (Helfer and Schliichter 1995).

Cl .5.2 Sirius Group outcrop at Mount Feather The characteristics of the Sirius Group deposit located along the eastern side of Mount Feather at 2750 m have been described by Brady and McKelvey (1979). The deposit, classified as a basal till, lies on striated pavement of the Beacon Supergroup. Striated clasts are common, dast lithology is dominated by dolerite, while clasts of coal, 65 sandstone, and Feather Conglomerate (Beacon Supergroup, see Table C1 .2), as well as metasedimentary basement rocks occur less frequently. Brady and McKelvey (1979) believed that some of these rule out a local origin from Mount Feather. Brady and McKelvey (1979) noted that the glacial deposit at Mount Feather had been 'cut and swept' by ice from the nearby cirque. They interpreted the deposit to lie in an older glacial valley that was later cut by glaciers which follow the present drainage system. They therefore estimated the age of the deposit to be early or middle Miocene. Since their investigation, late Pliocene marine diatoms (Harwood 1986), freshwater diatoms (Kellogg and Kellogg 1984, Harwood 1986), Eocene foraminfera, and silicoflagellates have been described from Sirius sediments at Mount Feather (Webb et al. 1984).

Cl .5.3 Sirius Group outcrop at Mount Fleming Bedrock mapping by Pyne (1984) delineated the distribution of the glacial deposits along the south side of Mount Fleming, now assigned to the Sirius Group (Webb et al. 1984). These deposits, also termed Mount Fleming Upper Valley Drift, recently have been the subject of detailed fabric and lithologic study (Helfer and Schluchter 1995, Stroeven et al. 1992, 1994). The basal till is massive and poorly sorted, the matrix being made up of sand and silt derived locally from the Terracotta Siltstone. Clast lithology reflects a local source as well; predominantly fine-grained dolerites and lesser Beacon sandstones (Helfer and SchlUchter 1995). Rat tail interpretations indicate an ice flow direction from the north northeast (Stroeven et al. 1992, 1994). They interpreted the Sirius ice to have come from the north and to predate the existing valley structure. Harwood (1986, Harwood and Webb 1986) described the presence of diatoms which are latest Miocene to early Pliocene in age in Mount Fleming sediments.

Cl .6 SUMMARY OF PART C We have determined exposure ages for samples collected from the surfaces of erratic boulders from three Sirius Group outcrops in the Dry Valleys. Data from Table Mountain and Mount Fleming are presented in Chapter C2 (published in Geology, November 1995). Also discussed in Chapter C2, are a suite of Beacon sandstone bedrock samples, whose exposure ages constrain the minimum time of formation of the plateau surface at Table Mountain. Chapter C3 includes additional data from the Sirius Group outcrop at Mount Fleming. The new data support the interpretation that that outcrop is early Pliocene in age or older. In the final chapter pertaining to Antarctica (C4), a detailed discussion of erosion, uplift and landscape evolution in light of our new exposure ages is presented. 66




Fig. C1 .3. Cross-section through Table Mountain (modified from Barrett and Powell 1962). 67


Susan Ivy-Ochs lnstitut fur Teilchenphysik and lngenieurgeologie, ETH Honggerberg, CH-8093 Zurich, Switzerland

Christian Schluchter Geologisches lnstitut, Universitilt Bern, CH-3012 Bern, Switzerland

Peter W. Kubik Paul Scherrer lnstitut c/o lnstitut fur Teilchenphysik, ETH Honggerberg, CH-8093 Zurich, Switzerland

Beate Dittrich-Hannen lnstitut fur Teilchenphysik, ETH Honggerberg, CH-8093 Zurich, Switzerland

)Org Beer EAWAG, CH-8600 Di.lbendorf, Switzerland

1 Geology (1995, v. 23, p. 1007-1010), reprinted with permission of the Geological Society of America. 68

C2.1 ABSTRACT Using accelerator mass spectrometry (AMS), we have measured 10Be and 20AI in quartz from granites and sandstones from Table Mountain and Mount Fleming, Antarctica. Our data show that the plateau surface at Table Mountain had formed at the latest by early Pliocene. Granites fringing but within the Sirius Group at Table Mountain give a minimum exposure age of 2.6 Ma for this deposit. A sandstone clast on the Ferrar dolerite surface just outside and below, and thus postdating, the Sirius Group has a minimum age of 2.9 Ma. Two samples from the Sirius Group at Mount Fleming have 10Be concentrations that have reached secular equilibrium. This deposit is at least 4.8 Ma old. The Sirius Group at Mount Fleming cannot have been deposited after 3.0-2.5 Ma, as implied by biostratigraphic data. Our dates contradict the hypothesis that in the Pliocene East Antarctica was deglaciated and the climate was significantly warmer and wetter. The preservation of these surfaces indicates a continuous cold desert in theDry Valleys since the beginning of the Pliocene. The high 10Be concentrations we have measured cannot be reconciled with uplift of the Transantarctic Mountains at a rate of 1 km/Ma during the past 3 Ma.

C2.2 INTRODUCTION The question of whether the East Antarctic ice sheet has been a stable feature of Antarctica since its formation or whether its size can vary significantly within a short time period is crucial with respect to possible greenhouse scenarios. The stable ice sheet point of view maintains that the East Antarctic ice sheet has existed roughly in its present form since at least Miocene time (Denton et al. 1993). The present landscape is largely relict, formed by scarp retreat and planation due to fluvial downcutting before 15 Ma (Denton et al. 1993, Sugden et al. 1995b). This includes the broad plateaux of the upper and intermediate surfaces (Sugden et al. 1995b) found in theDry Valleys (Fig. C2.1 ). Uplift and denudation were largely completed by middle Miocene (Sugden et al. 1995b). Deposition of the Sirius Group sediments, interbedded lodgement tills, glaciofluvial and glaciolacustrine fades, found scattered along the Transantarctic Mountains is thought to have occurred well before Pliocene time (e.g. Denton et al. 1993). Undisturbed in-situ ash deposits up to 15 Ma old (40Ar/39Ar dates), found in relict arid landcape features (such as desert pavement) indicate the continuous presence of hyper-arid cold-desert conditions in theDry Valleys (Marchant et al. 1993a). In the framework of thi~ dynamic ice sheet model (Webb and Harwood 1991, and references therein), the East 1\ntarctic ice sheet has undergone several episodes of expansion and contraction, the most recent being in Pliocene time. Marine diatoms reportedly found in Sirius Group sediments (McKelvey et al. 1991) in the Beardmore region and at Mount Feather in theDry Valleys (Fig. C2.1) indicate open-marine 69 conditions in inland Antarctica, between 4.8 and 3.0-2.5 Ma (Webb et al. 1984, Harwood and Webb 1986). The age ranges of the diatoms have been confirmed by an 40 9 Ar/' Ar age on volcanic ash of 2.8 ± 0.3 Ma in the CIROS 2 core (Barrett et al. 1992). According to the dynamic modd, after 3.0-2.5 Ma, the ice sheet once again expanded, entrained the diatoms, then deposited them with the subglacial sediments of the Sirius Group (Webb et al. 1984). Nothofagus (southern beech) found within Sirius Group sediments imply Pliocene temperatures 20-25 °C warmer than today's (e.g. Hill and Trusswell 1993). As this is unlikely at the present high elevation of the Sirius Group deposits (ca. 1800-2600 m in theDry Valleys), it has been postulated that the Transantarctic Mountains were nearer to sea level and were uplifted after 2.5-3 Ma, i.e. post-Sirius Group deposition (Webb et al. 1984, Behrendt and Cooper 1991, Webb and Harwood 1991 ).

C2.3 THEORY AND METHODS Terrestrial in-situ-produced cosmogenic nuclides (e.g. 10Be, half-life 1,500 ka and 26Al, half-life 716 ka) are produced within the mineral lattices in the upper decimeters of an exposed rock surface as a result of cosmic-ray bombardment (Cerling and Craig 1994, and references therein) according to


N is the number of atoms.g·' Si02, P is the production rate (atoms·g·1 Si02.yr' '), T is the exposure time (yr), A. is the decay constant (y( 'l, p is the rock density (g.cm·\ e is the 1 2 erosion rate (cm.y( ), and A is the cosmic ray attenuation length (g.cm. ). During the austral summers of 1991, 1992, and 1993, we collected sandstones from the plateau surface of the Beacon sandstone, granites from along the edge but within the Sirius Group and a sample of sandstone found on the Ferrar dolerite surface, at Table Mountain (Fig. C2.2). Also included in this discussion are two samples from the Sirius Group outcrop at Mount Fleming. Samples were collected under 'open-sky' conditions with dip angles of 0 ° to 10 °. Elevations were determined with global positioning system equipment, altimeters, and topographic maps; errors are estimated to be within 20 m. 70

E'2Lj Intermediate and upper surfaces B Rectlllnear slopes and ridges c::J Rolling Slopes and 1--="'-----...i.•.w..-----"'..... -----' valley benches @ Sirius Group E;iil Ice, lea-free valleys

Fig. C2.1. Map of part of the Dry Valleys (modified from Sugden et al. 1995, and Denton et al. 1993). We have grouped upper and intermediate surfaces (D. Sugden, personal communication) which include the plateau surface at Table Mountain. On the inset map, X shows location of Beardmore region, EAIS is East Antarctic ice sheet, WAIS is West Antarctic ice sheet, and TAM is Transantarctic Mountains.

Table Mountain

5:'!J Sirius Group

Table Mountain t2Zl plateau 0 , Rectilinear slopes Rolling slopes D and valley benches

Fig. C2.2. Enlarged schematic view of Table Mountain (modified from Sugden et al. 1995, and Denton et al. 1993). Black circles mark approximate sample locations. Escarpment at edge of Table Mountain plateau surface is at 1820 m above sea level. 71

We used the sample preparation methods described by Brown et al. (1991 ), Nishiizumi et al. (1991) and Kohl and Nishiizumi (1992) with minor modifications. Briefly, pure quartz, obtained by selective chemical dissolution, was completely dissolved together with 0.5 mg 9 Be carrier. Be and Al were separated by means of cation exchange and the hydroxides were precipitated, then baked to the oxides at 850 °C. 10Be and 26 Al were measured by accelerator mass spectrometry (AMS) at the Eidgenossische Technische Hochschule ZUrich/Paul Scherrer lnstitut (ETH/PSI) tandem accelerator facility in ZUrich.

C2.4 RESULTS The ages listed in Table C2. l were calculated with no erosion (withe= 0 in Eq. 11 ), thus many are conservative minimum ages (i.e., their age would increase by assuming even a very small erosion rate). Many of the rocks we analyzed had 10Be and 26AI concentrations at secular equilibrium, this is especially true for the sandstones with silicified surfaces. Secular equilibrium occurs when production and decay are balanced and the concentration of the isotope no longer increases with time. In such cases where the 1 a measurement uncertainty range includes the secular equilibrium value, we quote only a minimum age corresponding to the radioisotope concentration at the lower 1 a limit. Sandstones from the Sirius Group at Mount Fleming had 10Be concentrations in secular equilibrium. For this deposit, the minimum calculated exposure age is 4.8 Ma. The possibility that the Mount Fleming samples still contain meteoric 10Be is highly unlikely. This is based on comparison of 10Be concentrations from samples where sequential etching steps were analyzed (cf. Brown et al. 1991) and the coherence of the 26 10 AI and Be data for samples where both isotopes were measured. At Table Mountain, the 10Be mean exposure ages for granites found along the lower edge of the Sirius Group deposit range from 0.7 to 2.6 Ma. Displaying 10Be and 26 Al data on the erosion island plot (Fig. C2.3) often reveals complicated exposure and erosion histories, as a result of the difference in half-lives of 10Be and 26Al (Nishiizumi et al. 1991, Lal 1991). Four out of five of the results for the granites fall on or reach down to the steady-state (with respect to cosmogenic nuclides) erosion line in Figure C2.3, so that one may consider these ages to be minimum ages (erosion rates up to 70 cm/Ma). We note that with the stated errors and the location of the results on the erosion island, we cannot unequivocally determine secular equilibrium with erosion nor preclude brief periods of cover by ice or sediment for the granite samples. A silicified sandstone collected from the Ferrar dolerite surface slightly outside and below the Sirius Group outcrop (Fig. C2.2) has a mean exposure age of 2.9 Ma. This provides an additional minimum age for the Sirius Group at Table Mountain. Table C2.1. The AMS-measured 10Be and 26Al concentrations for Table Mountain and Mount Flem in~ sameles. 10 6 26 6 Sample Alt.• Samp.f Be x 10 Error Mean Lower Upper AI x 10 Error Mean Lower Upper Al Remark no. (m) (cm) a1/9 Sio, (%) Ase(Ma) limit limit at/9 SiOi (%) A9e (Ma) limit limit

Granite from margins of Sirius Group, Table Mountain 207 HF4 1840 5 112 6.5 1.20 1.06 1.35 164 207 Hf5 1840 5 104 6.0 1.04 0.94 1.16 118 207 HF6 1840 5 28.7 6.0 1.54 1.41 1.68 208 HF6 1840 2.5 21.3 5.8 0.99 0.92 1.07 90.6 6.3 0.81 0.73 0.89 129 209 HF5 1840 5 66.6 6.2 0.54 0.50 0.59 98 209 HF6 1840 5 15.9 7.8 0.71 0.65 0.78 70 5.9 0.58 0.53 0.63 172 217a HF6 1820 2 33.5 5.1 1.91 1.76 2.07 146 5.9 2.21 1.83 2.81 80 217b Hf6 1820 5 39.2 9.8 2.61 2.17 3.18 149 5.9 2.71 2.10 75

Table Mountain plateau, Beacon sandstone 213a HF2 2080 7 55.7 6.5 4.02 3.42 4.95 213a HF4 2080 7 57.0 5.9 4.30 3.63 5.28 214b HF4 2055 4 58.9 5.2 4.72 3.98 5.85 185 6.5 2.71 159 silicified 218 HF2 2170 7 33.7 10.3 1.42 1.22 1.64 N 218 HF3 2170 7 33.9 7.8 1.43 1.28 1.60 142 5.1 1.25 1.13 1.38 64 " 219a HF4 2100 3 37.6 7.1 1.70 1.53 1.89 172 10.5 2.04 1.52 3.12 43 219b HF4 2100 2 46.1 5.5 2.34 2.12 2.59 220 HF4 2090 2 66.2 7.1 4.75 226 5.5 4.48 56 silicified 323 Hf6 2060 3 62.0 5.2 5.48 4.46 198 11.2 2.48 39 silicified 325 HF6 2060 3 51.8 6.8 3.17 2.73 3.73 silicified 326 HF4 2060 1 65.4 5.6 5.05 silicified

Sirius Group at Mount Fleming, sandstone 320 HF4 2140 1 70.4 5.2 5.80 silicified 322 HF6 2140 2 67.4 5.2 6.15 4.83 silicified Note: Errors are at the 1a level, including the statistical (counting} error and the error due to the normalization to the standards and blanks. A sample processing reproducibility 16 26 1 error of 5% has been included. Exposure ages were calculated using the production rates of Nishiizumi et al. (1989), i.e. 6.0 Be atoms.g·• 5iQ1.yr' and 36.7 atoms Af.g" SiO,.yr" at sea level, and the elevation scaling fact°" of Lal (1991). The ages have been corrected for the thickness of sample processed (a correction of no more than 6%). No shielding corrections have been made. HF refers to the etching step from which the sample was made. Total Al was measured by inductively-coupled plasma atomic emission spectrometry using three standard additions. •Altitude. t Sample thickness. 73

The 26Al/10Be ratios for silicified sandstone slabs from the plateau surface at Table Mountain cluster at the lower right-hand side of the erosion island (Fig. C2.3), indicating 26 that both ' 0 Be and AI have reached secular equilibrium. Maximum erosion rates for the silicified sandstones calculated with the ' 0 Be concentrations are 0-5 cm/Ma.

C2.5 IMPLICATIONS FOR ANTARCTIC LANDSCAPE DEVELOPMENT Our exposure dates indicate that the Sirius Group sediments at Mount Fleming are 0 no less than 4.8 Ma old. Similarly, ' Be exposure ages of 2.6 ± 0.3 Ma and 3.8-0.9/+1.4 Ma have been measured by Brook et al. (1995b). These old exposure ages from the Sirius Group at Mount Fleming are clearly at odds with the idea that the Sirius Group was deposited after 3.0-2.5 Ma, as implied by Sirius Group diatom ages (Webb and Harwood 1991, Barrett et al. 1992). At Table Mountain, granites from the edge of the Sirius Group have 2.6 Ma minimum exposure ages. A sandstone from outside and below the Sirius Group, thus thought to postdate it, has a minimum age of 2.9 Ma. One might consider that the spread of the ages, 0.7 to 2.6 Ma, indicates that we have dated a lag deposit collecting on the Sirius Group at Table Mountain (e.g. Hallet and Putkonen 1994). Another possibility is that the granites do not belong to the main body of the Sirius Group, which otherwise contained no quartz-bearing clasts. Interesting in this context is the work of Barrett and Powell (1982), who described granitic clasts in a possibly younger moraine onlapping the Sirius Group deposit on the west side of Table Mountain. The situation at Table Mountain remains unresolved, it may be that more field work and exposure dating will help to answer these questions. In contrast to the uplift rates (1 km/Ma) suggested by Behrendt and Cooper (1991) for the Beardmore region, Brown et al. (1991) and Brook et al. (1995b) have presented a model showing that the concentration of 10Be in several of their samples from the Dry Valleys was too high for such an uplift rate. This is also true for our samples from the Sirius Group at Mount Fleming and the majority of the Table Mountain samples. These rock surfaces could not have achieved such high ' 0 Be concentrations in only 3 Ma, with the smaller production rates present at lower elevations. This agrees with estimates by Wilch et al. (1993) of a maximum of 300 m of uplift for the Dry Valleys in the past 3 Ma. Erosion rates calculated from our ' 0 Be and 20AI data from sandstones range from 0 to 30 cm/Ma (cf. Nishiizumi et al. 1991 ). On the basis of the elevation step-down from the sub-Sirius Beacon surface to the eroded-back Beacon platform, Marchant et al. (1993a) estimated that 0.5 to 3 m of sandstone had been eroded in about 10 Ma (5 to 30 cm/Ma). Sugden et al. (1995b) determined that prior to ca. 15 Ma a wedge of rock up to 4 km thick must have been removed in order for the Dry Valleys to have their present shape. Our erosion rates allow that in the past 3 Ma a maximum of about 1 m of sandstone has been removed from the plateau surface at Table Mountain. For the 74 sandstones with silicified surfaces, the calculated erosion rates are no more than 5 cm/Ma. Thus, the ventifacted siliceous crusts, direct evidence of a polar desert climate, had already formed by the early Pliocene. Our results contradict the idea that extreme uplift accompanied by intense erosion (Behrendt and Cooper 1991 J in a non-polar desert climate characterized the Dry Valleys during the Pliocene (Webb and Harwood 1991 ).

C2.6 CONCLUSIONS Our 10Be and 26AI data indicate that the Sirius Group sediments at Mount Fleming were deposited before 4.8 Ma. The plateau surface at Table Mountain was cut prior to the early Pliocene, as shown by the very old exposure ages for the silicified ventifacted sandstones. Neither of these features has been significantly eroded in the past several Ma; verified by the low erosion rates we have determined. A minimum age of 2.9 Ma can be assigned to the Sirius Group deposit at Table Mountain. Our Mount Fleming ages are not in agreement with the theory that the Sirius Group was deposited after 3.0-2.5 Ma (Webb and Harwood 1991 ). The stability and lack of erosion of these surfaces preclude the existence of a much more wet or temperate climate any time since the end of the Miocene. This contradicts the concept that between 4.8 and 3.0 Ma, East Antarctica was largely deglaciated and that the climate was warm enough to sustain vegetation. On the contrary, the exposure dates confirm that the Dry Valleys have had a continuous polar desert climate since before the Pliocene began (Denton et al. 1993, Marchant et al. 1993a, Sugden et al. 1995b). 75




2 6 1 0 1 0 Be atoms/gram

Figure C2.3. Plot of 26Al/ 10Be vs. 10Be. Concentrations have been normalized to sea level. All samples shown are from Table Mountain. Open diamonds indicate granites from the Sirius Group, solid diamonds indicate sandstones from the plateau surface, solid triangle indicates sandstone from atop the Ferrar dolerite surface. Banana-shaped area is known as steady-state erosion island (Lal 1991 }. Its upper boundary represents evolution with time of 261\l/10Be ratio with zero erosion. Lines splaying downward within island are paths of 26 Al/10Be ratio with time at various erosion rates (from left to right 30, 20, 10, and 5 cm/Ma). Lower boundary of erosion island connects final ratios achieved when rock is in steady-state erosion with respect to the two isotopes. We have used a density of 2.7 for both granite and sandstone and an average attenuation length of 150 g.cm-2 (Brown et al. 1992). 76


Susan Ivy-Ochs lnstitut !Or Teilchenphysik, ETH Honggerberg, CH-8093 Zurich, Switzerland

Christian SchlUchter Geologisches lnstitut, Universitlit Bern, CH-3012 Bern, Switzerland

Mike Prentice Department of Geology University or New Hampshire, Durham, New Hampshire, 03824, USA

Peter W. Kubik Paul Scherrer lnstitut c/o lnstitut !Or Teilchenphysik, ETH Honggerberg, CH-8093 Zurich, Switzerland

JOrg Beer EAWAG, CH-8600 Dubendorf, Switzerland

1 Terra Antarctica, 1996 (in press). 77

C3.1 ABSTRACT We have measured 10Be and 26AI concentrations in rock surfaces from several landscape elements in the Dry Valleys of Antarctica. These data have allowed us to determine minimum exposure ages and maximum erosion (down-wearing) rates for the surfaces investigated. Expo;;ure ages are considered to be minimum ages because of the difficulty in constraining absolute erosion rates, boulder weathering out, shifting or splitting. The plateau surface carved into the Beacon Sandstone at Table Mountain formed prior to the early Pliocene, based on the exposure ages greater than 5.5 Ma. This bedrock surface has experienced down-wearing at a rate of no more than 30 cm/Ma since at least the Miocene/Pliocene boundary. Clasts on top of the Sirius Group deposit at Table Mountain indicate a minimum age of 2.9 Ma for this deposit. We have determined exposure ages for two other Sirius Group deposits in the Dry Valleys, at Mount Fleming two clasts yielded minimum exposure ages of 4.8 and 5.8 Ma. A single embedded boulder from Mount Feather had a minimum age of 2.3 Ma. The low erosion rates combined with the old exposure ages from the Dry Valleys surfaces indicate a climate similar to today's in the Dry Valleys since or prior to the beginning of the Pliocene. Rapid uplift rates during the Quaternary must be discounted because of the high measured 10Be concentrations.

C3.2 INTRODUCTION The East and West Antarctic Ice Sheets lock up an immense quantity of water. For projections into climates warmer than today's the behavior of these ice sheets remains a crucial unknown. For example, it has been proposed that during the somewhat warmer climates of the early Pliocene, the East Antarctic Ice Sheet comprised less than half ils present size, and open seaways (Wilkes-Pensacola basin) existed within the Antarctic continent (Webb et al. 1984, Harwood and Webb 1986, Webb and Harwood 1987, 1991). The dynamic ice sheet model was an outgrowth of a younger than 3.0-2.5 Ma age for the glacigenic sediments of the Sirius Group. This age was based on the reported presence of Plio-Pleistocene diatoms (Webb et al. 1984,

Harwood and Webb 1986, McKelvey et al. 1991 1 Barrett et al. 1992). For the Sirius deposits to attain their present elevations (higher than ca. 1 BOO m) in the Transantarctic Mountains, rapid uplift (1000 m/Ma) since the Pliocene has been called upon (Behrendt and Cooper 1991). Although the relevance of the diatoms has been seriously questioned (Burckle and Potter 1996, Kellogg and Kellogg 1996), both the long term stability of the Antarctic Ice Sheets and the uplift history remain an issue. In direct contrast to the dynamic ice sheet hypothesis is the recent verification of the apparent antiquity of the Dry Valleys landscape based on the presence of 15 million year old in situ ash deposits within cold landscape features (Denton et al. 1993, Marchant et al. 1993a, Sugden et al. 1995b). 78

A good way to address the question of how long the climate in the Dry Valleys has been roughly similar to today's is to determine the age of rock surfaces and to estimate the rates of rock surface modification using in-situ produced cosmogenic nuclides. Our goals were to use cosmogenic isotopes to date directly the Sirius Group deposits and neighbouring landscape elements, and to place constraints on down- wearing of the investigated surfaces using the measured isotope concentrations. 10 26 In this paper, we present several new Be and AI exposure ages. This includes the first exposure dates from the Mount Feather Sirius Group deposit. We also report 26 here new 26AI data for the Mount Fleming Sirius Group deposit. This AI data prove that none of the 10Be can be attributed to meteoric contamination and verify the old exposure ages for the Mount Fleming Sirius Group sediments which were reported in Ivy-Ochs et al. (1995).

C3.3 METHODOLOGICAL DETAILS 10 26 Cosmogenic nuclide,;, e.g. the radionuclides Be (t112 =1,500,000 yr) and AI 21 (t112 = 716,000 yr) and the stable noble gases, 3He and Ne, are produced in the upper decimeters of an exposed rock surface as a result of cosmic-ray bombardment (Lal 1991, Cerling and Craig 1994a and references therein). The build-up of a radionuclide concentration can be expressed as:

N (12)

where N is the number ol atoms/gram Si02, P is the local production rate in atoms/g

Si02.yr, Tis the length oftime the surface has been exposed in yr, A. is the decay 1 3 constant in y( , p the rock density in g/cm , Ethe erosion rate in cm/yr, and A is the 2 cosmic ray attenuation length in g/cm • During the austral ,;ummers of 1991, 1992 and 1993, rock surfaces from several of the landscape elements in the Dry Valleys were sampled. The data presented here are from the following three sites Table Mountain, Mount Fleming and Mount Feather. Detailed sample locations can be found in Ivy-Ochs et al. (1995) and Bruno (1995). Quartz samples were prepared according to the methods described by Brown et al. (1991) and Nishiizumi et al. (1991 ). Pure quartz obtained by selective chemical dissolution (Kohl and Nishiizumi 1992) was completely dissolved together with 0.5 mg 9Be carrier. Be and Al were separated using cation exchange, the hydroxides were 10 26 27 precipitated then baked to the oxides at 850 °C. Be/9Be and Al/ Al were measured by accelerator mass spectrometry (AMS) at the ETH/PSI tandem accelerator facility in 79

Zurich. Stable Al was measured by inductively-coupled plasma atomic emission spectrometry (ICP-AES) using three standard additions per sample. In calculating exposure ages, we have used the presently accepted production rates which were determined empirically by Nishiizumi et al. (1989). These have been scaled to the sample altitudes (Lal 1991). The uncertainties of the production rates and scaling factors have been estimated at less than 15% (Lal 1991, Masarik and Reedy 1995). Recently, the validity of these production rates has been questioned. It is important to note that lower not higher rates have been suggested (Clark el al. 1995). If this were indeed the case, even more of our samples could be considered to be at saturation and our calculated maximum erosion rates would be even smaller. The listed exposure age:; (Tables C3.1 and C3.2) were calculated with t=O in Eq. 12 (i.e. by assuming no erosion):


We take erosion, to include weathering or down-wearing of the rock surface itself, as well as deflation of overlying fine-grained sediment. In Figure C3.1, the no-erosion 10Be concentration is represented by the upper curved line. Note that the 10Be concentration begins to flatten out after ca. 3 half lives. After that, the concentration of 10Be will not increase no matter how long the sample is exposed. The shaded band in Figure C3.1 shows the 10% error band on either side of the secular equilibrium (saturation) concentration. Sample 10Be concentrations lying within this band cannot be distinguished from the secular equilibrium concentration. The assumption of no erosion is a conservative one because most surfaces have probably experienced at least a little erosion or weathering. When the erosion rate is included, the actual exposure age for the rock surface is always older. Because one cannot know the actual erosion rate during the whole period of exposure, the ages listed in Table C3.1 (all calculated with t=O) must be considered to be minimum ages. There are additional possible geomorphologic processes that lead us as well to interpret the ages as minima. These include the possibility of variable erosion, weathering, or down-wearing rates; spalling, splitting, shifting, or rolling over of the boulders; or coverage by sediment, ice or snow. In general, processes that yield too young exposure ages are more likely than those yielding too old ages (e.g. Brook et al. 1993, Zreda 1994). 80

NO EROSION 5cm/Ma 7 10 10 cm/Ma 8"106 0 "' U5 6 30cm/Ma E 6x10 ~ ~ 50cm/Ma E 4"106 ~ l I I Q) al ;! mr---- 2"106 106 107 Time in years

Figure C3.1. Plot of 10Be concentration vs. time. The bold line shows the growth of the 10 Be concentration in the rock surface when the surface is not experiencing erosion or down-wearing. The series of lines below show the maximum possible concentration of 10 Be for the quoted erosion rat

If we were to make the assumption that the measured concentration has actually reached secular equilibrium wilh respect to the cosmogenic isotope (i.e. the horizontal part of the curve in Fig. CJ.1) then the exponential term in Eq. 12 is zero, and solving for i::, we get the maximum possible erosion rate for the measured 10Be concentration in a given rock:


The actual erosion rate could only have been smaller, not larger, otherwise the concentration of 10Be could not be so high. 10 26 By measuring both Be and AI even more information about the exposure history of a rock surface may be revealed. This data can be portrayed on the erosion island plot (Fig. C3.2) (Klein et al. 1986, Lal 1991, Nishiizumi et al. 1991 ). The upper 26 0 boundary of the erosion island follows the evolution of the Al/' Be ratio with time when there is zero erosion (continuous exposure). Along the bottom of the erosion island the rocks have reached secular equilibrium with erosion for both isotopes. Ratios plotting below the erosion island may have been exposed then buried (discontinuous exposure), although such scenarios can be difficult to uniquely constrain.

26 Table C3.1. AMS-measured HIBe and AI concentrations.

7.1x107 8 2.0 x 10 1.2 128 8 2.1x10 >3.1 161 Age ranges are calculated from the 1CJ measurement errors, which include normalization to the standards and blanks, as well as a 5 % sample reproducibility error.

C3.4 RESULTS ANO DISCUSSION C3.4.1 Beacon Sandstone Plateau Surface at Table Mountain Bedrock surfaces 220, 323, 325, 326 all had wind-formed, fluted silicified surfaces. These plates stood out a few centimeters, pedestal-like, from the surrounding 10 eroded rubble. Three out of four of the Be exposure ages for this type of sample have reached secular equilibrium. The actual exposure age of the Beacon Sandstone plateau 26 10 surface is older than the oldest minimum age (5.5 Ma). The Al/ Be ratios for samples 220, 323 and 326 cluster at the lower end of the erosion island confirming that their 10 26 Be and AI concentrations have reached secular equilibrium (Fig. C3.2). The 10Be 82 concentrations are so high in these rocks that they could have experienced only 5 cm of erosion per Ma, or less. Of these bedrock slabs, 325 had the thinnest silica crust and therefore the highest maximum erosion rate (10 cm/Ma). Samples 213a, 214b, 219a, and 219b comprise rounded loose cobbles or concretions, roughly 10 to 25 cm in diameter. Their surfaces lacked features of wind- molding. 10Be ages for these samples show a wide range, from 1.7 Ma to 4.3 Ma. Indeed the two oldest samples (213a and 214b) had thin silicified crusts, while the other two samples did not. We have measured one sample (218) chipped from a knob of Beacon Sandstone near the top of Table Mountain. The 10Be exposure age is 1.2 Ma and the 26AI age is 1.1 Ma. In contrast to the silicified slabs (e.g. 220), this bedrock sample has no silicified surfaces and could be easily disaggregated. Sandstone erosion rates range from 0-30 cm/Ma. These rates are similar to those noted in previous cosmogenic isotope-based determinations (Nishiizumi et al. 1991, Brown et al. 1991, Brook et al. 1993, 1995) and field interpretations (Marchant et al. 1993a). These rates are orders of magnitude less than denudation rates in more temperate climates (e.g. Saunders and Young 1983, Summerfield 1991). The erosion rates we have calculated are valid for the whole exposure period. This makes it unlikely that warmer or wetter conditions existed during that last 5.5 Ma. The lowest erosion rates are for the wind-polished ventifacted sandstones from the plateau surface at Table Mountain. This plateau surface is comprised of a series of platforms and verandas carved into the Beacon Sandstone. As exposure age simply correlates to the degree of silicification of the rock surface, erosion is the process dominating the inventory of th<• cosmogenic isotopes. All of the different steps of the Beacon surface are older than the minimum exposure ages we have measured.

C3.4.2 Sirius Group at Table Mountain The exposure ages for th(' granites from the edge of the Sirius Group surface at Table Mountain (207, 208, 209, 217a, 217b) range from 0.7 to 2.6 Ma. In the discussion of the Sirius Group at Table Mountain we also include results for sample 205 which was collected just outside the Sirius deposit on the Ferrar Dolerite bedrock surface. Our data allow three possible scenarios: 1) the granite clasts represent large boulders that split apart, 2) fine .. grained sediment has been eroding off of the clasts, and 3) the clasts were deposited around 3 Ma. Below we will discuss each of these along with the implications for the age of the Sirius Group sediments at Table Mountain. Tab e C3.2. Samp e lntormation, minimum exposure ages an d maximum erosion rates.

Sample Altitude Unit sampled Sampling location Rock Overall shape Silicified Surface 10Be Maximum No. (m) type and size surface ventifacted, Age• erosion rate fluted (Ma) cm/Ma

205 1820 dolerite surface Table Mountain sandstone round cobble > 1 cm. no 2.9 10 213a 2080 plateau surface Table Mountain sandstone round cobble < 1 cm. no 4.0 5

214b 2055 plateau surface Table Mountain sandstone round cobble > 1 cm. no 4.7 2

218 2170 plateau surface Table Mountain sandstone bedrock no no 1.4 30

...... ,,J ,_,..,l-.L-.1,,.. 219a 2100 p!ateau surface Ta.b1e :'Act.:ntair. sandstone I \JUI lU \..VUU'IC (10 no i.7 25 219b 2100 plateau surface Table Mountain sandstone round cobble < 1 cm. no 2.3 15 220 2090 plateau surface Table Mountain sandstone bedrock plate > 1 cm. yes >4.8 2 co 323 2060 plateau surface Table Mountain sandstone bedrock plate > 1 cm. yes >5.5 1 w

325 2060 plateau surface Table Mountain sandstone bedrock plate > 1 cm. ves 3.2 10

326 2060 plateau surface Table Mountain sandstone bedrock plate > 1 cm. yes >5.1 1 207 1840 Sirius Group Table Mountain granite cobble 1.5 30 208 1840 Sirius Group Table Mountain granite cobble 1.0 45

209 1840 Sirius Group Table Mountain granite cobble 0.7 70

217a 1820 Sirius Group Table Mountain granite boulder 1.9 20

217b 1820 Sirius Group Table Mountain granite boulder 2.6 15

309 2750 Sirius Group Mount Feather sandstone boulder < 1 cm. no 2.3 15 320 2140 Sirius Group Mount Fleming sandstone cobble > 1 cm. yes >5.8 1

322 2140 Sirius Group Mount Fleming sandstone boulder > 1 cm. yes >4.8 2 10 26 • Be and Al concentrations for all samples except 309 and 320 (see Table C3.1) are found in Ivy-Ochs et al. (1995). 84

3 Samples 207, 208 and 209 (volume of each ca. 0.3 m ) were taken from an area of the Sirius Group drift where a chaotic assemblage of small granitic clasts is found along the lower edge of the deposit (cf. Barrett and Powell 1982). Samples 217a and 217b were from two large gr

Group sediments deposited on that surface are considerably younger. Post- depositionally, the surface of the Sirius till may have been slightly modified by (cold- based?) glaciers at ca. 3-2 Ma. At that time, only a few clasts, (including granites) may have been deposited along the margins of the older deposit. Barrett and Powell (1982) noted that at Table Mountain granitic clasts can only be found along the lower margins of the Sirius Group outcrop. Notably, none of the granite clasts was found embedded within the till matrix.

C3.4.3 Sirius Group at Mount Fleming and Mount Feather We have measured 10Be in two sandstone boulders from Mount Fleming. Both 322 and 320 have high 10Be concentrations, minimum 10Be ages are 4.8 and 5.5 Ma, respectively. The new 26AI concentration reported here for sample 320 is at the secular equilibrium concentration. The 26AI exposure age is greater than 3.1 Ma. This validates the old 10Be exposure age for 320 and indicates none of the 10Be was of meteoric origin. These old ages have been confirmed as well by 21 Ne measurements in dolerite clasts collected nearby at the same Sirius Group outcrop on Mount Fleming (Bruno1995). Additionally, both the 10Be and 26AI data discussed here and the 21 Ne data of Bruno (1995) agree with 10Be and 26AI r+ieasurements of Brook et al. (1995) from Mount Fleming. In this paper we also report the first exposure dates from the Sirius Group at Mount Feather. The 10Be age is 2.3 Ma, the 26 Al age, 1.2 Ma. The location of the 2r'Al/10Be ratio on the erosion island plot (Fig. C3.2) and the given errors allow two possibilities. The rock has reached secular equilibrium with erosion (15 cm/Ma) or the boulder may have been covered (by snow, ice or sediment) following exposure (as it lies slightly below the erosion island). Geomorphologic evidence indicates clear rapid deflation of the till surface and possible flaking and spalling of the boulder surface. 21 Ne measurements are needed to verify prior exposure.

C3.4.4 Uplift of the Transantarctic Mountains Cosmogenic isotope concentrations can set limits on the amount of uplift which took place during a given interval of time (Brown et al. 1991, Brook et al. 1993, 1995). Cosmogenic radionuclides are especially suitable because the final saturation concentration is altitude dependent. One can conclude that rocks with very high 10Be concentrations, (in our case samples 214b, 220, 323, 325, 326, 320, 322) must have been at their present elevation for at least the last several million years (Brook et al. 1995). At Table Mountain and Mount Fleming, the Sirius Group deposits could neither have been deposited at lower elevation 3.0-2.5 Ma ago, nor have been rapidly uplifted since then (cf. Wilch et al. 1993). 86

C3.5 SUMMARY ANO CONCLUSIONS For time periods on the order of millions of years, cosmogenic isotopes have clear limitations. One cannot really constrain what has happened to a boulder since the time it was deposited. Simple processes like weathering, erosion or spalling of boulders always yield younger apparent exposure ages. Therefore, we can only quote minimum ages for the three Sirius Group deposits we have investigated. At Mount Fleming the minimum exposure age is 4.8-5.8 Ma, at Table Mountain 2.6-2.9 Ma and at Mount Feather 2.3 Ma. Because these are minimum ages we cannot preclude the possibility that the various Sirius Group deposits are noncontemporaneous (cf. McKelvey et al. 1991 ). But, cosmogenic isotopes can provide much more than just minimum age information. They also yield clues helpful for interpreting the evolution of the Dry Valleys landscape. One may calculate a maximum possible erosion rate corresponding to the measured 10Be concentration. We have determined erosion or down-wearing rates of 0-30 cm/Ma for Beacon Sandstone bedrock surfaces and 15-45 cm/Ma for sediments of the Sirius Group at Table Mountain. Overall, the low erosion rates combined with old exposure ages, especially forthe plateau surface at Table Mountain, indicate little modification of these surfaces has taken place for at least the last several million years. This confirms that at the elevation of our samples, it has not been significantly warmer or wetter since at least the earliest Pliocene. The high 10Be concentrations we have measured preclude rapid uplift of the Dry Valleys region of the Transantarctic Mountains from sea level to ca. 2000 m in just the last 3 Ma. 87


···~·····~ .. :::.~·-·

: : 5 ···i·····f ...... :... ······"'·

' ···i ···················!

3 ... , ..... , ... ······························"'·····················'·······1:··o"-+c·m····1·''M·t:'"a·

5 ~m/Mai

2 6 7 1 0 1 0 1 0 Be atoms/gram

Figure C3.2. Plot of 26Al/1°Be vs. 10Be. All measured sample concentrations have been normalized to sea level (Lal 1991 ). Open circles are sandstone clasts from the Sirius Group at Mount Fleming (320) and Mount Feather (309); filled circles are bedrock (218, 220, 323, 326) or concretion samples (213a, 214b, 219a, 219b) from the Beacon Sandstone plateau surface at Table Mountain; squares are granite clasts from the Sirius Group at Table Mountain; the triangle represents sandstone clast (205) from just outside the Sirius Group at Table Mountain but on a dolerite surface. An average attenuation length of 150 g/cm 2 (Brown et al. 1992) and a density of 2.7 g/cm3 for both sandstone and granite have been assumed. 88


C4.1 INTRODUCTION In this chapter, we bring together all the threads relating to uplift and weathering of rock surfaces in the Dry Valleys as well as other regions of the Transantarctic Mountains. Following that discussion, recent evidence on the origin of the diatoms reported from Sirius Group sediments is discussed. Finally, a summary of the evolution of glacial ice on Antarctica and the development of the Dry Valleys landscape are presented in light of our cosmogenic isotope results.

C4.2 UPLIFT 10 In Chapters C2 and C3, the fact that high Be concentrations preclude rapid uplift during the past several million years was discussed briefly. In this section, we would like to provide the full basis for such conclusions. The following discussion may be viewed in two ways. First, old exposure ages measured at various localities can be viewed as a precluding the possibility of rapid uplift in the last few million years. To this end, we discuss all the high elevation sites where high concentrations of 10Be have been measured. Second, one can take an assumed uplift rate based on empirically measured uplift rates and calculate rough minimum ages for the various Sirius Group sites.

C4.2.1 Maximum uplift constraints The cosmogenic radionuclide concentration measured in a rock can set limits on uplift rates. This topic was explored by Brown et al. (1991) and Brook et al. (1993, 1995b). The fact that radionuclides reach saturation or secular equilibrium means they will always reach a maximum concentration which is characteristic for the production rate at the elevation at which they are found. Since production rates are lower at sea level than that at 2000 m, the maximum concentration that can be reached at sea level is therefore less than that at 2000 m. For example, at sea level in Antarctica, 6 10Be atoms/g Si02 are produced every year, while at 2000 m elevation, 32 atoms/g are produced. For a 10 rock continuously exposed for at least 5 Ma, the saturation concentration of Be (maximum concentration) will be reached. As explained in chapter B1, the concentration will no longer increase with time after that. At sea level the saturation concentration of 10Be (maximum concentration) will be five times less than the saturation concentration for a rock exposed for the same amount of time at 2000 m. This is not true for the stable 21 cosmogenic isotopes, such as 3He and Ne, noble gases because they do not decay, and therefore are integrated over all periods of exposure. For those isotopes, no maximum saturation concentration exists. That means a long period of exposure at lower elevation 89

3 21 cannot be distinguished from a short one at higher elevation using He and Ne, or a combination of a long period at sea level followed by very rapid uplift. The model of Brown et al. (1991) and Brook et al. (1993, 199Sb) is based on this altitudinal dependence of the cosmogenic radionuclide production rate. Based on this model, the amount of 10Be accumulated during uplift from sea level to the present elevation of the sample can be calculated (illustrated in Fig. C4.1 ). The model is expressed by the following relation:


where, N is the number of atoms/gram SiOz,

P,_1, is the sea level production rate in atoms/g Si02.yr, T is the length of time the surface has been exposed in yrs and the length of the period of uplift from sea level to the present elevation of the sample, 1 A is the decay constant of the radionuclide in y( , 3 p is the atmospheric density in g/cm , U is the uplift rate in cm/yr, calculated for the present sample elevation by varying the period of uplift, and 2 2 A, is the cosmic ray attenuation length in the atmosphere in g/cm (142 g/crn , Brook et al. 1995b).

We have adjusted the decrease in atmospheric density because of increasing altitude during uplift by using the formula:


Where, 3 pis the atmospheric density in g/cm , 3 p,1 is the atmospheric density at sea level in g/cm , h is the elevation attained, and

H0 is the atmospheric scaling factor appropriate for 2000 m elevation (Kneubuehl 1982) 90

It must be noted that this atmospheric density correction is only an approximation, additionally we have not included the effect of temperature. For a detailed discussion of these calculations, see Brown et al. (1991 ), Lal (1991) and Kurz and Brook (1994).

8.00E+07 t! 0 7.00E+07 2 :;) er 6.00E+07 E 0 5.00E+07 Oi -;;, 4.00E+07 • E 3.00E+07 · 0 0 2.00E+07 © ...... ~ 1.00E+07 O.OOE+OO 0 co co co co co co 0 0 0 0 0 0 0 ~ + + s + + + + s + s+ + w w w w w UJ w UJ w 0 0 i1i0 0 0 0 0 i1i0 0 0 8 q 0 0 0 0 0 0 0 0 q d C\i t"'.i .f lfi <6 r...: cci o:i Time (years)

10 Figure C4.1. Increase of the Be concentration for a rock surface continuously exposed at 2000 m (thin line) and for a rock surface continuously exposed at sea level (dashed line). The bold line is for a rock surface that is uplifting during exposure. Point 1 indicates the 0 ' Be concentration for a rock which has been uplifting for 3 Ma. Point 2 shows the concentration we have measured in a sample at 2000 m (Mount Fleming).

10 In Figure C4.1, the thin line shows the increase of the Be concentration with time in a rock surface continuously exposed at 2000 m (Eq. 1, Chapter Bl) (i.e. no uplift). The dashed line shows the increase for a rock continuously exposed at sea level (also no uplift). The bold line was plotted according to Eq. 15 above and using 2000 m of uplift (from sea level to 2000 m). For all points on the bold line, the beginning of exposure coincides with the beginning of uplift. Note that for the concentrations modeled with uplift (bold line), along the x-axis, the amount of time of uplift is changing, and therefore the rate of uplift. For example, a rock that has been uplifted from sea level to 2000 m in 3 Ma, therefore has a corresponding uplift rate is 666 m/Ma. In a rock surface undergoing such uplift, the amount of ' 0Be that was produced during the period of uplift 7 (which is the same as the period of exposure) is ca. 2.5 x 1 0 atoms/gram, This is shown 10 by the point labeled 1. Point 2 indicates the measured Be concentration for rock 320 from Mount Fleming. We can see that the measured concentration is much greater than could be attained with the discussed uplift scenario. Similarly, the high concentrations 91 we have measured preclude uplift rates of 1000 m/Ma for the Transantarctic Mountains (Behrendt and Cooper 1991) during the last 3 Ma. Based on the above model, high 10Be concentrations measured in rock surfaces found at high elevation rule out rapid uplift for those surfaces. Therefore, one can identify tectonic blocks which could not have been uplifted during the past several 10 million years by measuring Be (Brook et al. 1995b). If these prove to have concentrations too high to have been at sea level only 3 Ma ago then rapid uplift is not a possibility. In Table C4.1, we have compiled all published 10Be measurements for various high elevation surfaces in the Transantarctic Mountains. The sampling locations are shown as numbers in Fig. C4.2.


Figure C4.2. Map of the Transantarctic Mountains showing Sirius Group outcrops as dots. The numbers are the sites where high 10Be concentrations have been measured, site details are located in Table C4.1. WAIS= West Antarctic Ice Sheet, EAIS East Antarctic Ice Sheet.

The measured rocks in Table C4.1 include samples from bedrock surfaces and clasts in the Sirius Group. The samples from the farthest south are two bedrock samples collected from the Queen Alexandra Range (Nishiizumi et al. 1991 ). The Queen Alexandra Range is adjacent to the Beardmore Glacier, where key outcrops of the Sirius Table C4.1. Tectonic blocks where no significant uplift could have occurred based on high 10Be concentrations, locations correspond to numbers in Fig. C4.2.

Map Location Rock type and association oldest present Reference 10 No. Be age elevation (Ma) (m)

1 Tillite Glacier, sandstone bedrock 3.9 2200 Nishiizumi et al. 1991 Queen Alexandra Range

2 Table Mountain, Dry Valleys sandstone from plateau surface, 4.8 2040 Chapter 2 bedrock

3 Mount Feather, Drv Valleys sandstone clast from Sirius Group 2.2 2175 Chapter 3

4 Mount Fleming, Dry Valleys sandstone clast from Sirius Group 4.8 2040 Chapters 2 and 3 3.8 2065

5 Allan Hills, Dry Valleys quartzite, loose rock 2.1 2050 Nishiizumi et al. 1991

6 Robert's Butte, quartz vein from Granite Harbour 4.0 2825 van der Wateren and North Victoria Land granite bedrock Verbers, 1992 93

Group are found (McKelvey et al. 1991 ). A single sample from an elevation of 2200 m 10 had a minimum Be age of 3.9 Ma. This is a bedrock sandstone sample collected from a glacially eroded ridge of the Buckley Formation (Beacon Supergroup, see Table C1 .2) 10 near the Ti Ilite Glacier. In North Victoria Land, a minimum Be age of 4 Ma indicates that this glacially striated bedrock surface at Robert's Mountain (van der Wateren and Verbers 1992) has been at 2825 m since at least the early Pliocene (Brook et al. 1995b). Data indicating that the highest elevation surfaces of the Dry Valleys could not have been much lower in the past few million years are more plentiful and were measured by three different groups (Brown et al. 1991, Nishiizumi et al. 1991, Brook et al. 1995b, and Chapters C2 and C3).

1.00E+08 t! 0 8.00E+07 er:;) E 6.00E+07 ~ Q) ';;; E 4.00E+07 E 0 Q) 2.00E+07 ¥' O.OOE+OO 0

10 Figure C4.3. Increase of the Be concentration when the rate of uplift is fixed at 100 10 7 m/Ma. The intersection of the measured Be concentration (7 x 10 atoms/gram quartz) in the Mount Feather sample 309 and this constant rate of uplift curve occurs at ca. 27 Ma.

C4.2.2 Minimum age from uplift One may also take the 'known' uplift rate and calculate a minimum exposure age. The actual rate of uplift can be assumed based on apatite fission track studies along the Transantarctic Mountains (Fitzgerald et al. 1986, Gleadow and Fitzgerald 1987, Fitzgerald 1992) and the age (2.98 Ma) of the sub-aerially erupted volcanic cones in the Dry Valleys (Wilch et al. 1993). In both cases the uplift rate is on the order of 100 m/Ma. Minimum exposure ages may, thus, be calculated based on the assumption that a deposit was originally lain down at an elevation nearer to sea level (which may not 94 necessarily be true). Using Eq. 15, we can then calculate how much time would be 10 needed to acquire the measured Be concentration. For discussion, we use the 10Be concentrations measured from Mount Feather and Mount Fleming (Table C2.1 Chapter C2 and Table C3.1 Chapter C3). It is important to remember that the ages suggested in this discussion are model ages. They are not actual measured exposure ages and are not meant to imply that 25 Ma exposure ages have been measured with 10Be. In Fig. C4.3, the empirically determined uplift rate of 100 m/Ma is used (Gleadow and Fitzgerald 1987, Fitzgerald 1992, Fitzgerald et al. 1986, Wilch et al. 1993). Again, the beginning of exposure coincides with the beginning of uplift. The line shows the 10 increase of the Be concentration with exposure time during uplift. In the case of the Mount Feather sample, the question is how long it takes to uplift it from 0 to 2750 m, and what is the corresponding final 10Be concentration? At a rate of 100 m/Ma, the uplift from 0 to 2750 m takes 27.5 Ma. The 10Be concentration calculated with this model for 7 this time span is 7 x 10 atoms/gram, which coincides with the actual value we have measured for the Mount Feather sample (309).

1.00E+08 ~ 0 ::i 8.00E+07 O' E e 6.00E+07 · .._Ol E 4.00E+07 ~ (I) 2.00E+07 ff' 0.00E+OO ,.... I'- I'- ,.... 8 ~ 0 t; 0 t; w+ w w+ w'l w+ w + w'l 0 0 0 0 i1i 0 8 C? ~ 'Il 0 If) ~ !.() 0 I.() <>i <>i ro C'.i Time (years)

Figure C4.4. The increase of the 10Be concentration in the rock surface which is exposed at 2000 m with no uplift (thin line) and when the rate of uplift from Oto 2000 m is fixed at 100 m/Ma {bold line) for 20 Ma. After 20 Ma, (the kink in the bold line) the rock has reached 2000 m and uplift stops. Then another ca. 5 Ma are still needed for the concentration in the rock surface to reach saturation level for that (new) elevation.

The Mount Fleming samples differ from the Mount Feather samples in that they have saturation concentrations for their altitude (2000 m). This means that they must 95 have been at that altitude for at least the last 5 Ma. If we once again assume that they were deposited near sea level and then uplifted at a rate of 100 m/Ma, 20 Ma are required to uplift to 2000 m1 and at least 5 Ma are needed to attain saturation levels for that elevation. This model is shown in Figure C4.4., the bold line presents the uplift part of the model. After the kink the surface has reached 2000 m and is continuously exposed at that elevation. One can see that ca. 5 Ma are still needed to reach the saturation concentration. For comparison, the upper thin line shows the increase of 10Be at 2000 m with no uplift. In both cases, Mount Feather and Mount Fleming, estimated model exposure ages are on the order of 25 Ma. Coincidentally, this is very similar to other age estimates of Oligocene to Miocene for the Sirius deposits (Brady and McKelvey 19791

Barrett and Powell 1982, Mercer 19721 Sugden 1992).

C4.3 EROSION In Antarctica, denudation by wind erosion, by glacial scouring, and by the action of moving water (mostly along the valley floors), progresses at relatively rapid rates, but very locally. At present the majority of the landscape is modified by simple in situ weathering or rock surface lowering. In Table C4.2, we have compiled rock surface lowering rates for Antarctica. Shown are both actual measurements and rates calculated from cosmogenic isotope concentrations. The calculation of erosion rates (including in situ weathering) using the measured rnBe concentrations has been discussed in detail in chapter B3. This calculation is based on the assumption that the concentration found in the rock surface is actually the saturation concentration. Then one may solve Eq. 1 (Chapter Bl) for£. In this way, we have calculated weathering rates from various rock surfaces in the Dry Valleys. Beacon sandstones that were armored with a silica rind had extremely low rates (0-5 cm/Ma). Notably, even friable Beacon Sandstones (New Mountain Sandstone at the top of Table Mountain) with no protective coating appear to be falling apart at rates less than 30 cm/Ma. The highest rates we determined were from the Sirius Group sediments at both Table Mountain and Mount Feather. In both cases it is not dear, if the process is simple in situ weathering or if the additional help of wind is involved. In comparison (Table C4.2), actual measured rates for wind erosion (15-2000 cm/Ma) and for weathering due to salt crystallization (2000 cm/Ma) in Antarctica are much higher than the rates determined using cosmogenic isotopes. The rates of wind erosion are extreme but may also be localized, for example, where katabatic winds are coming off the polar plateau. Indeed, we may be comparing two distinctly different processes; active erosion and in situ weathering. In both cases, the rates are markedly less than the denudation rates calculated for dry climates of low (5-35 m/Ma) or high relief (45-370 m/Ma) at other locations on the earth (e.g. Saunders and Young 1983, Summerfield 1991 ). If areas of rapid erosion are very localized, then most of Antarctica is f. h . T bl C4.2. a e companson o in situ weat en ng rates (a b ove t h e b 0 ldrine an d measure d erosion rates (b e Iowt h e b 0 ldrine. Process Lithologic characteristics Location Rate (cm/Ma) Reference in situ weathering unprotected sandstone Table Mountain 30 Chapter C2, C3 bedrock Allan Hills 50-140 Nishiizumi et al. 1991 in situ weathering (?) sandstone clasts, Altar and Arena Valley 5-27 Brown et al. 1991, Brook et al. 1993 Quartermain Tills in situ weathering silicified sandstone bedrock Table Mountain 1-8 Chapter C2, C3 Allan Hills 7 Nishiizumi et al. 1991 in situ weathering silicified clasts, Sirius Group Mount Fleming 1-2 Chapter C2, C3 in situ weathering sandstone clasts, Sirius Mount Fleming 6-13 Brook et al. 1995b in situ weathering sandstone clasts Mount Feather 15 Chapter C3 (and wind-deflation?) (and matrix ?) Sirius in situ weathering granite clasts Table Mountain 10-70 Chapter C2, C3 (and wind-deflation?) (and matrix ?) Sirius measured wind erosion dolerite Dry Valleys 1500 Malin 1985, 1988 basalt 2000 measured wind erosion gneiss and dolerite Vestfold Hills and 1500-2400 Spate et al. 1995 Larseman Hills salt erosion gneiss Vestfold Hills 2400 Gore and Calhoun, 1995 97 being modified only very slowly, and the landscape is relict (cf. Denton el al. 1993, Sugden et al 1995a). The calculated low weathering rates for most of the landscape agree with landscape modification processes inferred from the clay mineralogy in ocean sediments (Robert and Maillot 1990). After the Oligocene/Miocene boundary, the dominance of smectite verifies of the lack of chemical weathering of primary minerals. All of the data confirm the presence of little or no chemical weathering on Antarctica for the last ca. 5-6 Ma (cosmogenic isotopes-minimum ages), 15 Ma (ash ages, see section C4.5 below) or 35 Ma (clay composition).

C4.4 RECENT DATA CONCERNING THE DIATOMS The following diatom interpretations have been reported from the sites we have investigated: at Mount Feather late Pliocene marine diatoms (Webb et al. 1984, Harwood 1986), and at Mount Fleming marine diatoms latest Miocene to early Pliocene in age (Harwood 1986). No diatoms have been reported from the Sirius Group sediments at Table Mountain (Helfer and SchlOchter 1995). As pointed out by Denton et al. (1991 ), if Sirius ice had indeed eroded into near shore inland sea deposits and enclosed them in subsequent Sirius Group deposits, one would have expected to find macrofossils along with the diatoms. Recently, the relevance of the diatoms for the age of the Sirius Group has been seriously questioned (Burckle and Potter 1996). They have found diatoms of various ages, including modern ones, even in Beacon Sandstones (Devonian), Granite Harbour granites {Late Paleozoic) and Precambrian basement rocks. They suggest that the diatoms are probably windblown. A detailed study of diatom counts with depth in a pit dug on Mount Fleming has shown as well that the diatoms are a surface phenomenon (Stroeven 1994). Kellogg and Kellogg (1996) described diatoms all the way through an ice core taken from the East Antarctic Ice Sheet. They also concluded that the only explanation was that they were windblown. That makes it very unlikely that the ages of the diatoms have anything to do with the age of the Sirius Group outcrops.

C4.5 EVOLUTION OF ANTARCTIC ICE ANO LANDSCAPE In the following discussion we attempt to build a picture of ice and landscape development in the Dry Valleys region of Antarctica and our cosmogenic isotope results are included where appropriate. The picture is by no means complete, however. Before there was ice on Antarctica, the climate was semi-arid (Eocene or early Oligocene?). The Dry Valleys were carved in this semi-arid climate by the action of streams (Denton et al. 1993, Sugden et al. 1995a). As the Antarctic continent moved closer to a position right at the south pole, and as the circumpolar current became better established, the continent became colder and colder (Kennett 1977). Ice began to form, 98 first as alpine glaciers. This glacial activity may have carved into and modified the already established fluvial valleys. For these glacial periods, which occurred some time prior to ca. 15 Ma, there are as yet no direct age determinations on these glacial deposits. The only information we can glean is from southern ocean cores, which indicates that glacial activity began in the Eocene or Oligocene based on 0180 changes and the presence of IRD (e.g. Hambrey and Barrett 1993). For the time period after 15 Ma, glacial activity in the Asgard Range and Beacon Valley (Dry Valleys) has been fairly well constrained. A chronology was established by 40 39 Ar/ Ar dating of in situ ash layers (Marchant et al. 1993a, 1993b). If we can assume that the Dry Valleys provide an example of the climatic conditions of the Transantarctic Mountains over time, then we can begin to track the development of the climate. Marchant et al. (1993a) have recognized (at least?) one episode of wet-based alpine glaciation that is older than an ash layer dated at 15 Ma. The Sessrumnir and the Inland Forts Tills, whose ages are constrained to be> 15 Ma by ash layers contain evidence that the glaciation was wet-based. As the continent became even colder, alpine ice coalesced and ice sheets formed. 18 This may correspond to the change of 0 0 at 36 Ma (Kennett and Shackleton 1976, Shackleton and Kennett 1975, Kennett 1977). The ice created a positive feedback system and continued to grow. Eventually ice overtopped the Transantarctic Mountains which may have been lower in elevation at that time. Ash dates show that between 13.6 and 14.8/15.2 Ma ice flowing northeast (i.e. across the grain of present outlet glaciers) overrode the Asgard Range and presumably the Dry Valleys (Marchant et al. l 993a). The ice was polythermal, with patches of erosive wet-based ice. Ash dates also constrain the ages of the Asgard, Jotunheim and Nibelungen Tills. These tills were deposited during the early phases of overriding, and were subsequently planed by the thick overriding ice (Marchant et al. 1993a). When the ice sheet had reached its maximum thickness, the atmosphere became colder and drier, leading to a negative feedback. The East Antarctic Ice Sheet shrank down to a size more near its present size. Thus, it seems likely that Sirius Group deposits stem from either the time when local alpine ice dominated or from a time of thick overriding East Antarctic Ice Sheet. Clasts exclusively from rock derived locally point to a local origin by alpine ice for some of the Sirius Group deposits. This is the case for, e.g., Table Mountain (Helfer unpublished data) and Mount Fleming (Faure and Taylor 1983, Stroeven et al. 1992, Helfer unpublished data). Sirius Group deposits from overriding ice are represented by deposits found on striated pavement running in a different direction than the present glacial drainage or containing far-traveled erratics. Minimum age constraints for the Sirius Group come from the timing of carving of cross-cutting valleys (the present glacial drainage system). Evidence that the deposits were 99

truncated after deposition can be found at Table Mountain (Barrett and Powell 1982), Mount Feather (Brady and McKelvey 1979), and Mount Fleming (Stroeven et al. 1992, 1994) (see also chapter Cl) and at Allan Hills (Schlachter unpublished data). In many cases, the Sirius sediments appear to lie in a glacial valley that runs in a direction which is not compatible with the present glacial valleys. This seems to imply deposition from ice flowing in a direction unrelated to the present direction and therefore older. If the Sirius sediments antedate the establishment of the present glacial system, then minimum ages of the Dry Valleys yield minimum ages for the Sirius deposits. The minimum age of the present Dry Valleys can be placed at 7-10 Ma (diatoms in the fjord sediments of Wright Valley; Brady and McKelvey 1987, Barrett and Powell 1982) and 1 S Ma (dates for the ash deposits that fill many patterned ground expansion cracks; Marchant et al. 1993a, 1993b). Direct dating of the Sirius comes only from the enclosed wood fossils (not found in the Dry Valleys), which are Oligocene based on ocean cores (e.g. Truswell 1986) and 10 from exposure dating. As we believe that the concentrations of Be we have measured are affected by a small amount of weathering, we conclude that the exposure ages we have calculated are indeed minimum ages. Weathering pushes the attainment of saturation to a younger age. That means that the rock with the oldest measured age has been affected by the smallest amount of weathering and more nearly gives the minimum age for the deposit. We therefore take the oldest age from each outcrop to be the minimum age for that deposit in general, which are: • the Sirius Group at Mount Feather: > 2.3 Ma, • the Sirius Group at Mount Fleming:> 5.8 Ma, • the plateau surface at Table Mountain:> 5.5 Ma, and • the Sirius Group at Table Mountain:> 2.9 Ma. We interpret the Beacon Sandstone plateau surface at Table Mountain to have been carved not much earlier than the Sirius Group was lain down on top of it (Fig. C4.5). Therefore both are likely to be older than 5.5 Ma. Although it was warmer in Antarctica in the Eocene, it has been cold and dry since at least sometime in the Miocene (Marchant et al. 1993a). Hall et al. (1993) concluded that there had been minimal slope activity in Wright Valley since 4 Ma. Marchant et al. (1993a) also came to such a conclusion, but for the last ca. 15 Ma. This is in agreement with the high 10Be concentrations we have measured showing that there has been little erosive or weathering activity since at least the earliest Pliocene. During the exposure time of the samples that we have measured (ca. 5 Ma), rock surface lowering rates have been very slow {< 70 cm/Ma). These conclusions hold for our sampling sites at the Beacon plateau surface at Table Mountain, and the Sirius Group surface at Mount Fleming and Mount Feather. But one may assume that most of the Dry Valleys region 100 experienced similar climatic conditions. In either case, cold conditions have persisted in 10 40 the Dry Valleys since 5.8 Ma ( Be), or since 15.2 Ma ( Ar/39 Ar ash dating), respectively (Denton et al 1993, Marchant et al. 1993a, 1993b, Sugden et al. 1995a). SE

sandstone samples from Beacon bedrock ( 9) 1.4· >5. 5 IVla 2000 m

granites from edge of Sirius G-oup (5)

sample on dolerite surface (1) ~-~~~~~~t¥~s~~~~~~~~~~~~fff~TITITff~rrnt:f:tnn~-:-'sandstone 2.9 IVla

1000 m

~,. ~" ~" ~" ~" ~,, ~ ~~ ~,,~">:Granitic basement>~">>~,,>~">>, ....,t / " ... ",. " / ./' :,;I ,,,.> ,,> ,, ,, \ ,,.. i' ,, ..../' > " >/ >...... ,, / / " " / ... / ./' ,, ,, ,,. •////...... ' ' .,)',,/j'/t",i)'//' ...... ' ' ' ' ... ' ...... /j'///////,/////-'//,1'/i'i'./'.!'i'///-1' ' ' ...... ' .... ' ' ' ' ' ... ' ' ' ... ' ... ' ' ' ' ' ' , ...... ,,,,,.,,,,,,.,,,,,,./////<"/,/' ,,,,,,"'"'''''''"''''''''"''''''''' ... /?,,./'/;l/./'/Ji'/,,1///,/',t'//i'/i'/ • ,, ' """ ,/ ,, ,, / i' ,, ,, ,, / ,, ,, !' i' / i' ,, ,, ,, ,, ,, ,. ,. / ,, ,, ,, ,,. ,,. "" ,/' ,,. / ,. ,/ , ,, ... ,, ... '''''''''''''''''''''''''''''''''"''''''''''...... ~ ...... ,,,,,,,,,,,,,, Gacier ,,,,,. /?.f• ,,.,,,,.,,jf//,///i'/i',,i'///£'<',I'/// ... ,,/ • ., ...... '·'·'·'~',.' ... ' ... ',',,'... ' .. '·'·' ' ... ' '·'·'

Fig. C4.5. Cross-section of Table Mountain showing the measured 1 Ose exposure ages, the number of samples measured is shown in parentheses. 102



Dl.1 INTRODUCTION Ice Age glaciations in Switzerland have been recognized by the extension of ice tens to hundreds of kilometers outside the boundaries of the Alps themselves, where they spread out as broad glaciers. Such events have therefore been recorded in glacigenic sediments which cover the foreland of the Alps. By distinguishing the separate and distinct events and by correlating coeval events, the sequence of cold vs. warm periods portrayed in Fig. 01 .1 has been constructed (Sch!Uchter 1988, SchlUchter 1988- 1989, Schlilchter 1989, Schluchter 1992). Wherever possible this system of classification has been constrained with absolute dating (up to now for the most part radiocarbon and to a lesser extent Uffh dating) and by correlation with the deep sea timescale (e.g. lmbrie et al. 1984. Martinson et al. 1987, Shackleton 1987). The classical four- break-down of the ice ages as outlined by Penck and Bruckner (1901/1909) was based on the system of terraces found in southern Germany. In their system, the most recent glacial deposits (WUrm) are associated with the Niederterrassen (lower terraces). The Hochterrassen (high terraces) were then correlated to the Riss Glaciation. Prior to the Riss occurred the Mindel and GOnz Glaciations which are represented by the younger and older Oeckenschotter (cover or blanket ) deposits, respectively. This four-fold classical view of Swiss glaciations is beginning to be replaced with a more realistic and as well as more flexible multi-glaciation system (SchlUchter 1988) (Fig. 01 .1 ).

02.2 CHRONOLOGICAL OVERVIEW Modern classification of major ice advances has as its cornerstone two reference points, deposits of the Last Glacial Maximum (LGM) and those of the most extensive glaciation (Schlochter 1988, SchlUchter 1992). The extent of the glaciers at each of these glaciations is portrayed in Fig. D1 .2. The LGM is represented by the freshest, most recent glacially-related deposits and landforms. Such deposits overlie previous older glacial (or non-glacial) deposits but are related to the terraces found in the deepest part of the system of terraces. In contrast, the most extensive glaciation was, as its name implies, an event where the glaciers were even more extensive than during the LGM. The position in the classification system of other glaciations can therefore be located with respect to these two reference geometries. 103


10' 000 I--"'-+--

Schlieren Last Glacial Maximum Killwangen (Late Wilrm) 28,000 lnterstadial: Gossau

60,000 Early Worm 120,000 Eem Interglacial 145,000 Penultimate Glaciation


Extensive Glaciations


Most Extensive Glaciation

>1.5 Ma Deckenschotter Glaciations (up to 8 glaciations)

Wanderblock Formation Plio.?

Figure 01 .1. Classification system for the glaciations of the Swiss Alps (SchlOchter 1988, 1992). 104

In viewing the glacial deposits found in the northern Swiss foreland, it is important to note that they were lain down upon an uneven topography {cf. Schluchter 1988, 1992). Underlying the glacial sediments are Miocene conglomerates and sandstones of the Molasse. The time gap between the Molasse bedrock and the unconsolidated overlying glacial deposits contained in the erosional unconformity is therefore millions of years. The oldest deposits found in northern Switzerland which have been attributed to glacial activity are known as the Wanderblock Formation (Hantke 1978-1983 and references therein). Outcrops are located in the region of Basel. Lithologically, erratics some of which are of Black Forest origin (e.g. Black Forest granites and Buntsandstein) are present in a sandy to clayey matrix. The deposits are thought to be very weathered till (e.g. Schluchter and Rothlisberger 1995) and to stem from the early Pleistocene (Hantke 1978-1983). No absolute ages are available. Fluvioglacial sediments comprise the Deckenschotter deposits. The Deckenschotter gravels were deposited in association with glaciations after the Wanderblock Formation glaciations but before the most extensive glaciation of the Swiss Alps. In Switzerland, outcrops of Deckenschotter mantling Molasse-cored highlands form the plateau topography (ca. 100- 500 m above the present valleys) between Zorich and the Rhine River. The gravels were preserved as the plateaus were cut by later more recent ice streams. Paleomagnetic (cf. Schlllchter and Wohlfarth 1993) and fossil evidence (Bolliger et al. 1996) indicate that the Deckenschotter glaciations may lie as far back in time as pre-Brunhes/Matuyama paleomagnetic reversal (>780,000 years) or even(>) 1.8 Ma (Graf 1993). Clearly, the age of these deposits is only loosely constrained. Indeed, as the Deckenschotter deposits may record up to 8 separate glaciations (Graf 1993), they must represent a significant span of time. Possible tectonic adjustments and/or base level changes marked the transition to the glaciations which include, among others, the upper and lower terrace deposits (of Penck and Bruckner 1901/1909) and encompass the classical Riss and Wurm glaciations. The glacial advance in which ice reached its farthest extent from the boundary of the Alps is informally termed the most extensive glaciation (Fig. 01 .1) (Riss in the classical sense). This glacier extent during the most extensive glaciation has been delineated by glacial erratics of Alpine origin which were deposited high in the Jura Mountains (e.g. Hantke 1977) as well as north of the Rhine. The Alpine erratics located north of the Rhine are found intermingled with erratics of Black Forest crystalline rocks. This attests to the likely joining of the Alpine and Black Forest ice caps during the most extensive glaciation (such joining probably did not occur during the last glaciation). Tills (as found for example near Mi:ihlin) which are located outside of the ring of moraines which define the Last Glacial Maximum are interpreted to be contemporaneous with the aforementioned erratic deposition. Presently, the age of the most extensive glaciation is 105 unknown. In the classical system, the Riss (Hochterrasse) glaciation deposits are defined as both the glaciation prior to the Wurm (vorletzte) as well as the most extensive glaciation of the Swiss Alps. Glaciations which occurred between the Last Glacial Maximum and the most extensive glaciation are revealed in the sedimentary sequences of the northern foreland (Schluchter 1988-89, 1992, 1993, Schli.ichter and Rothlisberger 1995). There are several till sheets which indicate that glaciers advanced outside the northern border of the Alps a number of times between the LGM and the most extensive glaciation. These are simply termed 'extensive glaciations' (grosse Vergletscherungen) and are still to be delineated and correlated in detail (e.g. reference sites: Seon, Rothenturm, Langenberg). For the penultimate glaciation (correlated to oxygen isotope stage 6), the reference sites include Meikirch, Thalgut and Jaberg. The existence of these glaciations between the LGM and most extensive glaciation, makes the meaning of Riss in the sense glaciation prior to the last one, somewhat ambiguous (Schluchter 1988). Glaciations belonging to the last glaciation (or Wurm in the classical sense) include the two advances of glaciers from the Alps post-dating the Eem (Last) Interglacial. The last Interglacial lasted from 145 ka to 120 ka based on ocean core, i.e. the SPECMAP time scale (lmbrie et al. 1984, Martinson et al. 1987, Shackleton 1987) which is corroborated by a U/Th date of 115 ,000 years from Gondiswil (Schluchter 1989, 1992). The timing of the two ice advances following the Last Interglacial has been particularly well investigated at the Gossau gravel pit (Schluchter et al. 1987), where an ice advance early in the last glaciation (early Wurm), and a more extensive later ice advance (LGMJ are separated by an interstadial which lasted from ca. 60,000 to 28,000 14C yr BP. The interstadial is represented by compressed peat or lignite beds (Schieferkohle). Within the mountains themselves, during the height of the LGM, glacial ice was constrained to stream valleys so that part of the base of the ice was strongly erosive. During the last glaciation especially in granitic regions, highly polished surfaces were formed by the basal ice. In some places an older generation of polish can also be noted. The relation in time of these two generations of polish is not known. It may be that the older upper generation is related to the most extensive glaciation (Schluchter unpublished data). The positions reached by the major glaciers of the of the northern Swiss foreland at the height of the LGM is shown by the bold line in Figure 01 .2. After the maximum had been reached glacier terminus fluctuations are marked by generally two recessional/readvance moraines still in the region of the maximum, and which have been given the names shown in Table 01 .1. But one must caution that absolute dating is relatively sparse, so the correlation of these individual stands is still not clearly established. Even though the LGM lies well within the range of radiocarbon dating, the 106

lack of vegetation during the cold periods has limited the use of radiocarbon. Indeed, some of the organic material found in the gtacigenic sediments may be reworked and actually date to the mid-'Wilrm' interstadial (Gossau lnterstadial).

Table Dl .1. Nomenclature (informal) for the three major stadia at the end of the LGM (modified from Hantke 1978-1983, SchlUchter and Rothlisberger 1995, Keller and Kravss 1987).

Solothurner-Rhone Li nth-Rhine Rhine

(Bannwil) Killwangen Schaffhausen Wangen a.d. Aare Schlieren (Brastenberg) ZUrich Stein am Rhein Solothurn (Rapperswil) Konstanz

Moraines marking the classical post-glacial readvances of Gschnitz, Clavadel, Daun and Egesen are found deep within the mountains in the Alpine valleys, that is tens to hundreds of kilometers behind the LGM maximum positions. But the time of deposition of these moraines is difficult to correlate with the cold events recorded in pollen stratigraphy. The fluctuations of Swiss glaciers during the Holocene have been fairly well constrained by radiocarbon dating and eventually through historic records or artifacts (e.g. Holzhauser 1984, 1985, Zumbuhl and Holzhauser 1988, Haberli 1994), though many gaps still remain.

01 .3 SUMMARY OF PART 0 After the analysis of samples from Antarctica, we were confident that 10Be and 26 Al could be measured reproducibly from quartz with relatively high concentrations of those isotopes. The next step was to show that we could as well measure samples with much 10 26 lower se and AI concentrations. With rock surfaces of short exposure times (e.g. 10,000 years) the ratios we would be measuring would be much closer to the limit of the sensitivity of the measurement technique. The crucial limitation is the background signal 10 measured by AMS, which also includes the amount of Be contained in a chemistry 10 14 blank. In detail, the AMS blank ratio for BefBe is on the order of 1o· • We strove to 1 have sample ratios more than ten times that background ratio (i.e.>> 10· \ As shown in our results presented in the following chapters, the success we have attained by using large amounts of quartz, even for rocks with very low cosmogenic isotope concentrations, is very encouraging. 107

In this part of the dissertation, we present exposure age data from three sites in Switzerland. From the Swiss Alps, we have exposure dated six boulders from an Egesen moraine at Ju lier Pass. We chose this as the first problem to address in the realm of short exposure times, because the correlation of Egesen with Younger Dryas is fairly well accepted (for references see Chapter 02). Because there is no other way to date a glacially polished surface, another site we have investigated is Grimsel Pass. The age of the polish was unknown. Indeed as it appears there are two generations of polish with unknown ages, although we have addressed only the younger generation. A third problem we have addressed was the age of the maximum extent of the Rhone Glacier as delineated by the moraines and associated erratics in the region of Wangen a. d. Aare. Timing of individual lobes is required to assess the synchroneity or time offset of the fluctuations of each ice mass. Finally, in Chapter DS, the landscape development at the peak of and following the LGM are discussed in light of our new exposure dates. Figure D1 .2. The two reference geometries for the extent of glaciations from the Swiss Alps (SchlUchter 1988,1992 and references therein). The bold line shows the extent of glaciation at the LGM, the gray line delineates the most extensive glaciation. 109


Susan Ivy-Ochs lnstitut fur Teilchenphysik, lngenieurgeologie, ETH Honggerberg, CH-8093 Zurich, Switzerland

Christian Sch!Uchter Geologisches lnstitut, Universitat Bern, CH-3012 Bern, Switzerland

Peter W. Kubik Paul Scherrer lnstitut c/o lnstitut fur Teilchenphysik, ETH Honggerberg, CH-8093 Zurich, Switzerland

Hans-Arno Synal Paul Scherrer lnstitut c/o lnstitut fOr Teilchenphysik, ETH Honggerberg, CH-8093 Zurich, Switzerland

)Org Beer EAWAG, CH-8600 DObendorf, Switzerland

Hanns Kerschner lnstitut for Geographie, lnnrain 52, Universitat Innsbruck, A-6020, Austria

1 Eclogae Geologicae Helvetiae (1996, v. 89, p. 1049-1063), reprinted with permission. 110

Moraine, 0 2 km stadial as indicated •fill Historic glacier extent + Boulder location

Figure 02.1. Map of Switzerland showing the location of Ju lier Pass and some of the sites discussed in the text. Schematic map of the Ju lier Pass region with the historic glacier extents and the moraines formed during Late Glacial glacier expansions (modified from Cornelius 1935/1950/1951, Suter 1981, Maisch 1995). The outer lateral moraine is the western band in the Egesen complex. Sampled boulder locations are shown as crosses. 111

02.1 ABSTRACT The Egesen moraine complex on the east side of Julier Pass has a mean exposure age of 11, 100 years. The exposure age is based on accelerator-mass-spectrometry- measured concentrations of 10Be, 16Al and 36CI in six different boulders. Three of the boulders lie along the sharp-crested outer lateral moraine and have a mean exposure age of 11,800 years. This moraine wall represents the glacier's response to the climatic collapse at the very beginning of the Younger Dryas. The other three boulders are located along a hummocky boulder band that cross-cuts the outer moraine along its terminal front. The exposure age of the inner moraine is 10,400 years.

02.2 ZUSAMMENFASSUNG Der Egesenmoranenkomplex ostlich des Julierpasses besitzt ein mittleres Expositionsalter von 11, 100 Jahren. Das Expositionsalter basiert auf mittels 0 Beschleunigermassenspektrometrie bestimmten Konzentrationen an ' Be, 26AI and 36CI in sechs verschiedenen Blocken. Drei dieser Blocke befinden sich auf der steilgeboschten ausseren Seitenmorane und zeigen ein mittleres Expositionsalter von 11,800 Jahren. Dieser Moranenwall wurde durch den Gletscher zur Zeit des Klimakollapses am Beginn der Jungeren Dryas gebildet. Die drei anderen Blocke befinden sich auf einem hockrigen Blockstreifen, der die aussere Morane entlang ihrer Stirnseite schneidet. Das Expositionsalter dieser inneren Moriine betri:lgt 10,400 Jahre.

02.3 INTRODUCTION During the last Glacial Maximum, coalescing piedmont glaciers, e.g., the Rhine and Rhone glaciers, almost completely inundated the northern foreland of the Swiss Alps with ice (Jackli 1970). Following attainment of their maximum extent at around 20,000 years ago, glacial ice vanished rapidly from the foreland in only a few thousand years (Schluchter 1988). Radiocarbon dated sediments from foreland lakes indicate completely ice-free conditions no later than 14,600 at Zurichsee (Fig. D2.1) (lister 1988) and 14,200 at Soppensee (Hajdas et al. 1993) radiocarbon years. The Oldest Dryas vegetation zone (Fig. D2.2) marks the influx of pioneer tundra-type vegetation (e.g. Burga 1988). Oldest Dryas sites, with minimum dates on the order of 14,000 to 13,000 radiocarbon years (Furrer et al. 1987 and references therein), located in valleys deep within the Alps, indicate absence of glacial ice in many alpine valleys by that time. The Oldest Dryas was followed by a period of time warm enough for pine trees, the B0lling/Aller0d, which was cut short by the cold event known as the Younger Dryas. The extremely rapid decay of the Alpine glaciers was punctuated by minor readvances which were very much confined to the Alps proper. These stadia are classified by the moraines which were formed, for example Buhl (Penck & Bruckner 112

1901/1909), Steinach (Senarclens de Graney 1958), Gschnitz (Penck & Bruckner 1901/1909), Clavadel (Maisch 1981), Daun (Penck & Bruckner 1901/1909, Heuberger 1966, Mayr & Heuberger 1968), and Egesen (Kinzl 1929, Heuberger 1966) (Fig. 02.2). The moraine sequence is fairly consistent from glacier to glacier and has been extensively mapped in both the Swiss and Austrian Alps (e.g. Heuberger 1966, Mayr & Heuberger 1968, Kerschner 1978, Maisch 1981, 1987, Suter 1981, MUiier et al. 1981, Muller 1984). General correlation from valley to valley has been accomplished with carefully established geomorphologic characteristics combined with the inferred drops in equilibrium line altitude (ELA) from the 1850 reference level (Gross et al. 1978, Maisch 1992).

14C Time Pollen Zones 1 Glacier Advances2 ELA (years) Cm)

9000 Ci +. ·o Preboreal "' Cl 10,000 a'. 0 - - - - -Kromer" - - 60-90 Younger Dryas Ege sen Boe kt en 100-150 11,000 Maximum 170-240 Allerod

12,000 (Older Dryas)

Bolling 13,000 0 «:i 250-350 .Q Daoo ?! (.I) 14,000 Oldest Dryas Clavadel? 380-470 _, t 15,000 * ? Gschnitz ?t 600-700 16,000

17,000 :Q () Cl 18,000 0

Figure 02.2. Summary of the late Glacial vegetation zones !'Mangerud et al. 1974), glacier re-advances and the corresponding ELA depression ('Maisch 1982). 113

The Egesen moraines are probably the most marked and easily identifiable of the series. They were first described by Kinzl (1929) in the Stubai Mountains to the southwest of Innsbruck (Austria). There, the type locality is at the Egesengrat and its surroundings near the Oresdner HOtte. Originally, Kinzl did not consider the Egesen stade to have been an important advance. later, in 1966, Heuberger showed that Egesen moraines are widely distributed and were formed during a prominent glacier advance, which was clearly separated from the earlier Daun stade. During the B0lling/Aller0d milder climate period, the glaciers probably retreated back to roughly the 1850 level (Maisch 1982). In the central Alps Egesen moraines are classified by their location at the ca. 200 m ELA difference (Patzelt 1972, Gross et al. 1978, Maisch 1987) and by their fresh and frequently blocky form in comparison to the more subdued older Daun moraines (Heuberger 1966). Geomorphologically, Egesen moraines dearly represent the last cold event (i.e. Younger Dryas) prior to the many advances of Postglacial order of magnitude, which characteristically reached only the "Little Ice Age" - extent (Patzelt & Bortenschlager 1973, Holzhauser 1984, 1987). Egesen moraines do not show any evidence of permafrost or solifluction overprint but retain their fresh and blocky morphology, depending on lithology. This is in clear contrast to the older moraines at similar altitudes, particularly the Daun moraines (Heuberger 1966) which were modified by periglacial processes following deposition as a result of the very cold conditions during the Younger Dryas. Egesen moraines are characterized by several nested and slightly cross-cutting moraines (Heuberger 1966, Kerschner 1978, Maisch 1981, 1987). In many places three, sometimes four distinct series of moraines can be found. They indicate that the climatic deterioration was at least three phased. Basis radiocarbon dates from bog deposits in the tongue areas of several Egesen age moraines in the Austrian Alps are in the early Preboreal time range (Patzelt 1972, Kerschner 1978, Bortenschlager 1984). In the Val de Nendaz, the last phase of the Younger Dryas could be found by pollen analysis in lake deposits within the moraines of the second phase of the Egesen stade (KOttel 1979, MOiier et al. 1981 ). Egesen moraines thus correspond most probably to the Younger Dryas. There are no radiocarbon dates from the Vadret Lagrev Egesen moraine at Julier Pass. For the older moraines, organic material is lacking and 14C dating has not been possible. Thus no precise time period can be assigned to any of these moraines except the Egesen. The earlier stadia can only be placed in time somewhere between the last Glacial Maximum and the Younger Dryas. 114

02.4 EXPOSURE DATING 02.4.1 Sample preparation Cosmic rays encountering an exposed rock surface induce nuclear reactions within 10 26 the mineral lattices yielding "cosmogenic isotopes", e.g. Be (! 112=1,500,000 years), Al 36 (!112=716,000 years) and CI (t112=301,000 years) (e.g. Lal 1991, Cerling and Craig 1994, Kurz and Brook 1994, Zreda and Phillips 1994). The concentration of the cosmogenic isotope, which is measured by accelerator mass spectrometry (AMS), is a direct measure of the length of time the geologic surface has been exposed. Equation 17 describes the growth of a radionuclide concentration with time in the mineral grains of the exposed rock surface:

N=-p-(1 e-(A.+pe/A)TJ (17) A.+ pe A where N is the number of atoms/gram, P is the local production rate in atoms/g•yr, T is the length of time the surface has been exposed in yr, A. is the decay constant of the 1 3 radionuclide in yr" , p the rock density in g/cm (2.7 for granite and granodiorite), Ethe 2 2 erosion rate in cm/yr, and A is the cosmic ray attenuation length in g/cm (150 g/cm ; Brown et al. 1992). 10 26 Quartz is well suited to Be and Al exposure dating studies because it has a low 26 7 inherent Al content allowing measurement of Al/2 Al. Its tight crystal structure makes it a closed system to either loss or gain (other than in situ production) of the isotope of interest (Kohl and Nishiizumi 1992). We sampled the top surfaces of the largest boulders on the crests of the moraines (more than 1 meter higher than the surrounding sediment) by using a hammer and a chisel. Boulders with obvious signs of spalling or rolling over were avoided. Altitudes were determined from the topographic map and with an altimeter. The boulders were predominantly petrologic of the Ju lier Granite (Buhler 1983} including granodiorite ()8, Jl 0, Jl 04, )18), aplite (Jl 2), and diorite (J15). Thin section examination indicated an average quartz grain size of 0.5- 0.7 mm. The rocks are fairly fresh, with the highest degree of weathering in the feldspars (saussuritization), yielding the greenish cast typical of the Ju lier Granite. Ca. 0.5-1 .0 kg of each rock sample were crushed and sieved. The <0.4 mm size 36 fraction was used for CI sample preparation from whole rock, while the 0.4-1 .00 10 26 fraction was used for the separation of quartz to be used for Be and AI. A pure quartz fraction is gained by selective chemical dissolution (Kohl and Nishiizumi 1992). With this method the weak HF solution dissolves the feldspars and micas faster than the quartz, 115 purifying the quartz fraction. We use at least five HF steps so the quartz fraction is very pure, reflected as well by the low Al concentration (e.g. 79 ppm in Jl 8). The elegance of this method is that any contaminating meteoric 10Be from the atmosphere, which can be several orders of magnitude greater than the in situ component, is etched away from the surfaces of the quartz grains (Kohl and Nishiizumi 1992). 9Be carrier was added to the dried and weighed pure quartz (40-70 grams for each sample), which was then completely dissolved with concentrated HF. Be and Al were separated using cation exchange. The hydroxides were precipitated then baked to the oxides, Alz03 and BeO, at 850 °C. 10BefBe and 26Al/27 Al were measured by AMS at the ETH/PSI tandem facility in Zurich. Stable Al was measured by ICP-AES (inductively-coupled plasma atomic emission spectroscopy) using three standard additions for each sample. The <0.4 mm size fraction of the crushed whole rock was used for 36CI sample preparation. It is important to note that the exact same crushed rock sample was used for all three isotopes. Sample preparation for 36CI followed the method of Zreda (1994). To leach any possible meteoric 36CI, the <0.4 mm fraction was left overnight in a solution of weak HN03, then rinsed at least five times with 18 MO water. On the order of 70 g of dry whole rock were placed in teflon bottles with concentrated HN03 and HF. The bottles were kept in a 90 °C water bath until the rock had dissolved completely, which usually took several days. The contents of the bottle were centrifuged cold (10 °() in teflon centrifuge bottles in order to separate reprecipitated silica gel from the dissolved rock found in the HF solution. The liquid was decanted into teflon centrifuge tubes, then

AgN03 was added to the HF solution to precipitate AgCI. The solution was left at least 24 hours to allow nucleation of the AgCI. After centrifuging and decanting, the AgCI was redissolved with NH40H and unwanted hydroxides were precipitated, then separated with centrifuging. Ba(N03)i was added to the solution to remove sulfur by formation of 36 Ba504 ( 5 is an interfering isobar in AMS measurements) and left again at least 24 hours.

The BaS04 was removed by filtering, then the AgCI was reprecipitated, dried and pressed into tantalum target holders. 36Cl/CI ratios were measured by AMS at Zurich. Major elements, Gd, Band Cl were measured using XRF, prompt gamma neutron activation and ion selective electrode, respectively (Table 02.1 ). These elements were all measured on the rinsed sample material.

02.4.2 Exposure age Calculations Ages are calculated by solving Eq. 17 for time (T), and assuming no erosion of the rock surface ( =0):

(18) 116

10 26 We have used the production rate of 6.0 Be and 36.7 AI atoms per gram Si02 per year at high latitude (>60°) and sea level for the age calculations (Nishiizumi et al. 1989). 10 26 The uncertainty of the Be and AI production rates have been estimated at less than 10% (Masarik & Reedy 1995) and 5-7% (Lal 1991 ). For 36CI there are three main production mechanisms, i.e. spallation of Ca and K to 35 36 form 36CI and neutron capture by CI to form CI (e.g. Zreda and Phillips 1994, Zreda 1994). Therefore, the production rate must be calculated individually for each rock composition. We have used the elemental production rates determined by Phillips et al. 36 36 (1996a). They are 1280 CI atoms per kg rock per wt. % KzO per year and 530 CI atoms per kg rock per wt. % CaO per year (high latitude and sea level). The errors have been estimated at 9% (Zreda 1994). 36 35 In order to calculate the production of CI due to neutron capture by CI one must first calculate the fraction of neutrons absorbed by the rest of the rock. We have calculated this fraction using the measured major element concentrations as well as the concentrations of B and Gd (Table 02.1) which are significant neutron absorbers and the cross section data from Zreda (1994). The fraction available for production of 36CI from Cl is then multiplied by the neutron flux, 313,500 neutrons per kg rock per year (Zreda 1994) and the measured Cl concentration for that rock. The resulting 36CI production 36 rates for each rock range from ca. 11 to 14 CI atoms/gram rock per year. We have 36 corrected for subsurface (non-cosmogenic) production (Fabryka-Martin 1988) of CI based on the measured U and Th concentrations in J10. In all cases, this amounts to a correction of less than 2 %. The production rate for each of the three isotopes must then be scaled to the latitude (Ju lier Pass 46° 30') and altitude of the samples (Lal 1991 ). The uncertainty of the scaling is thought to be less than 10% (Lal 1991 1 Masarik and Reedy 1995). We have made two additional corrections, one for shielding and one for the thickness of the sample analyzed (Table 02.2). The production of cosmogenic isotopes is dependent on the cosmic rays that the rock surface actually sees. Therefore, reduction of the cosmic rays by ridges and mountain peaks directly surrounding the sampling site reduces the production rate accordingly. To the west and east from Ju lier Pass the shielding was zero, while the ridges related to Piz Goglia to the north of the moraine and Piz Lagrev to the south (Fig. 02.1) shielded the horizon by roughly 25 degrees. The precise angle to the horizon and the number of degrees of the quadrant which were affected were then used to calculate (Nishiizumi et al. 1989) the production rate reduction. This correction was only 3.5 % for all samples. For boulders whose original surface was dipping, the shielding corrections which includes the dip angle as well, were 117 also calculated. This only affects Jl 5 (32° dip to the southeast} making the total reduction of the production rate due to shielding 5.2 % for J15. The atoms which are produced in a rock surface falls off exponentially with depth due to the attenuation of cosmic rays penetrating matter. The production rate for each sample must be corrected based on the thickness of the piece of rock which was crushed 36 and dissolved (Table 02.2). For CI production, spallation components (production of 36CI from Ca and K) have been ccirrected for separately than neutron capture (based on Fabryka-Martin 1988 and Zreda 1994). No correction has been made for erosion of the surfaces, estimated to be small based on geomorphic appearance, for example lack of evidence of spalling. On the other hand, a fine-grained till matrix may have covered the boulders right after moraine formation. We hoped to avoid this problem by sampling boulders on the crests of the moraines. No correction has been made for intermittent snow cover. The position of the boulders on top of the moraine allows much of the snow that falls to be blown off. In recent years winter snow cover on these boulders has been nil.

Table 02.1. M a1or· an d minor e ement concentrations or t h e roe ks analyze Element 110 )12 )15 weight%

Si02 73.5 77.5 40.4

Al203 11.3 11.5 18.1

Fe,03 3.91 1.52 15.6 MnO 0.07 0.02 0.16 MgO 1.17 0.43 6.6 Cao 1.85 0.94 10.2 Na,O 1.94 2.99 1.71 I K,O 3.9 4.04 0.72

Ti02 0.38 0.12 2.09 P·Os 0.1 0 0.06 total 98.12 99.06 95.64 ppm B 13.5 10 13 Gd 5.5 12 4.5 Cl 189 115 213 u 2.8 Th 6.7 10 Table 02.2. AMS-measured concentrations of Be, 26AI and 36CI and calculated exposure ages.

Boulder lithology Thickness Altitude Isotope Atoms/gram Age Mean Moraine age

No. (cm) (m) (years) boulder age

5 Jl 8 granodiorite 4 2210 toBe 3.50 x 10 * 11,650±910 11,800 ± 840 Outer moraine: 6 26AI 2.29 x 10 * 12,640 ± 2150 11,820 ± 500 5 5 diorite 2 2195 36CI 6.84 x 10 11 ,490 ± 1 JSO 11,490 ± 1150 ··~· 10 5 ]12 aplite 4 2200 Be 3.60x 10 * 12,060 ± 1550 12,210±880

5 36CI 7.72 x 10 12,270 ± 1060

5 ]104 granodiorite 12 2185 to Be 2.86 x 10 • 10,390 ± 760 10,390 ± 760 Inner moraine:

5 J10 granodiorite 8 2185 ioBe 2.87 x 10 10,050 ± 770 10,480 ± 620 10,380 ±400 5 J&CI 8.67 x 10 11,260 ± 1030

5 ]8 granodiorite 6 2175 to Be 2.96x10 10,260 ± 700 10,260 ± 700

• data from Ivy-Ochs et al. 1995. 119

02.5 THE EXPOSURE AGE OF THE JULIER PASS EGESEN MORAINE COMPLEX We have determined nine exposure ages from six different boulders located on the crest of the Egesen moraine complex at Ju lier Pass (Table 02.2). The errors listed are the 1<1 measurement errors. They include both the statistical (counting) error and the error due to the normalization to the standards and blanks. A 5 % sample reproducibility error has been included in the error on each exposure age. This is based on the reprocessing of several different rock samples (see Ivy-Ochs 1996 for details). The 36CI age includes the 26 error from the total Cl measurement (8 %) and the AI age the total Al measurement (1%). 10 26 Exposure ages have been determined for sample J18 with both Be and AI from a quartz mineral separate. The two ages agree quite well within the given errors. We have only measured 36CI in sample )15 because it contained very little quartz. Samples J10 and 10 36 )12 have both Be and CI exposure ages which also show good agreement. The 36 10 agreement is noteworthy since CI was extracted from a whole rock sample, while Be was measured in quartz. Additionally, the published production rates for these three isotopes have been determined quite differently. Production rates for 10Be (and 26Al) were determined in a glacially polished surface where the age of the surface was 36 inferred based on regional correlations (Nishiizumi et al. 1989). While the CI production rates were calculated based on a suite of rock samples from several different sites and with ages ranging from 2,000 to 50,000 years (Phillips et al. 1996). For 10Be and 26AI we have used the presently accepted production rates (Nishiizumi et al. 1989) but recognize that reevaluation of these production rates may be called for {Clark et al. 1995). Based on our Ju lier Pass data, one might conclude that in this time range, the published production rates for all three isotopes and the latitude/altitude scaling factors are probably less than 10 % off. The exposure ages we have measured range from 10,050 to 12,640 years. When more than one isotope was de1:ermined from a single boulder, its exposure age was obtained by calculating the weighted mean {weighted based on the respective errors from each measurement). The overall mean age of the moraine complex is 11, 100 ± 700 years. The terminal zone of the: Egesen moraine at Ju lier Pass consists of several cross- cutting boulder bands and a tongue area filled with blocks, all apparently having been deposited relatively closely spaced in time. The outer lateral moraines are the highest and sharpest features. Their well-defined shape and considerable size indicates an abrupt glacier advance due to rapid t,emperature drop where the glacier remained in its maximum position for a rather long period of time. Boulders )18, J1 5, J12 lie along the crest of the sharp outer lateral moraine. This moraine hooks around to the east and is cross-cut by a later moraine made up of an ill-defined band of boulders (Fig. 02.3). Samples )8, J10 and J104 are from this inner moraine. The exposure ages from the three 120

Figure 02.3. Air photo of the te1·minal zone of the Vadret Lagrev glacier (east side of Ju lier Pass) during the Egesen stade. The outer moraine is found in the center foreground (photo: C. Schluchter). 121 rocks from the outer moraine give an average of 11,800 ± 500. For the three rocks from the inner moraine 10,400 ± 400 is obtained. The exposure ages are in line with the field data, as they indicate that the inner moraine is slightly younger than the outer moraine. The beginning of the YouPger Dryas was marked by an extreme cold snap (for example seen in 0180 in GRIP, Dansgaard et al. 1993). It may be that this correlates to the sharp outer moraine wall. The later inner moraine can actually be interpreted as a transitional form between a moraine and an ice-rich rock glacier, which, under permafrost conditions, might also have contained some remnant glacier ice. In any case, it is a typical feature of ice-rich permafrost. These landforms are characteristic for the later phases of the Egesen stade at small glaciers in the continental parts of the central Alps. They suggest that climatic conditions remained cold and became successively drier during the Younger Dryas. As a result, glaciers were starved and wide-spread permafrost features developed (Kerschner 1 982, 1985). Detailed field mapping combined with further exposure dating will shed more light on these complex relationships. As these boulders are most likely supra-glacial material originating from rockfalls onto the glacier surface, prior exposure is a definite possibility. This would happen when the rock has been exposed to cosmic rays either along the valley wall or cliff. Because of the coherence of the six boulder ages, we rule out prior exposure for the surfaces we sampled. A rock that had been exposed earlier would stick out as an older outliner.

02.6 COMPARISON WITH OTHER YOUNGER DRYAS RECORDS Our nine exposure ages, along with a band showing the approximate calendar age of the Younger Dryas, are pres·ented graphically in Fig. 02.4. Calendar years in ice cores are based on layer counting (e.g. annual changes in 0180 and dust) (Alley et al. 1993, Johnsen et al. 1992). At present the calendar boundaries of the Younger Dryas are slightly shifted to the older in ice cores as compared to several terrestrial records (Hajdas et al. 1995) (for example, Soppense•:;, Switzerland, Hajdas et al. 1993, Lake Van, Turkey, Landmann et al. 1996). In Switzerland, the beginning and end of the Younger Dpyas have been very precisely determined by AMS 14C dating of macrofossils from the small lake Soppensee (Hajdas et al. 1993.). Using the varve time scale as calendar time, the Younger Dryas as defined by the pollen signal, lasted from roughly 12, 100 to 11,000 calendar years ago. This is shown by the shaded band in Fig. D2.3. The cross-hatched band is for the Younger Dryas in ice cores. Given the uncertainties in the dating for each of these records, our moraine age of 11, 100 years corresponds quite well. Indeed, we are comparing very different signals (Broecker 1992) which may have different response times, i.e. pollen changes (in Soppensee), temperature change reflected by o18 0 shifts (in the ice cores) and a glacier's advance as recorded by a moraine, but they are all the result of the same marked drop in temperature. 122

. J 04 t-: 0 10Be 0 3sc1 <> 2sAI

0 0 c::> 0 0 0 0 0 0 0 0 c::> 0 0 0 0 0 0 0 q C::> 0 0 0 0 0 0 c::) ,... C') ..,; (0 co O> ,- ,... ,...C\i ,... ,... ,...LO ,...

Time (years)

10 26 36 Figure 02.4. Plot of Be, A! and CI exposure ages determined for six boulders from the Egesen moraine complex at Ju lier Pass. The shaded band represents the Younger Dryas as determined by varvE counting in sediment cores from Soppensee (Hajdas et al. 1993); the cross-hatched area shows the Younger Dryas as recorded in ice cores (Alley et al. 1993, Johnsen et al. 19S2). 123

Also contained in Table D2.3 are the ages of other moraines interpreted to correspond to the Younger Dryas which have been dated using cosmogenic isotopes. The Llyn ldwal moraine (Scotland) has been classified as Younger Dryas based on 36 radiocarbon dating (Phillips et al. 1994). We can see that the CI exposure ages for both the Llyn ldwal moraine (Phillips et al. 1994, 1996) and our Julier Pass Egesen moraine agree quite well, even though in each case the data set is small. Striking agreement is as 0 well obtained for the ' Be exposure age for the Inner Titcomb Lakes moraine in the Wind River Mountains (Wyoming, U.S.A., Gosse et al. 1995) and our Julier Pass Egesen moraine. Since in both cases the same production rates (Nishiizumi et al. 1989) and the same scaling factors (Lal 1991) were used, the two geographically widely-spaced events can be said to be contemporaneous. Indeed, the synchroneity of the time of construction 10 of the two moraines is proven by the coincidence of the measured Be concentrations at each site {Ivy-Ochs 1996). Thi:; is true irregardless of the 'actual' 10Be production rate. This indicates that the measurement of cosmogenic isotope concentrations is a powerful tool for determining synchroneity of glacial events world-wide. In this way one could circumvent the problem of calibration of the radiocarbon time scale (Hajdas et al. 1995) outside of the tree ring curve (Kromer and Becker 1993).

02.7 CONCLUSIONS The moraine complex we investigated at ju lier Pass provided an ideal test. It was classified as Egesen based its morphology and its ELA drop. Radiocarbon dates (though not specifically at this site) ind cate that Egesen moraines formed during the Younger Dryas. Therefore, the age of the Egesen complex at Julier Pass was generally accepted to be Younger Dryas, but until now there has been no method of directly dating it. Our exposure age of 11, 100 from six different boulders confirms that this moraine complex formed during the Younger Dryas. Additionally, we were able to unravel spatially and temporally close events such as the two events of this double-walled moraine. The outer moraine formed at the very onset of the Younger Dryas at 11 ,800; while the inner, more diffuse, moraine formed during the down-wasting and starving of the glacier at 10,400 years ago. Alpine glaciers do respond to marked cold pulses in climate and do react rapidly (lag times on the order of a hundred years, Haberli 1994). Thus, they are very sensitive climate indicators and do leave a precise record of past climatic down-turns. The method of exposure dating of moraines now allows establishment of precise for this most important terrestrial record allowing further regional and worldwide correlations (or not) of cold climatic intervals. Indeed, synchroneity of construction time of glacial features can be evaluated solEly on the basis of comparison of cosmogenic isotope inventories. Table 02.3. Comparison with severa other Younger Dryas records. Site Signal Beginning End Dating Reference method Soppensee, pollen change ca.10,700-11,000 ca. 9900-10,200 Hajdas et al. 1993 Switzerland Soppensee, pollen change -12,125 -11,000 varve Hajdas et al. 1993 Switzerland counting GiSP 2 12,940 ± 550 11,640 ± 250 iayer wu11li11g Alley el al. i 993 GRIP 12,700 ± 100 11,550 ± 70 Johnsen et al. 1992 36 Llyn ldwal, glacier advance 11,600 ± 1300 Ci # Phillips et al. 1994 Scotland moraine(s) 12,900 ± 2000 formed Titcomb lakes, glacier advance 11,000 ± 770 Gosse et al. 1995 Wyoming, U.S.A. moraine(s) formed Ju lier Pass, glacier advance 11, 100 ± 700 this work Switzerland moraine(s) (11,800 ± 500) (1 0,400 ± 400) formed #two rocks analyzed. *ten rocks analyzed. 125


03.1 CROSS-SECTION AT GRIMSEL PASS Grimsel Pass, located in the Berner Oberland, lies between the valleys of the Aare and the Rhone Rivers (Fig. 03.1 ). In periods of extensive glaciation, the Grimsel region was the source area for both the Aare and the Rhone Glaciers. Uniquely, during the LGM, the two glaciers which originated only hundreds of meters from each other, travelled separate paths for hundreds of kilometers only to rejoin in the foreland near Bern (see Fig. 04.1). The region around Gri msel Pass exhibits some of the most spectacular glacial polish found in the Swiss Alps. if not the world. Glacially-carved forms are well exemplified in the generally massive Aare granite. One example is the Hellen Platten, located in the Aare River vallE'Y just to the north of Grimsel Pass. This highly-polished broad flat rock surface still retains the carved footholds of the medieval pass crossing. The present cross section across the valley of the Aare River can be viewed as made up of two generations of polish (SchlGchter unpublished data). Below the trimlines found along the present valle•1 walls the most recent generation of glacial polish is found. It covers the floors and sides of the present Aare River valley, where stoss and lee morphologies such as roche moutonnee and flyggberg forms (e.g. Sugden and John 1976, Summerfield 1991) dominate. The topographically higher areas above the break in slope of the valley walls display what can be interpreted as an older generation of polish. Although the upper surface does not have the high gloss noted in the younger polish, stream-lined features indicating glacial molding are retained. One model suggested for the development of this landscape is shown is Fig. 03.2 (SchlUchter unpublished data). Prior to the onset of the Ice Ages, denudation of the Alps was dominated by uplift and associated fluvial down-cutting, where steep-gradient V- shaped valleys were formed (Fig. 03.2a). During an ice age prior to the LGM (the most extensive glaciation ?), stream valleys were filled with apparently broad, shallow ice masses (Fig. 03.2b) in comparison to the LGM ice streams. The ice cut into the mountain slopes and the rough form of the nunataks took shape. Broad U-shaped valleys were carved (Fig. 03 .2c). Followi "lg the glaciation(s), erosion by streams cut into the valley bottoms (Fig. 03.2d). As the next ice age began the deeper stream valleys became once again filled with ice. The valleys constrained the ice into thick and therefore more erosive ice streams (Fig. 03.2e). Since the vanishing of the LGM glaciers, stream action once again has dominated landscape modification (Fig. 03.21). 126

[ill Central Aare Qanite LGM ice limit G-imsel Q-anodiorite

Sout hem Aare Qanite

Alt krist allin Gleiss Schiefer Zwischenzone

Figure D3.1. Simplified of the Grimsel Pass region (modified from Stalder 1954, Labhart 1977) indicating the general location of the trimlines associated with the LGM ice extent (modified from Jiickli 1962, Hantke 1980). 127

It is difficult to assign ages to the events prior to the last glaciation, before the development of exposure datin~,, only relative dating had been available, The older generation of polish may indeecl originate from the time when Alpine glaciers extended farthest north from the Alpine front, (most extensive glaciation or in classical terms the 'Riss Glaciation'). In the foreland, the lateral extent of the ice masses during the most extensive glaciation is delineated by erratics which are outside of the LGM ice limits. Such erratics are also located at higher elevation than LGM erratics, which indicates the presence of a thicker ice mass in the foreland than during the LGM. However, for that time period it is difficu It to define the shape and thickness of possible local ice caps over the Swiss Alps.







Figure 03.2. Schematic representation of the evolution of the landscape at Grimsel Pass. 128

The present elevation of erratics which were left stranded by the LGM ice in the Grimsel Pass region, e.g. 2500 min the area around Juchlistock, delineates the maximum ice thickness attained by the ,!\are Glacier Oackli 1962, Hantke 1978-1983). At that time, to the south of Grimsel Pass, Rhone ice filled the Rhone valley to such an extent that ice backed up and flowed over Grimsel Pass and contributed to the ice mass of the Aare Glacier (e.g. Jackli 1962). In the region between the present Rhone Glacier and Grimsel Pass, chatter marks and crescentic gouges show that ice flowed over the pass from south to north (Schluchter unpublished data, Jackli 1962, Hantke 1978-1983). The Grimsel region was thus one of the most important centers of ice build-up and outflow in the western Alps at the height of che LGM. Exposure dating of the glacially polished surfaces can provide constraints for the timing of the ice build-up and breakdown. An alternative explanation for the morphology exhibited at Grimsel Pass is that the break-in-slope morphology represents an englacial thermal boundary as has been postulated for trimlines in Norway (e.g. Mccarrol and Nessje 1993, Brook et al. 1996). In such a model, the ice was thicker, temperate-based, below the 'trimlines' and therefore more erosive. Above the 'trirnlines' the ice was thinner, colder and in part frozen to its bed, which protected the underlying rock surface. In such a scenario, the ice was not restricted to the troughs below the trim lines. Sampling for exposure dating of the upper erosional surface is planned in order to investigate the possible alternative origins for that surface.

03.2 BEDROCK GEOLOGY AND SAMPLE DESCRIPTIONS The bedrock of Grimsel Pass belongs to the Aare Massif, which together with the Gotthard Massif, makes up the crystalline core of the Swiss Alps. These granitic to granodioritc rocks (Stalder 1964, Labhart 1977, Schaltegger 1990) were intruded into a metamorphic basement complex around 300 Ma. At ca. 40 Ma, Alpine Metamorphism resulted in mineralogical changes to the level of greenschist facies (e.g. Frey et al. 1980). The lithology found directly in the pass area is dominated by the Grimsel Granodiorite (Fig. D3.1 ). Its northern contact is gradational to the Aare Granite. The southern contact is complicated by a roughly east-west trending sliver or screen of the preintrusion metamorphic rocks, known as the Altkristallin or Gneiss-Schiefer Zwischenzone (GSZ) (Stalder 1964, Labhart 1977). Then to the south of this zone, the southern facies of the Aare Granite is found. Locally, the Grimsel Granodiorite exhibits strong foliation and may be termed a granodioritic or augen gneiss (Stalder 1964, Labhart 1977). The greenschist facies metamorphism which accorr panied Alpine tectonism resulted in formation of muscovite, chlorite, and epidote, among other minerals. In such deformed rocks, quartz is found predominantly as lenses of fine-grained mosaic quartz (so-called sand quartz, Stalder 129

1964). Individual quartz grains rarely reach 0.1 mm diameter. In less deformed areas of the Grimsel Granodiorite, the original magmatic texture is more recognizable and quartz grains can be larger, often up to 0.4 mm. A series of 13 samples were taken from rock surfaces below the trimlines at Grimsel Pass. All of the samples except G11 O were from glacially polished surfaces. Lithologically, all were either Grimsel Granodiorite or cross-cutting quartz veins. Details on the lithology of the samples we have analyzed are found in the thin-section tables of Appendix H2. Sample G1 was taken from the south side of a roche moutonnee due west of the pass. Rock samples G102, Gl 03, and G106 are from a single flat-lying rock surface which is part of the roche moutonnee landscape located between Grimsel Pass and the present Rhone Glacier. G102, G103 and G106 were from thin quartz veins (ca. 1-3 cm thick). Gl 11 was from a ca. 30-50 cm thick quartz vein comprised of large (up to several cm across) crystal-clear single crystals. The Gl 11 sampling site lies between the nunataks and Grimsel Pass proper, but at lower elevation than the G102/G103/G106 site. Rock sample G11 O was taken from the steep side of a nunatak overhanging the Rhone Glacier. Sampling sites are shown in Figure 03.1.

Ta bl e 03.1. Roe ks samples prepared rom Grimsel Pass area.

Rock lithology Altitude Thickness Shielding number (m) (cm) correction (%)

G1 11.ranodioritic 11.neiss 2180 5 * G102 quartz vein 2660 2 98.9 G103 quartz vein 2660 2 98.9 G106 quartz vein 2660 2 98.7

G110 augen i~neiss 2860 4 91.0 Gl 11 quartz vein 2400 8 92.5 *not calculated because this sample was not measured successfully.

03.3 DETAILS ON SAMPLE PR:EPARATION Six of the 13 collected samples were prepared and analyzed. Samples Gl, G102, G103, G106, Gl 10, and Gl 11 were prepared for measurement with 10se and 26AI, while Gl and Gl 1O were also prepared for 36CI measurement. As discussed below, difficulties 10 in the sample preparation led to only six successful AMS measurements (four Be measurements and two 26All. 130

Preparation of sample G102 was hampered by the presence of epidote. The epidote grains did not dissolve during the selective dissolution steps (purification of the quartz separate with weak HF). Using a binocular microscope, the larger green, columnar epidote crystals were picked out, but some fragments were too small to be removed. These fragments of epidote were subsequently dissolved with the quartz separate during the HF digestion step (with concentrated HF). Thus, a tiny amount of epidote contributed Fe, Ca, Mg, and Al in quantities which interfered with the Be and Al purification (see also Ochs and Ivy-Ochs 1996). After the problem with the epidote was noted, a heavy liquid step was added to the sample preparation procedure (see Chapter 82). Thereafter, Gl 03, Gl 06 and Gl 10 were measun~d successfully for 10Be with AMS. A whole rock sample of Gl was prepared for 36CI measurement. There was not enough AgCI after processing to measure with AMS. A quartz mineral separate from Gl 26 was dissolved for 10Be and AI measurement. From the resulting solution, we tried several times with various chemical extraction steps to purify the Be fraction. The resulting BeO was measured in four different '°Be runs, but very low current was obtained from the target material each time. This means that the Be had not been purified enough and/or too much of the Be had been lost during the various purification steps. Sample G1 was taken only 20 meters from one of the shear zones. Deformation of the granodiorite was evident in thin section. Quartz is present only as a mosaic of fine- grained quartz with no large crystals. Thin bands of muscovite wrap around the quartz grains and occur at the margins of the quartz lenses. It proved difficult to separate the fine-grained quartz from the intergrown muscovite and fine-grained epidote. As a result, the quartz mineral separate wzs apparently not pure enough. Although a heavy liquid step had been used to separate epidote, further mineral separation techniques clearly should be employed. Both mu!>Covite and epidote contributed Al, Ca, Fe and to a lesser extent Mg to the sample solution hampering Be and Al purification. Gl 10 was a rock similar in composition to Gl, but fortuitously there were rare large (0.4 mm) quartz grains present, probably because the rock was less deformed than Gl. It was necessary to use a large amount of crushed rock (on the order of 1 kg) because the large quartz grains were uncommon. Notably, G11 O had less muscovite than Gl. Even so, it was necessary to run the sample solution through the cation exchange column (20 ml of resin) three times to purify the Be fraction. Additionally, Gl 10 was prepared for 36 CI but no current was obtained from the extracted material.

03.4 THE EXPOSURE AGE OF THE YOUNGER GLACIAL POLISH AT GRIMSEL PASS The exposure ages of the samples we have measured range from 8270 to 11,560 years (Table D3.2). Very good agreement has been obtained between 10Be and 26AI when 131 measured in the same sample (Gl 06 and Gl 11 ). Several scenarios are possible to interpret our exposure ages and their meaning for the deglaciation of Grimsel Pass: 1. the oldest exposure age is c:losest to the actual time of deglaciation of Grimsel Pass, 2. most of the pass area was actually not ice free until the Preboreal, or 3. after the glaciers retreated, some of the sampled rock surfaces were covered by snow or sediment.

Tab e 03.2. AMS-measure dlOBe an A concentrations rom Grimse Pass samp Ie s.

Rock Lithology Isotope Al atoms/gram Exposure age number (ppm) SiO, (years)

G103 quartz vein 10Be 3.75 x 105 8940 ± 670 5 G106 quartz vein rnBe 3.46 x 10 8270 ± 600 6 26AI 11.6 2.23 x 10 8870 ± 640 10 5 G110 augen gneiss Be 3.76 x 10 8750 ± 640 Gl 11 quartz vein 10Be 3.43x105 10,990 ± 900 6 26AI 108.0 2.17 x 10 11,560 ± 2300

These relatively young exposure ages imply that all of the exposure occurred after the Last Glacial Maximum. This indicates that the ice which carved this landscape during the LGM was indeed thick and erosive enough to remove all or most of the rock which had been exposed before glaciation. For the spallation cosmogenic isotopes (i.e. 10Be and 26AI) the production rate at a point 2 meters into the rock is less than 3 % of its surface value. This depends not only on how much of the rock has been removed by the glacier, but how long the glacier covered the surface and how long the surface was exposed prior to glaciation. This has been &;cussed in more detail in Chapter B (section Bl .3.2). For the younger generation of glaciallv polished surfaces from this region of the Swiss Alps, the presence of inherited cosmogenic isotopes can probably be ruled out. This may not always be the case. For example, studies in the Sierra Nevada Mountains, California (Nishiizumi et al. 1989, Stone et al. 1996), and in the Wind River Range, Wyoming (Gosse et al. 1995a, 1995b), indicate that subglacial erosion may at times be inadequate to re-zero some rock surfaces. Based on the 10Be and 26AI exposure ages for quartz vein Gl 11, Grimsel Pass may have already been deglaciated, therefore, somewhere between roughly 14,000-10,000 years ago (including the lcr errors). Notably, the G11 ·1 surface was the lowest elevation sampling site, though itself still more than 200 m higher than the pass proper. This may mean that the landscape directly at the Pass was ice-free even earlier than the Gl 11 132 surface, especially if we consider local ice tongues as the mechanism of glaciation. Additionally, the Gl 11 quartz vein surface was extremely polished although the enclosing granodiorite countr)' rock had a rougher surface. This may mean that this rock surface had not been protected during exposure. Highly polished rock sJrfaces can often be found underneath a patch of soil or sediment where they would have been protected and preserved, In that light we note that the Gl 03, Gl 06 quartz veins we sampled did not stand in relief above the surrounding rock. That may mean that they were protected and had not been exposed to erosion. In contrast, in the area just west of Grimsel Pass, glacially-polished quartz veins (same generation of polish as our samples), stand-out 1-2 cm above the rougher surface of the enclosing gneiss. Quartz vein 1eight itself has been used as a semi-quantitative dating method (see for example Whitehouse et al. 1986, Birkeland 1982).


0 ... -·· ,. / 100 -- ,, / 200 / /' , / 300 v / 400 / I I I Depth (cm) 500 I : I 600 I : I 700 I I 800 I 900 I 1000 '

Figure 03.3. The decrease of the production rate (expressed as a percentage of its surface value) of spallation cosmogenic isotopes under a given thickness of rock, bold line (p=2.7 3 3 1 g/cm ); sediment, thin line (p=.2.0 g/cm ); ice, short dashed line (p=0.9 g/cm ); or snow, 3 long dashed line (p=0.3 g/cm ;; calculated using Eq. 4 (Chapter Bl). 133

The G103/G106 rock surface may have an apparent younger exposure age if it had been covered by sediment and/or snow following deglaciation. The decrease of the production rate due to coverage depends on the density and thickness of the covering material. In Figure 03 .3, the decrease of the production rates with depth under a given thickness of snow, ice, sediment or rock are shown. Under 2 m of snow the production rate is still almost 70 % of its value with zero coverage, while it takes only 20 cm of till to reduce the production rate by 1he same amount. In Table 03.3 are shown model 10Be exposure ages for the three quartz vein samples (G103, G106 and G 1 ·r o which include the effect of coverage of the surface by sediment or snow. We have calculated model exposure ages using a thickness of 10 or 20 3 cm of unconsolidated sediment with a density of 2.0 g/cm • We note that patches of basal lodgement till have been observed within several meters of the area where our samples were taken. Thick patches {ca. 20 cm) of soil and vegetation have also been noted covering well-polished rock surfaces directly above Totensee at Grimsel Pass (Fig. 03.1 ). The thicknesses we have used in our calculation therefore appear to be reasonable. Our calculations show that when ·: 0 cm of till covered the rock surface during the whole time of exposure, the exposure age would increase by more than 1000 years. When 20 cm covered the surface, then the surface is actually almost 3000 years older than the exposure age calculated assuming no coverage.

Tabl e 03.3. Modi e exoosure ages ca cu ate d for a rock surface covered by snow or till

Rock No. 10 cm till 20 cm till 1 m snow 2 m snow 3 m snow (6 months) (6 months) (6 months)

G103 10,200 11,700 9900 10,900 12,000

G106 9400 10,800 9100 10, 100 11,100

G111 12,500 14,300 12,100 13 400 14,800

In the last scenario, we calculated model exposure ages where the rock surfaces had been covered with snow for the last 10-15,000 years. As shown in Table 03.3, coverage by 1 to 3 m of snow for 6 months gives model exposure ages 1000-3000 years older than the age calculated without coverage. Coverage of the rock surface for at least part of the year by snow cannot be ruled out for the Gl03/G106 rock surface, as this surface is roughly horizontal. Moreover, the two different surfaces (the surface of Gl 03/G106 and the surface of Gl 11) need not have experienced the same coverage history. Clearly, it is impossible to really know by how much or with what the surfaces may have been covered. These estimates are made just to show that the effect of snow or sediment lying on top of th·~ rock surface can be significant. 134

We did determine an exposure age from one steeply dipping surface. Sample Gl 10 was taken from a steeply dipping slab along the lower margin of the nunataks directly overhanging the pres,ent Rhone Glacier. The high angle of the sample suggests that it was not covered with ice or sediment. On the other hand as the surface was not polished, it would be difficull to constrain the last time that a large slab broke off.

03.5 CONCLUSIONS The rock surface we have dated that was closest to Grimsel Pass proper had an 10 Be exposure age in the range of 14,000-10,000 years. Samples taken from a horizontal surface ca. 200 m higher up were more than 2000 years younger. We interpret this to show that the calculated ages are minimum ages and that the latter rock surface was covered by either snow or sediment (or both) during much of its exposure. Calculation of model ages using plausible amounts of either snow or till yielded exposure ages several thousand years older. More !,amples need to be analyzed to further elucidate the timing of deglaciation of high Alpine passes and the breakdown of the south to north transfluence of Rhone glacial ice over Grimsel Pass. 135


04.1 INTRODUCTION The glacial landscape of the Swiss foreland is characterized by huge erratics (Findlinge) which dot the surface. Unfortunately, many of those erratics have been 'dynamighted' for construction and road-building material. In the Alps, a rather wide variety of rock types are exposed, many of which occur very locally. A consequence of the lithologic diversity of the Alps is that specific lithologies are identified with certain glaciers and the precise origin of many erratics can be easily deciphered. These are so- called index erratics (Leiterratiker). Some of the more well-known index lithologies are listed in Table 04.1, while the original outcrop location and presumed path that an erratic might have been carried along by a given glacier are shown in Fig. D4.1.

Figure 04.1. Simplified map showing the origin and presumed path of some of the more well-known index erratics (e.g. MOiier et al. 1984 and references therein). Source location numbers refer to Table 04.1.

Index erratics are abo useful for deciphering the origin of a glacier associated with a given till sheet or outwash fan. This is especially crucial in regions which could have been covered by one of several glaciers, as is true for the central northern Swiss foreland. 136

Erratics provide information about the farthest lateral extent of a given glacier at a given time. The elevation at which an erratic was deposited helps to constrain the thickness of the ice mass at that time.

Table D4.1. Lithology and origin of a few index erratics (from Muller et al. 1984, original source Winterhalter and De 01ervain 1935). Glacier Index erratic Kanton of origin Rhine Julier-Albula Granite Graubunden Li nth-Rhine 2 Glarner Verrucano Glarus Rhone 3 Arkesine Wallis Rhone 4 Arolla Gneiss Wallis Vallorcine Conglomerate Wallis

Erratics of the northern Swiss foreland have been grouped into two broad categories: 1) those from the last glaciation (LGM), and 2) those from an earlier but more extensive glaciation (Riss in the classical sense), known as the most extensive glaciation. Erratics related to the LGM occur either atop LGM moraine walls or in association with LGM basal till deposits. They are found within the bounds of the LGM moraine belt, while erratics of the second type are characterized by being located external to the LGM moraine belt. In the Jura Mountains, such so-called Riss erratics are found at higher elevation than the LGM erratics (for example, 1000-1200 m while LGM erratics are at 500-600 m, Hantke 1977). It is presently not clear whether all such 'Riss' erratics date to the same ice age. Such a question can be attacked by exposure dating of erratic blocks. With exposure dating, we can also address the question of when each of the major glacial lobes of the northern foreland (the Geneva-Rhone, the Solothurn-Rhone, the Li nth-Rhine and the Rhine) reached their maximum extent, and whether the four lobes behaved synchronously. Here we present exposure ages obtained on four erratics. Three of these mark the maximum extent of the Solothurn lobe of the Rhone Glacier during the last glacial cycle, while the fourth erratic is from a deposit related to the most extensive glaciation.

04.2 GEOLOGIC SETTING AND SAMPLE DESCRIPTIONS The Rhone Glacier originates just south of Grimsel Pass (see Fig. D4.1 ). During the ice ages, the Rhone Glacier filled the Rhone Valley then spread out into the Swiss foreland. It split into a southwestern lobe known as the Geneva lobe and a northern lobe called the Solothurn lobe. The Jura Mountains deflected this glacial lobe to the east. 137


0 10 20km

+~ation of erratic boulder

~m LGM Rhonie Gacier moraine

Figure 04.2. Schematic map of the Wangen a. d. Aare region indicating the location of the erratics we have analyzed (ER1, ER?, ERB). The patterned band shows the area covered by landforms and deposits related to the terminal position of the Rhone Glacier at the LGM (modified from Nussbaum 1910, 1951, Staub 1950, Ledermann 1978). 138

At its maximum extent, the Rhone Glacier filled the Solothurn basin, and reached all the way to just past Wangen a. d. Aare (Fig. 04.2). The bedrock in the Solothurn region consists of an uneven erosional surface cut into the Molasse; to the north, the upturned Mesozoic beds at the front of the Jura Mountains are found. This ancient landscape contains north-south running dry stream valleys of enigmatic origin (i.E!. cross-cutting the present drainge system) (Nussbaum 1951 ), which separate small hills of Molasse. LGM Rhone Glacier basal till mantles this complex landscape. The moraine walls and associated erratic boulders on the crests of the Molasse hills mark the lateral and vertical Rhone LGM ice limits. The LGM landforms in the end-moraine region ofWangen a. d. Aare (Nussbaum 1910, 1951, Staub 1950, Ledermann 1978, Hildbrand 1990) have been subdivided into three sub-parallel moraine series (see Nussbaum 1910, 1951 for details). These have been termed the Wangen a. d. Aare maximum, the Brastenberg, and Solothurn recessional stadia of the LGM (Nussbaum 1910, 1951, Hantke 1977, 1978-1983, Ledermann 1978). The moraines and till deposits of the furthest extent of the Rhone Glacier at the LGM wind around Oberbipp and Bannwil continuing to Steinhof and Steinenberg. These two small Molasse hills are draped by a basal till which is capped by many large erratics. At Steinenberg the till, which contains on the order of 20 erratics, onlaps onto an older gravel deposit which outcrops at the top of the hill (Lederman 1978). We have sampled eight erratics at Steinenberg and two at Steinhof (Table 04.2). locations of the erratics inv•estigated so far are shown in Figure D4.2. The lithology of most of the Steinhof/Steinenberg erratics, including those we have sampled, is so-called arkesine, which originates from Val de Bagnes in Wallis (Ledermann 1978). Arkesine is a banded gneiss of the Dent Blanche Nappe. It is made up of alternating bands of chlorite- talc schist and a green granitic gneiss, which contain 1 cm long hornblende crystals. The quartz grains in the coarser grained granitic bands are up to 0.6 mm in size. Erratic sample ER 1 was taken from the top surface of the largest erratic at Steinhof (with an 3 estimated volume of 1200 m ). Our sample consists of a green foliated granite with large quartz, alkali feldspar, plagioclase, and hornblende crystals; although the lithology and grain-size vary on the centimeter scale. Sample ER7 and ERB are from the Steinenberg hill. ER 8 has a lithology similar to ER1 (foliated granite rich in hornblende), while ER7 is an aplitic that cross-cuts such a granite. Sample MLS was from a boulder ca. 1.5 m in diameter (sample location not shown). Lithologically, it is a weakly foliated granite. This boulder was found in association with glacial deposits along the western slope of the Napf region and therefore lies outside of the LGM moraine extent. It is postulated that the enclosing sediments are related to the most extensive glaciation (Lang 1991 ). 139

Ta b le D4.2. Erratic bOU Id ers analyzed

Rock Height of Lithology Altitude Sample number boulder (m) thickness (cm)

ER 1 lOm foliated granite 580 6

ER 7 1m aplite dike 620 4 ER 8 1.5 m foliated granite 610 3 ML5 1.5 m foliated granite 800 3

04.3 TIMING Of THE RHONE GLACIER MAXIMUM AT WANGEN A. 0. AARE 36 A whole rock sample from erratic ML5 was measured for CI. The resulting AgCI 36 did give enough current to be measurable but the Cl/CI ratio was at the level of a blank. That means that the surface we sampled from that boulder had not been exposed long 36 enough to accumulate measurable amounts of CI. Given its location at the base of a hill, it is likely that the boulder weathered out of the enclosing sedimentary matrix and rolled down the hill.

d 10 Ta bl e D4.3. AMS-measure Be an A concentrations an d ca cu ate d exposure a~ es.

Rock Al atoms/gram Si02 Exposure age Mean age of No. (oom) (years) erratic block

10 5 ER 1 Be 1.62 x 10 20,000 ± 1800 20,000 ± 1800

10 5 ER 7 Be 1.42 x 10 14,700±1200 15, 100 ± 1000 5 i&AI 11.8 9.58 x 10 16AOO ± 2100 10 5 ER 8 Be 1.50 x 10 16,700 ± 1400 17,300 ± 1200 6 26AI 66.8 1.11x10 18,900 ± 2400 15 ML5 i6CI (3x10" )* *The measured 36Cl/CI ratio corresponds to a blank.

10 16 The five exposure ages we have calculated for the Be and Al concentrations measured in the three •=rratic boulders from the LGM (ER1 1 ER7, and ERB) are given in 10 26 Table D4.3 and shown in Figure 04.3. Both Be and AI were measured in the erratics 0 26 ER7 and ERB. Within the listed errors, the ' Be and AI exposure ages for each rock agree well with each other. 140

a: ' I 6 i ; w j j l j : i ~ t-. ' . ' a: w I- •

00 : ~ a:' W! • I

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ..q-ft C\J co co 0 C\J~ ..q- ..- .,.... .,.... .,.... C\J C\J C\J

Time (years)

10 26 Figure 04.3. Plot of Be and AI exposure ages determined in three erratic boulders 26 from the Wangen maximum of the Rhone Glacier. Filled symbols are for AI open symbols for 10Be. 141

In light of the spread in ages, we take the viewpoint that the oldest exposure age of a suite of boulders more closely represents the actual age of the deposit (e.g. Zreda and Phillips 1995, Phillips et al. 1996b). As previously noted, erosion, spalling, rolling over shifting of the boulder or uncovering (eroding off of a till matrix), coverage by snow, or shielding of cosmic rays by twes, all act to make the calculated exposure age 'too young'. Such an interpretation is espedally applicable in our case because ER1 is much larger than the other two boulders (see Table 04.2). If the apparently younger ages are due to shielding of cosmic rays by forest and/or snow, then corrections might be called for, although we do not make these corrections here. Cerling and Craig (1994a) calculated a correction of 4 % decrease in production rate due to shielding of cosmic rays by forest (in the Puy de Dome region of France). Similar to the discussion pertaining to the Grimsel Pass samples, snow corrections of 1 m for 6 months of the year can increase an exposure age by as much as 10 %. In sum, shielding corrections for snow and forest cover could easily increase the ages of the smaller erratics several thousand years. One cannot rule out that the apparent younger ages of smaller erratics are due as well to post-depositional processes. The exposure age for ER1 indicates that at 20,000 ± 1800 years ago the Rhone Glacier reached the position where it left this moraine at Steinhof. Geomorphologically, this position is interpreted to be the maximum position for the Solothurn-Rhone lobe during the whole of the last glacial cycle (Nussbaum 1910, 1951, Staub 1950, Ledermann 1978, Hildbrand 1990). If: must be remembered that this age can as well be a minimum age, to a certain degree. It is unlikely that such a large erratic had been shielded by forest, on the other hand, erosion cannot be ruled out. This exposure age is important because it provides the first direct evidence that the Rhone Glacier reached its maximum extent in the late part of the last glacial cycle, that is to say not during the early Wurm. Our date also provides direct verification that the Rhone Glacier reached a maximum synchronous with the world-wide LGM ice volume maximum, dated at ca. 20,000-18,000 14C years (e.g. Chappell and Shackleton 1986). An alternate scenario that we can envision is that the spread in ages represents the actual time span that the Rhone Glacier fluctuated at this terminal position. Our exposure ages can be interpreted to mean that the Rhone Glacier reached its maximum extent at 20,000 ±1800 years ago, then retreated from its maximum position around 15,000 ±1000 years ago. It is worth noting that the huge erratic (ERl) actually lies directly on the outer moraine wall at Steinhof (Wangen stadia!) as mapped by Ledermann (1978). His map indicates that the Wa'.1gen outer moraine wall does not lie at Steinenberg (where ER7 and ER8 are from) but is located several kilometers to the east (south of Riedtwil). That means that the two smaller erratics may actually be located inside the outer wall and perhaps 142 stabilized somewhat later thar the Steinhof boulder. Note that all of these erratics still lie outside of the Solothurn (and less well-defined Brastenberg) stadia! moraine walls.

Ill II le ~ lb ~ c: ~ o la 0..


Figure 04.4. 3180 curve (Eicher 1979) with pollen zones (Welten 1947) for Burgaschisee. The pollen zones are as follows: la Oldest Oryas, lb, Bolling, le Older Dryas, II Alleroo, Ill Younger Oryas.

18 04.4 COMPARISON TO THE 8URGASCHISEE 3 0 RECORD Absolute ages bracketing the maximum extent of the Alpine glaciers in the foreland, prior to this investigation, have only been obtained with radiocarbon (cf. Schlachter and Rothlisberger 1995). Unfortunately, during the coldest periods, vegetation is so sparse that finding sample material later for radiocarbon dating becomes unlikley. On the other hand, palynological (and in many cases 3160) investigations of many of the lakes formed following the LGM, do provide a picture of the sequence of climatic events after retreat of the glaciers (also see Chapter 05). Fortuitously, both 3180 (Eicher 1979) and pollen (Welten 1947) were investigated at the small lake Burgaschisee (location shown in Fig. 04.2). Burgaschisee lies 2 km behind the moraine wall at Steinhof and was dammed by a smaller moraine. The location of Burgaschisee is unique in that it lies within the tongue region defined by the outer Wangen moraine belt but as well lies outside the Solothurn moraine belt. That means that the lake formed very close to the 143 time immediately following the withdrawal of the Rhone Glacier from its maximum position at Wangen. 18 In Fig. D4.4 we have reproduced the 0 0 curve for Burgaschisee from Eicher (1979, 1987), including the location of the pollen zones (after Firbas 1949) from Welten (1947). Although no radiocarbon dates are available, one can see that the record begins in the Oldest Dryas, well before the beginning of the B0lling , which began 14 around 13,000 C yr BP (Mangerud et al. 1974). Because sedimentation rates were high in the early parts of the Olde!;t Dryas, the amount of time contained in the Burgaschisee sediment underlying the onset of the B011ing is difficult to gauge. The deepest sediment described is a blue glacial clay, but basal till was never reached in either the early (Welten 1947) or later corings (Eicher 1979). Radiocarbon dates from Soppensee (Hajdas 1993) and ZOrichsee (Lister 1988) indicate that by at least 14,600 14C yr BP there was no ice left in the foreland. One might infer a similar chronology for Burgaschisee. Our direct dating of the erratic boulder at Steinhof (20,000 years) on the outer moraine at Wangen agrees well with the estimate that Burgaschisee had formed by at least 15,000 radiocarbon years ago. The oldest and youngest exposure ages may actually bracket the length of time (3,000-5,000 years) the Rhone Glacier stagnated at this terminal position.

04. 4 CONCLUSIONS The first direct measurement of the maximum extent of the Solothurn lobe of the Rhone Glacier has yielded an age of 20,000 ± 1800 years ago. This exposure age is from the largest erratic at Steinhof, which is located along the outermost moraine associated with the Wangen maximum. Ages of two smaller erratics located slightly behind the outer wall gave ages of 15, 100 and 17,300 years. These younger ages may be due to shielding or to post-depositional processes. Alternatively, the span of time the Rhone Glacier fluctuated at this maximum position may be on the order of 3000 to 5000 years. All three exposure ages restrict the maximum of the Rhone Glacier to the Last Glacial Maximum and not significantly earlier. 144

05 From the maximum of the Last Glaciation to the Preboreal: a comparison of new exposure dates with existing radiocarbon chronology

05.1 INTRODUCTION In this final chapter o' the section relating to exposure dating in the Swiss Alps, we present an overview of the timing of events, and therefore landscape evolution, from the height of the LGM to the onset of the Preboreal. The picture of landscape evolution that has been already constructed with: 1. glacial geomorphological information (e.g. Heuberger 1966, Kerschner 1978, Maisch 1982, 1987, 1992), 2. sedimentological fades studies of both lacustrine sediments (e.g. Lister 1988, Niessen et al. 1992) and coarse-grained deposits (e.g. SchlOchter and Wohlfarth-Meyer 1986, Graf 1993, MUiier 1994). 3. palynological investigations of lakes and bogs (e.g. Burga 1980, Burga 1988, Welten

1982, 19881 Ammann et al. 1994), 4. 0180 of lacustrine sediments (e.g. Eicher 1979, 1987, Lister 1988), and 5. high-resolution radiocarbon dating, wherever possible (Zbinden 1987, Hajdas 1993). Such information h

05.2 THE HEIGHT Of THE LGM At the height of the Last Glaciation, the northern Swiss foreland was more than 90% covered by broad pi•edmont glaciers of the Rhone, Reuss, linth-Rhine and Rhine Glaciers (see Fig. 04.1 ). The snow-line depression was estimated to be 1200 m compared to present day equilibrium line (Maisch 1982, 1987, 1992). Setting the exact time of the reaching of this maximum has been based on correlation with the established world-wide ice-volume maximum (e.g. Haberli and SchlUchter 1987) which occurred at 20,000- 18,000 14C years ago (e.g. Chappell and Shackleton 1986). Our exposure dates from erratic boulders located along the outer moraine of the Solothurn lobe of the Rhone Glacier indicate an age of 20,000 ± 1800 years for maximum of the Rhone Glacier, therefore roughly coincident with the world-wide LGM. In the discussion below, we will also compare the timing of the Solothurn lobe with the timing of variations of the Geneva lobe, the Linth-Rhine and the Rhine Glaciers. 145

At peak times of glaciation, the Rhone Glacier was made up of a Geneva lobe and a Solothurn lobe. For the Geneva lobe, correlation of moraines (at Beligneux) with radiocarbon-dated terraces (ages> 25,000 14C yr BP, Mandier 1988, Monjuvent and Nicoud 1988b, Evin et al. 1994) located downstream indicates that the maximum extent of the Rhone Glacier during the last glacial cycle apparently predated the LGM (summarized in Gillespie and Molnar 1995). A similar conclusion was reached based on tills and intervening lacustrine sediments outcropping and seen in cores taken south of Lake Geneva (Blavoux 1988, Monjuvent and Nicoud 1988a). Monjuvent and Nicoud (1988b) made the interpretation that the Last Glaciation was not actually two-phased, but was made up of one long cold phase with lesser climate fluctuations. Similar conclusions were reached from the Grand Pile record (e.g. de Beaulieu and Rei lie 1992) and from deposits near the Vosges (e.g. Seret et al. 1990), both in France. The aforementioned authors all maintain that in both the French Alps and the Vosges the maximum ice extent occurred early in the last glacial cycle (i.e. early Wurm - isotope stage 4 ?) not late. No direct dates are available for the time at which the Geneva lobe reached its maximum extent. Following this long glaciation, the Rhone Glacier retreated from the foreland 14 around 20,000-18,000 ( yr BP (Evin et al. 1994). Our new exposure dates provide the first absolute dates for the timing of the maximum extent of the Solothurn lobe. Based on the exposure age of the Steinhof erratic (ER 1), the maximum occurred at or slightly before 20,000 ± 1 800 years ago. The period of time corresponding to withdrawal of the Geneva lobe (20,000-18,000 14C yr BP, Evin et al. 1994) does roughly correspond to our date. A 14 recent radiocarbon dat·2 from the base of a proglacial sequence of 19,000 ( yr BP, described in cores fror1 the southern tip of Lake Geneva, indicates that the Rhone Glacier was still very nearby in the foreland at that time (Moscariello 1996). During the las't glacial cycle, the Li nth-Rhine Glacier reached its maximum extent at Killwangen. The timing of this advance has been constrained with radiocarbon dating (Schluchter and Rothlisberger 1995). Before the glacial advance which culminated in the Killwangen maximum, there was an interstadial period where the foreland itself was ice- free. This is based on the Gossau profile where lignite beds with radiocarbon ages from 60,000 to 28,000 14C yr BP (Schlochter et al. 1987). A minimum age for the Killwangen 14 till of 28,000 ( yr BP has also been obtained from a correlative lignite horizon in a core from Zurichberg (Susenbergstrasse; Schluchter and Rothlisberger 1995). The maximum age for the Killwangen advance was obtained from deltaic sediments overlying the Killwangen till (in '£he Zurich Hauptbahnhof drillcores) with two radiocarbon ages of 19,820 and 23,000 14C yr BP (Schluchter and ROthlisberger 1995). The time span therefore determined for the formation of the Killwangen moraine roughly overlaps the time at which the Solothurn-Rhone Glacier reached its maximum. 146

Radiocarbon dates for the advance and retreat of the Rhine Glacier in the region of Schaffhausen have been !;ummarized in Furrer (1991 ). Ice advanced onto the foreland to the north of Bodensee (Lake Constance) after 29 ,000 14C yr BP (Gross 1958). An age of 18,500 14C yr BP (Geyh and Schreiner 1984) was obtained from a broken mammoth tusk underlying basal till of the S1ein am Rhein stade (which postdates the Schaffhausen maximum, see also Table DI .1 ). An age of 15,900 14C yr BP (Gross 1958) has been interpreted as a minimum age for the retreat of the Rhine Glacier from the Schaffhausen terminal position (e.g. Furrer 1991 ). These dates clearly overlap our exposure dates. One should remember that direct comparison of exposure ages and radiocarbon dates is at best tenuous. Additionally, it seems likely that each lobe did actually stay at its maximum position for a few thousand years. This has, for example, been interpreted from the amount of sediment deposited and the size of the landscape features associated with the Wangen maximum (Nussbaum 1951 ). We may interpret our exposure age from the Steinhof erratic to approximai:e the time when the moraine stabilized right after the glacier withdrew from that position. Therefore, it would not be unexpected for such an exposure age to be slightly younger than a radiocarbon age from sediment underlying a till layer. That is, the radiocarbon date indicates the last vegetation just prior to a glacier's advance, while an exposure age from a moraine indicates the moment of glacier withdrawal. Indeed, we have few exposure dates from the foreland erratics of the LGM. First it is important to note that the date of 20,000 ±1800 years indicates that the Solothurn lobe of the Rhone Glacier did reach its maximum late in the last glacial cycle, i.e. coincident with the world-wide continental ice maximum. Second, we do therefore have the first hint of synchroneity for the three major glacial lobes of the northern Swiss fore land.

05.3 0EGLACIATION AND LANDSCAPE EVOLUTION FOLLOWING THE MAXIMUM EXTENT After the closely-spaced in time, multi-phased maximum position had been reached, the large piedmont glaciers rapidly melted back out of the foreland (Schluchter 1988-1989). In the case of the Li nth-Rhine lobe, a brief stabilization is marked by the Schlieren moraine, which may be correlated to the Stein am Rhein position of the Rhine Glacier. This was followed by a readvance to the Zurich position (Solothurn and Konstanz correlative?) (see Table 01 .1 ). Due to the general lack of absolute dates, the correlation of the recession/readvance moraines of each major lobe is still unclear. The exact sequence of events between the height of the LGM and complete deglaciation of the foreland are unclear. Several radiocarbon dates between 16,000 and 14,000 14C yr BP (Furrer 199-1

Steinhof erratics does actually represent the span of time the Rhone Glacier stagnated at the Wangen terminus, then that period lasted 3000-5000 years. We are wary of drawing conclusions on so few exposure dates, but point out that in such a scenario, the melt back of the foreland piedmont glaciers would have been incredibly rapid, as previously suggested by SchlUchter (1981l, 1992). This has been attributed to an increased continentality of climate following the glacial maximum reflected by a rapid equilibrium line rise (Haberli and SchlOchter 1987, Haberli 1991a, 1991b). The minimum time that the foreland was completely ice-free is based on a radiocarbon date of 14,600 14C yr BP from sediments of ZOrichsee which also dates the end of glacial meltwater influence on 0180 (Lister 1988). A similar date (14,200 14C yr BP) was obtained from Soppensee sediment (Hajdas 1992). During this transitional phase pollen studies provide a picture of pioneer grasses- type vegetation (artemesia) scantily covering the open ground between huge meltwater lakes and braided river systems. In many of the glacial troughs that had been cleaned out of older sediment by the LGM glaciers, very high sedimentation rates have been noted for that time period ( MOiier 1994). Large scale slope instability and erosion were important processes of landscape modification following glacial retreat (e.g. Furrer 1991). Loss deposition in the Lake Constance region was marked in the early part of the Late Glacial, then decreased markedly at 14,500 14C yr BP (Niessen et al. 1992). In the period just 'oefore 14,500 14C yr BP, sediment of foreland lakes show that the landscape is evolving from vegetation-free open ground to sparse colonization by grasses. Above basal gravels, the lowermost lacustrine sediment is dominated by the sand fraction (Wohlfarth et al. 1994). This is overlain by several meters of blue-gray clayey sediment, which contains a very low percentage of pollen, much of which is reworked (Wellen 1982, 1988). The first pollen zone (biozones of Firbas 1949, in the sense of Ammann and Lotter 1989) which has been recognized, following opening-up of the landscape as the glaciers withdrew, is the Oldest Dryas. This early phase with low pollen counts was followed bv a Helianthum (steppe tundra) phase then a Betula nana (dwarf shrub) phase (Ammann et al. 1994). The time of the transition between the second and third phase is ca. 13,5'JO 14C yr BP. By this later phase, the open landscape had been colonized by vegetation. Consequently loss deposition completely tailed off (Niessen et al. 1992). At roughly the same time (ca. 14,500 14C yr BP), the once extensive piedmont glaciers had been reduced to valley glaciers restricted to the Alps proper. Pulses of small scale glacier expansion are recorded in the Late Glacial recessional sequence Gschnitz, Clavadel, Daun, and Egesen (Suter 1981, Maisch 1982, 1987, 1992, for additional references see Chapter 02). Apart from the Egesen stade the associated moraines have 148 proven very difficult to date with radiocarbon. The Gschnitz, Clavadel, Daun cold events are postulated to lie between 14,500 14C yr BP ca. 12,000 14C yr BP. The sequence of events following deglaciation in the high Alpine regions has been provided by pollen investigatio1s (Burga 1980, 1987, 1988, Welten 1982, 1988, Furrer 1991, Schlochter and Wohlfarth 1993, Ammann et al. 1994, Wohlfarth et al. '1994). Therefore, the sequential deglaciation of areas at higher and higher elevation can be closely tracked by noting which pollen zone is found in the basal sediments at a given site (non-reworked) (Furrer 1991 and references therein). Within the main Alpine valleys (i.e. the Rhone and Rhine River valleys) the pollen record indicates normal lacustrine deposition began during the Oldest Dryas. Sites at slightly higher elevation have records beginning in the second phase (steppe tundra) or third phase (dwarf shrub) of the Oldest Dryas. The high passes were possibly still ice-covered during the Oldest Dryas (e.g. Gotthard, , Grosser St. Bernhard, Maloja) and then they finally became ice-free during the B0lling, as the major ice transfluences broke down (Heitz et al. 1982, Furrer et al. 1984, Burga 1987, Furrer 1991 ). In Figure D5.1, an interpretation of the increase of the snow and tree-lines following deglaciation is shown (Burga 1987 and references therein). We can see that the perennial snow-line did not go above 2500 m until after the beginning of the Preboreal. In Figure D5.1, the shaded band indicates the exposure ages we have measured from Grimsel Pass (Chapter D3). The altitude of our sampling sites ranges from 2200 to 2800, with older exposure ages coming from the lower elevation surfaces. Our oldest exposure age from a glacially polished quartz vein (at 2400 m) provides a minimum absolute age for an ice-free Grimsel Pass of ca. 12,000 years ago. As discussed in Chapter D3, this may itself be a minimum age becau~e it is difficult to constrain coverage by snow or sediment after deglaciation at our sampling sites. Based on pollen data, the passes themselves are thought to have become ice-free sometime during the B0lling. This has been verified by geomorphological data which indicate that the Egesen glaciers (Younger Dryas) did advance into an ice-free landscape (long lateral moraines) (e.g. Muller et al. 1981). Since vegetation was sparse during the Oldest Dryas, radiocarbon ages are rare. Indeed those that are found in tile literature were determined on bulk organic sediment and not on macrofossils (see Hajdas 1993 for a detailed discussion of the advantages of macrofossils over bulk material for radiocarbon dating). For example, a radiocarbon date from Hobschensee implies that Simplon Pass was ice-free by at least 12,580 radiocarbon years ago. Although it has been discussed that this age may be 'too old' (Welten 1982, Heitz et al. 1982), the enclosing pollen assemblage is indeed Aller0d (ca. 12,000-11,000 radiocarbon years BP), meaning the date could be only on the order of 500 years 'too old.' 149

3000 m


2000 m

1500 m ------

1000 m

14c time rn 17 1e 15 14 13 12 11 10 g s 7 e 5 4 3 2 calendar time g s 7 6 5 4 3 2 v

Figure DS.1. The increase in the elevation of the snow-line and the tree-line in the Swiss Alps at the end of the LGM (modified from Burga 1987). Time scales are in ka. The shaded band indicates the range of exposure ages we have obtained from Grimsel Pass (note that exposure time is plotted here as calendar time). Roman numerals indicate pollen assemblage zones as follows, la Oldest Dryas, lb B011ing, le Older Dryas, II Aller0d, Ill Younger Dryas, IV Preboreal, V Boreal.

The B0lling-Aller0d inte1-val is often termed an interstadial as it lies between two colder periods, the Oldest Drya~; and the Younger Dryas. Allemd climate amelioration is reflected in lithologic changes in lacustrine sediments from foreland lakes. The sediments became finer-grained and more rich in organic material after 13,300 14C yr BP. Deposition of lake marl began after 12,600 14C yr BP (e.g. Wohlfarth et al. 1994). The Laacher See Tephra (LST) layer is an important Allemd marker bed in sediments throughout central Europe. Recently dated at 11,200 (Hajdas et al. 1995), the LST occurs 200-300 varve years below the oeginning of the Younger Dryas cold event. A distinct increase in mineralogic and clay sedimentation marks the onset of Younger Dryas in lacustrine sediments (e.g. Schluchter and Wohlfarth 1993, Ammann et al. 1994, Wohlfarth et al. 1994). An increase of grasses (Artemesia) and a drop in tree pollen is seen in many records, especially from higher elevations (e.g. Welten 1982, 150

1988). The Younger Dryas was the most extreme cold event since the end of the last glaciation and resulted in abandonment of the landscape by vegetation at the higher altitudes, and led to increased slope erosion. Glaciers advanced markedly and the outer walls of the Egesen moraine complexes were formed at many locations in the western Alps (Kleiber 1974, Suter 1981, Maisch 1982, 1987, 1992). The exposure age (11, 100 years), we have determined from six boulders in a series of nested Egesen moraines at Julier Pass confirm that Egesen does equal Younger Dryas. We believe that the outer moraine wall records the Vadret Lagrev glacier's initial rapid response to the marked temperature drop at the beginning of the Younger Dryas. Our exposure date for this outer moraine wall is 11,800 years. The second group of samples is from a more diffuE.e boulder band which cross-cuts the sharper outer wall. Stabilization of this dump moraine occurred after the down-wasting of the glacier at the

Younger Dryas/Preboreal transition (101400 years ago). The synchroneity of the cold 4 event marked by pollen (1 c.datedl and the advance of the glaciers during the Egesen is notable. To summarize, our new exposure dates provide direct dates for several of the events at the height of and following the collapse of the Swiss LGM glaciers. The outer Wangen moraine which marks the maximum of the Solothurn lobe of the Rhone Glacier was constructed 20,000 years ago. If we interpret our date to indicate precisely when the glacier withdrew from its maximum position then temporal overlap of all three major foreland lobes may be inferred. The Rhone Glacier reached its maximum extent in the northern Swiss foreland coincident with the LGM but not earlier. A direct date from a glacially polished surface at Grimsel Pass indicates that the Rhone-Aare ice transfluence broke down no later than 12,000 years ago. Exposure dates from six boulders of an Egesen moraine provide di1·ect evidence that such moraines formed during the Younger Dryas.

16 05.4 A COMPARISON WITH SIMILAR 8E DATED EVENTS In the following brief discussion, we compare the 10Be atoms/gram quartz (scaled to high latitude and sea level with Lal 1991) for the maximum of the Rhone Glacier with the maximum extent of the Laurentide ice sheet at the LGM in eastern North America (Clark et al. 1995, Bierman and Larsen 1996) and the maximum extent of the Pinedale Glaciation (LGM) in the Wind River Range (Gosse et al. 1995b). We make such a comparison because, beyond the tree-ring calibration curve (>ca. 10,000 radiocarbon years, Kromer and Becker 1993), the exact calendar age that corresponds to a measured 14 radiocarbon age is difficult to ascertain. Presently accepted C calibration based on U-Th ages from corals (Bard et al. 1990, 1991, 1993) often in detail do not agree with other records (Hajdas et al. 1995). By comparing chronologies themselves constructed with 151

10Be we can circumvent this i:>roblem of radiocarbon calibration. We acknowledge that this type of comparison is only valid if past changes in magnetic field intensity or pole location did not result in sign ficant spatial variation of production rates. In each case, the

IUBe atoms/gram Si02 (thickness corrected and calculated back to sea level) are in the 5 5 5 neighborhood of 1.0 x 10 (i.e. 0.95-1.08 x 10 , our data; 0.95-1.2 x 10 , Gosse et al. 1995b; 1.02 x 105, Clark et al. 1995). This is striking. It indicates that irrelevant of the production rates used, one can use the atoms accumulated as a chronometer. It also means, perhaps just as importantly or perhaps more so, that an alpine glacier in Wyoming, an alpine glacier that had spread out into a piedmont glacier in Switzerland, and the eastern margin of the Laurentide ice sheet reached their maximum extents at something very near to precisely the same time. This is based on the amount of 10Be that has since accumulated in bou1ders that each glacier left after it retreated. A very similar case can be made for the Titcomb Lakes moraine (Wyoming) and our Julier Pass moraine. The ll'Be atoms/gram Si02 in both cases are very similar, i.e. our 4 4 5.7-6.8 x 10 to those of Gosse et al. (1995a) which are 6.1-6.9 x 10 • Indeed when one considers just the two outer moraine boulders from Julier Pass (concentrations 6.6 and 6.8 4 x 10 ) the agreement is once again remarkable. Without ever determining an actual age, one can immediately say that the Wyoming moraine is very nearly contemporaneous with the Julier Pass moraine. Since the latter can be shown by geomorphological evidence, as well as nearby radiocarbon dating, to be of Younger Dryas age, one can deduce that the Titcomb Lakes moraine is as well a Younger Dryas moraine. By measuring the amoL'nt of cosmogenic isotopes in boulders or surfaces from similar events worldwide one May compare their isotope inventories. Synchroneity of glacier fluctuations can then be checked without ever having to have an accurate production rate, and without ever determining an age. 152


E1 INTRODUCTION Knowledge of the most accurate production rates for the cosmogenic isotopes is essential. Otherwise comparison with dates from other records based on different 10 chronometers than Be (e.g. radiocarbon) will be meaningless. As the main goal in ice age research is to look at the magnitude and timing of geographically widely separated cold events (e.g. Broecker and Denton 1989), true comparison of chronologies is imperative. 10 26 In this section, we examine production rates for Be and Al which we have calculated using our data from Antarctica and from )ulier Pass, Switzerland. These data sets include very old and very young samples affording us a look at production rates in both time windows. We then compare the production rates calculated from our data with published production rates, which have been determined both empirically and from model calculations. Clearly the most difficult variable to constrain in production rate calculations from measured sample data is the actual age of the surface. With indirect methods it is difficult to constrain the age of a rock surface. Indeed, this is why exposure dating is such an indispensable tool. For example, even basal radiocarbon ages from a bog dammed by a moraine do not necessarily provide an accurate age for that moraine. There are many reasons why a date may be nonrepresentative. For example, the length of time of a glacier's stillstand at that position are unknown. Likewise, the lag time between glacial retreat and the time at which lake productivity was high enough that sufficient organic carbon is found in the sediment (in the case of bulk sediment dates), or for vegetation to be well-enough developed that rnacrofossils are washed in (in the case of macrofossil dating), are unknown factors. Too old radiocarbon ages from bulk lake sediment resulting from the hard-water effect (lake water containing more 'dead' carbon than the contemporaneous atmosphere due to underlying carbonate bedrock) are well-documented (e.g. Olsson 1986). A factor too often ignored, but often leading to unreliability in bulk sediment dating (Clark et al. 1995, 1996) is the washing in of older organic material in a catchment basin. This is especially likely when attempting to date a moraine built up in a cold period which followed a significant warm period, i.e. the Younger Dryas. At high elevations, the breakdown of organic matter proceeds especially slow. As the cold period sets in, the vegetation is rapidly killed off, bL:t it does not decompose and is eventually washed into 153

the newly dammed lakes. This process is also confirmed by the high percentages of reworked pollen noted in basal lake sediments from the Swiss foreland (Welten 1982, Ammann et al. 1994). Single-year terrestrial macrofossils are preferred for radiocarbon dating (Hajdas 1993). 14 The radiocarbon timescale is only confidently calibrated back to ca. 10,000 C yr BP (Hajdas et al. 1995). That means for older measured radiocarbon dates no calendar dates are available. Using a coral calibration (Bard et al. 1990, 1991, Stuiver and Reimer 1993) may be more dangerous than simply sticking with the radiocarbon date until an extended calibration scale is available. In that sense, production rate calculation with 14 samples from a lava flow that is 14,500 C yr BP (i.e. Tabernacle Hill) (Cerling and Craig 1994) may be flawed in that the actual (calendar) age of the lava flow is unknown. Therefore, radiocarbon dated surfaces are only useful when the age lies within the radiocarbon calibration scale (Kubik et al. 1996). It must be borne in mind that even a very precise radiocarbon age (error of< 1 %) can yield a rather broad calendar time range, if the measured age lies a long one of the many radiocarbon plateaux. A radiocarbon plateau is a period of time (calendar time) in which all the calendar ages are equal to a single radiocarbon age (e.g. Kromer and Becker 1993, Hajdas 1993). With such caveats in mind, we proceed with a discussion of production rate calculations from our data. In general, the data we use are as well-constrained, if not more so, than many of the data sets used thus far for the published production rates.

E2 ANTARCTICA Samples from Antarctica


26 10 The calculation of the Al production rate is straightforward because the Be ages 26 10 measured in the same samples prove that the AI is in saturation. For Be, the 21 assumption of saturation is not so straightforward. At Table Mountain, Ne was measured in sample 220 and an age of 6.0 Ma was obtained (Bruno 1995). If the surface had a 154 simple exposure history (no prior exposure), the 21 Ne measurement indicates that the plateau surface at Table Mountain is in saturation for 10Be. We then expand this assumption to include all sim lar silicified sandstone samples taken from the same geomorphologic surface (Chapter C2 and C3}. The 21 Ne measurements (Bruno 1995) support the conclusion that this surface is indeed very ancient. At Mount Fleming, 21 Ne (7.0 Ma} was measured in a dolerite clast found next to our sandstone clast (Bruno 1995). Although not (yet) directly measured in our samples, we assume that for Mount Fleming the 21 Ne ages indicate an age where one would expect 10Be to have reached saturation based on the antiquity of the Sirius Group surface.

Ta bl e E1.1. Pro d uction rates in quartz ca cu ate dfrom our Antarctic results. Rock Altitude ioBe 26AI No. (m) atoms/gram.yr atoms/gram.yr 205 1820 36.0 214b 2055 36.0 217b 1820 34.4 220 2090 5.7 42.2 320 2140 6.1 37.2 322 2140 5.6 323 2060 5.5 38.0 326 2060 5.7 Mean 5.7±0.1 37.3±1.1

Table E1 .1 contains the production rates for 26AI and 10Be we have calculated using Eq. 17. The given production rates have been scaled back down to sea level using the polynomials listed in Lal (1991 ), The concentrations have been corrected to surface values based on the thickness of the sample processed. All of the original atoms/gram quartz are given in Chapters C2 and C3. Six rocks were suitable for 26AI calculations. Five rocks were suitable for 10Be calculations. The rocks used for the 10Be calculation had 10 exposure ages> 4.8 Ma (using~. Be atoms/gram Si02 year as the production rate). For 26AI the individually calculated production rates vary between 36.0 and 42.2 26AI 10 atoms/gram Si02 year, with an average of 37.3 ± 1.1. For Be, the production rates from 10 each rock range from 5.5 to 6.1 Be atoms/gram Si02 year, and average 5.7 ± 0.1. The averages can be used to calculate an 26 Al;1°Be ratio of 6.5±0.2 which is in very good agreement with the measured va.lues of Nishiizumi et al. (1989) and Brown et al. (1991) (Table El .2), as well as our mearnrement of rock ]18 from Julier Pass, Switzerland (Table 1SS

El .3, see below). This agreement strengthens our confidence that also the 10Be data are saturation values and prior exposure has been negligible. One must recognize that we have not included erosion in our calculation. In that case, Eq. 17 would transform to:


10 All of the samples used for Be have very wind-polished surfaces. This indicates little erosion has affected those surfaces. We believe that these wind-polished surfaces formed more than ca. S million years ago. But one cannot rule out that very slow erosion took place during the early part of the exposure history, after which the wind-polished surfaces were formed. Inclusion of only 3 cm/Ma erosion into Eq. 18 yields an average 10Be production rate of 6.4. Therefore, because the actual erosion rates are unknown, combined with the fact that all of the samples need not have experienced the same erosion rate, these are considered to be minimum production rates.

Ta bl e El 2 Pro d uct1on rates or e an 1n quartz. 26 10Be 26AI Al/1°Be Time range Reference 2'.S.7±0.1 2:37.3±1.1 6.S±0.2 >4.8 Ma Antarctica, this work 6.1 >4.0 Ma Antarctica, Nishiizumi et al. 1991 6.4 41.7 6.S±l .3 >2.S Ma Antarctica, Brown et al. 1991 3S±2 >2.6 Ma Antarctica, Brook et al. 199S S.6±0.1 37.0# 6.6::1.2# 12,400 yr Julier Pass, Switzerland this work 6.03* 36.8 6.1 ±0.3 11,000 yr Sierra Nevada, CA Nishiizumi et al. 1989 s.809 13,000 yr Sierra Nevada, CA Nishiizumi et al. 1996 6.0±0.3 -10,000 yr water targets Nishiizumi et al. 1996 S.17±0.lS 30.4±1.01 S.9 21,SOO yr Bierman et al. 1996 4.74 28.9 6.1 14,000 yr Clark et al. 199S S.97 calculated Masarik and Reedy 199S jf calculated for Jl 8 where we have both 10 Be and 26 Al data. * using present geomagnetic latitude. 9 using geographic latitude. 156

The production rates we have calculated from our Antarctic samples are given with published values in Table El .2. The values which are presently accepted are those of 26 Nishii:wmi et al. (1989). Our AI value lies between those of other authors, while our 10Be value is somewhat lower. The value we have calculated is also slightly less than the value from model calculations by Masarik and Reedy (1995). We reiterate that we have not included erosion in these calculations. Brown et al. (1991) included an erosion rate of 10 3 cm/Ma in their calculation. If we also include a similar value, our Be production rate is increased to 6.4. This shows the sensitivity of the calculations to even very small erosion rates. In Fig. El .1 our calculated production rates are plotted vs. altitude. In this case the production rates have not been scaled back down to sea level, the values at the individual sampling altitudes are plotted. The curved lines indicate the variation of production rate with altitude based on Lal (1991 ), including a 10 % error band on either side. From this 26 10 figure we can see that our calculated production rates for both AI and Be lie within 26 the 10 % error band. The stated errors for the original 10Be and Al production rates were 5-7 % (Lal 1991 ). This figure indicates as well the agreement with the given altitude corrections for high latitude locations (Lal 1991 ).

E3 )ULIER PASS, SWITZERLAND We propose that the samples from the Ju lier Pass can be used to calculate a production rate because the ag€ of the Younger Dryas is known. As outlined irJ detail in Chapter D2, the Julier Pass moraine has been assigned to the Egesen stade based on equilibrium line altitude depres;.ion of the Vadret Lagrev, and geomorphologic: character (Suter 1981, Maisch 1982, 1992, 1995). We make the assumption that the outer moraine which we have dated formed at the beginning of the Younger Dryas and the ir.ner moraine formed near the end. Next we need to know the calendar age for thos1,; two boundaries. We can constrain the beginning of the Younger Dryas because we know the varve age of the LST ash layer is 12,560 years (Zolitschka 1996). The beginning of the Younger Dryas is known to have started ca. 200 varve years after the eruption of the LST (Hajdas et al. 1995a). This yields an age of 12350 for the beginning of the Younger Dryas. Since the Younger Dryas lasted 1100-1200 years, one can subtract 1150 from 12,350 and end up with 11,200 years ago for the end. 157



2 "E ~


0 0 50 100 150 200 250 300

26 Prcduction rate Al atoms/gram quartz



2 I Q) "O 1.5 E ~

Production rate 10ee atoms/gram quartz

26 10 Figure El.1 a for Al and Fig. El .1 b for Be. Production rate calculated using our measured data at the sampling location vs. altitude (modified from a figure shown by E. Brook at the Los Alamos workshop, 1996). The lines show the production rate plotted using the polynomials listed in Lal (1991) with 10% error on either side. 158

We assume the responsE1 time of the Vadret Lagrev (a small cirque glacier) was on the order of a few decades (e.g. Sugden and John 1976, Haberli 1994). We have therefore used 12,350 years for the beginning of the Younger Dryas and 11,200 years for the end. We also note that a radiocarbon age from a bog behind the second out of seven in a 14 series of moraine walls yielded a radiocarbon age of 10,700 C yr BP. (Heitz et al. 1982). This Egesen moraine series is located at Maloja Pass which is ca. 50 km from Julier Pass. This lends credence to the idea that the outer moraines at Ju lier Pass were formed at the beginning of the Younger Dryas. Since these samples are not in saturation, the exponential term in Eq. 1 (in Chapter B1) does not drop out. But since we know T, we can calculate a production rate by solving Eq. 1 for P and using the assumed time period and the measured concentration N for each rock ;


10 26 Production rates for Be and AI are listed in Table El .3. They have been scaled back to sea level and high latitude using Lal (1991 ). The concentrations used were corrected back for sample thickness and for shielding by surrounding mountain peaks. The present day geographic pole was used. Ohno and Hamano (1992, 1993) have shown that the location of the magnetic pole during the last 10,000 years is more closely approximated by the geographic pole than by the present magnetic pole. 10 We have two samples in which Be was measured from the outer moraine ()12 10 and Jl 8). They have yielded Be production rates of 5.6 and 5.8 atoms/gram Si02 year, 26 respectively. Additionally, from the AI concentration measured in J18, we have 26 26 calculated an AI production rate of 37.0 AI atoms/gram Si02 year. From the inner 10 10 moraine, three samples allowed Be production rate calculations. The average Be 10 production rate based on these five samples is 5.6 ± 0.1 Be atoms/gram Si02 year.

Table E1.3. Production rates calculated from )ulier Pass results Rock Altitude Age used 26Al/10Be ioBe 26AI No. (m) (yr)

J18 2210 12,350 6.54±1.2 5.6 37.0 Jl 2 2200 12,350 5.8 I J104 2185 11,200 5.7 )10 2185 11,200 5.4 )8 2175 11 200 5.5 Mean 5.6±0.1 159

The 10Be production rates from other sites with short exposure times listed in Table 0 E1 .2 vary between 4.74 and 6.03, with our values lying in between. Recently ' Be production has been measured in water targets and was recalculated to quartz, it is, as listed, 6.0 (Nishiizumi et al. 1996). The value suggested by Clark et al. (1995) is based on a reassessment of the time of deglaciation of the Sierra Nevada (California) and is not itself calculated from any new cosmogenic isotope measurements. They (Clark et al. 1995, Clark and Gillespie 1996) proposed that the Sierra Nevada production rates are too high because an age of 11,000 years was assumed for the deglaciation of the Sierra Nevada and thus the sampling site of Nishiizumi et al. (1989). They would suggest that the Sierra were deglaciated ca. 2000-3000 years earlier. This would actually be similar to the time at which the Alps were ice-free, that is prior to the Younger Dryas, and sometime in the B0lling or Aller0d, depending on elevation. Note that when the geographic latitude instead of the geomagnetic latitude is used in the original calculation, the production rate is 5.8, based on a 13,000 year age of deglaciation (Nishiizumi et al. 1996). In any case, the relationship of high-altitude polished surfaces to the age of up-valley Younger Dryas moraines is difficult to constrain. For example, the polished surfaces we investigated at Grimsel Pass (Chapter 03) lie well outside of the Younger Dryas position of the.Rhone Glacier. A fact which is not necessarily reflected in the obtained exposure ages. An additional data set was put forth, that from the southeastern margin of the Laurentide ice sheet (Clark et al. 1995, Bierman and Larsen 1996). Although the data are valuable for constraining minimum ice retreat ages for the Laurentide ice sheet, the correlations with the basal radiocarbon dates are tenuous. Such bulk sediment basal dates very often suffer from inclusion of old washed-in carbon. As a final note, we poirt out that calculating exposure ages from our Ju lier Pass data with the production rates of the above authors would yield ages in the range of 14,000 years ago. At this time all such cirque glaciers as the Vadret Lagrev were actually joined with the main trunk glaciers in the valleys (Suter 1981, Maisch 1982, 1987, 1992). Therefore such ages would be geomorphologically untenable. From Table E1 .2 it can be seen that production rates from Antarctica are slightly higher than from lower latitude locations. This may be due to the effect of the magnetic field. The production rates calculated from Antarctic samples reflect the integrated production rates for million-year time periods. The high latitude location of Antarctica makes it less sensitive to changes in paleomagnetic intensity. Instantaneous production rates may have varied significantly over time, this may be mostly attributable to changes in the strength of the earth's magnetic field in the past. Such variations do not affect high latitude locations. Variations in paleointensity are fairly well known for the last 10,000 years (McElhinny and Senar.ayake 1982, Meynadier et al. 1983, Creer 1988, Tric et al. 1992, Thouveny and Creer 1992). This is because data sets from archaeological samples, 160

lake sediment, and lava flows can be combined. Similarly, pole-wander over the last 10,000 years can be fairly well approximated (Ohno and Hamano 1992, 1993). From both sets of data, one might conclude that the periods of stronger paleointensity which occurred during the past 10,000 years resulted in a the smaller production rate of cosmogenic isotopes for that time interval.

E4 CONCLUSIONS Although the production rates we have calculated from our Antarctica and Ju lier Pass data are somewhat lower than the established values, they do not vary outside of the errors quoted for those values which are 5-7% (Nishiizumi et al. 1989). It is important to note that small changes in the latitude and altitude scaling factors can as well lead to such a change in the production rate. For example, Masarik and Reedy (1995) have suggested that more appropriate altitude scaling factors may be 5-10 'Yo less than the factors presently in use (Lal 1991 ). When one compares the best attainable measurement uncertainties (4-6 %) with the present understanding of paleomagnetic field strength and direction, and the latitude and altitude scaling factors, it is apparent that it may be difficult to distinguish variations of 5-10 % in accuracy of the presently used production rates for 10Be and 26AI. 161 f SUMMARY, CONCLUSIONS, AND OUTLOOK FOR FURTHER WORK

Conclusions drawn from our work can be divided into two broad categories; 1) conclusions about the timing of glacier advances and retreats that one might draw based on the exposure dates from the various sites we have investigated, and 2) the conclusions relating to the method of exposure dating itself. The outlook for further research is included within the discussion of each individual topic below.

Conclusions we have drawn with respect to the method of exposure dating of 26 36 rock surfaces with 10Be, Al, and CI pertain to the chemical extraction procedures and 10 26 the production rates of Be and Al which we have calculated from our data. 26 1) the extraction method for 10Be and Al from quartz which we have set-up and which was modeled after the method originally developed by Nishiizumi et al. (1989) and Kohl and Nishiizumi (1992), is extremely robust. It can be used reliably and reproducibly to separate quartz and then extract Be and Al from quartz from a variety of rock types. Although rather time-consuming, the extraction steps are simple and straightforward and not subject to failure due to small variations in the procedures. A number of sequential dissolution tests proved the reproducibility of the method. They 10 also showed with our extraction protocol all meteoric Be can be completely removed. Further development and refinement of the extraction procedure are needed in order to analyze rock samples having quartz grains which contain certain mineral inclusions and fine-grained quartz especially from metamorphic rocks. But there are several additional chemical extraction steps available which can be employed after testing. In general, steps to improving the purity of the final Be or Al fraction would lead to lower AMS uncertainties and therefore, lower uncertainties for the resulting exposure ages. 2) Detailed sequential dissolution tests of pyroxene (from Sirius Group·dolerite clasts) led to the conclusion that in some cases pyroxene can be irreversibly contaminated with 10 10Be from the atmosphere. In the atmosphere, Be is found in quantities orders of 10 magnitude higher than the expected in situ Be level in a mineral. As a result we were 10 not able to measure in situ Be in pyroxene, and therefore could not perform the 3 21 useful comparison with He and Ne (also in situ produced cosmogenic isotopes) which had been measured in aliquots of the same mineral separates (Bruno 1995). The contamination with meteoric 10Be may be unique to rock surfaces with old crystallization ages and that have been exposed for a long time. Pyroxenes from 162

10 recently erupted lava flows may prove suitable for Be exposure dating, especially if the individual grains have not been affected by any type of secondary alteration. 3) Although the majority of the exposure dates we have determined are from 10 Be and 26 Al measurements, we have also set-up the extraction procedure for Cl from whole rock. It was based on the method described by Zreda (1994) and Phillips et al. (1986). 36 Although CI exposure dating has the drawback that many elements must as well be determined in each rock sample (because of the multiple production mechanisms of 36 CI), it has the distinct advantage that any rock type can be used. That means that the rocks investigated need not be limited to those containing abundant large quartz grains 10 26 (preferred for Be and AI). The Cl extractions from whole rock (Julier Pass samples) which we completed yielded exposure ages in very good agreement with 10 Be ages which was measured indeprndently on quartz separated from the same crushed rock. This provides an important and successful intercomparison of the isotope systematics. More work along this line would be useful. By building-up a data set where all three 10 26 36 isotopes ( Be, AI, and CI) have been measured in the same rock, one may closely examine the production rate of each isotope. Perhaps more importantly, because of the difference in half-lives and the various production mechanisms of these three isotopes, intercomparison cc:n reveal the history of the rock surface itself, i.e. whether it has been periodically covered or how much erosion it has undergone. 10 26 4) Be and AI production rates calculated from our Antarctic and Julier Pass data are 5.7 and 5.6; and 37.3 and 37.0 atoms/gram quartz, respectively. These were calculated from the results from eight rocks from Antarctica and five rocks from Ju lier 10 21 Pass. In the former case, we assumed saturation for both Be (based on Ne 26 10 measurements, Bruno 1995) and Al (based on the Be ages), and no erosion. For Ju lier Pass we took the known age of the Younger Dryas and calculated production rates from the measured concentrations. We made no corrections for snow cover. The 10 Be values are slightly lower than the accepted production rates (6.0, Nishiizumi et al. 1989), but are within the stated errors for the production rates (5-7 %) and the scaling factors (10 %) (Lal 1991 ).

We addressed problem~: of the timing of glacier movements in two geographic areas: the Swiss Alps and Antarctica. The timing of glacier fluctuations deduced from our exposure dates as well as the additional information one may glean from the cosmogenic nuclide concentrations are outlined below. The question of the stability of the Antarctic Ice Sheets has important implications with respect to forecasts of global warming. This is because of the immense quantity of water they lock up. Our investigation went into not only the age of the Sirius Group deposits but of the antiquity of related Dry Valleys landforms. This provides basic 163 information about how old the landscape is and therefore how long the Antarctic Ice Sheets have existed in their pre!;ent form. 1) A single sandstone dast from the Sirius Group outcrop at Mount Feather was analyzed 10 26 for se and AI. The exposure age of 2.3 Ma is interpreted to be a minimum age. In this rock surface, the concentration of the cosmogenic isotopes is probably erosion controlled. The rate was less than 15 cm/Ma. Measuring a depth profile into the Mount Feather Sirius deposit may provide more information about erosion rates at this site. 2) A minimum exposure age of 2.9 Ma has been determined for the outcrop of the Sirius 10 26 Group on Table Mountain. This is based on 12 Be and Al exposure ages measured in samples from five different granite boulders from the margin of the Sirius outcrop and from one sandstone clast found just outside the outcrop on the eroded Ferrar dolerite surface. The concentration of large clasts on the top of the deposit may have formed as a lag deposit. From our data we have calculated erosion/weathering rates of no more than 70 cm/Ma, and no more than 45 cm/Ma when we neglect one sample that appears to have had a complicated exposure history. This provides an indication of the rate of wind-deflation of the finer grained sediment and the rate at which larger clasts are being uncovered. 3) Sixteen 10Be and 26AI ages were calculated from nine rock surfaces from the plateau at 10 Table Mountain. The minirrum Be exposure age for this surface is 5.5 Ma. The weathering rates of the Beacon Sandstone for this location range from 0 to 30 cm/Ma, with wind-polished silicified bedrock surfaces exhibiting the lowest rates. There is no geomorphologic evidence t:i indicate that the plateau surface at Table Mountain antedates the deposition of the Sirius till onto it, therefore the latter may be interpreted to be at least 5.5 million years old, as well. 10 4) Based on two sandstone sa1nples, a minimum Be exposure age of 5.8 Ma has been 26 determined for the Sirius Group outcrop at Mount Fleming. AI measurements 10 10 confirm that no meteoric Be contributed to the high Be concentrations we 10 measured from these samples. Such high Be concentrations do not allow erosion/weathering rates of more than 2 cm/Ma. These ages indicate that it is highly unlikely that the sediments of the Sirius Group were deposited at Mount Fleming after 3.0-2.5 million years ago, as required by the Pliocene deglaciation hypothesis (e.g. Webb and Harwood 1991) . 5) The very high 10Be concentrations we have measured at these three sites (Table Mountain, Mount Feather and Mount Fleming) are not compatible with the rapid uplift of these surfaces during the last 3 million years. For many of the rock surfaces the concentration of 10Be which we have measured must have accumulated at the present altitude and for a period of at least ca. 5 million years. Such measurements should be extended to other tectonic blocks of the Transantarctic Mountains to constrain uplift. 164

Because of the complex relationships of the coarse-grained glacial deposits of the northern Swiss foreland and as a result of a lack of organic material for radiocarbon dating, we have begun to apply surface exposure dating to help with the unraveling of the chronology of the glaciations of the Swiss Alps. The time range we have investigated thus far goes from the maximum extent of the Last Glaciation as recorded by the Rhone Glacier to the end of the Younger Dryas, based on an Egesen moraine complex at Ju lier Pass. 1) We have determined a minimum time for the deglaciation of Grimsel Pass of ca. 12,000 years ago. We interpret these exposure ages from glacially polished surfaces to be minimum ages because possible coverage by snow or sediment is difficult to constrain on the horizontal surfaces we sampled. Our oldest age does overlap with the postulated time of deglaciation of the high passes (i.e. B0lling) based on pollen data. Future sampling of high angle glacially polished surfaces should help to rule out the problem of shielding of cosmic rays due to coverage by snow or sediment. The age of the older generation of polished surfaces is an important but unknown parameter in reconstructions of ice volume for the extensive glaciations prior to the LGM that needs to be investigated. 2) The Egesen moraine compleK located at Julier Pass was constructed 11, 100 years ago 10 26 36 based on nine separate Be 1 Al and CI ages from six different boulders. We have

recognized an outer moraine which is 11 1 800 years old 1 and an inner moraine which is 10,400 years old. These msults provide additional evidence that Egesen moraines did indeed form during the Younger Dryas climatic deterioration. Our data from this site indicate that remarkable agreement can be obtained with exposure dates and more conventional means of dating (i.e. correlation with radiocarbon dated sites). 10 26 3) Five Be and AI exposure dates were measured in three erratics which mark the maximum extent of the Rhone Glacier near Wangen a. d. Aare. Based on these exposure dates, we have concluded that the Solothurn lobe of the Rhone Glacier stabilized at or just before 20,000 years ago and then had completely retreated from this site by 15,000 years ago. This provides evidence that the Rhone Glacier actually reached its farthest extent during the Last Glacial Maximum and not before. The Rhone and the Rhine Glaciers appear to have behaved roughly synchronously, although comparison with radiocarbon dates can be considered tenuous in this time range.

We have dated rock surfaces with both short (thousands of years) and long (millions of years) periods of exposure. Our sample preparation protocol can be used with equal reliability. Our results show the versatility and applicability of the method of rock surface dating with cosmogenic isotopes. By applying this method we have proven that, 165 for example, at Mount Fleming, the Sirius Group sediments there cannot be younger than 5.8 Ma. With regard to questions of timing of glacial expansion in Switzerland, we have provided absolute ages that verify that Egesen moraines did indeed form during the Younger Dryas climate reversal. Finally, the Solothurn lobe of the Rhone Glacier reached its maximum terminal position :;ynchronous with the world-wide ice volume maximum of the last glacial cycle. Surface exposure dating with cosmogenic isotopes is a powerful tool that allows the absolute dating of the surface of a rock and therefore opens a broad spectrum of possible problems one can address in geomorphology. 166


Alley, R. B., Meese, D. A., Shuman, C. A., Gow, A. J., Taylor, K. C., Gtootes, P. M., White, J. W. C., Ram, M., Waddington, E. D., Mayewski, P.A., and Zielinski, G. A., 1993, Abrupt increase in Greenland snow accumulation at the end of the Younger Dryas event: Nature v. 362, p. 527-529.

Amman, B., and Lotter, A. F., 1989, Late-glacial radiocarbon- and palynostratigraphy on the Swiss Plateau: Boreas, v. Hi, p. 1 09-126.

Amman, B., Lotter, A. F., Eicher, U., Gaillard, M.-)., Wohlfarth, B., Haberli, W., Lister, G., Maisch, M., Niessen, F., and Schluchter, C., 1994, The WUrmian Late-glacial in lowland Switzerland: Journal of Quaternary Science, v. 9, p. 119-125.

Andrews, J. N., Fontes, J.-C., Mic:helot, J.-L., and Elmore, D., 1986, In-situ neutron flux, 36CI production and groundwater evolution in crystalline rocks at Stripa, Sweden: Earth and Planetary Science Letters, v. 77, p. 49-58.

Andrews, J. N., Davis, S. N., Fabryka-Martin, J., Fontes, J.-C., Lehmann, B. E., Loosli, H. H., Michelot, J.-L., Moser, H., Smith, B., and Wolf, M., 1989, The in situ production of radioisotopes in rock matrices with particular reference to the Stripa Granite: Geochimica et Cosmochmica Acta, v. 5:~, p. 1803-1815.

Anthony, E. Y., and Poths, J., 19"2, 3He surface exposure dating and its implications for evolution in the Potrillo volcanic field, Rio Grande Rift, New Mexico, USA: Geochimica et Cosmo chm ica Acta, v. 56, ). 4105-41 08.

Askin, R. A., and Markgraf, V., 1986, Palynomorphs from the Sirius Formation, Dominion Range, Antarctica: Antarctic Journal of the United States, v. 21, p. 34-35.

Banfield, J. F., Ferruzzi, G. G., Casey, W. H., and Westrich, H. R., 1995, HRTEM study comparing naturally and e~perimentally weathered pyroxenoids: Geochimica et Cosmochmica Acta, v. 59, =i. 19-31 .

Banerjee, S. K., 1995, Chasing the Paleomonsoon over China: Its magnetic proxy record: GSA Today, v. 5, p. 93-97.

Bard, E., Hamelin, B., Fairbanks. R. G., and Zindler, A., 1990, Calibration of the 14C timescale over the past 30,000 years Jsing mass spectromic U-Th ages from Barbados corals: Nature, v. 345, p. 405-410.

Bard, E., Fairbanks, R. G., Arnold, M., and Hamelin, B., 1991, 230Th/234 U and 14 C ages obtained by mass spectrometry on corals from Barbados (West Indies), lsabela (Galapagos) and Mururoa (French Polynesia), in Bard, E., and Broecker, W. S., eds., The Last Deglaciation: Absolute and Radiocarbon Chronologies: Springer Verlag, Heidelberg, Germany, p. 103- 110.

Bard, E., Arnold, M., Fairbanks, R. G., and Hamelin, B., 1993, 2 ~h-234 U and 14C ages obtained by mass spectrometry on corals: Radiocarbon, v. 35, p. 191-199.

Barrett, P. J., 1991, The Devoni

Barrett, P. J., and Powell, R. D., 1982, Middle Cenozoic glacial beds at Table Mountain, southern Victoria Land, in Craddock, C., ed., Antarctic Geoscience Symposium on Geology and Geophysics, University of Wisconsin Press, Madison, USA, p. 1059-1067. 167

Barrett, P. J., Grindley, G. W., and Webb, P.-N., 1972, The Beacon Supergroup of East Antarctic, in Adie, R. J., ed., Antarctic Geology and Geophysics, Universitetsforlaget, Oslo, Noiway, p. 319-322.

Barrett, P. J., Adams, J. C., Mcintosh, W. C., Swisher Ill, C. C. and Wilson, G. S., 1992, Geochronological evidence supporting Antarctic deglaciation three million years ago: Nature, v. 359, p. 816-818.

Barrett, P. J., Hambrey, M. J., Harwood, D. M., Pyne, A. R., and Webb, P.-N, 1989, Synthesis, in Barrett, P. J., ed., Antarct c Cenozoic History from CIROS-1 Drill Hole McMurdo Sound, DSlR Bulletin, v. 245, p. 241-251, Wellington, New Zealand.

Barron, J. A., Baldauf, J. G., Barr.~ra, E., Caulet, J.-P., Huber, B. T. Keating, B. H., Lazarus, D., Sakai, H., Thierstein, H. R., and Wei, W., 1989, Biochronologic and magnetochronologic synthesis of ODP Leg 119 sediments from the Kerguelen Plateau and Prydz Bay, Antarctica, in Barron, J. A., et al., eds., Proc. Sci. Results, 119, Ocean Drilling Program, College Station, Texas, USA

Beck, P., ·1932, Ober den eiszei':lichen Aaregletscher und die Quartarchronologie: Verhandlungen der Schweizerischen Naturforschenden Gesellschaft Thun, v. 113, p. 189· 207.

Beer, J., Raisbeck, G. M., and Yiou, F., 1991, The variations of 10Be and solar activity, in Sonett, C. P., et al., eds., The Sun in Time: University of Arizona, Tucson, USA, p. 343-359.

Beer, J., Siegentaler, U., Bonani, G., Finkel, R. C., Oeschger, H., Suter, M., and Wolfli, W., 1988, Information on past solar activity and geomagnetism from 10Be in the Camp Century ice core: Nature, v. 331, p. 675-679.

Behrendt, J. C., and Cooper, A. K., 1991, Evidence of rapid Cenozioc uplift of the shoulder escarpment of the Cenozoic West Antarctic rift system and a speculation on possible climate forcing: Geology, v. 19, p. 315-319.

Berner, R. A., and Schott, J., 1982, Mechanisms of pyroxene and amphibole weathering II: Observations of soil grain•;: American Journal of Science, v. 207, p. 1214-1231.

Berner, R.A., Sjoberg, E.L, Veloel, M.A., and Krom, M.D., 1980, Dissolution of pyroxenes and amphiboles during weath,~ring, Science, v. 207, p. 1205-1206.

Bierman, P., and Larsen, P., lS,96, Refining estimates of 10Be and 26AI production rates: Radiocarbon, v. 38, p. 149.

Birkeland, P. W., 1982, Subdivision of Holocene glacial deposits, Ben Oahu Range, New Zealand, using relative dating methods: GSA Bulletin, v. 93, p. 433-449.

Blavoux, B., 1988, l'occupation de la Cuvette Lemanique par le glacier du Rhone au cours du Wurm: Bulletin de I' Association franc;:aise pour l'etude du Quaternaire, v. 1988-2/3, p. 69- 79.

Bloemendal, J., Liu, X. M., and Rolph, T. C., 1995, Correlation of the magnetic susceptibility of Chinese loess and the marine oxygen isotope record: Chronological and paleoclimatic implications: Earth and Planetary Science Letters, v .131, p. 371-380.

Bolliger, T., Fejfar, 0., Graf, H., and Kalin, D., 1996, Vorlaufige Mitteilung Ober Funde von pliozane Kleinsilugern a JS den hoheren Deckenschotter des lrchels (Kt. Zurich): Eclogae Geologicae Helvetiae, v. 89, p. 1043-8. 168

Bond, G.C., and Lotti, R., 1995, Iceberg discharges into the North Atlantic on millennial time scales during the last glaciation: Science, v. 267, p. 1005-1010.

Bond, G., Heinrich, H., Broecker, 'N., Labeyrie, L., McManus, J., Andrews, J., Huon, S., Jantschik, R., Clasen, S., Simec, C., Tedesco, K., Klas, M., Bonani, G., and Ivy, S., 1992, Evidence for massive discharges of icebergs into the North Atlantic ocean during the last glacial period: Nature, v. 360, p. 245-249.

Bond, G., Broecker, W., Johnsen,~ .. , McManus, J., Labeyrie, L., Jouzel, J., and Bonani, G., 1993, Correlations between climate records from North Atlantic sediments and Greenland ice: Nature, v. 365, p. 143-147.

Bortenschlager, S., 1984, Beitrage zur Vegetationsgeschichte Tirols I. lnneres btztal und unteres lnntal: Berichte des naturwissenschaftlich - medizinischen Vereins Innsbruck, v. 71,p.19 56.

Bourles, D. L., Brown, E.T., Raisbeck, G. M., Yiou, F., and Gieskes, J.M., 1992, Beryllium isotope of hydrothermally altered sediments: Earth and Planetary Science Letters, v. 109, p. 47-56.

Brady, H., and McKelvey, B., 1979, The interpretation of a Tertiary tillite at Mount Feather, Southern Victoria Land, Antarctica: Journal of Glaciology, v. 22, p. 189-193.

Brady, H., and McKelvey, B., 1982:, Some aspects of the Cenozoic glaciation of southern Victoria Land, Antarctica: Journal of Glaciology, v. 29, p. 343-349.

Breza, J.R., 1992, High resolution study of Neogene ice-rafted debris, ODP Leg 120 Site 751, southern Kerguelen Plateau, in Wise, S.W., et al., eds., Proceedings of Scientific Results 120, Ocean Drilling Program. College Station, Texas, USA, p. 207-221.

Breza, J.R., and Wise, S.W., 1992, lower Oligocene ice-rafted debris on the Kerguelen Plateau: Evidence for East Antarctic continental glaciation, in Wise, S.W., et al., eds., Proceedings of Scientific Results 120, Ocean Drilling Program, College Station, Texas, USA.

Broecker, W. S., 1992, Defining the boundaries of the BOA warm and YD cold isotope episodes: Quaternary Research, v. 38, p. 135-139.

Broecker, W. S., and Denton, G. H., 1989, The role of ocean-atmosphere reorganizations in glacial cycles: Geochimica et Cosmochimica Acta, v. 53, p. 2465-2501.

Brook, E. J., 1994, Surface exposure chronology using cosmogenic nuclides: Applications in Antarctic glacial geology: Ph.D. thesis, Massachusetts Institute of Technology-Woods Hole Oceanographic Institution Joint Program in Oceanography (WHOI Document 93-50), Cambridge and Woods Hole, USA, 230 p.

Brook, E. J., and Kurz, M. D., 1993, Surface-exposure chronology using in situ cosmogenic 3He in Antarctic quartz sandstone boulders: Quaternary Research, v. 39, p. 1-10.

Brook, E. J., Kurz, M. D., Ackert, R. P., Raisbeck, G. M., and Yiou, F., 1995a, Cosmogenic nuclide exposure ages and glacial history of late Quaternary Ross Sea drift in McMurdo sound, Antarctica: Earth and Planetary Science Letters, v. 131, p. 41-56.

Brook, E. J., Nesje, A., Lehman, S. J., Raisbeck, G. M., and Yiou, F., 1996, Cosmogenic nuclide exposure ages along a vertical transect in western Norway: implications for the height of the Fennoscandian ice sheet: Geology, v. 24, p. 207-21 O. 169

Brook, E. J., Brown, E.T., Kurz, M. D., Ackert, R. P., Raisbeck, G. M., and Yiou, F., 1995b, Constraints on age, erosion and uplift of Neogene glacial deposits in the Transantarctic Mountains using in situ cosmogenic 10Be and 26 Al: Geology, v. 23, p. 1063-1066.

Brook, E. J., Kurz, M. D., Ackert, IL,P., Denton, G. H., Brown, E. T., Raisbeck, G. M., and Yiou F., 1993, Chronology of Taylor Glader advances in Arena Valley, Antarctica, using in situ cosmogenic 3He and 10Be: Quaternary Research, v. 39, p. 11-23.

Brown, E.T., 1992, The geochemistry of beryllium isotopes: Applications in Geochronometry: Ph.D. thesis, Massachusetts Institute of Technology-Woods Hole Oceanographic Institution Joint Program in Oceanography (WHOI Document 90-47), Cambridge and Woods Hole, USA, 273 p.

Brown, E.T., Brook, E. )., Raisbeck, G. M., Yiou, F., and Kurz, M. D., 1992, Effective attenuation lengths of cosmic rays producing 10Be and 26 Al in quartz: Implications for exposure age dating: Geophysical Research Letters, v. 19, p. 369-372.

Brown, E.T., Stallard, R. F., Lars.m, M. C., Raisbeck, G. M., and Yiou, F., 1995, Denudation rates determined from the accumulation of in-situ produced 10Be in the Luquillo Experimental Forest, Puerto Rico: Earth and Planetary Science Letters, v. 129, p. 193-202.

Brown, E. T., Edmond, J. M., Raisbeck, G. M., Yiou, F., Kurz, M. D. and Brook, E. J., 1991, 10 Examination of surface exposure ages of Antarctic moraines using in situ produced Be and 26AI: Geochimica et Cosmochimica Acta, v. 55, p. 2269-2283.

Bruno, L., 1995: Datierung von Moranen und glazial uberpragten Oberflachen in der Antarktis mit in-situ produzierten ko:imogenen Edelgasen: ETH Dissertation No. 11 256, Zurich, Switzerland, 119 p.

Buhler, C., 1983, Petrographisc~e und geochemische Untersuchungen im Gebiet La Tscheppa- Lagrev (Julier-Bernina-Decke): Schweizerische mineralogische und petrographische Mitteilungen, v. 63, p. 457-477.

Bull, W. B., 1991, Geomorphic Responses to Climatic Change: , Oxford, UK.

Burckle, L. H., and Pokras, E. M., 1991, Implications of a Pliocene stand of Nothofagus (southern beech) within 500 kilometres of the South Pole: Antarctic Science, v. 3, p. 389- 403.

Burckle, L. H., and Potter Jr., N., 1996, Pliocene-Pleistocene diatoms in Paleozoic and Mesozoic sedimentary and igneous rocks from Antarctica: A Sirius problem solved: Geology, v. 24, p. 235-22-8.

Burga, C. A., 1980, Pollenanalystische Untersuchungen zur Vegetationsgeschichte des Schans und des San Bernardino-Passgebietes (Graubunden, Schweiz): Dissertationes Botanicae, v. 56, p. 1-194.

Burga, C. A., 1982, Oberblick zur palynologischen Erforschung GraubOndens 1929 - 1982, in Camper, M., ed., Physis<:he Geographie: Beitrage zur Quartarforschung In der Schweiz: Geographisches lnstitut der Universitat, Zorich, Switzerland, p.

Burga, C. A., 1988, Swiss vegetation history during the last 18000 years: New Phytologist, v. 110, p. 581-602. 170

Campbell, I. B., and Claridge, G. G. C., 1978, Soils and late Cenozoic history of the Upper Wright Valley area, Antarctica: New Zealand Journal of Geology and Geophysics, v. 21, p. 635-643.

Campbell, I. B., and Claridge, G. G. C., 1987, Antarctica: Soils, Weathering Processes and Environment: , Amsterdam, Netherlands.

Casey, W. H., Westrich, H. R., Banfield, J. F., Ferruzzi, G., and Arnold, G. W., 1993, Leaching and reconstruction at the surfaces of dissolving chain-silicate minerals: Nature, v. 366, p. 253-256.

Castagnoli, G., and Lal, D., 1980, Solar modulation effects in terrestrial production of carbon- 14: Radiocarbon, v. 22, p. 1.33-158.

Castagnoli, G., Albrecht, A., Beer,)., Bonino, G., Shen, C., Callegari, E., Taricco, C., Dittrich- Hannen, B., Kubik, P., Suter, M., and Zhu, G. M., 1995, Evidence for enhanced 10 Be deposition in Mediterranean sediments 35 kyr BP: Geophysical Research Letters, v. 22, p. 707-710.

Cerling, T.E., 1990, Dating geomorphologic surfaces using cosmogenic 3He: Quaternary Research, v. 33, p. 148-156.

Cerling, T. E., and Craig, H., 1994a, Cosmogenic 3He production rates from 39 °N to 46 °N latitude, western USA and France: Geochimica et Cosmochimica Acta, v. 58, p. 249-255.

Cerling, T. E., and Craig, H., 1994b, Geomorphology and in-situ cosmogenic isotopes: Annual Review of Earth and Planeta1·y Sciences, v. 22, p. 273-317.

Cerling, T. E., Poreda, R. )., and Rathburn, S. L., 1994c, Cosmogenic 3He and 21 Ne age of the Big Lost River Flood, Snake River Plain, Idaho: Geology, v. 22, p. 227-230.

Chappell,)., and Shackleton, N. )., 1986, Oxygen isotopes and sea level: Nature, v. 324, p. 137-140.

Claridge, G. G. C., and Campbell, I. B., 1978, Moraines of possible Miocene age, Dry Valleys, Antarctica: New Zealand Antarctic Records, v. 1, p. 1-5.

Clark, D. H., and Gillespie, A. R., 1996, Timing and significance of late-glacial and Holocene glaciation in the Sierra Nevada, California: Quaternary International (in press).

Clark, D. H., Bierman, P. R., and Larsen, P., 1995, Improving in situ cosmogenic chronometers: Quaternary Research, v. 44, p. 366-376.

Clark, P. U., and Bartlein, P. J., 1995, Correlation of late Pleistocene glaciation in the western United States with North Atlantic Heinrich events: Geology, v. 23, p. 483-486.

Clark, P. U., Clague, J. )., Curry, B. B., Dreimanis, A., Hicock, S. R., Miller, G. H., Berger, G. W., Eyles, N., Lamothe, M., Miller, B. B., Mott, R. )., Oldale, R. N., Stea, R. R., Szabo, J. P., Thorleifson, L. H., and Vincent, J.-S., 1993, Initiation and development of the Laurentide and Cordilleran ice sheets following the last interglaciation: Quaternary Science Reviews, v. 12, p. 79-114.

Colin, F., Nahon, D., Delvigne, J., and Michaux, L., 1993, Secondary-mineral formation during natural weathering of pyroxene: review and thermodynamic approach: American Journal of Science, v. 293, p. 111-1.34. 171

Cornelius, H.P., 1935/1950/1951: Geologie der Err-Julier-Gruppe. Erlauterungstext zur geologischen Karte. Kummerly und Frey, Bern, Switzerland.

Craig, H., and Poreda, R. J., 1986, Cosmogenlc 3He In terrestrial rocks: the summit of Maui: Proceedings of the National Academy of Science, v. 83, p. 1970-1974.

Creer, K. M., 1988, Geomagnetic field and radiocarbon activity through Holocene time, in Stephenson, F. R., and Wolfondale, A. W., eds., Secular Solar and Geomagnetic Variations in the Last 10,000 Years: Kluwer Academic Publishers, Dordrecht, The Netherlands, p. 381-397.

Dansgaard, W., Johnsen, S. J., Clausen, H. B., Dahl-Jensen, D., Gundestrup, N. S., Hammer, C. U., Hvidberg, C. S., Steffensen, J.P., Sveinbjornsdottir, A. E., Jouzel, )., and Bond, G., 1994, Evidence for general instability of past climate from a 250-kyr ice-core record: Nature, v. 364, p. 218-220.

Davis Jr., R., and Schaeffer, 0. A., '1955, Chlorine-36 in nature: Annals of the New York Academy of Science, v. 62, p. 105-122. de Beaulieu, J.-L., and Rellle, M .. 1992, The last climatic cycle at La Grande Pile (Vosges, France): A new pollen profile: Quaternary Science Reviews, v. 11, p. 431-438.

Denton, G. H., and Hendy, C. H., 1994, Younger Dryas age advance of Josef Glacier in the Southern Alps of New Zealand: Science, v. 264, p. 1434-1437.

Denton, G. H., Prentice, M. L., and Burckle, L H., 1991, Cenozoic history of the Antarctic Ice Sheet, in Tingey, R. J,, ed., The Geology of Antarctica: Oxford University Press, Oxford, UK, pp. 365-433.

Denton, G. H., Sugden, D. R., Marchant, D. R., Hall, B. L. and Wilch, T. I., 1993, East Antarctic Ice Sheet sensitivity to Pliocene climate change from a Dry Valleys perspective: Geograflska Annaler, v. 75A, p.155-204.

Denton, G. H., Armstrong, R. L., and Stulver, M., 1971, The Late Cenozoic glacial history of Antarctica, in Turekian, K. K., ed., The late Cenozoic Glacial Ages: Yale University Press, New Haven, p. 267-306.

Dep L., 1994, Cosmogenic nuclide production in terrestrial rocks: accelerator mass spectrometry measurements and monte carlo simulations: Ph.D. thesis, Purdue University, 155 p.

Dep, L., Elmore, D., Lipschutz, M.. Vogt, S., Phillips, F.M., Zreda, M. and Hossain, T.Z., 1994, Depth dependence of cosrnogenic neutron-capture-production 36Cl in a terrestrial rock: Nuclear Instruments and Methods, 11. B92, p. 301-307.

Dorn, R. I., and Phillips, F. M., 1991, Surface exposure dating: review and critical evaluation: Physical Geography, v. 12, p. 303-333.

Dowsett, H.J., and Cronin, T. M., 1990, High eustatic sea level during the middle Pliocene: Evidence from the southeastern U.S. Atlantic Coastal Plain: Geology, v. 18, p. 435-438.

Drewry, D.J., 1983, Antarctica: glaciological and geophysical folio: University of Cambridge, Cambridge, England.

Eggleton, R. A., 1986, The rela:.ion between crystal structure and silicate weathering rates, in Colman, S. M., et al., eds., Rates of Chemical Weathering of Rocks and Minerals, London, Academic Press, p. 21-40 172

Eicher, U., 1979, Die 180/1 60 unc 13C/12 C-lsotopenverhiiltnisse in spatglazialen SGsswasserkarbonaten und ihr Zusammenhang mit den Ergebnissen der Pollenanalyse. Ph.D. thesis, Universitat Bern, Bern, Switzerland, 203 p.

Eicher, U., 1987, Die spiitglaziale11 sowie die fruhpostglazialen Klimaverhaltnisse im Bereiche der Alpen: Sauerstoffisotopenkurven kalkhaltiger Sedimente: Geographica Helvetica, v. 2, p. 99-104.

Elmore, D., and Phillips, F. M., 1987, Accelerator mass spectrometry for measurement of long- lived rodioisotopes: Science, v. 236, p. 543-550.

Evin,)., Bintz, P., and Montjuvent. G., 1994, Human settlements and the last deglaciation in the French Alps: Radiocarbon, v. 36, p. 345-357.

Fabryk<1-Martin, J. T., 1988, Production of radionuclides in the earth and their hydrogeologk significance, with emphasis rn chlorine-36 and iodine-129: Ph.D. dissertation, University of Arizona, Tucson, USA.

Feige, Y., Oltman, B. G., and Kaisner, J., 1968, Production rates of neutrons in soils due to natural radioactivity: Journal of Geophysical Research, v. 73, p. 3135-3142.

Fink, D., Klein,)., Middleton, R., Vogt, S., and Herzog, G. F., 1990, 41 Ca in Iron Falls, Grant and Estherville: production rates and related exposure age calculations: Earth and Planetary Science Letters, v. ·107, p. 115-128.

Finkel, R., and Suter, M., 1993, AMS in the earth sciences: technique and applications: Advances in Analytical Geoc1emistry, v. 1, p. 1-114.

Firbas, F., 1949, Spat- und nachei,,zeitliche Waldgeschichte Mitteleuropas nOrdlich der Alpen: v. 1, , Jena, 480 p.

Fitzgerald, P.G., 1992, The Transantarctic Mountains of southern Victoria Land: The application of apatite fission track analysis to a rift shoulder uplift: Tectonics, v. 11, p. 634- 662.

Fitzgerald, P.G., Sandiford, M., Barrett, P. J., and Gleadow, A. J. W., 1986, Asymmetric extension asociated with upli't and sL1bsidence of the Transantarctic Mountains and Ross Embayment: Earth and Planetary Science Letters, v. 81, p. 67-78.

Frey, M., Bucher, K., Frank, E., and Mullis, J., 1980, Alpine metamorphism along the geotraverse Basel-Chiasso. A Reviwe: Eclogae Geolicae Helvetiae, v. 73, p . .527-546.

Friedmann, E. I., and Weed, R., 1987, Microbial trace-fossil formation, biogenous, and abiotic weathering in the Antarctic cold desert: Science, v. 236, p. 703-705.

Fronval, T., Jansen, E., Bloemenda:, J., and Johnsen, S., 1995, Oceanic evidence for coherent fluctuations in Fennoscandian and Laurentide ice sheets on millennium time scales, Nature, v. 374, p. 443-446.

Fullerton, D. S., and Richmond, G M., 1986, Comparison of the marine oxygen isotope record, the eustatic sea level record, and the chronology of glaciation in the United States of America: Quaternary Science Reviews, v. 5, p. 197-200.

Furrer, G., 1991, 25000 Jahre Gletschergeschichte: VerOffentlichungen der Naturforschenden Gesellschaft in Zurich, Neuja 1rsblatt auf das Jahr 1991, 52 p. 173

Furrer, G., Burga, C. A., Gamper, M., Holzhauser, H. P., and Maisch, M., 1987, Zur Gletscher-, Vegetations- und Klimageschichte der Schwelz selt der Spateiszeit: Geographica Helvetica, v. 42, p. 61-91.

Furrer, H., 1948, Das Quartir zwischen Solothurn und Wangen a. A.: Eclogae Geologica Helvetica, v. 41, p. 269-284.

Gaillard, M.-J., 1965, Late-glacial and Holocene environments of some ancient lakes in the Western Swiss Plateau: Dissertationes Botanicae, v. 87, p. 273-336.

Gamper, M., and Suter,)., 1982, Postglaziale Klimageschichte der Schweizer Alpen: Geographica Helvetica, v. 37, p. 105-114.

Geyh, M.A., and Schreiner, A., 1984, 14C-Datierungen an Knochen- und Stosszahnfragmenten aus Wllrmeiszeitlichen Abl.agerungen im westlichen Rheingletschergebiet (Baden- Wllrttemberg): Eiszeit!ater und Gegenwart, v. 34, p. 155-161.

Gillespie, A., and Molnar, P., 1995, Asynchronous maximum advances of mountain and continental glaciers: Reviews of Geophysics, v. 33, p. 311-364.

Gleadow, A.). W., and Fitzgerald, P.G., 1987, Uplift history and structure of the Transantarctic Mountains: New evidence from fission track dating of basement apatites in the Dry Valleys area, southern Victoria Land: Earth and Planetary Science Letters, v. 82, p. 1-14.

Gore, D. B., and Colhoun, E. A., 1995, Regional contrasts In terrain and glacial sediments, Vestfold Hills, East Antarctica: VII International Symposium on Antarctic Earth Sciences, Siena (abstracts), p. 163.

Gosse ). C., Evenson, E. B., Klein, J., Lawn, B., and Middleton, R, 1995a, Precise cosmogenic 10 Be measurements in we,;tern North America: Support for a global Younger Dryas cooling event: Geology, v. 23, p. 677-880.

Gosse, J. C., Klein, J., Evenson, E. B., Lawn, B., and Middleton, R., 1995b, Beryllium-10 dating of the duration and retreat of the last Pinedale glacial sequence: Science, v. 268, p. 1329- 1333.

Graf, H. R., 1993, Die Deckenschotter der zentralen Nordschwelz: ETH dissertation Nr. 10205, Zurich, 151 p.

Graf, T., Kohl, C. P., Marti, K., and Nlshllzuml, K., 1991, Cosmic-ray produced neon in Antarctic rocks: Geophy,;ical Research Letters, v. 18, p. 203-206.

Graham, R. F., and Harwood, D. M., 1995, In-situ 10Be and 26AI concentrations of glacial erratics collected in the uominlon Range, Antarctica: Implications for exposure history of the glacigene Sirius Group sediments, , VII International Symposium on Antarctic Earth Sciences Siena, (Abstracts), p. 167.

Greuel!, J. W., 1989, Glaciers and Climate: Ph.D. thesis, Rijksuniversiteit Utrecht, Utrecht, Netherlands, 181 p.

Grobe, H., FUtterer, D.K., and SplefS, V., 1990, Oligocene to Quarternary sedimentation processes on the Antarctic Continental Margin, ODP Leg 113, Site 693, in Barker, P.G., Kennett, J.P., et al., Proc. lnit. Repts, 113, Ocean Drilling Program, College Station, Texas, U.S.A. 174

Gross, G., Kerschner, H., and Pat~elt, G., 1978, Methodische Untersuchungen Ober die Schneegrenze in alpinen Gletschergebieten: Zeitschriit f(]r Gletscherkunde und Glazialgeologie, v. 12, p. 223-251.

Gross, H., 1958, Die bisherigen Ergebnisse von 14C-Messungen und palaolithischen Untersuchungen IUr die Gliederung und Chronologie des Jungpleistozans in Mitteleuropa und den Nachbargebieten: Eiszeitlater und Gegenwart, v. 9, p. 155-187.

Gunn, B. M., and Warren, G., 1962, Geology of Victorialand between the Mawson and Mullock Glaciers: New Zealand Geological Survey Bulletin, v. 71, 157 p.

Haberli, W., 1991 a, Glazialmorphologische und paleoglaziologische Model le, in Monbaron, M., and Haberli, W., eds., Modelle in der Geomorphologie Beispiele aus der Schweiz: Berichte und Forschungen Geographisches lnstitut fribourg/CH, v. 3, p. 7-10.

Haberli, W., 1991 b, Zur Glaziologie der letzteiszeitlichen Alpenvergletscherung, in Frenzel, B., ed., Klimageschichtliche Probleme der letzten 130000 Jahre: Paleoklimaforschung, v. 1, p. 409-419.

Haberli, W., 1994, Accelerated glacier and permafrost changes in the Alps, in Beniston, M., et al., eds., Mountain Environments in Changing Climates: London, , p. 91-107.

Haberli, W., and SchlUchter, C., 1987, Geological evidence to constrain modelling of the Late Pleistocene Rhonegletscher (Switzerland). The physical basis of ice sheet model Iing: IAHS Publications, v. 90, p. 333-'·46.

Hajdas, I., 1993, Extension of thE radiocarbon calibration curve by AMS dating of laminated sediments of Lake Soppensee and Lake Holzmaar: ETH Dissertation No. 10157, Zurich, 147 p.

Hajdas, I., Ivy-Ochs, S. D., and E:onani, G., 1995, Problems in the extension of the radiocarbon calibration curve (10-13 kyr BP): Radiocarbon, v. 37, p. 75-79.

Hajdas, I., Ivy-Ochs, S. D., Bonani, G., Lotter, A. F., Zolitschka, B., and SchlOchter, C., 1995, Radiocarbon age of the Laacher See Tephra: 11,230 ± 40 BP: Radiocarbon, v. 37, p. 149- 154.

Hajdas, I., Ivy, S. D., Beer, J., Bonani, G., Imboden, D., Lotter, A., Sturm, M., and Suter, M., 1993, AMS radiocarbon dating and varve chronology of Lake Soppensee: 6000 to 12000 14 C years: Climate , v. 9, p. 107-116.

Hall, B. L., Denton, G. H., Lux, D. R., and Bockheim, J. G., 1993, Late Tertiary Antarctic paleoclimate and ice-sheet dynamics inferred from surficial deposits in Wright Valley: Geografiska Annaler, v. 75A, p. 239-267.

Hallet, B., and Putkonen, J., 1994, Surface dating of dynamic landforms: young boulders on ageing moraines: Science, v.265, p. 937-940.

Hambrey, M. J., and Barrett, P. J., 1993, Cenozoic sediment and climatic record, Ross Sea region, Antarctica: in Kenn·ett, J. P. and Warnke, D. A., eds., The Antarctic Paleoenvironment: A Perspective on Global Change, 2: Antarctic Research Series, v. 60, p. 91-124.

Hantke, R., 1977, Eiszeitliche Stande des Rhone-Gletschers im westlichen Schweizer Mittelland, Berichte der N2.turforschenden Gesellschaft Freiburg i. Br., v. 67, p. 75-83.

Hantke, R., 1978-1983, Eiszeitalter. Die jungste Erdgeschichte der Schweiz und ihrer Nachbargebiete, Volume 1-3: Ott, Thun, Switzerland. 175

Haq, B. U., Hardenbol, J., and Vail, P. R., 1988, Mesozoic and Cenozoic and cycles of sea-level change, In Wilgus, D. K., et al., eds., Sea-level Changes - An Integrated Approach: Socie('I of Economic Paleontologists and Mineralogists Special Publication 42, p. 71-108.

Harrington, H.J., 1958, The nomenclature of rock units in the Ross Sea region: Nature, v. 182, p. 290.

Harwood, D. M., 1986, Diatom and paleoecology and a Cenozoic history of Antarctic ice sheets: Ph.D. dissertation, Ohio State University, Columbus.

Harwood, D. M., and Webb, P.·N., 1986, Recycled marine microfossils from basal debris-ice in ice-free valleys of southern Victoria Land: Antarctic Journal of the United States, v. 21, p. 87-88.

Hayes, D. E., and Frakes, L.A., 1975, General synthesis, Deep-Sea Drilling Project, Leg 28, in Hayes, D. E., et al., eds., hitial Reports of the Deep-Sea Drilling Project 28: U. S. Government Printing Office, Washington, D.C., U.S.A., p. 919-942.

Hays,). D., lmbrie, J., and Shackleton, N, J., 1976, Variations in the earth's orbit: Pacemaker of the ice ages: Science, v. 194, p. 1121-1132.

Heitz, A., Punchakunnel, P., and Zoller, H., 1982a, Zurn Problem der 14C-Datierung im Veltlin und Oberengadin, in Camper, M., ed., Physische Geographie: Beitrage zur Quartarforschung in der 'Schweiz: Geographisches lnstitut der Universitat, Zurich, Switzerland, p. 91-102.

Heitz, A., Punchakunnel, P., and Zoller, H., 1982b, Vegetations-, Klima- und Gletschergeschichte des Oberengadins, in Camper, M., ed., Physische Geographie: Beitrage zur Quartarforschung in der Schweiz: Geographisches lnstitut der Universitat, Zurich, Switzerland, p. '157-1 70.

Helfer, M., and Schluchter, C., 1995, The Antarctic Sirius Formation: Geological analysis of two sites: Vil International Symposium on Antarctic Earth Sciences, Siena, Italy (abstracts), p. 189.

Heller, F., and Liu, T.-S., 1932, Magnetostratigraphical dating of loess deposits in China: Nature, v. 300, p. 431-433.

Heuberger, H., 1966, Gletschergeschichtliche Untersuchungen in den Zentralalpen zwischen Sellrain und btztal: Wissenschaftliche Alpenvereinshefte, v. 20, 126 p.

Heuberger, H., 1968, Die 11lpengletscher irn Spat- und Postglazial: Eiszeitalter und Gegenwart 19.

Hildebrand, K., 1990, Das Endmorarengebiet des Rhonegletschers 0stlich von Wangen a. A.: diploma thesis, UnivNsity of Zurich.

Hill, R.S., 1989, fossil lea'', in Barrett, P.J., ed., Antarctic Cenozoic history from CIROS-1 drillhole McMurdo Sound: DSIR Bulletin, v. 245, p. 143-144,

Hill, R. S., and Truswell, L M., 1993, Nothofagus fossils in the Sirius Group, Transantarctic Mountains: Leaves and pollen and their climatic implications, in Kennett, J.P. and Warnke, D. A., eds., The Antarctic Paleoenvironment: A Perspective on Global Change, 2: Antarctic Research Series, v. 60, p. 67-73. 176

Hodell, D. A., and Venz, K., 1992, Toward a high-resolution stable isotope record of the Southern Ocean during the F'liocene-Pleistocene (4.8 to 0.8 Ma), in Kennett,). P., and Warnke, D. A., eds., The Antarctic Paleoenvironment: A Perspective on Global Change, 1: Antarctic Research Series, v. 56, p. 265-310.

Hofmann, H.J., Beer, J., Bonani, G., von Gunten, H. R,, Raman, S., Suter, M., Walker, R. L., Wolfli, W., and Zimmermann, D., 1987, 10Be: Half-life and AMS-standards: Nuclear Instruments and Methods, B:!9, p.32-36.

Holzhauser, H., 1982, Neuzeitliche Gletscherschwankungen: Geographlca Helvetica- Nr. 2, p.115-126.

Holzhauser, H., 1984, Zur Gescrichte des Aletsch-Gletscher und des Fieschergletschers: Physische Geographie, v. 13, 448 pp.

Holzhauser, H., 1985, Neue Ergebnisse zur Gletscher- und Klimageschichte des Spatmittelalters und der Neuzelt: Geographica Helvetica, v. 40, p. 168-195.

Holzhauser, H., 1987, Betrachtungen zur Gletschergeschichte des Postglazials: Geographica Helvetica, v. 42, p. 80-87. lmbrie, J., Hays, J. D., Martinson, D. G., Mcintyre, A, Mix, A L,, Morley, J. J., Pisias, N. G., Prell, W. L., and Shackleton, N. ),, 1984, The orbital theory of Pleistocene climate: Support 18 from revised chronology of the marine 0 record, in Berger, A. L,, et al., eds., Milankovich and Climate, Part 1: Reidel, Dordrecht, p. 269-305.

Ivy-Ochs, S,, SchlOchter, C., Kubik, P., and Beer, J., 1995, Das Alter der Egesenmorane am Julierpass: Geowissenschaften, v. 13, p. 313-315.

Ivy-Ochs, S., SchlOchter, C., Kubik, P. W., Dittrich-Hannen, B., and Beer,)., 1995, Minimum 10 Be exposure ages of early Pliocene for the Table Mountain plateau and the Sirius Group at Mount Fleming, Dry Valleys, Antarctica: Geology, v. 23, p. 1007-1010.

)ackli, H., 1962, Die Vergletsch·erung der Schweiz im Wurmmaximum: Eclogae Geologicae Helvetlae, v. 55, p. 285-294.

Jackli, A., 1970, Die Schweiz zur letzten Eiszeit: Atlas der Schweiz 6. Eidgenossiche Landestopographie, Wabe:·n-Bern.

Johnsen, S. J., Clausen, H. B,, Cansgaard, W., Fuhrer, K., Gundestrup, N. S., Hammer, C. U,, Iversen, P., )ouzel, J., Stauffer, B., and Steffensen, J. P., 1992, Irregular glacial interstadials recorded in a new Greenland ice core: Nature, v. 359, p. 311-313.

Jouzel, J., Lorius, C., Petit, J, R., Genthon, C., Barkov, N. I,, Kotlyakov, V. M., and Petrov, V. M., 1987, Vostok ice core: a continuous isotope temperature record overthe last climatic cycle (160000 years): Nature, v. 329, p. 403-407.

Jouzel, J., Vaikmae, R., Petit, J. R., Martin, M., Duclos, Y., Stievenard, M., Lorius, C., Toots, M., Melieres, M.A., Burckle, H., Barkov, N. I., and Kotlyakov, V. M,, 1995, The two-step shape and timing of the last deglaciation in Antarctica: Climate Dynamics, v. 11, p. 151- 161.

Jouzel, J., Barkov, N. I., Barnola, j, M., Bender, M., Chappelaz, J., Genthon, C., Kotlyakov, V. M., Lipenkov, V., Lorius, C., Petit, J. R., Raynaud, D., Raisbeck, G,, Ritz, C., Sowers, T., Stievenard, M., Yiou, F,, and Yiou, P., 1993, Vostok ice cores: extending the climate signal over the penultimate glacial period: Nature,v. 364, p. 407-412. 177

Keller, 0., and Krayss, E., 1987, Die hochwurmzeitlichen Rilckzugsphasen des Rhein- Vorlandgletschers und der erste alpine Eisrandkomplex im Spatglazial: Geographica Helvetica, v. 42, p. 169-178.

Kellogg, D. E., and Kellogg, T. B., 1996, Diatoms in South Pole ice: Implications for aeolian contamination of Sirius Group deposits: Geology, v. 24, p. 115-118.

Kellogg, T. B., and Kellogg, D. E., 1981, Pleistocene sediments beneath the Ross Ice Shelf: Nature v. 293, p. 130-133.

Kemp, E.M., 1978, Tertiary climatic evolution and vegetation history in the southeast Indian Ocean region: Paleoceanography, Paleoclimatology, and Paleoecology, v. 24, p. 169-208.

Kemp, E.M., and Barrett, P.J., 1975, Antarctic glaciation and early Tertiary vegetation: Nature, v. 258, p. 507 -508.

Kennett, J.P., 1977, Cenozo'c evolution of Antarctic glaciation, the Circum-Antarctic Ocean, and their impact on global paleoceanography: Journal of Geophysical Research, v. 82, p. 3843-3860.

Kennett, J. P., and Shackleton, N.J., 1976, Oxygen isotopic evidence for the development of the psychryosphere 38 m.y. ago: Nature, v. 260, p. 513-515.

Kennett,). P., and Barker, P.F., 1990, Climatic and oceanographic developments in the Weddell Sea, Antarcl ica, since the latest Cretaceous: an ocean-drilling perspective, in Barker, P.G., Kenner:, J.P., et al., Proc. lnit. Repts, 113, Ocean Drilling Program, College Station, Texas, U.S.A.

Kennett, J.P. and Hodell, D. A., 1993, Evidence for relative climatic stability of Antarctica during the early Pliocene: a marine perspective: Geografiska Annaler, v.75A, p. 205-220.

Kennett, J.P., and Hodell, D. A, 1995, Stability or instability of Antarctic Ice Sheets during warm climates of the Pliocene: GSA Today, v. 65, p. 1,2, 11,12, 13,22.

Kerschner, H., 1978, Untersuchungen zum Daun- und Egesenstadium in Nordtirol und Graubunden (melhodische Oberlegungen): Geographischer )ahresbericht aus Osterreich, v. 36, p. 26-49.

Kerschner, H., 1982, Zeugen der Klimageschichte im oberen Radurschltal - Alte Gletscherstande und Blockgletscher in der Umgebung des Hohenzollernhauses: Alpenvereinsjahrbuch 1982/83 (Zeitschrift des Deutschen und Osterreichischen Alpenvereins, v. 107, p. 23-27.

Kerschner, H., 1985, Quantitative paleoclimatic inferences from lateglacial snowline, timberline and rock glacier data, Tyrolean Alps, Austria: Zeitschrift fOr Gletscherkunde und Glazialgeologie, v. 21, p. 363-369.

Kleiber, H., 1974, Pollenanalytische Untersuchungen zum Eisrilckzug und zur Vegetationsg1'schichte im Oberengadin: Botanisches Jahrbuch Systematik, v. 94/1, p. 1-53.

Klein, J., Giegengack, R., Middleton, R., Sharma, P., Underwood, j. R. and Weeks, R. A., 1986, Revealing histories of exposure using in situ produced"'Al and 10Be in Libyan desert glass: Radiocarbon, v. 28, p. 547-.555.

Klein, j., Fink, D., Middleton, R., Vogt, S., Herzog, G. F., Reedy, R. C., Sisterson, ). M., Koehler, A. M., and Magliss, A., 1990, Average SCR flux during past 105 years: Inference from 41 Ca in Lunar rock 74275: Lunar and Planetary Science Conference, XXI. Lunar and Planetary Science Institute, Houston, Texas, USA, p. 635-636. 178

Kinzl, H., 1929, Beitrage zur Geichichte der Gletscherschwankungen in den Ostalpen: Zeitschrilt f(lr Gletscherkunde, v. 17, p. 66-121.

Knauer, M., 1994, Analyse der ~osmogenen Radionuklide 10 Be, 14C und 2bAI in Steinmeteoriten mittels Beschleunigermassenspektrometrie, Ph.D. Thesis, Universitat Hannover, Hannover, 181 p.

Kneubuehl, f., 1982, Repititorium der Physik: Teubner TaschenbUcher.

Kohl, C. P., and Nishiizumi, K., 1992, Chemical isolation of quartz for measurement of in-situ produced cosmogenic nuclides: Geochimica et Cosmochimica Acta, v. 56, p. 3583-3587.

Krantz, D. E., 1991, A chronology of Pliocene sea-level fluctuations: The U.S. Middle Atlantic Coastal Plain record: Qualernary Science Reviews, v. 10, p. 163-174.

Kromer, B., and Becker, B., 1993, German oak and pine 14C calibration, 7200 BC to 9400 BC: Radiocarbon, v. 35, p. 125-135.

Kubik, P. W., Ivy-Ochs, S., SchlOchter, C., and frank, M., 1996, Production rates for exposure dating: To be within the radiocarbon calibration curve: Radiocarbon, v. 38, p. 72-73.

Kukla, G., and An, Z. A., 1989, Loess stratigraphy in central China: Paleogeography, Paleoclimatology,and Paleoecology, v. 72, p. 203-225.

Kukla, G., Heller, f., Liu, X. M., Xu, T. C., Liu, T.-S., and An, Z, 1988, Pleistocene climates in China dated by magnetic susceptibility: Geology, v.16, p. 811-814.

Kurz, M. D., 1986, Cosmogenic helium in a terrestrial : Nature, v. 320, p. 435- 439.

Kurz, M. D., and Brook, E. )., ·1994, Surface exposure dating with cosmogenic nuclides, in Beck, C., ed., Dating in Exposed and Surface Contexts: University of New Mexico Press, Albuquerque, USA p. 139-159.

Kusakabe, M., and Ku, T.-L, 1984, Incorporation of Be isotopes and other trace metals into marine ferromanganese deposits: Geochimica et Cosmochimica Acta, v. 48, p. 2187- 2193,

KOttel, M., 1979, Pollenanalytische Untersuchungen zur Vegetationsgeschichte und zurn Gletscherruckzug in den westlichen Schweizer Alpen: Berichte der Schweizerischen Botanischen Gessellschaft, v. 89, p. 9-62.

Labhart, T. P., 1977, Aarmassiv und Gotthardrnassiv, in Gwinner, M. P., ed., Samrnlung Geologischer Fuhrer, Band 63, Gebrtider Borntrager, Berlin/Stuttgart, Germany.

Lal, D., 1988, In situ-produced cosmogenic isotopes in terrestrial rocks: Annual Review of Earth and Planetary Science, v. 16, p. 355-388.

Lal, D., 1991, Cosmic ray labeling of erosion surfaces: in situ nuclide production rates and erosion models: Earth and Planetary Science Letters, v. 104, p. 424-439.

Lal, D., and Peters, B., 1967, Cosmic-ray-produced radioactivity on the earth in Handbuch der Physik v. XLVl/2, Springer, Berlin, Germany p. 551-612.

Lal, D., and Arnold, ).R., 1985, Tracing quartz through the environment: Proceedings of the Indian Academy of Science (Earth and Planetary Sciences), v. 94, p. 1-5. 179

Landmann, G., Reimer, A., Lemcke, G., and Kempe, S., 1996, Dating late glacial abrupt climate changes in the l 4,570 years long continuous varve record of Lake Van, Turkey: PALAEOv.122,p.107-118.

Lang, G., 1985, Palynological research in Switzerland 1925-1985, in Lang, G., ed., Swiss Lake and Mire Environments during the last 15000 Years: Dissertationes Botanicae, v. 87, p. 11- 82.

Lang, G., and Tobolski, K., 1985, Hobschensee - late glacial and Holocene environment of a lake near the timberline, in Lang, G., ed., Swiss Lake and Mire Environments during the last 15000 Years: Dissertationes Botanicae, v. 87, p. 209-228.

Lang, R.M., 1991, Die Geologie zwishcen Briseck un Ettiswil: Diplom Arbeit ETH, Zurich.

Ledermann, H., 1978, Geologis.:her Atlas der Schweiz, Atlasblatt 72 (1127 Solothurn), mit Erlauterungen: Schweizerische Geologische Kommission, Bern, Switzerland, 36 p.

Lehman, B., Laj, C., Kissel, C. Mazaud, A., Paterne, M., and Labeyrie, L., I 996, Relative changes of the geomagnetic field intensity during the last 280 kyear from piston cores in the A~ores area: Physics of the Earth and Planetary Interiors, v. 93, p. 269-284.

Lichte, F. E., Hopper, S., and Osborn, T. W., 1980, Determination of silicon and aluminum in biological matrices by inductively coupled plasma emission spectrometry: Analytical Chemistry, v. 52, p. 120-1 24.

Lister, G. S., 1988, A 15,000-year isotopic record from Lake Zurich of deglaciation and climatic change in Switzerland: Quaternary Research, v. 29, p. 1 29-141.

Lorius, C., Jouzel, )., Ritz, C., Merlivat, L., Varkow, N. J., Korotkevich, Y. S., and Kotlyakov, V. M., 1985, A 150,000 year record from Antarctic ice: Nature, v. 316, p. 591-596.

Lotter, A. F., Eicher, U., Birks, H.J. B., and Siegenthaier, U., 1992, late-glacial climatic oscillations as recorded in Swiss lake sediments: Journal of Quaternary Science, v. 7, p. 187-204.

Lowell, T. V., Heusser, C. j., Andersen, B. G., Moreno, P. I., Hauser, A., Heusser, L. E., SchlOchter, C., Marchart, D. R., and Denton, G. H., 1995, lnterhemispheric correlation of Late Pleistocene glacial events: Science, v. 269, p. 1541-1549.

Maisch, M., 1981, Glazialmorphologische und gletschergeschichtliche Untersuchungen im Gebiet zwischen Landwasser- und Albulatal (Kt Graubunden, Schwelz): Physische Geographie, v. 3, 215 p.

Maisch, M., 1982, Zur Gle~cher- und Klimageschichte des alpinen Spatglazials: Geographica Helvetica, v. 37, p. 93-104.

Maisch, M., 1987, Zur GleNchergeschichte des alpinen Spatglazials: Analyse und Interpretation von Schneegrenzdater.: Geographlca Helvetica, v. 42, p. 63-71.

Maisch, M., 1992, Die Gletscher Graubundens. Rekonstruktion und Auswertung der Gletscher und deren Veranderungen seit dem Hochstand von 1850 Im Gebiet der ostlichen Schweizer Alpen (Bl:kdnerland und angrenzende Regionen). Tell A: Grundlagen- Analysen-Ergebnisse: Physische Geographie, v. 33, part A, 324 p., part B, 128 p.

Maisch, M., 1995, Gletscherschwundphasen Im Zeitraum des ausgehenden Spatglazials (Egesen-Stadium) und seit dem Hochstand von 1850 sowie Prognosen zum kOnftigen Eisruckgang in den Alpen, in Schweizerische Gletscherkommission, ed., Gletscher im 180

standigen Wandel: Zurich, Sv;itzerland, vdf Hochschulverlag AG an der ETH Zurich, p. 47-63.

Malin, M. C., 1985, Rates of geomorphic modification in ice-free areas, southern Victoria land, Antarctica: Antarctic Journal of the United States, v. 20, p. 18-20.

Malin, M. C., 1988, Abrasion in ice-free areas of southern Victoria land, Antarctica: Antarctic journal of the United States, v. 23, p. 38.

Mandier, P., 1988, Les problemes chronologiques poses par les phases de la recession WOrmienne dans la moyenne vallee du Rhone: Bulletin de I' Association frani;:aise pour l'etude du Quaternaire, v. 1988-2/3, p. 123-128.

Mangerud, J., Andersen, S. T., Berglund, B. E., and , J. J., 197 4, Quaternary stratigraphy of Norden, a proposal for terminology and classification: Boreas, v. 3, p. 109-127.

Marchant, D.R., Denton, G.H. and Swisher, C.C., Ill, 1993a, Miocene-Pliocene-Pleistocene glacial history of Arena Valley, Quatermain Mountains, Antarctica: Geografiska Annaler, v. 75A, p. 269-302.

Marchant, D. R., Denton, G. H., Sugden, D. E., and Swisher, C. C., Ill, 1993b, Miocene glacial stratigraphy and landscape =volution of the western Asgard Range, Antarctica: Geografiska Annaler, v. 75A, p. 303-330.

Marti, K., and Craig, H., 1987, Cosmic-ray-produced neon and helium in the summit lavas of Maui: Nature, v. 325, p. 335-337.

Martinson, D. G., Pisias, N. G., Hays, J. D., lmbrie, J., Moore Jr., T. C., and Shackleton, N. J., 1987, Age dating and the orbital theory of the ice ages: development of a high-resolution 0 to 300000-year chronostratigraphy: Quaternary Research, v. 27. p. 1-29.

Masarik, J., and Reedy, R. C, 1994, Effects of bulk chemical composition on nuclide production processes in meteorites: Geochimica et Cosmochimica Acta, v. 58, p. 5307-5317.

Masarik, J., and Reedy, R. C, 1995, Terrestrial cosmogenic-nuclide production systematics calculated from numerical simulations: Earth and Planetary Science letters, v. 136, p. 381- 396.

Masarik, J., and Reedy, R. C, 1996, Monte Carlo simulation of in-situ-produced cosmogenic nuclides: Radiocarbon (Abs.), v. 38, p. 163-164.

Matthews, R. K., and Poore, R. Z., 1980, Tertiary 180 record and glacioeustatic sea-level fluctuations: Geology, v. H, p. 501-506.

Mayewski, P.A., 1975, Glacial geology and late Cenozoic history of the Transantarctic Mountains, Antarctica: Institute of Polar Studies Report, No. 56, Ohio State University, Columbus, Ohio, USA.

Mayewski, P.A., and Goldthwaite, R. P., 1985, Glacial events in the Transantarctic Mountains: A record of the East Antarctic ice sheet, in Turner, M. D.,and Splettstoesser, J. R., eds., Geology of the Transantarctic Mountains, Antarctic Research Series, v. 36, p. 275-324.

Mayr, F., and Heuberger, H., ·1968, Type areas of Late Glacial and Post Glacial deposits in Tyrol, Eastern Alps, in Richmond, G.M., ed., Glaciation of the Alps: University of Colorado Studies, Series in Earth Sciences, v. 7, p. 143-165. 181

Mazaud, A., laj, C., Bard, E., Arnold, M., and Tric, E., 1991, Geomagnetic field control of 14C production over the last 80 ka: Implications for the radiocarbon timescale: Geophysical Research letters, v. 18, p. 1885-1888.

McCarrol, D., and Nesje, A., 1993, The vertical extent of ice sheets in Nordfjord, western Norway: Measuring degree of rock surface weathering: Boreas, v. 22, p. 255-265.

McElhinny, M. W., and Senanay;1ke, W. E., 1982, Variations in the geomagnetic dipole 1: The past 50,000 years: Journal of Geomagnetism and Geoelectricity, v. 34, p. 39-51.

McHargue, L. R., and Damon, P. E., 1991, The global beryllium 10 cycle: Reviews of Geophysics, v. 29, p. 141-158.

McHargue, L. R., Damon, P. E., and Donahue, D.)., 1995, Enhanced cosmic-ray production of 10Be coincident with the Mono Lake and laschamp excursions: Geophysical Research letters, v. 22, p. 659-662.

McKelvey, B. C., Webb, P.-N, Harwood, D. M., and Mabin, M. C. G., 1991, The Dominion Range Sirius Group- a record of the late Pliocene-Early Pleistocene Beardmore Glacier, in Thomson, M. R. A., et al., eds., The Geological Evolution of Antarctica: Cambridge University Press, Cambrid5e, UK, p. 675-682.

Meier, M. F., 1965, Glaciers and climate, in Wright, H. E., et al., eds., The Quaternary of the United States: Princeton University Press, Princeton, USA, p. 795-806.

Mercer, J. H., 1968, Glacial geology of the Reedy Glacier area, Antarctica: Geological Society of America Bulletin, v. 79, p. 471-486.

Mercer, J. H., 1972, Some observations on the glacial geology of the Beardmore Glacier area, in Adie, R. J., ed., Antarctic Geology and Geophysics: Universitetsforlaget, Oslo, Norway, p. 427-433.

Mercer, j. H., 1985, When did open-marine conditions last prevail in the Wilkes and Pensacola Basins, East Antarctica, and when was the Sirius Formation emplaced?: South African Journal of Science, v. B'r, p. 243-245.

Mercer, J. H., 1986, Southern Chile: A modern analog of the southern shores of the Ross Embayment during Pliocene warm intervals: Antarctic Journal of the United States, v. 21, p. 103-105.

Meynadier, L, Valet, J.-P., Weeks, R., Shackleton, N. J., and Hagee, V. l., 1992, Relative geomagnetic intensity of the field during the last 140 ka: Earth and Planetary Science Letters, v. 114, p. 39-57.

Mildenhall, D.C., 1989, Terrestrial palynology, in Barrett, P.J., ed., Antarctic Cenozoic history from CIROS-1 drillhole McMurdo Sound: DSIR Bulletin, v. 245, Wellington, New Zealand, p. 119-127.

Miller, K. G., Fairbanks, R G., and Mountain, G. S., 1987, Tertiary oxygen isotope synthesis, sea level history, and continental margin erosion: Paleoceanography, v. 2, p. 1-19.

Monaghan, M. C. Krishnaswami, S., and Thomas,). H., 1983, 10se concentrations and the long- term fate of particle-reactive nuclides in five soil profiles from California: Earth and Planetary Science letters, v. 65, p. 51-60. 182

Monjuvent, G., and Nicoud, G., 198Ba, Interpretation de la deglaciation Rhodanienne au Wurm, des moraines "intemes" a la Cuvette lemanique: Bulletin de I' Association franc;:aise pour i'etude du Qliaternaire, v. 1988-2/3, p. 129-140.

Monjuvent, G., and Nicoud, G., ·1988b, Modalites et chronologie de la deglaciation WOrmienne dans I' arc Alpin occidental et Jes Massifs franc;:ais: synthese et reflexions: Bulletin de I' Association fram;:aise pour l'etude du Quaternaire, v. 1988-2/3, p. 147-156.

Moscariello, A., 1996, Quaternary geology of the Geneva Bay (Lake Geneva, Switzerland): Sedimentary record, paieoewironmental and paleoclimate reconstruction since the last glacial cycle: Universite de Geneve dissertation No. 2814, 230 p.

MOiier, B., 1993, Zur Quartargeschichte de Seeztals (Kt. St. Gallen, Schweiz): ETH Dissertation No. 10184, Zurich.

MOiier, H. N., 1984, Spiitglaziale Gletscherschwankungen in den westlichen Schweizer Aipen (Simplon-SOd und Val de Nendaz, Wallis) und im nordislandischen Troilaskagi-Gebirge (Skidadalur): Nafels, Switzerland, Kung, 205 p.

MOiier, H., Kerschner, H., and l

MUiier, W. H .. Huber, M., Isler, A., and Kieboth, P., 1984, Geologische Karle der zentralen Nordschweiz (Geologishce Spezlalkarte Nr. 121 ), mit Erlauterungen: Schweizerische Geologische Kommission, Bern, Switzerland, 234 p.

Niedermann, S., Graf, T., Kim, J. S., Kohl, C. P., Marti K., and Nishiizumi, K., 1994, Cosmic- ray-produced 21 Ne in terrestrial quartz: the inventory of Sierra Nevada quartz separates: Earth and Planetary Science letters, v. 125, p. 341-355.

Niedermayer, M., 1995, Bestimmung van Tiefenprofilen der kosmogenen Radionuklide 1 OBe, 14C und 26AI in Quarz und Schwefel-Mineralen mittels Beschleunigermassenspektrometrie: Diplom Arbeit Technischen Universitat MOnchen, 56 p.

Niessen, F., Lister, G. and Giovanoll, F., 1992, Dust transport and paleoclimate during the Oldest Dryas in central Europe - implications from varves (lake Constance): Climate Dynamics, v. 8, p. 71-8'1.

Nishiizumi, K., Klein, J., Middleton R., and Craig, H., 1990, Cosmogenic 10Be, 26 AI and 3He in olivine from Maui lavas: Earth and Planetary Science letters, v. 98, p. 263-266.

Nishiizumi, K., Finkel, R. C., Klein, J., and Kohl, C. P., 1996, Cosmogenic production of 7Be and 10Be in water targets: Journal of Geophysical Research, v. 101, p. 22,225·22,232.

Nishiizumi, K., Lal, D., Klein, J., Middleton, R., and Arnold, J. R., 1986, Production of 10Be and 26 Al by cosmic rays in situ and implications for erosion rates: Nature, v. 319, p. 134-135.

Nishiizuml, K., Kohl, C. P., Arnold,). R., Klein, J., and Middleton, R., 1991, Cosmic ray produced 10Be and 26 Al in Antarctic rocks: exposure and erosion history: Earth and Planetary Science Letteri1, v. 104, p. 440-454.

Nishlizumi, K., Winterer, E. L., Kohl, C. P., Klein, J., Middleton, R., Lal, D., and Arnold, J. R., 1989, Cosmic ray production rates of 10Be and 26 Al in quartz from glacially polished rocks: Journal of Geophysical Research, v. 94, p. 17,907-17,915. 183

Nishiizumi, K., Kohl, C. P., Arnold, J. R., Dorn, R., Klein, J., Fink, D., Middleton, R., and Lal, D., 1993, Role of in situ cosmogenic nuclides 10Be and 26AI in the study of divers geomorphic processes: Earth Surface Processes and Landforms, v. 18, p. 407-425.

Nishiizumi, K., Arnold, J. R., Klein,)., Fink, D., Middleton, R., Kubik, P. W., Sharma, P., Elmore, D., and Reedy, R. C, 1991, Exposure histories of lunar meteorites: ALHA 81005, MAC88104, MAC88105, and yamato 791197: Geochimica et Cosmochimica Acta, v. 55, p. 3149-3155.

Nussbaum, F., 1910, Das Endmoranengebiet des Rhonegletschers von Wangen a. A.: Mitteilungen der Naturforschenden Gesellschaft Bern, v. 1761, p. 141-168.

Nussbaum, F., 1951, Kenntnis,; der Eiszeitbildungen der Umgebung von Solothurn: Mitteilungen der Naturforschenden Gesellschaft Solothurn, v. 16, p. 1-44.

Ochs, M., and Ivy-Ochs, S., 1996, The chemical behavior of Be, Al, Fe, Ca, and Mg during AMS target preparation from terrestrial silicates modeled with chemical speciation calculations: Nuclear Instruments and Methods (in press).

Ohno M. and Hamano, Y., 1992, Geomagnetic poles over the the past 10,000 years: Geophysical Re5€arch Letters, v. 19, p. 1715-1718.

Ohno M, and Hamano, Y., 1993, Global analysis of the geomagnetic field: Time variation of the dipole moment and the geomagnetic pole in the holocene: Journal of Geomagnetism and Geoelectricity, v. L,S, p. 1455-1466.

Olsson, I., 1986, Radiometric methods, in Berglund, B., ed., Handbook of Holocene Paleoecology and Paleohydrology: John Wiley & Sons, Chichester, USA, p. 273-312,

Patzelt, G., 1972, Die spatglazialen Stadien und postglazialen Schwankungen von Ostalpengletschern: Berichte der Deutschen Botanischen Gesselschaft, v. 85, p. 47-57.

Patzelt, G., and Bortensch!ager, S., 1973, Die postglazialen Gletscher- und Klimaschwankungen in der Venedigergruppe (Hohe Tauern, Ostalpen): Zeitschrift fur Geomorphologie Neue Folge, Supplement, v. 16, p. 25 - 72.

Pavich, M. J., Brown, L, Harden, J., Klein /., and Middleton, R., 1986, 10Be distribution in soils from Merced River terraces, California: Geochimica et Cosmochimica Acta, v. 50, p. 1727-1735. Peteet, D., 1995, Global Younger Dryas?: Quaternary International, v, 28, p. 93-104,

Peteet, D., Alley, R. B., Bond, G., Chappelaz, )., Clapperton, C., Del Genio, A., and Keigwin, l., 1993, Global Yc.unger Dryas?: EOS, Transactions, American Geophysical Union, v. 74, p. 586-588.

Penck, A., and Bruckner, E., 1901/09, Die Alpen im Eiszeitalter, v. 1-3: Leipzig, Germany, Tauchnitz.

Phillips, F. M., Zreda, M. G., Elmore, D., Sharma, P., 1996a, A reevaluation of cosmogenic 36CI production rates in terrestrial rocks: Geophysical Re5€arch Letters, v. 23, p. 949-952.

Phillips, F. M., Leavy, B. D., Jannik, N, 0., Elmore, D., and Kubik, P., 1986, The accumulation of cosmogenic chlorine-36 in rocks: A method for surface exposure dating: Science, v. 231, p. 41-43. 184

Phillips, F. M., Zreda, M. G., Smith, M. G., Kubik, P., Sharma, P., Elmore, D., 1990, A cosmogenlc chlorine-36 chronology for glacial deposits at Bloody Canyon, eastern Sierra Nevada, California: Science, v. 248, p. 1 529-1532.

Phillips, F. M., Zreda, M. G., Bemon, L. V., Plummer, M.A., Elmore, D., and Sharma, P., 1996b, Chronology for fluct~ations in late Pleistocene Sierra Nevada glaciers and lakes: Science, v. 274, p. 749-751.

Poreda, R. J., and Cerling, T. E., 1992, Cosmogenic neon in recent lavas from the western United States: Geophysical Research Letters, v. 19, p. 1863-1866.

Prael R. E., and Lichtenstein, H., 1989, User guide to LCS, The LAHET CodeSystem: Los Alamos National Laboratory Report, 1.A-UR-89-3014, 76 p.

Prentice, M. L., and Matthews, R. K., 1988, Cenozoic ice-volume history: development of a composite oxygen isotope record: Geology, v. 16, p. 963-966.

Prentice, M. L., Bockheim, J. G., Wilson, S. C., Burckle, L. H., Hodell, D. A., SchlUchter, C., and Kellog, D. E., 1993, Late Neogene Antarctic glacial history: Evidence from central Wright Valley, in Kennett, J. P. and Warnke, D. A., eds., The Antarctic Paleoenvironment: A Perspective on Global Change, 2: Antarctic Research Series, v. 60, p. 207-250.

Punchakunnel, P., 1982, Pollendiagramme aus dem Oberengadin: Physische Geographie, v. 6, p. 20-30.

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.

Raisbeck, G. M., Yiou, F., Bourles, D., Lorius, C., Jouzel, J., and Barkov, N. I., 1987, Evidence for two intervals of enhanced 10Be deposition in Antarctic ice during the last glacial period: Nature, v .326, p. 273-276.

Reedy, R. C., 1987, Nuclide production by primary cosmic-ray protons: Journal of Geophysical Research, v. 92, p. E697-E702.

Reedy, R. C., Nlshiizumi, K., and Arnold, J. R., 1990, Solar cosmic rays: fluxes and reaction cross sections: Lunar and Planetary conference, XX, Houston, p. 890-891.

Richmond, G. M., and Fullerton, D.S., 1986, Summation of Quaternary glaciations in the United States of America: Quaternary Science Reviews, v . 5, p. 183-196.

Robert, C., and Maillot, H., 1990, Evolution and significance of clay associations and inorganic geochemical data in the Wedoell Sea (Southern Ocean), Ocean Drilling Program Leg 113, in Barker, P.f., Kennett, J.P., et al., Proc. OOP, Sci. Results 113, Ocean Drilling Program, College Station, Texas, USA.

Robinson, C., Raisbeck, G.M., Ylou, f., Lehman, B., Laj, C., 1995, The realtlonship between '°Be and geomagnetic field strength records in central North Atlantic sediments during the last 80 ka: Earth and Planetary Science Letters, v. 136, p. 551-557.

Rott, H., Skvarca, P., and Nagler, T., 1996, Rapid collapse of Northern Larsen Ice Shelf, Antarctica: Science, v. 271, p, 788-792.

Sarafin, R., 1985, Anwendung radiochemischer Methoden und der Beschleuniger- Massenspectrometrie zur Bestimmung der tiefenabhangigen, langlebigen Spallatlons- Radionuklide 10Be, 26AI, 36Ci and 53Mn inStelnmeteorlten und Gedanken zum noch ungeklarten Ursprlng der Shergottite: Ph.D. dissertation Universitat zu Koln, 132 p. 185

Saunders, I., and Young, A., 1983, Rates of surface processes on slopes, slope retreat and denudation: Earth Surface P:·ocesses and Landforms, v. 8, p. 473-501.

Savin, S. M., Douglas, R. M., and Stehli, F. G., 1975, Tertiary marine paleotemperatures: Geological Society of America Bulletin, v. 86, p. 1499-1510.

Schaller-Donauer, A., 1925, Zekhensteine in Uri: Schweizer fOr Volkskunde, v. 26, p. 300-301.

Schaltegger, U., 1990, The Central Aar Granite: Highly differentiated calc-alkaline magmatism in the Aar Massif (Central Alps, Switzerland): European Journal of Mineralogy, v. 2, p. 245- 259.

Schluchter, C., 1986, The Quaternary glaciations of Switzerland, with special reference to the northern Alpine foreland: Quaternary Science Reviews, v. 5, p. 413-419.

Schluchter, C., 1987, Talgenese im Quartar - eine Standortbestimmung: Geographica Helvetica, v. 1987 Nr. 2, p. 109-115.

Sch!Uchter, C., 1988, The deglaciation of the Swiss Alps: A paleoclimatic event with chronological problems: Bulletin de I' Association fran<;:aise pour l'etude du Quarternaire, v. 1988-2/3, p. 141-145.

Schluchter, C., 1988-89, A ncn-classical summary of the Quaternary stratigraphy in the northern alpine foreland of Switzerland: Bulletin de la Societe neuchateloise de geographie, v. 32-33, p. 143-157.

SchlUchter, C., 1989, The most complete Quaternary record of the Swiss Alpine fore land: Paleogeography, Paleoclimatology, Paleoecology, v. 72, p. 141-146.

Schluchter, C., 1992, Terrestrial Quaternary stratigraphy: Quaternary Science Reviews, v. 11, p. 603-607.

Schluchter, C., and Wohlfar:h-Meyer, B., 1986, Till fades varieties of the Western Swiss Alpine Foreland, in van der Meer, J. J.M., ed., Tills and Gladotectonics: Balkema, Rotterdam, The Netherlands.

Schluchter, C., and Wohlfa1th, B., 1993, Geologie, in Die Schweiz vom Palaolithikum bis zum lruhen Mittelalter: Ver:ag Schweizerische Gesellschaft iOr Ur- und frUhgeschichte, Basel Switzerland, pp. 47-6S.

SchlOchter, C., and Rothlisberger, C., 1995, 100000 Jahre Gletschergeschichte, in Schweizerische Gletscherkommission, ed., Gletscher im standigen Wandel: Zurich, Switzerland, vdf Hochschulverlag AG an der ETH ZOrich, p. 47-63.

Schluchter, C., Maisch, M., Suter, J., fitze, P., Keller, W. A., Burga, C. A., and Wynistorf, E., 1987, Das Schieferkchlen-Profil von Gossau (Kanton Zurich) und seine stratigraphische Stellung innerhalb dEr letzten Eiszeit: Vierteljahrsschrift der Naturforschenden Gesellschaft in Zurich, v. 132/3, p. 135-174.

Schott, J., and Berner, RA, 1985, Dissolution mechanisms of pyroxenes and during weathering, in Drever, J. I., ed., The Chemistry of Weathering: Reidel, Dordrecht, Netherlands, p. 35-53.

Seidl, M., 199:i, Form and process in channel incision of bedrock, Ph.D. thesis, University of California, Berkeley, USA. 186

Senarclens de Graney, W., 1958, Zur Glazialgeologie des 6tztales und seiner Umgebung: Mitteilungen der Geologischen Gesellschaft Wien, v. 49, p. 257-313.

Seret, G., Dricot, E., and Wansard .. G., 1990, Evidence for an early glacial maximum in the French Vosges during the last glacial cycle: Nature, v. 346, p. 453-456.

Shackleton, N. J., 1987, Oxygen isotopes, ice volume and sea level: Quaternary Science Reviews, v. 6, p. 183-190.

Shackleton, N. )., and Opdike, N. D., 1973, Oxygen isotope and paleomagnetic stratigraphy of equatorial Pacific core V28-238: Oxygen isotope temperatures and ice volumes on a 105 and 106 year scale: Quaternary Research, v. 3, p. 39-55.

Shackleton, N. J., and Kennett, J. P., 1975, Paleoternperature history of the Cenozoic and the initiation of Antarctic glaciation: oxygen and carbon isotope analyses in DSDP sites 227, 279, and 281, in Kennett, J.P., Houtz, R.L, et al., lnit Repts. DSDP, v. 29., US Government Printing Office, Washington DC, USA, p. 743-756.

Sharma, P., and Middleton, R., 19H9, Radiogenic production of 10 Be and 26AI in uranium and thorium ores: Implications for studying terrestrial samples containing low levels of 10Be and 26AI: Geochimica et Cosmochimica Acta, v. 53, p. 709-716.

Shepard, M. K., Arvidson, R. E., Caffee, M., Finkel R., and Harris, L, 1995, Cosmogenic exposure ages of basalt flows: Lunar Crater voicanic field, Nevada: Geology, v. 23, p. 21- 24.

Smith R. M., and Martell, A. E., 1976, Critical Stability Constants, v. 4: Inorganic Complexes: Plenum Press, New York.

Sowers, T., Bender, M., Labeyrie, l. D., Jouzel, J., Raynaud, D., Martinson, D., and Korotkevich, Y. S., 1993, 135000 year Vostok-SPECMAP common temporal framework: Paleoceanography, v .8, p. 737-766.

Spate, A. P., Burgess, J. S., and Shevlin, J., 1995, Rates of rock surface lowering, Princess Elizabeth Land, Eastern Antarctica: Earth Surface Processes and Landforms, v. 20, p. 567- 573.

Stalder, H. A., 1964, Petrographisc'ie und mineralogische Untersuchungen im Grimselgebiet (Mittleres Aarmassiv): Schweizerische mineralogische und petrographische Mitteilungen, v. 44/1, p. 187-398.

Staub, W., 1950, Die drei Hauptstadien des Rhonegletschers im Schweizerischen Mittelland zur letzten Eiszeit: Berichte RObel, v. 1949, p. 110-117.

Staudacher, T., and Allegre, C. J., 1991, Cosmogenic neon in ultramafic nodules from Asia and in quartzite from Antarctica: Earth and Planetary Science Letters, v. 106, p. 87-102.

Staudacher, T., and Allegre, C. )., 1993, Ages of the second caldera of Piton de la Fournaise (Reunion) determined by cosmic ray produced 3He and 21 Ne: Earth and Planetary Science Letters, v. 119, p. 395-404.

Sternberg, R. S., 1992, Radiocarbon fluctuations and the geomagnetic field, in Taylor, R.E. et al., eds., Radiocarbon After Fcur Decades: Springer, New York, USA, p. 93-116.

Stroeven, A. P., Prentice, M. L., and Borns Jr., H. W., 1992, Sirius till at Mt. Fleming, Antarctica: Implications for early Pliocem• Antarctic glacial history: EOS, Transactions, American Geophysical Union, v. 73, p. 169. 187

Stroeven, A. P., Prentice, M. L., and Borns, Jr., H. W., 1992, Mt. Fleming upper valley drift: Evidence for Neogene glacial history of Antarctica: Antarctic Journal of the United States, v. 27, p. 51-53.

Stroeven, A. P., and Prentice, M. L, 1995, Recycled marine diatoms in Tertiary tills on Mt. Fleming, South Victoria Land, Antarctica: Complications for the hypothesized glacial conveyor: VII International Symposium on Antarctic Earth Sciences, Siena (abstracts), p. 364.

Sugden, D. E., 1992, Antarctic ice sheets at risk?: Nature, v. 359, p. 775-776.

Sugden, D. E., and John, B. S., 'i 976, Glaciers and Landscape: A Geomorphological Approach: Arnold, London.

Sugden, D. E., Denton, G. H., and Marchant, D. R., 1995a, Landscape evolution ol the Dry Valleys, Transantarctic Mountains: Tectonic implications: Journal of Geophysical Research, v. I 00, p. 9949-9967.

Sugden, D. E., Marchant, D. R., Potter, Jr., N., Sauchez, R. A., Denton, G. H., Swisher Ill, C. C., and Tison, J.-L., 1995b, Preservation ol Miocene glacier ice in East Antarctica: Nature, v. 376, p. 412-414.

Summerfield, M., 1991, Global Geomorphology: John Wiley and Sons, New York, USA, p. 371-402.

Suter, J., 1981, Gletschergeschichte des Oberengadins: Untersuchung von Gletscherschwankungen in der Err-Julier-Gruppe: Physische Geographie, v. 2, 170 p.

Suter, J., 1982, Zur klimatischen Interpretation der Gletschergeschichte, in Gamper, M., ed., Physische Geographle: Beltrage zur Quartarforschung in der Schweiz: Geographisches lnstitut der Universitat, Zurich, Switzerland, p. 37-48.

Swanson, T. W., Zreda, M., Phillips, F. M., Caffee, M., Finkel, R., and Southon, J., 1996, Determination of 36CI production rates from the deglaciation history ol Whidbey and Fidalgo Islands, Washington: Quaternary Research (In press).

Swanson, T. W., Caffee, M., Finkel, R., Harris, L., Southon, J., Zreda, M., and Phillips, F. M., 1993, Establishment of new production parameters for chlorine-36 dating based on the deglaciation history of Whidbey Island, Washington: GSA Abstracts with Programs, Annual Meeting, Boston, v. 25, p. A-461.

Synal, H.-A., Beer, I., Bonani, G., Lukasczyk, C., and Suter, M., 1994, 36CI measurements at the Zurich AMS facility: Nuclear Instruments and Methods, v. B29, p. 79-84.

Taylor, K. S., and Faure, G., 1983, Provenance dates of feldspar in glacial deposits, southern Victoria Land, Antarc':ica, in Oliver, R. L., et al., eds., Antarctic Earth Science: Australian Academy of Science, Canberra, Australia, p. 453-456.

Thiel, E. C., 1962, The amount of ice on the planet earth: Antarctic Research, Geophysical Monograph Series, v. 7, p. 172-175.

Thouveny, N., and Creer, K. M., 1992, Geomagnetic excursions in the past 60,000 years, ephemeral secular variation figures: Geology, v. 20, p. 399-402. 188

Thouveny, N., Creer, K. M. and Williamson, 0., 1993, Variations of the geomagnetic moment in Late Pleistocene times and their impact on cosmogenic isotope production: Global Planetary Change 7, p. 157-172.

Tingey, R. J., 1991, ed., The Geology of Antarctica: Oxford University Press, Oxford, UK.

Tric, E., Valet, J.-P., Tucholka, P., l'aterne, M., Labeyrie, L., Guichard, F., Tauxe, L., and Fontugne, M., 1992, Paleoint•ensity of the geomagnetic field during the last 80,000 years: Journal of Geophysical Research, v. 97, p. 9337-9351.

Truswell, E. M., 1986, Palynology, in Barrett, P. J., ed., Antarctic Cenozoic History from the MSSTS-1 Orillhole, McMurdc Sound: Wellington, New Zealand, Department of Scientific and Industrial Research Bullel:in, v. 237, p. 131-134.

Ugolini, F. C., 1986, Processes and rates of weathering in cold and polar desert environments, in Colman, S. M., and Dethie1', D. P., eds., Rates of Chemical Weathering of Rocks and Minerals: Academic Press, London, p. 193-235.

Ugolini, F. C., and Anderson, 0. M., 1973, Weathering and mineral synthesis in Antarctic soils: Soil Science, v.115, p. 461-470.

Valette-Silver, J. N., Tera, F., Klein, J., and Middleton, R., 1987, Beryllium 10 in hydrothermal vent deposits from the East Pacific Ridges: Role of sediments in hydrothermal processes: journal of Geophysycal Research, v. 92, p. 11,364-11,372. van den Bogaard, P., and Schmink•e, H.-U., 1985, laacher See Tephra: A widespread isochronous late Quaternary tephra layer in central and northern Europe: Geological Society of America Bulletin, v. 96, p. 1554-1571. van der Wateren, F. M., and Verbers, A. L. L. M., 1992, Cenozoic glacial geology and mountain uplift in northern Victoria Land, in Kaminuma, K., and Yoshida, Y., eds., Recent Progress in Antarctic Earth Science: Terra, Tokyo, Japan, p. 707-714.

Wardlaw, B. R., and Quinn, T. M., 1991, The record of Pliocene sea-level change at Enewetak Atoll: Quaternary Science Re1,iews, v. 10, p. 247-258.

Webb, P.-N., and Harwood, 0. M., 1987, Late Neogene terrestrial flora of Antarctica: Its significance in interpreting late Cenozoic glacial history: Antarctic Journal of the United States, v. 22, p. 7-11.

Webb, P.-N., and Harwood, D. M., 1991, Late Cenozoic glacial history of the Ross Embayment, Antarctica: Quaternary Science Reviews, v. 10, p. 21 5-223.

Webb, P.-N., Harwood, 0. M., Mcl

Welten, M., 1947, Pollenprofil Burgaschisee: Berichte Rubel, v. 1946, p. 101-111.

Welten. M., 1982, Pollenanalytische Untersuchungen im jlingeren Quartar des nordlichen Alpenvorlandes der Schweiz: lleitrage zur Geologischen Karte der Schweiz NF 156, Stampfii, Bern, Switzerland.

Wellen. M., 1988, Neue pollenanalytische Ergebnisse Ober das jungere Quartar des nordlichen Alpenvorlandes der Schweiz (Mittel- und Jungpleistozan): Beitrage zur Geologischen Ka rte der Schweiz, Neue Folge, Lfg. 162. 189

Westall, j. C., Zachary, j. L., and Morel, F. M. M., 1976, MINEQL- a computer program for the calculation of chemical equilbrium compostion of aquatic systems: MIT Technical Note 18.

Whitehouse, I. E., McSaveney, M. )., Kneupfer, P. L. K., Chinn, T. ). H., 1986, Growth of weathering rinds on Torlesse Sandstone, Southern Alps, New Zealand, in Colman, S. M., et al., eds., Rates of Chemirnl Weathering of Rocks and Minerals, London, Academic Press, p. 419-435.

Wilch, T. L., Lux, D. R., Denton, G. H., and Mcintosh, W. C., 1993, Minimal Plio-Pleistocene uplift in the dry valleys sector of the Transantarctic Mountains: Geology, v. 21, p. 841- 844.

Winterhalter, R. U., and de Quervain, F., 1935, Vergletscherungskarte der Schweiz 1:100,000, Beikartchen mit ausfOhrlicher Legende zur Geotechnischen Karle der Schweiz 1 :200,000, Bl. 2: Schweizerische Geologische Kommission, Bern, Switzerland.

Wise, S.W., Gombos, A.M., and Muza, J.P., 1985, Cenozoic evolution of polar water masses, southwest Antlantic Ocean, in HsO, K.J., and Weissert, H. (eds), South Atlantic Paleoceanography, p. 205-241, Cambridge University Press, Cambridge, UK.

Wise, S.W., Breza, j. R., Harwood, D. M., and Wei, W., Paleogene glacial history of Antarctica, in MacKenzie,)., et al., eds., Controversies in Modern Geology: Academic Press, London, p. 133-171.

Wohlfarth, B., Gaillard, M.-J., Haberli, W., and Kelts, K., 1994, Environments and climate in southwestern Switzerland during the last termination 15-10 ka BP: Quaternary Science Reviews, v. 13, p 361-394.

Woodruff, F., Savin, S. M., and Douglas, R. G., 1981, Miocene stable isotope record: a detailed deep Pacific Ocean study and its paleoclimatic implications: Science, v. 212, p. 665-668.

Wright, J. D., and Miller, K. G., 1993, Southern ocean influences on late Eocene to Miocene deepwater circulation, in Kennett, J. P. and Warnke, D. A., eds., The Antarctic Paleoenvironment: A Perspe·:tive on Global Change, 2: Antarctic Research Series, v. 60, p. 1-26.

Yokoyama, Y. )., Reyss, )., and Guichard, F., 1977, Production of radionuclides at mountain altitudes: Earth and Planetary Science Letters, v. 36, p. 44-50.

Zbinden, H., 1987, 14C Messungen an Pflanzenresten aus Sediment: Diploma Thesis, Physikalisches lnstitut der Universitat Bern, Bern, Switzerland.

Zolitschka, B., 1996, Paleoklimat1sche Bedeutung laminierter Sedimente: Habilitationsschrift, Universitat Potsdam, Germany, 196 p.

Zreda, M. G., 1994, Development and calibration of the 36CI surface exposure dating method and its application to the chronology of Late Quaternary glaciations: Ph.D. dissertation, New Mexico lnstitue of Mining and Technology, Soccorro, USA, 318 p.

Zreda, M. G., and Phillips, f. M., 1994, Surface exposure dating by cosmogenic Chlorine-36 accumulation, in Beck, C., ed., Dating in Exposed and Surface Contexts: University of New Mexico Press, Albuquerque.

Zreda, M. G., and Phillips, F. M., 1995, Insights into alpine moraine development from cosmogenic 36CI buildup daling: Geomorphology, v. 14, p. 149-156. 190

Zreda, M. G., Phillips, F. M., and Elmore, D., 1994, Cosmogenic 36CI accumulation in unstable landforms 2. Simulations and measurements on eroding moraines: Water Resources Research, v. 30, p. 3127-3136.

Zreda, M. G., Phillips, F. M., Kuc·ik, P., Sharma, P., and Elmore, D., 1993, Cosmogenic 36CI dating of a young basaltic eruption complex, Lathrop Wells, Nevada: Geology, v. 21, p. 57-60.

Zreda, M. G., Phillips, F. M., Elmore, 0., Kubik, P., Sharma, P., and Dorn, R., 1991, Cosmogenic chlorine-36 production rates in terrestrial rocks: Earth and Planetary Science Letters, v. 105, p. 94-109.

Zumbi.ihl, H. J., and Holzhauser, H., 1988, Alpengletscher in der Kleinen Eiszeit: Die Alpen, v. 3. Quartal 1988, p. 129-322. 191


Sample Dissolution Mineral Dissolved Isotope Chapter discussed No. Step or WR (g) /outcome Antarctica 4 HCI cpx 0.0478 toBe B3 10 4 HF1 CPX 1.9248 Be B3 4 HF2 cpx 0.3697 10Be B3 Sa HF2 cox 0.6905 10Be B3 Sa HF3 cpx 0.2691 toBe B3 Sa HF4 cpx 0.4376 toBe B3 Sa HFS cox 0.4301 ioBe 83 Sa HF6 cpx 2.4826 10Be 83 6b HF2 QZ 0.8198 10Be, 26AI • 6b HF2 QZ 5.201 ioBe, 26AI . 6b HF3 QZ 1.1673 toBe, 26AI . 205 HF4 QZ 7.736 toBe, 26AI C2 10 26 207 HF4 I qz 3.6814 Be, AI C2 207 HFS qz 8.1696 108e, 26AI C2 207 HF6 QZ 10.1267 toBe, 26AI C2 208 HF6 qz 6.5761 ioBe, 26AI (2 208 - WR 68.03 36(1 • 209 HF4 qz 4.5595 toBe, 26AI C2 209 HFS QZ 6.6311 10Be, 26AI C2 209 HF6 QZ 9.2767 10Be, 26AI (2 217a HF6 qz 19.4132 10Be, 26AI C2 217b HFS QZ 1.3971 10Be, 26AI C2 217b HF6 QZ 2.9749 10Be, 26AI 82,(2 217b HF7 qz 2.9935 toBe, 26AI B2,C2 10 26 217b HF8 QZ 5.1831 Be, AI B2,C2 217b HF6 qz 11.6875 10Be, 26AI B2,C2 217b - WR 72.20 36CI . 10 26 213a HCI QZ 0.1 Be, AI B2 213a HF2 qz 4.7063 ioBe, 26AI B2,C3 213a HF4 QZ 11.1623 ioBe, 26AI B2,C3 10 26 214b HF4 QZ 13.8094 Be, AI C2 192

218 HF2

J8 HF6 QZ 33.6807 ioBe, 26AI 02 )10 HF6 qz 40.29 10Be, 26AI 02 JlO WR 52.09 36CI 02 J104 HF7 qz 51.3097 ioBe, 26AI 02 J12 HF7 qz 47.8596 10Be, 26AI 02 )12 - w~: 65.85 36CI 02 J15 - WR 43.57 10Be, 26AI 02 J18 HF7 qz 40.3793 ioBe, 26AI 02 J18 - WR 86.49 36CI # G1 HF7 qz 45.2858 ioBe, 26AI 03 G1 - WR 52.14 36CI # 10 26 G102 HF6 QZ 54.1858 Be, AI 03 G103 HF5 qz 29.7941 10Be, 26AI 03 10 26 HF5 qz 42.4545 Be, AI 03 ~ HF6 qz 48.2215 ioBe, 26AI 03 G110 I - WR 66.21 36CI # G111 HF5 QZ 44.4013 ioBe, 26AI 03 ML5 - WR 70.97 36CI 03 ER1 HF5 qz 45.4067 ioBe, 26AI 04 ER1 - WR 55.02 36CI # ER7 HF5 QZ 54.3639 ioBe, 26AI 04 193

ER8 HFS tZ 61.0543 10Be, 26AI 04 FSL8 WR 55.03 36(1 #

FSL 10 HF5 tZ 49.3320 10Be, 26AI # FSL11 HF5 tZ 31.2272 10Be, 26AI * New Zealand NZ2 HF5 z 61.7945 NZ3 HF5 58.2422 Kofels landslide, Austria K04 HF5 Kubik et al. 1996 Ko6 HF5 75. Kubik et al. 1996 Ko7 HF6 67. Kubik et al. 1996 Ko8 HF5 45. Kubik et al. 1996 Drill core, Bavaria, German 5492 46.8139 lOBe, 26AI Niederma er 1995• 5495 31.0897 ioBe, 26AI Niederma er 1995 • EWI 05 60.0417 ioBe, 26AI Niederma er 1995• 10 26 EWI 10 66.339 Be, AI N iederma er 199 5 • EW101 19.7597 10Be, 26AI Niederma er 1995• PSI tar et 25.1773 ioBe, 26AI Niederma er 1995• *data not presented here, to be discussed in a later publication #not measured successfully •These samples were prepared by Niedermayer under the author's direction at EAWAG. APPENDIX H2 TABLE OF THIN SECTION ANALYSES.

Rock type Avg. Mineralogy Comments Formation name grain size Antarctica I TM 205 quartzite fn QZ qz overgrowths

TM 213a quartz sandstone med qz clay, Fe-oxide cement New Mountain Sandstone 214b " (well- qz>>ksp weakly silicified " 218 " rounded) qz clay, Fe-oxide cement " 219a " " II weakly silicified " 219b " II qz>>ksp weakly silicified II 220 II " qz silicified surface " 323 II " " sil icified surface II 325 " II II weakly silicified " 326 " II " silicified surface " TM 207 granite med-csc qz>ksp>plag>bio fresh, 208 " " II myrmekitic qz-ksp boundaries 209 " " " TM 217a granite csc qz>ksp>plag>bio plag slightly altered to clays 217b " " " Mfe 309 quartz sandstone csc qz weakly silicified

MFI 320 sandstone- csc qz»ksp arkosic, silicified throughout conglomerate

MFI 322 sandstone- csc qz>>ksp arkosic, silicified throughout conglomerate Ti"v1 4 do!er!te fn-med cpx>plag>>opx poikilitic, cores of px patchy Ferrar Dolerite alteration to clavs

TM Sa dolerite fn-med cpx>plag> >opx poikilitic, cores of px patchy Ferrar Dolerite alteration to clays Swiss Alps I G1 granodioritic gneiss fn-med qz>plag>ksp>bio strongly foliated, Grimsel Granodiorite >mu>ep mosaic quartz

G102 quartz vein med QZ>>ep

G103 quartz vein med QZ>>ep

G106 quartz vein med QZ>>ep

G110 augen gneiss med qz>ksp>bio>ep> bands of qz intergrown with Grimsel Granodiorite mu>chl mu

G111 quartz vein csc qz>>eP J8 granodiorite csc qz>plag>ksp julier Granite ..--.-----·········----~~~··~··------

JlO granodiorite csc qz>plag>ksp plag altered Julier Granite

Jl 2 aolite fn-med QZ, ksp Julier Granite Jl 5 diorite med hb, plag, >chi hb and olag very altered Julier Granite J18 rite csc qz>plag>ksp plag altered Julier Granite J104 granodiorite csc qz>olag>ksp Julier Granite ERl hornblende granite med qz>ksp>plag>hb weakly foliated Arkesine >chi ER? aolite fn-med qz>ksp>plag Arkesine ER8 hornblende granite med-csc qz>ksp>plag>hb weakly foliated I Arkesine >chi MLS foliated granite med-csc qz>kso>Plag>hb

List of abbreviations: Mineral abbreviations: TM: Table Mountain qz:quartz hb: hornblende fn: fine-grained MFe: Mount Feather ksp: alkali feldspars bio: biotite med: medium-grained MFI: Mount Fleming plag: plagioclase (incl. albite) mu: muscovite csc: coarse-grained G: Grimsel Pass cpx: clinopyroxene ep: epidote J: Ju lier Pass opx: orthopyroxene chi: chlorite ER and ML: northern foreland erratics CURRICULUM VITAE

May 8, 1958 Born in Evanston, Illinois, USA

1961 - 1965 Barbereaux Montessori School, Evanston IL 1965 -1968 Kingsley Elementary School, Evanston IL 1968 1971 Haven Junior High School, Evanston IL 1971 -1975 Evanston Township High School, Evanston IL

1977 -1980 Cypress Junior College, Cypress, CA 1980 Associate of Science Degree 1980 - 1982 Geological Assistant, Earth Science Consulting and Technology, Inc. Costa Mesa, CA 1980 - 1983 California State University Long Beach, Long Beach, CA 1983 Bachelor of Science Degree, Geology

1983 - 1984 University of Oregon, Eugene, OR 1983 - 1984 Teaching Assistant, Department of Geology (crystallography, mineralogy)

1984 - 1989 Oregon State University, Corvallis, OR 1984 - 1986 Teaching Assbtant, Department of Geology (optical mineralogy, petrography) 1988 Master of Science Degree, Geology

1987 - 1989 Research Assi:;tant, Department of Environmental Engineering, OSU 1989 Master of Science Degree, Environmental Engineering

1989 - 1992 Research Scientist, Lamont Doherty Earth Observatory, Columbia University, and lnstitut fur Teichenphysik, ETH Zurich

1993 - 1996 Doctoral Studies, Geologisches lnstitut and lnstitut fur Teilchenphysik, ETH Zurich