Surface exposure dating - IT Services of ETH Zurich

Transcription

Surface exposure dating:
A geologist’s view with examples from both
hemispheres
Inauguraldissertation
der Philosophisch-naturwissenschaftlichen Fakultät
der Universität Bern
vorgelegt von
Silvio Tschudi
von Luzern (LU)
Leiter der Arbeit:
Prof. Dr. Christian Schlüchter
Universität Bern
Dr. Peter W. Kubik
Paul Scherrer Institut & ETH Zürich
Von der Philosophisch-naturwissenschaftlichen Fakultät angenommen.
Bern, den 9.11.2000
Der Dekan:
Prof. Dr. P. Bochsler
Preface
This Ph.D. thesis is part of an interdisciplinary project between the University of Berne
(Institute of Geology, Quaternary Geology), the ETH Zurich (Institute of Particle Physics and
Institute of Isotope Geology and Mineral Resources) and the Paul Scherrer Institut. It is an
interdisciplinary approach to dating problems in Quaternary geology by using the method of
surface exposure dating (SED) with radioactive (10Be and 26Al) and stable (21Ne) cosmogenic
nuclides. My contribution to this project was the investigation and application of the
cosmogenic radionuclides 10Be and 26Al, which are measured with accelerator mass
spectrometry (AMS). The work was performed under the guidance of Christian Schlüchter,
Susan Ivy-Ochs and Peter W. Kubik. It was hosted at the Institute of Particle Physics at ETH
Zurich within the group of Martin Suter.
The thesis is organized into six chapters. The introducing Chapter 1 gives the motivation and
presents the main goals and questions of this dissertation. In Chapter 2, the method of
surface exposure dating (SED) is briefly discussed, progress and problems are reported and
the path of a SED sample from the field to the laboratory and finally to the AMS
measurement is followed. The data chapters, Chapter 3 and Chapter 4, are mainly
publications, already in press, submitted or close to be submitted. Chapter 3 deals with the
dating of young glacigenic deposits in the Northern Hemisphere. The reader will find a
publication about a Younger Dryas glacial formation in Finland, a pilot study on bedrock
surfaces on Wrangel Island, Russia, and two studies on moraine systems in eastern and
central Tibet. In Chapter 4, results from the Southern Hemisphere are shown. It is the “old
chapter”, as it discusses SED analyses with exposure ages up to a few million years from
Antarctic samples of Beacon Valley and the Allan Hills nunatak. If appropriate or necessary,
the publications are followed by a short appendix with further comments and/or additional
data. Chapter 5 provides the overall conclusions and the outlook. The „global“ reference list,
Chapter 6 and the Appendices A & B end this thesis. According to the given rules, the
layout of the thesis and parts of the text of chapter 3.4 are slightly modified compared to the
original thesis, presented on November 9th 2000 at the University of Berne.
Contents
ABSTRACT ..................................................................................................................................... 1
ZUSAMMENFASSUNG.................................................................................................................. 3
CHAPTER 1 INTRODUCTION ..................................................................................................... 7
1.1
1.2
1.3
1.4
1.5
General.............................................................................................................................. 7
Younger Dryas Salpausselkä I Formation, Finland............................................................. 8
Wrangel Island, Far Eastern Russia ................................................................................... 8
Eastern and central Tibet ................................................................................................... 9
Antarctica (Beacon Valley and Allan Hills nunatak) ........................................................ 10
CHAPTER 2 METHODS .............................................................................................................. 11
2.1
2.2
2.3
2.4
Basic principles of SED................................................................................................... 11
Application of SED in Quaternary geology...................................................................... 21
SED Sampling: Preparation and fieldwork....................................................................... 23
SED sample processing: Laboratories .............................................................................. 29
CHAPTER 3 NORTHERN HEMISPHERE................................................................................. 35
3.1
3.2
3.3
3.4
10
Be Dating of Younger Dryas Salpausselkä I Formation in Finland................................. 35
Constraints for the latest glacial advance on Wrangel Island, Arctic Ocean, from rock
surface exposure dating................................................................................................... 48
Preliminary 10Be and 26Al dating of moraines in the Kanding area, eastern Tibet .............. 57
The limited influence of glaciations in Tibet on global climate over the past 180ky.......... 63
CHAPTER 4 SOUTHERN HEMISPHERE ................................................................................. 71
4.1
4.2
Last major advance of Taylor Glacier into central Beacon Valley at least 4 Ma ago: New
indications from surface exposure dating on clasts from Granite drift .............................. 71
Cosmogenic nuclides in the Allan Hills nunatak, Antarctica: An approach to reconstruct the
local glacial history......................................................................................................... 84
CHAPTER 5 CONCLUSION AND OUTLOOK........................................................................ 103
CHAPTER 6 REFERENCES ...................................................................................................... 109
APPENDIX A USED ABBREVIATIONS................................................................................... 117
APPENDIX B COMPILATION OF ALL 10BE AND 26AL DATA ............................................ 118
DANKSAGUNG........................................................................................................................... 123
CURRICULUM VITAE .............................................................................................................. 125
ABSTRACT / ZUSAMMENFASSUNG
1
Abstract
The chronological classification of glaciations within a wide range of geographical and
geological regions, based on surface exposure dating is the core of this work. It is shown that
surface exposure dating (SED) is a suitable tool for solving Quaternary dating problems.
However, it is also shown that the key to successful SED studies lies in careful fieldwork and
the study of the local and regional geological setting. This is most important because the
relation of the sampled surface to the geological event (here: glaciations) has to be defined
and clearly understood.
Northern Hemisphere
The Finnish Younger Dryas glacial formation Salpausselkä I (see e.g. Rainio et al., 1995) was
dated using four boulders west of Lahti at the apex of the bend of Salpausselkä I. The
minimum 10Be exposure ages range between 11’050 ± 910 years and 11’930 ± 950 years with
an error-weighted mean of 11’420 ± 470 years. This corresponds to the Younger Dryas
interval in the Greenland ice cores GRIP and GISP2 (Alley et al., 1993; Johnsen et al., 1992).
For the first time, the assumed Younger Dryas age of Salpausselkä I is directly confirmed.
Both, coverage by vegetation or snow and uplift due to isostatic rebound have been taken into
account in the calculations. The study shows that SED can be used, where very low
cosmogenic nuclide concentrations are expected.
For Wrangel Island, Eastern Siberia, Russia, the presence of ice during the last glacial
maximum (LGM) has been discussed. On the one hand, Grosswald (1997) suggests an intense
glaciation with a thickness of about 1’000m in the Arctic region, covering also Wrangel
Island. On the other hand, Vartanyan et al. (1995) and Sulerzhitsky and Romanenko (1999)
report radiocarbon ages on mammoth teeth, bones and tusks from Wrangel Island, ranging
from 4’000 up to 30’000 years BP with some data between 20’000 and 22’400 years BP. SED
results of 10Be in bedrock samples argue against a glaciation of the island during LGM.
Assuming that an ice sheet on the island excavates previously unexposed bedrock by erosion
(Cuffey et al., 2000), the sampled surfaces indicate no significant glaciation since 64’600 ±
6’400 years before present. This indeed supports that the island was ice-free during LGM.
Glacial deposits at the eastern margin of Tibet (Kanding and Litang area) and the central part
of Tibet (Tanggula area) were sampled to establish an absolute chronology of glacial events
on the plateau. The presence of a huge ice sheet until the end of the last glacial cycle, as
suggested by Kuhle (1998), is ruled out by our two independent studies, which clearly
indicate the presence of valley glaciers during marine oxygen isotope stage 2 (MIS-2) and the
2
last glacial-interglacial transition (LGIT). The 10Be and 26Al data from Kanding vary between
11’440 ± 900 and 13’470 ± 1’030 years, whereas Litang data cluster between 14’600 ± 1’100
and 16’800 ± 1’300 years (10Be and 26Al). MIS-2 glacial deposits indicate that the eastern part
of Tibet shows synchronous behavior with other glacial advances in the Northern
Hemisphere. The discussion about synchronous (e.g. Lehmkuhl, 1998) or asynchronous
(Phillips et al., 2000) behavior with respect to Tibetan glaciations is indeed important for the
understanding of Tibet’s role within the global climate system. From the central part of Tibet,
the Tanggula area, we report minimum exposure ages confirmed by 10Be, 26Al and 21Ne of up
to 180ka for moraines supposed to be deposited by the penultimate glaciation (Lehmkuhl,
1998). Our data indicate that the Tanggula area was ice-free since at least 180ka ago. Again,
the presence of an ice-sheet, as suggested by Kuhle (1998), could not be supported.
Southern Hemisphere
The data from the Southern Hemisphere come from the Transantarctic Dry Valleys, where the
landscape is preserved for millions of years (see e.g. Denton et al., 1993; Schäfer et al., 1999;
Sugden et al., 1999; Summerfield et al., 1999). Erosion is low, and fingerprints of past glacial
events (e.g. deposits of glacial advances due to an expanding East Antarctic Ice Sheet, EAIS)
are still observable. The determination of the age of these events, and subsequently the timing
of past behavior of the EAIS, was the goal of two SED studies in Beacon Valley and the
Allan Hills nunatak. In Beacon Valley, clasts of Granite drift, a glacial deposit of a significant
glacial advance of Taylor Glacier into central Beacon Valley, were sampled. The minimum
age for this drift is 3.99 ± 0.15Ma. This age yields also a minimum age for the morphostratigraphically older Sirius Group at nearby Mt. Feather (a crucial glacial formation in the
discussion about the stability of EAIS, e.g. Miller and Mabin, 1998) and the remnant body of
ice buried under the valley floor (Sugden et al., 1995b). Consistent 10Be, 26Al and 21Ne data
indicate that the sampled Granite drift represents the last significant advance of Taylor
Glacier into central Beacon Valley. This implies that Taylor Dome, which is the feeding
source of Taylor Glacier, did not significantly thicken again after early Pliocene time.
Moreover, the data support previous dating of in situ ashes that give a minimum age for the
underlying body of ice of about 8Ma (Sugden et al., 1995b), which had been questioned by
Hindmarsh et al. (1998).
The Allan Hills nunatak bears the northernmost outcrop of Sirius Group deposits (Borns
unpublished, referenced in Denton et al., 1991). There, direct dating by SED yields a
minimum exposure age of 2.31 ± 0.31Ma for this glacial deposit. This age is in accord with
recent SED studies located in the Inner Dry Valleys (Brook et al., 1995; Bruno et al., 1997;
Ivy-Ochs, 1996; Ivy-Ochs et al., 1997; Schäfer et al., 1999; Tschudi et al., submitted).
Consisting 10Be, 26Al and 21Ne ages indicate a simple exposure history without significant
coverage since deposition. Younger glacial deposits trace a smaller expansion of EAIS into
3
the Allan Hills nunatak. SED indicates prior exposure of all samples, due to possible shallow
burial within uncovered bedrock during long periods of time before excavation by erosive
glacial activity. From model calculations on a multi-sampled erratic boulder and from the use
of the modified two-nuclide plot of Heisinger (1998), one can conclude that the nunatak was
ice-free since about 100ka before the present. The period of coverage by the advanced ice is
too short to be recognizable with SED (i.e. about 100ka or less). The young advance of EAIS
into the Allan Hills nunatak may correlate with the Taylor II advance into Arena Valley, Inner
Dry Valleys (Marchant et al., 1993a and references therein). This advance of Taylor Glacier
yields a minimum 10Be exposure age of about 120ka (Brook et al., 1993a; reevaluated with
production rates of Kubik et al., 1998).
Zusammenfassung
Die chronologische Einteilung von Vereisungen innerhalb von diversen geografischen und
geologischen Regionen, basierend auf der Methode der Oberflächenaltersbestimmung, bildet
den Kern der vorliegenden Arbeit. Es wird gezeigt, dass Oberflächenaltersbestimmung ein
geeignetes Arbeitsinstrument zur Lösung von Datierungsproblemen des Quartärs darstellt,
wenn den Datierungsstudien sorgfältige Feldarbeit zu Grunde liegt. Dies impliziert auch ein
Verständnis von lokalen und regionalen Einflüssen auf das zu datierende geologische Ereignis
(hier: eine vergangene Vergletscherung). Insbesondere muss der Zusammenhang zwischen
der beprobten Oberfläche und diesem Ereignis klar definiert und verstanden sein.
Nördliche Hemisphäre
Die Finnische Jüngere Dryas Glazialformation Salpausselkä I (siehe zB. Rainio et al., 1995)
wurde mit vier grossen Gesteinsblöcken westlich von Lahti am Scharnier von Salpausselkä I
datiert. Das minimale 10Be Expositionsalter variiert zwischen 11’050 ± 910 und 11’930 ± 950
Jahren mit einem fehlergewichtetem Mittel von 11’420 ± 470 Jahren, welches mit dem
Intervall der Jüngern Dryas in den grönländischen Eisbohrkernen GRIP und GISP2
übereinstimmt (Alley et al., 1993; Johnsen et al., 1992). Das vermutete Jüngere-Dryas-Alter
der Salpausselkä I konnte erstmals direkt bestätigt werden. Bedeckung durch Vegetation und
Schnee wurden ebenso in den Berechnungen berücksichtigt, wie Hebung auf Grund von
isostatischem Ausgleich. Die Studie zeigt, dass die Methode der Oberflächendatierung auch
an Proben mit sehr niedrigen Konzentrationen von kosmogenem Nukliden angewendet
werden kann.
Eine Eisbedeckung von Wrangel Island, Ostsibirien, Russland, während des letzteiszeitlichen
Maximums (LGM) wird intensiv diskutiert. Auf der einen Seite schlägt Grosswald (1997)
4
eine grossräumige Vereisung der arktischen Region mit Eisdicken von etwa 1’000 Metern
vor, welche auch Wrangel Island bedecken würde. Auf der anderen Seite haben Vartanyan et
al. (1995) und Sulerzhitsky and Romanenko (1999) auf Wrangel Island Zähne, Knochen und
Stosszähne von Mammuts mit der Radiocarbonmethode datiert. Ihre Altersbestimmungen
variieren zwischen 4’000 und 30’000 Jahren BP, mit einigen Daten zwischen 20’000 und
22’400 Jahren BP. Die Oberflächendatierung an anstehendem Gestein mit Hilfe von 10Be
bringt neue Argumente gegen eine Vereisung während des LGM. Unter der Annahme, dass
eine Vereisung der Insel vorher nicht exponiertes Gestein freilegen würde (Cuffey et al.,
2000), weisen die beprobten Oberflächen darauf hin, dass seit 64’600 ± 6’400 Jahren vor
Heute keine wesentliche Vereisung vorhanden war. Das wiederum unterstützt, dass die Insel
während des LGM eisfrei war.
Glazialablagerungen am östlichen Rand von Tibet (Gebiete in Kanding und Litang) und im
zentralen Bereich von Tibet (Tanggula) wurden für die Bestimmung einer absoluten
Chronologie von Glazialereignissen auf dem tibetanischen Hochplateau beprobt. Bisherige
Annahmen von Kuhle (1998), der die Präsenz eines grossflächigen Eisschildes bis zum Ende
des letzteiszeitlichen Zyklus vorschlägt, sind durch zwei unabhängige Studien, welche die
Existenz von Talgletschern während der marinen Sauerstoff-Isotopenstufe 2 (MIS-2) und der
Übergangszeit vom letzten Glazial zum Interglazial (LGIT) belegen, verworfen. Die 10Be und
26
Al Daten von Kanding variieren zwischen 11’440 ± 900 und 13’470 ± 1’030 Jahren,
während sich die Daten von Litang zwischen 14’600 ± 1’100 und 16’800 ± 1’300 Jahren
bewegen (ebenfalls 10Be und 26Al). Glazialablagerungen zur Zeit des MIS-2 zeigen, dass sich
der östliche Rand des Plateaus synchron zu anderen Gletschervorstössen auf der nördlichen
Hemisphäre verhalten hat. Die Diskussion über die Synchronität (siehe zB. Lehmkuhl, 1998),
bzw. die Asynchronität (Phillips et al., 2000) der tibetanischen Vergletscherungen sind
wichtig für das Verständnis der Rolle, welche das Plateau im globalen Klimarahmen inne hat.
Von Zentraltibet (Tanggula) sind minimale Expositionsalter an penultimativen Moränen
(Lehmkuhl, 1988) von bis zu 180’000 Jahren mit 10Be, 26Al, und 21Ne bestimmt worden. Die
übereinstimmenden Alter der drei gemessenen Nuklide deuten auf Eisfreiheit seit mindestens
180’000 Jahren hin. Die Präsenz eines mächtigen Eisschildes (Kuhle, 1998) kann nicht
bestätigt werden.
Südliche Hemisphäre
Die Daten der südlichen Hemisphäre stammen von den Transantarktischen Trockentälern, wo
jahrmillionenalte Landschaften vorhanden sind (siehe z.B. Denton et al., 1993; Schäfer et al.,
1999; Sugden et al., 1999; Summerfield et al., 1999). Tiefe Erosionsraten lassen Zeugen von
vergangenen Glazialereignissen (zB. Ablagerungen von Vorstössen des ostantarktischen
Eisschildes, EAIS) über lange Zeiträume bestehen. Die Bestimmung der Alter dieser
Ereignisse und damit die zeitliche Festlegung vergangener Oszillationen des EAIS waren die
5
Ziele von zwei Studien zur Bestimmung von Oberflächenaltern im Beacon Valley und auf
dem Allan Hills Nunatak. Im Beacon Valley wurden Findlinge der „Granite drift“,
Ablagerungen eines signifikanten Vorstosses des Taylor Gletschers in das zentrale Beacon
Valley, beprobt. Das Minimalalter dieser Ablagerung liegt bei 3.99 ± 0.15Ma. Dieses Alter
liefert auch ein Minimalalter für die morphostratigrafisch ältere Siriusformation auf dem
nahen Mt. Feather (der im Zusammenhang mit der Diskussion um die Stabilität des EAIS eine
Schlüsselrolle zukommt, zB. Miller and Mabin, 1998) und für das Toteis, welches unter dem
heutigen Talboden des Beacon Valley liegt (Sugden et al., 1995b). Übereinstimmende 10Be,
26
Al und 21Ne Daten deuten darauf hin, dass es sich bei der beprobten „Granite drift“ um die
Ablagerung des letzten signifikanten Vorstosses des Taylor Gletschers in das zentrale Beacon
Valley handelt. Diese Tatsache impliziert, dass die Mächtigkeit des Taylor Dome (Ursprung
des Taylor Gletschers) in der Zeit nach dem frühen Pliozän nie mehr signifikant dicker war.
Zudem werden Untersuchungen an in situ Aschen, welche das begrabene Toteis auf ein Alter
von etwa 8Ma datieren (Sugden et al., 1995b), und zuvor von Hindmarsh et al. (1998) in
Frage gestellt waren, unterstützt.
Am Allan Hills Nunatak ist der nördlichst gelegene Aufschluss der Siriusformation
beschrieben (Borns, unpubliziert, Referenz in Denton et al., 1991). Direkte Datierung mit
10
Be, 26Al und 21Ne liefert ein minimales Expositionssalter von 2.31 ± 0.31Ma für diese
Glazialformation. Dieses Alter ist konkordant zu anderen Oberflächenalterbestimmungen im
Bereich der Inneren Trockentäler (Brook et al., 1995; Bruno et al., 1997; Ivy-Ochs, 1996; IvyOchs et al., 1997; Schäfer et al., 1999; Tschudi et al., submitted). Die übereinstimmenden
Alter (10Be, 26Al und 21Ne) deuten auf eine einfache Bestrahlungsgeschichte ohne signifikante
Bedeckung nach der Ablagerung hin. Jüngere Glazialablagerungen zeugen von einem
kleineren Vorstoss des EAIS in den Nunatak. Die Oberflächendatierungen deuten auf eine
Vorbestrahlung aller Proben hin. Dies liegt an der Tatsache, dass die beprobten
Gesteinsoberflächen vor der Erosion durch den vorstossenden Gletscher möglicherweise im
anstehenden und unbedeckten Grundgestein während langer Zeit untief begraben waren.
Modelrechnungen an einem mehrfach beprobten erratischen Block und die Benutzung des
modifizierten Zwei-Nuklid-Plots von Heisinger (1998) lassen darauf schliessen, dass der
Nunatak seit rund 100ka eisfrei ist. Die Periode der Eisbedeckung vor dem Rückzug ist zu
kurz, um mit der Methode der Oberflächendatierung erkannt zu werden (d.h. ungefähr 100ka
oder weniger). Der junge Vorstoss des EAIS in den Allan Hills Nunatak kann eventuell mit
dem Taylor II Vorstoss in das Arena Valley der inneren Trockentäler (Marchant et al., 1993a
und Referenzen darin) korreliert werden. Dieser Vorstoss wurde mit Hilfe von 10Be auf ein
Alter von rund 120ka datiert (Brook et al., 1993a; reevaluiert mit Produktionsraten von Kubik
et al., 1998).
CHAPTER 1: INTRODUCTION
7
Chapter 1
Introduction
1.1 General
The discussion about global climate change and therefore the future path of the Earth’s
climate is among the most important topics we have all to face (see for example United
Nations Framework Convention on Climate Change, New York (1992) and Kyoto (1998),
http://www.unfccc.de/resource/convkp.html). Climate prediction is mainly based on climate
computer models (i.e. general circulation models and coupled ocean-atmosphere models).
However, reliable predictions need accurate input parameters and the models have to be
checked and verified against past climate changes (Alverson et al., 2000). Checking and
verification is only possible, if the climate changes can be identified in time and space within
the geologic archives. Climate models also compute the extensions of past glaciers and ice
sheets during distinct time slices. Therefore, the extension of former glaciers and ice sheets
traced and mapped with glacial deposits (e.g. moraines, pro-glacial deltas or large bodies of
gravel) and the chronological identification (i.e. the relation of an extension to a distinct time
slice) are keys for the model verification. A wide variety of different dating methods is
available for this chronological identification (e.g. 14C, dendrochronology, pollen analysis,
varve counting, optical stimulated luminescence and paleomagnetic chronologies). Recently,
this was extended by the method of surface exposure dating (SED), which can provide
absolute direct dating of glacial formations within a time frame ranging from millions of
years down to a few thousands of years. Together with careful fieldwork, which is needed for
the understanding of the results, this method is a promising instrument to improve
chronological knowledge of glacial extensions during the past. Within this work, the method
of SED is applied to several geologic problems of the Quaternary and even beyond. These
problems were chosen to yield information for the solution of the “big climate puzzle”. In the
following, each sampling area is introduced and the main goals of this thesis are described.
8
1.2 Younger Dryas Salpausselkä I Formation, Finland
The last glacial-interglacial transition (LGIT) was not a steady and continuous change from
cold to warm climate conditions (e.g. Lowe et al., 1995). During LGIT, a climate reversal,
defined as the Younger Dryas Chronozone lasting from 11’000 to 10’000 14C years BP
(Mangerud et al., 1974), brought back cold conditions. The classic Younger Dryas (YD) area
was first limited to north-western Europe, but Greenland and Antarctic ice core investigations
(Alley et al., 1993; Johnsen et al., 1992; Jouzel et al., 1991) as well as terrestrial records in
North and South America (e.g. Gosse et al., 1995; Hooghiemstra and van der Hammen, 1995;
Peteet, 1995) and New Zealand (Denton and Hendy, 1994; Denton and Hendy, 1995; IvyOchs et al., 1999) showed that a YD signal is also visible outside this area in the Northern and
in the Southern Hemisphere. There are therefore indications that the YD cold reversal was a
global event during LGIT. Still, all investigated archives record a climate change in a timedepending way. To distinguish “real” time lags from “archive-depending” or “methoddepending” time lags, Bard and Kromer (1995) suggest the use of different methods applied
on the available archives. This should improve the global temporal framework of the YD, and
one may be able to identify a possible synchronous or asynchronous behavior of both
hemispheres (e.g. Ivy-Ochs et al., 1999 and references therein).
So far, YD glacial formations have been successfully investigated with SED applied to
boulders lying on moraines (Gosse et al., 1995, North America; Ivy-Ochs et al., 1996, Swiss
Alps; Ivy-Ochs et al., 1999, Southern Alps New Zealand). The SED investigation of the
Salpausselkä I (Ss I) glacial formation in southern Finland yields new chronological
information on a terrestrial YD archive in the Northern Hemisphere that was previously only
identified by pollen investigation, counting methods (i.e. varve counting in lake sediments)
and little 14C data (Tschudi et al., 2000 and references therein). It also shows that SED is
robust towards environmental impact (e.g. vegetation, isostatic uplift, and littoral erosion) and
that it can also be used with very large samples (necessary for very low altitudes and young
ages). The results are presented in Chapter 3.1.
1.3 Wrangel Island, Far Eastern Russia
The extension of ice sheets during the Last Glacial Maximum (LGM) is an important
parameter for testing global climate models. Within the region of Far Eastern Russia,
intensive and thick coverage by ice is suggested by Grosswald (1988, 1997). However, this
hypothesis is questioned by radiocarbon dated mammoth bones, teeth and tusks found on
Wrangel Island, a remote island in Far Eastern Russia close to Bering Strait (Sulerzhitsky and
Romanenko, 1999; Vartanyan et al., 1995). Their data indicate ice-free conditions during
9
LGM on the island. Moreover, Gualtieri et al. (2000) showed by using SED that the extension
of glaciers in the region of Far Eastern Russia was restricted to a much smaller extent than
suggested by Grosswald (1988, 1997).
During an expedition in 1997 on Wrangel Island, geologic traces of former glaciations were
mapped and samples for SED were taken to tackle the question whether there was an ice
coverage during LGM or not. Unfortunately, no distinct glacigenic formations were observed,
which could have been dated with SED (e.g. boulders on moraines). Instead of this, bedrock
surfaces were sampled and analyzed to determine the time interval from the end of the last
coverage by ice to the present. Although the interpretation of these data would not be
unambiguous, I felt encouraged to work on the Wrangel Island samples. This was also due to
the successful investigation on SED samples from similar environmental conditions in
southern Finland (i.e. the dating of Salpausselkä I, Tschudi et al., 2000). The preliminary
results of the Wrangel Island study are presented in Chapter 3.2 and in Karhu et al. (2000).
1.4 Eastern and central Tibet
The Tibetan Plateau with an average elevation of > 4’500m above sea level holds an unknown
potential for Quaternary glaciations. Various publications present different hypotheses with
either huge glaciations of the whole plateau (Kuhle, 1988, 1994 and 1998) or with locally
restricted glaciations (Benxing and Rutter, 1998; Derbyshire et al., 1991; Lehmkuhl, 1998).
Their conclusions are mainly based on geomorphological field evidence. Absolute
chronological information on glacial formations is limited to a few spots (Benxing and Rutter,
1998 and references therein; Lehmkuhl, 1998; Phillips et al., 2000).
Because the knowledge of the volume and the extent of large ice bodies during the past is
essential for the verification of global climate models (Felzer et al., 1998), two field
expeditions were initiated with Chinese partners. The goal was to investigate and sample
different glacial formations for SED and to establish an absolute glacial chronology. The
sampling areas are located in the eastern part of Tibet (Litang and Kanding area) and in the
central part of Tibet (Tanggula area). Nucleogenic neon contribution was significant for the
Litang samples. There, only 10Be and 26Al exposure ages were used for the chronological
identification of the sampled Litang moraines. The preliminary results of a small sample set
from the Kanding area are presented in Chapter 3.3. Preliminary data of the Litang and
Tanggula sample set are presented in Schäfer (2000), whereas the final results appear in
Chapter 3.4 and Schäfer et al. (submitted).
10
1.5 Antarctica (Beacon Valley and Allan Hills nunatak)
The understanding of coupling mechanisms between the Antarctic climate system and the
climate system of the rest of the Earth is one of the key questions in recent global climate
research. An important part within this discussion is the timing of the Sirius Group, a
glacigenic deposit of unknown age found at relatively high elevations at various locations
along the Transantarctic Mountains (Denton et al., 1991). Its age is a crucial factor in the
discussion about the stability of the East Antarctic Ice Sheet (EAIS) (see e.g. Miller and
Mabin, 1998 and references therein). The suggestion of a partial meltdown of EAIS during
the Pliocene warming (Webb et al., 1984) is opposed by a proposed long-term stability of the
Antarctic landscape (e.g. Denton et al., 1993; Marchant et al., 1993a; Sugden et al., 1995a;
Schäfer et al., 1999). Evidence for the latter is based on geomorphological assumptions and
on absolute chronological information from distinct geological formations. This chronological
framework is however questioned in parts (e.g. Harwood and Webb, 1998; Hindmarsh et al.,
1998) and more independent data are needed either to confirm or to disprove previous
datings.
For this purpose, different glacial formations were sampled for SED during several field
expeditions to the Dry Valleys. Within Beacon Valley (lying in the vicinity of McMurdo
Sound), Taylor Glacier, which is an outlet glacier from EAIS, deposited various generations
of glacial drifts. One of these drifts, the Granite drift, seems closely related to buried remnant
ice (Sugden et al., 1995b), which is supposed to be Miocene in age (Sugden et al., 1995b).
However, this chronology is questioned by Hindmarsh et al. (1998). Our SED analyses on
clasts of Granite drift yield new and independent information on this problem (Tschudi et al.,
submitted), confirming recent investigations on nearby Sirius Group clasts on Mt. Feather
(Ivy-Ochs, 1996; Ivy-Ochs et al., 1997; Schäfer et al., 1999). Moreover, multi-nuclide studies
with 10Be, 26Al and 21Ne are used to shed light on the past glacial history of Beacon Valley.
The Allan Hills nunatak, marking a northwestern outpost of the Transantarctic Mountains,
was also investigated by SED, because Sirius Group sediments are outcropping at this ice-free
spot within EAIS (Mayewski and Goldthwait, 1985 and references therein). Moreover,
younger traces of glacial activity are visible at the nunatak. Both deposits were sampled for
SED to establish an absolute chronology. Again, a multi-nuclide approach was made to obtain
information on prior exposure and periods of shielding by the advancing EAIS. The results
are presented in Chapter 4.2, a manuscript, which is close to be submitted (Tschudi et al., to
be submitted).
CHAPTER 2: METHODS
11
Chapter 2
Methods
2.1 Basic principles of SED
In this chapter, the physical principles of SED and the sample processing from the field to the
measurement are briefly described. It shall give an idea of the current progress of SED and it
shall point to the unsolved problems of the method. Chapter 2 focuses on the radioactive
cosmogenic nuclides 10Be and 26Al, as their analysis forms the main part of this thesis. For
further reading, I recommend the articles of Lal and Peters (1967), Lal (1991) and Cerling and
Craig (1994).
Production of cosmogenic nuclides in rock samples
Cosmic ray particles entering the Earth’s atmosphere produce a shower of secondary particles
(mainly neutrons, protons and muons for SED purposes) (Lal and Peters, 1967). Those
particles that reach the terrestrial surface can produce so called cosmogenic nuclides. The
ideal terrestrial SED target mineral for the production of 10Be, 26Al and 21Ne is quartz,
because it has a simple chemical composition (SiO 2), it is abundant in many different
lithologies, and it is quite resistant to erosional processes. All analyses presented in this thesis
were performed on quartz as target mineral. The following statements and discussion are
therefore restricted to quartz.
12
CHAPTER 2: METHODS
Table 2.1: Main target elements and production mechanisms for 10Be, 26Al and 21Ne in quartz.
Cosmogenic
Major production through
nuclide
10
Be
26
21
Al
Ne
•
Spallation
•
negative muon capture
•
fast muon reactions
•
Spallation
•
negative muon capture
•
fast muon reactions
•
Spallation
Major target elements
Half-life used for calculations in
in quartz (SiO2)
this thesis [years]
O
1’510’000 (Hofmann et al., 1987)
Si
716’000 (Samworth et al., 1972)
Si
stable
The concentration of the radioactive cosmogenic nuclides 10Be and 26Al produced in the
Earth’s surface is described with equation Eq. 2.1, where N is the number of cosmogenic
nuclides per gram quartz, P is the production rate function in units of atoms per year and
gram quartz and λ is the particular decay constant in units of per year. Note that the
concentration of N, the production rate P and also the spatial variable x are time-dependent.
Eq. 2.1
dN(x(t),t)
= − N(x(t),t)λ + P(x(t),t)
dt
The crucial parameter in this formula is the function P(x(t),t). Basically, one would like to
have a well-defined numerical value of a reference surface production rate P0, which is
scalable to any place on Earth and to any time during the past. P0 is defined as the production
rate at sea level and at geomagnetic latitudes = 60°, where the Earth’s magnetic field has no
effect on the production of cosmogenic nuclides (Lal and Peters, 1967). Fortunately, it is
possible, in general, to separate the scaling function into a purely spatial and a purely
temporal part. The next sections discuss the various parts one at a time.
Spatial: Atmospheric scaling function
For the incoming cosmic rays, the geomagnetic field and the atmosphere act as filters. Its
properties are mainly dependent on two parameters: the geomagnetic latitude and the altitude
above sea level (see e.g. Cerling and Craig, 1994). The reference production rate P0 has
therefore to be scaled according to the sample site’s geomagnetic latitude and altitude above
sea level. Based on experimental data (Lal and Peters, 1967), Lal (1991) presents a scaling
formalism, which assumes a pure dipole geomagnetic field and standard atmosphere
conditions all over the planet. This formalism is widely used within SED studies. It is
supposed to bear uncertainties of about 10% (Lal, 1991). Recently, Dunai (2000) presented a
reevaluation of the scaling factors using also non-dipole components for the description of the
CHAPTER 2: METHODS
13
geomagnetic field. The differences between the two scaling methods range from -15% to up
to 30%. Below 40° latitude, the 0% level ( = no difference) ranges from 2km altitude up to
4km at about 30° latitude. At latitudes higher than 40° it lies at about 2km altitude. For all
samples discussed in this thesis, the differences are between -5% and +5%. Note that these
differences are calculated with a standard pressure-altitude relationship for the atmosphere
(see Fig. 6 in Dunai, 2000). Determination of in situ production rates at different altitudes and
latitudes would shed light onto this discussion. Such experiments are under way (Graham et
al., 1999; Schäfer, 2000). The use of standard atmosphere condition for the atmospheric
scaling is however an approximation. It is a good approach for most places at mid-latitudes,
but not for high latitudes (e.g. Antarctica) (comment of Monaghan in Reedy et al., 1994;
Nishiizumi et al., 1999; Dunai, 2000). In Antarctica, the atmosphere shows a pressure-altitude
relationship with lower pressure for all altitudes than the standard distribution predicts (see
Fig. 2.6 in Schäfer, 2000). This means that due to this effect the production rate of
cosmogenic nuclides would be higher in Antarctica than presently thought. Another crucial
point in the discussion of the scaling function is the contribution of the muonic component to
the production of the radionuclides 10Be and 26Al (Table 2.1). Lal (1991) assumes that the
muonic component contributes about 16% for 10Be and 17% for 26Al, respectively, to the total
production rate of these nuclides at sea level and high latitude. Heisinger et al. (1997) and
Stone et al. (1998) on the contrary argue that the muonic contribution is less than 5% for both
10
Be and 26Al. This has the effect of reducing the production rates of 10Be and 26Al. For
Antarctica that would mean a reduction in the increase of production rates caused by the
special altitude-pressure relationships. Dunai (2000) and Stone (2000) recently presented new
scaling formalisms, which take these effects into account. Overall, the new scaling would
change the results presented in the Chapters 4.1 and 4.2. This might have an effect also on the
interpretation of the data. The discussion about the atmospheric scaling of the production rate
P0 is however presently ongoing and further investigations are needed, to be able to adjust
scaling to different “types” of atmospheres.
Spatial: Depth dependency of production rates in rock material
The discussion about depth dependency is important for two reasons: (1) The in situ
production rate used in Eq. 2.1 just represents the surface value. A distinct sample thickness
requires a corrected production rate. (2) Estimation of prior exposure may need the
calculation of production rates at greater depths (e.g. buried samples within bedrock, cf.
Chapter Beacon Valley rockfall). Presently, there are two different models, which describe
the depth dependency of the production rate within rock material. Lal (1991) assumes that the
production rate decreases directly from the surface on with an exponential function. At the
depth x within a rock the production rate is calculated according to Eq. 2.2, where PS is the
14
CHAPTER 2: METHODS
local production rate at surface (atoms/yr·g SiO 2), ρ is the rock density (g/cm3) and Λ is the
1/e attenuation length for production of cosmogenic nuclides.
Eq. 2.2
P(x) = PS exp(−
ρx
)
Λ
If this function is integrated over the sample thickness (e.g. from surface down to depth x),
one obtains the thickness-corrections for surface samples (Eq. 2.3).
Eq. 2.3
P(0 to x) = PS
Λ
ρx
1− exp(− )
ρx 
Λ 
Another depth dependency model for the in situ production within rocks is that of Masarik
and Reedy (1995). It proposes a nearly constant production rate for the nucleonic component
from the surface of the rock down to a level of about 12g/cm2 within the rock, due to the
boundary effects of the air-surface interface. This means that the production rate is constant
for the first 4.6cm, if a typical rock density of ρ = 2.6g/cm3 is assumed. Note that Masarik and
Reedy (1995) use an attenuation length Λ = 157g/cm2, instead of the traditional value of
150g/cm2 of Brown et al. (1992) and Lal (1991). The model of Masarik and Reedy (1995) has
further implications for SED studies, because it bears a decreased sensitivity of the actual
production rate towards coverage of the sampled surface with biomass or snow.
Both depth-dependencies described neglect nuclide interactions other than spallations, which
limits the relevant "production layer" within rocks to the uppermost centimeters. However, a
more realistic model would consider other reactions too (e.g. muonic). Heisinger et al. (1997),
Heisinger (1998) and Stone et al. (1998) describe the production mechanisms of muons
within rock material. Below a surface layer of about 3-4m, muonic processes dominate the
production of cosmogenic nuclides. From this depth on, the decrease of the production rate
with depth is not following Eq. 2.2 anymore (see Heisinger, 1998 for details). The integration
of muonic production mechanisms greatly impacts on the data interpretation of SED studies
in case of dating, where prior exposure is considered or bedrock samples are analyzed, or in
case of erosion studies. However, the influence of muonic production is accepted but not yet
widely in use within SED studies, because its handling is rather complex and not yet fully
agreed on. Therefore, more studies are needed to confirm (or also to disconfirm) the
importance of the "muonic component" within the production mechanisms of cosmogenic
nuclides.
CHAPTER 2: METHODS
15
Spatial: Impact on local surface production rates by shielding and coverage
Shielding of the sample by surrounding mountains or a dipping surface, and coverage by
biomass, soil or snow decrease the incoming cosmic ray flux and therefore the local surface
production rate. The shielding by surrounding mountains can either be directly measured in
the field using a clinometer or it can be determined after fieldwork with the help of a
topographic map (manually or with a Geographic Information System, GIS). The shielding
angle is given in degrees from the horizontal for distinct parts of the total 360° of azimuth.
Corrections are computed according to the angular distribution of the cosmic ray particles in
the troposphere (Lal, 1958; see Nishiizumi et al., 1989 for details). For e.g. 15° shielding
angle for 45° of the azimuth and 10° shielding angle for the rest of the azimuth, the surface
production rate would be reduced by 0.5%.
The effect of biomass (e.g. trees, bushes and moss) can be approximated by estimating a
covering layer of vegetation with the appropriate density of biomass on the sampled surface
(Kubik et al., 1998). But, biomass is not constant in time, and no information is available
about past distribution. If the effect of biomass on the local surface production rates is
estimated with present day’s parameters, one can easily over- or underestimate the coverage
of the past. For samples from the LGIT, one can make the assumption that past biomass was
equal or less during the period of exposure than today’s biomass (if the past climate was less
optimal for plant-grow). Therefore, using today’s coverage has the maximal effect on the
local surface production rates. Note that the magnitude of this effect is dependent on the used
depth-dependency model for the production rate. If the “classic” exponential decrease is used,
moss coverage of 5cm with a density of 0.2g/cm3 (determined at Salpausselkä I) yields a
reduction of the production rate of less than 1%. If, on the other hand, the model of Masarik
and Reedy (1995) is used, then the production rate is not affected at all by this moss coverage,
because the production rate is flat down to a depth of 12g/cm2, which equals 60cm of moss
coverage(!). The impact of snow on the local surface production rate is similar to the effect
caused by biomass. Again, the past distribution of the coverage is unknown. Moreover, even
the present day’s thickness is often uncertain (consult local authorities or see regional climatic
time series and averages, if available). For Finland (Chapter 3.1), the snow thickness was
estimated to be 30cm during 3 months of the year. There, own field observations during
wintertime showed that the boulders are not just covered by snow, but also by an ice layer of
up to 10cm. Depending on the model (“classic” exponential or Masarik and Reedy, 1995), the
influence of this layer ranges of the order of a 0-2% decrease of the theoretical production rate
at the rock surface (calculated with an estimated snow density of 0.2g/cm3 and an ice density
of 0.9g/cm3). Compared to shielding by mountains or a dipping surface (both measurable), the
effect of biomass or snow/ice is difficult to estimate and one may over- or underestimate it,
16
CHAPTER 2: METHODS
because past distribution of the parameters is unknown. However, the maximal effects seem
to vary only within a few percent.
Temporal: Time variations of the geomagnetic field
The temporal variation of the reference production rates P0 is mainly dependent on variations
of the strength of the geomagnetic field, because this results in a varying cosmic ray flux on
the terrestrial surface (e.g. Cerling and Craig, 1994). These variations are smoothed out
because the reference production rate P0 is the integrated rate over the whole time interval of
the calibration period (e.g. about 10ky for the values of Kubik et al., 1998 or several years for
artificial targets). In principle one needs a calibrated production rate P0 for all expected
exposure periods of his samples. If no suitable calibration is available for the expected
exposure interval, one may apply corrections for the used production rate. Cerling and Craig,
(1994) present variations in the integrated production rate of 3He. Their calculations show two
major aspects: (1) the influence of a changing magnetic intensity is decreasing with increasing
geomagnetic latitude and is insignificant for latitudes > 60° and (2) the effect of a changing
magnetic intensity on the integrated production rate is smoothed out at about 50ky ago. In a
recent work, Licciardi et al. (1999) suggest that the effect of a variation of the paleointensity
on the integrated production rate is much smaller for the past 10ky at mid-latitude than
expected. Beside a changing geomagnetic intensity, the geomagnetic field shows secular
variations, which let the geomagnetic dipole axis circle around the geographic pole. This
process results in a difference between geomagnetic and geographic latitudes. If a long-term
integrated reference production rate is used (> 10ka), this spatial variation of the geomagnetic
dipole is smoothed out (Ohno and Hamano, 1992, 1993). In that case one can use geographic
latitude instead of the geomagnetic latitude for the scaling of the reference production rate to
the local surface rate.
Most samples of this work originate from latitudes > 60°. Therefore, no corrections were
necessary (Cerling and Craig, 1994). The correction for the Tibet samples (sample site at
about 30°N) was estimated to be less than 2% of the actual surface production rate (Schäfer et
al., submitted). It was therefore neglected.
Spatial and temporal: Changes in production rate due to uplift
Tectonic uplift changes the sample’s altitude with time. It therefore changes the local
production rate of the sample. This effect depends on the uplift rate as a function of time. For
a first estimation, uplift is neglected and the present day’s altitude is taken as constant during
the period of exposure. The assumed uplift rate (from literature, if available) is taken to
calculate the former altitude of the sample (using the exposure age as time interval). If the
CHAPTER 2: METHODS
17
uplift rate is constant in time, a new averaged production rate is easily calculated as the mean
value from the former and the present day’s value, using the appropriate altitude scaling. The
exposure age becomes older with the newly averaged production rate and one has to repeat
the steps recursively. Commonly, tectonic uplift is slow and limited to a maximum rate of a
few millimeters/yr. Then, no large effect on the production rate is expected (Ivy-Ochs et al.,
1999; Schäfer, 2000) and the use of the present day’s altitude for the determination of the
local production rate is justified (note that the age is still a minimum age, because the present
day’s production rate is higher than the true one). However, the impact may be significant, if
uplift is rapid and intense. In Finland, for example, uplift dramatically modified the landscape
during the past 10ky. There, the crust experienced isostatic rebound (Mörner, 1980) after the
Scandinavian Ice Sheet melted away. The speed of this rebound was fast at the beginning and
slowed down towards the present (see Fig. 5, page 259, in Mörner, 1980). Such uplift data
were determined from the analysis of shoreline displacements. They were used to reconstruct
the former absolute elevation of the sample site and to evaluate the average local production
rates for the whole period of exposure (Tschudi et al., 2000).
Spatial and temporal: Erosion
Erosion affects the local production rate both spatially and temporally. The effect of erosion
will be discussed later as it is best viewed in the context of determining exposure ages.
Reference production rate P0
The sections above explained how to calculate the local cosmogenic nuclide production rate
by scaling a reference rate appropriately in space and time. This section concentrates again
only on the determination of P0 for 10Be and 26Al.
Traditionally, P0 is a numerical value for standard atmosphere sea level altitude and high
geomagnetic latitudes (≥ 60°). However, the calibration sites for P0 were at quite different
locations (e.g. Sierra Nevada, 38°N, > 2'000m.asl Nishiizumi et al., 1989; Köfels, 47°N,
> 1'600m.asl Kubik et al., 1998). Moreover, shielding or coverage have affected the sample
sites and corrections were inevitably made during data reduction. Therefore, the published
reference production rates depend on the applied models for atmospheric scaling, depth
dependency and estimation of shielding and coverage. A first set of reference production rates
for 10Be and 26Al was published by Nishiizumi et al. (1989). These values are based on a high
muon contribution to the total production of the cosmogenic nuclides and a “classic”
exponential depth dependency and shielding correction. For the scaling to sea level and high
geomagnetic latitudes, the formalism of Lal and Peters (1967) was used. Lal (1991)
formulated a scaling table for altitude and geomagnetic latitude corrections, based on the
18
CHAPTER 2: METHODS
values of Nishiizumi et al. (1989). Later on, Nishiizumi et al. (1996) recalculated the original
production rates to take into account a new estimate of the exposure time of their calibration
sites and to replace geomagnetic with geographic latitude (see section Temporal: Time
variations of the geomagnetic field). Both original and revised production rates as published
by Nishiizumi et al. (1989) and Nishiizumi et al. (1996) are listed in Table 2.2. This table also
lists the estimate of Lal (1991) that is used in many SED studies to date and the values of
Kubik et al. (1998), which is used in this thesis. Kubik et al. (1998) used the same scaling
formalism as in Lal (1991), but taking into account that geomagnetic latitude should be
replaced with geographic latitude for samples with exposure ages of 10ka or longer (Kubik et
al., 1998 and references therein), and the same shielding functions (mountains, dip of sample
surface) as Nishiizumi et al. (1989). However, the reference values of Kubik et al. (1998) are
based on the “flat” depth dependency of Masarik and Reedy (1995) instead of the “classic”
exponential decrease.
Note that the use of reference production rates implies the use of certain data reduction
models (e.g. Lal’s formalism for atmospheric scaling, “classic” or “flat” depth dependencies).
Within the data reduction of the samples, the same models have to be used to be consistent. If
one wants to use other models (e.g. “flat” depth dependency with Nishiizumi et al.’s values),
then the reference production rates have to be reevaluated according to these models.
Table 2.2: Production rates P0 of cosmogenic nuclides in quartz at sea level and latitudes = 60°.
Cosmogenic nuclide
Nishiizumi et al. (1989)
P(10Be) [atoms/yr·g SiO2]
P(26Al) [atoms/yr·g SiO2]
6.03 ± 0.29
36.8 ± 2.7
5.994
36.67
5.80
(35.4)*
5.75 ± 0.24
37.4 ± 1.9
Lal (1991), (based on
Nishiizumi et al., 1989)
Nishiizumi et al. (1996)
Kubik et al. (1998)
*calculated with the ratio given in Nishiizumi et al., 1989
For this thesis, two different “sets” of production rates are used for the calculation of
exposure ages. For samples from Beacon Valley, Antarctica, the values of Lal (1991), based
on Nishiizumi et al. (1989) were used, because previously published data in the vicinity of our
sample site are calculated with these values. For the other studies (Finland, Wrangel Island,
Tibet and Allan Hills) the recently published values of Kubik et al. (1998) were used. The
application of their values is justified, because the calibration is independent based on dendrocalibrated radiocarbon datings and because they use the more physical depth dependency of
Masarik and Reedy (1995), instead of the exponential decrease. In addition, their production
rates and the revised rates of Nishiizumi et al. (1996) are in very good agreement (Table 2.2).
CHAPTER 2: METHODS
19
The use of Kubik et al. (1998) has also another justification. Presently, there is an ongoing
discussion about the use of different 26Al standards at different AMS facilities. If 26Al
measurements of the Zurich AMS facility are concerned, the use of reference 26Al production
rates of Kubik et al. (1998) elegantly eliminates any problem, because both, the
measurements of Kubik et al. (1998) and the sample measurements, are then calibrated to the
same 26Al standard. This yields results that are independent of the 26Al standard.
Age calculation and erosion
The formalism that is used to calculate exposure ages with the radionuclides 10Be and 26Al is
based on Eq. 2.1. If the “classic” exponential depth dependency for the production rate is
used, the integration yields Eq. 2.4, where N is the number of produced cosmogenic nuclides
(atoms/g SiO 2), PS the local production rate (atoms/yr·g SiO 2), λ is its decay constant (yr -1),
ρ is the rock density (g/cm3), ε is a so-called steady-state erosion rate (cm/yr) having affected
the sample surface over the whole time T of exposure, Λ is the 1/e attenuation length for
production of cosmogenic nuclides, T is the exposure time (yr) and N0 is the inherited amount
of cosmogenic nuclides.
Eq. 2.4
PS
N=
ρε
λ+
Λ

ρε 

− λ +  T 
1 − e  Λ   + N0


The parameters PS, λ, Λ and ρ are considered to be known, whereas N is measured with the
appropriate method (AMS in the case of radionuclides 10Be and 26Al). ε and T are unknowns.
Note that λ is equal to zero, if the equation is used for the stable 21Ne. The value of Λ is
chosen according to the used reference production rates. For Kubik et al. (1998), a value of
157g/cm2 was chosen (Masarik and Reedy, 1995), whereas the reference values of Lal (1991)
and Nishiizumi et al. (1989) require a value of 150g/cm2 (Brown et al., 1992). Eq. 2.4 is
underdetermined and one cosmogenic nuclide is not sufficient for the determination of both
erosion rate ε and exposure time T. Cerling and Craig (1994) suggest the use of the two
limiting cases, where (1) a minimum time of exposure and (2) a maximum erosion rate are
calculated. On the one hand, the erosion rate ε is set to zero to obtain a minimum exposure
age of the surface. This "zero-erosion" approach is often used, as most SED studies are
launched to yield chronological information on distinct geologic formations. On the other
hand, the exposure time is set to infinity and the equation enables a calculation of maximum
in situ erosion rates. “Realistic” exposure ages can be determined, if erosional processes are
considered. If the “classic” exponential depth dependency is used, the exposure ages can
easily be determined by using Eq. 2.4. If on the other hand the “flat” depth dependency of
Masarik and Reedy (1995) is used (as for most of our data), Eq. 2.1 does not lead to an
20
CHAPTER 2: METHODS
analytical solution. The function for the calculation of exposure ages with erosion rates has
then to be solved with numerical methods.
Steady state erosion
Steady state erosion ε can directly be used within Eq. 2.4. The determination of a local steady
state erosion rate is indeed difficult, and only little information is available. Most erosion rate
studies deal with long-term erosion phenomena and regional run-off estimations (Clayton and
Megahan, 1986; Saunders and Young, 1983). However, these studies give a possible interval
for erosion rate values found at different places on Earth. The problem of erosion rate
determination can also be tackled with cosmogenic nuclides, when a multi-nuclide approach
is performed. There, a second or even third cosmogenic nuclide is determined within a single
sample. If the local steady state erosion rate ε is assumed to be smaller than 1m/Ma (Small et
al., 1997) (e.g. in the Dry Valleys), and the sample plots within the “erosion island” of the
two-nuclide-plot (26Al/10Be versus 10Be) (Klein et al., 1986; Lal, 1991; Nishiizumi et al.,
1991), then a steady state erosion rate can be determined, by simultaneously solving the
system of equations for the erosion rate ε and the exposure time T. However, an exact
analytical determination of these values is mostly not performed, because the uncertainties are
too large (for an error estimation see Lal, 1991). Instead of analytical techniques, one can also
use a graphical approach with the two-nuclide-plot, where trajectories of distinct steady state
erosion rates are plotted and compared to the data. This technique enables an estimate of a
maximum steady state erosion rate. It was used in Chapter 4.1 (Tschudi et al., submitted).
Episodic loss of mass
One could argue that steady state erosion is not realistic for a particular sample, but that
episodic loss of mass should be considered as the dominant process. Small et al. (1997) model
an example where a rock surface loses a surface layer of 15cm every 100ky. After a time of
300ky, the resulting amount of 10Be and 26Al is about to be equal to the amount, which is in
equilibrium with a steady state erosion of 5m/My. Steady state erosion is therefore a good
approximation for episodic loss of mass, if long periods of exposure (i.e. several 100ky) with
only a few loss events can be assumed. Again, Eq. 2.4 can be used to estimate exposure ages.
For shorter periods of exposure (e.g. 12ka for Younger Dryas), an episodic loss of mass
would be dramatic, because the sampled surface would then not reflect the true top surface
right after the period of exposure began. The resulting SED would be wrong (i.e. exposure
ages would be much too young, see Fig. 2.4). It is therefore important to chose appropriate
samples in the field (whenever possible) that do not show obvious erosional features, like
spalling or flaking. We had good experience with samples that bear rock knobs, which are
more resistant to erosional processes than the surrounding rock surface (see Fig. 2.5). To
CHAPTER 2: METHODS
21
calculate “realistic” exposure ages for these surfaces, independently determined steady state
erosion rates (e.g. Clayton and Megahan, 1986; Saunders and Young, 1983) were used within
Eq. 2.4.
Conclusion
For any determination of exposure ages it is absolutely important to give reference to the
following parameters: reference production rates used, the scaling and depth dependency
models applied and the corrections used for shielding, coverage and geomagnetic variations.
This is vital for an easy application of future improvements of any kind (e.g. adjusted scaling
or new reference production rates) to the published data.
2.2 Application of SED in Quaternary geology
The time-dependency of the concentration of cosmogenic nuclides (Eq. 2.4) and the limited
penetration depth of cosmic rays within rocks (see Chapter 2.1) allow the use of in situ
produced cosmogenic nuclides as a chronometer, if applied to an appropriate rock surface.
Erosional and depositional glacial processes, as sketched in Fig. 2.1, may produce such
surfaces that are crucial for the determination of an absolute glacial chronology. The key for a
suitable rock is indeed the relation between its exposure to cosmic rays and the glacial process
that is to be dated.
Fig. 2.1: Schematic discrete erosion by advancing ice with subsequent deposition. Both samples are
directly related to the glacial advance. SED delivers the timing of the glacial retreat, which marks the
beginning of the exposure to cosmic rays.
22
CHAPTER 2: METHODS
In principle, the surfaces sketched in Fig. 2.1 fulfill this relation, because they were shielded
from exposure within bedrock before the glacier advanced. Dating of the bedrock sample
therefore yields the age of the glacier’s retreat. The dating of glacial deposits (e.g. moraines,
tills or erratic boulders) allow different interpretations. On the one hand, moraines mark a
stable glacial position for a certain time period. Their dating therefore directly yields an age
of a glacial stage. Tills on the other hand pinpoint the presence of a glacier. Their first
exposure to cosmic rays is related to the deglaciation (e.g. in the case of subglacial till).
Erratic boulders, not related to a moraine ridge “float” chronologically seen somewhere inbetween, because their deposition may occur at an arbitrary time and place within the close
vicinity of a glacier (e.g. (1) at the glacier’s front, they date the maximum extent, (2) at lateral
position, they date the presence of a glacier at that time, deposited as a drop stone (3), they
mark the disappearance of water covering the deposit). One crucial question is, however,
whether the sampled surface bears any inherited cosmogenic nuclides. There is no or
negligible inheritance if the sample was shielded from cosmic rays by several meters of rock
or ice before it was excavated and deposited again. The shielding depth within rock, which is
necessary to prevent significant prior exposure, can be estimated with Eq. 2.2.
An example: a sample that was buried at a depth of 3m, has a local production rate of < 0.5%
of the surface value (if only production through spallation is considered). If the muonic
component is considered too (Heisinger et al., 1997), then the production rate at this depth is
about 2% of the surface value. Still, this production rate is low and the effect of inheritance is
rather small. In another example taken from Chapter 3.1 (Younger Dryas) one can estimate
the inheritance for a boulder lying on a YD moraine (exposure period about 12ka, measured
10
Be concentration about 7x104 atoms/g SiO 2 with a production rate of about 6atoms/yr·g
SiO2, for sea level at geomagnetic latitudes = 60°). The amount of inherited 10Be depends on
the previous location of the boulder (i.e. depth of burial before it was excavated) and on the
erosion rate of the landscape. Assuming saturation, an original depth of about 3m and no
erosion, one would then expect about 1x105 atoms/g SiO 2, which is larger than the measured
amount of 10Be. Thus, the sample must have been buried at greater depth or the erosion rate
must have been larger. If an erosion rate of 20mm/ka is assumed (a realistic value, see
Saunders and Young, 1983), then, the subsurface saturation level for 10Be yields only about
2% of the measured concentration of 7x104 atoms/g SiO 2.
Boulders seem to be a good host for suitable surfaces, because they often originate from
rockfalls, where it is likely that the sample was buried at greater depth. Boulders were
successfully used to apply SED (Gosse et al., 1995; Ivy-Ochs et al., 1996; Ivy-Ochs et al.,
1999; Phillips et al., 1997; Tschudi et al., 2000). Sampling fresh bedrock surfaces that are
related to glacial advances (e.g. striated bedrock or excavations, Fig. 2.1) could be more
CHAPTER 2: METHODS
23
troubled with inheritance, if the advancing glacier did not erode enough covering material
away. Unfortunately, little is known about a glacier’s capability of eroding bedrock, and
many unknown parameters influence this feature (e.g. temperate or cold based, lithology and
concentration of debris within the ice). Preservation of old polished surfaces and previous
existing striae might occur in a cold-based glacier environment, but this is questioned by the
recent work of Cuffey et al. (2000), who discusses the possibility that also cold-based glaciers
can cause significant erosion. In general, glacial deposits should be suitable SED surfaces.
However, the problem of inheritance of cosmogenic nuclides has to be considered, especially
when bedrock samples are analyzed.
2.3 SED Sampling: Preparation and fieldwork
Careful preparation of the field campaign and a proper sampling strategy are probably the
most important things within a study on cosmogenic nuclides. In the following, a short and
general guideline on "how to get a sample" will be presented. However, this chapter is by no
means a complete list and it shall not be a step-by-step recipe. Too variable are the geologic
problems addressed and too different are the sampling areas. Every SED campaign has its
own and individual setting.
Theoretical estimate
Let us assume that we would like to date a distinct glacial formation, e.g. an end moraine.
Two questions should be answered before the first strike of a hammer is performed:
•
•
What is the altitude range of the sampling area?
What is the expected age of the glacial formation?
Both answers are essential for an estimate of the amount of quartz needed to get satisfying
results during the AMS measurement. This estimate can be based on a simple calculation,
which includes the sample’s altitude and its estimated exposure age (Eq. 2.5, which is derived
from Eq. 2.4).
c⋅
Eq. 2.5
NA
⋅r
mBe
w= p
Be
⋅(1 − e ( −T ⋅λ ) )
λ
Here w is the amount of quartz in grams, c is the amount of 9Be carrier in milligrams (added
during the processing), NA is the Avogadro number (NA = 6.022x1023), mBe is the atomic mass
of 9Be, r is a limiting AMS 10Be/9Be ratio, pBe is the local production rate for the sample site
24
CHAPTER 2: METHODS
in atoms/yr·g SiO 2, λ is the decay constant for 10Be (4.59x10-7 per year) and T is the expected
exposure time in years. The limiting 10Be/9Be ratio is estimated from the AMS efficiency and
from typical AMS Be blank ratios. Together with an assumed quartz content of the sampled
lithology, the total amount of sample material needed can then be estimated. Note that the
calculations result in an approximate amount of quartz needed as many parameters, which
influence the production of 10Be, are not considered (e.g. shielding of the sample, cover by
vegetation or snow). However, Eq. 2.5 is quite useful under field conditions. A typical output
of the calculation is given in Fig. 2.2, where an example for a sample-area lying at
geographical latitude of 40° North (or South) and at altitudes ranging from 0 to 4’000m above
sea level is shown. Using this estimate is no guarantee for successful SED results, of course,
but the calculation prevents at least 10Be concentrations too low for AMS measurements.
Similar estimates can be made for other cosmogenic nuclides (e.g. 26Al and 21Ne).
Fig. 2.2: Altitude-weight graph, where the amount of quartz needed is shown for a sample at latitude
40°N and within an altitude range of 0 to 4’000m.asl. The amount of quartz is calculated for assumed
exposure ages between 5 and 11ky. The calculations are based on production rates of Kubik et al.
(1998) and the scaling formalism of Lal (1991) with a 9Be carrier addition of 0.3mg.
Select a sample
SED is the hunt for the most representative samples of a distinct geologic formation that shall
be dated, where disturbing processes, which could also affect SED (see Fig. 2.3), are
minimized. By own experience, this attempt is optimized is increased by consulting as much
information as possible about the target object and area. Small-scale mapping of the
CHAPTER 2: METHODS
25
surrounding area before the field trip prevents geologic traps, hidden to the eye of a first time
visitor of a sample site. It is important that geologic background knowledge and basic
understanding of geologic processes are present in the mind of the sampling crew.
Fig. 2.3: Many parameters affect the method of SED. Shielding, vegetation, snow and tectonic uplift
influence the local incoming cosmic ray flux and therefore the local production rate. Earth surface
dynamics, vegetation, erosion and other unknown processes may disintegrate the target rock and
change its original orientation. Example: Erratic boulder, Kola Peninsula, Russia.
During the past field parties, various criteria of suitability crystallized. They are subjective
and weighted by my own experience. They are adjusted to bouldery deposits, as I was mostly
dating glacial formations by the use of boulders. To fulfill the required representativity of a
sample, a boulder should have had a stable geomorphic position during the time of exposure.
Unfortunately, it is almost impossible to make statements about geomorphic stability from the
present day situation. Nonetheless, we can derive a rough estimation of stability from the
boulder’s position within a formation (e.g. moraine). We expect that boulders on slopes are
more likely to have been affected by solifluction, than boulders from top positions. Boulders,
perching on moraines, are therefore preferred, whereas boulders on slopes should be avoided,
because solifluction might have turned them upside down. Furthermore, the boulders should
have a certain size. We assume that small ones (less than 0.5m3) are more mobile and,
therefore, are more affected by surface processes. In such cases, SED does not necessarily
date the oldest surface of a boulder. However, dating of such surfaces still leads to a
minimum age of the formation, but of course, it would not be the oldest possible age. Indeed,
also big boulders of sufficient size can be moved or turned over. We just assume that it is less
likely than for smaller ones. Moreover, some small boulders with a top surface at less than
26
CHAPTER 2: METHODS
about 50cm above ground may be covered by snow (from field experience in Finland during
wintertime). Finally, erosion and weathering are considered, as they directly affect the
nuclide’s concentration in the rock (see also Chapter “Age calculation and erosion”). The
boulder should yield a certain resistance to in situ weathering and erosion. Boulders with
obvious erosional features like spalling or breaking apart of the surface should be avoided.
The results could just pinpoint the erosional event (Fig. 2.4).
Fig. 2.4: Quaternary deposit in Litang county, Tibet, showing exciting erosion processes. Despite its
huge appearance, the dating of this boulder did not yield satisfying results compared to the existing
stratigraphy and other SED investigations within the area. Its surface age may pinpoint the breakingapart event.
Another erosional feature is shown in Fig. 2.5. No obvious breaking apart is visible, but some
parts of the boulder seem to be more resistant to in situ weathering than the rest of the rock.
There, considerable erosion has to be taken into account. Little erosion on hard rocks can be
expected when the rock’s surface is significantly polished (e.g. by wind erosion). The grain
size of the sampled lithology is also an indicator for potential erosion: Whereas fine grained
lithologies often yield low erosion rates, coarse grained rocks show an increased permeability
and therefore an decreased resistance to erosional processes.
CHAPTER 2: METHODS
Fig. 2.5: Close up of a boulder in Litang county, Tibet. Considerable erosion of several centimeters has
to be assumed. Next to the hammer, a knob of rock material is more resisting to erosion than the rest of
the boulder. Without this feature, erosion would be underestimated.
Fig. 2.6: A boulder of the Salpausselkä I formation in southern Finland showing the presence of in situ
grown peaty soil. Shielding by vegetation decreases the production of cosmogenic nuclides and the
dating would result in a too young age. This boulder seems not suitable for SED. Fig. 3.2 shows
another example, where surrounding vegetation decreases the in situ production of cosmogenic
nuclides.
27
28
CHAPTER 2: METHODS
Vegetation is another important environmental parameter of suitable SED samples. The case
that trees and bushes surround or even cover the boulders is often encountered in the field
(e.g. moraine in a forest, Fig. 3.2). A less common case is in situ growing of vegetation on a
boulder’s surface (Fig. 2.6). Both effects reduce the production rate of cosmogenic nuclides.
Not taking this into account will result in a calculated exposure age that is lower than the true
one (see section “Spatial: Impact on local surface production rates by shielding and coverage”
for quantification). However, the effect is rather small compared to other uncertainties.
If a boulder seems to be suitable, the sampling can start following the steps given in Table
2.3. More than one sample is taken for a particular geological feature, if suitable boulders are
found. All samples presented within this work were taken according to these guidelines. Some
may argue that drill tools might be easier and better than hammers and chisels. During the
past fieldwork, it proved that simple and straight equipment is most convenient for surface
exposure samples. The sample sites are quite often remote places, without car access.
Additional heavy tools, beside the described material, would be a logistical challenge for the
whole party. Moreover, on-site supply of (cooling-) water and fuel is difficult.
Table 2.3: Guideline for sampling SED samples.
Step
•
Sketch and describe the boulder, include the boulder’s dimensions
•
Take pictures of the boulder and its surroundings
•
Measure the shielding of surrounding hills and mountains
•
Describe vegetation on and around the boulder
•
Determine position and altitude (topographic map, GPS and altimeter)
•
Determine the sample’s geometry and measure the dip angle of the surface
•
Mark the top position of the sample
•
Mark the direction of the surface dipping
•
Take the sample
•
Check the sample’s weight and consider the altitude-weight graph
•
Take pictures of the taken sample
•
Note the sample’s thickness (if the sample is not broken apart)
CHAPTER 2: METHODS
29
2.4 SED sample processing: Laboratories
The basic principle of getting 10Be and 26Al out of a rock sample has been described in IvyOchs (1996), Kohl and Nishiizumi (1992) and Nishiizumi et al. (1989). The processing of the
samples for this thesis followed the path outlined in (Fig. 2.7).
Fig. 2.7: Flowchart from the raw SED sample to the AMS measurement.
Quartz purification
After the sample was carefully described in its surficial features (e.g. sealing or oxidation of
the surface) the rock was mechanically treated with a rock-crusher, hammer and metal mortar
to a final grain size, similar to the mineral grain size of quartz within the sample. For our
samples, this size ranged from 1 to 2mm. The finer fraction (clay and silt) was rinsed away
with water. The sand was then filled into PE bottles, where the quartz was separated and
etched. All processing steps were carried out with ultrapure water (Millipore “Milli-Q
gradient” water purification system) with a conductivity of 18MΩ/cm. Separation was done
with hydrofluoric acid (HF < 5%) in several steps (up to 5 steps with rinsing and replacement
of the acid). This process was stopped when the quartz had reached a purity of about ≥95%
(checked under the microscope). The etching removed from the quartz the meteoric 10Be
(Brown et al., 1991; Ivy-Ochs, 1996). Meteoric 10Be is produced in the atmosphere and
washed out with precipitation to the Earth’s surface. There, it penetrates rocks and absorption
30
CHAPTER 2: METHODS
is possible within microscopic cracks and intergranular fractures. 10Be deposition rates lie
between 1.5 and 4.5x10-2 atoms/cm2·s (measured in rainfall, deep-sea and lake sediments, and
ice cores) (McHargue and Damon, 1991 and references therein). The accumulation per year
and cm2 is therefore much higher (on the order of 1x105) than the in situ production of 10Be.
Note that the use of ultrapure water and ultrapure chemicals (HCl, HNO3, NH3) during the
processing is absolutely necessary to prevent a re-contamination of the sample material with
meteoric 10Be contained in normal water.
Adding the spike (9Be and 27Al)
AMS does not measure absolute concentrations, but isotopic ratios. These are, in the case of
SED, 10Be/9Be and 26Al/27Al, respectively. We therefore have to determine the amount of 9Be
and 27Al contained in the AMS sample to get the concentration of 10Be and 26Al. The amount
of 9Be in quartz is very small (< 2ppm for whole rock). Therefore, the technique of “spikeaddition” is used, where a known amount of 9Be is added to the sample. The addition of an Al
spike is normally not required, because quartz bears enough Al for successful processing.
However, whenever only a small amount of quartz was used (< 5g), Al carrier were added to
prevent failure during sample processing. Beryllium and, if necessary Aluminum carrier
solutions, 9Be(NO3)2 and 27Al(NO3)3, respectively (Merck ICP solutions with concentrations
of 1000mg/l), are added with Eppendorf pipettes to the dry and weighed samples into a dry
Teflon beaker (usually 250ml, if required 500ml). The amount of carrier was adjusted within
limits according to the expected exposure age and the available quartz (cf. page 23). Usually,
a volume of 0.3 - 0.4ml carrier solution was added to the samples (which equals to 0.3 0.4mg of 9Be or 27Al).
Blank preparation
Parallel to the samples, blank solutions were prepared to check, whether non in situ 10Be or
26
Al might contaminate the samples during the processing. This contamination could occur
through chemicals and labware, atmospheric dust in the laboratory or cross-contamination
from other samples, which were processed at the same time. The blank solutions were
prepared and treated in the same manner as the “real” samples. All samples therefore had their
corresponding blanks. Both carrier solutions were also tested on their 10Be/9Be and 26Al/27Al
ratios with typical values of 0.02x10-12 for 10Be/9Be and 0.01x10-12 for 26Al/27Al.
HF digest
The HF digest was the next step in the sample processing. The acid (HF 48%) was carefully
poured on the quartz in the Teflon beaker just so that the quartz is covered. The digestion was
accelerated by heating on a hot plate. Note that the heat was only turned on after a certain
CHAPTER 2: METHODS
31
time to prevent strong initial reactions between the acid and fine-grained quartz. The final
standard temperature level was 90°C. After the digest (HF dissolves about 10g of quartz per
day), the remaining SiF4 was fumed off with the help of HCl and HNO3. Both acids were
added to the dry sample and fumed off on the hot plate at 90°C. This process was first
repeated three times for HCl and then three times for HNO3.
Al aliquot and ICP-AES
Because the total Al content of the quartz is the sum of in situ Al and the added spike (where
necessary), it had to be determined independently, so that the 26Al concentration could be
calculated with the help of the AMS measured 26Al/27Al ratio. To avoid problems of
inhomogeneous Al distribution within the minerals, the total Al content was measured in a
solution of the dissolved quartz. The Al aliquot was taken as early as possible during the
chemical processing, right after the fuming of SiF4. The sample was dissolved with HCl and
rinsed into a 100ml flask. This flask was filled up to the mark with ultrapure water and then
shaken thoroughly. Approximately 5ml of the solution were taken out with a pipette as the Al
aliquot. The measurement with ICP-AES (Inductively Coupled Plasma Atomic-Emission
Spectroscopy) was made according to the method described in Ivy-Ochs (1996). For one
sample to be measured with ICP-AES, four different sub-samples were prepared with
increasing standard additions (commercial Al standard solution from Merck with an Al
concentration of 100ppm). The first sub-sample was measured without standard addition,
whereas the following had increasing additions of 100µl, 200µl and 300µl, respectively. The
addition of standard solutions to the sample minimized the influence of unstable ICP-AES
measurements.
In general, the measured aluminum concentrations ranged from about 30 to about 450ppm,
depending on the quartz. For the ICP-AES measurement, the Al aliquot was diluted in an
appropriate ratio. I mostly used a dilution factor of 10. This means that 500µl of the Al aliquot
was poured in a tube, which was then filled up to 5ml with water. If the sample could have
had a high Al concentration (for samples > 80g), then the dilution factor was increased to 30.
The ICP-AES measurement uncertainties are less than 1%, based on reproducibility and
measurement of solutions of known Al concentrations (Kubik et al., 1998). The 26Al
concentration errors are always > 5%. Therefore, the < 1% ICP-AES error was not included in
the exposure age calculations. Nonetheless a note of caution: For a sample close to 26Al
saturation (which is the case for sample AL9704, Chapter 4.2), the 26Al exposure age is very
sensitive to errors in the determination of the total amount of Al. If furthermore the total
amount of Al is low (e.g. 42ppm for AL9704), a small decrease of about 2%, which is within
the ICP-AES variability at low concentrations, changes the exposure age enormously
32
(AL9704: from 2.90Ma down to 2.46Ma). The
therefore to be taken cautiously.
CHAPTER 2: METHODS
26
Al exposure ages of such samples are
Cation extraction (Precipitation and Fe exchange)
Before the dissolved sample is loaded onto the cation exchange column, interfering cations
should be removed, as they would decrease the efficiency of the following ion exchange.
Ochs and Ivy-Ochs (1997) showed that Be2+ and Al3+, our target elements precipitate at a pH
of 8, whereas other cations like Ca2+, Mg3+, Na+ or K+ are still in solution. A precipitation of
the amorphous Be(OH)2 and Al(OH)3 at pH 8 therefore allows a removal of Ca2+, Mg3+, Na+
and K+. The next step was a specific Fe exchange. This exchange was not performed for all
samples but only for those, where the solution had a bright yellow color, which showed the
presence of Fe. Fe can either be removed with an anion exchange column or with MIBK
extraction (methyl isobutyl ketone), which was described in Ivy-Ochs (1996) and Knauer
(1994). The latter treatment method was used for our samples.
Cation exchange
The processing with cation exchange columns followed closely Ivy-Ochs (1996). Two
different column volumes with 14ml and 50ml of Biorad analytical grade AG50W X8 cation
exchange resin, respectively, were used. The cation exchange was sometimes repeated for a
second or even third time, if a larger amount of quartz was used (> 20g). The large column
(50ml resin) was only used for the samples from Salpausselkä I (Tschudi et al., 2000) and
from Wrangel Island (Karhu et al., 2000).
Precipitation
The cation exchange divided the sample into two separates, a Be solution and an Al solution,
respectively. The finishing steps, the precipitation and the oxidation, were similar for both
fractions. After the sample was dried again, precipitation of Be(OH)2 and Al(OH)3 at pH 8
was performed. The amount of precipitate in the tube was a measure for the quality and purity
of the cation exchange, because the maximum amount of Be(OH)2 was known. This was
derived from a simple experiment: the corresponding amount of carrier, directly precipitated
at pH 8, yielded the maximum volume of Be(OH)2. If more precipitate was present, then the
Be(OH)2 was not pure enough and it may have contained Al(OH)3. Then, the cation exchange
had to be repeated.
Oxidation and pressing
If the precipitation yielded a satisfying result, the hydroxides had to be transferred into a
suitable chemical form for the AMS measurement. This means that water had to be removed
CHAPTER 2: METHODS
33
by oxidation of Be(OH)2 and Al(OH)3 to BeO and Al2O3, which was achieved in a muffle
furnace at 850° for 2 hours (Ivy-Ochs, 1996). The procedure for pressing the oxides into the
AMS targets followed Ivy-Ochs (1996).
Encountered problems during the processing
A common problem was the presence of non- (or not readily) soluble solids at the end of the
HF digest. These powders, mostly whitish, seldom pinkish, were encountered with some of
the samples of Salpausselkä I, Wrangel Island and Tibet (there, large quantities of quartz
> 80g were used). The powder was probably a mixture of fluoride and relict minerals (e.g.
zircon and accessories). We suspected that this solid could have acted as a sink for Al and Be.
This indeed could have affected the Al measurement, if Al was removed by precipitation into
the powder before the Al aliquot was taken. We therefore performed a qualitative analysis
with the method of Rutherford Backscattering Spectrometry (for basic principles of RBS see
Chu et al., 1978) to determine the atomic composition of the unknown solid. First preliminary
results are shown in the appendix of Chapters 3.1 and 3.2.
AMS measurement
In general, the Zurich AMS facility still corresponds to the description of Synal et al. (1997).
Updated parameters are shown in Table 2.4. Standards are used according to Ivy-Ochs
(1996). For further understanding of the AMS method and the physics “behind” it, I
recommend the reading of Finkel and Suter (1993).
Table 2.4: Zurich AMS characteristics for 10Be and 26Al
10
Ion source
AMS
Be
26
Al
•
Extracted ions
BeO-
Al-
•
Current
3µA
200nA
•
Interference
10
B (mainly from dust and
-
labware (Brown et al., 1992))
•
Precision
3-5%
-14
3-10%
•
Chemistry blank ratio
2x10
1x10-14
•
System background ratio
< 10-14
< 10-14
CHAPTER 3: NORTHERN HEMISPHERE
35
Chapter 3
Northern Hemisphere
3.1
10
Be Dating of Younger Dryas Salpausselkä I
Formation in Finland1
Abstract
Boulders of the Younger Dryas Salpausselkä I (Ss I) formation west of Lahti, southern
Finland, were sampled for surface exposure dating. The 10Be concentrations, determined by
accelerator mass spectrometry, yield minimum exposure ages of 11’930 ± 950, 11’220 ± 890,
11’050 ± 910 and 11’540 ± 990 years, using recently published production rates scaled for
latitude and elevation. This includes a correction to the production rate resulting from postglacial uplift of the Fennoscandian lithosphere (i.e. changing elevation) during the time of
exposure. The error-weighted mean exposure age of 11’420 ± 470 years of the analyzed
boulders agrees with previous varve dates of Ss I, which range from 11’680 to 11’430
calendar years BP. However, erosion has to be taken into account as a process affecting rock
surfaces and therefore influencing exposure ages. Available information suggests an erosion
rate of 5 mm/kyr, which increases the error-weighted mean exposure age to a value of 11’610
1
BOREAS, 29, pp. 287 - 293
Silvio Tschudi, Institute of Geology, University of Berne, c/o Institute of Particle Physics, ETH Zürich, 8093
Zürich, Switzerland
Susan Ivy-Ochs, Institute of Geology, University of Berne, c/o Institute of Particle Physics, ETH Zürich, 8093
Christian Schlüchter, Institute of Geology, University of Berne, 3012 Berne, Switzerland
Peter W. Kubik, Paul Scherrer Institut, c/o Institute of Particle Physics, ETH Zürich, 8093 Zürich, Switzerland
Heikki Rainio, Geological Survey of Finland, 02151 Espoo, Finland
Corresponding author: Silvio Tschudi, [email protected]
36
± 470 years. Within the errors, the formation of Ss I in the Vesala area west of Lahti falls into
the Younger Dryas time bracket, as defined by the GRIP and GISP 2 ice core (Greenland).
Introduction
The Younger Dryas (YD) margin of the Scandinavian Ice Sheet is one of the most impressive
glacial formations on Earth (Donner, 1978; Hyvärinen, 1975). In Finland, studies on these
formations have a long scientific tradition (Rainio, 1996 and references therein) providing key
information for the understanding of geologic processes during the last glaciation and its
termination in northern Europe. The Finnish Salpausselkä (Ss) ice-marginal formations, two
sub-parallel main ridges Ss I and Ss II, are part of this geomorphic feature, which can be
mapped over a length of about 600km (Fig. 3.1). In the east, Ss II and probably also Ss I find
their continuation in the Koitere end moraine (Rainio, 1995). The deposition of the Russian
Rugozero ice-marginal formation, a physical continuation of the Koitere end moraine, has
been suggested to have been contemporaneous with Ss I (Ekman and Ilyin, 1991; Lukashov
and Ekman, 1982 and references therein). In the west, the traces of the Ss are lost in the Gulf
of Finland.
Fig. 3.1: Sketch map of the sampling area. The samples were taken west of Lahti at the apex of the
bend of Ss I (Map modified after Okko, 1962). The inlet gives a general overview of the Scandinavian
Ice Sheet during the YD (modified after Björck et al., 1996).
37
The Ss complex comprises glacial, glacio-fluvial and glacio-lacustrine sediments (e.g.
Glückert, 1977, 1995; Okko, 1962; Rainio et al., 1995). Studies of lake level records of the
Baltic Ice Lake (BIL) along Ss I and Ss II allowed Donner (1969, 1982) to conclude that Ss I
defined the shoreline of the BIL during level B I, the highest level of this freshwater lake,
which was dammed by ice at Billingen, Sweden. With the retreat of the Scandinavian Ice
Sheet, the lake level dropped from B I to the levels B II and B III, and finally to the level of
the Yoldia Sea (e.g. Bodén et al., 1997). These level changes are closely related to the
processes of formation of Ss I and Ss II. Okko (1962) proposed the Heinola deglaciation
followed by the Salpausselkä readvance, with the deposition of Ss I as the terminal formation
(e.g. Rainio, 1996 and references therein).
The timing of these processes, and thus the age of the Ss formations, has always been an
important question. In southern Finland, first estimates were given by Sauramo (1918, 1923
and 1929) using the technique of varve chronology construction. Varved lake deposits are
correlated with the help of stratigraphic data, for example marker beds or grain size analysis;
or macrofossil and pollen analysis. A relative chronology for the deposition of Ss I can be
derived, if varve analysis is performed on sediment from lakes located inside and outside of
Ss I, since the composition of the lake sediments is dependent on the relative position to the
ice. Together with marker beds (e.g. the drainage varve of the BIL-level B III to the Yoldialevel), deposition of Ss I can be tied to the floating varve chronology. During the 20th century,
the Finnish varve chronology was revised several times (e.g. Niemelä, 1971; Sauramo, 1958)
and also linked to the Swedish varve chronology (e.g. Cato, 1987; Strömberg, 1990), but the
dating of Ss I and Ss II is still indirect and relative. The currently accepted time-frame for the
deposition of Ss I in the area of Lahti, based on the varve chronology of Niemelä (1971) and
the correlation with the Swedish system by Cato (1987) and Strömberg (1990), ranges
between 11’680 and 11’430 years BP. Recent published data indicate that 300 - 1’000 years
of varve layers may be missing during the Holocene in the Swedish varve chronology (e.g.
Björck et al., 1996; Wohlfarth et al., 1997). This would shift the Swedish varve chronology to
older ages, and as the Finnish varve chronology is directly linked to the Swedish one, it would
also affect the estimated age of the deposition of Ss I. An error of 300 - 1’000 years would
therefore shift the time-frame of the deposition to values between 12’700 and 11’700 calendar
years BP.
As far as radiocarbon data are concerned, very few data are available. Radiocarbon dated lateglacial and early post-glacial pollen stratigraphy at lake Varrassuo, a peat-bog located west of
Lahti (Fig. 3.1), just marks the beginning of vegetation after the retreat of the ice from the Ss I
position at about 8’810 ± 240 radiocarbon years BP (Donner, 1966). However, studies on the
vegetational history supported by pollen analysis (Hyvärinen, 1972, 1973) and the occurrence
38
of periglacial features related to Ss I (Aartolahti, 1970) let Donner (1978) conclude that Ss I
and Ss II were formed during the Younger Dryas chronozone. This conclusion is mainly
based on a change from an Artemisia to a Birch pollen assemblage, observed in sediment
cores taken outside of the Ss formations. This transition, from open and pioneer vegetation to
a birch forest was interpreted as a change from the cold YD period to warmer Preboreal
period (Donner, 1978 and references therein).
If we follow Okko (1962) and Rainio (1996), the Salpausselkä readvance and its termination
Ss I can be identified as the readvance of the Scandinavian Ice Sheet during the Younger
Dryas chronozone. The identification is, however, still based on indirect dating and
correlation, and absolute data are missing. This lack can be filled by the technique of
exposure dating using cosmogenic nuclides (e.g. 10Be, 21Ne, 26Al and 36Cl), performed on
rock surfaces (e.g. Lal, 1991). This method has been successfully used to date Quaternary
landforms. Analysis of bedrock surfaces yields information about the timing and extent of
glaciations (e.g. Brook et al., 1996; Bierman et al., 1999), whereas boulders, deposited on
moraines, can be used to date glacial events (e.g. Gosse et al., 1995; Ivy-Ochs et al., 1999;
Phillips et al., 1997).
To obtain preliminary absolute age information for this important YD ice-marginal formation,
ten boulders of Ss I were sampled during summer 1997 for 10Be exposure dating. The set of
samples is divided into two subsets, yielding one set of six samples along Ss I between
Kouvola and Lappeenranta and another set of four samples taken west of Lahti at the bend of
Ss I. The latter was chosen as a test series and its results will be presented here.
Sample Site
The samples of the western subset were taken in the Vesala section of Ss I, as defined by
Okko (1962), north of the highway Tampere-Lahti (Fig. 3.1). This section occupies the apex
of the bend of Ss I, which represents a transitional area between two re-advancing ice lobes,
one flowing from NW to SE, the other flowing from N to S (Punkari, 1980). Moraine terrain
occurs in the northwestern part of this section with WSW to ENE trending chains of moraine
ridges, which are likely to have been deposited by the northwestern glacier (Okko, 1962). The
four samples west of Lahti were taken out of two different areas with two samples for each
locality. Boulders thought to be affected by geomorphic processes, like rolling, sliding or
spalling, were avoided, since these processes could change the initial position of the top
surface (i.e. decrease the exposure time of the sample taken). We have chosen boulders
greater than several m3 in volume, located in stable positions on moraine ridges (Fig. 3.2). All
samples were taken by removing the uppermost centimeters of the top surface of boulders.
39
Fig. 3.2: View of a boulder at locality 1, N61.00° E025.39°. Vegetation is limited to light boreal forest,
but some moss growing on the boulders may occur. During wintertime, snow and ice covers the
surfaces (up to 30cm). The hammer on the boulder is 50cm long and the size of the boulder is roughly
10m3.
The sampled granitic boulders Sal 1 and Sal 3 from locality 1 (N 61.00° E 025.39°) lie on a
boulder-rich distal ridge with a height of about 5m, which represented the BIL shoreline
during deposition. Here the deposit is clast-supported, with fine material completely lacking.
All sampled boulders lie above the ancient B I shoreline and there is no evidence for either
periglacial disturbance or large-scale littoral erosion, although finer material might have been
washed out by wave erosion. Locality 2 lies just south of the proximal slope of Ss I (N 61.02°
E 025.40°). Both samples, Sal 4b and Sal 5, perch on a WSW to ENE trending moraine ridge.
Sal 4b was taken from quartz veins of a flat-lying granitic boulder, whereas sample Sal 5,
located a few tens of meters away, was taken from a huge granitic boulder.
Methodology
General
As a result of secondary cosmic-ray bombardment, 10Be and other nuclides (e.g. 21Ne, 26Al or
36
Cl) are in situ produced significantly only within the uppermost decimeters of a rock (e.g.
Lal, 1991). The measured concentration of the radioactive isotope 10Be can be used to
calculate the exposure age of a sample. Our calculations are based on the production rate
depth dependency of Masarik and Reedy (1995), as this model was also used for the
determination of the 10Be surface production rate by Kubik et al. (1998).
40
Quartz is an ideal target mineral to determine the concentration of in situ 10Be. The amount of
quartz needed for samples near sea level and of exposure ages of about 12’000 years (as valid
for Ss I) was estimated to be more then 160g. This ensured that the effect of chemistry blank
corrections on the measured 10Be/9Be ratios were small, and that counting statistics during the
AMS measurement were good. The purification of quartz out of the granitic rock samples and
the chemical separation of 10Be was done according to the method described in Nishiizumi et
al. (1989), Kohl and Nishiizumi (1992) and Ivy-Ochs (1996). The 10Be concentrations were
measured by accelerator mass spectrometry (AMS) at the Zurich tandem accelerator facility
of the Paul Scherrer Institut and ETH.
Age calculation
To calculate exposure ages, we used the recently published 10Be surface production rate of
Kubik et al. (1998). In the time range near the YD, this production rate is the only one where
the calibration event has been independently constrained (dendrocalibrated radiocarbon age).
For the Salpausselkä sample locality at a geographic latitude of 61° and a present altitude of
160m above sea level, the production rate of 5.75 ± 0.24atoms/yr·g SiO 2, valid for sea level
and latitudes = 60°, was scaled according to Lal (1991). We have assumed that the
geomagnetic latitude was roughly the same as the geographic latitude when averaged over the
last 10’000 years (Ohno and Hamano, 1992 and 1993; Sternberg, 1996). This production rate
was corrected for the sample thickness by using the calculated flat depth profile for the first
12g/cm2 below the surface (Masarik and Reedy, 1995). No corrections were necessary for the
sample geometry, since all samples were flat lying surfaces. Geometric shielding of
surrounding hills could be excluded, since the Salpausselkä formation itself defines the
highest elevation of the adjacent area.
Deglaciation of the Scandinavian Ice Sheet caused rapid and intensive uplift of the
Fennoscandian lithosphere due to isostatic rebound. According to Mörner (1980), this caused
an absolute uplift of the surface of about 250m during the past 11’000 years at the region of
Lahti. Taking sea level changes into account, this results in a present altitude of 160m above
sea level. Thus, the elevation of the sampled boulders and therefore also the production rate of
10
Be was changing from the beginning of their exposure until present time. Estimated uplift
values from Mörner (1980) with an estimated error of 20%, which results in an uncertainty
for the integrated production rate of 1%, yield an integrated production rate for 10Be of
6.31 ± 0.27atoms/yr·g SiO 2 for the sample site. This value is more than 5% lower than the
value corresponding to the present elevation. We used this lower production rate to calculate
exposure ages of Ss I, because the encountered uplift is strongly related to the deglaciation of
the Scandinavian Ice Sheet. It is well defined and has been studied by independent methods.
41
Shielding of the samples by snow coverage (about 30 ± 30cm during 4 months, personal
communication, Lahti Tourist Information) has been considered and estimated. The large
uncertainty reflects that the given snow cover is today’s value and might not be a
representative value for the last 10ka. Using the flat profile of Masarik and Reedy (1995),
snow yields corrections of the production rate of less than 1%. Present vegetation (light boreal
forest and moss growing on the boulders, Fig. 3.2) affects the production rate in the same
order (< 1%), but it may not accurately reflect the past coverage with biomass. To avoid an
overestimation of shielding by vegetation, no correction was performed.
As a summary, the following sources of uncertainties were considered for the age calculation:
AMS 1σ measurement error (Table 1), uncertainties for the chemical processing and
reproducibility in the laboratory (5%, see Ivy-Ochs, 1996), the production rate (4.2%, Kubik
et al., 1998) and the uplift rates (20%, which results in an uncertainty for the integrated
production rate of 1%) and uncertainties in the determination of the sample thickness
(d = ± 0.5cm) and the snow coverage (snow = 30 ± 30cm). For the presented data, the total
uncertainty was determined according to accepted error propagation calculations. The mean
value was calculated as an error-weighted mean value with a 1σ error of the mean.
Results and conclusion
The 10Be exposure ages of the sampled boulders range between 11’050 ± 910 and 11’930 ±
950 years (Table 3.1), where all cited sources of uncertainties are included. The overall
minimum exposure age of the Ss I formation west of Lahti, calculated as an error-weighted
mean age of all four samples, is 11’420 ± 470 years (Fig. 3.3). This age agrees with the
present varve dating of Ss I, which defines a deposition between 11’680 and 11’430 calendar
years BP (Cato, 1987; Niemelä, 1971; Strömberg, 1990).
Taking erosion into account would increase the exposure ages. If it would be possible to
determine the erosion rate of a particular sample, a close-to-real exposure age could be
calculated. Unfortunately, there is almost no precise information about erosion rates on rock
surfaces available. Saunders and Young (1983) review reported values. In their work,
weathering rates for rhyodacite determined on rock tablets in Colorado, range from 4 to
5mm/kyr. Values for erosional and chemical denudation of a batholith, determined with river
loads, are given in Clayton and Megahan (1986). They report values between 8 to 9mm/kyr.
Although rhyodacite is volcanic and its porosity is likely to be higher than the one of the
sampled granite, we are interested in the actual rate of rock surface weathering rather than
rates of landscape denudation. Therefore, we have used 5mm/kyr of Saunders and Young,
(1983) as an estimate.
42
Table 3.1: AMS measured 10Be concentrations with calculated minimum exposure ages of Ss I west of
Lahti. We used a production rate of P = 5.75 ± 0.24atoms/yr·g SiO2 for 10Be at latitudes = 60° and sea
level (Kubik et al., 1998), scaled for elevation and latitude according to Lal (1991). Within these
corrections, uplift due to isostatic rebound was considered. For the calculations, a rock density of ρ =
2.7g/cm3 and an attenuation length of cosmic rays Λ = 157g/cm2 (Masarik and Reedy, 1995) were
used. 10Be concentrations are chemistry blank corrected. The age uncertainties are explained in the
text.
Sample Lithology Size of Height Thick- Sample
boulder above
ness
size,
3
[m ]
ground
[cm] Qtz [g]
level [m]
Sal 1
Sal 3
Sal 4b
Sal 5
Granite
Granite
Qtz vein
Granite
10
2
6
14
3
0.5
1.5
3.5
4
3
2
2
207.11
210.71
165.52
160.76
10
Present
Be
Error [%] Minimum age
altitude Concentration
[years]
[m above
[Atoms/g
sea level]
SiO2]
160
7.50x104
6.7
11’930 ± 950
160
7.07x10
4
6.8
11’220 ± 890
6.95x10
4
7.0
11’050 ± 910
7.26x10
4
7.4
11’540 ± 990
160
160
Fig. 3.3: Exposure ages of the Ss I samples compared with previous varve dating (Niemelä, 1971;
Cato, 1987; Strömberg, 1990) and the Younger Dryas (YD) time interval defined by the GISP 2 (Alley
et al., 1993) and GRIP ice core (Johnsen et al., 1992). The two sets of exposure ages were calculated
with different erosion rates of 0mm/kyr and 5mm/kyr, respectively, with (×) for data points Sal 1, 3, 4b
and 5 and (s) for the error-weighted mean. The included uncertainties are explained in the text. The
black bars give the time range of the varve and ice core dating, while the white ones give the
uncertainties.
43
Exposure ages using this erosion rate were calculated based on the flat profile of Masarik and
Reedy (1995). The ages then increase to 12’250 ± 980, 11’430 ± 910, 11’180 ± 920 and
11’690 ± 1’000 years, with an error-weighted mean age of 11’610 ± 470 years. Again, this
result is consistent with the present Finnish varve chronology. Within the errors, the
formation of Ss I in the Vesala area west of Lahti also falls into the Younger Dryas time
bracket, as defined by the GRIP and GISP 2 ice core (Greenland). There, two different
investigations based on layer counting and isotope analysis are available. In the GISP 2 ice
core, the YD lasted from 12’940 ± 550 to 11’640 ± 250 calendar years BP (Alley et al., 1993),
whereas in the GRIP ice core it was from 12’700 ± 100 to 11’550 ± 70 calendar years BP
(Johnsen et al., 1992).
However, exposure ages are minimum ages and they are dependent on many factors. We have
only considered uplift, coverage by snow and an estimated erosion rate of 5mm/kyr. Other
limiting factors, such as coverage by soil or intensive growth of vegetation during the
exposure time, might have occurred and the calculated exposure age would then be too
young. One could also argue that the oldest calculated age of our sample set is more likely to
represent the ”real” minimum exposure age of the sampled formation. Then, Ss I would have
been deposited at 11’930 ± 950 years ago (without consideration of erosion) or 12’250 ± 980
years ago (with an assumed erosion rate of 5mm/kyr).
Acknowledgements
This work was funded by Swiss National Science Foundation grants 21-043469.95/1. We
thank all of the Zurich tandem crew for maintaining the good performance of the AMS
facility.
44
Comments and appendix to Chapter 3.1
General remarks
The main difficulty with the samples from the Younger Dryas formation Salpausselkä I was
the low location altitude and the short duration of the exposure of the boulders to cosmic rays.
Initially, more than 200g of quartz were used to determine the exposure ages. Parallel
processing of the raw rock made it possible to obtain the pure quartz within useful time. After
reducing the 9Be carrier addition from 0.4mg to 0.3mg, the amount of quartz needed could be
decreased by 20% form 200g to 160g.
RBS measurements of insoluble solids
During the HF digest of the large Salpausselkä samples (> 160g quartz), insoluble whitish and
pinkish material with unknown composition remained in the Teflon beaker. Quartz is never
100% chemically pure and zircon or other minerals (e.g. titanite, apatite) may occur and
partly resist the chemical purification with HF. Therefore, certain compounds with Zr, Ti and
others can be expected in the solid. The solid could also contain Al from the sample, which
would affect the determination of the total Al content, because it is determined after the HF
digest. Incomplete fuming (e.g. incomplete removing of Fluorides after the HF digest) could
be another source for the observed solid, as they are not well soluble.
To determine the qualitative atomic composition of the unknown solid, we performed RBS
analyses (see Chu et al., 1978 for details) on the solids of samples Sal 3, Sal 4b and Sal 5. A
small quantity of the solid was suspended in water (about 0.1ml in total) and put on a pure
substrate of graphite. It was dried on a hot plate, so that a thin layer of solid remained. The
samples were then measured with RBS at the accelerator facility of Zurich.
The results of the RBS are presented in Fig. 3.4, Fig. 3.5 and Fig. 3.6. As expected, the solids
reflect the atomic composition (Table 3.2) of common accessory minerals (e.g. apatite Ca5[(F,
OH, Cl) | (PO4)3], titanite CaTi[O | SiO 4], barite BaSO4 or zircon Zr[SiO 4]). Al is not
observed in sample Sal 3, but is abundant in the samples Sal 4b and Sal 5. This Al could
originate from the dissolved quartz. The determination of the total Al concentration (by ICPAES) and the subsequent calculation of the 26Al exposure age would then be flawed.
Fig. 3.4: RBS energy spectrum of Sal 3 with normalized yield. The black line, overlying the measured
spectrum, is the calculated best-fit elemental simulation to this spectrum. The unknown solid consists
mainly of O, P and Ti. Minor amounts of Zr and Ba are present. Al is not observed in significant
quantities (Table 3.2).
Fig. 3.5: RBS energy spectrum of Sal 4b. The black line, overlying the spectrum, is the calculated bestfit elemental simulation to this spectrum. The unknown solid consists mainly of O, Al and Si. Minor
amounts of Ti and Fe are present (Table 3.2).
45
46
Fig. 3.6: RBS energy spectrum of Sal 5. The black line, overlying the spectrum, is the calculated bestfit elemental simulation to this spectrum. The unknown solid consists mainly of O, Al and Si. Minor
amounts of Ti and Zr are present (Table 3.2).
26
Al ages, calculated with the measured ICP-AES Al concentrations, agree with the 10Be ages
(Tschudi et al., 2000 and Table 3.3), which means that the Al in these solids does not come
from the quartz fraction, but from remaining and undissolved minerals. These minerals have
to be insoluble in acids (i.e. HF) and rich in Al. One promising candidate is corundum
(Al2O3). But then, the occurrence of Si is not explained. Alumosilicate (i.e. staurolite, kyanite
and andalusite) would deliver Al, Si and O, but they are soluble in HF. For hard evidence, the
solid will have to be analyzed and identified with other methods before any final conclusion
can be drawn.
Table 3.2: Quantitative estimate of the elemental abundance in the unknown solids of sample Sal 3, Sal
4b and Sal 5.
Sal 3
Element
Sal 4b
Atom percent
Element
[%]
Sal 5
Atom percent
Element
[%]
Atom percent
[%]
O
68.8
O
64.0
O
68.2
P
17.7
Al
21.9
Si
21.0
Ti
11.4
Si
12.8
Al
10.0
Zr
0.8
Fe
0.5
Zr
0.6
Ba
0.4
Ti
0.5
Cu
0.3
others
0.9
Cl
0.4
U
0.01
47
26
Al exposure ages
Exposure ages based on 26Al AMS and Al ICP-AES measurements are presented in Table 3.3.
The ages confirm the 10Be results of Tschudi et al. (2000) within the errors.
Table 3.3: AMS measured 26Al concentrations of the Salpausselkä I samples of Tschudi et al. (2000).
We used the production rate of Kubik et al. (1998), scaled according to Lal (1991) and corrected for
isostatic uplift. For the calculations, an attenuation length of cosmic rays Λ = 157g/cm2 (Masarik and
Reedy 1995) were used. 26Al concentrations are chemistry blank corrected. The error includes 1σ
measurement uncertainties and a 5% variability for the chemical processing (Ivy-Ochs, 1996). The age
uncertainties are explained in the text (Chapter 3.1).
26
Al
Error [%] Minimum 26Al exposure Minimum 10Be exposure
Concentration
age
age (Tschudi et al., 2000)
[Atoms/g
[years]
[years]
SiO2]
Sample
Total Al
content
[ppm]
Sal 1
125
5.54x105
10.9
13’560 ± 1’480
11’930 ± 950
Sal 3
176
4.37x105
17.3
10’690 ± 1’850
11’220 ± 890
Sal 4b
157
4.81x10
5
11.7
11’780 ± 1’380
11’050 ± 910
Sal 5
122
4.71x105
12.0
11’530 ± 1’380
11’540 ± 990
A note of caution should be added, as far as the determination of the total Al content with
ICP-AES is concerned. For the samples Sal1 and Sal3, 500µl of the sample aliquot were
diluted within 5ml of water, which yields a dilution factor of 10. During the ICP-AES
measurement, it turned out that this dilution factor is too small for samples with several
hundreds of grams of quartz. Therefore higher dilution factors were used for the samples
Sal 4b and Sal 5 (150µl of sample aliquot was diluted within 5ml of water, i.e. a dilution
factor of 30 was used instead of 10).
48
3.2 Constraints for the latest glacial advance on Wrangel
Island, Arctic Ocean, from rock surface exposure
dating2
Abstract
During an expedition to Wrangel Island in 1997, three samples of quartz veins were collected
for surface exposure dating. Two were bedrock surface samples from different parts of the
island, one was from a local boulder. The samples were collected where former coverage by
valley glaciers can be excluded. Especially the bedrock surface samples were expected to
provide new constraints for the chronology of glaciations on the island and on the arctic
continental shelf in the Beringian region. Minimum exposure age estimates for these bedrock
samples, calculated on the basis of 10Be concentrations, are 64’600 ± 6’400 and 26’400 ±
2’100 years. These estimates suggest that no major glaciations have affected Wrangel Island
or the adjacent shelf area after 64’600 ± 6’400 years. Specifically, these dates seem to rule out
any major glaciation during the Last Glacial Maximum (LGM). The estimates should
however be regarded as preliminary, as only two bedrock surface samples were analyzed, but
are compatible with the absence of glacial morphology and sediments, and with radiocarbon
dates from mammoth remains.
2
Global and Planetary Changes, in press.
Juha Karhu, Geological Survey of Finland, P.O. BOX 96, FIN-02151 Espoo, Finland
Silvio Tschudi, Institute of Geology, University of Berne, c/o Institute of Particle Physics, ETH Hönggerberg,
8093 Zurich, Switzerland
Matti Saarnisto, Geological Survey of Finland, P.O. BOX 96, FIN-02151 Espoo, Finland
Peter W. Kubik, Paul Scherrer Institut, c/o Institute of Particle Physics, ETH Hönggerberg, 8093 Zurich,
Switzerland
Corresponding author: Juha A. Karhu, [email protected]
49
Introduction
Wrangel Island in the Arctic Ocean is an isolated, exposed part of the continental shelf of
northeastern Russia. Any information about the chronology of glaciations on the island is
useful for understanding past glacial advances and retreats on the arctic continental shelf in
the Beringian region. However, only limited data exist, and, especially, numerically
constrained investigations are few.
Wrangel Island is well known for the abundance of mammoth skeletal remains. A number of
bones, teeth and tusks have been dated by radiocarbon methods at different laboratories and
the ages seem to cover the whole period from about 30’000 to 4’000 years BP (Sulerzhitsky
and Romanenko, 1999; Vartanyan et al., 1995). Some bones and teeth have yielded
radiocarbon ages such as 20’000 ± 110 (LU-2807, Vartanyan et al., 1995), 22’400 ± 200
(GIN-8257, Sulerzhitsky and Romanenko, 1999) and 22’400 ± 300 years BP (GIN-8259,
Sulerzhitsky and Romanenko, 1999), which are coincident with the Last Glacial Maximum
(LGM). These results suggest that during the LGM this part of the continental shelf was icefree, although globally the period is characterized by the maximum ice volume of the last
glacial cycle (Shackleton, 1987). However, different scenarios have also been suggested.
Grosswald (1988) proposed an extensive East Siberian Ice Sheet which, according to him,
covered most of the northeastern Russian shelf during the LGM, including Wrangel Island. A
similar view is also expressed in a recently published compilation of the extent of the late
Pleistocene ice sheets in the north-east of Russia (Grosswald, 1997), implying an ice
thickness exceeding 1000m in the region of Wrangel Island and the adjacent shelf. These
interpretations are clearly in conflict with the information obtained from mammoth remains.
The inconsistency is not restricted to Wrangel Island, as both the New Siberian Islands
(Sulerzhitsky and Romanenko, 1999) and Severnaya Zemlya (Makeyev et al., 1981) have
yielded mammoth bones dated by the radiocarbon method to the LGM.
To obtain new, independent data on the chronology of glaciations in the region, we collected
rock surface samples for surface exposure dating. The rock samples used for the analysis of
cosmogenic nuclides were taken during an expedition to Wrangel Island in the summer of
1997. In this study we report 10Be concentration analyses and discuss their implications. The
description of Quaternary landforms is also based on fieldwork on Wrangel Island in 1997
(Saarnisto et al., 1998).
Geological setting
Wrangel Island is situated 140km north of mainland Chukotka, between the East Siberian Sea
in the west and the Chukchi Sea in the east, at a distance of about 700km from the Bering
50
Strait. The island consists of a southern and a northern coastal plain and an east-west trending
mountainous region in the middle. Mountains generally have a maximum elevation of about
500m, with the highest point reaching 1’095m on Mount Sovetskaya. Wrangel Island has a
core of Precambrian basement, outcropping in the central mountains. Elsewhere the basement
is covered by Paleozoic and Mesozoic strata, metamorphosed during the Chukotkan Orogeny
at the end of the Mesozoic Era. Tertiary deposits, a few meters thick, cover much of the
northern part of the island (Kos'ko et al., 1993). Glacigenic depositional landforms cannot be
recognized on the island. The landscape is strikingly different compared to the mainland of
Chukotka, where a glaciated terrain with eskers and deltas in the valleys contrasts with
unglaciated, more elevated areas with v-shaped valleys and boulder fans. On Wrangel Island
moraines, eskers etc. are absent, as well as striae or striated clasts. Similarly, no diamicton,
which could be interpreted as glacial sediment, was found in elevated terrain. Elongated
bedrock forms are structural features of sedimentary rocks, or formed by periglacial processes
which are very active. Thoroughly broken bedrock surfaces, boulder fields, solifluction banks
and -stripes and landslide fans are common. However, u-shaped valleys at lower elevations
and provenance studies of diamictons, interpreted as glacial till, suggest that at least the major
valleys have been glaciated. Extensive outwash plains in river valleys contain abundant, fartraveled and completely rounded boulders, which may indicate long fluvial transport times
and more active fluvial systems, i.e. higher precipitation, or perhaps glaciofluvial transport.
For the Late Weichselian, however, field evidence seems to rule out extensive glaciation,
although local valley glaciers cannot be excluded.
Samples
Samples Wra 1 (71° 9.50’ N, 178° 50.86’ E, 20m.asl) and Wra 2 (71° 9.41’ N, 178° 51.11’ E,
20 m.asl) were collected from massive, several meters wide quartz veins in Permian siltstone
at Ptichii Bazar on the northwestern coast of the island (Fig. 1). These quartz veins are
exposed on small hills at a distance of about 10m (Wra 1) and 70m (Wra 2) from a wave-cut
cliff at the shoreline. Sample Wra 1 represents an outcrop, whereas Wra 2 was collected from
a large, 70x40cm boulder in a local boulder field of quartz-vein-bearing boulders. The local
character of the boulder field is clearly demonstrated by a sharp contrast in color between the
field of white quartz-vein boulders surrounded by boulders consisting entirely of dark gray
siltstone.
Sample Wra 4 (71° 00’ N, 178° 29’ W, 160m.asl) was collected from a 2cm thick quartz vein
in an outcrop of Triassic slate on the southern coast of the island (Fig. 3.7). The outcrop is
located on a small hilltop about 2km north of the village Ushakovsky and separated from
higher mountains in the north by the valley of the Nasha River.
51
Fig. 3.7: The Bering Strait region and a sketch map of Wrangel Island (W). Samples Wra 1 and Wra 2
were taken at the coastline in the northwestern part of the island, whereas Wra 4 is from the top of a
hill between the village Ushakovsky and the Nasha River valley.
Methods
General
As a result of cosmic-ray bombardment, 10Be and other nuclides (e.g. 3He, 21Ne, 26Al, 36Cl)
are in situ produced within the uppermost decimeters of a rock surface (Cerling and Craig,
1994; Lal, 1991). The concentration of these nuclides in a rock surface is time dependent. It is
therefore a measure for the time interval of exposure to cosmic rays. If we assume that the
exposure of a sampled rock surface is related to a geologic event (e.g. glaciation), then surface
exposure dating directly yields an age estimate for this event.
Quartz is an ideal target mineral to determine the concentration of in situ 10Be. Its purification
out of the rock samples and the chemical separation of 10Be were done according to the
method described in Ivy-Ochs (1996); Kohl and Nishiizumi (1992) and Nishiizumi et al.
(1989). The 10Be concentrations were measured by accelerator mass spectrometry (AMS) at
the Zurich tandem accelerator facility of the Paul Scherrer Institut and the ETH.
52
Age calculation
Our calculation of exposure ages follows the formalism of Lal (1991). For the production
rates of in situ 10Be, we have chosen the recently published value of Kubik et al. (1998),
which is dendrocalibrated with 14C for the last 10ka (5.75 ± 0.24atoms/yr·g SiO 2 for
geomagnetic latitudes = 60°). All our calculations are based on the production rate depth
dependency of Masarik and Reedy (1995), as this model was also used for the determination
of the 10Be surface production rate by Kubik et al. (1998). This model suggests a “flat” depth
dependency of the production near the surface, instead of a straight exponential decrease of
production with depth. The production rate was scaled for geographic latitude and present
altitude of the samples (N 71°, 20m and 160m.asl, respectively) according to Lal (1991) and
corrected for the sample thickness. No corrections were necessary for the sample geometry,
since all samples were from flat-lying surfaces. Geometric shielding of surrounding hills, as
well as Quaternary uplift of Wrangel Island due to tectonics or isostatic rebound (e.g.
McManus and Creager, 1984) influences the production rate in the order of < 1% and are
therefore neglected here. Shielding by vegetation and snow was estimated for the examined
samples. Present vegetation does not have an effect on the ages, as it is limited to local grass,
not covering the samples. Due to climatic conditions, past vegetation is assumed to be equal
or even less intensive than today’s. Snow coverage, estimated to about 5cm during 8 months,
yields a decrease of less than 1% for the production rate. The influence of vegetation and
snow was therefore neglected. The calculated ages are generally expressed as minimum ages,
since erosion is unknown and therefore set to the value of zero. If erosion is taken into
account, the exposure ages increase according to the used model of in situ production of
cosmogenic nuclides within the rock surface. In our case however, erosion does not
significantly change the exposure ages, as the “flat” model from Masarik and Reedy (1995)
minimizes the effect of erosion (Tschudi et al., 2000). We therefore just present minimum
surface exposure ages.
Results and discussion
The quartz vein samples analyzed from the Ptichii Bazar area comprise two types of rock
surfaces and the minimum exposure age estimates seem to vary accordingly. Sample Wra 1
represents a bedrock surface, and it has a minimum exposure age of 64’600 ± 6’400 years. In
contrast, Wra 2 is a local boulder, and it yields a much younger minimum exposure age of
12’400 ± 1’000 years (Table 3.4).
The minimum exposure age estimate for the bedrock surface suggests no major glacial
advances in this area later than 64’600 ± 6’400 years. This age, however, could also reflect a
complex exposure history with repeated periods of exposure and shielding during the past.
Bedrock samples reflect the deglaciation time of the last glacial event only if the advancing
53
glacier eroded enough material from the bedrock surface to reset the cosmogenic
“chronometer”. No statement can be drawn about this erosional activity from our data. Yet,
recent investigations suggest that even cold-based glaciers abrade their beds and entrain basal
material (Cuffey et al., 2000). Therefore, the survival of cosmogenic nuclides within the
uppermost decimeters of the rock surface over a glacial interval may be rather unlikely. The
relatively young minimum exposure age estimate for Wra 2 could indicate that the sampled
boulder has suffered disintegration and possibly rotation due to frost action. Because the age
is a strict minimum estimate, no definitive age can be given for the disintegration event.
Sample Wra 4, collected from the bedrock surface close to the southern coastline of the
island, has a minimum exposure age estimate of 26’400 ± 2’100 years. Field evidence
excludes a former coverage by valley glaciers. Thus, any major advances of glacial ice sheets
should have occurred before this date. These data are consistent with sample Wra 1, as
exposure ages are always minimum ages calculated with zero erosion, when no independent
information exists on the magnitude of erosion.
Table 3.4: AMS measured 10Be concentrations with calculated minimum exposure ages for two
bedrock surfaces and a boulder at Wrangel Island. We used a production rate of
pBe=5.75±0.24atoms/yr·g SiO2 for 10Be at latitudes = 60° and sea level (Kubik et al., 1998). Corrections
for elevation and latitude were done according to Lal (1991). The age uncertainties include the
measurement and processing error (1σ AMS measurement error and a 5% error of the chemical
processing) and a 4.2% error of the production rate. A corresponding blank was subtracted.
Sample
Lithology
Wra 1
Qtz vein in
Permian siltstone;
Bedrock
Qtz vein in
Wra 2
10
Thickness
[cm]
Present
altitude
[m.asl]
Be
Concentration
[Atoms/g SiO2]
Error
[%]
Age
[years]
10
20
3.57x105
8.8
64’600 ± 6’400
7
20
7.14x104
7.3
12’400 ± 1’000
5
160
1.77x105
6.5
26’400 ± 2’100
Permian siltstone;
Boulder
Wra 4
Qtz vein in
Triassic slate;
Bedrock
Only three samples from Wrangel Island have been analyzed for the concentration of
cosmogenic 10Be, and the results should thus be regarded as very preliminary. However, one
bedrock surface sample from the northwestern coast and another from the southern coastal
area yielded pre-LGM minimum exposure ages. They are not in conflict with the radiocarbon
data from mammoth skeletal remains and also compare favorably with the absence of
glacigenic landforms and sediments. These minimum exposure age estimates indicate that,
excluding local valley glaciers, no major growth of glacial ice occurred on Wrangel Island or
54
on the adjacent continental shelf during the Late Weichselian (Sartan) glaciation and possibly
even during the Middle Weichselian. This is consistent with recent studies on the northeastern
Russian mainland (Hopkins et al., 1998), where surface exposure dating demonstrates that
only local valley glacier complexes existed during the LGM.
Acknowledgements
This work was partly funded by Swiss National Science Foundation grants 21-043469.95/1.
We thank all of the Zurich tandem crew for keeping the machine running. A special thanks
goes to Susan Ivy-Ochs, who was always there when we encountered problems with the
sample preparation. The expedition to Wrangel Island in 1997 by J.K. and M.S. was funded
by the Geological Survey of Finland and the Academy of Finland. We thank J. Chlachula,
P.M. Grootes, C. Hjort and an anonymous reviewer for useful comments.
26
Al exposure ages
The interpretation of the 10Be data from Karhu et al. (2000) is not unambiguous. Prior
exposure and subsequent shielding by advancing ice would not be visible within a singlenuclide-study. The analysis of 26Al can provide new information about the exposure history of
the Wrangel Island samples. Agreeing 10Be and 26Al ages would indicate a simple exposure
history without periods of coverage. Exposure ages based on AMS 26Al measurements are
presented in Table 3.5.
Table 3.5: AMS measured 26Al concentrations of the Wrangel Island samples of Karhu et al. (2000).
We used the production rate of Kubik et al., (1998), scaled according to Lal (1991). 26Al
concentrations are chemistry blank corrected. The error includes the 1σ AMS measurement error and a
5% error of the chemical processing (Ivy-Ochs, 1996). For the age calculation an additional 5.1% error
of the production rate (Kubik et al., 1998) was included.
26
Al Concentration Error Minimum 26Al exposure age Minimum 10Be exposure age
[Atoms/g SiO2]
[%]
[years]
(Karhu et al., 2000)
[years]
Sample
Total Al
[ppm]
Wra 1
25
5.06x104
15.8
1’400 ± 200
64’600 ± 6’400
Wra 2
33
4.25x10
5
7.5
11’500 ± 1’000
12’400 ± 1’000
Wra 4
58
1.05x106
7.0
24’500 ± 2’100
26’400 ± 2’100
Sample Wra 1 excluded, the new 26Al data agree within the uncertainties with the 10Be data
measured in the same rock samples (Karhu et al., 2000). This indeed indicates that the rock
55
surfaces experienced a single period of exposure, which lasted at least 11’500 ± 1’000 or
24’500 ± 2’100 years, respectively. The data suggest that no significant coverage occurred
during this time-period. Sample Wra 1 yields a 26Al exposure age of 1’400 ± 200 years, which
is much younger than the Wra 1 10Be age with 64’600 ± 6’400 years (Karhu et al., 2000).
Various causes could lead to this difference:
•
•
•
•
Contamination of 10Be samples during chemical processing
Loss of Al during the processing
Measurement problems (AMS, ICP-AES)
Unknown processes and factors
If the young 26Al age would be accepted as the true age of the sample, then, the high amount
of 10Be could be explained by contamination of the 10Be sample during the processing. But a
contamination seems unlikely, because the corresponding 10Be blank, which was processed at
the same time, does not differ from other 10Be blanks.
The loss of Al during the processing would yield to an underestimate of the total Al content in
the quartz sample (cf. Appendix Chapter 3.1). Similar to the observation during the HF digest
of the large Salpausselkä samples, the Wrangel Island sample Wra 1 yielded a large amount
of insoluble pinkish solid with unknown composition in the Teflon beaker. To determine the
qualitative atomic composition of the unknown solid, we performed a RBS analysis (see Chu
et al., 1978 for details) on the solid of sample Wra 1 (see appendix Chapter 3.1 for details).
The RBS results are presented in Table 3.6 and Fig. 3.8. Ca and F are the major elements that
are present in the solid. The abundance of Ca is also indicated by the ICP-AES analysis of the
dissolved part of sample Wra 1. The source of Ca is calcite that occurs as hydrothermal
deposit within the quartz. F precipitates together with Ca (i.e. CaF2, a poorly soluble salt). Al
is not observed within the sample. It was therefore not lost within the analyzed solid.
Table 3.6: Quantitative estimate of the elemental abundance in the unknown solid of sample Wra 1.
Note that the main part of the solid is CaF2 from abundant CaCO3 and HF fuming.
Wra 1
Element
Atom percent [%]
F
61.8
Ca
30.9
Si
3.1
O
3.1
Fe
0.8
others
0.3
56
Fig. 3.8: RBS energy spectrum of Wra 1 with normalized yield. The black line, overlying the
spectrum, is the calculated best-fit elemental simulation to the measured spectrum. The unknown solid
consists mainly of Ca and F. Minor amounts of Si and Fe are present. Al is not observed in significant
quantities (Table 3.6).
Irregularities in the AMS measurement of 10Be and 26Al can be excluded, because the samples
were re-measured and both the 10Be/9Be and the 26Al/27Al ratios were confirmed. Moreover,
the measured total Al content is in the range of the other two samples (Table 3.5). The true Al
content would have to be at least an order of magnitude larger for an agreement of the 10Be
and 26Al ages. It seems very improbable for the ICP-AES measurement to be so far off. The
large difference between the 10Be and the 26Al data in sample Wra 1 cannot yet be explained.
New sample material will be processed and measured.
57
3.3 Preliminary 10Be and 26Al dating of moraines in the
Kanding area, eastern Tibet
Due to its unique geomorphological configuration (large area at high altitude), the Tibetan
Plateau acts as an important part of the world’s climate system. However, the coupling of the
Tibetan Plateau’s climate to the rest of the system is still poorly understood. Global signals
(e.g. LGM, Younger Dryas cold reversal) have been traced in Tibetan ice cores and lake-cores
at the western and northeastern part of the plateau (Gasse et al., 1991; Thompson et al., 1989;
Thompson et al., 1997), but little is known about the reaction of Tibetan glaciers and ice caps
with respect to these events. Whereas Burbank and Cheng (1991) suggest a synchronous
advancing of glaciers at the southern margin of the plateau, Gillespie and Molnar (1995) and
Phillips et al. (2000) find evidence for asynchronous behavior during marine oxygen isotope
stage 2 (MIS-2) in the same region. The eastern margin of the Tibetan Plateau on the other
hand, shows synchronous glacial advances during MIS-2 (Lehmkuhl, 1997; Schäfer et al.,
submitted). However, one should keep in mind that the problem of synchronicity of glacial
behavior with respect to Northern Hemisphere glaciations is also closely related to the paleotemperature and paleo-precipitation patterns on and around the Tibetan Plateau. Both, a
certain drop of temperature and a sufficient amount of precipitation are necessary to let
glaciers advance. Absolute dating of distinct glacigenic deposits is the key to the
understanding of the glacial behavior of the plateau during the Quaternary.
Fig. 3.9: Overview of the Tibetan Plateau. Stars mark the different SED sample sites of our field
campaigns 1997, Litang (Li) and Kanding (Ka), and 1998, Tanggula (Ta).
58
Apart from the question of synchronicity of former ice advances in Tibet with the rest of the
world, there is also an ongoing discussion about the extent of paleo-glaciations on the Tibetan
Plateau. Various opposing hypotheses exist to explain and reconstruct the glacial history of
this unique, high altitude feature of the Earth during the Quaternary (see Rutter, 1995 or
Benxing and Rutter, 1998 for a summary). The theories range from huge masses of ice, as
suggested by Kuhle (1998), to restricted ice-sheets and valley glaciations of local character
(Derbyshire et al., 1991; Lehmkuhl, 1998; Zheng, 1989). Again, the key to the verification of
these hypotheses and models lies in the understanding of the stratigraphical order. For this
reason, absolute dating is necessary, as it provides the possibility for correlation and
identification of different glacial stages. Present day dating is mainly based on thermoluminescence dating (TL), optical stimulated luminescence dating (OSL) or, to a lesser extent,
on radiocarbon dating (for an overview see Benxing and Rutter, 1998). In addition,
knowledge of absolute chronology is limited to a few places within Tibet, because suitable
material is often lacking. SED could fill this gap. We therefore initiated two sample
campaigns to collect sample material and to obtain chronological information on defined
glacigenic deposits.
The campaigns were on the eastern margin of Tibet and in the central part of the plateau (Fig.
3.9). They were expected to provide the following information:
•
•
•
•
Is a correlation between different parts of Tibet (here: central and eastern part) possible?
Do we find pre-LGM deposits in Tibet?
Can we identify glacial advances, which correlate with LGM or Younger Dryas advances
in the Northern Hemisphere?
Is there evidence for a huge ice sheet (i.e. shielding of SED samples)?
The results from the eastern margin (Litang) and from the central part are summarized in
(Schäfer et al., submitted), found in Chapter 3.4 of this thesis. Additional data from the
eastern margin (from the Kanding area between Litang and Chengdu, Fig. 3.9) are discussed
in the following section.
Methodology
Geologic setting
The sampling area (30°05N, 101°50E) is about 240km east of Litang near the city of Kanding
at an altitude of 4’250m.asl (Litang is at about 4’600m.asl). The samples originate from a
system of north-south trending medial or lateral moraines found on a shoulder on the
northeastern side of the northwest to southeast trending main valley. The shoulder is
terminated by a steep slope to the west and south going down to the valley floor that is about
59
20 to 40m below. The assumed paleo-ice-flow direction of the local valley glaciers is from
north to south (arrows in Fig. 3.10) with an estimated distance from the feeding cirque to the
moraines of 2km to 3km. Field evidence suggests at least two generations of moraines: an
older moraine that is laterally overlain by a younger glacigenic deposit. Both features are
covered by a large number of boulders. Reworking of the moraines by later glacial advances
seems unlikely, whereas permafrost activity cannot be excluded. The older moraine was
supposed to be pre-LGM, whereas the younger moraine was thought to be LGM (pers.
communication Wu, 1997). Note that detailed mapping of the Quaternary deposits is missing
in the sampling area and no absolute chronology of glacial events is available.
Fig. 3.10: Sample location at Kanding area, eastern Tibet. The samples were taken close to the
Chengdu - Lhasa highway, about 15km west of the city of Kanding. The arrows mark the assumed
flow direction of the former glaciers, which would have deposited the sampled moraines.
Two samples were taken for preliminary absolute ages of these moraines. Sample K1 lies on
the flat shoulder of the older moraine. Boulder K2 was taken from the top of the younger
moraine just at the beginning of the westward slope. There, we have looked for signs of
downslope movement, but we believe that the boulder is still in situ. To minimize the effect
of erosion, boulders with resistant knobs (“Härtlinge”) at top positions were chosen (see also
Fig. 2.5). Vegetation is limited to small bushes and grass. Snow coverage is assumed to
reduce the production rate of cosmogenic nuclides by less than 1%.
60
Method
Sampling procedures, laboratory processing, AMS analysis and the age calculation are
analogous to the SED studies already described in this thesis. The production rates of Kubik
et al. (1998) and the scaling formalism of Lal (1991) were used.
Results
The results and minimum exposure ages of the two sampled moraines are presented in Table
3.7 a & b. Note that the concentrations of both 10Be and 26Al, yield corresponding ages for the
sampled surfaces. The 10Be and 26Al minimum exposure ages for K1 are 13’470 ± 1’030 and
12’590 ± 950 years, respectively and for K2 11’440 ± 900 years (10Be) and 11’470 ± 920
years (26Al). The errors of the exposure ages include the 1σ AMS measurement errors, a 5%
variability of the chemical processing (Ivy-Ochs, 1996), a 4.2% (10Be) and 5.1% (26Al)
uncertainty for the used production rate (Kubik et al., 1998), an uncertainty of 30% for the
determination of the sample thickness and an uncertainty for a top coverage (e.g. vegetation
and snow) of 2g/cm2. The ages indicate that moraine K1 is slightly older than moraine K2,
which is supported by field evidence. The preliminary character of the study does however
not allow final conclusions about the relative stratigraphy of this moraine system.
Table 3.7a: AMS measured 10Be and 26Al concentrations for Kanding boulders. The error includes a
1σ AMS measurement error and a 5% variability of the chemical processing (Ivy-Ochs, 1996). The Al
concentration was measured with ICP-AES.
Sample
10
Be concentration Error [%] Al concentration
[Atoms/g SiO2]
26
Al concentration
[ppm]
[Atoms/g SiO2]
Error [%]
K1
9.165x105
6.4
202
5.496x106
6.3
K2
7.865x105
6.6
365
5.058x106
6.8
Table 3.7b: 10Be and 26Al minimum exposure ages. The production rates of Kubik et al. (1998) were
scaled according to Lal (1991). Shielding by surrounding mountains was less than 1%. Production rate
corrections due to finite sample thickness were made according to Masarik and Reedy (1995). See text
for error description.
Sample
Lithology
Thickness
Present
[cm]
altitude [m.asl]
10
Be age [years]
26
Al age [years]
K1
Granite
1.5
4240
13’470 ± 1’030
12’590 ± 1’020
K2
Granite
1.3
4260
11’440 ± 900
11’470 ± 980
Both boulders show erosional effects. Although the samples represent resistant parts of the
boulders (knobs), we have to keep in mind that the presented ages are minimum ages. To get
61
an idea about the effect of continuous erosion, we tried to estimate erosion rates appropriate
for the encountered climate and lithology. Unfortunately, there is little independent
information about erosion rates. Saunders and Young (1983) give a value of 5mm/kyr,
derived from studies on rock tablets in montane terrain. With such a steady state erosion rate
of 5mm/kyr and a rock density of 2.6g/cm3, the ages increase slightly to 13’620 ± 1’040 and
12’700 ± 1’030 years (10Be and 26Al) for K1 and 11’510 ± 900 and 11’540 ± 980 years for K2
(10Be and 26Al).
Discussion
The results clearly indicate an age younger than 20ka for both moraines, even if moderate
continuous erosion is taken into account. The deposits do therefore not represent glacial
advances, which are correlated to LGM with respect to the Northern Hemisphere (e.g.
Wisconsinan continental ice sheets, Gillespie and Molnar, 1995). However, the deposition of
the two moraines falls within late glacial time at the transition of MIS-2 to the Holocene. Our
preliminary results therefore support a synchronous behavior of the glaciers at the eastern
margin of the Tibetan Plateau, with advancing glaciers during LGIT. This is in contrast to
SED results for the southern margin of the Tibetan Plateau (Himalayas) (Phillips et al., 2000),
which indicate asynchronous growing of glaciers during MIS-2. On the other hand, our results
from Kanding confirm SED dating from Schäfer et al. (submitted, see chapter 3.4) in the
geographical “neighborhood” in the area of Litang. There, exposure ages also indicate glacial
advances during MIS-2. Kuhle (1998) proposes that the huge ice sheet, covering the Tibetan
Plateau, was built and melted again during the last glacial maximum (end of MIS-2). Because
the moraines of Kanding clearly indicate a deposition by advancing valley glaciers, we
conclude that the proposed ice sheet, if it existed at all, had to be melt down before 12’970 ±
600 years ago at the latest (if we take the weighted mean age of K1). Otherwise, the
deposition of the sampled moraines would not have been possible at that time.
Gasse et al. (1991) investigated a lake-core and the shorelines at Sumxi Co, western Tibet,
using δ18O and CH4. They find evidence for a return to cold and dry climate conditions
between 11’000 and 10’000 14C years BP (estimated ages, based on radiocarbon dating). This
(cold) signal is taken as a proxy for the Younger Dryas (YD) cold reversal. Ice core data from
the Guliya ice cap on the Tibetan Plateau (Thompson et al., 1997) show the presence of a
signal during the process of deglaciation after the last glacial sequence within MIS-2. This
signal is also interpreted as a proxy for the YD. One could therefore argue that the YD cold
reversal event is recorded on the Tibetan Plateau. If the precipitation during this cooling event
was sufficient, then glacial advances could be expected. So far, no such advances are reported
on the Tibetan Plateau and no moraines were identified as deposits of the YD yet. However,
our preliminary minimum ages and the exposure ages calculated with estimated erosion rates
62
(5mm/kyr) of the two moraines K1 and K2 fall within the limits of the accepted YD time
brackets from the Greenland ice cores (Fig. 3.11). The GISP2 ice core puts the YD in the
range of 12’940 ± 550 to 11’640 ± 250 (Alley et al., 1993), while the GRIP ice core yields a
range from 12’700 ± 100 to 11’550 ± 70, (Johnsen et al., 1992).
Fig. 3.11: Preliminary results of SED at Kanding area, eastern margin of Tibetan Plateau in
comparison with YD ice core time brackets from GISP2 (Alley et al., 1993) and GRIP (Johnsen et al.,
1992). Crosses (x) mark 10Be ages, filled circles (•) mark 26Al. The sample set is shown for zero
erosion and for an estimated steady state erosion rate of 5mm/kyr.
However, many factors influencing SED ages would have to be tested before one can clearly
identify the moraines to be the fingerprint of the YD cold reversal at the eastern margin of the
Tibetan Plateau. Continuous permafrost activity for example could have overturned the
sampled boulders after deposition. If the sampled surfaces had not been at the top position
since their deposition, then, the calculated ages would be too young. Verification with more
data from the same moraines could exclude such processes. Moreover, the geomorphological,
stratigraphical and topographical context of the area would have to be investigated with
respect to the position of the terminal moraines as part of the medial or lateral moraines
sampled.
63
3.4 The limited influence of glaciations in Tibet on global
climate over the past 180’000 years3
Abstract
Extensive ice cover on the Tibetan Plateau would significantly influence Earth’s climate in
general and the Asian monsoon system in particular, but extent and timing of Quaternary
glaciations in Tibet remain highly controversial. We date moraines with cosmogenic nuclides
showing that glacial advances were restricted to few ten kilometers during the last 180kyrs in
Central Tibet and during Oxygen-Isotope-Stage 2 (~24-13kyrs ago) in Eastern Tibet.
Advances of Tibetan glaciers were much less prominent than elsewhere in the northern
hemisphere, a proposed ice-dome covering the entire Plateau can be excluded. Thus, albedo
increase of Tibet did trigger neither northern hemisphere ice ages nor paleomonsoon changes
during the last two glacial cycles.
Introduction
To identify regions on Earth having a driving influence on the planet’s climate system is at
the forefront of modern climatology. Thereby, high altitude-low latitude areas are of
outstanding interest, because they have the potential to drastically change the global
circulation patterns (Ose, 1996; Raymo and Ruddiman, 1992). It is therefore amazing that the
role of the Tibetan Plateau within the Earth’s climate system is still poorly understood.
Particularly relevant in this context is the history of glaciations: on the one hand glacial
advances mirror climate changes to cold and/or humid conditions, on the other hand
glaciations in Tibet have potentially a high impact on the global radiation budget. Extended
snow and ice cover on the huge Plateau area of more than 2.5*106 km2 elevating 4.5km on
average would dramatically change the albedo of Eurasia, thereby modifying the temperature
and air pressure gradients between continent and ocean. This would have direct impact on the
3
Science, submitted
Jörg M. Schäfer, Isotope Geochemistry and Mineral Resources, ETH Zürich, 8092 Zürich, Switzerland
Zhizong Zhao, Institute of Geomechanics, Chinese Academy of Geological Sciences, Beijing, China
Xihao Wu, Xi’an Laboratory of Loess and Quaternary Geology, Chinese Academy of Sciences, Xi’an, China
Rainer Wieler, Isotope Geochemistry and Mineral Resources, ETH Zürich, 8092 Zürich, Switzerland
Heinrich Baur, Isotope Geochemistry and Mineral Resources, ETH Zürich, 8092 Zürich, Switzerland
Corresponding author: Jörg M. Schäfer, [email protected]
64
Asian monsoon (Vernekar et al., 1995) and possibly even on the circulation and climate on a
hemispheric to global scale (Kuhle and Herterich, 1989; Ye and Wu, 1998).
There is some evidence for a correlation between climate changes in China (Porter and An,
1995; Wang et al., 1999) and in Tibet itself (Thompson et al., 1997) on the one hand and
excursions in the climatic key area of the North Atlantic on the other, where the formation of
North Atlantic Deep Water drives the enormous energy pump called thermohaline circulation
(e.g. the cooling during the Last Glacial Maximum (LGM) of the northern hemisphere some
20kyrs ago). However, the response of the Tibetan glaciers to the LGM and other Quaternary
climate signals is still subject to controversy. Today, less than 1% of the Plateau’s surface is
covered by glaciers (Häberli et al., 1989), but how extensive were the ice caps in the past?
Many relative glacial stratigraphies have been proposed for the Tibetan Plateau (Burbank and
Cheng, 1991; Derbyshire et al., 1991; Rutter, 1995; Lehmkuhl, 1997; Benxing and Rutter,
1998), however absolute dating is lacking. Most of these authors suggest several main glacial
advances of various extensions during the last glacial cycle (Burbank and Cheng, 1991;
Derbyshire et al., 1991; Lehmkuhl, 1997; Lehmkuhl and Haselein, 2000), whereas Kuhle
(1998) presents a scenario of an enormous ice-dome covering the entire Plateau and even its
flanks during late Pleistocene. According to this author, the increased albedo of Eurasia due
to this Tibetan ice-dome triggered the climate changes that culminated in the huge northern
hemisphere glaciations during LGM. Furthermore, it has been inferred that the Asian
monsoon system was repeatedly switched to a mode with reduced intensity of precipitationrich summer monsoon and stronger winter monsoon, respectively (Sarkar et al., 1990; Emeis
et al., 1995). These modifications occurred synchronously to glacial times in the northern
hemisphere and are ascribed to an extensive snow and ice cover of Tibet. These latter studies
propose inter-connection between glacial advances covering a substantial area of the Tibetan
Plateau and climate cooling in the North Atlantic region. In contrast, recent work from the
Karakorum, western margin of the Plateau (Phillips et al., 2000; Richards et al., 2000), report
absence of glacial advances during northern hemisphere LGM times, but glacial advances in
the Holocene synchronous to the period of increased summer monsoon of Southwest Asia
(Phillips et al., 2000). This suggests that at least the paleoglaciations in the Karakorum were
related to higher precipitation caused by the modified monsoon (contemporary to northern
hemisphere inter-glacials) rather than to cooler temperatures in glacial periods of the North
Atlantic system. However, it is unclear whether the Karakorum area is representative for the
whole Himalayas or even the entire Tibetan Plateau (Benn and Owen, 1998).
Summarizing, two crucial questions need to be answered to evaluate the importance of Tibet
within Earth’s climate system: (i) When did glaciers in Tibet advance and is there a causal
relationship between Tibetan glacial events and climate cooling in the North Atlantic on the
65
one hand and higher precipitation in the Asian monsoon area on the other? (ii) Did an
extensive ice-sheet exist on the Plateau in the late Quaternary influencing the circulation
patterns on a regional (monsoon) to hemispheric scale? The main reason for this ignorance is
the lack of absolute dating of the past glacial advances on the Tibetan Plateau, especially in
the key areas Central and Eastern Tibet.
The goal of this work is to provide such age information by investigating glacial erratics on
top of moraines in the Litang Area in East Tibet and in the Tanggula Area, Central Tibet. We
date these erratics and therefore the glacial advances by means of cosmogenic 10Be, 26Al, and
21/22
Ne. This allows us to present first reconstructions of Quaternary glacial advances in
Eastern and Central Tibet for the last almost 200kyrs. Dry Central Tibet is the source area of
the hypothetical ice-dome and therefore key to the conflict about its existence and the
underlying question whether Tibet’s paleoclimate triggered the northern hemisphere ice-ages
and switched the modes of the Asian monsoon system, respectively. The hardly investigated
Eastern margin of the Tibetan Plateau is an appropriate area to evaluate whether glacial
advances have been synchronous to those of the northern hemisphere or not. Synchronicity
would point to a transcontinental climate link between Tibet and the North Atlantic region,
asynchronicity would rise the question whether the glacial advances in Tibet are related to an
alternative climatic feature, e.g. to periods of increased summer monsoon causing higher
precipitation rates as already proposed for the western Plateau margin (Phillips et al., 2000).
Geological setting and methods
Sampling sites are shown in Fig. 3.9 (see p. 57). Litang County, eastern margin of the Plateau,
is situated in the humid part of Tibet, whereas Tanggula Pass Area, forms part of the dry
region of the Central Tibetan Plateau. Any extensive ice sheet covering significant parts of the
Tibetan Plateau would cap the Tanggula Area, the ice-dome proposed by Kuhle (1998) would
require more than 1’000m of ice overlying the Tanggula Mountains.
All samples are glacial erratics on top of clearly defined glacial deposits some 15km (Litang,
Fig. 3.12) and 30km (Tanggula) away from the present glacier, respectively. Note that any
significant glacial overriding of the moraines after their initial deposition would have
removed the original glacigenic landscape elements (Cuffey et al, 2000), i.e. the sampled
erratics. In Litang County (30°17’N, 99°32’E), we sampled in a glacially U-shaped valley
(Fig. 3.12) at ~4’600m. Samples Lit 3-6 are from the moraine representing the most extensive
glacial advance visible in this valley. The moraine of sample Lit 7 is about 500m closer to the
present glacier. The lithology is a quartz-rich, hard granite. In the Tanggula area (32°30’N,
91°50’E), the two moraine systems belonging to the most extensive glacial advances visible
south of Tanggula Pass were sampled. All Tanggula moraine ridges present permafrost
66
features. Due to melting of the active layer in summer, the erratics have sunken into the
ground since their deposition. The lithology is a hard quartz-andesite bearing about 10%
quartz.
Quartz separates were analyzed for cosmogenic noble gases and radionuclides according to
Niedermann et al. (1993) and Kohl and Nishiizumi (1992), respectively, using the method of
SED. Analyses were performed at the AMS facility PSI/ETH Zürich, and the noble gas
laboratories at ETH Zürich. SED is particular suitable to establish a glacial chronology on the
Tibetan Plateau, because the production rates of cosmogenic nuclides are ~20 times higher
than at sea level and because suitable glacial deposits are abundant, i.e. quartz-bearing glacial
erratics. Alternative methods, like 14C and optically stimulated luminescence dating are
hampered by rareness of samples and a dating limit of < 40’000 yrs and < 100’000 yrs
(Richards, 2000), respectively. However, several potential complexities of SED have to be
considered, like: Complex exposure4, erosion5, post depositional coverage or overturning 6 and
nucleogenic neon7.
4
A first exposure of the sample followed by significant periods of burial and re-exposure is the only process that
might yield older-than-real exposure ages (erosion and post depositional cover of the surface yield too young
ages; see below). Here, all 10Be and 26Al ages (and for the Tanggula samples also the 21Ne ages) agree with each
other for a given sample, which strongly indicates a simple and continuous exposure history. Furthermore, the
ages of four different boulders Lit 3, 4, 5, and 6 are all the same, again strongly implying a simple exposure
history, as identical preexposure histories of four boulders are highly unlikely.
5
Erosion leads to underestimation of the cosmogenic nuclide concentration. The ages underlying all conclusions
made here assume conservatively a range of erosion rates from 0-20mm/ka for the samples from the rather
humid alpine environment of Litang, whereas for the samples from the dry Tanggula area zero erosion is
assumed yielding minimum exposure ages. Most likely erosion rates are 10mm/ka for the Litang samples (Small
et al., 1997) and 1mm/ka for the Tanggula samples, respectively (see Table 3.8).
6
The boulders from Litang are 10-50m3 in size, deposited on flat moraine ridges. The undisturbed soil around
the boulders indicates absence of sliding or turning of the rocks. The boulder size makes substantial snow cover
unlikely. Parts of the Tanggula samples stick out of the permafrost soil by 1m or less. In the current arid climate
snow is hardly covering the rocks, however this cannot be excluded in the past. In addition, the rock’s sinking
could have prevented continuous exposure of the sampled surface. Like erosion, these processes would increase
the real age of our samples. The reported ages of the Tanggula rocks are therefore considered as minimum ages.
The most likely explanation for the inconsistent ages of the two boulders TAN 4 and 5 from the same moraine is
a rotation of TAN 4 after deposition, yielding a too young exposure age (see also Table 3.8).
7
A stepwise heating procedure was performed for Ne analysis and the < 600°C steps are used to calculate
exposure ages (Niedermann et al., 1993). However, a mixture of atmospheric, cosmogenic, and nucleogenic
neon even in the low-temperature steps of the Litang samples is indicated by the neon isotope ratios. A reliable
correction for nucleogenic neon is problematic and we therefore use the 21Ne excess above air to calculate an
upper limit of the exposure ages of the Litang samples. These values (given in brackets in Table 3.8) argue
against a longer period of preexposure. No nucleogenic neon was observed in the Tanggula samples.
67
Results and implications
Litang samples: Cosmogenic radionuclide concentrations of Lit 3-6 are identical within errors
(Table 3.8). These boulders have been deposited by a glacial advance certainly between
14kyrs and 22kyrs ago. This age interval considers a conservative estimate of the possible
range of erosion rates for the sampled rocks from 0-20mm/kyr (Small et al., 1997). The
presumably most reasonable erosion rate of 10mm/kyr (Small et al., 1997 and footnote 5)
yields the most likely deposition age for the dated boulders of 16.4 ± 1.5kyrs ago. In any case,
the dated ~15km glacial advance in Litang County occurred during Oxygen Isotope Stage 2.
Sample Lit 7 shows a slightly lower age, possibly indicating a younger and less extensive
glacial advance in accordance with the position of this sample (Fig. 3.12). However, since
uncertainties are too high, we concentrate on samples Lit 3-6.
Fig. 3.12: Litang County sampling site, Eastern Tibet. Positions of samples Lit 3-7 are shown. In the
background: retreated glacier that formed the U-shaped valley. The moraines are less than 15km away
from the present glacier. One example for the huge, glacially rounded granitic erratics is indicated by
the arrow to the lower right. Note the clearly defined quaternary geomorphology of this sampling site.
Tanggula Area: The geological setting of the Tanggula Pass sampling site is comparable. However,
the two moraines south of Tanggula Pass are some 30km (TAN 2) and 25km (TAN 4, 5), respectively,
away from today`s glacial position. The sample north of Tanggula Pass (TAN 7) is less than 10km
away from the present glacier. A geological sketch map of the Tanggula area including the two
sampled moraines south of the pass is published in Lehmkuhl (1997).
Tanggula Samples: The erratics from the two moraines south of Tanggula Pass have
strikingly high ages and thus were deposited by the so far oldest directly dated glacial events
on the Tibetan Plateau. Consistent with its position, the most extensive glacial advance is the
68
oldest one with an age of 181 ± 22kyrs, defined by sample TAN 2 (ages stated here for
Tanggula samples are minimum values; see Table 3.8 and footnotes 5 and 6). Sample TAN 5
yields a minimum age of 173 ± 19kyrs for the second moraine system south of Tanggula Pass
(the age of TAN 4 is most likely too low, see footnote 6). The younger age of 70 ± 4kyrs for
the erratic TAN 7 on the northern side of Tanggula Pass reflects its position relatively close to
the present glacier. In summary, glacial advances in the Tanggula area have deposited the
most extensive moraine some 30km from the present glacier more than ~180kyrs ago and a
second moraine 25km from the present glacier tongue earlier than ~170kyrs ago (a deposition
of these two moraines by the same event is possible within uncertainties). Since at least
180kyrs, the moraines south of Tanggula Pass have not been overridden by a glacier. A
smaller advance occurred at least 70kyrs ago, which deposited TAN 7.
Table 3.8: 10Be, 26Al and 21Ne exposure ages for the Eastern and Central Tibetan erratics calculated
with production rates of Niedermann (2000) and Kubik et al. (1998), scaled after Lal (1991). The
exposure age for the main glacial event (bold) in Litang is based on the 10Be and 26Al data of the
samples Lit 3-6. For the Ne-ages given in brackets see footnote 7. The “conservative age interval”
represents the mean values of radionuclide ages based on 0 and 20mm/ka erosion, respectively. The
additional „best“ value is based on 10 mm/ka (Small et al., 1997). The mean minimum ages (zero
erosion assumed) for the Tanggula samples are based 10Be, 26Al and 21Ne data (with the exception of
TAN 7, where the uncertainty of the Ne-age is unreasonably high). The TAN 2 age is exemplary also
given with an erosion rate of 1 mm/ka, to give a feeling for sensitivity of exposure ages to erosion. For
the TAN 4 age given in brackets see footnote 7. All ages used for further implications (bold) rely on
the conservative age interval (Litang) and the minimum ages (Tanggula), respectively. Radionuclide
errors are 2σ and include analytical and blank errors. Neon uncertainties represent 2σ of the 21Neexcess above air. Calibration uncertainties are about 3 %. The overall uncertainty of given exposure
ages (including production rate uncertainty) should be lower than 15%.
Altitude [m]
Min. 10Be-ages [ka]
Min. 26Al-ages [ka]
Lit 3
4’560
13.9 ± 1.4
15.5 ± 1.4
(29.5 ± 6.0)
Lit 4a
4’560
17.9 ± 2.0
15.5 ± 1.7
(28.4 ± 6.4)
Lit 4bc
4’560
15.0 ± 1.7
15.0 ± 1.7
Lit 5a
4’610
15.3 ± 1.6
13.8 ± 1.5
Lit 5b
4’610
15.3 ± 1.5
13.8 ± 3.0
Lit 6
4’570
15.6 ± 1.4
15.0 ± 2.0
Sample
21
Ne-ages [ka]
EASTERN MARGIN; LITANG
Conservative age interval
(52.9 ± 9.2)
(21.4 ± 4.7)
14.0 – 22.0
erosion 0 – 20 mm/ka
16.4 ± 1.5
“Best” value, 10 mm/ka
Lit 7
4’480
13.3 ± 1.4
14.3 ± 3.0
(27.5 ± 7.7)
69
Table 3.8 continued
Minimum 10Be-ages Minimum 26Al-ages
CENTRAL PLATEAU,
TANGGULA
Tan 2
5’015
[ka]
[ka]
162.3 ± 19.4
171.5 ± 17.2
21
Minimum
Mean ages
Ne-ages [ka]
[ka]
210.0 ± 20.5
181.3 ± 22.0
200.1 ±25.1
TAN 2, erosion: 1 mm/ka
Tan 4
4925
91.1 ± 9.2
74.2 ± 9.0
102.6 ± 15.1
(89.3 ± 15.9)
Tan 5
4925
164.9 ± 16.6
161.8 ± 16.2
192.0 ± 17.8
172.9 ± 19.0
Tan 7
5120
70.5 ± 7.0
68.4 ± 8.2
(77.0 ± 45.1)
69.5 ± 4.3
The first absolute datings for glacial advances at the Eastern margin and the Central Plateau
of Tibet, respectively, indicate that the most extensive advance of some 30 km relative to
present glacier position occurred > 180kyrs ago and therefore much earlier than Last Glacial
Maximum time. All later glacial advances in Central Tibet have been smaller. Thus, climate
excursions to cold and humid conditions in Central Tibet have been much less dramatic than
in other parts of the northern hemisphere at least in the last ~180kyrs. This finding excludes
the Tibetan Plateau as a pacemaker of northern hemisphere ice ages by an initial plateau
glaciation (Kuhle, 1998). Similarly, our data strongly contradict the idea that the switch of the
Asian monsoon mode during northern hemisphere glacial times, e.g. during LGM, are caused
by an extensive ice-cover on the Tibetan Plateau (Emeis et al., 1995; Sarkar et al., 1990). It is
beyond the scope of this paper to discuss alternative processes having the potential to modify
the monsoon circulation synchronously to glacial times in the northern hemisphere.
In East Tibet, glaciers advanced during Oxygen Isotope Stage 2. However, this event has been
much smaller than in other parts of the northern hemisphere, where glacial advances during
the Last Glacial Maximum have been on the order of ~100km, e.g. Rutter (1995). No more
extensive glacial advance occurred on the Eastern Plateau margin later, i.e. neither the Eastern
nor the Central Plateau glaciers advanced significantly during the Holocene monsoon
precipitation maximum of South Asia. Therefore, the paleoglaciation history of the Himalayas
at the western margin of the Plateau reported by Richards et al., (2000) and Phillips et al.
(2000), i.e. absence of any glacial event during Oxygen Isotope Stage 2 but advances in the
Holocene, seems to be out of phase with the rest of Tibet. Our study suggests a correlation of
the North Atlantic and East Tibet climate, respectively, most likely established by the effect
of westerly winds. This finding is in agreement with earlier studies reporting such a climatic
link between the North Atlantic on the one hand and northwest Tibet (Thompson et al., 1997)
and China (Porter and An, 1995; Wang et al., 1999) on the other. The moderate amplitude of
the Tibetan glacial advances and underlying climate changes in the last 180kyrs in general
and during LGM time in particular is consistent with the expectation that climate coolings
70
during glacial times have been most pronounced around the North Atlantic and smaller
elsewhere, e.g. Alley and Clark (1999), as the weak (or even switched-off) Deep Water
formation during glacials reduced (or stopped) the thermohaline circulation and therefore the
oceanic heat transfer to the North Atlantic. In summary, this study implies that ice extent on
the Tibetan Plateau has been very limited during at least the last two glacial cycles which
excludes the albedo of the Asian continent as a critical parameter for regional to hemispherewide climate changes. However, a detailed glacial chronology especially in the western part
of the Tibetan Plateau on time scales of several 100kyrs still needs to be established by
studies similar to the one presented here.
Acknowledgments
We thank Walter Wittwer for mineral separation and Max Döbeli for his analytical help and
acknowledge the support of Swiss National Foundation (Grant 2000-053942.98/1), Chinese
Academy of Sciences and Chinese Academy of Geological Sciences.
To be consistent with the other data chapters in this thesis, the data of chapter 3.4 were
reevaluated (Table 3.9). All results are given at 1σ confidence level and uncertainties of the
reference production rate are now included in the exposure age uncertainties.
Table 3.9: Reevaluated exposure ages of Table 3.8. All ages are given at 1σ level with included 3%
error for calibration gas uncertainty (maximum estimate) and measurement uncertainties of the
reference production rates: 4.2% for 10Be, 5.1% for 26Al and 13% for 21Ne (Kubik et al., 1998;
Niedermann et al., 1994 updated by Niedermann, pers. comm.).
Sample
Minimum 10Be-ages [ka]
Minimum 26Al-ages [ka]
Minimum 21Ne-ages [ka]
Lit 3
13.9 ± 1.2
15.2 ± 1.3
(29.5 ± 4.9)
Lit 4a
18.0 ± 1.6
15.5 ± 1.4
(28.4 ± 5.0)
Lit 4bc
15.0 ± 1.3
15.0 ± 1.3
-
Lit 5a
15.3 ± 1.3
13.8 ± 1.2
(52.9 ± 8.4)
Lit 5b
15.3 ± 1.2
13.8 ± 1.8
-
Lit 6
15.6 ± 1.1
15.0 ± 1.4
(21.4 ± 3.7)
Lit 7
13.3 ± 1.1
14.3 ± 1.8
(27.5 ± 5.3)
Tan 2
162.3 ± 11.3
171.5 ± 13.8
210.0 ± 29.8
Tan 4
91.1 ± 6.3
74.2 ± 6.2
102.6 ± 15.6
Tan 5
164.9 ± 11.9
161.8 ± 12.9
192.0 ± 27.1
Tan 7
70.5 ± 5.0
68.4 ± 5.6
(77.0 ± 24.8)
CHAPTER 4: SOUTHERN HEMISPHERE
71
Chapter 4
Southern Hemisphere
4.1 Last major advance of Taylor Glacier into central
Beacon Valley at least 4 Ma ago: New indications
from surface exposure dating on clasts from Granite
drift8
Abstract
Surface exposure dating with cosmogenic nuclides provides new absolute chronological
information for Granite drift, which was deposited by an expanded Taylor Glacier into central
Beacon Valley, Antarctica. Our analysis yields a minimum exposure age for the Granite drift
of 4.0Ma. Based on a multi-nuclide analysis (10Be, 26Al and 21Ne), which indicates a simple
exposure history for the sampled drift, this is also a minimum age for the last major advance
of Taylor Glacier into central Beacon Valley. Dating Granite drift also yields a minimum age
8
Geology, submitted.
Jörg M. Schäfer, Institute of Isotope Geology and Mineral Resources, ETH Zürich, 8092 Zürich, Switzerland
Noel Potter, Department of Geology, Dickinson College, Carlisle, PA 17013, USA
George H. Denton, Department of Geological Sciences and Institute for Quaternary Studies, University of
Maine, Orono, ME 04469, USA
David R. Marchant, Department of Earth Sciences, Boston University, Boston, MA 02215, USA
Corresponding author: Silvio Tschudi, [email protected]
72
of 4.0Ma for the debated Sirius Group deposits at nearby Mt. Feather, located at the head of
Beacon Valley (e.g. Marchant et al., 1993a). We also conclude that Taylor Dome, which
feeds Taylor Glacier, did not thicken significantly in the last 4.0Ma. Lastly, the study supports
the contention that previously dated ashes in Beacon Valley are indeed in situ and that buried
remnant ice underlying the floor of central Beacon Valley is at least of Miocene age (Sugden
et al., 1995b).
Introduction
Beacon Valley, Antarctica (78°S and 161°E) is a crucial site for the reconstruction of
Antarctic paleoglaciations and for the understanding of Antarctic landscape evolution. Beacon
Valley links Mt. Feather, one of the key locations of the Sirius Group at ~2’900m.asl
elevation (Brady and McKelvey, 1979; Brady and McKelvey, 1983), with low-elevation
glacigenic sediments (< 1’400m.asl) in the western sector of the Dry Valleys tectonic block of
the Transantarctic Mountains. This is based on the concept that down cutting of Beacon
Valley post-dates deposition of the Sirius Group sediments at Mt. Feather (Denton et al.,
1993). Beacon Valley is nearly surrounded by mountains, although it opens to Taylor Glacier
at the valley mouth. Taylor Glacier is an outlet glacier that drains Taylor Dome, a small ice
rise on the East Antarctic Ice Sheet about 35km west of Beacon Valley. Expansion of Taylor
Dome does result in an advance of Taylor Glacier into lower and central Beacon Valley
(Denton et al., 1993; Marchant et al., 1994; Marchant et al., 1993a). Beacon Valley therefore
acts like a small “overflow basin” for this part of the East Antarctic Ice Sheet (EAIS). This
makes glacial sediments in Beacon Valley an important archive for past fluctuations in Taylor
Glacier, Taylor Dome, and nearby sectors of the EAIS. Although a relative stratigraphy of
surficial sediments in Beacon Valley is well defined (Denton et al., 1993; Linkletter et al.,
1972; Marchant et al., 1993b; Marchant et al., 1993a; Potter and Wilson, 1983), an absolute
chronology of Taylor Glacier deposits in Beacon Valley is lacking.
A granite-bearing till, known informally as Granite drift, represents a major advance of
Taylor Glacier into central Beacon Valley (Sugden et al., 1995b). This drift was deposited as
ablation till on the floor and eastern slope of Beacon Valley (Sugden et al., 1995b). It includes
from 1 to 10% granite cobbles and boulders foreign to Beacon Valley. The age of Granite
drift is poorly constrained, but stratigraphic and geomorphic investigations suggest a possible
correlation to the Quartermain I and II drifts found in nearby Arena Valley. These drifts have
minimum ages of 7.4Ma and 11.3Ma, respectively, based on 40Ar/39Ar dating of associated
volcanic ashes (Marchant et al., 1993a). Underneath Granite drift lies a buried relict body of
ice, which is thought to represent the last vestiges of the glacier that deposited Granite drift
(Sugden et al., 1995b). Radiometric dates of volcanic ashes that lie stratigraphically above the
ice indicate a minimum age for the glacier relict of ~8.0Ma (Sugden et al., 1995b). This age is
73
based on the assumption that the ashes overlying the ice are in situ, a condition questioned by
Hindmarsh et al. (1998). Hindmarsh suggests that the ashes overlying the buried ice in Taylor
Valley may have been reworked and transported by younger advances of Taylor Glacier into
Beacon Valley and therefore should not be used to date the buried ice. On the other hand,
Schäfer et al. (2000) argued on the basis of cosmogenic studies that valley floor sediments in
Beacon Valley have not been reworked during the last 2.3Ma.
The goal of this study is to date Granite drift by means of surface exposure dating with
cosmogenic 10Be, 26Al and 21Ne. The new chronology does provide a minimum age for
deposition of the stratigraphically older deposits of the Sirius Group at Mt. Feather (e.g.
Marchant et al., 1993a) and, in turn, date periods of ice thickening at Taylor Dome. Moreover,
dating Granite drift yields a minimum age for the underlying remnant ice.
Fig. 4.1: Overview of Beacon Valley in the Dry Valleys, Antarctica, modified after (Sugden et al.,
1995b). The samples were taken along a transect from 1’380m.asl to 1’500m.asl on the eastern flank
of the valley 6km to the southwest of the present Taylor glacier (1). The (x) marks the site, where a
remnant buried body of ice was found (Sugden et al., 1995b).
Methods
We sampled six granitic clasts on the valley floor from the surface of Granite drift near its
outer margin (Fig. 4.1). The altitude of the samples range from 1’380m.asl to about
1’500m.asl above sea level, which is > 450m above the present level of Taylor Glacier at the
mouth of Beacon Valley. All sampled boulders have a size of several m3. The boulders
74
exhibit different degrees and types of erosion (e.g. spalling, flaking and pitting, Fig. 4.2), as
well as reddish oxidation rind. The sampled granite has been weakly metamorphosed, because
biotite (partly altered to chlorite) shows kink bands, and undulose extinction is abundant in
the mosaic quartz grains.
Fig. 4.2: Photograph of a sampled boulder “342” showing the effect of considerable erosion. Note that
this particular boulder bears the oldest age of our sample set.
Quartz was separated according to Kohl and Nishiizumi (1992) and the chemical separation
of 10Be and 26Al was done following Ivy-Ochs (1996); Kohl and Nishiizumi (1992) and
Nishiizumi et al. (1989). The concentrations of the cosmogenic nuclides 10Be and 26Al were
of the Paul Scherrer Institut (PSI) and the ETH. For the noble gas analysis, the quartz was
handpicked from the chemically separated quartz fraction. The processing for the 21Ne
analysis followed Bruno (1995). It was performed on a mass spectrometer (MS) at the
Institute of Isotope Geology and Mineral Resources (IGMR) of the ETH Zurich.
Results
Minimum exposure ages for Granite drift range from 0.9 ± 0.1Ma to 4.0 ± 0.2Ma (Table 4.1
and Table 4.2). All ages are strictly minimum ages, as no erosion or coverage (e.g. with finer
material or snow) is taken into account in the calculations. Note that the 10Be and 26Al results
all fall within errors into the “erosion island” (Klein et al., 1986; Lal, 1991; Nishiizumi et al.,
1991) of the two-nuclide-plot (26Al/10Be versus 10Be, Fig. 4.3). In addition, the 21Ne ages
75
agree with the radionuclide results. This rules out significant periods of prior exposure and
indicates a simple exposure history of the sampled glacial deposit.
Table 4.1: AMS measured 10Be and 26Al concentrations of Granite drift clasts in central Beacon
Valley, with calculated minimum exposure ages according to the equation given in Lal (1991). The
production rates and scaling are according to Lal (1991). The production rates were corrected for the
given sample thickness. Correction for shielding of surrounding mountains reduces the local
production rate by about 1%. The errors include 1σ AMS measurement errors for samples and blanks
and a 5% uncertainty for possible chemical processing variability (Ivy-Ochs, 1996).
Sample Lithology
339
Thick-
Present
10
Be conc.
Error
Total Al
ness [cm]
altitude
[Atoms/g
[%]
[ppm]
[m.asl]
SiO2]
1’380
1.81x107
Granite
1.4
26
10
26
Al conc.
Error
Be age
[Atoms/g
[%]
[106 years]
[106 years]
6.4
1.25 ±0.11
1.25
Al age
SiO2]
6.5
269
8.34x107
+0.17
-0.15
340
Granite
1.4
1’380
-
-
182
6.98x107
7.4
-
-
0.92
+0.12
-
341
Granite
0.8
2.52x107
1’400
6.5
442
11.26x107
10.5
1.98
-0.10
+0.22
-*
2.74
-0.20
342
Granite
0.5
3.59x107
1’430
6.1
94
11.62x107
5.8
3.87
-0.90
+0.77
2.83
+1.94
-0.57
343
Granite
0.9
2.77x107
1’450
7.3
105
10.64x107
6.4
2.19
-0.63
+0.29
1.94
+0.45
-0.26
344
Granite
0.8
1.68x107
1’500
6.5
257
8.39x107
8.4
1.01
-0.31
+0.08
1.06
+0.17
-0.08
-0.15
*26 Al in saturation, no age calculation possible
Table 4.2: MS measured
21
Ne concentrations of Granite drift clasts in central Beacon Valley,
Antarctica, with calculated minimum exposure ages. We used a production rate of pNe = 20 atoms/yr·g
SiO2 at latitudes > 50° and sea level (Niedermann et al., 1994 updated by Niedermann, pers. comm.),
which was scaled for elevation and latitude according to Lal (1991). The errors of 20Ne and the noble
gas ratios are given at 2σ level, including statistical, sensitivity and mass-discrimination errors.
Exposure ages are given at 1σ confidence level and errors due to uncertainties of calibration gas
amounts are included (3% as maximum estimation).
Sample
20
Ne
21
Ne/20Ne
22
Ne/20Ne
21
21
Ne (cosmogenic)
Error
Ne Age
[108 atoms/g SiO2]
[%]
[106 years]
[109 atoms/g SiO2]
[10-3]
342
27.37 ± 0.15
14.03 ± 0.49
0.115 ± 0.002
2.8967
3.7
3.99 ± 0.15
343.1
50.44 ± 0.33
6.66 ± 0.33
0.106 ± 0.002
1.7638
5.4
2.39 ± 0.13
343.2
96.39 ± 0.47
4.97 ± 0.13
0.105 ± 0.001
1.8376
4.4
2.49 ± 0.11
76
Fig. 4.3: Two-nuclide plot (26Al/10Be versus 10Be) of all measured samples. All concentrations are
normalized to sea level and high latitude, using the production rates and the scaling of Lal (1991). All
samples fall into the “erosion-island” (Klein et al., 1986; Lal, 1991; Nishiizumi et al., 1991), where the
gray area marks possible erosion rates ranging from zero erosion up to about 40cm/m.y. Sample 342
yields a maximum erosion rate of 7cm/m.y. Error bars are taken from Table 4.1.
All of the sampled boulders show some erosion. The scatter of the calculated exposure ages
reflects the observed different degrees of erosion. This means that the oldest exposure age of
4Ma (sample 342) of our sample set is still itself only a minimum age for deposition of
Granite drift. As the assumption of zero erosion is not realistic, in order to calculate “real”
exposure ages, erosion rates have to be estimated. Approximate steady-state erosion rates can
be derived graphically from the 26Al/10Be versus 10Be plot (e.g. Lal, 1991; Nishiizumi et al.,
1993). Our sample set covers erosion rates ranging from 0 up to about 40cm/m.y. (Fig. 4.3).
Sample (342) bearing the oldest age indicates a value between 0 - 7cm/m.y. This value is very
low, but comparable with other estimations derived from cosmogenic investigation (Brook et
al., 1995; Ivy-Ochs et al., 1997; Nishiizumi et al., 1993; Schäfer et al., 1999). An erosion rate
of 7cm/m.y. shifts the minimum age of 4Ma to a value of about 7.4 Ma. Although this number
may be closer to the true age of the Granite drift, we base all the following conclusions on the
minimum age of 4.0Ma.
77
Implications and conclusions
Taylor Glacier deposited the Granite drift, which represents a major advance of Taylor
Glacier into central Beacon Valley sometime prior to 4.0Ma. This is consistent with the
suggested correlation of Granite drift to Quartermain I and II drifts in Arena Valley. The
minimum age of 4Ma for the Granite drift suggests an early Pliocene or older age for the
Sirius Group deposits at Mt. Feather. Such a conclusion is in accord with recent age
estimations based on cosmogenic nuclide analyses from clasts of the Sirius Group at nearby
Mt. Feather, which yields minimum exposure ages of up to 5.3Ma (Ivy-Ochs, 1996; Ivy-Ochs
et al., 1997; Schäfer et al., 1999) for the deposition of Sirius Group. The age of 4Ma also
yields a new and independent minimum age for the underlying remnant body of ice.
Our multi-nuclide analysis supports a simple exposure history, indicating no coverage of our
samples since deposition. Therefore, Taylor Glacier never reached this part of Beacon Valley
again and Granite drift represents the last major advance. This conclusion also implies that
Taylor Dome, the feeding source of Taylor Glacier did not thicken significantly after early
Pliocene time, because this would result in a major advance into Beacon Valley. This
contradicts the idea of an overriding of high-altitude sites by the EAIS within the Dry Valleys
at 3.1Ma ago, as proposed by Webb et al. (1984). No glacial advance into central Beacon
Valley after deposition of Granite drift also means that the local valley floor sediments have
not been reworked significantly since at least 4Ma. These data strongly support the
observation that ashes on top of Granite drift are in situ and that the underlying remnant ice is
at least Miocene in age (Marchant et al., 1993a; Sugden et al., 1995b).
Acknowledgements
This work was partly funded by Swiss National Science Foundation grants 21-043469.95 and
21-053942.98 and by the United States Division of Polar Programs of the National Science
Foundation. Logistical support in Antarctica was given by the US-NAVY VXE-6. We thank
the Zurich tandem and noble gas crew for keeping the MS systems in great shape.
Usually, 10Be and 26Al were analyzed on the same part of the sample. Due to problems during
the laboratory processing of the Beacon Valley samples, 10Be and 26Al could not be measured
on the same part of the sample. This means that two different amounts of quartz were
processed (see Appendix B). However, this did not affect the results. The exposure ages of
both analyses are in good agreement to each other.
78
Application of new production rates (10Be and
drift samples
26
Al) for Granite
For the calculation of the minimum exposure ages in Chapter 4.1, the production rates of Lal
(1991) were used for 10Be and 26Al. The 21Ne ages were calculated with the production rates
of Niedermann et al. (1994 updated by Niedermann, pers. comm.). This was done to be
consistent with previously published SED data in the vicinity of Beacon Valley. Here, the
Granite drift data were re-evaluated with the 10Be and 26Al production rates of Kubik et al.
(1998) to be consistent within this thesis and for reasons discussed in Chapter 2.1. Moreover,
the uncertainties of the various production rates have been taken into account.
Table 4.3: Measured 10Be and 26Al concentrations of Granite drift clasts in central Beacon Valley and
calculated minimum exposure ages. We used production rates of Kubik et al. (1998), scaling of Lal
(1991) and the depth dependency of Masarik and Reedy (1995). The error includes a 1σ AMS
measurement uncertainty for sample and blank and a 5% uncertainty for possible chemical processing
variability (Ivy-Ochs, 1996). For further sample information see Table 4.1. Age errors include
production rate errors.
Sample
339
10
Be concentration Error
[Atoms/g SiO2]
[%]
1.81x107
6.5
26
Al concentration Error
[Atoms/g SiO2] [%]
8.34x107
6.4
10
Be age
6
26
Al age
6
[10 years]
[10 years]
1.32
1.19
±0.14
+0.20
21
Ne age
[106 years]
-
-0.17
340
-
-
6.98x107
7.4
-
-
0.88
-
341
2.52x107
6.5
11.26x107
10.5
2.12
+0.30
3.59x107
6.1
11.62x107
5.8
4.40
+1.40
2.47
2.77x107
7.3
10.64x107
6.4
2.34
+0.39
2.56
1.68x107
6.5
8.39x107
8.4
1.06
+0.11
-0.20
-
*
3.99 ± 0.54
-0.64
1.81
-0.33
344
*
-0.80
-0.86
343
-
-0.12
-0.26
342
+0.13
1.02
+0.51
2.39 ± 0.34
-0.34
2.49 ± 0.34
+0.19
-0.16
26
* Al in saturation, no age calculation possible
If the values of Kubik et al. (1998) are used, the results change only slightly (compare Table
4.1 and Table 4.3). The minimum exposure age for Granite drift clasts becomes now
4.40 +1.4/-0.86Ma, instead of 3.99 ± 0.15Ma. Within the stated uncertainties, the conclusions
of Tschudi et al. (submitted) do not need to be redrawn. The data still indicate a simple
history of exposure (Fig. 4.4) and Granite drift still corresponds to the deposit of the last
major advance of Taylor Glacier into central Beacon Valley (see Tschudi et al., submitted, for
details).
79
Fig. 4.4: Two-nuclide plot (26Al/10Be versus 10Be) of all measured samples. See Fig. 4.3 for details.
Beacon Valley rockfall
Introduction and methods
During the sampling campaign in Beacon Valley an undisturbed fan of rockfall debris that is
stratigraphically resting on the valley drift at the western valley wall (sample locality lies in
the vicinity of the x in Fig. 4.1) was sampled for SED (Fig. 4.5). The dating of this debris
should yield a minimum age for the last major advance of Taylor Glacier into central Beacon
Valley (see Tschudi et al., submitted), because the rockfall debris has not been transported
since initial deposition. We sampled the highest point of the largest accessible sandstone
boulder of the rockfall deposit (Fig. 4.6).
Prior exposure of the sampled surface cannot be excluded and calculated exposure ages do
not necessarily reflect the minimum age of the rockfall itself. To tackle this problem, a pilot
multi-nuclide study with two radionuclides (10Be and 26Al) and one stable nuclide (21Ne), all
measured within the same sample, was made. First results are presented in the following.
80
Fig. 4.5: Westward view across Beacon Valley towards the undisturbed debris fan of the sampled
rockfall. Under the valley floor in front of the rockfall, a remnant body of ice is buried (Sugden et al.,
1995b).
81
Fig. 4.6: Sampled boulder (size > 400m3) from the rockfall debris fan in Beacon Valley, Antarctica.
The altitude of the sample site is 1’345m above sea level.
Results
Ignoring prior exposure for the moment, minimum exposure ages can be calculated from the
10
Be, 26Al and 21Ne analyses. These ages range between 0.80 ± 0.07Ma and 0.87 ± 0.10Ma
(Table 4.4).
Table 4.4: Measured 10Be, 26Al and 21Ne concentrations for sample BV9714 from the rockfall debris in
central Beacon Valley, Antarctica. Minimum exposure ages were calculated with the production rates
of Kubik et al. (1998) and Niedermann et al. (1994 updated by Niedermann, pers. comm.). Scaling
according to Lal (1991) and correction for sample thickness (3.5cm) according to Masarik and Reedy
(1995). Surrounding mountains decrease the local production rates (97% of the “open sky” value). For
10
Be and 26Al, the exposure ages are given with 1σ AMS uncertainty, a 5% uncertainty for possible
chemical processing variability (Ivy-Ochs, 1996) and the uncertainty of the production rate (4.2% and
5.1%, Kubik et al., 1998). The 21Ne exposure age is given within 1σ confidence level with included
errors due to uncertainties of calibration gas amounts (3% as maximum value) and 13% error of the
production rate (Niedermann et al., 1994 updated by Niedermann, pers. comm.).
10
Be conc.
Error
10
Be min. Total Al
26
Al conc.
Error
[Atoms/g SiO2]
[%]
exp age [Ma]
[ppm]
[Atoms/g SiO2]
[%]
1.17x107
5.4
0.80±0.07
82
6.62x107
5.3
26
Al min.
21
Ne
exp age [Ma] [Atoms/g SiO2]
0.87±0.10
56.60x106
Error
21
Ne min.
[%]
exp age [Ma]
3.7
0.86±0.12
82
Discussion
As mentioned before, prior exposure and therefore inherited cosmogenic nuclide
concentrations have to be considered. The results of Table 4.4 may therefore not reflect a true
minimum exposure age of the rockfall. The following three scenarios were tested for
consistency with our measured concentrations:
•
•
•
only inheritance at the surface of the original location before a very recent rockfall
inheritance and exposure after rockfall
exposure to cosmic radiation only after rockfall
Basically, the amount of inherited cosmogenic nuclide concentration depends on the sample
position before the rockfall and on the time period it was exposed in this position. For the
original position of the boulder before the rockfall the following two assumptions were made:
•
•
The original position lies about 300m above the actual valley floor because sandstone is
outcropping at that altitude (see light lithology in Fig. 4.5).
For the calculation of the geometric shielding, a vertical wall was assumed. This yields a
reduction of the “open sky” production rate of 50%.
Dated in situ ashes (Sugden et al., 1995b) and clasts (Schäfer et al., 2000; Tschudi et al.,
submitted) on the floor of Beacon Valley suggest that the valley is a very old landscape (i.e.
in the order of 8 millions of years or older). Had the sample been at the surface in the original
boulder position before the rockfall (scenario a), the 21Ne concentration would have been
much larger than the measured, if the surface had not undergone erosion. The measured
concentration of 21Ne can however be used to determine the steady-state erosion rate, which is
necessary so that sample BV9714 was at the surface of the original boulder position for at
least 8Ma. Note that for simplicity reasons the exponential depth dependency was used in the
calculations. The estimated erosion rate is about 45cm/m.y., which is a rather high value
compared with other erosion rate estimations for the Dry Valleys (Brook et al., 1995; IvyOchs et al., 1997; Nishiizumi et al., 1993; Schäfer et al., 1999). However, it is not
unreasonable, because the sample’s lithology is a weak sandstone. The radionuclide
concentrations of 10Be and 26Al do however not reach their measured value with this erosion
rate. Scenario (a) is therefore not consistent with our data.
Scenario (b) describes the case of partial inheritance, where the measured concentrations are
divided into inheritance and “real” exposure of the sampled surface after the rockfall event.
Again, a minimum exposure time of 8Ma was used for the original position. Two different
cases are considered here: no erosion and a steady state erosion of 20cm/m.y. (arbitrary
83
chosen; it has to be less than 45cm/m.y. and more than zero). One can now calculate a critical
depth at which the 21Ne concentration reaches 100% of the measured concentration within the
given time period of 8Ma. This critical depth is located at about 1m for the no erosion case
and at about 0.5m for an erosion rate of 20cm/m.y. As the radionuclides do not reach the
measured values within the given time at these depths, one has to conclude that the samples
must originate from greater depths, where the production of 21Ne is lower and where 10Be and
26
Al are also produced by muonic reactions (Heisinger et al., 1997; Heisinger, 1998) (muon
production of 21Ne is unknown, but assumed to be negligible, pers. comm. Kubik). A
calculation of the inherited part, which would allow the determination of the minimum
exposure age of the rockfall, is not appropriate, because too many parameters (e.g. actual
production rate and actual erosion rate) are unknown. The lower depth limit of scenario (b) is
defined by scenario (c).
For scenario (c), spallation and muon reactions (Heisinger et al., 1997; Heisinger, 1998) are
considered for the estimation of the production rate of 10Be and 26Al. The production rates
were extracted from Fig 5.1, p. 96 of (Heisinger, 1998), whereas for 21Ne only spallation was
taken into account. For both cases (i.e. no erosion and steady-state erosion over a period of
8Ma), the concentrations of all nuclides at about 4m depth within the rockwall reach values of
less than 1% of concentrations measured. The inheritance becomes negligible and the
measured concentrations reflect the period of exposure after the rockfall event. The calculated
ages (Table 4.4) are thus minimum exposure ages of the rockfall.
Conclusions
Dating of the rockfall debris in Beacon Valley, Antarctica using SED is not straightforward,
as prior exposure cannot be excluded. In the case of prior exposure, the measured
concentrations of the cosmogenic nuclides are the sum of inherited nuclides and nuclides that
were produced after the rockfall event. The inheritance depends on the original source locality
of the sampled surface (scenarios a, b and c). Scenario (a), where the sampled surface
originates from the top position, is not consistent with our data. For scenario (b), where the
sampled surface originates from 0.5m/1m (depending on the assumed erosion rate) up to a
depth of 4m within the rockwall, the measured concentrations are the sum of inherited
cosmogenic nuclides (saturation level for all nuclides, reached within the wall) and from
cosmogenic nuclides that were produced after the rockfall event. An exact calculation of the
inheritance is not possible, as too many parameters are unknown. If however the sampled
surface was lying at depths greater than 4m within the rockwall, then the calculated ages
presented in Table 4.4 are the minimum exposure ages of the rockfall. No definitive
interpretation and conclusion can be given at this preliminary stage of the study.
84
4.2 Cosmogenic nuclides in the Allan Hills nunatak,
Antarctica: An approach to reconstruct the local
glacial history9
Abstract
Two different generations of glacial deposits observed in the Allan Hills, Antarctica, have
been dated by means of surface exposure dating, using cosmogenic 10Be, 26Al and 21Ne.
Samples of the Sirius Group formation, at altitudes above 1’710m.asl, yield minimum
exposure ages of 2.31 ± 0.31Ma. Analyses on surfaces of the latest significant expansion of
the East Antarctic Ice Sheet into the Allan Hills nunatak indicate ice-free conditions there
since about 100ka. The beginning of this expansion is assumed to have happened at about
200ka ago. This advance has produced a penetration depth into the nunatak of < 1’500m with
a minimum ice thickness of 60m. The latest significant advance of the East Antarctic Ice
Sheet into the Allan Hills nunatak is comparable in amplitude and chronology to Taylor II
advances of Taylor Glacier into Arena Valley, Inner Dry Valleys (Brook and Kurz, 1993b;
Marchant et al., 1993a).
Introduction
The Allan Hills nunatak (AHN), Antarctica (77°S and 168°W), is a unique archive for glacial
deposits from an expanded East Antarctic Ice Sheet (EAIS). Gentle slopes rising from the
northern Manhaul Bay of the nunatak towards the south (Fig. 4.7) enable the southward flow
of ice (Drewry, 1982; Nishio and Annexstad, 1979). Soft Beacon sandstone, which is
outcropping all over the nunatak, offers a large reservoir of source material for glacial
deposits. At least two different generations of glacial events are observable. In the AHN the
northernmost outcrop of Sirius Group deposits is found (see Borns unpublished, referenced in
Denton et al., 1991), a key formation in the discussion about ice sheet dynamics of Antarctica.
The dating of these deposits is crucial to answer the question about a melt down of EAIS
during global Pliocene warming (e.g. Miller and Mabin, 1998 and references therein). The
Sirius Group is mapped all along the Transantarctic Mountains (Fig. 1 in Hall et al., 1997),
but absolute dating is difficult. Harwood and Webb (1998) and Webb et al. (1984) use
diatoms from Sirius Group sediments to determine a Pliocene deposition age. This indeed is
questioned by e.g. Stroeven et al. (1998), who believe that the diatoms are windblown and
therefore not in situ.
9
Publication in preparation with following authorship: Tschudi, S., Schäfer, J.M., Borns, H., Ivy-Ochs, S.,
Barrett, P., Kubik, P.W., and Schlüchter, C.
85
Surface exposure dating (SED), using cosmogenic nuclides sheds light on the age of Sirius
Group sediments. Recent studies have produced ages up to 10Ma (Schäfer et al., 1999) for
Sirius Group clasts on Mt. Fleming, which clearly contradict a Pliocene deposition. So far,
age determination for Sirius Group deposits is restricted to the Inner Dry Valleys (Brook et
al., 1995, Bruno et al., 1997, Ivy-Ochs et al., 1995; Ivy-Ochs et al., 1997; Schäfer et al., 1999;
Tschudi et al., submitted). No age information of Sirius Group deposits is available for the
AHN.
In the AHN, the Sirius Group deposits are found on the northern slope towards Manhaul Bay
and on a plateau to the west (“Sirius West” in Fig. 4.7). They consist mainly of thin grayish
and semi-lithified tillite overlain by numerous boulders with a wide variety in size and
lithology (i.e. local dolerite and Beacon sandstone, but also foreign granite). Weathering is
prominent and boulders disintegrate in situ. Most of the boulders are just overlying the drift,
but some are partly embedded. This let us assume that the original drift was thicker than it is
today, and that we probably observe an erosional relict.
Besides Sirius Group deposits, other glacial sediments of a significant glacial advance into
AHN are observed. These deposits are stratigraphically younger than the Sirius Group
deposits, as they consist partly of re-worked Sirius Group sediments re-deposited along north
to south trending small-scale flute-like deposits. These features are observed close to the
present ice front of EAIS and on Beacon Sandstone cliffs that are covered by Sirius Group
drift.
On the other hand, the advancing EAIS eroded soft Beacon Sandstone bedrock and redeposited it as boulder trains, push moraines or other glacigenic features (e.g. as single
boulders or clusters of erratic material). The fingerprints of this glacial advance are found up
to an altitude of 1’705m.asl, which is about 60 meters above of the present ice-surface. The
trending of deposition and therefore of the paleo-flow direction is generally towards the
south. A similar flow direction is also derived form striae and rat-tails, which are found on
bedrock surfaces near the present limit of the EAIS and the central dolerite dyke. The
determined paleo flow direction is in accordance with the present flow direction of the EAIS
near AHN (from contour lines in Drewry, 1982). No absolute age information is available
about this advance.
86
Fig. 4.7: Schematic and aerial overview of the sampling area at the Allan Hills nunatak (AHN) with
sample locations. The East Antarctic Ice Sheet (EAIS) had earlier extended to the south onto the
nunatak. The thin gray lines correspond to bedrock cliffs of varying heights (2m up to 10m). Note that
the scale is only qualitatively estimated.
87
The goal of this study was to date the Sirius Group and the younger advance of EAIS into
AHN by means of surface exposure dating with cosmogenic 10Be, 26Al and 21Ne. This
chronology could provide direct age information of the northernmost outcrop of Sirius Group
deposit. Dating the young advance could yield information on the latest significant expansion
of EAIS into AHN. This local glacial dynamics may also correlate to other glacial events
within the Dry Valleys (e.g. expansion of Taylor Glacier into Arena Valley). Two field
campaigns to AHN were initiated for sampling the Sirius Group and younger glacial deposits.
First preliminary results are presented in the following.
Methods: General and sample description
In the superarid landscape of Antarctica surface features are preserved for millions of years
(see e.g. Denton et al., 1993; Schäfer et al., 1999) and the dating of glacial formations by
means of surface exposure dating is often not a straightforward approach. Both, loose rock
and bedrock samples may have experienced prior exposure and periods of unknown coverage
e.g. by ice. The measured amount of cosmogenic nuclides is likely to represent a complex
history of exposure. The major source of the Allan Hills samples is local bedrock (soft
sandstone, either fine grained or coarse grained up to conglomerate). If we assume long-term
stability of the landscape, this bedrock should be in saturation for cosmogenic nuclides. This
saturation concentration depends on the original depth at which the samples were buried
before excavation and on the steady state erosion rate of the actual lithology. However, both
parameters are unknown and data interpretation is thus not unambiguous. The measurement
of more than one cosmogenic nuclide, the subsequent analysis within the two-nuclide plot
(26Al/10Be versus 10Be) and a consideration of muonic contribution to the production of
cosmogenic nuclides (Heisinger, 1998) may shed light onto this problematic interpretation.
The sample set from AHN is used to contribute to this discussion.
The sample set is divided into several subsets. The older glacial event is looked at with two
samples (227 and AL9704), whereas for the younger advance ten samples (225, 228, 232,
AL9708, AL9709, AL9710, AL9711a, AL9711b, AL9713a, AL9713b, AL9713c and
AL9714) are available. Additional samples (AL9702, AL9711c and AL9712b) were collected
to look the coverage of the AHN by an expanded EAIS. Detailed sample description is found
in the following and in the appendix of this chapter (Table 4.6). The sample locations are
marked in Fig. 4.7.
Samples of the Sirius Group
For the dating of the Sirius Group deposit, we sampled clasts, which were either overlying the
Sirius Group drift or which were partly buried within the deposit. The results should yield
minimum ages for the Sirius Group. We have chosen two boulders of the western Sirius
88
Group sediments for our SED analysis (“Sirius West” on Fig. 4.7). Sample 227 is a rather
small (< 1m3), well-rounded and hard quartzite. The surface is very compact and weathering
is not obvious. Sample 9704 is a coarse-grained sandstone boulder that is embedded into the
Sirius Group drift to a depth of about 10cm. The sampled top surface is about 30cm above
ground level. The total size of the boulder is ~1m3. Weathering and erosion has to be
assumed, because the lithology seems quite soft.
Samples of the young advance
The younger glacial event offers a large variety of glacial deposits and suitable rock surfaces
for SED. All samples are assumed to represent erosional (i.e. bedrock excavation) and
depositional (i.e. single erratic or boulder train) subglacial features (Table 4.6). Sample 225
originates from a sandstone boulder, which is member of a boulder train 200m south of the
main dyke. The sample site is close to the present day’s cliff and we assume a short
transportation distance. It is a soft and coarse grained Beacon sandstone with coarse quartz
grains. Sample 228, a coarse-grained sandstone boulder, was found within a push-moraine
deposit at about 200m distance from the present ice front at Manhaul Bay. We assumed that
the young glacial advance deposited this moraine. The material of the deposit indeed may
consist of reworked boulders that originate from the ancient surface of AHN towards the
EAIS. Sample 232 is smooth and fine-grained sandstone. The sample is taken from striated
bedrock at the main east-west trending dyke. The striae are interpreted to originate from the
latest significant advance of the EAIS into AHN. Because the striae are still visibly fresh,
little erosion can be expected.
Samples AL9708 and AL9709 originate from a large-scale boulder train that has been
deposited from 349° to 169° azimuths on the eastern flank of the AHN (Fig. 4.8). The flutelike deposit is traceable over a length of about 150m. For sampling, we have chosen the two
largest boulders of this deposit. The lithology is fine-grained sandstone that shows weathering
and erosional features. Sand is partly surrounding the clasts. Transport distance is about
300m. Sample AL9710 is an excavated bedrock sample. The advancing ice eroded bedrock
material from the cliff and excavated a fresh outcrop. Inheritance by prior exposure is likely,
as the depth of prior burial is only about 30cm. The geological position of the samples
AL9711a and AL9711b are similar to AL9708 and AL9709 (i.e. from a boulder train).
Samples AL9711a and AL9711b are taken from small (< 1m3) and coarse grained sandstone.
This outcrop was the southernmost observed fingerprint of the young advance during our field
survey.
89
Fig. 4.8: Large-scale sandstone boulder train, from which samples AL9708 and AL9709 were taken.
Note the human at the end of the alignment as scale (black arrow).
The sample set AL9713a, AL9713b and AL9713c was taken from a huge boulder (about
100m3) that was lifted to a new position on a bedrock cliff (see also Fig. 4.11 for close up) to
the east of the investigated area. The source of the boulder lies a few tens of meters to the
north. The boulder’s sedimentological texture (e.g. fore-set structures and fining upward
sequences) and the weathering features (e.g. erosional forms, color and oxidization) were
used to determine the original top surface within the old cliff. This was done to prevent
sampling the previous (old) surface, as this position could bear massive inherited cosmogenic
nuclides. We have taken two samples from the top of the boulder (AL9713a and AL9713b)
and one sample (AL9713c) from a shielded position underneath the boulder, where only little
exposure (about 1% of the top value of boulder AL9713) is expected. This “background
amount” of cosmogenic nuclides shall help to crosscheck the minimum exposure age of the
top samples AL9713a and AL9713b. Erosion has to be assumed, as loose sand is lying
underneath the boulder. Sample AL9714 is similar to the sample set AL9713, but from a
smaller boulder (> 6m3) that is atop of another huge boulder (> 40m3).
90
Miscellaneous samples
Additional samples have been taken to provide information about the duration of coverage of
the AHN near Manhaul Bay by the most recent advance of the ice sheet. Coverage by ice
would reduce the cosmic ray flux by a factor of 100 for 20m of ice atop a sample and
radioactive decay reduces already accumulated concentration of 10Be and 26Al. The smallest
traceable period of coverage depends on the half-life of the radionuclides used. In our case,
the shorter half-life of 26Al is relevant (with t½ = 716ka). If we assume a measurement
uncertainty of about 10%, then a period longer than about 100ka would be recognized by this
approach. Sample AL9702 was taken at a close distance (< 100m) to the present ice position
at the edge of the gentle main depression of AHN towards Manhaul Bay in the north. The
clast is supposed to be in situ, because no signs of re-working by the young advance were
visible. However, re-working cannot be excluded. The boulder is of similar lithology as
sample 227 (hard and fine-grained quartzite) and is of about 1m3 in size. Sample AL9712b, a
hard quartzite cobble, was taken close to the site of AL9710, at the center of the AHN. The
size of the clast does not exclude re-working by a glacial advance. Both samples, AL9702 and
AL9712b, are expected to have been covered by the ice-sheet of the young advance. Sample
AL9711c was taken within the group of the samples AL9711a and AL9711b and is of the
same coarse-grained sandstone lithology.
Quartz was separated according to Kohl and Nishiizumi (1992) and the chemical separation
of 10Be and 26Al was done following Nishiizumi et al. (1989), Kohl and Nishiizumi (1992)
and Ivy-Ochs (1996). The concentrations of the cosmogenic nuclides 10Be and 26Al were
of the Paul Scherrer Institut (PSI) and the ETH. For the noble gas analysis, the quartz was
handpicked from the chemically separated quartz fraction. The processing for the 21Ne
analysis followed Bruno (1995). It was performed on a mass spectrometer (MS) at the
Institute of Isotope Geology and Mineral Resources (IGMR) of the ETH Zurich.
Results
All results are shown in the appendix of this chapter, Table 4.7 and Table 4.8. All data are
shown in a two-nuclide plot (26Al/10Be versus 10Be, Fig. 4.9). Within the errors, samples
AL9708, AL9709, AL9710, AL9711a, AL9712b, AL9713a, AL9713b, AL9714, 225 and 232
plot below the erosion saturation line, which indicates that periods of burial or episodic loss
of mass occurred. The results are discussed as follows:
91
Sirius Group
The samples 227 and AL9704 yield minimum exposure ages ranging from 0.80Ma (10Be
exposure age of sample 227) to 2.90Ma (26Al exposure age of sample AL9704)(see appendix
of this chapter, Table 4.7 and Table 4.8). For sample AL9704, 26Al is close to saturation,
which means that this age is very sensitive to uncertainties in the determination of the total
amount of Al. If the total amount of Al would be reduced from 42ppm to 41ppm (which is
within the ICP-AES variability at low concentrations), the exposure age would be reduced to
a value of 2.46 + 1.28 / - 0.55Ma. The 26Al exposure age of AL9704 was therefore not
considered for the conclusions (see also section: Al aliquot and ICP-AES in Chapter 2). The
upper limit of the interval of minimum ages of the Sirius Group is then given by the 21Ne
exposure age of sample AL9704 (2.31 ± 0.31Ma). The 10Be and 21Ne ages of AL9704 are
consistent within errors, whereas 227 has a slightly higher 21Ne age compared to 10Be and
26
Al. Increased 21Ne concentrations are not in contradiction to lower 10Be and 26Al
concentrations, as all analyzed samples are sedimentary rock. This means that cosmogenic
21
Ne is inherited from the whole “lifetime” of the sampled surface. Deep burial of the sample
in its earlier “life” would not affect the stable nuclide, whereas 10Be and 26Al would have
decayed by radioactive processes.
Fig. 4.9: Two-nuclide plot (26Al/10Be versus 10Be) of all measured samples, except AL9713c. All
concentrations are normalized to sea level and high altitude, using the production rates of (Kubik et al.,
1998) and the scaling of (Lal, 1991). Error bars include 1σ error of the AMS measurement and a 5%
uncertainty for possible chemical processing variability (Ivy-Ochs, 1996).
92
Young advance
The minimum exposure ages for the samples of the young advance are presented in Fig. 4.10
and the appendix of this chapter (Table 4.7 & Table 4.8). Their ages range from
0.12 ± 0.01Ma (sample AL9713a, 10Be) to 1.94 ± 0.26Ma (sample 228, 21Ne). The eastern
samples (AL9708, AL9709, AL9713a, AL9713b, and AL9714) cluster below 200ka, whereas
the western samples (AL9711a and AL9711b) lie between 0.78 ± 0.09Ma and 0.95 ± 0.09Ma.
Samples 225, 228, 232, AL9710 and AL9711a show large differences between their
minimum exposure ages (especially if 21Ne ages are considered).
Fig. 4.10: Minimum exposure ages for the “young-advance” samples from the Allan Hills nunatak
AL9713c is excluded (see text).
Discussion
Sirius Group samples
Sirius Group at AHN bears a minimum exposure age of 2.31 ± 0.31Ma. A reworking of the
sampled surfaces seems unlikely, because AL9704 is partially embedded in the ground. The
10
Be and 21Ne data are consistent within the errors. This indicates a simple exposure history
without periods of significant burial (e.g. by advanced ice). Furthermore, no geologic traces
of the young advance (e.g. boulder trains, striations and erratics) were observed at altitudes of
the sampled Sirius Group clasts (i.e. at altitudes above 1’730m.asl). We therefore conclude
that the area around the sampled deposit was a nunatak during the last significant glacial
93
advance of EAIS into AHN. Thus, our ages reflect true minimum ages of the Sirius Group
within the AHN.
Young advance samples
Minimum exposure ages from samples of the young advance show a wide range, with
different sub-groups being distinguishable using the two-nuclide plot. Note that this division
does not correspond to differences between the lithologies (see Table 4.6). The results are
therefore discussed according to the group division. Simple model calculations were used in
the reconstruction of the exposure histories of the samples. Note that these models allow a
wide variety of solutions. Field evidence (described below) was used to support the likelihood
of the selected scenario.
(a) AL9708, AL9709, AL9713a, AL9713b, AL9713c, AL9714, 225 and 232
Single erratic boulders, as well as boulders from boulder trains yield similar results. In the
two-nuclide plot all samples cluster within errors below the “erosion island” (Fig. 4.9).
According to (Nishiizumi et al., 1991), data below the “erosion island” indicate periods of
burial and coverage from cosmic radiation (e.g. coverage by advancing ice). Note that
Nishiizumi et al. only consider spallation as a major production mechanism. Heisinger et al.
(1997), Heisinger (1998) and Heisinger and Nolte (2000) offer a new point of view, as they
state that data points below the “erosion island” could also indicate that the sampled surfaces
originate from certain depths within bedrock, where low saturation concentrations due to
muons for 26Al and 10Be are reached before excavation (here: by a glacial erosive advance)
followed by exposure at the surface (see figure 5.10 on p.109 in Heisinger, 1998).
Sample-set AL9713 justifies a detailed discussion, because two samples (samples a&b) were
taken from the top of this boulder and one sample (sample c) was taken from underneath the
boulder at a shielded position (Fig. 4.11). Field evidence indicates burial depths of about 1m
for samples AL9713a&b and about 2m for AL9713c, respectively (derived from the boulder’s
sedimentological texture, such as fore-set structures and graded bedding, and the weathering
features, such as erosional forms, color and oxidization). The availability of three samples of
the same boulder but from various depths should allow estimating both steady state erosion
and exposure time after excavation. Model calculations for 10Be and 26Al concentrations were
preformed with a period of prior exposure and erosion, followed by exposure and erosion at
the surface. Prior exposure was taken as saturation at the depth taken from field evidence.
Erosion before and after excavation was taken to be at the same rate. A “starting value” of
about 100ka for the exposure time after excavation was estimated from Fig. 4.12. The model
parameters (erosion and exposure time) were varied until the deviation of the calculated 10Be
and 26Al concentrations to the measured values were minimum (< 3% for 10Be and < 8% for
94
26
Al). These conditions were established with a steady state erosion of about 200cm/m.y.,
which is rather high if compared to other erosion rate estimates in Antarctic environments
(Brook et al., 1995; Ivy-Ochs et al., 1997; Nishiizumi et al., 1993; Schäfer et al., 1999;
Tschudi et al., submitted). However, the lithology of AL9713 is a rather weak coarse-grained
sandstone, which could allow higher erosion rate.
Fig. 4.11: Close up of AL9713. Lines mark the stratigraphic bedding of the Beacon sandstone. The
ancient top surface is at the backside of the boulder (arrow). This original top surface was determined
by weathering features and sedimentological textures. From that, original burial depths of the top
samples AL9713a&b (about 1m) and the bottom AL9713c (about 2m) were derived.
The determined period of exposure after excavation is 100ka. The measured 21Ne
concentration does however not fit in this best fit model. Inheritance of 21Ne, produced during
the “lifetime” of the sediment, could be an explanation. However, one should keep in mind
the uncertainties of the calculated concentrations of 5-10% due to uncertainties of the
production rates. On the other hand a detailed investigation into the effect of scaling with Lal
(1991) or with either Dunai (2000) or Stone (2000), which affects the production of 10Be and
26
Al differently from that of 21Ne due to muon production could change the production rate
ratios of e.g. 10Be/21Ne and thus affect the accumulated (and modeled) concentrations.
95
Fig. 4.12: Modified two-nuclide plot without erosion after Heisinger (1998). The dotted lines mark
depth of burial before excavation and the solid lines mark periods of exposure after excavation.
Sample group (a) clusters around the line of 100ka of exposition with varying burial depths (50mwe to
2mwe, which equals to about 20m to 0.5m of bedrock material). Erosion changes this figure in such a
way that the data points correspond to smaller burial depth and longer exposure times (see Fig. 5.9 on
p. 108 in Heisinger, 1998). The data points were re-scaled with reference production rates of
Nishiizumi et al. (1989) to accommodate the use of the figures of Heisinger (1998).
All data of subset (a) (except the shielded sample AL9713c) are shown in Fig. 4.12. They
cluster along the same exposure curve as sample AL9713. If we assume the simplest case,
then all samples correspond to the same glacial event. They would therefore have the same
exposure age and show roughly the same erosion rate but would come from different depths
before excavation. Sample 225, with an increased 26Al/10Be ratio, fits also into this picture, if
erosion is considered. The estimated exposure period reflects the time since deglaciation from
the sample’s position, as we assume subglacial deposition of subset (a).
(b) AL9710 and 228
From field evidence, the sampled surface AL9710 was buried at a depth of about 40cm before
excavation (observable excavation at the edge of the cliff). A best fit (< 3% deviation to the
measured values) for 10Be and 26Al is reached (cf. modeling of sample set AL9713) for a
steady state erosion rate of 180cm/Ma used during prior exposure and an exposure time of
100ka after excavation. The calculated 21Ne concentration again, reaches only about 80% of
96
the measured concentration with these model parameters. Inheritance from prior exposure
during the “lifetime” of the sampled sediment can be a reason.
Sample 228, a clast in a push moraine related to the young advance close to the present ice
sheet, seems to be an outlier. 10Be and 26Al agree with each other, but the concentration of
21
Ne is higher compared to 10Be and 26Al. We assume the exposure of this surface is not only
related to the young advance, but that it is a reworked boulder of the ancient nunatak surface.
The sample is therefore discussed in the section “Miscellaneous samples”.
(c) AL9711a and AL9711b
Two samples of small boulders were taken at the southernmost outcrop of the “young
advance” deposits. Exposure ages of both samples cluster between 0.78 ± 0.09Ma and 0.95 ±
0.09Ma. The samples plot inside the “erosion island” of the two-nuclide plot (Fig. 4.9), which
indicates a simple exposure history (Klein et al., 1986; Lal, 1991; Nishiizumi et al., 1991). A
correlation with an exposure period of 100ka is possible, if we apply the same simple model
calculation, as for sample AL9713. A best fit for these two samples (< 1% deviation to the
measured values, except 10Be for AL9711a, which has a large deviation of 16%) is reached
for shallow burial at 5cm within bedrock with an estimated erosion rate of 50cm/Ma.
In a summary, all samples that are closely related to the young advance yield a common
history of exposure, where prior exposure at certain depths (i.e. buried in bedrock) is followed
by excavation (i.e. glacial erosion) and subsequent re-exposure to cosmic radiation. The time
of re-exposure (about 100ka) corresponds to the time of glacial retreat, if we assume a
subglacial deposition of the sampled features, which means that the sampled surfaces were
shielded from cosmic radiation until the retreat of the ice.
Miscellaneous samples
This sample set (AL9702, 228, AL9711c and AL9712b) should determine the duration of the
young advance, which ended at about 100ka ago, because these surfaces were covered by ice
and therefore shielded from cosmic radiation. Sample 228, taken from a “young advance”
deposit is likely to be a reworked ancient surface of the AHN. Its geomorphic position is
therefore similar to AL9702. Samples 228 and AL9711c cluster in the two-nuclide plot (Fig.
4.9). This indicates a similar exposure history for these samples. 10Be and 26Al ages agree,
whereas 21Ne exposure ages are higher. Sample AL9712b plots on the steady state erosion
line of the “erosion island”. This sample could have a different exposure history than the
other samples. Moreover, 10Be and 26Al ages are not corresponding. Note that the 26Al
exposure age of AL9702 could not be determined due to problems with the ICP-AES
measurement.
97
The subgroup “miscellaneous samples” yields corresponding radionuclide ages (except
AL9712b) for different lithologies (hard quartzite, weak coarse-grained sandstone) at
different locations within the AHN that was covered by an expanded EAIS. The thickness of
ice is assumed to have been at least 60 meters, the difference in elevation between the present
ice front and the highest outcrop of “young advance” deposits. This thickness would reduce
the production of cosmogenic nuclides to about 0.1% of the surface value. Production during
coverage is therefore negligible compared to the measured concentrations. Any long-term
coverage (longer than 100ka) would lead to a decreased 26Al concentration compared to 10Be,
due to the different half-lives of the two nuclides. Corresponding exposure ages and positions
within the “erosion island” of the two-nuclide plot therefore indicate a period of coverage
shorter than 100ka. Sample AL9712b shows slightly lower 26Al concentration. The model
calculations (cf. sample AL9713) were refined with a period of burial by thick ice, which
results in a period with negligible production of new cosmogenic nuclides. Best fit for all
cosmogenic nuclides (< 1% for 10Be and 26Al and < 6% for 21Ne) corresponds to a prior
exposure time of less than 3Ma, due to the assumption that the sample was already at the
surface before the ice advanced (this however is speculative, as the sample is a small clast of
<< 1m3, where reworking by advancing ice cannot be excluded. Then, no information about
past location, exposure and shielding is available). The model erosion rate is rather small
(5cm/Ma), but it corresponds to the hard quartzite lithology. For the time of burial, a period of
about 100ka is calculated. This time however, is close to the time span that our method can
distinguish.
Conclusion
For the northernmost outcrop of Sirius Group deposits within the Transantarctic Mountains a
minimum exposure age of 2.31 ± 0.31Ma is calculated. This minimum age suggests a Late
Pliocene or older age for the Sirius Group deposits at the AHN. Such a conclusion is in
accord with recent age estimations based on cosmogenic nuclide analyses from clasts of the
Sirius Group within the Dry Valleys (Brook et al., 1995; Bruno et al., 1997; Ivy-Ochs, 1996;
Ivy-Ochs et al., 1997; Schäfer et al., 1999; Tschudi et al., submitted). No erosion was taken
into account for the age calculation, although such processes are observed on boulder
AL9704. If erosion would be considered, the exposure age would be older than the minimum
exposure ages given here.
SED analyses performed on clasts, which are related to the younger advance of EAIS into
AHN, indicate that the northern part of AHN was covered with ice up to an altitude of about
1705m.asl during an unknown time period until about 100ka ago. This age is based on model
calculations on a boulder, which was excavated and deposited by this young advance. Prior
exposure and erosion was taken into account for these calculations. The depth of burial, at
98
which the samples were buried before excavation, was determined by field evidence. When
the data of the young advance samples are plotted on the modified two-nuclide plot of
Heisinger (1998), they cluster on the same exposure curve. This curve describes prior
exposure at different depths with subsequent exposure at about 100ka. Samples AL9710,
AL9711a and AL9711b do also fit this model, if shallow burial depths are assumed.
The subgroup “miscellaneous samples” yields corresponding radionuclide ages for different
lithologies at different locations within the area of AHN that was covered by the advanced
ice. This excludes coverage of the samples during periods longer than 100ka. This conclusion
is also confirmed for sample AL9712b, where 26Al is slightly lower than 10Be. In summary,
our data indicate a glacial history shown in Table 4.5.
Table 4.5: Sketched glacial history of Allan Hills nunatak.
Glacial event
Chronology
Deposition of Sirius Group sediments
> 2.3Ma
?
?
Advancing EAIS into northern part of Allan Hills nunatak
= about 200ka
Ice covered nunatak
= about 100ka
Retreat form advanced position
about 100ka
The Allan Hills nunatak was repeatedly covered by ice of the EAIS during the past. Our data
indicate that it is ice-free since about 100ka. The duration of this latest significant overriding
of the nunatak is = about 100ka, because most of the samples were covered by ice but yield
corresponding exposure ages for 10Be and 26Al and their position within the two-nuclide plot
may be explained by other processes than periods of burial (e.g. prior exposure at shallow
depth and episodic erosion). Previous SED analyses from a bedrock transect in the Allan Hills
(Nishiizumi et al., 1991) also indicate that no significant periods of burial occurred, if an
extended production model is applied (see p. 109 in Heisinger, 1998). Heisinger estimates the
period of exposure (which equals to the ice-free period) with Nishiizumi et al. data (1989) at
about 500ka. However, no erosion was considered in this reevaluation.
In Arena Valley, Taylor II stage moraines of Taylor Glacier, a small outlet glacier of a
peripheral Taylor Dome of EAIS, were dated with SED to a mean minimum exposure age of
117 ± 51ka (Brook and Kurz, 1993b; Brook et al., 1993a). This dating is based on production
rates of Nishiizumi et al. (1989). Re-evaluation with the values of Kubik et al. (1998) would
slightly increase the mean age to about 120ka. This age is still a minimum exposure age for
the Taylor II moraine in Arena Valley. It corresponds to a depositional age of the moraine and
a correlation to MIS-5 is suggested by (Brook and Kurz, 1993b). The geometric expansion of
99
Taylor II advance into Arena Valley is up to 750m to the south of present day’s Taylor
Glacier and up to an elevation of 200m above present ice front. Field observation in the Allan
Hills nunatak indicates that the latest significant advance of EAIS into the nunatak reached
altitudes up to 1710m.asl at a distance of about 1.5km from the present ice front. This
expansion is similar to the Taylor II advance of Taylor Glacier into Arena Valley. However,
more field investigation and mapping is needed to obtain definitive geometric parameters of
the expansion. The presented exposure scenario suggests a time correlation of the expansion
of EAIS into the Allan Hills nunatak with Taylor II moraine in Arena Valley. However, a
correlation with MIS-5 is not definite as the advance could have also occurred during MIS-6.
More data along bedrock transects and depth profiles are needed to confirm or reject the
postulated young and short period of ice expansion in the Allan Hills.
Acknowledgements
This work was funded by Swiss National Science Foundation grants 21-043469.95 and 21053942.98 and by the United States Division of Polar Programs of the National Science
Foundation. Logistical support in Antarctica was given by Antarctica New Zealand, the Royal
New Zealand Air Force and the US-NAVY VXE-6. We thank the Zurich tandem and noble
gas crew for keeping the MS systems in great shape.
100
Appendix
Table 4.6: Description of Allan Hills nunatak samples. The samples are divided into sub-groups (Sirius
Group, young advance and miscellaneous).
Sample
Lithology
Altitude
Thick-
Quartz [g]
Comment
[m.asl] ness [cm]
227
Quartzite
1’730
1.25
9.11
Clast
AL9704
Coarse sandstone
1’745
1.5
7.04
Embedded clast
225
Coarse sandstone
1’670
1.5
5.45
Clast, single erratic
228
Coarse sandstone
1’670
0.8
7.88
Clast, out of push moraine
232
Fine sandstone
1’635
1.0
5.84
Bedrock, striated
AL9708
Fine sandstone
1’690
1.0
5.42
Clast, boulder train
AL9709
Fine sandstone
1’705
1.0
5.28
AL9710
Fine sandstone
1’675
3.0
4.52
Bedrock, excavated
Young
AL9711a
Coarse sandstone
1’705
1.3
7.62
advance
AL9711b
Coarse sandstone
1’705
1.0
7.13
AL9713a
Coarse sandstone
1’665
4.0
10.58
AL9713b
Coarse sandstone
1’665
3.5
2.05
AL9713c
Coarse sandstone
1’660
3.0
1.53
AL9714
Coarse sandstone
1’595
3.5
1.88
AL9702
Quartzite
1’645
0.5
5.05
Clast
AL9711c
Coarse sandstone
1’705
0.8
7.89
Bedrock
AL9712b
Quartzite
1’635
0.5
3.03
Clast
Sirius Group
Miscellaneous
101
Table 4.7: AMS measured 10Be and 26Al concentrations of the Allan Hills nunatak, with calculated
minimum exposure ages according to the equation given in (Lal, 1991). Prior exposure (see text) is not
included here. We used production rates of pBe = 5.75 ± 0.24 and pAl = 37.4 ± 1.9 atoms/ yr·g SiO2 at
latitudes ≥60° and sea level (Kubik et al., 1998), which were scaled for elevation and latitude
according to (Lal, 1991). The production rates are corrected for the given sample thickness. Correction
for shielding is included. The error includes a 1σ AMS measurement error for sample and blank and a
5% uncertainty for possible chemical processing variability (Ivy-Ochs, 1996). The age uncertainties
include the uncertainties of the reference production rates (Kubik et al., 1998).
Sample
10
Be [107
Error
Atoms/g SiO2] [%]
227
1.61
6.8
Total Al
[ppm]
26
Al [107
Atoms/g SiO2]
243
8.79
Error
[%]
6.9
10
Be minimum age
[106 years]
0.80
26
Al minimum age
[106 years]
±0.08
0.82
+0.11
-0.10
AL9704
3.30
5.3
42
15.21
5.1
2.13
+0.26
2.90
-0.23
225
1.03
6.5
141
4.51
6.7
0.50
228
2.24
6.3
130
11.38
6.6
1.29
±0.04
+0.14
*
-0.78
0.36
1.40
-0.13
±0.04
+0.28
-0.22
232
0.76
7.9
2839
3.09
9.1
0.37
±0.04
0.24
±0.03
AL9708
0.43
6.4
467
2.31
7.6
0.19
±0.02
0.17
±0.02
AL9709
0.41
9.8
331
2.15
6.4
0.18
±0.02
0.15
±0.01
AL9710
0.51
6.1
70
3.05
5.8
0.23
±0.02
0.23
±0.02
AL9711a
1.82
5.6
112
8.37
5.1
0.95
+0.09
0.78
-0.08
+0.09
-0.08
AL9711b
1.58
6.6
90
8.56
5.2
0.80
±0.08
0.81
±0.09
AL9713a
0.26
6.8
99
1.66
7.1
0.12
±0.01
0.12
±0.01
AL9713b
0.30
9.1
201
1.63
6.4
0.13
±0.01
0.12
±0.01
AL9713c
0.02
60.2
255
0.12
18.3
G
AL9714
0.33
11.2
455
1.84
7.7
0.16
AL9702
2.75
8.1
?
?
?
1.78
G
±0.02
+0.27
G
0.14
?
-0.24
AL9711c
2.65
5.3
114
12.72
5.1
1.58
+0.16
3.08
6.7
86
11.62
5.6
2.29
+0.35
-0.30
±0.01
?
?
1.71
-0.15
AL9712b
G
+0.38
-0.28
1.67
+0.38
-0.28
*26Al in saturation, no age calculation possible
G See text for details concerning age calculation of AL9713c
? The total Al content of sample AL9702 could not be determined due to problems during the ICP-AES measurement.
Therefore, no 26Al exposure age could be calculated.
102
Table 4.8: MS measured 21Ne concentrations for boulders in the Allan Hills, Antarctica, with
calculated minimum exposure ages. We used a production rate of pNe = 20 atoms/ yr·g SiO2 at latitudes
> 50° and sea level with an uncertainty of 13% (Niedermann et al., 1994 updated by Niedermann, pers.
comm.), which was scaled for elevation and latitude according to (Lal, 1991). The errors are given
within 1σ confidence level, including statistical, sensitivity, mass-discrimination errors and errors due
to uncertainties of calibration gas amounts with a maximum value of 3%. The age errors include
furthermore the uncertainty of the reference production rate.
21
Sample
Ne
Error [%]
7
Ne Age
[106 years]
[10 Atoms/g SiO2]
227
21
9.60
4.8
1.05 ± 0.15
21.36
4.1
2.31 ± 0.31
225
7.58
10.3
0.87 ± 0.14
228
16.95
3.4
1.94 ± 0.26
232
7.67
4.0
0.90 ± 0.12
AL9710
3.12
6.4
0.36 ± 0.05
AL9713a
1.39
7.2
0.16 ± 0.02
AL9702
22.39
4.1
2.61 ± 0.36
AL9711c
17.13
3.8
1.91 ± 0.26
AL9712b
20.11
5.0
2.43 ± 0.34
Al9704
CHAPTER 5:CONCLUSION AND OUTLOOK
103
Chapter 5
Conclusion and outlook
The chronological classification of glaciations in a wide range of geographical and geological
regions, based on surface exposure dating, is the core of this work. The worldwide
distribution of study areas may be confusing to the reader but it has to be seen from the
perspective of the project “Chronology of Northern and Southern Hemisphere Glaciations “
(Swiss National Foundation, #21-043469.95), in which this thesis is embedded. It was shown
that SED is indeed a suitable tool for attacking these dating problems and that the key to
successful SED studies lies in careful fieldwork. This does not only mean that all relevant and
sometimes seemingly irrelevant details are to be noted during actual sampling, but it also
requires careful studies of the local and regional geological setting. This is most important
because the relation of the sampled surface to the geological event (here: glaciations) has to
be clearly understood. Only then can the concentrations of cosmogenic nuclides yield
absolute ages of the sampled geologic event.
Northern Hemisphere
The Younger Dryas Salpausselkä I formation is a clearly defined glacigenic deposit in
southern Finland. Boulders atop this formation are related to the ice front at the time of
deposition. Knowledge of the geologic environment, like lake level changes, isostatic rebound
of the crust or littoral processes is included in the study and enables a clear relation of the
minimum exposure age to the geologic process. The postulated Younger Dryas age (Donner,
1978) was confirmed (Tschudi et al., 2000). But this work was not only a pure “dating
exercise”. It is a significant methodological progress. The Salpausselkä I study was the first
SED study to be performed on samples with a short period of exposure of a few thousand
years at very low elevation (below 200m.asl). The Salpausselkä I study has proven that SED
is a robust and suitable dating tool for glacigenic Quaternary deposits, even where low
concentration of 10Be and 26Al are to be expected. On the other hand this does also mean that
rather young exposure times at higher elevations can be measured. An example: For a
104
Holocene moraine deposited 5’000 years ago at 2’000m.asl one would need about 80g of
quartz to be measured, well less than the 160g for the Salpausselkä study (see Chapter 2, Fig.
2.2).
Whether the eastern Arctic region was significantly covered by ice during the Last Glacial
Maximum is brought into question by radiocarbon dating of mammoth teeth on Wrangel
Island, East Siberia, which yield ages for mammoths living during LGM (Sulerzhitsky and
Romanenko, 1999; Vartanyan et al., 1995). The method of SED was chosen to clarify this
discussion and to obtain absolute chronological information about past glaciations on this
island. The island lacks glacial deposits, which could be related to a significant glaciation
(e.g. coverage during LGM) (Karhu et al., 2000). Thus, SED samples were taken from
outcropping bedrock and a single boulder, to determine the period of time since the last
deglaciation. A strict relation from the sample to the glaciation cannot be given. However, the
data indicate that the island was ice-free since at least 65ka ago (Karhu et al., 2000). Shortterm ice coverage cannot be excluded, although 26Al analyses are not indicating significant
coverage (see Appendix to Chapter 3.2).
Boulders from moraines from the eastern margin of the Tibetan Plateau (Litang area) were
sampled to date distinct glacial positions (Schäfer et al., submitted). The data scatter below
20ka, even if erosion of the rock surfaces (obvious from field observations) is considered. It
could therefore be shown that limited glacial advances occurred in eastern Tibet during
MIS-2. Whether these advances correspond to the LGM cannot be determined by dating
alone. Local geologic mapping in the investigated valleys is needed for final identification
and correlation. Kanding data cluster around the Younger Dryas time period as defined from
Greenland ice cores (Alley et al., 1993; Johnsen et al., 1992). However, this correlation is
speculative, as only two boulders were analyzed yet and little is known about the local
geologic and paleo glaciologic setting. Nonetheless, the hypothesis of Kuhle (1994, 1998),
where a huge ice-sheet covering the entire Tibetan Plateau until the end of the last glacial
cycle is suggested, is in contradiction with the presented data sets from Litang and Kanding.
Data from the Tanggula area, central Tibet, indicate that there was no such ice coverage since
at least 180ka ago.
In the Northern Hemisphere, young glaciations (up to the penultimate glaciation) were
investigated. To date these glaciations terminal glacigenic formations were sampled (except
Wrangel Island), which reflect distinct glacial stages of ice sheets or valley glaciers. The
sampled surfaces were taken from boulders lying on these deposits and they are therefore
clearly related to these positions. Though, the determined minimum exposure ages are
minimum ages for these glacial stages. The samples originate from an active environment,
105
where fast erosional and depositional processes steadily modify the landscape. In Tibet, the
high relief assures a constant supply of fresh material out of rockfalls. In Scandinavia, ice
coverage and therefore subglacial processes dominated the recent geologic past. A relict
landscape that would bear high cosmogenic inheritance is rare in these areas. Significant prior
exposure, which would lead to inheritance of cosmogenic nuclides in the measured samples
seems therefore not very probable. The data should reflect close to true minimum exposition
ages of the sampled surface.
The relation of a local glacigenic archive (here: a terminal glacigenic formation) to the global
climate system has to be shown first before a correlation with global climate changes can be
accepted. Every single archive has its own “transfer function” from global input (e.g.
changing solar radiation, changing ocean circulation) into local output (e.g. advancing
glaciers). All archives therefore reflect climatic changes in different ways and this reflection
is time dependent itself. An example: an arbitrary terminus moraine is thought to be a
fingerprint of the YD cold reversal. How can we prove this? Of course, we can investigate the
archive (here: the moraine) and try to date it. But, what is the relation of the archive to the YD
climate event? How long did it take for the glacier to establish the investigated stable
position? What is the time lag of its formation to the YD cold reversal minimum temperature,
as recorded in Greenland ice cores, for example? To solve these problems, a dating campaign
should always be based on more than one single dating method. If this is not possible (and
this might be the usual case), then one has to be careful with correlating the investigated
glacial stage into the global context.
Southern Hemisphere
The investigated areas of the Southern Hemisphere are located in one of the least understood
regions on Earth with respect to geologic processes: Antarctica. There, the landscape is likely
to be preserved for millions of years (see e.g. Denton et al., 1993; Schäfer et al., 1999; Sugden
et al., 1999; Summerfield et al., 1999). Fast erosional and depositional processes are unlikely
and geologic deposits, such as glacigenic formations may be preserved almost untouched for
a long time. Cold and arid desert conditions dominate disintegration and erosion of the
outcropping formations. Geologic work in Antarctica demands another time dimension in the
understanding of these ongoing landforming processes.
Dating Antarctic glaciations by means of SED is different from dating glaciations in regions,
where moderate climate conditions occur. Relict glacial deposits are common (e.g. Sirius
Group deposits or Granite drift), but “classic” terminal formations are missing, because coldbased glaciers dominate the depositional regime. For SED, we have to rely on boulders that
are lying on drifts to determine exposure ages. This kind of dating is based on the principle of
106
superposition, where overlying stratigraphic units are said to be younger than underlying
strata. It was applied on Granite drift clasts in Beacon Valley and on clasts lying on Sirius
Group deposits in the Allan Hills. However, the sampled clasts of both studies yield close
relation to the glacial advances that are to be dated: Within Beacon Valley, granite is foreign
(i.e. erratic) and its presence is just due to glacial processes and in the Allan Hills, boulders
were partially embedded within glacial deposits. Corresponding exposure ages from multinuclide analyses (10Be, 26Al and 21Ne) indicate simple exposure histories without periods of
significant shielding for both studies. The dated Granite drift therefore represents the last
major glacial advance of Taylor Glacier into central Beacon Valley. Moreover, the study
yields indirect chronological information about the Sirius Group deposits on the nearby Mt.
Feather and it supports dating of the remnant body of ice (Sugden et al., 1995b), questioned
by Hindmarsh et al. (1998).
The dating of the younger glacial advance in the Allan Hills confronted SED with a key
problem: prior exposure of the sampled surface. This is mainly due to the slow Antarctic
landscape development, where no running water and vegetation is present. This implies that
the “exchange rate of landscape”, i.e. the process of erosion and deposition, is much smaller
than under moderate climate conditions. Glacial debris, like reworked soft bedrock is likely to
carry inherited cosmogenic nuclides, as the landscape did not significantly change during the
time window considered. In other words, the analyzed debris originates from shallow
positions at nearby outcropping bedrock. These outcrops are likely to have existed for a long
period of time, which is sufficient to build up significant concentrations of cosmogenic
nuclides. It has been shown that this process is not necessarily a disadvantage for data
interpretation. The application of Heisinger’s model, where muonic contribution to the
production of cosmogenic nuclides is considered (Heisinger, 1998), together with careful field
observation enabled the estimation of prior burial depths and the evaluation of the time of
exposure of the samples after deposition. As the samples were taken from subglacial deposits,
this period represents the time since deglaciation. Furthermore, the duration of ice coverage
by this advance was tackled with samples from inside the former ice expansion. There, prior
exposure and subsequent coverage must have occurred. Indeed, the samples yield
corresponding exposure ages for 10Be and 26Al, which indicates that no significant periods of
coverage happened. Therefore, the period of coverage was too short (= 100ka) to be
recognized by means of SED. Information from the stable 21Ne analysis has to be handled
with care, because the sampled rocks may have experienced even a more complex exposure
history, where periods of exposure, burial and re-exposed occurred for several times. The
radioactive nuclides 10Be and 26Al may have decayed during these repeated processes,
whereas 21Ne is accumulated.
107
The presented Southern Hemisphere investigations were concentrated in the Dry Valleys,
Antarctica. Within two localities, Beacon Valley and Allan Hills, several generations of
glacial deposits were dated. The older generation, the Sirius Group deposits, is thought to
have been deposited before the down-cutting of the Dry Valleys occurred (Denton et al.,
1993). The age of this deposit is crucial for the discussion about the past behavior of the EAIS
(see e.g. Miller and Mabin, 1998). Our data showed that the minimum exposure age of these
formations lies at 4.0Ma in the Beacon Valley and at 2.3Ma in the Allan Hills. Recent SED
studies, based upon stable nuclides 21Ne and 3He, yield ages of up to 10Ma for Sirius Group
clasts on Mt. Fleming (Schäfer et al., 1999).
The deposition of Granite drift and the young advance into the Allan Hills are likely to have
happened into a landscape that was already quite similar to today’s. Paleo-flow direction
derived from striae and rat-tails indicate similar flow directions like the present ice movement
in the Allan Hills (Drewry, 1982; Nishio and Annexstad, 1979). Granite drift, Beacon Valley,
is thought to be deposited as an ablation till on the floor and eastern slope of the valley by an
advanced Taylor Glacier (Sugden et al., 1995b). From the perspective of an ice sheet, both
localities, Beacon Valley and Allan Hills nunatak, act like an “overflow basin” of East
Antarctic Ice Sheet (EAIS): A thickening of ice would cause an expansion into these ice-free
areas. The feeding sources of these two systems are likely to be closely coupled, as both
belong to the East Antarctic Ice Sheet. Furthermore, the distance between the two “overflow
basins” is only about 130km. A connection between the Allan Hills and Beacon Valley /
Arena Valley is therefore likely. However, a drift, similar to the Beacon Valley Granite drift,
has not yet been found in the Allan Hills. Young and rather small expansions of Taylor
Glacier into Beacon Valley and Arena Valley have on the other hand been observed and even
partly dated by means of SED (Brook and Kurz, 1993b; Brook et al., 1993a). However, other
than in the Allan Hills, more than ten different advances of Taylor Glacier into Arena Valley
are traced (Marchant et al., 1993a). One of these, Taylor II advance, is of special interest, as it
yields a minimum exposure age of about 120ka (Brook et al., 1993a, recalculated with Kubik
et al., 1998). This age is roughly in agreement with the determined age of the young advance
into the Allan Hills less than 200ka ago with a subsequent deglaciation at about 100ka ago.
Moreover, both advances have similar geometric dimensions. Taylor II advance is suggested
to be correlative to MIS-5 (Brook and Kurz, 1993b). However, a correlation of the young
advance at the Allan Hills nunatak with MIS-5 is not definite, as the advance could have also
occurred during MIS-6. A correlative link to the global climate system is beyond the data’s
capability.
However, it is clear that the EAIS was not just a static and stable mass of ice, but that it
showed small and relatively short-termed oscillations during the past several hundred
108
thousands years. These oscillations are not comparable in expansion with a Sirius advance or
a Granite drift advance, of course. But nonetheless, Antarctica seems to react to global
signals, although the coupling mechanisms are not yet fully understood.
Outlook
By now, SED is an established method for dating geologic surface features. It offers the
unique possibility to date Quaternary deposits, from moraines of Younger Dryas age to
boulders related to Pliocene glaciations. Ivy-Ochs (1996) said that the method is very robust
during processing. This statement was again proven, as samples with very low concentrations
of cosmogenic nuclides and of different lithologies were successfully processed. The
Salpausselkä study for example showed that SED is robust towards low 10Be and 26Al
concentrations in large samples. This implies that the practical age limit for SED, which used
to be about 10ka at high altitudes, is much less now. Holocene moraines (e.g. deposited 5ka
ago at 2’000m.asl at mid-latitudes) are now dateable with SED. The study on the Earth’s
climate system and its driving forces is currently focused on the questions of how ice ages are
caused. Studies on the last glacial cycle (i.e. intensive mapping and precise absolute dating of
distinct glacial deposits) are key for this understanding. Whether the hemispheres reacted
synchronous or asynchronous to each other is still open, but crucial. Within the ongoing
studies, SED is supposed to act as a dating tool, providing absolute glacial chronologies,
where no other dating methods are usable. However, SED results still yield an uncertainty of
up to 10%, which is rather large with respect to questions about the reaction times of glaciers
towards global climate changes (e.g. Younger Dryas cold reversal or LGM). Exact
synchronicity or small time lags between the hemispheres, for example, may not yet be
determined with SED. The main source of uncertainty is still in the determination of the
correct local production rate of cosmogenic nuclides. Improving the scaling functions will
definitively make SED a more precise tool of absolute dating (i.e. decrease of uncertainty
from 10% to 5% as a goal). Global correlations of glacial advances can then be performed and
even local and small scale correlations may be identified by means of SED (e.g. different
substages of one single glacial advance or a progressive advance of the front of an ice sheet).
Production rate experiments and investigations of the scaling functions are underway (e.g.
Schäfer, 2000). Therefore, present day SED data have to be presented in a way that these
improvements can easily be incorporated. However, the key to successful SED is not only in
methodological improvements, but also in the understanding of the method, which implies (1)
appropriate sample strategies and (2) an understanding of the relation of the sample to the
glacial feature that shall be dated. If theses conditions are fulfilled, one should be encouraged
to further use this beautiful method of dating glacigenic formations for the benefit of
Quaternary earthscience!
CHAPTER 6: REFERENCES
109
Chapter 6
References
Aartolahti, T., 1970, Fossil ice-wedges, tundra polygons and recent frost cracks in southern Finland,
Annales Academicae Scientiarum Fennicae, v. 107, p. 26.
Alley, R.B., Meese, D.A., Shuman, C.A., Gow, A.J., Taylor, K.C., Grootes, 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.
Alley, R.B., and Clark, P.U., 1999, The deglaciation of the Northern Hemisphere: A global
perspective: Annual Reviews of Earth and Planetary Sciences, v. 27, p. 149-189.
Alverson, K., and Oldfield, F., 2000, Past global changes and their significance for the future, an
introduction, in Alverson, K.D., Oldfield, F., and Bradley, R.S., eds., Past global changes and their
significance for the future, Quaternary Science Reviews, v. 19, p. 1-5.
Bard, E., and Kromer, B., 1995, The Younger Dryas: absolute and radiocarbon chronology, in
Troelstra, S.R., van Hinte, J.E., and Ganssen, G.M., eds., The Younger Dryas, p. 161-166.
Benn, D.I., and Owen, L.A., 1998, The role of the Indian summer monsoon and the mid-latitude
westerlies in Himalayan glaciations: review and speculative discussion.: Journal of Geological
Society, London, v. 155, p. 353-363.
Benxing, Z., and Rutter, N., 1998, On the problem of Quaternary glaciations, and the extent and
patterns of Pleistocene ice cover in the Qinghai-Xizang (Tibet) Plateau: Quaternary International, v.
45/46, p. 109-122.
Bierman, P.R., Marsella, K.A., Patterson, C., Davis, P.T., and Caffee, M., 1999, Mid-Pleistocene
cosmogenic minimum-age limits for pre-Wisconsinan glacial surfaces in southwestern Minnesota
and southern Baffin Island; a multiple nuclide approach: Geomorphology, v. 27, p. 25-39.
Björck, Kromer, Johnson, Bennike, Hammarlund, Lemdahl, Possnert, Rasmussen, Wohlfarth,
Hammer, and Spurk, 1996, Synchronized Terrestrial-Atmospheric Deglacial Records Around the
North Atlantic: Science, v. 274, p. 1155-1160.
Bodén, P., Fairbanks, R.G., Wright, J.D., and Burckle, L.H., 1997, High-resolution stable isotope
records from southwest Sweden: The drainage of the Baltic Ice Lake and Younger Dryas ice margin
oscillations: Paleoceanography, v. 12, p. 39-49.
Brady, H., and McKelvey, B., 1979, The interpretation of a Tertiary tillite at Mount Feather, southern
Victory Land, Antarctica: Journal of Glaciology, v. 22, p. 189-193.
Brady, H., and McKelvey, B., 1983, Some aspects of the Cenozoic glaciation of southern Victoria
Land, Antarctica: Journal of Glaciology, v. 29, p. 343-349.
Brook, E.J., Kurz, M.D., Ackert, R.P., Denton, G.H., Brown, E.T., Raisbeck, G.M., and Yiou, F.,
1993a, Chronology of Taylor Glacier advances in Arena Valley, Antarctica, using in situ
cosmogenic 3He and 10Be: Quaternary Research, v. 39, p. 11-23.
Brook, E.J., and Kurz, M.D., 1993b, Using in situ 3He in Antarctic quartz sandstone boulders for
surface exposure chronology: Quaternary Research, v. 39, p. 1-10.
110
Brook, E.J., Brown, E.T., Kurz, M.D., Ackert, R.P., Raisbeck, G.M., and Yiou, F., 1995, Constraints
on age, erosion, and uplift of Neogene glacial deposits in the Transantarctic Mountains determined
from in situ cosmogenic 10Be and 26Al: Geology, v. 23, p. 1063-1066.
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. 207-210.
Brown, E.T., Edmond, J.M., Raisbeck, G.M., Yiou, F., Kurz, M.D., and Brook, E.J., 1991,
Examination of surface exposure ages of Antarctic moraines using in situ produced 10Be and 26Al:
Geochimica et Cosmochimica Acta, v. 55, p. 2269-2283.
Brown, E.T., Brook, E.J., Raisbeck, G.M., Yiou, F., and Kurz, M.D., 1992, Effective attenuation
lengths of cosmic rays producing 10Be and 26Al in quartz: Implications for exposure age dating:
Geophysical Research Letters, v. 19, p. 369-372.
Bruno, L., 1995, Datierung von Moränen und glazial überprägten Oberflächen in der Antarktis mit in
situ produzierten kosmogenen Edelgasen [Ph.D. thesis]: Zurich, ETH Zurich, 140 p.
Bruno, L.A., Baur, H., Graf, T., Schlüchter, C., Signer, P., and Wieler, R., 1997, Dating of Sirius
Group tillites in the Antarctic Dry Valleys with cosmogenic 3He and 21Ne: Earth and Planetary
Science Letters, v. 147, p. 37-54.
Burbank, D.W., and Cheng, K.J., 1991, Relative dating of Quaternary moraines, Rongbuk Valley,
Mount Everest, Tibet: Implications for an ice sheet on the Tibetan Plateau: Quaternary Research, v.
36, p. 1-18.
Cato, I., 1987, On the definitive connection of the Swedish time scale with the present, Sveriges
Geologiska Undersökning, v. 68: Uppsala, p. 55.
Cerling, T.E., and Craig, H., 1994, Geomorphology and in situ cosmogenic isotopes: Annual Review
Earth Planetary Sciences, v. 22, p. 273-317.
Chu, W.K., Mayer, J.W., and Nicolet, M.A., 1978, Backscattering Spectrometry: Orlando, Academic
Press, 384 p.
Clayton, J.L., and Megahan, W.F., 1986, Erosional and chemical denudation rates in the southwestern
Idaho batholith: Earth surface processes and landforms, v. 11, p. 389-400.
Cuffey, K.M., Conway, H., Gades, A.M., Hallet, B., Lorrain, R., Severinghaus, J.P., Steig, E.J.,
Vaughn, B., and White, J.W.C., 2000, Entrainment at cold glacier beds: Geology, v. 28, p. 351-354.
Denton, G.H., Prentice, M.L., and Burckle, L.H., 1991, Cainozoic history of the Antarctic ice sheet, in
Tingey, R.J., ed., The geology of Antarctica: Oxford, Oxford University Press, p. 365-433.
Denton, G.H., Sugden, D.E., Marchant, D.R., Hall, B.L., and Wilch, T.I., 1993, East Antarctic ice
sheet sensitivity to Pliocene climatic change from a Dry Valleys perspective: Geografiska Annaler,
v. 75A, p. 155-204.
Denton, G.H., and Hendy, C.H., 1994, Younger Dryas age advance of Franz Josef Glacier in the
Southern Alps of New Zealand: Science, v. 264, p. 1434-1437.
Denton, G.H., and Hendy, C.H., 1995, Age of the Waiho Loop glacial event: A reply: Science, v. 271,
p. 868-869.
Derbyshire, E., Yafeng, S., Jijun, L., Benxing, B., Shijie, L., and Jingtai, W., 1991, Quaternary
glaciation of Tibet: The geological evidence: Quaternary Science Reviews, v. 10, p. 485-510.
Donner, J.J., 1966, The Late-glacial and early Post-glacial pollen stratigraphy of southern and eastern
Finland: Societas Scientiarum Fennica, Commentationes Biologicae, v. 29, p. 1-24.
Donner, J.J., 1969, Land/sea level changes in southern Finland during the formation of the
Salpausselkä endmoraines: Bull. Geol. Soc. Finland, v. 41, p. 135-150.
Donner, J.J., 1978, The dating of the levels of the Baltic Ice Lake and the Salpausselkä moraines in
South Finland: Commentationes Physico-Mathematicae, v. 48, p. 11-38.
Donner, J.J., 1982, Fluctuations in water level of the Baltic Ice Lake, Annales Academiae Scientiarum
Fennicae, v. 134, p. 13-28.
Drewry, D.J., 1982, Ice flow, bedrock, and geothermal studies from radio-echo sounding inland of
McMurdo Sound, Antarctica, in Craddock, C., Loveless, J.K., Vierima, T.L., and Crawford, K.A.,
eds., Antarctic Geoscience, v. Series B, Nr 4: Madison, International union of geological sciences,
p. 977-983.
Dunai, T., 2000, Scaling factors for production rates of in situ produced cosmogenic nuclides: a
critical reevaluation: Earth and Planetary Science Letters, v. 176, p. 157-169.
Emeis, K.C., Anderson, D.M., Doose, H., Kroon, D., and Schulz-Bull, D., 1995, Sea-surface
temperatures and the history of monsoon upwelling in the Northwest Arabian Sea during the last
500'000 years: Quaternary Research, v. 43, p. 355-361.
Ekman, I.M., and Ilyin, V., 1991, Deglaciation, the Younger Dryas end moraines and their correlation
in the Karelian ASSR and adjacent areas, in Rainio, H., and Saarnisto, M., eds., Eastern
Fennoscandian Younger Dryas and moraines, v. 32: Espoo, Geological Survey of Finland, p. 73-99.
Felzer, B., Webb, T., and Oglesby, R.J., 1998, The impact of ice sheets, CO2, and orbital insolation on
late Quaternary climates: Sensitivity experiments with a general circulation model: Quaternary
Science Reviews, v. 17, p. 507-534.
Finkel, R.C., and Suter, M., 1993, AMS in the earth science: Technique and applications, Advances in
Analytical Geochemistry, v.1, p. 1-114.
Gasse, F., Arnold, M., Fontes, J.C., Fort, M., Gibert, E., Huc, A., Bingyan, L., Yuanfang, L., Qing, L.,
Mélières, F., Van Campo, E., Fubao, W., and Qingsong, Z., 1991, A 13'000-year climate record
from western Tibet: Nature, v. 353, p. 742-745.
Gillespie, A., and Molnar, P., 1995, Asynchronous maximum advance of mountain and continental
glaciers: Reviews of Geophysics, v. 33, p. 311-364.
Glückert, G., 1977, On the Salpausselkä ice-marginal formations in southern Finland: Zeitschrift für
Geomorphologie, v. 27, p. 79-88.
Glückert, G., 1995, The Salpausselkä End Moraine in southwestern Finland, in Ehlers, J., Kozarski, S.,
and Gibbard, P., eds., Glacial Deposits in North-East Europe: Rotterdam, Balkema, p. 51-56.
Gosse, J.E., Evenson, E.B., Klein, J., Lawn, B., and Middleton, R., 1995, Precise cosmogenic 10Be
measurements in western North America. Support for a global Younger Dryas cooling event:
Geology, v. 23, p. 877-880.
Graham, I.J., Barry, B.J., Ditchburn, R.G., and Whitehead, N.E., 1999, Validation of cosmogenic
nuclide production rate systematics through direct measurement, Abstract of Accelerator mass
spectrometry conference (AMS-8): Vienna. p. 185.
Grosswald, M.G., 1988, An Antarctic-style ice sheet in the Northern hemisphere: Toward a new
global glacial theory: Polar Geography and Geology, v. 12, p. 239-267.
Grosswald, M.G., 1997, Former glaciation, in Kotlyakov, V.M., ed., World Atlas of Snow and Ice
Resources: Moscow, Russian Academy of Sciences, p. 385-386.
Gualtieri, L., Glushkove, O., and Brigham-Grette, J., 2000, Evidence for restricted ice extent during
the last glacial maximum in the Koryak Mountains of Chukotka, far eastern Russia: GAS Bulletin,
v. 112, p. 1106-1118.
Häberli, W., Bösch, H., Scherler, K., Ostrem, G., and Wallén, C.C., 1989, World glacier inventory,
Status 1988.
Hall, B.L., Denton, G.H., Lux, D.R., and Schlüchter, C., 1997, Pliocene paleoenvironment and
Antarctic ice sheet behavior: evidence form Wright Valley: The Journal of Geology, v. 105, p. 285294.
Harwood, D.M., and Webb, P.N., 1998, Glacial transport of Diatoms in the Antarctic Sirius Group:
Pliocene refrigerator: GSA Today, v. 8, p. 1-8.
Heisinger, B., Niedermayer, M., Hartmann, F.J., Korschinek, G., Nolte, E., Morteani, G., Neumeier,
S., Petitjean, C., Kubik, P., Synal, A., and Ivy-Ochs, S., 1997, In situ production of radionuclides at
great depths: Nuclear Instruments and Methods in Physics Research, v. B 123, p. 341-346.
Heisinger, B.P., 1998, Myonen-induzierte Produktion von Radionukliden: [Ph.D. thesis]: München,
Technische Universität München, 153 p.
Heisinger, B., and Nolte, E., 2000, Cosmogenic in situ production of radionuclides: Exposure ages and
erosion rates, Nuclear Instruments and Methods, v. B 172, in press.
Hindmarsh, R.C.A., Van der Wateren, F.M., and Verbers, A.L.L.M., 1998, Sublimation of ice through
sediment in Beacon Valley, Antarctica: Geografiska Annaler, v. 80A, p. 209-219.
Hofmann, H.J., Beer, J., Bonani, G., von Gunten, H.R., Raman, S., Suter, M., Walker, R.L., Wölfli,
W., and Zimmermann, D., 1987, 10Be: Half-life and AMS-standards: Nuclear Instruments and
Methods in Physics Research, v. B29, p. 32-36.
Hooghiemstra, H., and van der Hammen, T., 1995, A Younger Dryas climatic event in Columbia, in
Troelstra, S.R., van Hinte, J.E., and Ganssen, G.M., eds., The Younger Dryas, p. 47-48.
111
112
Hopkins, D.M., Brigham-Grette, J., Gualtieri, L., Glushkova, O., Ivanov, V.F., Lozhkin, A., Heiser,
P.A., and Hamilton, T.D., 1998, The consensus view: Limited glacial ice extent across N.E. Russia
during LGM: AGU, 1998 Fall Meeting. Eos, Transactions, American Geophysical Union, v. 79, p.
F491.
Hyvärinen, H., 1972, Flandrian regional pollen assemblage zones in eastern Finland: Societas
Scientiarum Fennica, Commentationes Biologicae, v. 59, p. 25.
Hyvärinen, H., 1973, The deglaciation history of eastern Fennoscandia - recent data from Finland:
Boreas, v. 2, p. 85-102.
Hyvärinen, H., 1975, Late glacial paleogeography of Fennoscandia: Terra, v. 87, p. 155-166.
Ivy-Ochs, S., Schlüchter, C., Kubik, P.W., Dittrich-Hannen, B., and Beer, J., 1995, Minimum 10Be
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.
Ivy-Ochs, S., 1996, The dating of rock surfaces using in situ produced 10Be, 26Al, and 36Cl, with
examples from Antarctica and the Swiss Alps [Ph.D. thesis]: Zurich, ETH Zurich, 197 p.
Ivy-Ochs, S., Schlüchter, C., Kubik, P.W., Beer, J., and Kerschner, H., 1996, The exposure age of an
Egesen moraine at Julier Pass measure with the cosmogenic radionuclides 10Be, 26Al and 36Cl:
Eclogae Geologicae Helvetiae, v. 89, p. 1049-1063.
Ivy-Ochs, S., Schlüchter, C., Prentice, M., Kubik, P.W., and Beer, J., 1997, 10Be and 26Al Exposure
Ages for the Sirius Group at Mount Fleming, Mount Feather and Table Mountain and the Plateau
Surface at Table Mountain, in Ricci, C.A., ed., The Antarctic Region: Geological Evolution and
Processes: Siena, Terra Antartica Publication, p. 1153-1158.
Ivy-Ochs, S., Schlüchter, C., Kubik, P., and Denton, G., 1999, Moraine exposure dates imply
synchronous Younger Dryas glacier advances in the European Alps and in the Southern Alps of
New Zealand: Geografiska Annaler, v. 81A, p. 313-323.
Johnsen, S.J., Clausen, H.B., Dansgaard, W., Fuhrer, K., Gundestrup, N., Hammer, C.U., Iversen, P.,
Jouzel, 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., Petet, J.R., Barkov, N.I., Barnola, J.M., Chappellaz, J., Ciais, P., Kotkyakov, V.M., Lorius,
C., Petrov, N., Raynaud, D., and Ritz, C., 1991, The last deglaciation in Antarctica: evidence for a
"Younger Dryas" type climatic event, in Bard, E., and Broecker, W.S., eds., Absolute and
Radiocarbon Chronologies, v. 2: NATO ASI Series: Berlin, Springer-Verlag, p. 229-266.
Karhu, J.A., Tschudi, S., Saarnisto, M., Kubik, P., and Schlüchter, C., 2000, New constraints for latest
glacial advances on Wrangel Island, Arctic Ocean, from rock surface exposure dating: Global and
Planetary changes, in press.
Klein, J., Giegengack, R., Middleton, R., Sharma, P., Underwood, J.R., and Weeks, R.A., 1986,
Revealing histories of exposure using in situ produced 26Al and 10Be in Libyan desert glass:
Radiocarbon, v. 28, p. 547-555.
Knauer, M., 1994, Analyse der kosmogenen Radionuklide 10Be, 14C und 26Al in Steinmeteoriten
mittels Beschleunigermassenspektrometrie [Ph.D. thesis]: Hannover, Universität Hannover, 181 p.
Kohl, C.P., and Nishiizumi, K., 1992, Chemical isolation of quartz for measurement of in situproduced cosmogenic nuclides: Geochimica et Cosmochimica Acta, v. 56, p. 3583-3587.
Kos'ko, M.K., Cecile, M.P., Harrison, J.C., Ganelin, V.G., Khandoshko, N.V., and Lopatin, B.G.,
1993, Geology of Wrangel Island, between Chukchi and East Siberian Seas, northeastern Russia:
Geological Survey of Canada, Bulletin, v. 461, p. 101.
Kubik, P.W., Ivy-Ochs, S., Masarik, J., Frank, M., and Schlüchter, C., 1998, 10Be and 26Al production
rated deduced from an instantaneous event within the dendro-calibration curve, the landslide of
Köfels, Ötz Valley, Austria: Earth and Planetary Science Letters, v. 161, p. 231-241.
Kuhle, M., 1988, The Pleistocene glaciation of Tibet and the onset of ice ages - An autocycle
hypothesis: GeoJournal, v. 17, p. 581-595.
Kuhle, M., and Herterich, K., 1989, On the ice age glaciation of the Tibetan Highlands and its
transformation into a 3-D model.: GeoJournal, v. 19.2, p. 201-206.
Kuhle, 1994, Present and Pleistocene Glaciation on the North-Western margin of Tibet between the
Karakorum main Ridge and the Tarim Basin, Supporting the Evidence of a Pleistocene Inland
Glaciation in Tibet: GeoJournal, v. 33, p. 133-272.
Kuhle, M., 1998, Reconstruction of the 2.4 Million km2 Late Pleistocene ice sheet on the Tibetan
Plateau and its impact on the global climate: Quaternary International, v. 45/46, p. 71-108.
Lal, D., 1958, Investigations of nuclear interactions produces by cosmic rays: [Ph.D. thesis] Bombay,
Tata Institute of Fundamental Research, 90 p.
Lal, D., and Peters, B., 1967, Cosmic ray produced radioactivity on the Earth, Handbuch der Physik,
v. 46/2, Springer, p. 551ff.
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.
Lehmkuhl, F., 1997, Late Pleistocene, late-Glacial and Holocene glacier advances on the Tibetan
Plateau: Quaternary International, v. 38-39, p. 77-83.
Lehmkuhl, F., 1998, Extend and spatial distribution of Pleistocene glaciations in eastern Tibet:
Quaternary International, v. 45/46, p. 123-134.
Lehmkuhl, F., and Haselein, F., 2000, Quaternary paleoenvironmental change on the Tibetan Plateau
and adjacent areas (Western China and Western Mongolia): Quaternary International, v. 65/66, p.
121-145.
Licciardi, J.M., Kurz, M.D., Clark, P.U., and Brook, E.J., 1999, Calibration of cosmogenic 3He
production rates from Holocene lava flows in Oregon, USA, and effects of Earth’s magnetic field:
Earth and Planetary Science Letters, v. 172, p. 261-271.
Linkletter, G., Bockheim, J., and Ugolini, F.C., 1972, Soils and glacial deposits in the Beacon Valley,
southern Victoria land, Antarctica: New Zealand Journal of Geology and Geophysics, v. 16, p. 90108.
Lowe, J.J., Coope, G.R., Lemdahl, G., and Walker, M.J.C., 1995, The Younger Dryas climate signal
in land records from NW Europe, in Troelstra, S.R., van Hinte, J.E., and Ganssen, G.M., eds., The
Younger Dryas, p. 3-25.
Lukashov, A.D., and Ekman, I.M., 1982, General information on geology and geomorphology of the
Karelian ASSR, 4-17, p. 46-56.
Makeyev, V.M., Arslanov, K.A., and Garutt, V.E., 1981, Age of Severnaya Zemlya mammoths and
some problems of Late Pleistocene paleogeography: Doklady-Earth Science Sections, v. 245, p. 5255.
Mangerud, J., Andersen, S.T., Berglund, B.E., and Donner, J.J., 1974, 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.I., 1993a, Miocene-Pliocene-Pleistocene glacial
history of Arena Valley, Quartermain Mountains, Antarctica: Geografiska Annaler, v. 75A, p. 269302.
Marchant, D.R., Denton, G.H., Sugden, D.E., and Swisher, C.C.I., 1993b, Miocene glacial
stratigraphy and landscape evolution of the western Asgard Range, Antarctica: Geografiska
Annaler, v. 75A, p. 303-330.
Marchant, D.R., Denton, G.H., Bockheim, J.G., Wilson, S.C., and Kerr, A.R., 1994, Quaternary
changes in level of the upper Taylor Glacier, Antarctica: implications for paleoclimate and East
Antarctic Ice Sheet dynamics: Boreas, v. 23, p. 29-43.
Masarik, J., and Reedy, R.C., 1995, Terrestrial cosmogenic-nuclide production systematics calculated
from numerical simulations: Earth and Planetary Science Letters, v. 136, p. 381-395.
Mayewski, P.A., and Goldthwait, R.P., 1985, Glacial Events in the Transantarctic Mountains: A
Record of the East Antarctic Ice Sheet, in Turner, M.D., and Splettstoesser, J.F., eds., Geology of
the Central Transantarctic Mountains Antarctic Research Series, v. 36, American Geophysical
Union, p. 275-338.
McHargue, L.R., and Damon, P.E., 1991, The global Beryllium-10 cycle: Reviews of geophysics, v.
29, p. 141-158.
McManus, D.A., and Creager, J.S., 1984, Sea-level data for parts of the Bering-Chukchi Shelves of
Beringia from 19’000 to 10’000 14C yr BP: Quaternary Research, v. 21, p. 317-325.
Miller, M.F., and Mabin, M.C.G., 1998, Antarctic Neogene landscapes - In the refrigerator or in the
deep freeze?: GSA Today, v. 8, p. 1-3.
Mörner, N.-A., 1980, The Fennoscandian Uplift: Geological Data and their Geodynamical
Implication, in Mörner, N.-A., ed., Earth Rheology, Isostasy And Eustasy: Chichester, John Wiley
& Sons, p. 251-285.
113
114
Niedermann, S., Graf, T., and Marti, K., 1993, Mass spectrometric identification of cosmic-rayproduced neon in terrestrial rocks with multiple neon components: Earth and Planetary Science
Letters, v. 118, p. 65-73.
Niedermann, S., Graf, T., Kim, J.S., Kohl, C.P., Marti, K., and Nishiizumi, K., 1994, Cosmic-rayproduced 21Ne in terrestrial quartz: the neon inventory of Sierra Nevada quartz separates: Earth and
Planetary Science Letters, v. 125, p. 341-355.
Niedermann, S., 2000, The 21Ne production rate in quartz revisited: Earth and Planetary Science
Letters, v. 183, p. 361-364.
Niemelä, J., 1971, Die quartäre Stratigraphie von Tonablagerungen und der Rückzug des Inlandeises
zwischen Helsinki und Hämeenlinna, Geological Survey of Finland Bull., v. 253: Otaniemi, 79 p.
Nishiizumi, K., Winterer, E.L., Kohl, C.P., Klein, J., Middleton, R., Lal, D., and Arnold, J.R., 1989,
Cosmic ray production of 10Be and 26Al in Quartz from glacially polished rocks: Journal of
geophysical research, v. 94, p. 17907-17915.
Nishiizumi, K., Kohl, C.P., Arnold, J.R., Klein, J., Fink, D., and Middleton, R., 1991, Cosmic ray
produced 10Be and 26Al in Antarctic rocks: exposure and erosion history: Earth and Planetary
Science Letters, v. 104, p. 440-454.
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 26Al in the study of diverse geomorphic
processes: Earth Surface Processes and Landforms, v. 18, p. 407-425.
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. B 101, p. 22'225-22'232.
Nishiizumi, K., Finkel, R.C., and Caffee, M.W., 1999, Measurements of absolute production rates of
in situ produced cosmogenic nuclides, Abstract of Accelerator mass spectrometry conference
(AMS-8): Vienna. p. 183.
Nishio, F., and Annexstad, J.O., 1979, Glaciological survey in the bare ice area near the Allan Hills in
Victoria Land, Antarctica, Fourth symposium on Antarctic meteorites, v. Special Issue. 15: Tokyo,
Memoirs of National Institute of Polar Research, p. 13-23.
Ochs, M., and Ivy-Ochs, S., 1997, 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 Physics Research, v. 123, p. 235-240.
Ohno, M., and Hamano, Y., 1992, Geomagnetic poles over the past 10000 years: Geophysical
Research Letter, 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
Geolelectricity, v. 45, p. 1455-1466.
Okko, M., 1962, On the development of the first Salpausselkä, West of Lahti, Bulletin de la
commission géologique de Finlande, v. 202: Helsinki, 162 p.
Ose, T., 1996, The comparison of the simulated response to the regional snow mass anomalies over
Tibet, Eastern Europe and Siberia: Journal of Meteorological Society of Japan, v. 74, p. 845-866.
Peteet, D.M., 1995, North American evidence for the Younger Dryas cooling-the last decade of
research, in Troelstra, S.R., van Hinte, J.E., and Ganssen, G.M., eds., The Younger Dryas, p. 37-46.
Phillips, F.M., Zreda, M.G., Gosse, J.C., Klein, J., Evenson, E.B., Hall, R.D., Chadwick, O.A., and
Sharma, P., 1997, Cosmogenic 36Cl and 10Be ages of Quaternary glacial and fluvial deposits of the
Wind River Range, Wyoming: Geological Society of America Bulletin, v. 109, p. 1453-1463.
Phillips, W.M., Sloan, V.F., Shroder, J.F., Sharma, P., Clarke, M.L., and Rendell, H.M., 2000,
Asynchronous glaciation at Naga Parbat, northwestern Himalaya Mountains, Pakistan: Geology, v.
28, p. 431-434.
Porter, S.C., and An, Z., 1995, Correlation between climate events in the North Atlantic and China
during the last glaciation.: Nature, v. 375, p. 305-308.
Potter, N., and Wilson, S.C., 1983, Glacial geology and soils in Beacon Valley: Antarctic Journal, v.
18, p. 100-103.
Punkari, M., 1980, The ice lobes of the Scandinavian ice sheet during the deglaciation in Finland:
Boreas, v. 9, p. 307-310.
Rainio, H., 1995, Large ice-marginal formations and deglaciation in southern Finland, in Ehlers,
Kozarski, and Gibbard, eds., Glacial Deposits in North-East Europe: Rotterdam, Balkema, p. 57-66.
Rainio, H., Saarnisto, M., and Ekman, I.M., 1995, Younger Dryas end moraines in Finland and NW
Russia: Quaternary International, v. 28, p. 179-192.
Rainio, H., 1996, Late Weichselian end moraines and deglaciation in eastern and central Finland:
Espoo, Geological Survey of Finland, 73 p.
Raymo, M.E., and Ruddiman, W.F., 1992, Tectonic forcing of late Cenozoic climate: Nature, v. 359,
p. 117-122.
Reedy, R.C., Tuniz, C., and Fink, D., 1994, Report on the Workshop on Production Rates of
Terrestrial In situ-produced Cosmogenic Nuclides, Nuclear Instruments and Methods, v. B 92, p.
335-339.
Richards, B.W.M., 2000, Luminescence dating of Quaternary sediments in the Himalayas and High
Asia: a practical guide to its use and limitations for constraining the timing of glaciation.:
Quaternary International, v. 65/66, p. 49-61.
Richards, B.W., Owen, L.A., and Rhodes, E.J., 2000, Timing of Late Quaternary glaciations in the
Himalayas of northern Pakistan.: Journal of Quaternary Science, v. 15, p. 283-297.
Rutter, N., 1995, Problematic ice sheets: Quaternary International, v. 28, p. 19-37.
Saarnisto, M., Karhu, J., and Vartanyan, S., 1998, The Wrangel Island expedition 1997: Searching for
mammoths, ice wedges and evidence of the extent of glaciations., Quaternary Environments of the
Eurasian North (QUEEN), Second QUEEN Workshop, v. 42: St. Petersburg, Russia.
Samworth, E.A., Warburton, E.K., and Engelbertink, G.A.P., 1972, Beta decay of the 26Al ground
state: Physical Review C, v. 5, p. 138-142.
Sarkar, A., Ramesh, R., Battacharya, S.K., and Rajgopalan, G., 1990, Oxygen isotope evidence for a
stronger winter monsoon current during the last glaciation.: Nature, v. 343, p. 549-551.
Saunders, I., and Young, A., 1983, Rates of surface processes on slopes, slope retreat and denudation:
Earth Surface Processes and Landforms, v. 8, p. 473-501.
Sauramo, M., 1918, Geochronologische Studien über die Spätglazialzeit in Südfinnland, Fennia, v. 41
(1), 44 p.
Sauramo, M., 1923, Studies on the Quaternary varve sediments in southern Finland, Fennia, v. 44 (1),
164 p.
Sauramo, M., 1929, The Quaternary Geology of Finland: Helsinki, Geologiska Kommissionen I
Finland, 110 p.
Sauramo, M., 1958, Die Geschichte der Ostsee, Annales Academiae Scientiarum Fennicae, v. 51, 522
p.
Schäfer, J.M., Ivy-Ochs, S., Wieler, R., Leya, I., Baur, H., Denton, G.H., and Schlüchter, C., 1999,
Cosmogenic noble gas studies in the oldest landscape on earth: surface exposure ages of the Dry
Valleys, Antarctica: Earth and Planetary Science Letters, v. 167, p. 215-226.
Schäfer, J.M., Baur, H., Denton, G.H., Ivy-Ochs, S., Marchant, D.R., Schlüchter, C., and Wieler, R.,
2000, The oldest ice on Earth in Beacon Valley, Antarctica: New evidence from surface exposure
dating: Earth and Planetary Science Letters, v. 179, p. 91-99.
Schäfer, J.M., 2000, Reconstruction of landscape evolution and continental paleoglaciations using in
situ cosmogenic nuclides. Examples from Antarctica and the Tibetan Plateau [PhD thesis]: Zürich,
ETH Zürich, 121 p.
Schäfer, J.M., Tschudi, S., Zhao, Z., Wu, X., Ivy-Ochs, S., Wieler, R., Baur, H., Kubik, P.W., and
Schlüchter, C., submitted, The limited influence of glaciations in Tibet on global climate over the
past 180'000 years: Science.
Shackleton, N.J., 1987, Oxygen isotopes, ice volume and sea level: Quaternary Science Reviews, v. 6,
p. 183-190.
Small, E.E., Anderson, R.S., Repka, J.L., and Finkel, R., 1997, Erosion rates of alpine bedrock summit
surfaces deduced from in situ 10Be and 26Al: Earth and Planetary Science Letters, v. 150, p. 413425.
Sternberg, R., 1996, Paleomagnetic averages of geomagnetic latitude: Radiocarbon, v. 38, p. 169-170.
Stone, J.O.H., Evans, J.M., Fifield, L.K., Allan, G.L., and Cresswell, R.G., 1998, Cosmogenic
chlorine-36 production in calcite by muons: Geochimica et Cosmochimica Acta, v. 62, p. 433-454.
Stone, J., 2000, Air pressure and cosmogenic isotope production: Journal of Geophysical Research, v.
105, p. 23'753-23'759.
115
116
Stroeven, A.P., Burckle, L.H., Kleman, J., and Prentice, M., 1998, Atmospheric transport of diatoms
in the Antarctic Sirius Group: Pliocene deep freeze: GSA Today, v. 8, p. 1-5.
Strömberg, B., 1990, A connection between the clay varve chronologies in Sweden and Finland,
Annales Academiae Scientiarum Fennicae, v. 154: Helsinki, 32 p.
Sugden, D.E., Denton, G.H., and Marchant, D.R., 1995a, Landscape evolution of the Dry Valleys,
Transantarctic Mountains: Tectonic implication: Journal of Geophysical Research, v. 100, p. 9'9499'967.
Sugden, D.E., Marchant, D.R., Potter, N., Souchez, R.A., Denton, G.H., Swisher, C.C., and Tison, J.L., 1995b, Preservation of Miocene glacier ice in east Antarctica: Nature, v. 376, p. 412-414.
Sugden, D.E., Summerfield, M.A., Denton, G.H., Wilch, T.I., McIntosh, W.C., Marchant, D.R., and
Rutford, R.H., 1999, Landscape development in the Royal Society Range, southern Victory Land,
Antarctica; stability since the mid-Miocene: Geomorphology, v. 28, p. 181-200.
Sulerzhitsky, L.D., and Romanenko, F.A., 1999, The "twilight" of the mammoth fauna in the Asiatic
Arctic: Ambio, v. 28, p. 251-255.
Summerfield, M.A., Stuart, F.M., Cockburn, H.A.P., Sugden, D.E., Denton, G.H., Dunai, T., and
Marchant, D.R., 1999, Long-term rates of denudation in the Dry Valleys, Transantarctic Mountains,
southern Victoria Land, Antarctica based on in situ-produced cosmogenic 21Ne: Geomorphology, v.
27, p. 113-129.
Synal, H.-A., Bonani, G., Döbeli, M., Ender, R.M., Gartenmann, P., Kubik, P.W., Schnabel, C., and
Suter, M., 1997, Status report of the PSI/ETH AMS facility: Nuclear Instruments and Methods, v. B
123, p. 62-68.
Thompson, L.G., Mosley-Thompson, E., Davis, M.E., Bolzan, J.F., Dai, J., Yao, T., Gundestrup, N.,
Wu, X., Klein, L., and Xie, Z., 1989, Holocene-Late Pleistocene climatic ice core records from
Qinghai-Tibetan Plateau: Science, v. 246, p. 474-477.
Thompson, L.G., Yao, T., Davis, M.E., Henderson, K.A., Mosley-Thompson, E., Lin, P.-N., Beer, J.,
Synal, H.-A., Cole-Dai, J., and Bolzan, J.F., 1997, Tropical climate instability: The last glacial cycle
from a Qinghai-Tibetan ice core: Science, v. 276, p. 1821-1825.
Tschudi, S., Ivy-Ochs, S., Schlüchter, C., Kubik, P., and Rainio, H., 2000, 10Be dating of Younger
Dryas Salpausselkä I Formation in Finland: Boreas, v. 29, p. 287 - 293.
Tschudi, S., Schäfer, J.M., Potter, N., Ivy-Ochs, S., Kubik, P.W., Denton, G., Marchant, D., and
Schlüchter, C., submitted, Last major advance of Taylor Glacier into central Beacon Valley at least
4 Ma ago: New indications from surface exposure dating on clasts from Granite drift: Geology.
Tschudi, S., Schäfer, J.M., Borns, H., Ivy-Ochs, S., Barrett, P. Kubik, P.W., and Schlüchter, C., to be
submitted, Cosmogenic nuclides in the Allan Hills nunatak: An approach to reconstruct the local
glacial history: The Journal of Geology.
Vartanyan, S.L., Arslanov, K.A., Tertychnaya, T.V., and Chernov, S.B., 1995, Radiocarbon dating
evidence for mammoths on Wrangel Island, Arctic Ocean, until 2'000 BC.: Radiocarbon, v. 37, p.
1-6.
Vernekar, A.K., Zhou, J., and Shukla, J., 1995, The effect of Eurasian snow cover on the Indian
monsoon: Journal of Climate, v. 8, p. 248-266.
Wang, L, Sarnthein, M., Erlenkeuser, H., Grimalt, J., Grootes, P., Heilig, S., Ivanova, E., Kienast, M.,
Pelejero C., and Pflaumann, U., 1999, East Asian monsoon climate during the Late Pleistocene:
high-resolution sediment records from the South China Sea: Marine Geology, v. 156, p. 245-284.
Webb, P.N., Harwood, D.M., McKelvey, B.C., Mercer, J.H., and Stott, L.D., 1984, Cenozoic marine
sedimentation and ice-volume variation on the East Antarctic craton: Geology, v. 12, p. 287-291.
Wohlfarth, B., Björck, S., Cato, I., and Possnert, G., 1997, A new middle Holocene varve diagram
from the river Angermanälven, northern Sweden: indications for a possible error in the Holocene
varve chronology: Boreas, v. 26, p. 347-353.
Ye, D.Z., and Wu, G.X., 1998, The role of the heat source of the Tibetan Plateau in the general
circulation: Meteorology and Atmospheric Physics, v. 67, p. 181-198.
Zheng, B., 1989, Controversy regarding the existence of a large ice sheet on the Qinghai-Xizang
(Tibetan) Plateau during the Quaternary period: Quaternary Research, v. 32, p. 121-123.
APPENDIX
117
Appendix A
Used abbreviations
AHN
AMS
BP
EAIS
GIS
ICP-AES
Ka / ky
LGIT
LGM
Ma / my
m.asl
MIS
RBS
SED
Ss I / II
YD
Allan Hills nunatak, Dry Valleys, Antarctica
Accelerator mass spectrometry
Before present
East Antarctic ice sheet
Geographic information system
Inductively coupled plasma atomic emission spectroscopy
thousand years
last glacial-interglacial transition (from about 15ka to 10ka)
last glacial maximum (about 24 to 22ka BP)
million years
meters above sea level (altitude)
Marine oxygen isotope stage
Rutherford backscattering spectrometry
Surface exposure dating
Salpausselkä I (YD glacial formation in southern Finland)
Younger Dryas (climate reversal during LGIT)
118
APPENDIX
Appendix B
Compilation of all 10Be and 26Al data
Chapter 3.1, Salpausselkä I
Sample Description
Sample
Sal 1
Sal 3
Sal 4b
Sal 5
Lithology
Thickness
[cm]
4
3
2
2
Granite
Granite
Quartz vein
Granite
Present altitude
[m.asl]
160
160
160
160
Remarks
Clast on Salpausselkä I glacial formation
Results
Sample
Sal 1
Sal 3
Sal 4b
Sal 5
Run label
Quartz [g] Be Carrier
[mg]
4.98-ZB0598
6.98-ZB0613
3.99-ZB0770
3.99-ZB0771
207.11
210.71
165.52
160.76
0.4
0.4
0.3
0.3
10
Be Conc.
[Atoms/g
SiO2]
7.50x104
7.07x104
6.95x104
7.26x104
Error
[%]
Run label
6.7
6.8
7.0
7.4
2.98-ZA0171
1.99-ZA0210
2.99-ZA0233
2.99-ZA0234
26
Al content
Al Conc.
[ppm]
[Atoms/g SiO2]
5.54x105
4.37x105
4.81x105
4.71x105
125
176
157
122
Error
[%]
10.9
17.3
11.7
12.0
Chapter 3.2, Wrangel Island
Sample Description
Sample
Lithology
Wra 1 Quartz vein; bedrock
Wra 2 Quartz vein; boulder
Wra 4 Quartz vein; bedrock
Thickness
[cm]
10
7
2
Present altitude
[m.asl]
20
20
160
Remarks
Bedrock
Loose clast
Bedrock
Results
Sample
Run label
Wra 1
Wra 2
Wra 4
5.99-ZB0832
5.99-ZB0833
1.00-ZB0931
[mg]
120.01
119.67
72.25
0.3
0.3
0.3
10
Be Conc. Error
[Atoms/g
[%]
SiO2]
35.73x104 8.8
7.14x104 7.3
1.77x104 6.5
Run label
1.00-ZA0309
1.00-ZA0310
1.00-ZA0311
26
Al content
Al Conc.
[ppm]
[Atoms/g SiO2]
Error
[%]
0.51x105
4.25x105
10.52x105
15.8
7.5
7.0
25
33
58
The error includes a 1σ AMS measurement uncertainty for sample and blank and a 5% uncertainty for possible chemical processing
variability (Ivy-Ochs, 1996). Run label is an internal label from the Zurich AMS facility.
APPENDIX
119
Chapters 3.3 Kanding
Sample Description
Sample
Lithology
K1
K2
Granite
Granite
Thickness
[cm]
1.5
1.3
Present altitude
[m.asl]
4’240
4’260
Remarks
Clast on moraine
Clast on moraine
Results
Sample
Run label
K1
K2
1.99-ZB0719
6.98-ZB0615
[mg]
22.777
25.011
0.4
0.4
10
Be Conc.
[Atoms/g
SiO2]
9.17x105
7.87x105
Error
[%]
Run label
6.4
6.6
1.99-ZA0214
1.99-ZA0215
26
Al content
Al Conc.
[ppm]
[Atoms/g SiO2]
202
365
54.96x105
50.58x105
Error
[%]
6.3
6.8
Chapter 3.4, Litang & Tanggula
Sample Description
Sample
Li1a
Li1b
Li3
Li4a
Li4bc
Li5a
Li5b
Li6
Li7
Tan2
Tan4
Tan5b
Tan7
Lithology
Granite
Granite
Granite
Granite
Granite
Granite
Granite
Granite
Granite
Quartzite
Sandstone
Dacite
Dacite
Thickness
[cm]
1.5
5
2
2.5
1.5
2.5
2.5
3
1.5
3
1
1
3
Present altitude
[m.asl]
4’890
4’890
4’560
4’560
4’560
4’610
4’610
4’570
4’480
5’015
4’925
4’925
5’120
Remarks
Clast on moraine
Clast on moraine
Clast on moraine
Clast on moraine
Clast on moraine
Clast on moraine
Clast on moraine
Clast on moraine
Clast on moraine
Clast on moraine
Clast on moraine
Clast on moraine
Clast on moraine
Results
Sample
Li1a
Li1b
Li3
Li4a
Li4bc
Li5a
Li5b
Li6
Li7
Tan2
Tan4
Tan5b
Tan7
Run label
4.98-ZB0590
4.98-ZB0591
6.98-ZB0614
4.98-ZB0592
4.98-ZB0593
4.98-ZB0594
4.98-ZB0595
4.98-ZB0596
4.98-ZB0597
5.99-ZB0834
1.00-ZB0933
1.00-ZB0934
1.00-ZB0935
[mg]
27.977
29.492
11.288
18.403
22.976
24.897
27.953
30.031
19.521
10.789
11.661
6.015
5.195
0.4
0.4
0.35
0.4
0.4
0.4
0.4
0.4
0.35
0.3
0.3
0.3
0.3
10
Be Conc. Error
[%]
[Atoms/g
SiO2]
8.21x105
7.0
11.57x105
7.0
11.08x105
6.6
14.32x105
6.9
11.98x105
7.0
12.53x105
6.7
12.47x105
6.4
12.48x105
6.3
10.21x105
6.7
163.73x105
5.2
90.27x105
5.2
160.68x105
5.5
77.09x105
5.5
Run label
2.98-ZA0171
2.98-ZA0172
1.99-ZA0213
2.98-ZA0173
2.98-ZA0174
2.98-ZA0175
2.98-ZA0176
2.98-ZA0177
2.98-ZA0178
1.00-ZA0312
1.00-ZA0313
1.00-ZA0314
1.00-ZA0315
26
Al content
Al Conc.
[ppm]
[Atoms/g SiO2]
Error
[%]
94.71x105
54.57x105
77.85x105
79.23x105
76.55x105
72.27x105
72.44x105
77.07x105
70.24x105
1063.21x105
465.62x105
973.45x105
472.24x105
6.8
6.6
6.4
7.1
7.1
7.1
11.7
8.1
11.2
5.3
6.1
5.3
6.0
276
215
148
287
262
266
208
194
138
200
283
111
197
120
APPENDIX
Chapter 4.1, Beacon Valley
Sample Description
Sample
339
340
341
342
343
344
BV9714
Lithology
Granite
Granite
Granite
Granite
Granite
Granite
Sandstone
Thickness
[cm]
1.4
1.4
0.75
0.5
0.9
0.8
3.5
Present altitude
[m.asl]
1’380
1’380
1’400
1’430
1’450
1’500
1’345
Remarks
Clast from Granite drift
Clast from rockfall
Results
Sample
Run label
339
340
341
342
343
344
BV9714
5.97-ZB0436
2.97-ZB0395
5.97-ZB0437
5.97-ZB0438
2.97-ZB0396
6.98-ZB0617
Quartz
[g]
5.012
6.160
5.781
5.007
6.190
8.017
Ber
Carrier
[mg]
0.4
0.4
0.4
0.4
0.4
0.4
10
Be Conc. Error
Run label
[Atoms/g [%]
SiO2]
1.81x107
6.5 1.98-ZA0103
- 1.98-ZA0104
2.52x107
6.5 1.98-ZA0105
3.59x107
6.1 1.98-ZA0106
7.3 1.98-ZA0107
2.77x107
1.68x107
6.5 1.98-ZA0108
1.17x107
5.4 1.99-ZA0226
Quartz
[g]
2.462
2.584
0.899
5.000
5.548
1.819
8.017
Al content
[ppm]
269
182
442
94
105
257
82
26
Al Conc.
Error
[Atoms/g
[%]
SiO2]
8.34x107
6.4
6.98x107
7.4
11.26x107 10.5
11.62x107
5.8
10.64x107
6.4
8.39x107
8.4
6.62x107
5.3
Chapter 4.2, Allan Hills nunatak
Sample Description
Sample
225
227
228
232
AL9702
AL9704
AL9708
AL9709
AL9710
AL9711a
AL9711b
AL9711c
AL9712b
AL9713a
AL9713b
AL9713c
AL9714
Lithology
Sandstone
Quartzite
Sandstone
Sandstone
Quartzite
Sandstone
Sandstone
Sandstone
Sandstone
Sandstone
Sandstone
Sandstone
Quartzite
Sandstone
Sandstone
Sandstone
Sandstone
Thickness
[cm]
1.5
1.25
0.8
1.0
0.5
1.5
1.0
1.0
3.0
1.3
1.0
0.8
0.5
4.0
3.5
3.0
3.5
Present altitude
[m.asl]
1’670
1’730
1’670
1’635
1’645
1’745
1’690
1’705
1’675
1’705
1’705
1’705
1’635
1’665
1’665
1’660
1’595
Remarks
Clast, single erratic, young advance
Sirius Group clast
Clast, push moraine, young advance
Bedrock striated
Clast
Sirius Group clast, embedded
Clast, boulder train, young advance
Bedrock, excavated, young advance
Bedrock
Clast
APPENDIX
121
Results
Sample
Run label
Quartz Be Carrier
[g]
[mg]
10
Be Conc.
Error
Run label
[Atoms/g
[%]
SiO2]
225
1.97-ZB0355
5.451
0.5
1.03x107
6.5
1.97-ZA0077
227
1.97-ZB0356
9.109
0.5
1.61x107
6.8
1.97-ZA0078
228
1.97-ZB0357
7.877
0.5
2.24x107
6.3
1.97-ZA0079
232
1.99-ZB0727
5.836
0.3
0.76x107
7.9
1.99-ZA0225
AL9702
7.98-ZB0687
5.047
0.3
2.75x107
8.1
*
AL9704
1.99-ZB0723
7.041
0.3
3.30x107
5.3
1.99-ZA0217
AL9708
1.99-ZB0725
5.423
0.3
0.43x107
6.4
1.99-ZA0218
AL9709
1.99-ZB0726
5.284
0.3
0.41x107
9.8
1.99-ZA0219
AL9710
6.98-ZB0616
4.524
0.4
0.51x107
6.1
1.99-ZA0220
AL9711a 3.99-ZB0776
7.616
0.3
1.82x107
5.6
2.99-ZA0235
AL9711b 3.99-ZB0777
7.129
0.3
1.58x107
6.6
2.99-ZA0236
AL9711c 3.99-ZB0778
7.895
0.3
2.65x107
5.3
2.99-ZA0237
AL9712b 2.99-ZB0753
3.028
0.3
3.08x107
6.7
2.99-ZA0238
AL9713A 7.98-ZB0688 10.583
0.3
0.26x107
6.8
1.99-ZA0221
AL9713B 7.98-ZB0689
2.046
0.3
0.30x107
9.1
1.99-ZA0222
AL9713C 7.98-ZB0690
1.532
0.3
0.02x107 60.2
1.99-ZA0223
AL9714
7.98-ZB0691
1.880
0.3
0.33x107 11.2
1.99-ZA0224
* No 26Al determination due to problems during the ICP-AES measurement
Al content
[ppm]
141
243
130
2839
*
42
467
331
70
112
90
114
86
99
201
255
455
26
Al Conc.
[Atoms/g
SiO2]
4.51x107
8.79x107
11.38x107
3.09x107
*
15.21x107
2.31x107
2.15x107
3.05x107
8.37x107
8.56x107
12.72x107
11.62x107
1.66x107
1.63x107
0.12x107
1.84x107
Error
[%]
6.7
6.9
6.6
9.1
*
5.1
7.6
6.4
5.8
5.1
5.2
5.1
5.6
7.1
6.4
18.3
7.7
DANKSAGUNG
123
Danksagung
Mit Danksagungen ist es so eine Sache: Nie kann man sie allen recht machen und immer
lassen sich Personen finden, die übergangen wurden. Deshalb sei hier und jetzt, an allererster
Stelle allen unterstützenden Personen und Organisationen (besonders dem Schweizerische
Nationalfonds SNF) ein herzliches Dankeschön gewidmet. Die Druckkosten der vorliegenden
Arbeit wurden teilweise vom Gletschergarten Luzern, dem Naturmuseum Luzern, sowie der
Sand AG in Neuheim (ZG) übernommen. Danke für diese namhaften Beiträge!
Dem Leiter dieser Arbeit, Christian Schlüchter, möchte ich für seinen Enthusiasmus
gegenüber der Methode der Oberflächenaltersbestimmung und für seine unermüdliche Lust
an Feldarbeit danken. Seine Begeisterung für Quartärgeologie hat sich auf mich übertragen
und die gemeinsamen Expeditionen und die damit verbundenen Erfahrungen sind
unbeschreiblich wertvoll. Danke Christian!
Die gute Seele, ohne die diese Arbeit nie zu Stande gekommen wäre ist Susan Ivy-Ochs. Sie
hat mich im Labor in die Geheimnisse der komplexen Probenaufbereitung eingeweiht, und sie
half mir zu erkennen, dass wir noch weit davon entfernt sind alles über die spannende
Methode der Oberflächendatierung zu wissen. Ich wünsche ihr alles Gute in ihrer Zukunft.
Thanxx a billion times Susan!
Der Physiker an meiner Seite, Peter Kubik, half mir die geologischen Probleme mit
physikalischen Ansätzen zu lösen. Er weihte mich in die Geheimnisse der MASCHINE ein und
weckte die Neugier auf Physik. Nie werde ich den Moment vergessen, als wir zusammen vor
dem Hamster sassen und gespannt den Zahlen zuschauten, die sich zur Jüngeren Dryas
formten. Er war stets ein kritischer Leser, der mich trainierte den Zahlen einen Fehler zu
geben und masslosen Behauptungen auf den Grund zu gehen. Danke Peter für Deinen
fruchtbaren Einsatz der massgeblich zum Gelingen dieser Arbeit beigetragen hat!
Mit Jörg Schäfer, dem Mitdoktoranden von der Edelgasseite, war ich zweifelsohne im
Wettbewerb. Aber, es war ein fruchtbarer und konstruktiver Wettbewerb! Die Highlights auf
den unendlichen Höhen des tibetanischen Plateau sind unvergesslich und ich hoffe nicht, dass
wir uns aus den Augen verlieren, auch wenn Jörg nach Westen und ich in den Nordosten
ziehe. Danke Jörg für Deine Ehrlichkeit!
124
DANKSAGUNG
Der Gruppe auf dem Hönggerberg, die mir stets das Gefühl gab, ich sei einer von ihnen, sei
hiermit ein dickes Danke gesprochen. Sie haben mich als Geologen mit offenen Armen
empfangen und sie haben akzeptiert, dass es Leute gibt, die für Steine um den halben Planeten
reisen. Danke Georges Bonani, Irka Haydas, Rainer Mühle, Ruedi Pfenniger, Christof
Schnabel, Martin Suter, Arno Synal. Nicht zu vergessen die MASCHINENCREW René Gruber,
Peter Kägi und Jürg Thut, ohne die das Ding ja eh’ nicht laufen würde. Max Döbeli möchte
ich für seine sprudelnde Motavation und seinen spontanen RBS Einsatz danken. Weiterhin
bin ich Wolfgang Gruber zu tiefem Dank verpflichtet, da er stets zu Diensten war, wenn im
Labor etwas mechanisches nicht mehr so ganz wollte, wie es sollte. Und, er hat mir seinen
schnellen Apfel überlassen. Und da wären noch die lieben Mitstreiter im Club der
Doktoranden: Steff, Colin, Jürgen, Henning und Philipp. Hier ein Schwatz und da ein Glas,
hier mal Tanzen und da mal Reisen. Es war spassig mit euch!
Vielen Dank an die EAWAG (Jürg Beer, Silvia Bollhalder, Alfred Lück, Caroline Stengel),
wo ich mich als Gast in den Laborräumlichkeiten breit machte und wo ich unter der Leitung
von David Kistler das ICP-AES für die Aluminiummessungen benutzen durfte.
Ein Dankeschön-Kiitos-Thanks-
geht an alle, die im Feld in irgendeiner Art und Weise
mitgeholfen haben. Ohne ihre logistische Unterstützung wäre keine einzige Probe aus
Finnland, New Zealand der Antarktis oder China in die Schweiz gelangt. Dankeschön der
Scott Base Crew für die Gastfreundschaft und den KIWI-14 und VXE-6 Piloten für die
erbrachten Flugleistungen in der Antarktis.
Dann möchte ich noch allen Freunden und Bekannten ein Dankeschön widmen, ohne die ich
die Dissertationszeit wohl nicht über die Runden gebracht hätte.
Meinen Eltern, die mir während der gesamten Studienzeit in allen Belangen zur Seite standen
und mich immer unterstützten gebührt der grösste Dank:
Mami und Papi, merci velmol!
CURRICULUM VITAE
125
Curriculum vitae
Angaben zur Person
Name
Eltern
geboren
Silvio Tschudi
Rolf und Ursula Tschudi-Bruha
am 16. Februar 1971 in Luzern (Schweiz)
Ausbildung
1977-1983
1983-1990
1990
1990-1991
1991
1991-1996
1996-2000
Primarschule in Kriens, Luzern
Kantonsschule Alpenquai Luzern
Matura Typus C
Studium der Physik an der ETH Zürich
1. Vordiplom in Physik
Studium der Erdwissenschaften an der ETH Zürich mit Vertiefung
Ingenieurgeologie / Angewandte Geophysik. Diplomarbeit zum Thema
“Observed pressure variations and their causes in the geothermal field
of Kuzuluk/NW Turkey”.
Promotion über das Thema “Surface exposure dating: A geologist’s
view with examples from both hemispheres” am Geologischen Institut
der Universität Bern bei Prof. Dr. Christian Schlüchter und am Institut
für Teilchenphysik der ETH Zürich bei Dr. Peter W. Kubik.
Berufliche Erfahrungen
1993-1996
1996-1997
1996-2000
Hilfsassistent in der Lehre am Geologischen Institut der ETH Zürich
Freier Mitarbeiter im Ingenieur- und Geologiebüro Wanner AG, Aathal.
Tätigkeiten: Baugrund- und Altlastenuntersuchungen, Erarbeitung von
Störfallberichten und Feuerwehr-Einsatzplänen.
Wissenschaftlicher Assistent am Geologischen Institut der Universität
Bern und am Institut für Teilchenphysik der ETH Zürich. Tätigkeiten:
Assistenz in der Lehre, Radioisotopenmessung mittels AMS.

Surface exposure dating - IT Services of ETH Zurich

Transcription

Similar documents

COOLVIEW - PyroGraphics

Processes of Glacial Erosion

Ice Vault Arena SkateFest 12-10-13

October 2014

MIS33-Skylight Ice Arena.qxp - The Ives Architecture Studio

5776 Kadima Ice Skating Final

betriebsinfo E f1a

Värsta istiden - Student Portal

Why The Yaktrax Pro Works

Save at Aldo`s! - Aldo`s Restaurant Block Island