Assessing the Potentials of
Calcium Isotope Thermometry in Tropical and
Polar Oceans
Inauguraldissertation
der Philosophisch-naturwissenschaftlichen Fakultät
der Universität Bern
vorgelegt von
Dorothee Hippler
aus Deutschland
Leiter der Arbeit:
Dr. Th. F. Nägler
Institut für Geologie, Gruppe Isotopengeologie, Universität Bern
Prof. Dr. Anton Eisenhauer
Leibniz-Institut für Meereswissenschaften (IfM-GEOMAR), Kiel
Assessing the Potentials of
Calcium Isotope Thermometry in Tropical and
Polar Oceans
Inauguraldissertation
der Philosophisch-naturwissenschaftlichen Fakultät
der Universität Bern
vorgelegt von
Dorothee Hippler
aus Deutschland
Leiter der Arbeit:
PD Dr. Th. F. Nägler
Institut für Geologie, Gruppe Isotopengeologie, Universität Bern
Prof. Dr. Anton Eisenhauer
Leibniz-Institut für Meereswissenschaften (IfM-GEOMAR), Kiel
Von der Philosophisch-naturwissenschaftlichen Fakultät angenommen.
Bern, den 2. November 2004
Der Dekan:
Prof. Dr. P. Messerli
„Wir sehen in der Natur nicht Wörter,
sondern immer nur Anfangsbuchstaben von Wörtern,
und wenn wir alsdann lesen wollen, so finden wir,
dass die neuen sogenannten Wörter wiederum
bloß Anfangsbuchstaben von andern sind.“ *
* Georg Christoph Lichtenberg
Acknowledgements
I want to pronounce my sincere gratitude and appreciation to many people who made this
PhD-project possible.
•
•
•
•
•
Firstly, to my supervisor Thomas Nägler who introduced me patiently and clearly to such
an interesting area of research. For his encouragement pursuing my interests and the
chance to participate in workshops and conferences all around the world.
My special gratitude goes also to Anton Eisenhauer for his enthusiam and appreciate
advice concerning all aspects of Ca isotopes. Thanks for offering the possibility to
particitpate on the research cruise on RV Sonne (SO164) from May to June 2002 in the
Caribbean Sea.
My sincere gratitude to Kate Darling for the great collaboration in the very exciting N.
pachyderma-project.
I want to thank Jan Kramers and Igor Villa for the benefit from their huge analytical and
scientific knowledge.
Thanks to Thomas Stocker who honoured the position as Koreferent.
My greatest thanks to:
•
•
•
•
All the collaborators of our familiar and increasing Ca network for the good cooperation
and the fun on many conferences: Anne-Désirée Schmidt, Peter Stille, Jan Fietzke,
Reinhard Kozdon, Volker Liebetrau, Nikolaus Gussone, Alexander Heuser, Katharina
Kreissig and Adrian Immenhauser.
All the colleages and friends at the institute with a special mention to Bärbel, Ilka, Chris,
Martin, Nicola, Franziska, Katja, Schorsch and Dänu for having a good time at the
institute and Borgi Hebeisen for helping with the lab equipment
I wish to thank all my best friends outside the institute, especially Sebi, Petra, Marianne
and Thomas for cycling and skiing the mountains up and down all over switzerland.
A very special thank to Annett, my lovely roommate, for being a great friend and
invaluable in many situations. Stay as you are!
Dear mum and dad thanks that you have always believed in me and supported me in many
ways.
Dear Pius, thank you very much for having the feeling that you are always on my side!
Thanks to all of you!
Preface and Summary
The PhD-thesis presented here focuses on the relation between ambient seawater
temperatures and the mass-dependent shift of calcium isotope abundances (δ44/40Ca) in calcite
shells of surface-dwelling marine organisms (foraminifera and mollusks). Subsequent to the
refinement and establishment of the analytical technique, δ44/40Ca-temperature relationships
were calibrated for different species. δ44/40Ca is positively correlated to temperature with
increasing temperatures result in higher 44Ca/40Ca ratios. Thus, Ca isotopes are a direct proxy
for sea surface temperature (SST). This hypothesis has been tested in tropical and polar
oceans on different timescales. These topics are addressed in five chapters. The chapters are
in form of manuscripts that have been either published, accepted for publication, submitted
for review or are in preparation.
The first two chapters of this thesis provide an overview of Ca isotope studies in the literature
and point to the problems in comparing results from different groups In particular, there was
no agreement on an international Ca-standard used as reference material. Most of the results
published so far were normalised to laboratory internal reference materials. Subsequent to a
series of workshops concerning stable isotopes (e.g. Int. Goldschmidt Conference 2002,
Davos), a compilation of different calcium reference materials has been initiated by three
European groups (Kiel, Strasbourg, Bern). This inter-laboratory comparison of different
reference materials showed general agreement. Consequently, the use of NIST SRM 915a as
reference material was recommended for future Ca isotope studies.
The third chapter presents the refinement of the published δ44/40Ca-temperature calibration of
Nägler et al. (2000). We investigate modern specimens of planktonic foraminifera
Globigerinoides sacculifer and re-assess the robustness of Ca isotopes as sea surface
temperature (SST) proxy. The δ44/40Ca values of modern G. sacculifer show a clear
temperature dependent Ca isotope fractionation allowing estimates of relative temperature
changes and absolute temperatures. The application of this δ44/40Ca-temperature relationship
to a tropical East Atlantic sediment core render SST changes at the last two glacialinterglacial transitions of about 3-4°C that is in excellent agreement with literature data. To
obtain information on magnitude and timing of tropical SST changes we adopt a multiproxy
approach (δ18O, transfer function, Mg/Ca and δ44/40Ca). Our findings reveal new evidence for
the important role of the tropics in triggering global climate change.
The fourth chapter summarises the results of an interdisciplinary project (geo- and
biosciences) exploring the potential geochemical consequences of cryptic genetic diversity of
planktonic foraminifera Neogloboquadrina pachyderma (sinistral) on the Ca isotopic
signature. This species inhabiting dominantly cool-water environments is an important
ecological and geochemical proxy of past ocean conditions. A comparison of Ca isotopic
composition of different genotypes shows exactly the same temperature-dependence. Thus,
Preface and Summary
genetic diversity does not influence the δ44/40Ca-temperature relationship. We suggest the use
of Ca isotopes as an additional proxy for the reconstruction of sea surface temperatures in
high-latitude oceans. Furthermore, Ca isotopes in N. pachyderma (sin.) provide new insights
in biomineralisation processes.
The fifth chapter is dealing with climate reconstruction in the Upper Cretaceous. Seasonal
temperature fluctuations of Tethyan coastal waters are recorded in a ≈80Ma-old rudist shell
from Turkey. Three different palaeo-temperature proxies (δ18O, Mg/Ca and δ44/40Ca) were
applied to assess the critical factors other than seawater temperature (e.g. seawater pH, CO32concentrations, diagenetic alteration and metabolic processes) that might have influenced the
shell geochemical record.
The Appendix briefly summarises the analytical procedures concerning Ca isotope analyses
performed on a single collector thermal ionisation mass spectrometer (TIMS). It gives
additional information about the calibration of the 43Ca-48Ca double spike that was used for all
Ca isotope analyses presented in this work.
Financial support for this work, including the graduate student assistantship to the author, was
provided by the Swiss National Science Foundation (grants 21-61644.00 and 21-60987.00) to
Thomas F. Nägler. The work included the participation on the research cruise on RV Sonne
(May 22 to June 28, 2002, SO164) within the framework of RASTA (Rapid climate changes
in the western tropical Atlantic – Assessment of the biogenous and sedimentary record). The
cruise SO164 was founded by the Ministry of Education and Research (BMBF) under project
No. 03G0164 to GEOMAR.
Introduction
Calcium (Ca) is an alkaline earth metal and is widely distributed in many geological and
biological reservoirs. It is one of the most abundant elements in the bulk silicate earth (1.1 wt
%) as well as a major component of the world oceans (≈ 400ppm). It is also an essential
constituent in animal and plant tissues. On Earth, Ca generally exists in chemical compounds
in particular carbonates, sulphates, phosphates or silicates. Furthermore, Ca plays an
important role in different geological (e.g. diagenesis, weathering and carbonate precipitation)
and biological processes (e.g. calcification, bone formation and metabolism).
Ca (atomic number Z = 20) has six naturally occurring stable isotopes, with 40Ca being by far
the most abundant: 40Ca (96.941%), 42Ca (0.647%), 43Ca (0.135%), 44Ca (2.086%), 46Ca
(0.004%) and 48Ca (0.187%). The mass range of eight atomic mass units (amu) corresponds to
20% mass difference between 40Ca and 48Ca. The large relative mass difference of Ca isotopes
and the wide distribution of Ca suggest a large isotope fractionation in natural processes.
Early studies focussing on Ca isotope fractionation of terrestrial samples found no
fractionation of Ca isotopes or ambigious results. In several studies on Ca isotope systematics
and chemical purification Heumann and coworkers reported that natural fractionation of Ca
isotopes is relatively small (cf. Heumann et al. 1970, 1982, Heumann & Lieser 1972, Stahl &
Wendt 1968) see Fietzke et al. 2004) requiring high analytical precision to be resolved.
A major improvement in the precise Ca isotope measurement technique was the application
of the double spike technique (Coleman 1971, Russell et al. 1978). This method revealed
promising results for the use of Ca isotopes in various fields of environmental sciences. The
double spike technique was previously used for lead (Pb) isotope analysis (Compsten &
Oversby, 1969). Russell et al. (1978) expressed their Ca isotope data as δ(40Ca/44Ca) =
[(40Ca/44Ca)sample/(40Ca/44Ca)standard–1]*1000 [‰] using a natural calcium flouride (CaF 2) sample
as reference material. They avoided to use industrialy refined calcium which can be extremely
fractionated (up to ≈13% in the 40Ca/44Ca ratio) relativ to natural occuring calcium. Skulan et
al. 1997 redefined the notation using the same convention as in light stable isotopes (e.g.
δ18O, δ13C, δ15N) by normalising the measured 44Ca intensity to 40Ca: δ44Ca [‰] =
[(44Ca/40Ca)sample/(44Ca/40Ca)standard–1]*1000. As reference materials an internal calcium
carbonate standard was used by Skulan et al. 1997, Skulan & DePaolo 1999, De La Rocha &
DePaolo 2000, seawater by Zhu & MacDougall 1998 and a natural CaF2 was employed as
reference material by Nägler et al. 2000 and Nägler & Villa 2000. Due to the lack of an
international consensus different groups used different Ca isotope notations and reference
materials. Most of the studies dealing with Ca isotopes used a single collector thermal
ionisation mass spectrometer (TIMS) and a double spike technique. Although it is a timeconsuming method with small sample throughput good internal statistical precision can be
achived. Few publications report the successful use of a multi collector TIMS (Fletcher et al.
1997, Heuser et al. 2002). Other mass spectrometers like multicollector inductively coupled
Introduction
plasma mass spectrometry (MC-ICP-MS) (Halicz et al. 1999) have also been tested for Ca
isotope analysis. The main problem of MC-ICP-MS measurements arises from isobaric
interferences e.g. 40Ar+ (which is commonly used as transfer medium) on mass 40 or 12C16O2+
on mass 44. Therefore, Halicz et al. (1999) did not measure 40Ca and rather reported 42Ca/44Ca
ratios as δ44Ca [‰] = [(44Ca/40Ca)sample/(44Ca/40Ca)standard–1]*1000 relative to the NIST SRM
915a carbonate reference material. In order to overcome this problem, recent progress in the
measurement of Ca isotopes using ICP-MS has been presented by Boulyga & Becker (2001)
using collision or reaction cells to remove interfering 40Ar+ , however without achieving the
necessary precision for geological applications and the promising „cool-plasma“ MC-ICP-MS
technique (Fietzke et al. 2004). In contrast to the TIMS technique, Ca isotope analysis on
ICP-MS can be performed by bracketing standard technique without the requirement of a Ca
double spike.
In terrestrial material, variations of the natural isotopic composition of Ca isotopes occur due
to mass-dependent fractionation during biological and physico-chemical processes or due to
radiogenic ingrowth (Russell et al. 1978). The production of 40Ca resulting from the β- decay
of 40K (half-life: 1.277*109 years) has been used in geochronology (K/Ca-dating) of igneous
and metamorphic rocks as well as minerals (Coleman 1971, Marshall & DePaolo 1982,
Fletscher et al, 1997, Marshall et al. 1986, Nägler & Villa, 2000). In igneous petrology Ca
isotopes have also been used to discuss the origin of granites (Marshall & DePaolo 1989).
Particularly, in marine geo- and biosciences new applications of Ca isotope fractionation
arose during the last seven years. Several studies dealt with Ca isotopes to model the global
marine Ca cycle (Zhu & MacDougall 1998, De La Rocha & DePaolo 2000, Schmitt et al.
2003a). The Ca isotopic composition of modern seawater is homogeneous in oceans
worldwide (Zhu & MacDougall 1998, De La Rocha & DePaolo 2000, Schmitt et al. 2001,
Hippler et al. 2003). This homogeneity is expected in view of the long Ca residence time (≈
1Ma, Broecker & Peng 1982) compared to the mixing time of ocean water (≈ 1.5 ka). Secular
variations in the Ca isotopic composition of seawater have been reported for the last 160Ma
(Skulan et al. 1997) and 80Ma (De La Rocha & DePaolo, 2000) based on studies of marine
carbonate samples. Since bulk sediment may yield ambiguous results, Schmitt et al. (2003b)
and Soudry et al. (2004) based their study exclusively on authigenic marine phosphates to
record Ca isotopic variations over the last 24Ma and for the Aptian-Eocene period (between
≈110-40Ma), respectively.
Further studies focused their attention on the biological control of Ca isotope fractionation in
marine and terrestrial organisms, investigating intraorganismal differences and trophic level
relationships (Skulan et al. 1997, Skulan & DePaolo 1999, Clementz et al. 2003). Comparing
the Ca isotopic composition of different marine organisms Skulan et al. (1997) postulated that
the temperature effect is small in comparison to effects they attributed to the tropic level. Zhu
& MacDougall 1998 showed that among foraminifera, Ca isotopes exhibit species-dependent
fractionation, but also appear to display systematic temperature-dependent fractionation
within individual species. A more systematic study of δ44/40Ca-temperature relationships in
cultured and fossil planktonic foraminifera G. sacculifer demonstrated a clear temperature
dependence (Nägler et al. 2000). Based on their findings the latter authors suggest that Ca
isotopes are potentially a new proxy for past sea surface temperatures (SST). In contrast, De
La Rocha & DePaolo (2000) claimed that Ca isotopic composition of intertidal foraminifera
Glabratella ornatissima is independent of temperature variations over a relatively small
temperature range (8.5 to 10.8°C). The observation of species-dependent Ca isotope
fractionation is supported by Gussone et al (2003) who investigated the temperaturedependent Ca fractionation in calcite shells of cultured Orbulina universa and inorganically
precipitated aragonite. The different fractionation trends of O. universa and G. sacculifer they
ascribed different biochemical fractionation processes related to their calcite precipitation
mechanism. Although Mariott et al. (2004) favours equilibrium dynamics to describe Ca
isotope fractionation studying inorganic precipitates, they explained the stronger temperature
dependence of G. sacculifer as the result of a superimposed additional biological fractionation
related to biomineralisation.
Objectives
One of the first objectives of this study was to suggest a common Ca isotope standard since
there was no uniform way of standardisation. Direct comparison of published data sets was
hampered by the use of various reference materials.
Specifically,
• the Ca isotopic composition of several reference samples (including seawater) used in
previous studies was determined
• reference samples (including seawater) were to be cross-calibrated
• the Ca isotope fractionation between NIST SRM 915a (CaCO3 reference powder) and
seawater, representing a major natural Ca reservoir was to be defined.
Concerning Ca isotope fractionation in surface-dwelling marine organisms the first goal of
this research was to verify and refine the δ44/40Ca-temperture relationship of Nägler et al.
(2000).
In particular,
• we investigate the relation between ambient seawater temperatures and the massdependent shift of Ca isotope abundance of modern specimen of planktonic foraminifera
G. sacculifer in order to re-assess the potential of Ca isotopes as an independent sea
surface temperature (SST) proxy for tropical and subtropical environments.
• to test the refined relationship we performed additional Ca isotope analyses of fossil G.
sacculifer on the down-core record from a tropical east Atlantic sediment core covering
the last 140.000 years, i.e. the last two glacial-interglacial transitions.
Introduction
The second important objective was to study the implications for Ca isotope thermometry
resulting from cryptic genetic diversity in planktonic foraminifera. We choose N. pachyderma
(sin.) inhabiting preferably cool-water environments, since five distinct genotypes have been
identified recently.
Therefore,
• the genotype and Ca isotopic composition was determined on the same individual of N.
pachyderma (sin.).
• the relation of genotype, δ44/40Ca and ambient seawater temperature was investigated.
• the potential of Ca isotope thermometry in high-latitutde oceans was explored.
Further, after experience gained in the Quaternary, Ca isotope thermometry was applied to a
≈80Ma-old low-Mg calcite rudist shell in order to investigate the possibilities and limits of
the new δ44/40Ca-temperature proxy.
Therefore,
• δ44/40Ca was measured on low-Mg calcite rudist shell with sub-annual resolution.
• Rudist δ44/40Ca was compared with other proxies (δ18O, Mg/Ca) and trace element
concentrations in order to assess whether shell geochemistry reflects the temperature of
the ambient seawater, seawater pH, diagenetic alteration or metabolic processes.
References
Boulyga S. F. and Becker J. S. (2001)
ICP-MS with hexapole collision cell for isotope ratio measurements of Ca, Fe, and Se.
Fresenius J. Anal. Chem. 370, 618-623.
Broecker W. S. and Peng T.-H. (1982)
Tracers in the Sea. Eldigo Press (New York), 690pp.
Clementz M. T., Holden P. and Koch P. L. (2003)
Are calcium isotopes a reliable monitor of trophic level in marine settings? Int. Journal
Osteoarchaeology, 13, 29-36.
Coleman M.L. (1971)
Potassium-calcium dates from pegmatic micas. Earth and Planetary Science Letters, 12, 399405.
Compston W. and Oversby V. M. (1969)
Lead isotope analysis using a double spike. Journal of Geophysical Research, 74(17), 43384348.
De La Rocha C. L. and DePaolo D. J. (2000)
Isotopic evidence for variations in the marine Calcium cycle over the Cenozoic. Science, 289,
1176-1178.
Fietzke J., Eisenhauer A., Gussone N., Bock B., Liebetrau V., Nägler Th. F., Spero H. J.,
Bijma J. and Dullo C. (2004)
Direct measurement of 44Ca/40Ca ratios by MC-ICP-MS using the cool plasma technique.
Chemical Geology, 206, 11-20.
Fletscher I. R., McNaughton N. J., Pidgeon R. T. and Rosman K. J. R. (1997)
Sequential closure of K-Ca and Rb-Sr systems in Archaean micas. Chemical Geology, 138,
289-301.
Gussone N., Eisenhauer A., Heuser A., Dietzel M., Bock B., Böhm F., Spero H.J., Lea D. W.,
Bijma J. and Nägler Th. F. (2003)
Model for kinetic effects on calcium isotope fractionation (δ44Ca) in inorganic aragonite and
cultured planktonic foraminifera. Geochimica et Cosmochimica Acta 67(7), 1375-1382.
Halicz L., Galy A., Belshaw N. S. and O'Nions R. K. (1999)
High-precision measurements of calcium isotopes in carbonates and related materials by
multiple collector inductively coupled plasma mass-spectrometry (MC-ICP-MS). Journal of
Analytical Atomic Spectrometry, 14, 1835-1838.
Heumann K. G., Lieser K. H., et al. (1970)
Difficulties in measuring the isotopic abundances of calcium with a mass spectrometer.
Recent Dev. Mass spectrometry, 457-459.
Heumann K. G. and Lieser K. H. (1972)
Untersuchung von Calciumisotopieeffekten bei heterogenen Austauschgleichgewichten. Z.
Naturforsch. 27b (2), 126-133.
Heumann K. G., Klöppel H. and Sigl G. (1982)
Inversion der Calcium-Isotopenseparation an einem ionenaustauscher durch Veänderung der
LiCl-Elektrolytkonzentration. Z. Naturforsch. 37b , 786-787.
Heuser A., Eisenhauer A., Gussone N., Bock B., Hansen B. T. and Nägler T. F. (2002)
Measurements of calcium isotopes (δ44Ca) using a multicollector TIMS technique.
International Journal of Mass Spectrometry, 220, 385-397.
Hippler D., Schmitt A.-D., Gussone N., Heuser A., Stille P., Eisenhauer A. and Nägler Th. F.
(2003)
Ca isotopic composition of various standards and seawater. Geostandards Newsletter, 27(1),
13-19.
Introduction
Mariott C.S., Henderson G. M., Belshaw N. S. and Tudhope A. W. (2004)
Temperature dependence of δ7Li, δ44Ca and Li/Ca during growth of calcium carbonate. Earth
Plant. Sci. Lett. 222,615-624.
Marshall B. D. and DePaolo D. J. (1982)
Precise age determination and petrogenetic studies using the K-Ca method. Geochimica et
Cosmochimica Acta, 46, 2537-2545.
Marshall B. D., Woodard H. H., Krueger H. W. and DePaolo D. J. (1986)
K-Ca-Ar systematics of authigenic sanidine from Waukau, wisconsin, and the diffusivity of
argon. Geology, 14, 936-938.
Marshall B. D. and DePaolo D. J. (1989)
Calcium isotopes in igneous rocks and the origin of granite. Geochim. Cosmochim. Acta, 53,
917-922.
Nägler T. F. and Villa I. M. (2000)
In pursuit of the 40K branching ratios: K-Ca and 39Ar-40Ar dating of gem silicates. Chemical
Geology, 169, 5-16.
Nägler T. F., Eisenhauer A., Müller A., Hemleben C. and Kramers J. (2000)
The δ44Ca-temperature calibration on fossil and cultured Globigerinoides sacculifer: New tool
for reconstruction of past sea surface temperatures. Geochemistry Geophysics Geosystems, 1,
2000GC000091.
Russell W. A., Papanastassiou D. A. and Tombrello T. A. (1978)
Ca isotope fractionation on the earth and other solar system materials. Geochimica et
Cosmochimica Acta, 42, 1075-1090.
Schmitt A.-D., Bracke G., Stille P. and Kiefel B. (2001)
The Calcium isotope composition of modern seawater determined by Thermal Ionisation
Mass Spectrometry. Geostandards Newsletter, 25, 267-275.
Schmitt A. D., Chabaux F. and Stille P. (2003a)
The Ca riverine and hydrothermal isotopic fluxes and the oceanic calcium mass balance.
Earth and Planetary Science Letters, 213(3-4), 503-518.
Schmitt A. D., Stille P. and Vennemann T. (2003b)
Variations of the 44Ca/40Ca ratio in seawater during the past 24 million years: Evidence from
δ44Ca and δ18O values of Miocence phosphates. Geochimica et Cosmochimica Acta 67(14),
2607-2614.
Skulan J., DePaolo D. J. and Owens T. L. (1997)
Biological control of calcium isotopic abundances in the global calcium cycle. Geochimica et
Cosmochimica Acta, 61, 2505-2510.
Skulan J. and DePaolo D. J. (1999)
Calcium isotope fractionation between soft and mineralized tissues as a monitor of Calcium
use in vertebrates. Proc. Natl. Acad. Sci., Biochemistry, 96: 13709-13713.
Soudry D., Segal I., Nathan Y., Glenn C. R., Halicz L., Lewy Z. and VonderHaar D. L.
(2004)
44
Ca/42Ca and 143Nd/144Nd isotope variations in Cretaceous-Eocene Tethyan francolites and
their bearing on phosphogenesis in the southern Tethys. Geology, 32 (5), 389-392.
Stahl W. and Wendt L.(1968)
Fractionation of calcium isotopes in carbonate precipitation. Earth planet. Sci. Lett. 5, 184186.
Zhu P. and MacDougall J. D. (1998)
Calcium isotopes in the marine environment and the oceanic calcium cycle. Geochimica et
Cosmochimica Acta, 62, 1691-1698.
Table of Contents
Chapter 1
Calcium isotopic composition of various reference materials and seawater
1. Introduction............................................................................................................... 2
2. Sample Material........................................................................................................ 3
3. Data acquisition and presentation............................................................................. 5
4. Results and discussion .............................................................................................. 6
5. Summary and conclusions ........................................................................................ 7
Appendix 1: Purification of seawater samples in Bern ................................................ 8
Chapter 2
Proposal for international agreement on Ca notation resulting from discussions at workshops
on stable isotope measurements held in Davos (Goldschmidt 2002) and Nice (EGS-AGUEUG 2003)............................................................................................................................... 11
Chapter 3
Tropical SST history inferred from Ca isotope thermometry over the last 140ka
1. Introduction............................................................................................................. 18
2. Sample and core locations ...................................................................................... 19
3. Sample preparation and Ca isotope analyses.......................................................... 20
4. Ca isotope thermometry.......................................................................................... 21
4.1 δ44/40Ca-temperature calibrations ............................................................. 21
4.2 Inter-calibration comparison.................................................................... 22
4.3 Preservation of the primary temperature signal....................................... 23
5. Application of Ca isotopes as SST proxy ............................................................... 24
5.1 Down-core variations on fossil G. sacculifer of core GeoB1112............ 24
5.2 SSTd44/40Ca estimates and comparison to recent conditions ...................... 27
5.3 Past sea surface temperature reconstruction in the tropical Atlantic....... 27
5.4 Phasing of SST proxy signals.................................................................. 30
6. Summary and conclusion........................................................................................ 31
Appendix 3: SST estimates derived from oxygen isotopes ........................................ 32
Supplementary data................................................................................ 33
Table of contents
Chapter 4
Genetic diversity and implications for Ca isotope thermometry
1. Introduction............................................................................................................. 44
2. Results and discussion ............................................................................................ 46
Appendix 4.................................................................................................................. 50
Chapter 5
A critical assessment of mollusk 18O/16O, Mg/Ca, and
Cretaceous seawater temperature seasonality
44
Ca/40Ca ratios as proxies for
1. Introduction............................................................................................................. 58
2. Carbonate material, methods and sampling strategy .............................................. 58
2.1 Sampling locality and age of carbonate material..................................... 58
2.2 Methods ................................................................................................... 60
2.3 Assessment of shell preservation............................................................. 62
2.4 Rudist growth-rates and sampling strategy.............................................. 64
3. Results..................................................................................................................... 65
3.1 Oxygen and carbon isotope ratios............................................................ 65
3.2 Elemental composition ............................................................................ 67
3.3 Magnesium/calcium molar ratios ............................................................ 68
3.4 Ca isotope ratios ...................................................................................... 68
4. Discussion............................................................................................................... 69
4.1 18O/16O and 13C/12C ratios......................................................................... 69
4.2 Mg/Ca molar ratios.................................................................................. 73
4.3 44Ca/40Ca ratios......................................................................................... 74
5. Conclusion .............................................................................................................. 76
Appendices
A.I Analytical protocol for the determination of Ca isotopic composition ...................I
A.II Publications and conference abstracts................................................................VII
A.III Curriculum vitae ................................................................................................ XI
Chapter 1: PhD-thesis Dorothee Hippler, Bern (2004)
Ca isotopic composition of various reference materials and seawater
Dorothee Hippler1, Anne-Désirée Schmitt2, Nikolaus Gussone3, Alexander Heuser3, Peter
Stille2, Anton Eisenhauer3 and Thomas F. Nägler1
1
2
3
Institute of Geological Sciences, University of Bern, Switzerland
Centre de Géochimie de la Surface (CNRS), Strasbourg Cedex, France
GEOMAR, Forschungszentrum für Marine Geowissenschaften, Kiel, Germany
Geostandards Newsletters 2003, Vol. 27 (1), 13-19.
Abstract - A compilation of δ44/40Ca (δ44/42Ca) data sets of different calcium standard material is
presented, based on measurements in three different laboratories (Institute of Geological Sciences,
Bern; Centre de Géochimie de la Surface, Strasbourg; GEOMAR, Kiel) to support the establishment of
a Ca isotope reference standard. Samples include a series of international and internal Ca reference
materials, including NIST SRM 915a, seawater, two calcium carbonates and a CaF2 reference sample.
The deviations in δ44/40Ca for selected pairs of reference samples have been defined and are consistent
within statistical uncertainties in all three laboratories. Emphasis has been placed on characterising
both NIST SRM 915a as an internationally available high purity Ca reference sample and seawater as
representative of an important and widely available geological reservoir. The difference between
δ44/40Ca of NIST SRM 915a and seawater is defined as -1.88 ± 0.04‰ (δ44/42CaNIST SRM 915a/Sw = -0.94 ±
0.07‰). The conversion of values referenced to NIST SRM 915a to seawater can be described by the
simplified equation δ44/40CaSa/Sw = δ44/40CaSa/NIST SRM 915a –1.88 (δ44/42CaSa/Sw = δ44/42CaSa/NIST SRM 915a – 0.94).
We propose the use of NIST SRM 915a as general Ca isotope reference standard, with seawater being
defined as major reservoir with respect to oceanographic studies.
Ca isotopic composition of reference materials and seawater
1. Introduction
Calcium (Ca) is an abundant element both in marine and in terrestrial systems (Broecker &
Peng 1982). Previous investigations have shown that calcium isotopes are a powerful tool for
studying geological, biological and climate-related processes (Russell et al. 1978, Skulan et
al. 1997, Zhu & MacDougall 1998, Skulan & DePaolo 1999, Nägler et al. 2000, De La Rocha
& DePaolo 2000). The significance of calcium isotope data for palaeoceanography has
recently been demonstrated (Nägler et al. 2000, De La Rocha & DePaolo 2000).
Variations of the natural isotopic composition of Ca isotopes occur due to mass-dependent
fractionation or may be introduced by the production of 40Ca from the radioactive decay of
40
K. This latter process has been used in geochronology of igneous rocks as well as minerals
(Coleman 1971, Marchall & DePaolo 1982, Fletscher et al. 1997, Nägler & Villa 2000). By
applying high precision mass-spectrometric methods and a double spike technique relatively
small natural mass-dependent fractionation of calcium isotopes can be resolved (Russell et al.
1978). The latter represents a reconnaissance study on the determination of calcium isotopic
fractionation in meteoritic and terrestrial samples. Subsequent work deals with the biological
control of the calcium isotopic abundance and its significance for biochemistry (Skulan et al.
1997, Skulan & DePaolo 1999). Studies on cultured and natural single species planktonic
foraminifera have shown that temperature-dependant biological fractionation occurs. Based
on these investigations calcium isotopes are established as an alternative sea-surface
temperature proxy (Zhu & MacDougall 1998, Nägler et al. 2000). Furthermore the calcium
isotope ratio of modern seawater is homogenous in oceans worldwide within analytical
uncertainties (Zhu & MacDougall 1998, De La Rocha & DePaolo 2000, Schmitt et al. 2001).
This homogeneity is expected in view of the long residence time of calcium (1Ma) compared
to the mixing time of ocean water of about 1.5 ka (c.f. Broecker & Peng 1982). In addition,
the knowledge of Ca isotopic composition of seawater, river, terrestrial and biological
samples has made it possible to use Ca isotopes as a tool for quantifying the marine calcium
cycle (Zhu & MacDougall 1998, De La Rocha & DePaolo 2000).
An outstanding problem is that direct comparison of published data sets is hampered by the
use of various reference materials. At present there is no uniform way of standardisation and
the apparent differences between the reference samples in use is in the same order of
magnitude as natural variations. Therefore, a definition of a common Ca isotopic standard is
highly desirable. Previous studies dealing with Ca isotopes refer mostly to Ca salts as
laboratory standards. Russell et al. (1978) gave a precisely defined reference value for the
42
Ca/44Ca ratio of 0.31221±0.00002 deduced from two Ca terrestrial reference materials, two
lunar samples and four meteorites. Skulan et al. (1997) normalised the unspiked sample to
this 42Ca/44Ca standard value. They also set an arbitrary, but reasonable, value for their
44
Ca/40Ca ultrapure CaCO 3 standard ratio of 0.0217 and referred all data to this value. The net
result is that quite different isotopic composition values for seawater, which is the only
common and comparable sample for several Ca isotope studies, have been published (Schmitt
2
Chapter 1: PhD-thesis Dorothee Hippler, Bern (2004)
et al. 2001). Russell et al. (1978) and Nägler and Villa (2000) used natural fluorite as a Ca
reference material free of industrially produced or biological isotopic fractionation. In order to
avoid the problem of interlaboratory bias Zhu & MacDougall (1998) and Schmitt et al. (2001)
proposed the use of seawater as a common reference material.
Seawater is widely available and represents a major natural Ca reservoir. However, the
inconvenience is that Ca isotope measurements of seawater require a calcium separation
technique prior to analyses. The CaCO3 reference powder of the National Institute of
Standards and Technology (NIST) SRM 915a, used by Halicz et al. (1999), would be an
alternative and is proposed here as the future reference standard.
An interlaboratory Ca workshop in Autumn 2001 held at the GEOMAR research institute in
Kiel, Germany encouraged us to compile our data of international and in-house reference
materials. The main goal of this study was twofold: (a) Several reference samples (including
seawater) used in previous studies (Russell et al. 1978, Skulan et al. 1997, Zhu & MacDougall
1998, Nägler et al. 2000, Schmitt et al. 2001) were to be cross-calibrated; (b) the Ca isotope
fractionation between NIST SRM 915a and seawater was to be defined. The co-operation of
the three institutes at Bern, Strasbourg and Kiel allowed inter-comparisons to be made that
were independent of particular laboratory techniques.
2. Sample material
NIST SRM 915a is a Ca certified reference material provided by the National Institute of
Standards and Technology, USA. Aliquots of Johnson Matthey Ca carbonate reference
samples Lot. 9912 and Lot. 4064 were provided by courtesy of Dr. A. Papanastassiou. These
reference samples have been previously measured at the CALTECH by Russell et al. (1978).
Large Ca fractionation has been observed for Lot. 4064 because the CaCO3 had been
industrially purified (distillation process) from Ca metal that may not have an unfractionated
initial composition (Russell et al. 1978). These reference samples were included to link our
data to those of Russell et al. (1978). Because of their wide isotopic range (≈12‰ δ44/40Ca),
they would also magnify any potential systematic interlaboratory bias.
A solution obtained from a natural fluorite (CaF2) sample used as a reference sample in Bern
(Nägler et al. 2000) and Kiel (Heuser et al. 2002) was added for comparison. IAPSO
(International Association for the Physical Sciences of the Oceans) reference seawater (batch
P135) is a reference seawater material from the “Ocean Scientific International” and is used at
GEOMAR as a seawater reference. Various seawater samples were taken from different
localities and depths. Seawater samples measured in Bern are listed in Table 3. Seawater
samples from Southern England, Mauritania, Canary Islands and Japan were measured at
Strasbourg (Schmitt et al., 2001).
3
4
±0.18
±0.05
±0.06
±0.25
±0.07
±0.11
-12.78
-0.90
0.48
CaF2
Johnson Matthey Lot. 4064
Johnson Matthey Lot. 9912
Mean seawater
0.24
4
-0.45 2
-6.39 2
29
-0.74 8
0.00
-1.41
0.41
-0.83
-12.78
0.00
Ca
-11.88
Johnson Matthey Lot. 4064 - 9912
-0.98
-5.94
±0.18
Ca
±0.07
-11.95 ±0.34
-1.82
44/40
Kiel
Ca δ
±0.09
δ
44/42
δ
Ca
-5.98
-0.91
44/42
±0.13
±0.18
0.29*
±0.29
±0.30
2σ
Ca
-12.07
-1.89
δ
44/40
0.21
-0.42
-6.39
0.00
-0.71
±0.25
δ
Ca
-5.99
-0.93
44/42
5
3
1
88
109
δ44/42Ca n
±0.07
Strasbourg
±0.06
±0.10
±0.03
±0.03
2 σ SE
-11.95
-1.88
Mean δ
±0.17
±0.15
±0.34
±0.08
2σ
±0.13
Calculated
0.00
-0.71
-6.75
-0.95
0.00
-0.69
-6.68
-0.93
-5.97
-0.94
±0.05
±0.07
±0.11
±0.14
±0.07
±0.07
δ44/42Ca δ44/42Ca** 2 σ
Mean δ44/42Ca 2 σ SE
±0.04
±0.09
±0.24
±0.06
2 σ SE
±0.04
Ca error
44/40
Weighted
0.00
-1.42
-13.49
-1.89
δ44/40Ca
Strasbourg
Uncertainties are the 2σ standard error propagated from Table 1.
Propagated errors for Johnson matthey standards are larger due to limited number of analyses (see Table 1).
δ44/42Ca values are calculated for Bern and Kiel (conversion per atomic mass unit, see text for discussion) and measured at Strasbourg (see Table 1).
-1.95
NIST SRM 915a - Seawater
44/40
δ
Bern
Table 2: Relativ difference of δ44/40Ca and δ44/42Ca of reference samples
2σ: 2σ standard deviation
* Single analysis: external standard reproducibility given
2σ SE: 2σ standard error
δ44/40Ca Values of bern and Kiel are normalised to CaF2 reference sample
δ44/40Ca Values of Strasbourg are normalised to seawater
δ44/42Ca Values are calculated (conversion per atomic mass unit, see text for discussion)
** δ44/42Ca Values are measured and normalised to seawater
n Number of analyses
±0.07
±0.04
±0.19
±0.19
-1.47
0.00
NIST SRM 915a
δ44/40Ca
δ44/42Ca n
δ44/40Ca 2 σ
2 σ SE
Kiel
Bern
Table 1: Calcium isotopic composition of different reference materials
15
3
2
2
n
Ca isotopic composition of reference materials and seawater
Chapter 1: PhD-thesis Dorothee Hippler, Bern (2004)
3. Data acquisition and presentation
All three laboratories used 43Ca-48Ca double spikes. Absolute isotope ratios defined with the
double spike technique are dependent on the spike calibration procedure and the
normalisation standard. Thus, we present our data as per mil differences relative to a given
reference value. The isotopic composition of calcium is expressed in the δ-notation (δ44/40Ca
and δ44/42Ca), with the heavy isotope in the nominator using the same convention as in light
stable isotope analysis (e.g. δ18O):
δ44/40Ca = [(44Ca/40Ca)sample / (44Ca/40Ca)reference – 1] * 1000
[1]
δ44/42Ca = [(44Ca/42Ca)sample / (44Ca/42Ca)reference - 1] * 1000
[2]
Presenting results in both notations is necessary here, as both are currently in use. 40Ca and
44
Ca are the two most abundant Ca isotopes and provide a mass range of four atomic mass
units, i.e. a relative mass difference of 10%. Most publications to date have applied these two
isotopes to present their data. Halicz et al. (1999) proposed the use of 42Ca instead of 40Ca.
Their instrumental set up (multi-collector ICP-MS) did not allow accurate definition of 40Ca
due to the presence of 40Ar from the Ar plasma. An advantage of the use of 42Ca instead of
40
Ca is the absence of any radiogenic increase in that isotope (see below) allowing the
definition of the natural fractionation using a single measurement on any terrestrial sample.
On the other hand, the relative isotope abundance of 42Ca is low (0.647% of total Ca).
Moreover, the lower mass difference between 42Ca and 44Ca results in lower natural variation
of the 42Ca/44Ca ratio and so reduces the resolution of the method to one half.
Radiogenic increase in 40Ca caused by 40K decay in a given sample can be determined by
measuring an unspiked aliquot and normalise the 40Ca/44Ca ratio to another stable Ca isotope
ratio. However, due to the low relative isotopic abundance of 40K (0.0117% of total K) and
the high relative isotopic abundance of 40Ca (96.941% of total Ca), only minerals or rocks
with K concentrations exceeding Ca concentrations show a significant radiogenic 40Ca
increase. Based on these theoretical expectations, samples with Ca concentrations greater than
K concentrations are generally assumed to be void of radiogenic 40Ca variations (e.g. Skulan
et al. 1997, De La Rocha and DePaolo 2000). Furthermore, none of the samples measured by
Russell et al. (1978) or Zhu and MacDougall (1998) showed any isotopic variation except
natural mass dependent fractionation, noting that unspiked runs were executed in these
studies.
In the present study, unspiked runs were performed in Bern and Kiel on all reference samples
except seawater. None of them revealed any significant increase in 40Ca. Further, δ44/42Ca
values measured directly in Strasbourg (including seawater) were identical to δ44/42Ca values
calculated from δ44/40Ca assuming the absence of radiogenic 40Ca: δ44/42Ca =
{(δ44/40Ca)/(43.956-39.963)}*(43.956-41.959). Therefore, significant discrepancies in Ca
5
Ca isotopic composition of reference materials and seawater
isotopic composition due to radioactive decay of 40K can be excluded for the samples
analysed here. Details of analytical techniques are given elsewhere (Nägler et al. 2000,
Schmitt et al. 2001, Heuser et al. 2002) The method of seawater purification used in Bern is
described in the Appendix 1.
4. Results and Discussion
The compiled results of different measurements on reference samples are presented in Table
1. Uncertainties are given at the 2σ level. Data sets from Bern and Kiel are normalised to the
CaF2 reference sample, whereas the data set of Strasbourg is referred to an average value of
ocean water. Mean seawater measurements at Bern and Kiel resulted in δ44/40Ca values of
0.48±0.11 and 0.41±0.13 relative to the CaF2 reference sample, respectively. The apparent
bias when compared with the data from Strasbourg in Table 1 is the direct consequence of
these different normalisation procedures. When referred to the same reference sample, values
of all three laboratories are identical within analytical uncertainties, precluding any
interlaboratory bias.
Table 2 shows the relative deviations in δ44/40Ca ( δ44/42Ca) of two pairs of selected reference
samples relative to each other. The deviations are defined for NIST SRM 915a and seawater
and for the two Johnson Matthey (Lot. 4064 and 9912) reference samples. Uncertainties
reflect the 2σ standard error propagated from Table 1. Propagated errors for the difference of
the two Johnson Matthey reference samples are larger due to the limited number of analyses.
Table 3: The Ca isotopic composition of modern seawater measured at the Institute of Geological
Sciences, Bern
Sample
Depth (m)
δ44/40Ca
2 σ SE
δ44/42Ca
Indian Ocean (Southern Arabian Sea)
24 CTD
Deep Sea
0.52
±0.12
0.26
Atlantic Ocean (S. England)
ATL-OC
Surface
0.42
±0.15
0.21
Pacific Ocean (North Rift Zone)
52 CTD
1637
0.45
±0.15
0.23
IAPSO standard seawater (batch P135)
IAPSO
-
0.54
±0.19
0.27
Location
2σ SE 2σ standard error of the measurement
δ44/40Ca values of modern seawater are normalised to CaF2 reference sample
δ44/42Ca values are calculated (conversion per atomic mass unit, see text for discussion)
Sample ATL-OC was previously analysed by Schmitt et al. (2001)
The weighted average of the difference in δ44/40Ca between NIST SRM 915a and seawater is
calculated to be -1.88±0.04‰ and -11.95±0.13‰ for the two Johnson Matthey reference
samples. The corresponding means and uncertainties (2σ) in δ44/42Ca are -0.94±0.07‰ and
–5.97±0.05‰, respectively. Again, the results of all three laboratories are identical within
uncertainties. In addition, the deviation of both Johnson Matthey reference samples measured
6
Chapter 1: PhD-thesis Dorothee Hippler, Bern (2004)
by Russell et al. (1978, δ44/40Ca = -11.83±0.28‰) is identical within uncertainties to the
weighted average of the three laboratories.
Present and previous measurements at Bern (Table 3), GEOMAR and Strasbourg of Ca
isotopic composition of modern seawater, including IAPSO, support earlier observations of
Ca-isotopic homogeneity between the Earth’s oceans and at different water depths.
Based on the general δ-equation the δ-values of samples can be easily converted to the
different reference standards as long as the conversion factor of both standards is known.
Accordingly, the δ44/xCa values normalised to NIST SRM 915a can be converted to δ44/xCa
values related to seawater in the following manner:
δ44/xCaSa/Sw = [(10-3 δ44/xCaSa/NIST SRM 915a +1) (10-3 δ44/xCaNIST SRM 915a/Sw +1) -1]*103
[3]
The subscripted indices in the formula correspond to the deviation sample-seawater
(δ44/xCaSa/Sw) and sample-NIST SRM 915a (δ44/xCaSa/NIST SRM 915a). X represents either mass 40 or
42. As listed in Table 2, the weighted average value to describe the conversion factor between
NIST SRM 915a and seawater (δ44/40CaNIST SRM 915a/Sw) is -1.88 (δ44/42CaNIST SRM 915a/Sw = -0.94).
Within the existing analytical precision, equation [3] can accurately be approximated by
equations [4] and [5]:
δ44/40CaSa/Sw = δ44/40CaSa/SRM 915a – 1.88
[4]
δ44/42CaSa/Sw = δ44/42CaSa/SRM 915a – 0.94
[5]
The definition of the difference between NIST SRM 915a and seawater provides a necessary
reference for published and future oceanographic studies.
5. Summary and Conclusion
A set of Ca reference materials provides consistent δ44/40Ca ratios (and corresponding
calculated and measured δ44/42Ca ratios) within statistical uncertainties in three European
laboratories (Bern, GEOMAR, Strasbourg). The relative deviation of δ44/40Ca (δ44/42Ca) of two
pairs of reference samples (NIST SRM 915a and seawater, Johnson Matthey Lot. 4064 and
Lot. 9912) shows very good correspondence, the latter being further compatible with previous
results of Russell et al. (1978). Therefore, no interlaboratory bias correction is necessary and
the Ca isotope data of the three laboratories are directly comparable with each other when
expressed in δ44/40Ca or δ44/42Ca notation.
In conclusion, NIST SRM 915a as a widely available Ca certifies reference material, and
seawater, as an important geological reservoir, have been characterised here and NIST SRM
7
Ca isotopic composition of reference materials and seawater
915a is proposed as the international standard to which measurements should be referenced.
The difference of δ44/40Ca of NIST SRM 915a to seawater is defined as -1.88±0.04‰
(δ44/42CaNIST SRM 915a/Sw = -0.94 ± 0.07‰). The simplified descriptive equation for this relation is
δ44/40CaSa/Sw = δ44/40CaSa/NIST SRM 915a – 1.88 (δ44/42CaSa/Sw = δ44/42CaSa/NIST SRM 915a – 0.94).
Acknowledgements
We are greatly indebted to D.A. Papanastassiou for providing standard material. D.H. and
T.F.N. wish to thank Igor Villa for fruitful discussions. Constructive review by two
anonymous referees is gratefully acknowledged. Ca isotope work at Bern was supported by
the Swiss National Science Foundation (grants 21-61644.00 and 21-60987.00). The work of
N. Gussone, A. Heuser and A. Eisenhauer was supported by a grant from the “Deutsche
Forschungsgemeinschaft, DFG (Ei272/12-1)”. A.-D. Schmitt profited from an MRT grant.
This is the EOST contribution N° 2003.602-UMR7517.
Appendix 1
Purification of seawater samples in Bern
5µl seawater was processed through chemical separation. A 43Ca-48Ca double spike was added
to the samples prior to purification of Ca. Microcolumns containing 10µl AP-MP-50 (200400 mesh) cation exchange resin and a HCl chemistry were used. This was necessary to
remove remaining interfering elements such as potassium and strontium that might act as
potential isobaric interferences. The chemical separation efficiency for Ca was close to 100%.
Total procedural blanks in our work were below 1ng and therefore had a negligible effect on
the isotopic data.
8
Chapter 1: PhD-thesis Dorothee Hippler, Bern (2004)
References
Broecker W. S. and Peng T.-H. (1982)
Tracers in the Sea. Eldigo Press (New York), 690pp.
Coleman M.L. (1971)
Potassium-calcium dates from pegmatic micas. Earth and Planetary Science Letters, 12, 399405.
De La Rocha C. L. and DePaolo D. J. (2000)
Isotopic evidence for variations in the marine Calcium cycle over the Cenozoic. Science, 289,
1176-1178.
Fletscher I. R., McNaughton N. J., Pidgeon R. T. and Rosman K. J. R. (1997)
Sequential closure of K-Ca and Rb-Sr systems in Archaean micas. Chemical Geology, 138,
289-301.
Halicz L., Galy A., Belshaw N. S. and O'Nions R. K. (1999)
High-precision measurements of calcium isotopes in carbonates and related materials by
multiple collector inductively coupled plasma mass-spectrometry (MC-ICP-MS). Journal of
Analytical Atomic Spectrometry, 14, 1835-1838.
Heuser A., Eisenhauer A., Gussone N., Bock B., Hansen B. T. and Nägler T. F. (2002)
Measurements of calcium isotopes (δ44Ca) using a multicollector TIMS technique.
International Journal of Mass Spectrometry, 220, 385-397.
Marshall B. D. and DePaolo D. J. (1982)
Precise age determination and petrogenetic studies using the K-Ca method. Geochimica et
Cosmochimica Acta, 46, 2537-2545.
Nägler T. F. and Villa I. M. (2000)
In pursuit of the 40K branching ratios: K-Ca and 39Ar-40Ar dating of gem silicates. Chemical
Geology, 169, 5-16.
Nägler T. F., Eisenhauer A., Müller A., Hemleben C. and Kramers J. (2000)
The δ44Ca-temperature calibration on fossil and cultured Globigerinoides sacculifer: New tool
for reconstruction of past sea surface temperatures. Geochemistry Geophysics Geosystems, 1,
2000GC000091.
Russell W. A., Papanastassiou D. A. and Tombrello T. A. (1978)
Ca isotope fractionation on the earth and other solar system materials. Geochimica et
Cosmochimica Acta, 42, 1075-1090.
9
Ca isotopic composition of reference materials and seawater
Schmitt A.-D., Bracke G., Stille P. and Kiefel B. (2001)
The Calcium isotope composition of modern seawater determined by Thermal Ionisation
Mass Spectrometry. Geostandards Newsletter, 25, 267-275.
Skulan J., DePaolo D. J. and Owens T. L. (1997)
Biological control of calcium isotopic abundances in the global calcium cycle. Geochimica et
Cosmochimica Acta, 61, 2505-2510.
Skulan J. and DePaolo D. J. (1999)
Calcium isotope fractionation between soft and mineralized tissues as a monitor of Calcium
use in vertebrates. Proc. Natl. Acad. Sci., Biochemistry, 96: 13709-13713.
Zhu P. and MacDougall J. D. (1998)
Calcium isotopes in the marine environment and the oceanic calcium cycle. Geochimica et
Cosmochimica Acta, 62, 1691-1698.
10
Chapter 2: PhD-thesis Dorothee Hippler, Bern (2004)
Proposal for International agreement on Ca notation resulting from
discussions at workshops on stable isotope measurements held in
Davos (Goldschmidt 2002) and Nice (EGS-AGU-EUG 2003)
Anton Eisenhauer1, Thomas Nägler2, Peter Stille3, Jan Kramers2, Nikolaus Gussone1, Barbara
Bock1, Jan Fietzke1, Dorothee Hippler2 and Anne-Désirée Schmitt3
1
Leibniz-Institut für Meereswissenschaften (IfM-GEOMAR), Kiel, Germany
2
3
Isotopengeologie, Institut für Geologie, Universität Bern, Switzerland
Centre de Géochimie de la Surface (CNRS), Strasbourg Cédex, France
Geostandards and Geoanalytical Research (2004) Vol. 28 (1), 149-151.
Abstract - A proposal is made to standardise the reporting of Ca isotope data to the δ44/40Ca notation (or
δ44/42Ca) and to adopt NIST SRM 915a as the reference standard.
Proposal for International Agreement on Ca notation
The purpose of this short note is to summarise the results of discussions held during the
Goldschmidt Conference (2002) in Davos and the EGS meeting (2003) in Nice at stable
isotope reference materials workshops. Among other people interested in the measurement of
stable isotopes, representatives of the University of Cambridge, the University of Oxford, the
Scripps Institute of Oceanography, the University of Bern, the University of Strasbourg and
the Leibniz Institute for Marine Sciences in Kiel contributed actively to the ongoing
discussion concerning the nomenclature of Ca isotope measurements.
At present, calcium (Ca) isotope fractionation can be measured either by thermal ionisation
mass spectrometry (TIMS) or by the multiple collector-inductively coupled plasma-mass
spectrometry (MC-ICP-MS) technique. Applying the TIMS technique, Ca isotopes of interest,
usually 40Ca (96.9%), 42Ca (0.65%), 43Ca (0.14%), 44Ca (2.09%) and 48Ca (0.19%), as well as
interfering isotopes can be determined by peak jumping on a single Faraday cup or by a
multicollector technique as presented by (Heuser et al. 2003). Using TIMS, a double spike,
either 43Ca/46Ca (Fletcher et al. 1997a), or 42Ca/48Ca (Russell et al. 1978, Zhu and MacDougall
1998; Skulan and DePaolo, 1999; Bullen et al. 2003) or 43Ca/48Ca (Nägler et al. 2000) is
required. With a MC-ICP-MS either the so-called “hot-plasma-technique” (HPT) (Halicz et
al. 1999) or the “cold-plasma-technique” (CPT) (Fietzke et al. 2003) can be used; both
determine Ca isotope fractionation by the bracketing reference sample technique. The
advantage of the CPT (Fietzke et al., 2003) over the HPT is that all Ca isotopes and in
particular 40Ca can be measured by MC-ICP-MS without the use of a collision cell. This
means that 40Ca and 44Ca can be measured in the same manner as with TIMS due to a
significant reduction of 40Ar interferences. Further important advantages of CPT arise from a
significant suppression of doubly-charged 88Sr and 84Sr isotopes, which usually interfere with
44
Ca and 42Ca, respectively. As a consequence, for carbonate samples no chemical purification
has to be performed prior to the measurement. This significantly reduces the laboratory effort
per sample and the uncertainty associated with natural Ca isotope fractionation, introduced by
the use of ion chromatography.
In order to increase the comparability between laboratories using different techniques and
notations, we propose that Ca isotope data based on TIMS and CPT MC-ICP-MS
measurements are presented in the following δ44/40Ca notation defined in equation [1].
Alternatively, laboratories that must or prefer to measure 42Ca and 44Ca should present their
data in the analogous δ44/42Ca notation (equation [2]).
δ44/40Ca = [(44Ca/40Ca)Sample / (44Ca/40Ca)Reference standard –1] *1000
[1]
δ44/42Ca = [(44Ca/42Ca)Sample / (44Ca/42Ca)Reference standard –1] *1000
[2]
At the moment a wide range of reference samples are used, including calcium fluoride (CaF2,
Nägler et al. 2000), seawater (Zhu and Macdougall 1998, Schmitt et al. 2001, 2003a) or
various carbonate reference materials such as NIST SRM 915a (Halicz et al. 1999, Heuser et
12
Chapter 2: PhD-thesis Dorothee Hippler, Bern (2004)
al. 2002), Berkeley Carbonate Standard (Skulan et al. 1997, Skulan and DePaolo 1999, De La
Rocha and DePaolo 2000) and ANU Tridacna standard (Fletcher et al. 1997a, b). Obviously,
an agreement on data presentation and material to serve as the reference standard is highly
desirable. The reference standard recommended by IUPAC is NIST SRM 915a (Coplen et al.
2002), and we encourage the use of this reference material. Below, a conversion is given to
show how to recalculate from seawater-normalised δ-values to SRM 915a-normalised values
(Hippler et al. 2003). However, in order to avoid chemical purification of Ca from the
designated reference material prior to TIMS or HPT analysis, a pure IRMM metal reference
material may additionally be recommended for future interlaboratory calibration work.
Disagreement exists also on the subject of which isotope to use as the reference isotope (40Ca
or 42Ca). Using 40Ca has several advantages: 40Ca and 44Ca have the highest abundances and
therefore can easily be measured by TIMS and CPT MC-ICP-MS (Fietzke et al. 2003).
Furthermore, due to the mass difference of 4 atomic mass units (amu) between 44Ca and 40Ca,
mass fractionation is statistically better resolvable in this pair of isotopes compared to 42Ca
and 44Ca. However, problems in using 40Ca as the reference isotope may arise from the fact
that 40Ca is a radiogenic isotope and its abundance depends on the radioactive decay of 40K
with a half-life of about 1.28*109 years. In natural samples, contributions of 40K are only
relevant in very old material with extremely high K/Ca ratios, because 40K is the least
abundant K isotope (0.01%). Furthermore, any deviation of the measured Ca isotope variation
from mass dependent fractionation can easily be verified by measuring Ca isotope
fractionation in the samples via spiked and unspiked aliquots. Up to now no deviation from
normal mass dependent isotope fractionation has been reported in marine carbonate samples
(Schmitt et al. 2003a, b).
The radiogenic ingrowth problem arising from the use of 40Ca as normalising isotope can be
overcome by the use of 42Ca as a reference isotope. The use of the 44Ca and 42Ca isotope pair
is necessary for the HPT MC-ICP-MS measurements because 40Ca intensities are impossible
to determine due to isobaric interferences with 40Ar. However, the measurement of the 44Ca
and 42Ca pair is more difficult with TIMS and CPT MC-ICP-MS because of the low 42Ca
intensity (only 0.65%). This means that the smaller fractionation effect (only two amu
difference) has to be identified using less precise measurements. A further problem is that
42
Ca is often used as spike isotope (see above).
We recommend that the calculated δ44/42Ca values should also be published. The equation
used for that conversion, published by (Hippler et al. 2003), is given below (equation [3])
δ44/40Ca = δ44/42Ca*[(43.956-39.963)/(43.956-41.959)]
[3]
Similarly, Ca isotope fractionation based on different reference samples can also be converted
using equations [4] and [5], originally given in Hippler et al. (2003):
13
Proposal for International Agreement on Ca notation
δ44/40CaSa/Sw = δ44/40CaSa/NIST SRM 915a - 1.88
[4]
δ44/42CaSa/Sw = δ44/42CaSa/NIST SRM 915a - 0.94
[5]
Where Sa/Sw is the sample normalised to seawater and Sa/NIST SRM 915a is the sample
normalised to NIST SRM 915a.
References:
Bullen T. D., Kim S. T. and Paytan A. (2003)
Ca isotope fractionation during Ca-carbonate precipitation: There’s more to it than
temperature. Goldschmidt Conference Abstracts A 49.
Coplen T. B., Böhlke J. K., De Bievre P., Ding T., Holden N. E., Hopple J. A., Krouse R.,
Lamberty A., Peiser H. S., Revesz K., Rieder S. E., Rosman K. J. R., Roth E., Taylor P. D. P.,
Vocke R. D. J. and Xiao Y. K. (2002)
Isotope-abundance variations of selected elements (IUPAC Technical Report). Pure and
Applied Chemistry 74(10), 1987-2017.
De La Rocha C. and DePaolo D. J. (2000)
Isotopic evidence for variations in the marine calcium cycle over the Cenozoic. Science 289,
1176-1178.
Fietzke J., Eisenhauer A., Liebetrau V., Bock B., Gussone N., Nägler Th. F., Dietzel M.,
Spero H. J., Bijma J. and Dullo C. (2003)
A new method for 44Ca/40Ca determination using cool plasma MC-ICP-MS. Geophysical
Research Abstracts 5, EAE03-A-11280.
Fletcher I. R., Maggi A. L., Rosman K. J. R. and McNaughton N. J. (1997a)
Isotopic abundance measurements of K and Ca using a wide-dispersion multi-collector mass
spectrometer and low-fractionation ionisation techniques. International Journal of Mass
Spectrometry and Ion Processes 163, 1-17.
Fletcher I. R., McNaughton N. J., Pidgeon R. T. and Rosman K. J. R. (1997b) Sequential
closure of K-Ca and Rb-Sr isotopic systems in Archean micas. Chemical Geology 138, 289301.
Gussone N., Eisenhauer A., Heuser A., Dietzel M., Bock B., Böhm F., Spero H.J., Lea D. W.,
Bijma J. and Nägler Th. F. (2003)
Model for kinetic effects on calcium isotope fractionation (δ44Ca) in inorganic aragonite and
cultured planktonic foraminifera. Geochimica et Cosmochimica Acta 67(7), 1375-1382.
14
Chapter 2: PhD-thesis Dorothee Hippler, Bern (2004)
Halicz L., Galy A., Belshaw N. and O´Nions R. (1999)
High-precision measurement of calcium isotopes in carbonates and related materials by
multiple collector inductively coupled plasma mass spectrometry (MC-ICP-MS). Journal of
Analytical Atomic Spectrometry 14, 1835-1838.
Heuser A., Eisenhauer A., Gussone N., Bock B., Hansen B. T. and Nägler T. F. (2002)
Measurement of calcium isotopes (δ44Ca) using a multicollector TIMS technique.
International Journal of Mass Spectrometry 220, 385-397.
Hippler D., Schmitt A.-D., Gussone N., Heuser A., Stille P., Eisenhauer A. and Nägler Th. F.
(2003)
Ca isotopic composition of various standards and seawater. Geostandards Newsletter, 27(1),
13-19.
Nägler T., Eisenhauer A., Müller A., Hemleben C. and Kramers J. (2000)
The δ44Ca-temperature calibration on fossil and cultured Globigerinoides sacculifer: New tool
for reconstruction of past sea surface temperatures. Geochemistry, Geophysics, Geosystems
1(2000GC000091).
Russel W. A., Papanastassiou D. A. and Tombrello T. A. (1978)
Ca isotope fractionation on the Earth and other solar system materials. Geochimica et
Cosmochimica Acta 42, 1075-1090.
Schmitt A.-D., Bracke G., Stille P. and Kiefel B. (2001)
The calcium isotope composition of modern seawater determined by thermal ionisation mass
spectrometry. Geostandards Newsletter 25(2-3), 267-275.
Schmitt A. D., Chabaux F. and Stille P. (2003a)
The Ca riverine and hydrothermal isotopic fluxes and the oceanic calcium mass balance.
Earth and Planetary Science Letters, 213(3-4), 503-518.
Schmitt A. D., Stille P. and Vennemann T. (2003b)
Variations of the 44Ca/40Ca ratio in seawater during the past 24 million years: Evidence from
δ44Ca and δ18O values of Miocence phosphates. Geochimica et Cosmochimica Acta 67(14),
2607-2614.
Skulan J. and DePaolo D. J. (1999)
Calcium isotope fractionation between soft and mineralized tissues as a monitor of calcium
use in vertebrates. Proc. Nat. Acad. Sci. USA 96(24), 13709-13713.
Skulan J. L., DePaolo D. J. and T.L. Owens (1997)
Biological control of calcium isotopic abundances in the global calcium cycle. Geochimica et
Cosmochimica Acta 61, 2505-2510.
15
Proposal for International Agreement on Ca notation
Zhu P. and MacDougall D. (1998)
Calcium isotopes in the marine environment and the oceanic calcium cycle. Geochimica et
Cosmochimica Acta 62(10), 1691-1698.
16
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
Tropical SST history inferred from Ca isotope thermometry
over the last 140ka
Dorothee Hippler1, Anton Eisenhauer2 and Thomas F. Nägler1
1
2
Institute of Geological Sciences, University of Bern, Bern, Switzerland
Leibniz-Institut für Meereswissenschaften (IfM-GEOMAR), Kiel, Germany
Manuscript submitted to Geochimica et Cosmochimica Acta
A b s t a r c t - We present a refinement of the published δ44/40 Ca-temperature calibration investigating
modern specimens of planktonic foraminifera Globigerinoides sacculifer and apply this to sea surface
temperature (SST) reconstructions over the last two glacial-interglacial cycles. Reproduced
measurements of modern Globigerinoides sacculifer collected from surface waters describe a linear
relationship for the investigated temperature range (19.0 to 28.5°C): δ44/40 Ca [‰] = 0.22(±0.05)*SST
[°C] -4.88. Thus a change of δ44/40 Ca [‰] of 0.22 (±0.05) corresponds to a relative change of 1°C. The
refined δ44/40 Camodern -calibration allows the determination of both relative temperature changes and
absolute temperatures. This δ44/40 Camodern -calibration for G. sacculifer has been applied to the tropical
East Atlantic sediment core GeoB1112 for which other SST proxy data are available. Comparison of
the different data sets gives no indication for any secondary overprinting of the δ44/40 Ca signal. The Ca
isotope record follows glacial-interglacial cycles and suggests that sea surface temperature in the tropical
Atlantic increased by 3.0 to 4.0°C at the last two glacial-interglacial transitions. For both terminations
the increase in SST occurs approximately 3ka before the Northern Hemisphere ice caps melted as
reflected by the oxygen isotope record. The latter reveal new evidence for the important role of the
tropics in triggering global climate change, based on a new independent SST proxy.
Tropical SST history inferred from Ca isotope thermometry
1. Introduction
Sea surface temperature (SST) is one of the most important variables for the Earth’s
climate system. It controls and influences the air-sea gas exchange,
evaporation/precipitation patterns and primary biological production. Further it is the
dominant variable controlling salinity, seawater density and deep-ocean circulation driven
by the thermo-haline circulation (conveyor-belt). In particular, the thermo-haline
circulation is partly driven by the temperature difference between the tropics and the
polar oceans. In earlier studies, climate models always emphasised the role of the northern
North Atlantic as a pacemaker of global climate change although there is increasing
evidence that the tropical oceans are at least as important as the high-latitude oceans for
the strength of the conveyer-belt. A key to answer the question of the specific role of the
tropical and/or high-latitude oceans is to reconstruct the differences in the temperature
dynamic of tropical and polar surface waters on glacial-interglacial time scale. In order to
follow this approach a variety of SST proxies, including δ18O and Mg/Ca of foraminiferal
calcite, foraminiferal transfer function and alkenone unsaturation index (UK’37) have been
developed serving as tools in palaeoceanography. Although an ideal proxy is only
influenced by one environmental parameter, recent studies clearly showed that the SST
proxies are effectively influenced by several other parameters apart from temperature.
Thus the results of only one SST proxy might be biased and have to be independently
confirmed by at least one additional SST proxy. Beside this multiproxy approach to
reconstruct past SST variations, the successful development of new and in particular of
isotope methods is still highly desirable.
Ca isotopic composition in primary foraminiferal calcite might serve as such an additional
SST proxy. Skulan et al (1997, 1999) were the first to investigate the biological influence
on the Ca isotopic abundance in a variety of organisms. Comparing the δ44/40 Ca of benthic
and planktonic foraminifera they postulated that the temperature effect is small in
comparison to effects they attributed to the trophic level. In particular Zhu & MacDougall
(1998) showed that among foraminifera, Ca isotopes exhibit species-dependent
fractionation. Some species appeared to vary significantly with ocean water temperature
or depth. Zhu & MacDougall (1998) observed a δ44/40 Ca difference of 0.6‰ for G.
sacculifer between samples of the Holocene and the Last Glacial Maximum for the
equatorial Pacific. Following recent publications on glacial-interglacial SST change in
equatorial Pacific (Lea et al. 2000, Visser et al. 2003), this 0.6‰ difference would reflect
a temperature change of around 3°C. Nägler et al. (2000) published a direct δ44/40 Catemperature calibration on cultured G. sacculifer. Here a shift in δ44/40 Ca of 0.24‰ would
correspond to a temperature change of 1°C. In a first attempt these authors applied this
calibration to an equatorial East Atlantic sediment core (GeoB1112, 5°46.7’S,
10°45.0’W). The decrease of δ44/40 Ca of 0.71±0.24‰ at the Holocene-Last Glacial
Maximum boundary is in line with Zhu & MacDougall observations and corresponds to a
change in SST of ≈3.0 ± 1.0°C. Although δ44/40 Ca and Mg/Ca values generally correlate,
18
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
δ44/40 Ca seemed to predict higher SSTs. Gussone et al. (2003) developed a thermodynamic
model describing the temperature-dependent Ca isotope fractionation behaviour. This
model is based on investigations of the δ44/40 Ca-temperature relationship of inorganically
precipitated aragonite and cultured planktonic foraminifera Orbulina universa in
combination with the findings of Nägler et al. (2000) on cultured G. sacculifer. One
major conclusion is that the temperature-dependent fractionation of Ca isotopes in the
studied carbonate species could be explained by kinetic fractionation and that the degree
of temperature sensitivity cannot be generalised among different species of planktonic
foraminifera. The shallow slope of inorganic precipitates and O. universa reflects the
involvement of Ca 2+-aquacomplexes that are less susceptible to temperature-dependent
kinetic isotope fractionation. In contrast the steep slope of G. sacculifer indicates that the
Ca2+-aquacomplexes are actively dehydrated and that biologically mediated processes at
the seawater-cell interface are responsible for the temperature-controlled transport of pure
Ca2+-ions prior to calcification. The recently published work by Marriott et al. (2004)
favours equilibrium dynamics to describe Ca isotope fractionation, assuming a weaker
bonding of Ca in the carbonate structure relative to that in aqueous structure. The similar
fractionation behaviour of inorganic precipitates and O. universa they interpreted as
direct precipitation of calcite from seawater by a simple process. The stronger temperature
dependence of G. sacculifer is explained as the result of a superimposed additional
biological fractionation effect assuming different biomineralisation processes for both
species.
The objective of this study is twofold: First, in continuation of the earlier work of Nägler
et al. (2000) we attempt to verify the older calibration and to derive a more precise
δ44/40 Ca-temperature relationship based on modern specimens of G. sacculifer captured at
known SST conditions evaluating both relative and absolute temperatures. Second in
order to test the refined relationship we performed Ca isotope analyses of fossil G.
sacculifer on the down-core record from tropical eastern Atlantic (GeoB1112) covering
the last 140.000 years, i.e. the last two glacial-interglacial transitions. We compared the
SSTδ44/40Ca record with available SST proxy data (δ18O, transfer function, Mg/Ca)
determined on the same core material.
2. Sample and Core Locations
Modern samples of G. sacculifer were collected in the South Atlantic Ocean and in
subtropical and tropical North Atlantic on board of RV Meteor during leg M46-2 and leg
M38-1. SST ranges from 19 to 22°C and 25 to 27°C, respectively. Further samples were
obtained on board of RV Sonne (SO 164) in the Caribbean Sea with SST ranging from
27 to 29°C. The samples were collected from the surface water layer by constant
pumping. The seawater was filtered using a plankton net (HYDROBIOS). Samples were
rinsed with freshwater and ethanol. Individuals of G. sacculifer larger than 150µm were
19
Tropical SST history inferred from Ca isotope thermometry
selected, dried and stored for isotope analysis. We focused on individuals with neither
final stage kummerform nor sac-like chamber.
Figure 1: Sample localities of modern
specimens of G. sacculifer collected from surface
waters on vessel cruises in the Caribbean Sea
(SO164), the tropical Atlantic (M38-1) and the
South Atlantic offshore Brazil (M46-2).
Numbers indicate sample numbers (see Table
1). Sediment core GeoB1112 (5°46.2‘S,
10°44.7‘W, 3122m) was recovered from the
Guinea Basin and is located in the low
productive South Equatorial Current (SEC).
Inset: Map of sea surface temperatures of the
eastern tropical Atlantic.
Shells of fossil G. sacculifer were obtained from sediment core GeoB1112 (5°46.2’S,
10°44.7’W). It is located in the low productive South Equatorial Current (SEC) and was
recovered from the Guinea Basin in 3122m water depth. Major information about the
stratigraphy, sedimentology and geochemistry of this core can be found in Wefer et al.
(1996). Stable isotope geochemistry (δ18O), foraminiferal assemblage analyses and
Mg/Ca-derived palaeo-SSTs were previously published by Wefer et al. (1996) and
Nürnberg et al. (2000). Some samples were previously measured by Nägler et al. (2000)
for δ44/40 Ca.
3. Sample Preparation and Ca Isotope Analyses
For Ca isotope analyses on core material about five to ten shells of G. sacculifer were
selected from the 250 to 500µm size fraction and pre-cleaned following the procedure
described in Nürnberg et al. (2000). Around five tests larger than 150µm were picked
from samples of G. sacculifer collected by pumping. In both cases the calcite shells were
dissolved in 2.5 N ultrapure HCl. An aliquot corresponding to 0.5 to 1.0µg Ca was mixed
with a 43Ca-48Ca double spike to correct for isotopic fractionation during the measurement.
In order to eliminate any remaining organic impurities modern samples were treated with
20
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
a H 2O2-HNO 3 solution (Hippler et al. 2004). The sample-spike mixture was dried and
recovered in about 1µl 2.5 N HCl and then loaded on previously washed and outgassed
single Re filament together with 1µl of a Ta 2O5-activator solution (Birck 1986).
All Ca isotope measurements were carried out at the Institute of Geological Sciences at the
University of Bern, Switzerland on a modified single cup AVCO® mass spectrometer
equipped with a Thermolinear® ion source. It was operated in positive ion mode with a
7.6 kV acceleration voltage and a 10 11 Ω resistor setting for the Keithley® electrometer.
Ca isotopes are measured successively in peak jumping mode in descending sequence
(masses 48, 44, 43, 40). During data acquisition K+ and Sr 2+ were continuously monitored
on mass 41 and 43.5 to trace possible isobaric interferences on masses 40, 43 and 44. No
such interferences have been observed at measuring temperatures.
Measured ratios were corrected for analytical isotope fractionation online using a threedimensional data reduction based on an exponential fractionation law (Siebert et al.,
2001). The isotope variations of Ca are expressed in the δ-notation (δ44/40 Ca [‰] =
[( 44Ca/40Ca)sample /(44Ca/40Ca)standard – 1] * 1000). The measured 44Ca/40Ca ratios were
compared to the 44Ca/40Ca ratio of a calcium fluorite standard (CaF2, Hippler et al. 2003).
The 44Ca/40Ca ratios of our internal CaF2 standard reproduced within 0.11‰ (2σ standard
deviation, n=30). The 2σ-reproducibility of our sample δ44/40 Ca [‰] is 0.15 determined
by repeated aliquot measurements. We present our Ca isotope data re-normalised relative
to NIST SRM915a as recently proposed by Eisenhauer et al. (2004) and Coplen et al.
(2002). The δ44/40 Ca value of our internal CaF 2 standard is 1.47±0.04 (‰ SRM915a).
Furthermore, the calculated (δ44/42 Ca values are presented in the tables and are converted
according to the equation published by Hippler et al. (2003). The cross-calibration of
various Ca reference materials and seawater of Hippler et al. (2003) allows direct data
comparison with previous published work.
4. Ca Isotope Thermometry
4.1 δ44/40 Ca -temperature calibrations
The δ44/40 Ca values of G. sacculifer collected from surface waters range from -0.74 to
+1.51 (in ‰ SRM 915a) (Figure 2a and Table 1, see supplementary data). For the
investigated temperature range (19.0 to 28.5°C) we observe a linear trend, although the
individuals were sampled from waters of different salinity, varying from 34.5 to 36.5‰.
Data points represent the weighted means of at least two to three independent analyses.
The resulting δ44/40 Ca-temperature regression is expressed as
δ44/40 Ca [‰ SRM 915a] = 0.22(± 0.05) * SST [°C] - 4.57
[1]
Thus a change of δ44/40 Ca of 0.22‰ corresponds to a relative temperature change of 1°C.
21
Tropical SST history inferred from Ca isotope thermometry
Within statistical uncertainties the temperature gradient of this calibration is identical to
that of Nägler et al. (2000) based on cultured specimen of G. sacculifer. Adjusted to the
SRM 915a standard, the latter calibration can be expressed as:
δ44/40 Ca [‰ SRM 915a] = 0.24 (±0.02) * SST [°C] – 6.53
[2]
In general, calibrations based on cultures have the advantage that foraminifera calcify
under pre-set conditions and known temperatures although the controlled laboratory
environment may not fully reproduce natural growth. The study of Nägler et al. (2000)
demonstrated that Ca isotope fractionation in G. sacculifer tend to be mainly controlled
by temperature (samples were cultivated under constant salinity (36‰), see Hemleben et
al. 1987), now confirmed by the consistency of two independent calibrations.
Figure 2 (a) The δ44/40 Ca ratios [‰ SRM 915a] of modern G. sacculifer collected from surface
waters in comparison to Nägler et al. (2000, dashed line). Every data point represents the weighted
means of either two or three independent measurements. The observed δ44/40 Ca-temperature relationship
results from the linear fit to the data (bold line). The error bars show the statistical uncertainties (2σ
s.e.). (b) Comparison of the δ44/40 Camodern -temperature calibration and the δ44/40 Ca-SST Mg/Ca crosscalibration. The shaded area mark the range of data defining the corresponding trend line (N = 53)
(dotted-dashed line).
4.2 Inter-calibration comparison
We compared the δ44/40 Camodern -SST calibration of this study with the δ44/40 Caculture-SST
calibration of Nägler et al. (2000). The two temperature gradients of 0.22 (± 0.05) and
0.24 (± 0.02) are identical and estimated relative temperature changes would not deviate
by more than 1.0°C within the investigated temperature range (19 to 28.5°C). Thus the
application of either the δ44/40 Camodern - or δ44/40 Caculture-temperature calibration as SST proxy
over time would propose the same magnitude of cooling or warming. However, the study
shows that there is an apparent offset between SST calculated from δ44/40 Camodern and
δ44/40 Caculture concerning absolute temperature reconstruction. The observed temperature
offset between both thermometers is 4.0 to 4.5°C. As a result the δ44/40 Caculture-calibration
22
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
(Nägler et al. 2000) would predict much higher absolute sea surface temperatures than the
refined δ44/40 Camodern -calibration.
The δ44/40 Ca is in particular useful in combination with Mg/Ca. Comparison of the first
δ44/40 Ca-temperature calibration to Mg/Ca records indicated that both thermometers
simultaneously reflect seawater temperature changes. However, the δ44/40 Ca-thermometer
tend to predict about 2.5°C higher absolute SSTs compared to Mg/Ca derived
temperatures (Nägler et al., 2000). Figure 2 b) illustrates the matching of the δ44/40 Camodern temperature calibration and the δ44/40 Ca-SSTMg/Ca cross calibration, i. e. linking δ44/40 Ca
values of fossil G. sacculifer to Mg/Ca based temperatures using the calibration of
Nürnberg et al. (2000). Both calibrations show the same temperature gradients of 0.22 (±
0.05)‰ and 0.23 (±0.07)‰ per 1°C respectively, and the range of data (including
uncertainties) defining the respective relationship is overlapping. These observations
imply that Mg/Ca and Ca isotopes give similar absolute temperatures if the δ44/40 Camodern SST calibration is applied.
4.3 Preservation of the primary temperature signal
Several lines of evidence indicate that the Ca isotopic composition of foraminiferal calcite
is well preserved in our samples, therefore the primary temperature signal is not
overprinted by secondary alteration. In particular, concerning past SST reconstruction the
preservation of the primary signal is of major concern. Among others e.g. dissolution and
the addition of gametogenic calcite are two processes that have to be considered in this
respect. First preliminary results of a partial dissolution experiment on shells of G.
sacculifer provide evidence that there is no relationship between the degree of dissolution
and the corresponding δ44/40 Ca values (Hönisch, 2002). In the case of Mg/Ca thermometry
Dekens et al. (2003) were able to show in a multi-species comparison of core-top
foraminiferal Mg/Ca ratios that G. sacculifer is among the species most resistant to
dissolution. This result is in agreement with previous studies of selective dissolution
susceptibility (Bé, 1977). Second, modern samples reflecting the δ44/40 Camodern -temperature
calibration are collected from surface waters that are generally supersaturated in respect to
calcium carbonate, whereas cross check calibrations based on fossil material have the
advantage that they reflect the material buried. The core depth (3122m) is well above
both the present-day (≈ 4800m) and the glacial-age lysocline depth (deeper ≈ 3800m,
Bickert & Wefer (1996)) With respect to carbonate concentration also the deepest parts of
the Guinea Basin are dominated by slightly supersaturated North Atlantic Deep Water
(NADW) (Broecker & Peng, 1982).
In order to reproduce G. sacculifer moves far below the euphotic zone. During
gametogenesis it secretes a layer of secondary calcite which may contain a subsurface
temperature component. Bijma & Hemleben (1994) guess that gametogenic calcite is
secreted in 80-100m while Duplessy et al. (1981) postulate even deeper water depth
23
Tropical SST history inferred from Ca isotope thermometry
within the main thermocline (300-800m). Obviously the δ44/40 Ca of individuals cultured
under constant temperature can not account for a subsurface temperature component. In
contrast the δ44/40 Ca of modern and fossil samples of G. sacculifer comprises the natural
variability related to their life cycles (including gametogenesis) and therefore might
contain a muted surface signal. As shown above the δ44/40 Camodern -temperature and δ44/40 CaSSTMg/Ca cross calibration of this study result within uncertainties in identical absolute SST
estimates, with modern calculated core top SST values corresponding to Levitus & Boyer
(1994). The observed temperature signal in foraminiferal calcite of G. sacculifer seems to
be primary and a muted surface signal could be precluded. Concerning Mg/Cathermometry Dekens et al. (2003) quoted that G. ruber is the most accurate recorder of
sea surface temperature, while G. sacculifer records temperatures below the surface at 2030m. As long as the calcification depth of G. sacculifer is controversial, the evidence
whether δ44/40 Ca in G. sacculifer represents surface or subsurface temperatures can only be
rendered by depth habitat tracking of both species. Comparing three independent δ44/40 Catemperature calibrations we can conclude that dissolution and the potential addition of
gamatogenetic calcite have no significant impact on the reliability of Ca isotopes as SST
proxy. Summarising our results we propose the use of the refined δ44/40 Camodern -temperature
relationship for future palaeoceanographic applications.
5. Application of Ca Isotopes as SST Proxy
5.1 Down-core variations on fossil G. sacculifer of core GeoB1112
We measured δ44/40 Ca ratios on fossil calcite shells of G. sacculifer from sediment core
GeoB1112 covering the last about 140.000 years (marine isotope stages (MIS) 1 to 6)
thus comprising two glacial-interglacial transitions (Figure 3 and Table 2, see
supplementary data). Throughout the investigated time interval δ44/40 Ca values range
between -0.17 and +1.18 (‰ SRM 915a). Given the observations that the Ca isotopic
composition of seawater is homogenous throughout modern oceans (Zhu & MacDougall
1998, Schmitt et al. 2001, Hippler et al. 2003) and no secular variations occur over the
last 19Ma (De La Rocha & DePaolo 2000, Schmitt et al. 2003, Heuser et al. subm.) these
variations do not reflect changes in seawater Ca isotopic composition. The observed
δ44/40 Ca down-core variations coincide with the Mg/Ca record for G. sacculifer (Nürnberg
et al. 2000) and reflect glacial-interglacial cycles. Highest δ44/40 Ca values occur during
interglacial times while samples related to glacial times tend to have relatively low δ44/40 Ca
values. We found pronounced δ44/40 Ca gradients at stage boundaries Holocene to Last
Glacial Maximum (LGM) and MIS 5e to MIS 6, respectively. δ44/40 Ca values in the
Holocene (MIS 1) range from 0.60 to 1.15‰. At stage boundary MIS 1/2 (Termination
I) δ44/40 Ca drops to values of around 0.00±0.20‰. LGM δ44/40 Ca values are low with smallscale variations between -0.17 and 0.17‰. During MIS 3 and MIS 4 the δ44/40 Ca ratio
undulates between 0.00 and 1.02‰ describing an overall increasing trend towards the last
interglacial. Measured δ44/40 Ca values in MIS 5 vary over a wide range with peak values of
24
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
1.18‰ at MIS 5e. Stage boundary MIS 5/6 (Termination II) exhibits a pronounced
decrease to values of around 0.11‰.
Figure 3: Down-core record of δ44/40 Ca values [‰ SRM 915a] of fossil G. sacculifer and reconstructed
sea surface temperatures (SST). SSTδ44/40Ca are calculated applying the refined δ44/40 Camodern -temperature
calibration (δ44/40 Ca [‰] = 0.22 (±0.05) per 1°C) introduced in this study. Shaded bands represent the
Holocene and the last interglacial period. Numbers on top of the diagram reflect marine oxygen isotope
stages (MIS) after Martinson (1987).
Using the δ44/40 Camodern -temperature calibration (see equation [1]) past temperatures could
be calculated using the following equation:
SST [°C] = 4.76 * δ44/40 Ca [‰] + 28.76 = SSTδ44/40Ca [°C]
[3]
The corresponding δ44/40 Camodern -derived temperature record of sediment core GeoB1112 is
displayed in Figure 3. During the Holocene calculated SST δ44/40Ca values range between
24.0 and 26.5°C. Minimum SST δ44/40Ca between 21.0 and 23.0°C persists during the Last
Glacial Maximum and MIS 6, while peak SSTδ44/40Ca of 27°C is reached during the Last
Interglacial MIS 5e. The overall glacial-interglacial amplitude for the investigated time
interval is 3.5 ± 1.0°C. whereas the temperature change at the Termination II is more
distinct than at the Holocene-LGM transition.
25
Tropical SST history inferred from Ca isotope thermometry
26
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
5.2. SST δ44/40Ca estimates and comparison to recent conditions
Applying the δ44/40 Camodern -temperature calibration, estimates of absolute SST of core top
sample in core GeoB1112 correspond to the upper limit of what is expected from Levitus
& Boyer (1994). The calculated SST δ44/40Ca signal is in agreement with the temperature
situation averaged for the entire year within the uppermost 50m of the water column, the
assumed habitat of pregametogenic G. sacculifer. According to Levitus & Boyer (1994)
the core location of GeoB1112 is characterised by rather constant temperatures
throughout the year. The adjacent equatorial upwelling lasting from May to August has
no significant impact on seasonal temperature variations in the core-collected region.
Considering further the low temporal resolution of the core, our data does not allow a
more detailed seasonal assignment. The previous work of Nürnberg et al. (2000)
additionally included an upwelling-influenced sediment core (GeoB1105, 1°39.9’S,
12°25.7’W, 3225 m). From the comparison of these two Mg/Ca derived SST-records the
authors concluded that for the area of investigation the observed SST Mg/Ca signal of G.
sacculifer reflects the austral low-latitude fall/winter upwelling situation within the
uppermost 50m of the water column.
Table 3: Core top SST estimates for various proxies in comparison to Levitus & Boyer (1994)
Reference
Method
SST [°C]
Remarks
23.3-25.5 1)
annual
SST δ44/40Ca
25.6
core top
SST δ18O 2)
22.8
(")
25.4-27.3
(")
26.1
(")
Levitus & Boyer (1994)
This study
Wefer et al. (1996)
SST-TFcold-warm
Nürnberg et al. (2000)
SST Mg/Ca
1)
The Levitus & Boyer (1994) water temperatures represent an average value of the upper 50m of the water column for the
entire year
2) 18
δ O record of Meinecke (1992) and calculated SST based on the δ18 O-temperature relationship published by Mulitza et al.
(2003)
5.3 Past sea surface temperature reconstruction in the tropical Atlantic
We compared the δ44/40 Ca record with the available δ18O, TF and Mg/Ca records of core
GeoB1112 (Figure 4a-c). The results broadly coincide and confirm postulated trends of
tropical palaeo-reconstruction.
The SSTδ18O record was calculated (see Appendix) from the available δ18O-record of
Meinecke (1992) by applying the δ18O-temperature relationship of Mulitza et al. (2003).
The SSTδ18O record reflects the main glacial-interglacial cycles. However, SSTδ18O is
27
Tropical SST history inferred from Ca isotope thermometry
considerably lower than the corresponding SSTδ44/40Ca . The observed offset is 2.0±1.0°C
(Figure 4 a). According to the detailed discussion of Mulitza et al. (2003) SST δ18O based
on calibrations of G. sacculifer should record a muted surface signal. The depth ranking
of four examined species based on δ18O analyses by Spero et al. 2003 suggests that G.
sacculifer inhabits a slightly deeper/cooler environment. These findings are in accordance
with relationships derived from pH-controlled culturing experiments of Bemis et al.
(1998). The uncertainty on calculated absolute palaeotemperatures resulting from the use
of either the δ18O-temperature relationship of Mulitza et al. (2003), Bemis et al. (1998) or
Spero et al. (2003) would amount to less than 1.5°C. The latter authors note that the
palaeothermometers proposed by Shackleton (1974) or Erez & Luz (1983) give reliable
estimates for relative temperature change but seem to overestimate absolute temperatures.
Apart from the observed offset in absolute temperatures the general pattern of calculated
SSTδ18O and SST δ44/40Ca is in good agreement. Even the smaller-scale temporal amplitudes
of both records reveal similar amounts of tropical temperature change. Apparently the
greater deviation during MIS 3 and MIS 4 might be a direct consequence of the lower
sample resolution of the SSTδ44/40Ca record.
The SSTδ44/40Ca record postulates similar overall SST than the corresponding palaeosummer
SST-TF record of Wefer et al. (1996) derived from foraminiferal transfer functions
(Figure 4 b). However in detail the record shows positive and negative excursions of 12°C around the TF warm estimates. During the late Holocene SSTδ44/40Ca matches fairly with
SST-TF cold , while peak temperature and early Holocene estimates are in accordance to
SST-TF warm . During the last glacial period (MIS 2 to 4) TFwarm shows only small
temperature variation. SSTδ44/40Ca and TF cold are offset both reflecting the main cooling
trend and stronger temporal amplitudes. The Last Interglacial (MIS 5) is characterised by
higher SST exceeding mean Holocene temperature. Observed temperature drops in MIS 5
recorded by TF warm and more pronounced by TF cold are not shown to the same extent by
SSTδ44/40Ca . Further, these negative excursions are not exactly in phase. However, both
TFwarm and SSTδ44/40Ca likewise illustrate the temperature decrease of ≈4°C at the boundary
MIS 5e to MIS 6. The glacial-interglacial SSTδ44/40Ca amplitude over the investigated time
interval is 3.0-4.0°C, whereas the SST-TF amplitudes are larger and often exceed 5°C. The
larger amplitudes might result from the fact that SST-TF integrates temperatures from the
surface down to the thermocline when referring to the entire planktonic foraminiferal
assemblage (Meinecke, 1992). Further, vertical water column hydrography and nutrient
distribution rather than SST alone might also influence the foraminiferal abundance
(Ravelo & Fairbanks, 1992; Sikes & Keigwin, 1994).
The temporal pattern of the palaeo-SST records calculated from Mg/Ca (Nürnberg et al.
2000) and δ44/40 Ca is in overall agreement, although the SSTMg/Ca appear generally higher
in comparison to SSTδ44/40Ca . This apparent offset (1-2°C) becomes negligible considering
the uncertainties of the respective temperature record (SST δ44/40Ca and SSTMg/Ca ).
Furthermore, several species-specific Mg/Ca-temperature relationships were published
28
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
recently e.g. by Dekens et al. 2003 and Anand et al. 2003. Application of the latter
calibrations would also result in similar absolute temperatures within errors. In particular,
the consistency of the temporal phasing of both records (δ44/40 Ca and Mg/Ca) is evident in
the point-to-point comparison of both records at the transitions Holocene-LGM and MIS
5 to MIS 6. This feature is apparently not reached to the same extent during MIS 3 and
MIS 4 because the sample resolution of both records is not equivalent. Throughout both
SST records amplitudes of short temporal variations are more distinct for SSTδ44/40Ca
especially during interglacials. Common to both down-core records is the glacialinterglacial amplitude of 3-4°C for the last climatic changes with less pronounced
Holocene-LGM (2.5-3.5°C) temperatures compared to the interglacial stage 5 to glacial
stage 6.
Table 4: Magnitude of tropical SST cooling at the transition Holocene-LGM
Reference
SST [°C]
Method
Global Oceans:
CLIMAP (1981)
1.0-2.0
Faunal assemblages
This study
2.5-3.5
Van Campo (1990)
3.0-4.0
SSTδ44/40Ca
Global Circulation Model
Atlantic Ocean:
Guilderson et al. (1994)
≈5.0
SSTSr/Ca (corals)
Hastings et al. (1998)
2.6±1.3
SSTMg/Ca
Wolff et al. (1998)
2.0-3.0
SSTδ180
SSTUK'37
Rühlemann et al. (1999)
3.5
3.0-4.0
≈2.5
SSTδ180
SSTMg/Ca
2.6±0.5
SSTMg/Ca
Lea et al. (2000)
≈3.0
SSTMg/Ca
Kienast et al. (2001)
≈3.0
SSTUK'37
Visser et al. (2003)
3.5-4.0
SSTMg/Ca
Rosenthal et al. (2003)
2.3±0.5
SSTMg/Ca
Nürnberg et al. (2000)
Lea et al. (2003)
Pacific Ocean:
The CLIMAP study (1981) initially suggested a drop in SST in the tropics of less than
2°C at the Holocene-Last Glacial Maximum transition. In contrast, there is a growing
consensus about glacial temperatures in the tropics from both the Atlantic and the Pacific
Ocean. Temperature reconstruction based on various proxies using different marine
archives (δ18O, Mg/Ca, Sr/Ca, UK’37) and modelling studies point to a cooling of surface
water masses of 2.5 to 5.0°C (Table 4).
29
Tropical SST history inferred from Ca isotope thermometry
5.4 Phasing of SST proxy signals
Investigating δ44/40 Ca and δ18O on the same samples allows the determination of the relative
phasing between tropical SST variability and high-latitude deglaciation. Assuming that the
Ca isotope signal of G. sacculifer monitors past SST variability and that a considerable
amount of the δ18O signal reflects global ice volume change, the timing of surface
temperature warming and global ice volume changes can be deduced. We observed that
for the last terminations I and II the changes of both isotopic signals are not synchronous
(Figure 5). For both terminations low or decreasing δ18O values persist whereas the δ44/40 Ca
record indicates that SST already reached interglacial values. The estimated time lag of ice
volume response amounts to approximately 3.000 years. Tropical SST variations
preceding global ice volume changes have already been suggested for the equatorial
Atlantic (Rühlemann et al. 1999, Nürnberg et al. 2000), the Indian (Bard et al. 1997,
Cayre & Bard, 1999) and the Pacific Ocean (Lea et al. 2000, Visser et al. 2003) on the
basis of δ18O, Mg/Ca, transfer functions and alkenones data. Thus tropical surface water
warming at the transition from the LGM to Holocene appears to have been a roughly
synchronous global-scale feature. It is important to note here, that our findings are based
on an additional and independent SST proxy.
Figure 5: G. sacculifer oxygen isotope (diamonds) and SSTδ44/40Ca (dots) records for core GeoB1112
plotted versus time for (a) Holocene and Last Glacial Maximum and (b) MIS 5e and 6. For both
terminations the warming of tropical surface waters occurs approximately 3000 years before the melting
of the Northern Hemisphere ice sheets is reflected by the δ18O record.
Our findings support the hypothesis that the tropical Atlantic plays a dynamic part in
triggering global climate changes. The atmosphere is very sensitive to changes in tropical
SST, even small changes (0.5 to 1.0°C) can have large effects on the moisture budget,
storminess and the overall atmospheric circulation (Rind 1990, Walser & Grahm 1993,
Schneider et al. 1996), and particularly on fluxes of heat and greenhouse gases (water
vapour, CO2). Other players are sunlight intensity and the deep circulation of the North
Atlantic. Atmospheric greenhouse forcing might explain the roughly synchronous
30
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
tropical temperature changes, whereas the warming of the tropical Atlantic might result
from slowing of the thermohaline overturn in the North Atlantic along with an expected
decrease of northward heat transport (Crowley 1992, Manabe & Stouffer 1997). The low
time-resolution of our core does not allow the precise assignment whether the timing of
SST changes is synchronous or non-synchronous to high-latitude Atlantic temperatures
over the last glacial terminations. A geographical pattern has recently emerged in which
the western part of the tropical Atlantic (Rühlemann et al. 1999) observed temperature
changes that are out of phase to high-latitude North Atlantic temperatures, while Zhao et
al. (1995) obtain synchronous changes in the eastern low-latitude North Atlantic. These
studies point to the sparse palaeoceanographic evidence that is currently available about
the linkage between low- and high-latitude regions. Providing an independent and robust
tool (δ44/40 Ca) to estimate the magnitude and phasing of tropical SST changes, our study
supports the growing evidence that the tropics are an important controlling factor for
global climate change.
6. Summary and conclusion
Systematic Ca isotope analyses on tests of modern G. sacculifer covering a temperature
range between 19 and 29°C provide an additional and robust tool for the reconstruction
of past sea surface temperatures. A change in δ44/40 Ca [‰] of 0.22(±0.05) corresponds to a
temperature change of 1°C. Secondary processes like dissolution or gametogenesis have
no significant impact on the temperature-dependent Ca isotope fractionation as shown by
the comparison of three independent calibrations. Estimates of relative temperature
changes are compatible with previous results of Nägler et al. (2000). Furthermore the
accuracy of absolute temperatures is supported by the conformity of calculated
SSTδ44/40Ca from core top sample and modern annual SST according to Levitus &
Boyer (1994). This improvement over the initial calibration (Nägler et al. 2000) points to
a δ44/40 Ca bias in cultured G. sacculifer of the latter.
We propose the use of the refined δ44/40 Camodern -temperature calibration (δ44/40 Ca [‰] =
0.22(±0.05) * SST [°C] – 4.88) for future applications in palaeoceanography.
The strong temperature dependence of Ca isotope fractionation observed for biogenic
calcite tests of G. sacculifer could provide new insights in biomineralization processes. As
shown in this study the complex biology does not unsettle the confidence in the potential
of Ca isotopes as robust SST proxy when using selective planktonic foraminifera species.
Our results indicate that the equatorial Atlantic was about 3 to 4°C colder than present-day
conditions during the last two glacial periods. At the last terminations I and II the timing
of SST change precedes the descent of ice volume by several thousand years
(approximately 3ka) at this site, suggesting that tropical warming played a major role in
forcing global climate change.
31
Tropical SST history inferred from Ca isotope thermometry
Acknowledgements
We thank R. Schneider and S. Mulitza who generously provided foraminiferal tests from
sediment core GeoB1112 and from different cruises (M38-1, M46-2, SO164). We are
grateful to Anne Müller, Kate Darling and Jan Kramers for their helpful comments on
improving this manuscript. D. Hippler and Th. F. Nägler benefit from a research project
founded by the Swiss National Science Foundation (grants 21-61644.00 and 2160987.00). The work of A. Eisenhauer and N. Gussone was supported by DFG grant
Ei272/12-1.
Appendix 3: SST estimates derived from oxygen isotopes
Within the scope of our multi-proxy comparison we determined past SSTs from the
available δ18O record of G. sacculifer (Meinecke 1992) by applying the species-specific
temperature-δ18O relationship established by Mulitza et al. (2003):
T [°C] = -4.35 (δ18Ocal - δ18Osw) + 14.91
[4]
Here δ18Ocal stands for the oxygen isotopic composition of the calcite (in ‰ Peedee
Belemnite (PDB)) and δ18Osw is the oxygen isotopic composition of seawater (in ‰
Standard Mean Ocean Water (SMOW)). The conversion of δ18Osw from the SMOW to the
PDB scale is performed according to Hut (1987) proposing a correction factor of 0.27‰. Estimating SST from the δ18O of foraminiferal calcite, local salinity effects on the
δ18Osw have to be considered. To do this we chose estimates of δ18Osw from the Global
Seawater Oxygen-18 Database (G. A. Schmidt, G. R. Bigg and E. J. Rohling available at
(http://www.giss.nasa.gov/data/o18data/) (Schmidt, 1999). According to Levitus & Boyer
(1994) salinities of approximately 35.5‰ prevail at 0-100m water depth in low-latitude
Atlantic. For the core location of GeoB1112 we obtained representative δ18Osw values for
the tropical east Atlantic of 0.79±0.10‰. This value is in good agreement with what
would be calculated using δ18O versus salinity relationships proposed in the literature
(Duplessy et al. 1991, Wang et al. 1995). Investigating glacial-interglacial cycles secular
variations of seawater δ18O driven by ice volume changes have to be taken into account.
We considered the so-called “ice effect” by subtracting the mean δ18Osw record of
Vogelsang, 1990 (modified after Labeyrie et al. 1987). The mean δ18Osw increase of 1.1‰
during the LGM is in accordance with direct measurements of oxygen isotopes of relict
glacial pore water (Schrag et al. 1996), high-resolution δ18O records of benthic
foraminifera (Labeyrie et al. 1987) and combined studies of Shackleton, 2000.
32
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
Supplementary data
Table 1: The δ44/40 Ca of modern specimen of G. sacculifer
Locality
Sample
T [°C]
11
28.2
9
28.1
Central tropical Atlantic Ocean (M38-1)
13
20
7
South Atlantic Ocean, offshore Brazil (M46-2)
Western tropical Atlantic Ocean, Caribbean
Sea (SO164 Leg 1-3)
44/40
Ca
2
SE
44/42
Ca
1.51
0.82
1.09
1.15
0.11
0.16
0.11
0.15
0.755
0.410
0.545
0.575
27.2
27.0
26.1
1.16
1.29
1.03
0.74
0.16
0.20
0.15
0.19
0.580
0.645
0.515
0.370
35
22
22.1
21.1
32
19.4
-0.02
-0.41
-0.30
-0.10
-0.48
-0.74
0.15
0.16
0.12
0.21
0.08
0.20
-0.010
-0.205
-0.150
-0.050
-0.240
-0.370
δ44/40 Ca and calculated δ44/42 Ca values are given in [‰] relative to NIST SRM 915a.
δ44/42 Ca values are calculated according to Hippler et al. (2003)
Uncertainties reflect the 2σ standard error of single analysis
33
Tropical SST history inferred from Ca isotope thermometry
Table 2: Down-core variations in δ44/40 Ca of fossil G. sacculifer from core GeoB1112
Depth [cm]
Age [kyr] 1)
3
8
13
18
23
28
33
38
43
48
53
58
63
68
73
78
83
1.71
3.52
5.34
7.14
8.95
10.77
12.57
14.38
16.19
18.00
19.96
21.92
23.88
25.83
27.79
29.75
31.71
0.80
0.75
0.69
0.63
1.15
0.77
0.62
0.60
0.55
0.05
0.17
0.00
0.06
-0.17
0.19
0.38
0.17
0.16
0.13
0.11
0.20
0.16
0.14
0.19
0.10
0.27
0.19
0.10
0.11
0.07
0.34
0.23
0.13
0.10
0.40
0.38
0.35
0.32
0.58
0.39
0.31
0.30
0.28
0.03
0.09
0.00
0.03
-0.09
0.10
0.19
0.09
Number of
Analyses 5)
3*
1
1
1
1*
1
1*
1
1*
2
1
3
1
2*
1
2*
2
93
98
103
108
113
118
35.63
37.58
39.54
41.50
43.46
45.42
0.33
0.26
0.09
0.56
0.49
0.16
0.16
0.12
0.18
0.12
0.11
0.11
0.17
0.13
0.05
0.28
0.25
0.08
1
2
1
1
1
2
128
138
148
158
168
178
188
198
49.33
53.25
57.17
61.08
65.00
70.50
76.00
81.50
1.19
0.68
1.00
0.18
0.05
0.77
0.52
0.41
0.18
0.14
0.16
0.12
0.18
0.16
0.20
0.19
0.60
0.34
0.50
0.09
0.03
0.39
0.26
0.21
1
1
1
1
1
2
1*
2
203
213
223
84.25
88.91
92.73
0.65
0.86
1.02
0.20
0.36
0.14
0.33
0.43
0.51
1
1
1
228
233
238
243
248
253
258
94.64
96.55
98.45
100.36
102.27
104.18
106.09
0.74
0.26
0.48
0.88
1.26
0.54
-0.02
0.07
0.21
0.34
0.49
0.13
0.14
0.22
0.37
0.13
0.24
0.44
0.63
0.27
-0.01
1
1*
1
1*
1
1
1
268
278
109.69
113.06
0.77
0.51
0.12
0.14
0.39
0.26
1
1
283
293
114.75
118.13
0.69
0.78
0.13
0.45
0.35
0.39
1*
298
303
308
313
119.81
121.50
123.19
124.88
0.67
0.73
1.18
0.79
0.14
0.30
0.30
0.09
0.34
0.37
0.59
0.40
3*
1
1
1
323
328
333
128.25
129.94
131.63
1.07
0.89
0.59
0.23
0.17
0.30
0.54
0.45
0.30
1
1
1
348
136.45
0.11
0.12
0.06
2*
δ44/40 Ca [‰]
2)
Uncertainty 3)
δ44/42 Ca [‰]
4)
1
1)
Age model taken from Meinecke (1992)
2)
δ44/40 Ca values in [‰] relative to NIST SRM 915a. For reproduced analyses a weighted average is given.
3)
Uncertainties reflect the 2σ standard error for single and the weighted mean error for reproduced analyses, respectively.
4)
δ44/42 Ca values in [‰] relative to NIST SRM 915a are calculated according to Hippler et al. (2003)
5)
Number of analyses: * Data after Nägler et al. (2000)
34
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
References:
Anand P., Elderfield H. and Conte M. H. (2003)
Calibration of Mg/Ca thermometry in planktonic foraminifera from a sediment trap time
series. Paleoceanography 18(2).
Bard E., Rostek F. and Sonzogni C. (1997)
Interhemisperic synchrony of the last
palaeothermometry. Nature 385, 707-710.
deglaciation
inferred
from
alkenone
Bemis B. E., Spero H. J., Bijma J. and Lea D. W. (1998)
Re-evaluation of the oxygen isotopic composition of planktonic foraminifera:
Experimental results and revised paleotemperature equations. Paleoceanography 13(2),
150-160.
Bé A. W. H. (1977)
An Ecological, Zoogeographic
Foraminifera. Academic Press.
and
Taxonomic
Review of
Recent
Planktonic
Bickert T. and Wefer G. (1996)
Late Quaternary deep water circulation in the South Atlantic: Reconstruction from
carbonate dissolution and benthic stable isotopes. In The South Atlantic: Present and Past
Circulation (ed. G. Wefer), Springer-Verlag, New York, 599-620.
Bijma J. and Hemleben C. (1994)
Population dynamics of the planktonic foraminifer Globigerinoides sacculifer (Brady)
from the central Red Sea. Deep Sea Research Part I, 41, 485-510.
Birck J. L. (1986)
Precision K-Rb-Sr isotopic analysis: application to Rb-Sr chronology. Chemical Geology
56, 73-83.
Broecker W. S. and Peng T.-H. (1982)
Tracers in the Sea. Eldigo Press.
Cayre O. and Bard E. (1999)
Planktonic foraminiferal and alkenone records of the last deglaciation from the Eastern
Arabian Sea. Quaternary Research 52, 337-342.
CLIMAP Project Members (1981)
Climate: Long-range investigation, mapping, and prediction (CLIMAP), seasonal
reconstructions of Earth's surface at the Last Glacial Maximum. In GSA Map and Chart
Series. Geological Society of America.
35
Tropical SST history inferred from Ca isotope thermometry
Coplen T. B., Böhlke J. K., Bièvre P. D., Ding T., Holden N. E., Hopple J. A., Krouse R.,
Lamberty A., Peiser H. S., Revesz K., Rieder S. E., Rosman K. J. R., Roth E., Taylor P. D.
P., Vocke R. D. J. and Xiao Y. K. (2002)
Isotope-abundance variations of selected elements (IUPAC Technical Report). Pure and
Applied Chemistry 74, 1987-2017.
Crowley, T. J. (1992)
North Atlantic Deep Water cools the southern hemisphere. Paleoceanography, 7, 489-497.
Dekens P. S., Lea D. W., Pak D. K. and Spero H. J. (2002)
Core top calibration of Mg/Ca in tropical foraminifera: Refining paleotemperature
estimation. Geochemistry, Geophysics, Geosystems 3(4), 10.1029/2001GC000200.
De La Rocha C. L. and DePaolo D. J. (2000)
Isotopic evidence for variations in the marine calcium cycle over the Cenozoic. Science
289, 1176-1178.
Duplessy J.-C., Labeyrie L., Juillet-Leclerc A., Maitre F., Duprat J. and Sarntheim M.
(1991)
Surface salinity reconstruction of the North Atlantic Ocean during last glacial maximum.
Oceanologica Acta 14(4), 311-324.
Eisenhauer A., Nägler T. F., Stille P., Kramers J., Gussone N., Bock B., Fietzke J., Hippler
D. and Schmitt A.-D. (2004)
Proposal for international agreement on Ca notation resulting from discussions at
workshops on stable isotope measurements held in Davos (Goldschmidt 2002) and Nice
(EGS-AGU-EUG 2003). Geostandards Newsletter: The Journal of Geostandards and
Geoanalysis 28.
Erez J. and Luz B. (1983)
Experimental paleotemperature equation for planktonic foraminifera. Geochimica et
Cosmochimica Acta 47, 1025-1031.
Guilderson T. P., Fairbanks R. G. and Rubenstone J. L. (1994)
Tropical temperature variations since 20,000 years ago: Modulating interhemispheric
climate change. Science 263, 663-665.
Gussone N., Eisenhauer A., Heuser A., Dietzel M., Bock B., Böhm F., Spero H. J., Lea D.
W., Bijma J. and Nägler T. F. (2003)
Model for kinetic effects on calcium isotope fractionation (δ44Ca) in inorganic aragonite
and cultured planktonic foraminifera. Geochimica et Cosmochimica Acta 67, 1375-1382.
36
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
Hastings D. W., Russell, A. D. and Emerson, S. R. (1998)
Foraminiferal magnesium in G. sacculifer as a paleotemperature proxy in the equatorial
Atlantic and Carribean surface oceans. Paleoceanography, 13(2), 161-169.
Hemleben C., Spindler M., Breitinger I. and Ott R. (1987)
Morphological and physiological response of Globigerinoides sacculifer (Brady) under
varying laboratory conditions. Marine Micropaleontology 12, 305-324.
Hippler D., Schmitt A.-D., Gussone N., Heuser A., Stille P., Eisenhauer A. and Nägler T.
F. (2003)
Calcium isotopic composition of various reference materials and seawater. Geostandards
Newsletter: The Journal of Geostandards and Geoanalysis 27, 13-19.
Hippler D., Villa I. M., Nägler T. F. and Kramers J. D. (2004)
A ghost haunts mass spectrometry: real isotope fractionation or analytical paradox?
International Goldschmidt Conference 2004, Copenhagen, 2.7.P12.
Hönisch B. (2002)
Stable isotope and trace element composition of foraminiferal calcite -from incorporation
to dissolution. PhD Thesis, Universität Bremen, Bremen, 118 pp.
Hut G. (1987)
Consultants group meeting on stable isotope reference samples for geochemical and
hydrological investigations, Report to the Director General, International Atomic Energy
Agency, pp. 42.
Kienast M., Steinke S., Stattegger K. and Calvert S. E. (2001)
Synchronous tropical South China Sea SST change and Greenland warming during
deglaciation. Science, 291, 2132-2134.
Labeyrie L. D., Duplessy J.-D. and Blanc P. L. (1987)
Variations in mode of formation and temperature of oceanic deep waters over the past
125,000 years. Nature 327, 477-482.
Lea D. W., Pak D. K. and Spero H. J. (2000)
Climate impact of Late Quaternary equatorial Pacific sea surface temperature variations.
Science 289, 1719-1724.
Lea D. W., Pak D. K., Peterson L. C. and Hughen K. A. (2003)
Synchroneity of tropical and high-latitude Atlantic temperatures over the last glacial
termination. Science 301, 1361-1364.
Levitus S. and Boyer T. P. (1994)
World Ocean Atlas. US Department of Commerce.
37
Tropical SST history inferred from Ca isotope thermometry
Manabe, S. and Stouffer, R. J. (1997)
Coupled ocean-atmosphere model response to freshwater input: comparison to Younger
dryas event. Paleoceanography, 12, 321-3336.
Marriott C. S., Henderson G. M., Belshaw N. S. and Tudhope A. W. (2004)
Temperature dependence of δ7Li, δ44Ca and Li/Ca during growth of calcium carbonate.
Earth and Planetary Science Letters 222, 615-624.
Martinson, D. G., Pisias, N. G., Hays, J. D., Imbrie, J., Moore, T. C. J. and Shackleton
(1987)
Age dating and the orbital theory of the ice ages: Development of a high-resolution 0 to
300.000-year chronostratigraphy. Quaternary Research, 27, 1-29.
Meinecke G. (1992)
Spätquartäre Oberflächenwassertemperaturen im östlichen äquatorialen Atlantik. Berichte
Fachbereich Geowissenschaften 29, pp. 181.
Mulitza S., Boltovskoy D., Donner B., Meggers H., Paul A. and Wefer G. (2003)
Temperature:δ18O relationships of planktonic foraminifera collected from surface waters.
Palaeogeography, Palaeoclimatology, Palaeoecology 202, 143-152.
Nägler T. F., Eisenhauer A., Müller A., Hemleben C. and Kramers J. (2000)
The δ44Ca-temperature calibration on fossil and cultured Globigerinoides sacculifer: New
tool for reconstruction of past sea surface temperatures. Geochemistry, Geophysics,
Geosystems 1, 2000GC000091.
Nürnberg D., Müller A. and Schneider R. R. (2000)
Paleo-sea surface temperature calculations in the equatorial east Atlantic from Mg/Ca in
planktonic foraminifera: A comparison to sea surface estimates from Uk'37, oxygen
isotopes, and foraminiferal transfer functions. Paleoceanography 15, 124-134.
Ravelo A. C. & Fairbanks R. G. (1992)
Oxygen isotopic composition of multiple species of planktonic Foraminifera; recorders of
the modern photic zone temperature gradient. Paleoceanography 7(6), 815-831.
Rind D. (1990)
Palaeoclimate puzzles from the tropics. Nature 346, 317-318.
Rosenthal Y., Oppo D. W. and Linsley B. K. (2003)
The amplitude and phasing of climate change during the last deglaciation in the Sulu Sea,
western equatorial Pacific. Geophysical research letters, 30(8), 1428, doi:
10.1029/2002GL016612.
38
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
Rühlemann C., Mulitza S., Müller P. J., Wefer G. and Zahn R. (1999)
Warming of the tropical Atlantic Ocean and slowdown of thermohaline circulation during
the last deglaciation. Nature 402, 511-514.
Schmidt G. A., Bigg G. R. and Rohling E. J. (1999)
Global seawater oxygen-18 database. http://www.giss.nasa.gov/data/o18data/.
Schmitt A.-D., Bracke G., Stille P. and Kiefel B. (2001)
The calcium isotope composition of modern seawater determined by thermal ionisation
mass spectrometry. Geostandards Newsletter: The Journal of Geostandards and
Geoanalysis 25, 267-275.
Schmitt A.-D., Stille P. and Vennemann T. (2003)
Variations of the 44Ca/40Ca ratio in seawater during the past 24 million years: Evidence
from δ44Ca and δ18O of Miocene phosphates. Geochimica et Cosmochimica Acta 67(14),
2607-2614.
Schneider N., Barnett, T., Latif, M. and Stockdale, T. (1996)
Warm pool physics in a coupled GCM. Journal of Climate, 9, 219-239.
Schrag D. P., Hampt G. and Murray D. W. (1996)
Pore fluid constraints on the temperature and oxygen isotopic composition of the glacial
ocean. Science 272, 1930-1932.
Shackleton N. J. (1974)
Attainment of isotopic equilibrium between ocean water and the benthonic foraminifera
genus Uvigerina: Isotopic changes in the ocean during the last glacial. Colloquium
International C.N.R.S. 219, 203-209.
Shackleton N. J. (2000)
The 100,000-year ice-age cycle identified and found to lag temperature, carbon dioxide,
and orbital eccentricity. Science 289, 1897-1902.
Siebert C., Nägler T. F. and Kramers J. D. (2001)
Determination of molybdenum isotope fractionation by double-spike multicollector
inductively coupled plasma mass spectrometry. Geochemistry, Geophysics, Geosystems 2,
2000GC000124.
Sikes E. L. and Keigwin L. D. (1994)
Equatorial Atlantic sea surface temperature for the last 30 kyr: a comparison of Uk'37,
δ18O, and foraminiferal assemblage temperature estimates. Paleoceanography 9(1), 31-45.
39
Tropical SST history inferred from Ca isotope thermometry
Skulan J., DePaolo D. J. and Owens T. L. (1997)
Biological control of calcium isotopic abundances in the global calcium cycle.
Geochimica et Cosmochimica Acta 61, 2505-2510.
Skulan J. and DePaolo D. J. (1999)
Calcium isotope fractionation between soft and mineralized tissues as a monitor of
calcium use in vertebrates. Proceedings of the National Academy of Science 96, 1370913713.
Spero H. J., Mielke K. M., Kalve E. M., Lea D. W. and Pak D. K. (2003)
Multispecies approach to reconstructing eastern equatorial Pacific thermocline
hydrography during the past 360 kyr. Paleoceanography 18(1), 1022,
doi:10.1029/2002PA000814.
Van Campo, E., Duplessy, J. C., Prell, W. L., Barratt, N. and Sabatier, R. (1990)
Comparison of terrestrial and marine temperature estimates for the past 135 ky off
southeast Africa: a test for GCM simulations of paleoclimate. Nature 348, 209-212.
Visser K., Thunell R. and Stott L. (2003)
Magnitude and timing of temperature change in the Indo-Pacific warm pool during
deglaciation. Nature 421, 152-155.
Vogelsang E. (1990)
Paläo-Ozeanographie des Europäischen Nordmeeres anhand stabiler Kohlenstoff- und
Sauerstoffisotope. Reports SFB 313, 23, University of Kiel, pp. 136.
Walser, D. E. and Grahm, N. E. (1993)
Journal of Geophysical Research, 96, 12881-12893.
Wang L., Sarntheim M., Duplessy J.-C., Erlenkeuser H., Jung S. and Pflaumann U. (1995)
Paleo sea surface salinities in the low-latitude Atlantic: The δ18O record of Globigerinoides
ruber (white). Paleoceanography 10(4), 749-761.
Wefer G. (1996)
Late quaternary surface circulation of the south Atlantic: The stable isotope record and
implications for heat transport and productivity. In The South Atlantic: Present and Past
Circulation (ed. G. Wefer), Springer-Verlag, New York, 461-502.
Wolff T., S. M., Arz H., Pätzold J. and Wefer G. (1998)
Oxygen isotopes versus CLIMAP (18 ka) temperatures: A comparison from tropical
Atlantic. Geology 226(8), 675-678.
40
Chapter 3: PhD-thesis Dorothee Hippler, Bern (2004)
Zhao M., Beveridge N. A. S., Shackleton N. J., Sarntheim M. and Eglinton G. (1995)
Molecular stratigraphy of cores off northwest Africa: Sea surface temperature history over
the last 80 ka. Paleoceanography 10(3), 661-675.
Zhu P. and MacDougall J. D. (1998)
Calcium isotopes in the marine environment and the oceanic calcium cycle. Geochimica
et Cosmochimica Acta 62, 1691-1698.
41
Chapter 4: PhD-thesis Dorothee Hippler, Bern (2004)
Genetic diversity and implications for Ca isotope thermometry
in polar oceans
Dorothee Hippler1, Kate F. Darling2, Anton Eisenhauer3 and Thomas F. Nägler1
1
2
Institute of Geological Sciences, University of Bern, Bern, Switzerland
Grant Institute of Earth Science and Institute of Evolutionary Biology, University of Edinburgh,
Edinburgh, United Kingdom
3
Leibniz-Institut für Meereswissenschaften (IFM-GEOMAR), Kiel, Germany
Manuscript in preparation for submission to Science
Abstract – Assessing the potential geochemical consequences of cryptic genetic diversity of
planktonic foraminifera has become a crucial element in the validation of palaeoceanographic
proxies. We incorporated the genetic component into a new sea surface temperature
geochemical proxy developed for high latitudes. We demonstrate that the calcium isotopic
composition (δ44/40Ca) of calcite from genotyped planktonic foraminifera Neogloboquadrina
pachyderma (sin.) is strongly related to temperature and independent of genotype. This
provides an additional well-constrained proxy for the multiproxy approach to the
reconstruction of sea surface temperatures in the high-latitudes.
Genetic diversity and implications for Ca isotope thermometry
1. Introduction
Sea surface temperatures (SSTs) contribute a vital element to our understanding of past and
future climate dynamics. Changes in SST have a strong impact on the thermohaline
circulation, which forms a significant component of the global conveyer belt mechanism (1).
In this context, the polar oceans are of major importance as they represent sensitive key
locations of hydrographic activity within the system. The accurate reconstruction of the SST
history in high-latitude settings is therefore of particular interest.
The different SST proxies based on bio-statistical or geochemical data successfully applied in
the tropics (2-5) are problematic for high-latitude application and are associated with
considerable uncertainty. In particular, interpretation of foraminiferal δ18O is confounded by
the fact that seawater δ18O is altered significantly by fluctuating meltwater discharge related
to complex ice-sheet dynamics (e.g. 6). Further, none of the available δ18O temperature
relationships (7, 8) were calibrated below 4°C. Although attempts to use other proxies such as
the Mg/Ca ratios of foraminiferal calcite resulted in reliable magnitudes of SST changes in
the Subantarctic region (9), Mg/Ca thermometry failed in the North Atlantic, by
overestimating SST (10). Similarly, alkenone proxies also suggest much higher temperatures
for these regions (11), probably due to the presence of ice-rafted, ancient alkenones masking
the autochthonous molecular signal (12). Application of transfer functions is complicated by
the nearly monospecific composition of foraminiferal assemblages at highest latitudes and
shell size dependency (13).
An unforeseen constraint comes from the recently discovered cryptic genetic diversity in
planktonic foraminifers (14, 15, 16), further questioning the reliability of higher latitude SST
proxies. Phylogenetic studies have revealed that many morphologically defined species
(morphospecies) of planktonic foraminifera in fact represent complexes of different and often
highly divergent genetic types (genotypes) which may have different environmental
adaptations. Palaeoceanographers may be unknowingly pooling different genotypes with
unique ecologies. Apparent single species records used in palaeoceanographic reconstructions
in the North Atlantic have already been shown to contain a change in species concurrent with
environmental change (17). Current SST proxies therefore require reassessment and these
inherent shortcomings highlight the requirement for an independent tool for SST
reconstruction in cool temperature environments.
In polar high-latitude oceans Neogloboquadrina pachyderma (sin.) is the dominating
morphospecies in modern and fossil foraminiferal assemblages (18). Due to its bipolar
character, N. pachyderma (sin.) provides the major ecological and geochemical proxy of past
polar ocean conditions in both hemispheres. To date, five distinct SSU genotypes (19) of N.
pachyderma (sin.) have been identified (15, 20) in the Atlantic and Southern Ocean with
different biogeographical distributions. Such differing ecologies imply that
palaeoceanographic proxies based on this taxon should be calibrated independently which
44
Chapter 4: PhD-thesis Dorothee Hippler, Bern (2004)
would require morphological discrimination in the fossil record. This problem can be
bypassed by the use of a geochemical proxy shown to be independent of genotype.
To this end we have developed a new cool water SST proxy based on the Ca isotopic
composition (δ44/40Ca) (21) of genetically characterised individuals of planktonic foraminifera
N. pachyderma (sin.) (22), collected from surface waters. The spatial distribution of sampling
sites covers all major surface water masses of the Atlantic Ocean inhabited by this
morphospecies (Figure 1) (23).
Figure 1: Sampling localities (★) of high-latitude and Benguela N. pachyderma (sin.). The pictorial
map (main figure) shows the five zoogeographical planktonic foraminiferal faunal provinces (24) and
surface currents of the North and South Atlantic. The planktonic morphospecies N. pachyderma (sin.)
used in this study is found predominantly in polar and subpolar provinces, and consequently exhibits a
bipolar distribution. Inset a) and b) highlight the sample sites and distribution of different genotypes in
the Fram Strait and Norwegian Sea (Type I) and in the subpolar/polar Antarctic (Type II, III, IV).
Samples from the Benguela system belong to Type V (see notes). (B and C, large arrows = Benguela
and Canary Current)
45
Genetic diversity and implications for Ca isotope thermometry
2. Results and Discussion
The δ44/40Ca [‰] values of N. pachyderma (sin.) vary between –0.47 and 0.72 in Arctic
specimen (Type I), and between –0.18 and 0.91 in Antarctic specimen (Type II, III, IV).
Samples of the Benguela system (Type V) yield values around 1.75 (Table 1, see Appendix
4). Given the observations that Ca isotopic composition of seawater is homogeneous
throughout modern oceans (25, 26, 27) these variations do not reflect regional differences in
seawater Ca isotopic composition.
Figure 2: δ44/40Ca of N.
pachyderma (sin.) (●) Samples
from the N. Atlantic (Type I),
(▲) the S. Atlantic (Type
III,IV) (▼) the Benguela
system (Type V) are positively
correlated to temperature and
describe a clear temperaturedependence independent of the
genotype. Color (grey or black)
according to used pre-analytical
procedures. Open symbols mark
samples from extreme low
salinity conditions (see Table 1
for data and supplementary
information).
The Ca isotopic composition of N. pachyderma (sin.) is positively correlated to sea surface
temperature (Figure 2). Supporting evidence is given by (25) who observed variation in
δ44/40Ca of 0.5‰ in N. pachyderma (sin.) most probably related to temperature. We observe a
clear linear trend over the temperature range between 2.0 and 14.0°C. The resulting δ44/40Catemperature relationship can be expressed as follows: δ44/40Ca [‰] = 0.19 (±0.01)*SST [°C] –
0.65. Thus a change in δ44/40Ca of 0.19 (±0.01)‰ corresponds to a relative temperature change
of 1°C.
Several lines of evidence indicate that the Ca isotopic composition of foraminiferal calcite is
well preserved in our samples and records the primary temperature signal. The major concern
of secondary alteration or overprint can also be discounted, confirming the preservation of the
primary signal and strongly supports the case that Ca isotope data is a robust SST proxy. In
1998, (28) investigated foraminiferal distribution and ecology and found N. pachyderma (sin.)
resistant to dissolution. Further evidence of primary signal retention is provided by partial
dissolution experiments on shells of N. pachyderma (sin.) which demonstrate that the degree
of dissolution has no impact on the δ44/40Ca values (29). In addition, for the calibration
46
Chapter 4: PhD-thesis Dorothee Hippler, Bern (2004)
described here, the δ44/40Ca-temperature relationship is derived from modern samples where
surface waters are generally supersaturated in respect to calcite, precluding dissolution.
There is no general consensus on the vertical distribution of N. pachyderma (sin.). The first
direct study of the depth habitat of N. pachyderma (sin.) demonstrate that it lives at <100m
depth north of 83° in the Arctic Ocean (30) Peak abundances of N. pachyderma (sin.) were
found in the surface 20-80m, in conjunction with the chlorophyll maximum zone (31) but are
also reported at the depth of 300-500m offshore Oman and Somalia (32). However, their
stable isotope data indicated that N. pachyderma (sin.) calcified in the upper 25m of the water
column (32) confirmed by (33) offshore Greenland. Most evidence suggests that N.
pachyderma (sin.) calcifies at depth similar to other morphospecies used as recorders of SST.
The determination of the genotype and δ44/40Ca on the same shell of N. pachyderma (sin.)
provides direct evidence that there is no genotype dependence concerning the temperature
sensitivity of Ca isotope fractionation. Ca isotope fractionation in planktonic foraminifera is
known to be strongly species dependent (25, 34, 35, 36). Eliminating genetically induced
uncertainties is therefore of particular importance for Ca isotopic use as a SST proxy. Arctic,
Antarctic and Benguela genotypes both describe the same relationship within analytical
uncertainties. A direct consequence of this result is the global validity of the δ44/40Catemperature calibration on the morphospecies level.
However, values from samples collected in true polar surface waters characterised by
temperatures below 2.0°C and low salinities (<33.5‰) and values from samples characterised
only by low salinities (<33.0‰) plot offset from the trendline and overestimate local
temperatures. Our data indicate that the breakdown of the δ44/40Ca-temperature relationship
only occurs in these extreme low-salinity environments, with smaller shell size being a
potential diagnostic factor. These extreme hydrographic conditions are exclusively
characterised by the occurrence of monospecific foraminifera assemblages (13, 18). Therefore
concerning past temperature reconstruction an otherwise unsupported increase in the δ44/40Ca
record associated with reduced shell size reflect a potential salinity crises in the water mass.
In all other open marine high-latitude settings SST reconstruction based on N. pachyderma
(sin.) remains highly robust.
We observed the identical temperature gradient within uncertainties as the one defined for
cultured and modern specimens of planktonic foraminifera G. sacculifer (34, 37.) (Figure 3).
According to these calibrations based on N. pachyderma (sin.) and G. sacculifer respectively,
the magnitude of temperature changes is recorded to the same extent. The offset between the
respective calibrations in predicting absolute temperatures emphasises the significance of
species-specific calibrations. The particular temperature range covered by each of these
calibrations coincides with the preferred SST-habitat of N. pachyderma (sin.) (38) and G.
sacculifer, respectively. The fact that both species obtain the same temperature gradient
indicates identical temperature-dependent Ca isotope fractionation mechanisms, which are
47
Genetic diversity and implications for Ca isotope thermometry
subject to similar biomineralisation processes. In contrast, planktonic foraminifera Orbulina
universa, which is widely distributed with respect to temperature, shows only a weak
temperature dependence (36), similar to inorganic precipitates (36, 39).
Figure 3: The δ44/40Ca-temperature sensitivity of N. pachyderma (sin.) in comparison to δ44/40Catemperature relationships defined for G. sacculifer (37) and O. universa (36). Identical temperature
gradients of N. pachyderma (sin.) and G. sacculifer point to same Ca fractionation mechanisms
suggesting similar biocalcification processes in contrast to O. universa. Samples from extreme low
salinity conditions tend to plot on the extrapolation of the O. universa trendline being potentially
indicative for a switch in biocalcification mode.
To describe the similar Ca isotope fractionation pattern of inorganic carbonate precipitates
and O. universa, (39) postulate equilibrium dynamics. The stronger temperature dependence
of G. sacculifer is interpreted as the result of a superimposed additional biological
fractionation effect assuming different biomineralisation processes for both species. In
contrast, (36) explained the temperature-dependent fractionation of Ca isotopes in the studied
carbonate species as a result of kinetic fractionation. The shallow slope reflects the
involvement of Ca2+-aquacomplexes that are less susceptible to temperature-dependent kinetic
fractionation. The steep slope of G. sacculifer indicates an active dehydration of Ca2+aquacomplexes preceding temperature-controlled biological mediated processes involving the
pure Ca2+-ion prior to calcification. Our results demonstrate that Ca isotope fractionation in N.
pachyderma (sin.) is highly likely to be controlled by similar biologically mediated processes
as those proposed for G. sacculifer.
48
Chapter 4: PhD-thesis Dorothee Hippler, Bern (2004)
Systematic Ca isotope analyses on genetically characterised individuals of modern N.
pachyderma (sin.) covering a temperature range between 2 and 14°C provide an additional
and robust tool for the reconstruction of sea surface temperatures in high-latitude settings
independent of cryptic genetic diversity and secondary alteration processes. The strong
temperature sensitivity of Ca isotope fractionation observed for biogenic calcite tests of N.
pachyderma (sin.), is in excellent agreement with the temperature sensitivity obtained for
tropical planktonic foraminifera G. sacculifer. Therefore our findings render strong evidence
that biocalcification in these two species is related to comparable biochemical mechanisms
controlling Ca isotope fractionation.
Acknowledgement
We are grateful to Jan Kramers for his helpful comments on improving this first version of
the manuscript. D. Hippler and Th. F. Nägler benefit from a research project founded by the
Swiss National Science Foundation (grants 21-61644.00 and 21-60987.00). Kate Darling is
supported by the NERC UK (Award: NER/J/S/2000/00860).
49
Genetic diversity and implications for Ca isotope thermometry
Appendix 4:
Table 1: Calcium isotopic composition (δ44/40Ca) of N. pachyderma (sin.)
Northern Atlanic Ocean, Polarstern (ARK XV/1, 1999)
Station No.
Genotype
SAL [‰]
SST [°C]
δ44/40Ca [‰]1)
Uncertainty2)
δ44/40Ca [‰]3)
A5
3
3
4
Type I
(")
(")
33.5
33.5
33.5
1.0
1.0
1.0
0.55*
0.46*
0.67*
0.28
0.18
0.20
0.20
0.11
0.32
B5
1
2
(")
(")
33.4
33.4
2.0
2.0
0.04*
-0.01*
0.19
0.16
-0.31
-0.36
C5
3
4
(")
(")
33.4
33.4
2.0
2.0
-0.04*
-0.12*
0.21
0.18
-0.39
-0.47
D5
1
(")
33.7
2.0
-0.04*
0.19
-0.39
F5
2
3
(")
(")
32.7
32.7
2.5
2.5
0.27**
0.36**
0.13
0.30
0.27
0.36
H5
3
(")
34.8
4.0
0.45*
0.23
0.10
*
0.22
0.19
I5
2
(")
34.8
4.0
0.54
J5
1
2
2
(")
(")
(")
34.7
34.7
34.7
3.5
3.5
3.5
-0.04**
0.31*
0.23
0.17
-0.04
-0.04
5
(")
35.0
6.5
1.07*
0.24
0.72
*
0.20
0.72
0.18
0.52
K5
L5
1
(")
35.1
6.5
1.07
M5
1
(")
35.0
6.0-6.5
0.87*
Southern Atlantic Ocean, British Antarctic Survey James Clark Ross (JR 48, 2000)
Station
No.
Genotype
SAL [‰]
SST [°C]
δ44/40Ca [‰]1)
Uncertainty2)
δ44/40Ca [‰]3)
8
69
71
Type III
Type III
34.1
34.1
7.0
7.0
0.89*
1.05*
0.25
0.16
0.54
0.70
10
121
129
129
Type III
Type III
Type III
33.9
33.9
33.9
6.0
6.0
6.0
0.88*
0.95*
0.89*
0.25
0.14
0.15
0.53
0.57
0.54
13
165
171
Type III
Type III
33.9
33.9
4.0
4.0
0.53*
0.24
0.18
17
223
231
Type III
Type III
33.7
33.7
3.5
3.5
0.47*
0.20
0.12
23
265
Type III
33.7
2.0
0.17*
0.30
-0.18
30
297
Type III
34.3
2.0-2.5
0.27*
0.30
-0.08
*
36
309
313
321
Type IV
Type IV
Type III
32.4
32.4
32.4
0.0
0.0
0.0
0.67
0.57*
0.52*
0.19
0.21
0.14
0.32
0.22
0.17
85
649
649
651
Type III
Type III
Type III
34.0
34.0
34.0
8.5
8.5
8.5
0.80**
0.79**
0.91**
0.14
0.16
0.15
0.80
0.79
0.91
δ44/40Ca [‰]1)
Uncertainty2)
δ44/40Ca [‰]3)
Benguela System (offshore Namibia), RV Welwitscha (2001)
Station No.
Genotype
SAL [‰]
SST [°C]
10
13
23
Type V
Type V
35.0
35.0
13.5
13.5
1.67**
1.79**
0.22
0.26
1.67
1.79
20
31
35
Type V
Type V
35.0
35.0
14.0
14.0
1.85**
0.29
1.85
δ Ca values are given in [‰] relative to NIST SRM 915a. 2) Uncertainties reflect the 2σ standard
error of single analysis. 3) Preferred δ44/40Ca values (see details below). * and ** refer to different preanalytical techniques which have been applied.
1)
50
44/40
Chapter 4: PhD-thesis Dorothee Hippler, Bern (2004)
Improved pre-analytical procedures (prior to measurements) have been developed since the substantial
suspicion arose that residual organics (related to the EDTA-buffer solution used for DNA-extraktion)
could not be totally removed by column chemistry and a single oxidation step. Therefore, samples were
treated twice with a oxidising HNO3-H2O2-solution (40). Note, that the following considerations have
not been relevant for the study of modern individuals of G. sacculifer (collected from surface water
catches) since the application of both pre-analytical procedures resulted in the same Ca isotope ratios.
Therefore we concluded that the EDTA-buffer is accountable for the analytical-induced offset in
δ44/40Ca values.
In consideration of different pre-analytical treatments the δ44/40Ca-temperature relationship (excluding
samples from low-salinity conditions) can be expressed as follow:
δ44/40Ca-temperature relationship defined for samples treated through column chemistry with only one
oxidising step prior to analyses (*):
δ44/40Ca [‰] = 0.22 (±0.03)*SST [°C] – 0.43 (MSWD = 0.54)
[1]
δ Ca-temperature relationship defined for samples treated according to modified analytics (column
chemistry + 2*HNO3-H2O2) (**):
44/40
δ44/40Ca [‰] = 0.18 (±0.03) SST [°C] – 0.68 (MSWD: 0.48)
[2]
When we consider relationship [1] and [2] separately than both relationships describe (sub-) parallel
trendlines. The gradients are identical within analytical uncertainties and therefore would predict the
the same magnitude of temperature change, while absolute temperature estimates would differ
insignificantly. The observation of gradient identity and remeasurement of sample J5 allowed the
empirical correction of -0.35‰ to minimise the consistant offset (δ44/40Ca values in the last column). In
accordance with these results our prefered suggestion is the application of the δ44/40Ca-temperature
relationship defined for pooled samples (old and improved analytics) for past temperature
reconstructions:
δ44/40Ca [‰] = 0.19 (±0.03) SST [°C] – 0.65 (MSWD: 0.95)
[3]
The mathematically determined uncertainty on the gradient of ±0.01 has been set to the more
realistic value of ±0.03 considering slope uncertainties of [1] and [2] and the empirical offset
correction.
References and notes
1. Broecker W. S. and Peng T.-H. (1982)
Tracers in the Sea. Eldigo Press.
2. Cayre O. and Bard E. (1999)
Planktonic foraminiferal and alkenone records of the last deglaciation from the Eastern
Arabian Sea. Quaternary Research 52, 337-342.
3. Rühlemann C., Mulitza S., Müller P. J., Wefer G. and Zahn R. (1999)
Warming of the tropical Atlantic Ocean and slowdown of thermohaline circulation during the
last deglaciation. Nature 402, 511-514.
51
Genetic diversity and implications for Ca isotope thermometry
4. Lea D. W., Pak D. K., Peterson L. C. and Hughen K. A. (2003)
Synchroneity of tropical and high-latitude Atlantic temperatures over the last glacial
termination. Science 301, 1361-1364.
5. Visser K., Thunell R. and Stott L. (2003)
Magnitude and timing of temperature change in the Indo-Pacific warm pool during
deglaciation. Nature 421, 152-155.
6. Jones G. A. and Keigwin L. D. (1988)
Evidence from Fram Strait (78°N) for early deglaciation. Nature, 336, 56-59.
7. Kim S. T. and O’Neil J. R. (1997)
Equilibrium and nonequilibrium oxygen isotope effects in synthetic carbonates. Geochim.
Cosmochim. Acta, 61, 3461-3475.
8. Bemis B. E., Spero H. J., Bijma J. and Lea D. W. (1998)
Re-evaluation of the oxygen isotopic composition of planktonic foraminifera: Experimental
results and revised paleotemperature equations. Paleoceanography 13(2), 150-160.
9. Mashiotta T. A., Lea D. W. and Spero H. J. (1999)
Glacial-interglacial changes in Subantarctic sea surface temperature and δ18O-water using
foraminiferal Mg. Earth Planet. Sci. Lett., 170, 417-432.
10. Kandiano E. S., Bauch H. A. and Müller A. (2004)
Sea surface temperature variability in the north Atlantic during the last two glacialinterglacial cycles: comparison of faunal, oxygen isotopic, and Mg/Ca-derived records.
Palaeo3, 204, 145-164.
11. Rosell-Mele A. and Comes P. (19999
Evidence for a warm Last Glacial Maximum in the Nordic Seas or an example of
shortcomings in UK37‘ and UK37 to estimate low sea surface temperatures. Paleoceanography,
14 (6), 770-776.
12. Weaver P. P. E., Rutledge D., Chapman M. R., Eglinton G., Read G. and Zhao M. (1999)
Combined coccolith, foraminiferal, and biomarker reconstruction of paleoceanographic
conditions over the last 120 kyr in the northern Atlantic (59°N, 23°W). Paleoceanography, 14,
336-349.
13. Huber R., Meggers H., Baumann K.-H., Raymo M. E. and Henrich R. (2000)
Shell size variation of the planktonic foraminifer Neogloboquadrina pachyderma sin. in the
Norwegian-Greenland Sea during the last 1.3Myrs: implications for paleoceanographic
reconstructions. Palaeo3, 160, 193-212.
52
Chapter 4: PhD-thesis Dorothee Hippler, Bern (2004)
14. de Vargas C., Norris R., Zaninetti L., Gibb S. W. and Pawlowski J. (1999)
Molecular evidence of cryptic speciation in planktonic foraminifers and their relation to
oceanic provinces. Proc. Nat. Acad. Sci., 96, 2864-2868.
15. Darling K., Wade C. M., Stewart I. A., Kroon D., Dingle R. and Brown A. J. L. (2000)
Molecular evidence for genetic mixing of arctic and Antarctic subpolar populations of
planktonic foraminifers. Nature, 405, 43-47.
16. Stewart I. A., Darling K., Kroon D., Wade C. M. and Troelstra S. R. (2001)
Genotypic variability in subarctic Atlantic planktic foraminifera. Marine Micropaleontology
43, 143-153.
17. Bauch D., Darling K., Bauch H. A., Erlenkeuser H. and Kroon D. (2003)
Palaeoceanographic implications of genetic variation in living North
Neogloboquadrina pachyderma. Nature, 424, 299-302.
Atlantic
18. Pflaumann U., Duprat J., Pujol C. and Labeyrie L. D. (1996)
SIMMAX: A modern analog technique to deduce Atlantic sea surface temperatures from
planktonic foraminifera in deep-sea sediments. Paleoceanography, 11(1), 15-35.
19. DNA extraction, amplification by polymerase chain reaction (PCR) and cloning or direct
automated sequencing of an ~1000-b.p. region of terminal 3’ end of the foraminiferal SSU
rRNA gene were as described previously (15, 16). Genetic determinations of N. pachyderma
(sin.) have been derived from a consensus sequence amplified from the gene family of a
single individual using a direct sequencing approach.
20. Darling K., Kucera M., Pudsey C. J. and Wade C. M. (2004)
Molecular evidence links cryptic diversification in polar planktonic protists to Quaternary
climate dynamics. Proc. Nat. Acad. Sci., 101 (20), 7657-7662.
21. Ca isotope variations are expressed in the δ-notation: δ44/40Ca [‰] = {(44Ca/40Ca)sample /
(44Ca/40Ca)standard – 1} * 1000, where the standard is NIST SRM 915a (27, 41) A detailed
compilation of reference materials is given by (27). The techniques for the chemical
separation and mass spectrometric analysis using a 43Ca-48Ca double-spike technique of
calcium are modified from those described in (16, 27, 37. Prior to and after column chemistry
it is essential that samples are treated with a HNO3-H2O2 solution to remove residual organics
(40). Ca isotope analyses were all performed at the University of Bern (CH). The 2σ-sample
reproducibility of δ44/40Ca is 0.2‰.
22. An important advantage of our combined approach is that genetic determination and Ca
isotope analysis were performed on the same individual. The sensitivity of the Ca isotope
method allows replicate analysis of single shells resulting from the ability to measure very
small quantities of Ca (200-500ng).
53
Genetic diversity and implications for Ca isotope thermometry
23. Samples were collected on board of RV Polarstern (ARK XV/1, 1999) in the Nordic seas
at 75°N (from 13°W to 13°E). South Atlantic samples were collected along transects between
the Falkland Islands (53°21‘S, 58°20‘W) and the Antarctic Peninsula (65°36‘S, 77°39‘W) on
board of RRS James Clark Ross (JR48) and samples from the Benguela system were taken
offshore Namibia (23°S) on RV Welwitschia (November 2001). Samples were obtained either
by pumping continually from the surface water layer (6m, 63-µm filter) or from vertical
plankton tows (≤100m, 63µm mesh). Sea surface temperature was between –1.0 and 14.0°C
and salinity vary between 32.4 and 35.0‰. For more details see (20).
24. Bé A. W. H. (1977)
An Ecological, Zoogeographic and Taxonomic Review of Recent Planktonic Foraminifera.
Academic Press.
25. Zhu P. and MacDougall J. D. (1998)
Calcium isotopes in the marine environment and the oceanic calcium cycle. Geochimica et
Cosmochimica Acta 62, 1691-1698.
26. Schmitt A.-D., Bracke G., Stille P. and Kiefel B. (2001)
The calcium isotope composition of modern seawater determined by thermal ionisation mass
spectrometry. Geostandards Newsletter: The Journal of Geostandards and Geoanalysis 25,
267-275.
27. Hippler D., Schmitt A.-D., Gussone N., Heuser A., Stille P., Eisenhauer A. and Nägler T.
F. (2003)
Calcium isotopic composition of various reference materials and seawater. Geostandards
Newsletter: The Journal of Geostandards and Geoanalysis 27, 13-19.
28. Martinez J. I., Taylor L., De Dekker P. and Barrows T. (1998)
Planktonic foraminifera from eastern Indian Ocean: distribution and ecology in relation to the
Western Pacific Warm Pool (WPWP). Marine Micropaleontology, 34, 121-151.
29. Hönisch B. (2002)
Stable isotope and trace element composition of foraminiferal calcite - from incorporation to
dissolution. PhD Thesis, Universität Bremen, Bremen, 118 pp.
30. Carstens J. and Wefer G. (1992)
Recent distribution of planktonic foraminifera in the Nansen basin, Arctic Ocean, Deep Sea
Res., 39 (S2), S507-S524.
31. Kohfeld K. E., Fairbanks, R. G., Smith S. L. and Walsh I. D. (1996)
Neogloboquadrina pachyderma (sinistral coiling) as paleoceanographic tracers in polar
oceans: evidence from Northeast Water Polynya plankton tows, sediments traps, and surface
sediments. Paleoceanography, 11(6), 679-699.
54
Chapter 4: PhD-thesis Dorothee Hippler, Bern (2004)
32. Ivanova E. M., Conan S. M.-H., Peeters F. J. C. and Troelstra S. R. (1999)
Living Neogloboquadrina pachyderma (sin.) and its distribution in the sediments from Oman
and Somalia upwelling areas. Marine Micropaleontology, 36, 91-107.
33. Simstich J. (1999)
Variations in the oceanic surface layer of the Nordic Seas: The stable-isotope record of polar
and subpolar planktonic foraminifera. Ber.-Rep., Inst. für Geowiss. Universität Kiel, Nr. 2, p.
96.
34. Nägler T. F., Eisenhauer A., Müller A., Hemleben C. and Kramers J. (2000)
The δ44Ca-temperature calibration on fossil and cultured Globigerinoides sacculifer: New tool
for reconstruction of past sea surface temperatures. Geochemistry, Geophysics, Geosystems
1, 2000GC000091.
35. Skulan J., DePaolo D. J. and Owens T. L. (1997)
Biological control of calcium isotopic abundances in the global calcium cycle. Geochimica et
Cosmochimica Acta 61, 2505-2510.
36. Gussone N., Eisenhauer A., Heuser A., Dietzel M., Bock B., Böhm F., Spero H. J., Lea D.
W., Bijma J. and Nägler T. F. (2003)
Model for kinetic effects on calcium isotope fractionation (δ44Ca) in inorganic aragonite and
cultured planktonic foraminifera. Geochimica et Cosmochimica Acta 67, 1375-1382.
37. Hippler D., Eisenhauer A. and Nägler T. F. (subm.)
Tropical SST history infered from Ca isotope thermometry over the last 140ka. Geochimica et
Cosmochimica Acta.
38. Hilbrecht H. (1997)
Morphologic gradation and ecology in Neogloboquadrina pachyderma and N. dutertrei
(planktic foraminifera) from core top sediments. Marine Micropaleontology, 31., 31-43.
39. Marriott C. S., Henderson G. M., Belshaw N. S. and Tudhope A. W. (2004)
Temperature dependence of δ7Li, δ44Ca and Li/Ca during growth of calcium carbonate. Earth
and Planetary Science Letters 222, 615-624.
40. Hippler D., Villa, I. M., Nägler Th. F. and Kramers J. D. (2004)
A ghost haunts mass spectrometry: real isotope fractionation or analytical paradox?
Geochimica et Cosmochimica Acta, 68, A215.
41. Eisenhauer A., Nägler T. F., Stille P., Kramers J., Gussone N., Bock B., Fietzke J.,
Hippler D. and Schmitt A.-D. (2004)
Proposal for international agreement on Ca notation resulting from discussions at workshops
on stable isotope measurements held in Davos (Goldschmidt 2002) and Nice (EGS-AGUEUG 2003). Geostandards Newsletter: The Journal of Geostandards and Geoanalysis 28.
55
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
A critical assessment of mollusk 18O/16O, Mg/Ca, and 44Ca/40Ca ratios
as proxies for Cretaceous seawater temperature seasonality
Adrian Immenhauser1, Thomas F. Nägler2, Thomas Steuber3 and Dorothee Hippler2
1
Vrije Universiteit, Department of Earth and Life Sciences, Amsterdam, The Netherlands
2
University of Bern, Institute of Geological Sciences, Isotope Geology, Bern, Switzerland
3
Ruhr-Universität Bochum, Institute of Geology, Mineralogy and Geophysics, Bochum, Germany
Palaeogeography, Palaeoclimatology, Palaeoecology (Article in press)
Abstract - Seasonal temperature fluctuations of Tethyan coastal water are recorded in a pristine, ≈80
Myr-old rudist shell from Turkey. In order to critically assess factors other than seawater temperature
that might have influenced the shell geochemical record, three different palaeo-temperature proxies
(δ18O, Mg/Ca and δ44/40Ca) were applied. In a qualitative manner, all three proxies reflect the same
cyclical trends and thus yield robust evidence for seasonal fluctuations in Late Cretaceous surface
seawater temperature. This suggests the successful application of the new δ44/40Ca-temperature proxy to
fossil mollusk calcite. The direct comparison of the three data sets, however, demonstrates that all
proxies are fraught with problems. Similar to other studies, 18O/16O ratios point to warmer summer
temperature maxima than those in comparable coastal settings today. Nevertheless, the 18O/16O proxy is
subject to environmental factors that might lead to an overestimation of peak temperatures. Moreover,
Mg/Ca molar ratios are less sensitive to environmental factors but were strongly affected by the ion
regulating capability of the rudist bivalve that responded to the low Mg/Ca ratio of Cretaceous
seawater. Similarly, uncertainties of the δ44/40Ca composition of Cretaceous seawater and the complex
bio-calcification of mollusks presently limit the interpretation of δ44/40Ca values in terms of absolute
seawater temperature. The multiproxy approach applied here, however, documents that these
limitations do not obscure the dominant patterns of seasonal sea surface temperature variations
recorded in the biogeochemical rudist archive.
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
1. Introduction
Understanding the past climate evolution is essential for modelling future climate
development. Amongst the most fundamental parameters of the Earth’s climate system are
high-resolution sea surface temperature records (SST; Crowley, 1991, Barron, 1995; Wilson
and Opdyke, 1996; Wefer et al., 1999, Norris et al., 2001, Wilkinson and Ivany, 2002). The
reconstruction of palaeo-SST, however, needs a temperature-dependent quantity that can be
measured (SST proxy), and a geologic archive that preserved a record of this quantity.
For the past 103-104 years, aragonitic coral skeletons represent one of the most time-resolved
(weeks to centuries) marine archives of past climate variability and corals are widespread in
(sub) tropical seas (Gagan et al., 2000 and references therein). Cretaceous corals, however,
did not preserve a comparable palaeoclimate record as the primary geochemical composition
of their skeletons was significantly altered during diagenesis. Therefore, diagenetically more
stable archives of the Cretaceous climate with a high temporal resolution are needed. At
present, most Cretaceous SST reconstructions are based on the temperature-dependent 18O/16O
ratios of planktonic foraminifera tests, for instance from Ocean Drilling Project core material
(Norris et al., 2001). Foraminiferal records, however, do not resolve sub-annual temperature
fluctuations. In contrast, well-preserved low-Mg calcite shells of rudist bivalves, that
flourished in the shallow (sub) tropical Cretaceous seas next to corals, might represent such
time-resolved archives (Steuber, 1996).
The purpose of this paper is twofold. First, the possibilities and limits of the new 44Ca/40Catemperature proxy (Nägler et al., 2000) applied to Cretaceous skeletal calcite are explored;
second, the compatibility of multi-proxy palaeoclimate records from low-Mg calcite rudist
shells is tested in a critical and qualitative manner. The ultimate goal of this approach is the
evaluation and quantification of factors that limit the application of three independent palaeotemperature proxies responding differently to environmental, biological, and diagenetic
influences.
2. Carbonate material, methods and sampling strategy
2.1 Sampling locality and age of carbonate material
The hippuritid rudist Vaccinites ultimus (Milovanovic) (Steuber et al., 1998; Steuber, 1999)
used for this study is derived from a Campanian carbonate platform fronting onto a central
Tethyan seaway (North Central Turkey; ≈25ºN palaeo-latitude), i.e. from the northern margin
of the tropical realm (Figure 1a).
The sampling locality of V. ultimus lies ca. 5km WSW of Amasya near the Höbek Tepe in
exposures of the Upper Cretaceous Lokman Formation (N40°39’14’’/E35°54’21’’; Alp,
1972; Steuber et al., 1998). At this locality, the Lokman Formation is built by transgressive-
58
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
regressive sequences of which rudist floatstone facies overlies a first transgressive unit of
lithobioclastic packstones (Figure 1; Steuber et al., 1998). The Upper Cretaceous sedimentary
deposits unconformably (but autochthonously) rest on trusted ophiolitic imbricates on top of a
rifted Jurassic-Early Cretaceous carbonate platform (Steuber et al., 1998).
Figure 1a) Locality map of the eastern Mediterranean region with indication of the rudist sampling
locality (✩, Amasya area, N40°39’14’’/E35°54’21’’). b) Schematic stratigraphic section of the
Mesozoic and Cenozoic deposits of the Amasya area. Lower Campanian platform limestones
comprising the rudist V. ultimus are shaded grey (modified after Steuber et al. 1998). c) Campanian
stratigraphy. The age of the rudist V. ultimus is 81.26 (± 0.63 ) Ma based on Sr-isotope stratigraphy.
The rudist shell was preserved in vertical life position and associated with other rudist species
and colonial corals within a tabular lithosome (Steuber et al., 1998). Similar to most other
occurrences of rudist bivalves, Vaccinites grew in the shallow subtidal, i.e., an estimated
depth range of 3-15 meters (Gili et al., 1995). The rudist shell therefore recorded seawater
temperature and geochemistry of a bathymetric interval reaching from near sea surface to
shallow depths of about 15 meters, i.e. above the annual (summer) thermocline.
59
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
The specimen investigated has a stratigraphical age of 81.26 Ma ± 0.63 Ma based on Srisotope stratigraphy (Steuber 2003). This age refers to the Early Campanian (Figure 1b).
2.2 Methods
Two hundred and fifty-nine carbonate powder samples (20 to 80µg on average) were analysed
on a ThermoFinnigan MAT 252 ratio mass spectrometer for carbon and oxygen isotope ratios
(Figure 2). Repeated analyses of carbonate standards show a reproducibility of better than
0.1‰ for δ18O and better than 0.05‰ for δ13C. Duplicate samples scatter on the order of ±
0.1‰ or less for δ18O and δ13C. All isotope results are reported in ‰ relative to the VPDB
standard in the conventional manner.
Eighty-five samples were analysed for Ca, Mg, Mn, and Fe elemental concentrations (Figure
2). In order to remove possible organic material from calcite powder samples, we used a
customised version of the Cambridge cleaning method for planktic foraminifera (Elderfield
and Ganssen, 2000). The cleaning of these shell samples was comparably simple, as they
were free of clay minerals, a common source of contamination in foraminifera tests. Data
from 20 duplicate measurements using untreated sample powder, however, differed only
within the limits of the analytical precision from values obtained from cleaned samples. This
implies that the preserved organic matter content in these samples is low. Element
concentrations were measured by inductively coupled plasma-atomic emission spectrometry
(ICP-AES) after dissolution of ≈ 0.8 to 1mg of powdered sample in 1 N HNO3 with
subsequent dilution to a ≈0.1 N HNO3 sample solution. The sample size made accurate
measurements of powder weights difficult. We thus converted raw solution data to Mg/Ca
ratios by assuming that all Ca and Mg are present as carbonate-bound cations, a method
similar to the one applied by Klein et al. (1996a). The resulting analytical precision of
repeated analyses is in the order of ±1.5% for Ca and ±8% for Mg, Sr, Mn and Fe.
Fourteen samples were selected for δ44/40Ca analysis, nine of which from intervals with lowest
and highest δ18O values (presumably Tmax and Tmin), respectively (Figure 2). The small
number of samples selected for 44Ca/40Ca analysis reflects the significantly more time
consuming analytical procedure involved resulting in a slow sample throughput (one analysis
per day on average). The δ44/40Ca values were measured applying a 43Ca/48Ca double spike
technique. Calcium runs were done on a modified single cup AVCO® mass spectrometer
(Nägler et al., 2000) with a new Thermolinear® source. External reproducibility was 0.2‰
(2σ
standard
deviation).
The
δ-value
is
calculated
as
δ44/40Ca
=
44
40
44
40
*
[( Ca/ Casample)/( Ca/ Castandard) -1] 1000, where the standard is the same dissolved fluorite as
in Nägler et al. (2000). The δ44/40Ca of the international NIST SRM 915a standard is -1.44 ±
0.03‰ (2σ weighted error) relative to the fluorite standard. The δ44/40Ca of modern seawater
is +0.44 ± 0.04‰ relative to fluorite and +1.88 ± 0.04‰ relative to NIST SRM 915a (Hippler
et al., 2003).
60
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
Figure 2: Sampling strategy. a) Schematic drawing of right valve of Vaccinites ultimus cut
longitudinally. Grey area represents portion of shell that was mounted on glass plate and polished. b)
Drawing of V. ultimus shell. Dashed lines indicate sampled portion of shell. Numerals refer to
geochemical maxima and minima respectively as depicted in Figure 5 and Table 1. Sampling localities
and type of geochemical analyses are indicated. c) and d) Enlargement of shell portions with indication
of sampling localities and geochemical analyses made. Note growth increments.
61
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
Thin sections of the rudist shell were investigated under a cold-stage cathode luminescence
microscope operating at 10 to 14kV accelerating voltage, 200 to 300µA beam current, and a
beam diameter of 4mm. Backscatter images of the rudist shell were obtained using an electron
microprobe operating at 20kV, beam current 0.015mm, and 13mm beam diameter (Figure 3;
detection limits for Sr = 200ppm, Mg = 100ppm, Mn = 200ppm, Fe = 200ppm). Images
shown in Figure 3 were taken by use of a field emission Jeol 6301 scanning electron
microscope.
2.3 Assessment of shell preservation
Following the approaches established by previous authors (e.g., Brand and Veizer, 1980; AlAasm and Veizer, 1986), the fossil low-Mg calcite shell used in this study has been tested for
its degree of preservation, i.e. the possible impact of diagenetic alteration on shell
geochemistry. The following methods were applied: (i) Scanning electron microscopy in
order to asses the preservation of the ultra-structure only recognised in pristine (i.e. nonrecrystallized) shell material (Figure 3); (ii) cathode luminescence microscopy and
microprobe backscatter imaging in order to detect possibly altered zones; (iii) ICP-AES
analysis of shell elemental abundances to detect geochemical evidence of diagenetic
(meteoric and burial) alteration, particularly elevated Mn and Fe and lowered Sr and Mg
values characteristic for meteoric diagenesis (Figure 4).
Figure 3a) and b) Scanning-electron images of V. ultimus used in this study. View is parallel to longest
growth axis. Growth increments (gr) and, c-axes of calcite crystals (ca) are indicated. Note excellent
preservation of rudist low-Mg calcite. Black holes are open void (v) formerly filled by organic
material. c) and d) Electron microprobe backscatter images of boundary between inner, formerly
aragonitic shell and outer, low-Mg shell used in this study. Note excellent preservation. Arrows points
upward or in growth direction respectively.
62
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
Scanning electron microscopy of the V. ultimus specimen indicates a well-preserved shell
structure, with the exception of alteration of about 1mm-wide zones in the direct vicinity of
sediment-filled boreholes that penetrate the outer shell at two localities. The high degree of
shell preservation is also indicated by backscatter imaging (Figure 3). Bored areas were
consequently avoided when collecting shell material for chemical analysis. Under cathode
luminescence, the shell appeared non-luminescent (black), again with the exception of thin
rims of patchy luminescence near borings.
Results of ICP-AES analyses of rudist shell (Sr, Mg, Mn, and Fe concentrations) are shown in
Figures 4 and 5 and in Table 1. Strontium and Mn values obtained from the low-Mg calcite
outer shell plot well into the range of recent marine bivalves (Figure 4a) and close to the line
of modern (pristine) biotically precipitated calcite (Figure 4b; Carpenter and Lohman, 1992).
Figure 4: Geochemical preservation of rudist low-Mg calcite of V. ultimus. a) Strontium and Mn
values of outer low-Mg calcite rudist shell plotting well within the trace elemental composition of
recent marine bivalves (Steuber, 1999). b) Range of rudist Mg and Sr elemental composition plotting
very close to the pristine biotic calcite line of Carpenter and Lohmann (1992).
Nevertheless, it has been demonstrated that some of these screening techniques might fail
under exceptional geological situation. For example, Barbin et al. (1991) found luminescent
intervals in recent bivalve shells, i.e. such that clearly had not been altered diagenetically.
Moreover, Banner and Hanson (1990) showed that under specific circumstances,
extraformational brines could lower carbonate δ18O values without affecting carbonate trace
elemental concentrations. This type of alteration, however, usually leads to a highly variant
δ18O signature with typical values of around -8 to –10‰ as opposed to the regular, cyclical
isotope shifts observed in V. ultimus. Similarly, a pervasive meteoric diagenetic overprint
characteristically results in an invariant δ18O and a variant δ13C signature (the ‘meteoric water
line’ of Lohman, 1988). Finally, it is of significance that δ13C shows an anti-cyclical pattern
with respect to δ18O (cf. sections II, III and IV in Figure 5). This is not indicative of
diagenetic alteration as for instance meteoric fluids, in combination with soil-zone CO2,
63
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
causes both δ18O and δ13C to shift towards lower values (see discussion in Immenhauser et al.,
2002).
It is therefore concluded, that in combination, all optical and geochemical screening methods
point to well-preserved low-Mg calcite of the V. ultimus specimen under study.
Sample
H 584 13 o
H 584 12 o
H 584 11 o
H 584 10 o
H 584 8 m
H 584 7 m
H 584 6 m/a
H 584 6 m/b
Ru6
H 584 5 m
H 584 4m
H 584 3 u/a
H 584 3 u/b
H 584 2 u
H 584 1 u
Ru4
Cycle number
(Fig. 5)
12
11
10
9
8
7
6-7
6-7
6
5
4
3
3
2
1-2
1
44
Ca
‰
-1.14
-1.68
-1.17
-1.35
-0.73
-1.84
-1.41
-1.40
-0.73
-1.38
-1.31
-1.76
-1.76
-1.23
-1.30
-1.35
±2
±0.07
±0.09
±0.09
±0.17
±0.08
±0.11
±0.20
±0.17
±0.20
±0.17
±0.21
±0.22
±0.17
±0.08
±0.14
±0.13
18
O
‰
-4.8
-4.0
-4.7
-4.2
-4.9
-3.8
-5.0
-5.0
-4.3
-4.3
-4.7
-3.7
-3.7
-4.7
-4.4
-4.1
13
C
‰
2.3
1.8
2.3
1.9
2.2
1.9
2.1
2.1
2.3
1.9
1.8
1.8
1.8
2.2
1.9
1.6
Mg/Ca
x1000
Mg
(ppm)
Sr
(ppm)
Fe
(ppm)
Mn
(ppm)
9.3
8.6
9.2
6.6
9.7
7.1
7.9
n.d.
n.d.
6.6
n.d.
6.7
n.d.
8.5
7.1
n.d.
2268
2089
2238
1594
2360
1734
1926
n.d.
n.d.
1596
n.d.
1624
n.d.
2060
1732
n.d.
1431
1541
1401
1447
1627
1423
1514
n.d.
1459
1363
1420
1566
n.d.
1396
n.d.
1566
80
65
69
135
120
168
90
n.d.
38
34
162
16
n.d.
6
n.d.
35
b.d.l.
b.d.l.
b.d.l.
6
5
b.d.l.
5
n.d.
b.d.l.
b.d.l.
20
b.d.l.
n.d.
b.d.l.
n.d.
b.d.l.
Table 1: Overview of δ44Ca, δ18O, Mg/Ca, Mg, Sr, Mn and Fe analytical results from splits of large
powder samples collected in sclerochronological Section I (Figure 2). Cycles with odd numbers are
shaded grey. A detailed list with all geochemical data is available from the corresponding author on
request.
2.4 Rudist growth rates and sampling strategy
As laid out in detail in Steuber et al. (1998), the growth rates of V. ultimus were in the order
of 35-40 mm/year without any indication of adult growth deceleration. Estimates of rudist
growth rates are based on seasonal variations of rudist sclerochronical isotope sections
(Steuber, 1996). Judging from geochemical evidence (Figure 5), the V. ultimus specimen
investigated here recorded perhaps 8-10 seasonal cycles of which 5 or 6 were sampled (Figure
2).
The shell was cut longitudinally and a 4-mm-thick polished slice was mounted on a glass
plate for sampling (Figure 2a). Samples in three sclerochronological profiles were drilled with
a dental burr parallel to the growth axis (Figure 2b). Two sclerochronological sections were
drilled from the left side of the shell (sections I and II; Figures 2b-d and 5). Section I was
sampled in a low temporal resolution and for large powder volumes from which splits were
analysed for δ44Ca, δ18O, δ13C values and trace-elemental compositions (Figure 2c; Mg, Sr,
Mn, Fe). Hence, in terms of temporal resolution, each sample represents several weeks. The
resulting data are listed in Table 1.
64
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
Samples for Section I have been drilled at δ18O maxima and minima of Section II (Figure 5).
Section II provided small, densely spaced powder samples resulting in a highly time-resolved
δ18O record with sharp maxima and minima, but sample volumes were too small for other
analyses except for δ13C. The sampling increments are in the order of 1mm or less. Given the
rudist growth rate of 35-40 mm/year this implies that each sample covers approximately one
or two weeks in the life history of V. ultimus.
The third sclerochronological profile was drilled from the right side of the rudist valve
(Figures 2b, d, 5). Here the sampling resolution was lower (increments of ca. 3mm) compared
to section II as larger sample volumes for combined measurements of δ18O (section III) and
δ13C values (section IV) and trace elements (sections IV-VII) were needed (Figure 5). This
results in a sample-time ratio of approximately 3 to 4 weeks per sample point.
3. Results
3.1. Oxygen and carbon-isotope ratios
Oxygen-isotope analyses of the rudist specimen investigated reveal a pronounced δ18O
cyclicity with values ranging between -3.2 and -5.1‰ (section II; Fig. 5). In the
complementing section III, values range between -3.7 and -5.1‰. In order to estimate
seawater temperature from δ18O values, two different assumptions are used for the δ18O of
(sub-) tropical Late Cretaceous and Early Cenozoic seawater; the more commonly used -1‰
SMOW (e.g., Pearson et al., 2001) and -0.5‰ SMOW (Zachos et al., 1994). Applying the
temperature equation of Anderson and Arthur (1983), that is specifically based on bivalve
calcite, to the highest (-3.2‰) and lowest (-5.0‰) rudist δ18O values of section II (Figure 5),
temperature minima (Tmin) of 28.1°C and maxima (Tmax) of 34.7°C result for-1‰ δ18O SMOW
of seawater. Applying the geologically perhaps less reasonable δ18O of -0.5‰ SMOW of
Cretaceous seawater to these values, Tmin of 26.7°C and Tmax of 37.3°C results. Cycle
amplitudes of up to 6.6ºC result, irrespective of the absolute values of seasonal SST.
When comparing the oxygen isotope data of section II, with those of section III (Figure 5), it
becomes evident that the amplitudes differ by ≈0.5‰. Particularly, the isotope peaks with odd
numbers (1, 3, 5, etc.) are usually higher in section II by about 0.5‰ compared to the relative
peaks in section III. This feature reflects differences in sampling resolution and the amount of
carbonate powder drilled from one sampling spot, respectively. Section II was sampled with a
high spatial resolution and with small sample volumes. Each peak thus represents the isotopic
composition of a few sub-millimetre-thick rudist growth increments representing a time
interval of perhaps one or two weeks.
65
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
66
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
Section III was sampled with a relatively low resolution and sample volumes were larger as
material for δ18O and trace elemental analysis was collected (Figure 5). Each sample point
thus reflects a geochemical average across numerous growth increments in total representing
perhaps 3 to 4 weeks.
Peaks 1 and 9 of section III (Figure 5) were re-sampled with a spatial resolution and sample
sizes that are comparable to the sampling technique used in section II. The resulting δ18O
values (white circles to the right of section III in Figure 5) are very similar to those of section
I. The same feature is also observed when comparing the δ18O values of large-volume
samples drilled for calcium and oxygen-isotope analysis and Mg/Ca ratios (black bars in
section II) with high-resolution δ18O values plotted in section II (Table 1). This clearly
illustrates that the sampling resolution and the sample size are important aspects that must be
considered when comparing different data sets (Wilkinson and Ivany, 2002). This is a similar
observation to that experienced in corals (Gagan et al., 2000).
Carbon-isotope data from V. ultimus (section IV) range between 1.5 and 2.5‰ with a mean of
2.1‰ (standard deviation (s) = 0.22). As illustrated in Figure 5, δ13C shifts in a very regular
cyclical manner but anti-correlates with δ18O, i.e. δ13C maxima correlate to δ18O minima and
vice versa.
3.2. Elemental composition
The rudist shell has been investigated for its Mg, Mn, Sr and Fe elemental composition
(Figures 2, 4, 5 and Table 1). Here, we refer to Mn, Sr and Fe whereas the shell Mg/Ca molar
ratios are discussed in the following chapter.
Strontium varies in a cyclical manner with pronounced cycles found in the lower shell
interval (Figure 5, section VI, cycles 1-3) and decreasing cycle amplitudes upsection. In
general, maxima are as high as 1628ppm and minima are at 1249ppm (average = 1441ppm; (s
= 80ppm). The Sr values obtained are clearly within the range of Sr values of recent marine
bivalves (Figure 4a) and match those of other well-preserved Cretaceous rudist shells (see
examples in Steuber, 1999; 2002).
It is perhaps of interest to add that meteoric diagenesis typically results in decreasing Sr levels
(e.g., Al-Aasm and Veizer, 1986) due to the low Sr content of rainwater. Nevertheless, in the
rudist shell investigated, increasing Sr levels coincide with decreasing δ18O values, i.e. the
opposite of what would be expected if diagenesis fluids were the agent of isotope and
elemental shifts.
With reference to Fe elemental compositions, maxima are 738ppm with minima below
detection limit and an average of 95ppm (s=106ppm). As shown in Figure 5 (section VII), Fe
displays no recognisable cyclicity such as observed in the case of Mg or Sr but a rather
67
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
invariant pattern. Exceptions are some elevated values at the base of section VII and a minor
trend to higher concentrations towards the top of the section. In general, the Fe concentrations
are very similar to average values of many well-preserved rudist shells as summarised in
Steuber (1999). Meteoric diagenesis usually results in increased Fe levels, which is clearly not
the case here. The same accounts for Mn, which is typically elevated by diagenetic fluids but
here, most analyses were below detection limit (Table 1) and Mn is thus not shown in Figure
5.
3.3. Magnesium/Calcium molar ratios
Mg/Ca molar ratios (*1000) from the rudist shell range between 5.36 and 12.89 (average =
8.11; s = 1.28) and are plotted in section V (Figure 5). The cyclicity displayed by Mg/Ca
ratios matches cycles of δ18O in sections I and III, although high-amplitude δ18O cycles are
not necessarily accompanied by high-amplitude Mg/Ca ratio cycles. Again this might be a
sampling problem. Applying the Mg/Ca ratio temperature equation for modern bivalves of
Klein et al. (1996a) to our data, Tmin of 11.3ºC and Tmax of 30.2ºC result. Judging from this,
the seasonal variations of SST (Tmax-Tmin) based on the Mg/Ca molar ratio proxy thus perhaps
had amplitudes of up to 18.9ºC. Although comparative studies of modern (Mytilus) and fossil
(Vaccinites) organisms (including their metabolism) are regularly performed, such
considerations are based on circumstantial evidence and thus must be treated with care. It is
thus not clear if the Klein et al. (1996a) temperature equation is adequate for application to
skeletal calcite of fossil rudist bivalves. Amongst other constraining factors, it is perhaps only
valid for a temperature range of 5 to 23°C. Within this range, the temperature-Mg/Ca
relationship is linear but this might not be the case for temperatures higher than 23°C.
Moreover, Vander Putten et al. (2000) have found that modern Mytilus edulis covaries
reproducibly with seawater temperature during spring months but this covariation was found
to be abruptly interrupted after the spring phytoplankton bloom. These authors thus
concluded, in contrast to the work of Klein et al. (1996a), that the absence of a constant Mgtemperature relationship over the year hampers the use of mollusk shells as a direct seawater
temperature proxy.
3.4. Calcium-isotope ratios
The δ44/40Ca values of the rudist shell range from -0.73 to -1.84‰ relative to standard fluorite.
The δ44/40Ca seawater composition has been shown to be homogeneous in modern oceans, as
would be expected from the long Ca residence time of more than 1 Ma in seawater (Zhu and
MacDougall, 1998; De La Rocha and DePaolo, 2000; Schmitt et al., 2001; Hippler et al.,
2003). Therefore, published carbonate data sets can be compared directly to each other, when
referred to seawater (sw). The respective rudist data are δ44/40Ca [‰ sw] = -1.17 to -2.28.
68
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
Published δ44/40Ca values from mollusk calcium carbonate are still scarce. Nevertheless,
Russell et al. (1978) reported a δ44/40Ca value of -1.3‰ from a Strombus shell. Similar values
of δ44/40Ca [‰ sw] = -1.35 were measured when sampling the shallow marine tropical
gastropod Conus puncticulatus and a shell of the deep-sea limpet Bathyacnaea sp. δ44/40Ca [‰
sw] = -1.32 (data from Skulan et al., 1997; all values recalculated relative to standard
seawater).
While these values are within the range of the rudist data, the rudist δ44/40Ca average is
significantly lower (δ44/40Ca [‰ sw] = -1.74‰). This could be due to different biological
fractionation of 44Ca in gastropods and bivalves, or biological differences in Ca isotope
calcification may become evident on the genus-level. Such genus level (and even species
level) differences are known for instance from planktic foraminifera (Nägler et al., 2000;
Gussone et al., 2003; Hippler et al., 2002). An alternative explication is that the δ44/40Ca of
Cretaceous seawater was different from that of modern seawater. According to De La Rocha
and DePaolo (2000), however, at 80 Ma seawater δ44/40Ca has been similar to, or even up to
0.5‰ higher than, modern seawater. Taken at face value, this would render seawater δ44/40Ca
changes as the reason for the lower average of the rudist shell as rather improbable. In
summary, the present state of knowledge of the biogeochemistry of Ca isotope fractionation is
insufficient. Therefore, an unambiguous interpretation of δ44/40Ca values in terms of
palaeotemperatures is not possible.
For relative variations in δ44/40Ca within a bivalve shell, however, changes of ocean water
composition are not an issue; δ44/40Ca of ocean water must have been stable for the average
rudist lifetime (a few years) due to the long residence time of calcium in ocean water.
Furthermore, the Ca concentration of (average) river water is on average more than an order
of magnitude lower than that of seawater, suppressing freshwater influence on δ44/40Ca in
epicontinental seas. Accordingly, near-shore surface-water samples from the Pacific
(California; De La Rocha and DePaolo, 2000) and the Atlantic coasts (Portugal; Hippler et al.,
2003) are identical to average ocean water in terms of their calcium isotopic composition.
This implies that the seawater δ44/40Ca of the epeiric Cretaceous seas that were inhabited by
rudist bivalves should have had a value identical to that of open oceanic water masses.
4. Discussion
4.1 18O/16O and 12C/13C ratios
The oxygen-isotope values obtained from the rudist calcite (-3.2 to -5.0‰) are low when
compared to those from subtropical open marine Campanian planktonic foraminifers (-1 to 2.5‰; Norris et al. 2001), but more similar to the range of about -3.5 to -4.1‰ indicated by
Middle (Wilson and Norris, 2001; Norris et al., 2002) and Late Cretaceous (Pearson et al.,
2001) subtropical and tropical planktonic foraminifera. Open oceanic δ18O values of -4.5‰
were suggested for the Middle Cretaceous thermal maximum (Wilson et al., 2002).
69
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
Depending on the temperature equation and the assumed seawater δ18O used, the resulting
inferred rudist SST max of either 34.7ºC (or, less likely, of 37.3ºC for a seawater δ18O of -0.5‰
SMOW; temperature equation of Anderson and Arthur, 1983), are high relative to present-day
warm settings such as the Red Sea (SSTmax ≈30ºC). Recent compilations of Cretaceous SST
propose maximum values of between 28-36ºC for tropical open marine settings (Wilson and
Opdyke, 1999; Wilson and Norris, 2001; Pearson et al., 2001). These temperature maxima are
also reflected by high SST values of 32-36ºC based on Aptian to Turonian archeal membrane
lipids (Schouten et al., 2003). Given the fact that all optical and chemical screening methods
suggest that diagenetic alteration was not an issue in the case of the V. ultimus specimen used
here, the low rudist δ18O values merit discussion.
In this context, it is perhaps of interest that V. ultimus coexisted with Early Campanian
scleractinian corals (Steuber et al., 1998), as many modern corals have an upper seawater
temperature tolerance limit of about ≈30ºC. Nevertheless, some modern species seem to
tolerate transient temperature maxima in excess of 30ºC as documented by coral reefs in the
Gulf of Papua where temperatures above 30ºC are reached for many weeks each year
(personal communication, B. Opdyke, 2003). Moreover, it has been suggested (Wood, 1999
and references therein) that the tolerance limit of Cretaceous corals in terms of seawater
temperature and turbidity was higher compared to most recent species. Hence the comparative
study of modern and fossil corals is perhaps useful to some degree but remains largely
speculative.
The temperature estimates based on the δ18O record V. ultimus might thus point to high
coastal seawater temperatures, that are comparable to those deduced from the TEX86 proxy
(Schouten et al., 2003), but 2-5°C higher (when converting to the same seawater δ18O) than
those reconstructed from open oceanic planktonic foraminifera (Wilson and Opdyke, 1999;
Wilson and Norris, 2001; Pearson et al., 2001; Norris et al., 2002; Wilson et al., 2002). In this
context, it is important to recall that the rudist geochemical record reflects δ18O variations in a
much higher temporal resolution than one to three foraminifera tests lumped into one bulk
data point. The foraminifera data thus reflect average annual seawater δ18O and hence average
annual seawater temperatures. In contrast, the rudist δ18O maxima of -5.0‰ represent peak
temperatures during warmest summer weeks. Mean annual δ18O values of V. ultimus are -4.1
(section I, Figure 5) and not -5.0‰ and thus overlap with the range of other data sets.
Similar to recent scleractinian corals, rudists generally lived in a very shallow to shallow
bathymetric setting and thus presumably in most cases above an annual (summer)
thermocline. In modern shallow epeiric seas the annual (summer) thermocline is typically at
depths around 35 meter (see examples in Immenhauser and Scott, 2002), but these depths are
not necessarily applicable to Cretaceous shallow marine seas. Nevertheless, SST estimates
from the rudist low-Mg shell proxy should perhaps not be compared to those from for
instance planktonic foraminifera that typically live in at depths between 0 and 100m in open
oceanic settings and thus potentially in a wider vertical temperature range. In fact,
70
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
temperature differences between modern open-ocean SST records of planktonic foraminifera
and shallow marine coral records typically are in the order of a few degrees Celsius and coral
δ18O records from the same locality differ in the order of ≈5.0‰ when comparing corals from
the wave-swept reef margin and such from the shallow, proximal (and warmer) lagoon
(Mayotte Island; Negendank et al., 2003). When dealing with fossil planktonic foraminifera
for SST reconstructions (e.g. from core material), it is often assumed that the species with the
lowest δ18O values represent the warm surface waters (perhaps 0-20m); whereas other
species, with higher δ18O values, expectedly lived in deeper water closer to the thermocline or
even beneath it. Obviously, this raises the question of depth estimates of Cretaceous
thermoclines and the error bar related to this. But does this suggest that Cenomanian shallow
coastal settings reached summer seawater temperature maxima of nearly 35°C?
Perhaps not. This as several constraining factors limit the interpretation of absolute SST
values as deduced from the δ18O proxy. In particular, coastal seawater has a primary
variability in terms of temperature, salinity and geochemistry that is significantly greater than
that of open oceanic settings (see discussion in e.g., Patterson and Walter, 1994 or
Immenhauser et al., 2002). High (low) volumes of summer fresh-water runoff, for instance,
might have resulted in a δ18O value of the seawater-fresh water ratio that was lower (higher)
during the humid (arid) summer and higher (lower) during the dry (humid) winter. This
relation can lead to an overestimation (underestimation) of peak temperatures and temperature
seasonality (Steuber, 1999). This was illustrated by Klein et al. (1996a), investigating a recent
specimen of Mytilus trossulus from a tidally mixed inlet in British Columbia, Canada.
Assuming a normal and constant seawater salinity of 35‰ the calculated seawater
temperature based on the Mytilus δ18O overestimated in situ measured seawater temperatures
by ≈10 to 18°C because the seawater salinity at this locality in fact fluctuated between 23.5
and 29.0‰ (Klein et al., 1996). Poor estimates of coastal seawater salinity and potential
seasonal changes in meteoric influx or evaporation, however, are not the only problem that
affects the reliability of the δ18O-temperature proxy.
Other factors that may lead to an over- or underestimation of palaeo-temperatures are vital
effects leading to disequilibrium calcification (McConnaughey, 1989) and variable seawater
pH and CO32- concentrations. Zeebe (2001) argued, that poor estimates of Cretaceous
seawater pH and CO32- concentrations might lead to an underestimation of ocean surface
temperatures by about 2 to 3.5°C. In view of the high seawater temperatures indicated by the
rudist δ18O, however, seawater pH and CO32- concentrations could not have been dominant
factors.
The quantification of vital effects of a now extinct taxon, such as rudists, is difficult and
assumptions based on the study of modern bivalve mollusks might be biased by speciesdependent factors that are difficult to quantify. Nevertheless, while vital effects have been
documented for a single rudist genus, the species investigated here belongs to group of Late
Cretaceous genera, which show a very similar cyclity of δ18O values, and for which vital
71
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
effects have been considered unlikely (Steuber, 1999). Particularly, the comparison of the V.
ultimus δ18O record with that of other coeval, pristine rudist shells from the Pontid Mountains
in Turkey (Yvaniella alpani; Vacinites oppeli) shows overlapping oxygen-isotope ranges with
typical minima of -5‰ (Figure 6). This observation is clearly not in agreement with a species
or even specimen-specific vital control on the δ18O record. Furthermore, Wefer and Berger
(1991) demonstrated that even in the case of the few marine invertebrates that show a
metabolic influence on their shell δ18O record, annual temperature variations are still reliably
recorded although a certain offset in δ18O occurs (either to higher or lower values).
Figure 6: Comparison of range of δ18O
maxima and minima of V. ultimus (this
study, ✰) with other Campanian
rudists from Turkey. Dots, diamonds
and squares respectively indicate mean
maxima and minima and the relative
range per specimen (horizontal bars).
The δ18O maxima and minima of V.
ultimus plot well into the range of
other screened low-Mg calcite rudist
shells indicating that specimen-specific
metabolic processes did not affect its
oxygen-isotopic composition.
In summary, it is concluded that the pronounced cyclical variability in rudist δ18O is largely
the expression of fluctuating seawater temperature above the shallow carbonate platform.
This is clearly supported by the very similar data obtained from other coeval, pristine rudist
specimen from this region (Steuber, 1999). Diagenetic alteration is rejected here as an
explanation of the low values of -5‰. Hence, a number of alternative factors should be
considered that might potentially lead to an overestimation of these peak temperatures. Most
prominent are seasonal variations in meteoric water runoff and evaporation that have a direct
impact on the salinity and oxygen isotopic composition of the ambient seawater.
With reference to the δ13C values of V. ultimus, the anti-correlation with δ18O merits attention.
Similar cyclical δ13C records, either covarying or not with δ18O are observed in many fossil
and modern marine and freshwater bivalves (e.g., Steuber 1999; Vander Putten et al., 2000;
Kaadorp et al., 2003; Ivany et al., 2004 and references therein). The controlling factors of this
δ13C variability are often poorly understood due to an incomplete knowledge of
biomineralisation processes during shell secretion (McConnaughey and Whelan, 1997 and
references therein). Two main lines of interpretation are commonly presented. One argues
that the variant δ13C record is an expression of bivalve-specific metabolic activity (respiration
rate) such as in the case of Mytilus trossulus (e.g., Klein et al., 1996b; Vander Putten et al.,
72
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
2000). In contrast, the other line of reasoning explains the shell δ13C variability with
seasonally fluctuating of the dissolve inorganic bicarbonate pool of the ambient seawater
(e.g., Kaadorp et al., 2003) respectively as a function of seasonal upwelling pattern (e.g.
Killingly and Berger, 1979). Clearly, an adequate discussion of this issue is beyond the scope
of this paper and a more detailed interpretation of the underlying mechanisms must be the
topic of further work.
4.2 Mg/Ca molar ratios
Magnesium and calcium have residence times of several million years in the oceans. Annual
to decadal shifts in seawater Mg/Ca ratios are thus negligible. Furthermore, the Mg/Ca ratio
of seawater is rather insensitive to freshwater dilution. Mg/Ca ratios of foraminifera calcite
are thus regarded as a robust and reliable SST-proxy for the Quaternary (Elderfield and
Ganssen, 2000). In the case of the Campanian Mg/Ca molar ratios of rudist low-Mg calcite,
however, a number of problems are evident.
According to Hardie (1996), Lowenstein et al. (2001), and Horita et al. (2002), the Mg/Ca
molar ratio in the Late Cretaceous seawater was about 1, with Ca2+ concentrations that were
much higher (≈30 mmol/kg compared to about ≈10 mmol/kg H2O of modern seawater) and
Mg2+ concentrations that were much lower than in modern oceans. This high Ca2+
concentration of Late Cretaceous seawater, however, is debated (Steuber and Veizer, 2002).
The Mg/Ca ratios obtained from the Campanian rudist shell (5.36 to 12.89) are higher than
those reported from the recent shell of M. trossulus (3.55 to 10.78; Klein et al., 1996a). The
linear extrapolation of the bivalve data set of Klein et al. (1996a) to a temperature range of
≈25 to ≈35°C, as indicated by the rudist δ18O, suggests a theoretical Mg/Ca ratio of between
9.5 to 13. The measured range of the rudist Mg/Ca ratios (5.36 to 12.89), however, extends to
much lower values than 9.5, a fact that might be related to the lower Mg/Ca ratio of Late
Cretaceous seawater. In any case, a linear relation between the Mg/Ca molar ratio of ≈1,
suggested for the Late Cretaceous global seawater (Hardie, 1996) and the rudist
geochemistry, cannot explain the measured data. Also, the Mg-δ18O covariance of Cretaceous
rudist shells reflects the changes of the Cretaceous seawater Mg/Ca ratio as predicted by
Hardie (1996), but show that the Mg/Ca ratio of rudist calcite is not a precise proxy for
seawater temperature during the Late Cretaceous (Steuber, 2002). This implies that factors
other than the geochemistry and temperature of the ambient Cretaceous seawater have
influenced the Mg/Ca molar ratio of the rudist shell (Vander Putten et al., 2000). These
factors require consideration despite the fact that synchronous shifts of the rudist Mg/Ca
molar ratios with δ18O might indicate a relation between seawater temperature and shell
geochemistry.
Klein et al. (1996a) argued that the skeletal Mg content of mollusk bivalves is controlled in a
predictable way by the metabolic activity of the animal secreting the shells. This metabolic
activity of the bivalve, in turn, is influenced by the temperature of the ambient seawater.
73
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
Experiments with the cultured bivalve M. trossulus (Rosenberg and Hughes, 1991), however,
have shown that the utility of Mg/Ca ratios in mollusk shells appears to be limited by the ion
regulating capability of these animals. The degree of this biologic ion regulation depends on
the Mg/Ca molar ratio of the ambient Late Cretaceous seawater geochemistry (Wilkinson and
Algeo, 1989; Lowenstein et al., 2001; Steuber, 2002). In addition, taxon-specific ontogenetic
patterns of shell geochemistry might further bias the resulting data sets (Rosenberg and
Hughes, 1991). We thus consider absolute SST based on the rudist Mg/Ca molar ratios, as
well as the extreme temperature fluctuations of nearly 19°C that result from theoretical
assumptions, ambiguous because these values are controlled by several factors that are
difficult to constrain.
4.3 44Ca/40Ca ratios
As a result of uncertainties of the δ44/40Ca composition of Cretaceous seawater and the species
dependence of Ca isotope fractionation, absolute palaeo-temperatures cannot be derived from
the rudist δ44/40Ca data shown in Figure 3. A semi-quantitative evaluation, however, can be
modelled from existing calibrations and theoretical considerations. Different, temperature
dependent, Ca-isotope fractionations have been reported from foraminifera (Nägler et al.,
2000; Hippler et al., 2002; Gussone et al., 2003) and inorganic precipitates (Gussone et al.,
2003). Basically, they can be subdivided in steep gradients of 0.20-0.24‰ δ44/40Ca per 1ºC
temperature change in the case of the foraminifers Globigerinoides sacculifer and
Neogloboquadrina pachyderma (sin.) and gradients more than an order of magnitude lower (<
0.02‰ δ44/40Ca per 1ºC temperature change in the case of Orbulina universa and inorganic
precipitates). Applying the low temperature gradient of e.g., the foraminifera O. universa to
the rudist V. ultimus would result in a temperature seasonality in excess of 60ºC, an
unrealistic assumption that is not supported by other data.
More reasonable assumption of temperature seasonality result from applying the steep
gradient of 0.20-0.24‰ δ44/40Ca per 1ºC temperature change such as know from G. sacculifer
to the rudist data. Differences between δ44/40Camax at sample points 6 and 8 and δ44/40Camin at 3
and 7 of the V. ultimus shell (Figure 5) translate into relative temperature variations of about
5ºC. In the upper part (samples 10-12) seasonal temperature variations would be close to 2ºC.
These values could be established as minimum temperature seasonality, provided even
steeper gradients could be ruled out on a more general basis. The model for Ca isotope
fractionation in inorganic aragonite and cultured planktonic foraminifera of Gussone et al.
(2003) relates these gradients to the mass of the calcium ion or molecule that is dominant
during the fractionation process. Low slopes are best explained as reflecting transport of
higher molecular masses of the calcium molecules (Ca(HCO3)2 (162 atomic mass units (amu))
or the Ca2+-aquocomplex Ca*32H2O (617 amu; Langmuir, 1997). Steep gradients (0.200.24‰ δ44/40Ca per 1ºC), indicate fractionation of the pure ionic form of calcium (40 amu).
These considerations have two important implications for the present study: Firstly, steeper
temperature gradients than the one reflected by G. sacculifer (Nägler et al., 2000) are not
74
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
expected, as the pure ion is the smallest possible mass of calcium. Secondly, the minimum
seawater temperature seasonality reflected in the V. ultimus data can therefore be defined
using a gradient of 0.20-0.24‰ δ44/40Ca per 1ºC. This is despite the fact that the biocalcification of an extinct mollusk such as V. ultimus is poorly understood. Seasonal changes
in the epicontinental sea surface water masses of the Late Cretaceous would thus have at least
reached amplitudes of 5ºC. This amplitude, however, is most likely a minimum estimate,
biased by the low sampling resolution for δ44/40Ca analyses when compared to that for the
oxygen-isotope sections II and III (Figure 5).
Figure 7 a) Diagram of δ18O versus δ44Ca.
The good correlation of the data precludes
significant post-formational alteration. Filled
squares: δ18O = Averages of subsets of δ18O
data from Figure 5 representing the locations
of the respective δ44Ca samples. Open
circles: δ18O from dry aliquots of samples
analysed for δ44Ca. Both δ18O sample sets
give similar values with the exception of
four (arrows connect δ18O from both
procedures). The two samples plotting the
most off the correlation line (6, 6b) are from
the steep gradient between temperature
maxima 6 and minima 7 as indicated in
Figure 5. Thus, offsets most probably
represent sampling biases. b) and c)
Diagrams of δ13C versus δ44Ca and Mg/Ca
versus δ44Ca. Correlation is fair in both
cases. Error ellipses are at the 2σ-level.
75
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
Another factor that requires discussion is the potential of diagenetic alteration of the rudist
shell. The SEM analysis of shell structures (Figure 3) as well as concentrations of Sr, Mn and
Fe in shells from the sampled locality indicate excellent preservation of original chemical and
isotopic compositions (Figures 4, 5 and Table 1; Steuber, 1996; 1999; Steuber et al., 1998).
Significant post-formational alteration of the Ca isotope composition of the rudist shell are
not expected as calcium is the major component of the rudist skeleton, and the major source
of calcium in diagenetic pore fluids is the carbonate shell itself. In contrast, the main source
of oxygen during stabilisation of marine carbonates is diagenetic fluids. This makes oxygen
much more sensitive to diagenetic resetting than calcium. In spite of these different
sensitivities to diagenesis, δ44/40Ca and δ18O values are correlated (Figure 7a). This
observation is at odds with a significant diagenetic overprint and supports the assumption of
primary δ44/40Ca values. Therefore, the δ44/40Ca variations within the rudist shell investigated
(Figure 5) most probably reflect short-term environmental changes within its life habitat.
More specifically, the inverse correlation of δ18O and δ44/40Ca (Figures 5, 7) indicates a
temperature-dependent fractionation process (Gussone et al., 2003).
5. Conclusions
Co-variant shifts in the rudist geochemistry reflect predominantly temperature seasonality of
ambient seawater as opposed to variable seawater pH and CO32- concentrations, diagenetic
alteration of the shell, or metabolic processes. With reference to the δ18O temperature proxy,
however, changes in freshwater influx and evaporation can lead to an over- or
underestimation of temperature maxima and minima, particularly in shallow coastal water
masses. Judging from geochemical evidence, amplitudes of temperature shifts range from at
least 5°C, as theoretically inferred from the δ44/40Ca proxy, to 6.6°C as suggested by the δ18O
proxy.
In terms of absolute seawater temperatures, the interpretation of Mg/Ca molar ratios and
44
Ca/40Ca ratios is difficult due to the poorly constrained ion regulation and the complex biocalcification of the rudist bivalve. Therefore, an unambiguous interpretation based on the
44
Ca/40Ca proxy in terms of absolute palaeotemperatures is not possible. Furthermore, the
Cretaceous seawater geochemistry differed significantly from modern oceans so that the
extrapolation of experimental data based on modern bivalves to the Campanian rudist is
ambiguous.
Comparing the rudist δ44/40Ca data with the other two proxies suggests that the calcium
isotope fractionation within the mollusk shell be mainly controlled by the temperature of the
ambient seawater. This is an important result and, using a fossil mollusk shell, demonstrated
here for the first time.
The multiproxy approach applied here clearly and successfully demonstrates a dominant
pattern of seasonal sea surface temperature variations in a quantitative manner. At the same
76
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
time, these data also raise critical and obvious questions concerning the significance of singleproxy seawater temperature reconstructions, particularly when applied in a qualitative
manner.
Acknowledgments
We acknowledge helpful comments by G. Davies, J. Erbacher, G. Ganssen, S. Jung, and J. D.
Kramers. B. Opdyke and two anonymous Palaeo3 reviewers are thanked for reviews. The
editorial advice of P. DeDecker is greatly appreciated and S. Troelstra is thanked for
negotiating in the background. Ca-isotope work in Bern is supported by Swiss National
Science Foundation grant 21-61644 to Th. F. Nägler and D. Hippler. T. Steuber
acknowledges DFG grants Ste 670/8 and Ste 670/9.
References
Al-Aasm, I.S., Veizer, J., 1982.
Chemical stabilization of low-Mg calcite: An example of brachiopods. J. Sed. Pet. 52, 11011109.
Alp, D. 1972.
Amasya Yöresinin Jeolojisi. Isanbul Universitesi, Fen Fakultesi Monografileri, 22, 100 p.
Anderson, T.F., Arthur, M.A., 1983.
Stable isotopes of oxygen and carbon and their application to sedimentologic and
palaeoenvironmental problems. SEPM Short Course 10, 151 p.
Banner, J.L., Hanson, G.H., 1990.
Calculation of simultaneous isotopic and trace element variations during water-rock
interaction with applications to carbonate diagenesis: Geochim. Cosmochim. Acta 54, 31233137.
Barbin, V., Ramseyer, K., Debenay, J.P., Schein, E., Roux, M., Decrouez, D., 1991.
Cathodoluminescence of Recent biogenic carbonates: an environmental and ontogenetic
fingerprint. Geol. Mag., 128, 19-26.
Barron, E.J., 1995. Tropical climate stability and the implications for the distribution of life.
In: S.M. Stanley (Ed.), Effects of past global change on life. National Academy Press,
Washington, pp. 108-117.
Brand, U., Veizer, J., 1980.
Chemical diagenesis of a multicomponent carbonate system - 1: Trace elements: J. Sed. Res.
50, 1219-1236.
77
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
Carpenter, S.J., Lohmann, K.C., 1992.
Sr/Mg ratios of modern marine calcite: empirical indicators of ocean geochemistry and
precipitation rate. Geochim. Cosmochim. Acta 56, 3749-3764.
Crowley, T.J., 1991.
Past CO2 changes and tropical sea surface temperatures. Paleoceanography 6, 387-394.
De La Rocha, C., DePaolo, D.J., 2000.
Isotopic evidence for variations in the marine calcium cycle over the Cenozoic. Science 289,
1176-1178.
Elderfield, H., Ganssen, G., 2000.
Past temperature and δ18O of surface waters inferred from foraminiferal Mg/Ca ratios. Nature
405, 442-445.
Erez, J., Luz, B., 1983. Experimental Paleotemperature Equation for Planktonic-Foraminifera.
Geochim. Cosmochim. Acta 47, 1025-1031.
Gagan, M.K. Ayliffe, L.K., Beck, J.W., Cole, J.E., Druffel, E.R.M., Dunbar, R.B., Schrag,
D.P. 2000.
New views of tropical paleoclimates from corals. Quat. Sci. Rev. 19, 45-64.
Gili, E., Masse, J.-P., Skelton, P.W., 1995.
Rudists as gregarious sediment-dwellers, not reef-builders, on Cretaceous carbonate
platforms. Palaeogeogr. Palaeoclimatol. Palaeoecol. 118, 245-267.
Gussone, N., Eisenhauer, A., Heuser, A., Dietzel, M., Bock, B., Boehm, F., Spero, H., Lea,
D.W., Bijma, J., Nägler, T.F., 2003.
Model for kinetic effects on calcium isotope fractionation (δ44Ca) in inorganic aragonite and
cultured planktonic foraminifera: Geochim. Cosmochim. Acta 67, 1375-1382.
Hardie, L.A., 1996.
Secular variation in sea water chemistry: An explanation for the coupled secular variation in
the mineralogies of marine limestones and potash evaporites over the past 600 m.y. Geology
24, 279-283.
Hippler, D., Gussone, N., Darling, K., Eisenhauer, A., Nägler, T.F., 2002.
δ44Ca in N. pachyderma (left): A promising tool for SST-reconstructions in high-latitude
oceans. Eos Trans. AGU, 83(47), Fall Meet. Suppl., Abstract PP51A-0305.
Hippler, D., Schmitt, A.-D., Gussone, N., Heuser, A., Stille, P., Eisenhauer, A., Nägler, Th.F.,
2003.
Ca isotopic composition of various standards and seawater. Geost. Newslett. 27, 13-19.
78
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
Horita, J., Zimmermann, H., Holland, H.D., 2002.
Chemical evolution of seawater during the Phanerozoic: Implications from the record of
marine evaporites. Geochim. Cosmochim. Acta. 66, 3733-3756.
Immenhauser, A., Scott, R.W., 2002.
An estimate of Albian sea-level amplitudes and its implications for the duration of
stratigraphic hiatuses. Sed. Geol. 152, 19-28.
Immenhauser, A., Kenter, J. A. M., Ganssen, G., Bahamonde, J. R., van Vliet, A., Saher, M.
H., 2002.
Origin and significance of isotope shifts in Pennsylvanian carbonates (Asturias, NW Spain). J.
Sed. Res. 72, 82-94.
Ivany, L. C., Wilkinson, B. H., Lohmann, K. C., Johnson, E. R., McElroy, B. J., Cohen, G.J.,
2004.
Intra-annual isotopic variations in Venericardia bivalves: Implications for Early Eocene
temperature, seasonality, and salinity on the U.S. Gulf Coast. J. Sed. Res. 74, 7-19.
Kaadorp, R. J. G., Vonhof, H. B., Del Busto, C., Wesseligh, F. P., Ganssen, G. M., Marmol,
A. E., Pittman, L. R., van Hinte, J. E., 2003.
Seasonal stable isotope variations of the modern Amazonian freshwater bivalve Anodontites
trapesialis: Palaeogeogr. Palaeoclimatol. Palaeoecol. 194, 339-354.
Killingly, J.S., Berger, W.H., 1979.
Stable isotopes in a mollusk shell: detection of upwelling events. Science 205, 186-188.
Klein, R.T., Lohmann, K.C., Thayer, C.W., 1996a.
Bivalve skeletons record sea-surface temperature and δ18O via Mg/Ca and
Geology 24, 415-418.
18
O/16O ratios.
Klein, R.T., Lohmann, K.C., Thayer, C.W., 1996b.
Sr/Ca and 13C/12C ratios in skeletal calcite of Mytilus trossulus: covariation with metabolic
rate, salinity, and carbon isotopic composition of seawater. Geochim. Cosmochim Acta. 60,
4207-4221.
Langmuir, D., 1997.
Aqueous environmental geochemistry. Prentice Hall.
Lohman, K.C., 1988.
Geochemical Patterns of Meteoric Diagenetic Systems and Their Application to Studies of
Paleokarst, in James, N.P., Choquette, P.W. (Eds.), Paleokarst. Berlin, Springer, p. 58-80.
79
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
Lowenstein, T.K., Timofeef, M.N., Brennan, S.T., Hardie, L.A., Demicco, R.V., 2001.
Oscillations in Phanerozoic Seawater Chemistry: Evidence from Fluid Inclusions. Science
294, 1086-1088.
McConnaughey, T., 1989.
13
C and 18O isotopic disequilibrium in biological carbonates: II. In vitro simulation of kinetic
isotope effects. Geochim. Cosmochim Acta. 53, 163-171.
McConnaughey, T. A., Whelan, J. F., 1997.
Calcification generates protons for nutrient and bicarbonate uptake: Earth-Sci. Rev. 42, 97117.
Nägler, T.F., Eisenhauer, A., Müller, A., Hemleben, C., Kramers, J., 2000.
δ44Ca-temperature calibration on fossil and cultured Globigerinoides sacculifer: New tool for
reconstruction of past sea surface temperatures. Geochemistry, Geophysics, Geosystems, 1.
Negendank, J.F.W., Schwab, M.J., KIHZ-Consortium, 2003.
Natürliche Klimavariationen in historischen Zeiten bis 10.000 Jahre vor heute.
Abschlussbericht 2000-2003, GFZ-Potsdam, 280pp.
Norris, R.D., Kroon, D., Huber, B.T., Erbacher, J., 2001.
Cretaceous-Paleogene ocean and climate change in the subtropical North Atlantic. In: D.
Kroon, R.D., Norris, Klaus, A. (Eds.), Western North Atlantic Paleogene and Cretaceous
Paleoceanography. Geological Society, London, Spec. Publ. 183, 1-22.
Norris, R.D., Bice, K.L., Mango, E.A., Wilson, P.A., 2002.
Jiggling the tropical thermostat in the Cretaceous hothouse. Geology 30, p. 299-302.
Patterson, W.P. Walter, L.M., 1994.
Depletion of 13C in seawater pCO 2 on modern carbonate platforms: Significance for the
carbon isotopic record of carbonates. Geology 22, 885-888.
Pearson, P.N., Ditchfield, P.W., Singano, J., Hracourt-Brown, K.G., Nicholas, C.J., Olsson,
R.K., Shackleton, N.J., Hall, M.A. 2001.
Warm tropical sea surface temperatures in the Late Cretaceous and Eocene epochs. Nature
413, 481-487.
Rosenberg, G.D., Hughes, W.W., 1991.
A metabolic model for the determination of shell composition in the bivalve mollusk, Mytilus
edulis. Lethaia 24, 83-96.
Russell, W.A., Papanastassiou, D.A., Tombrello, T.A., 1978.
Ca isotope fractionation on the Earth and other solar system materials. Geochim. Cosmochim.
Acta 42, 1075-1090.
80
Chapter 5: PhD-thesis Dorothee Hippler, Bern (2004)
Schmitt, A.-D., Bracke, G., Stille, P., Kiefel, B., 2001.
The Ca isotope composition of modern seawater determined by thermal ionization mass
spectrometry. Geostandards Newsletter 25, 267-275.
Schouten, S., Hopmans, E.C., Forster, A., v. Breugel, Y., Kuypers, M.M.M., Sinninghe
Damsté, J.S., 2003.
Extremely high sea-surface temperatures at low latitudes during the middle Cretaceous as
revealed by archeal membrane lipids. Geology 31, p. 1069-1071.
Skulan, J., DePaolo, D.J., Owens, T.L., 1997.
Biological control of calcium isotopic abundances in the global calcium cycle. Geochim.
Cosmochim. Acta 61, 2505-2510.
Steuber, T., Yilmaz, C., Löser, H., 1998.
Growth rates of Early Campanian rudists in a siliclastic-calcareous setting (Pontid Mts.,
North-Central Turkey). Geobios Mem. Spec., 22, 385-401.
Steuber, T., 1996.
Stable isotope sclerochronology of rudist bivalves: Growth rates and Late Cretaceous
seasonality. Geology 24, 315-318.
Steuber, T., 1999.
Isotopic and chemical intra-shell variations in low-Mg calcite of rudist bivalves (MolluscaHippuritacea): disequilibrium fractionations and Late Cretaceous seasonality. Int. J. Earth Sci.
88, 551-570.
Steuber, T., 2002.
Plate tectonic control on the evolution of Cretaceous platform-carbonate production. Geology
30, 259-262.
Steuber, T., Veizer, J., 2002.
Phanerozoic record of plate-tectonic control of seawater chemistry and carbonate
sedimentation. Geology 30, 1123-1126.
Steuber, T., 2003.
Strontium isotope stratigraphy of Cretaceous hippuritid rudist bivalves: rates of
morphological change and heterochronic evolution. Palaeogeogr. Palaeoclimatol. Palaeoecol.
200, 221-243.
Vander Putten, E., Dehairs, F., Keppens, E., Bayens, W., 2000.
High-resolution distribution of trace elements in the calcite shell layer of modern Mytilus
edulis: Environmental and biological controls. Geochim. Cosmochim. Acta 64, 997-1011.
81
Mollusk 18O/16O, Mg/Ca and 44Ca/40Ca as proxies for Cretaceous STS
Wefer, G., Berger, W.H., 1991.
Isotope Paleontology: Growth and composition of extant calcareous species. Mar. Geol. 100,
207-248.
Wefer, G., Berger, W.H., Bijma, J., Fischer, G., 1999.
Clues to Ocean History; a Brief Overview of Proxies. In: G. Fischer, Wever, G. (Eds.), Use of
Proxies in Paleoceanography: Examples from the South Atlantic. Springer, Berlin,
Heidelberg, pp. 1-68.
Wilkinson, B.H., Algeo, T.J. 1989.
Sedimentary carbonate record of calcium-magnesium cycling. Am. J. Sci. 289, 1158-1194.
Wilkinson, B.H., Ivany, L.C., 2002.
Paleoclimatic inference from stable isotope profiles of accretionary biogenic hardparts - a
quantitative approach to the evaluation of incomplete data. Palaeogeogr. Palaeoclimatol.
Palaeoecol. 185, 95-114.
Wilson, P.A., Opdyke, B.N. 1996.
Equatorial sea surface temperatures for the Maastrichtian revealed through remarkable
preservation of metastable carbonate. Geology, 24, 555-558.
Wilson, P.A., Norris, R.D., 2001.
Warm tropical ocean surface and global anoxia during the mid-Cretaceous period. Nature
412, 425-428.
Wilson, P.A., Norris, R.D., Cooper, M.J., 2002.
Testing the Cretaceous greenhouse hypothesis using glassy foraminiferal calcite from the core
of the Turonian tropics on Demerara Rise. Geology 30, 607-610.
Wood, R., 1999.
Reef Evolution: New York, Oxford University Press, 414 p.
Zachos, J.C., Stott, L.D., Lohmann, K.C., 1994.
Evolution of early Cenozoic marine temperatures. Paleoceanography 9, 353-387.
Zeebe, R.E., 2001.
Seawater pH and isotopic paleotemperatures of Cretaceous oceans. Palaeogeogr.
Palaeoclimatol. Palaeoecol. 170, 49-57.
Zhu, P., MacDougall, J.D., 1998.
Calcium isotopes in the marine environment and the oceanic calcium cycle. Geochim.
Cosmochim. Acta 62, 1691-1698.
82
Appendix A.I: PhD-thesis Dorothee Hippler, Bern (2004)
A.I Analytical protocol for the determination of Ca isotopic
composition in natural samples
1. Mass spectrometric calibration of a 43Ca-48Ca double spike
1.1 Analytical considerations
Generally, the determination of natural isotope fractionation of an element requires the
possibility to distinguish between natural fractionation and fractionation occurring during
chemical purification and mass spectrometric analysis. It has been shown that Ca fractionate
to a large extent during mass spectrometric analysis and chemical purification (e.g. Heumann
1972, Heumann & Lieser 1972, Heumann et al. 1970, 1982, Russell et al. 1978). In order to
measure the original Ca isotopic composition and achieve an enhanced reproducibility using
thermal ionisation mass spectrometry we apply the Ca double spike technique, originally
introduced for Ca isotope analysis by Russell et al. (1978).
1.2 The “double-spike cocktail”
Double spiking has been used as a tool to monitor isotopic fractionation during mass
spectrometry for a number of elements when a fixed normalising ratio cannot be assumed
(Eugster 1969, Compston & Oversby 1969, Russell et al. 1978). We have chosen a double
spike consisting predominantly of 43Ca and 48Ca for the following reasons: (1) the relative low
natural abundance of 43Ca and 48Ca (see above). (2) The large mass difference of the tracer
isotopes would deamplify uncertainties in the instrumental mass fractionation per atomic
mass unit during analysis of a spiked sample. The double spike was prepared at GEOMAR
(Heuser et al. 2002, Gussone N. 2003) and calibrated in Bern.
Table 1: Ca spike isotopic composition
Isotope ratio
Value
40
Ca/44Ca
2.7211
43
44
Ca/ Ca
16.3866
48
44
21.8632
Ca/ Ca
1.3 Double spike calibration
Since large Ca isotope fractionation occurs during mass spectrometric analyses the double
spike cannot be directly determined and has to be defined relative to a Ca standard. Natural
flourite (CaF 2, cf. Russell et al. 1978) was employed as Ca standard and was supposed to be
free of technical fractionation (Nägler & Villa 2000, Nägler et al. 2000). Calibration of the
I
Analytical protocol for the determination of Ca isotopic composition
isotopic composition of the double spike closely followed an iterative approach, previously
published by Siebert et al. (2000) for Mo-isotope double spike correction based on routines of
Hofmann (1971) and Johnson & Beard (1999).
In this algorithm the linear fractionation terms have been replaced by terms which allow a
correction following the exponential law. The use of an exponential law was proposed by
Russell et al. (1978) to be more suitable for fractionation correction of the Ca isotope system.
Based on a three-dimensional approach the algorithm describes isotope fractionation in a
three-dimensional vector space, which is defined by three isotope ratios with the same mass
as common denominator (40Ca/44Ca, 43Ca/44Ca and 48Ca/44Ca).
Figure 1: Three-dimensional sketch of the Ca isotope composition of the spike, sample and
spike/sample mixtures. “Mixture” corresponds to the true composition of the spike-sample mixture and
has to be calculated from the “measured value” point (spike-sample mixture after instrumental
fractionation) by an iterative approach. (dashed line) Spike-sample mixing line = intersection line,
(dotted lines) fractionation lines, approximate direction of natural fractionation relative to the standard.
For calculation the calibrated Ca isotope ratios of the “spike” are used and for the “sample” ratios as
start parameters the CaF2 value of Russell et al (1978) is used as a fixed point.
It is necessary to measure the pure spike, the unspiked standard and mixtures of the spike and
standard. For data reduction an X-Y-Z coordinate system is used in which axis are assigned to
isotope abundance ratios. In the three-dimensional data reduction, straight lines and flat
II
Appendix A.I: PhD-thesis Dorothee Hippler, Bern (2004)
planes are used to determine intercepts, which yield fractionation factors. Fractionation
factors are denoted for natural fractionation relative to the standard and for instrumental
fractionation. This approach allows simple mathematics but does not correspond to the
exponential fractionation law, in which fractionation is described by a curve (see, e.g.
Johnson & Beard 1999). This problem has been solved by finding lines and flat planes by
iteration, for which intercepts correspond to the true intercepts of the fractionation curve. A
detailed description of mathematics is given in Siebert et al. (2000).
2. Experimental methods for Ca isotope analyses using TIMS
2.1 Cleaning procedure and blank levels
Ca is ubiquitous and even working in the clean hood is no absolute guarantee for blank-free
conditions. To avoid procedural blanks a careful handling of samples, laboratory materials
and solutions is important.
For the sample preparation teflon beakers are used. After use, tapes and remaining clue are
removed from the outside walls with acetone. The beakers and the caps are filled up with
distilled water and left overnight. The inside walls are then rubbed with distilled water and
Kimwipes®, before rinsing them three times with distilled water. The vials are filled half with
2 N HCl, tightly closed and placed on the heating plate at a temperature of ≈100°C for at least
2-3 days. Thereafter, the 2 N HCl is discarded and the beakers are rinsed three times and
filled with Mili-Q® water, capped and placed on the heating plate at a temperature of ≈100°C
for another two days. After discarding the water the beakers are stored. Before use, the vials
are rinsed again with two beaker volumes of 2.5 N HCl.
All < 6 N HCl solutions needed during sample preparation and column chemistry are diluted
out of doubly distilled HCl (30% HCl suprapur from Merck). Blank measurements have been
performed for safety also for HNO3, conc. H2O2, Ta2O5-activator solution, NH4-ac and the
DNA-extraction buffer solution.
2.2 Sample preparation, chromatography and sample loading
Ca isotope analyses were performed on different reference and sample material. For the
individual sample preparation see the respective chapter.
Generally, the 43Ca-48Ca double spike (see above) was added to each sample aliquot
(corresponding to between 0.4 and 1.0µg Ca) resulting in a spike to sample ratio ( 43Ca/44Ca)
of about 2 to 5. The spike-aliquot mixture was evaporated to dryness. In order to eliminate
any organic impurities, in some cases (e.g. samples of N. pachyderma (sin.)) spiked aliquots
were treated with a HNO3-H2O2 solution (Hippler et al. 2004).
For the ion exchange procedure one-way microcolumns containing 10µl AP-MP-50 (200-400
mesh) cation exchange resin and a chemical extraction procedure exclusively based on 1.5 N
HCl was used. This is necessary to remove elements such as potassium, strontium,
III
Analytical protocol for the determination of Ca isotopic composition
magnesium and sodium that might surpress ionisation of Ca during mass spectrometer run
and might act as potential isobaric interferences. First, the resin has to be cleaned as follows:
(1) 100µl Mili-Q® water (2) 100µl ammoniumacetate (NH4-ac) (3) 100µl Mili-Q® water (4)
6 N HCl, and is subsequently conditioned with three column volumes of 1.5 N HCl. Samples
are loaded onto the column. After rinsing with 210µl, the Ca is eluted with 120µl. The
procedure shows a >90% yield on standard solutions. Total procedural blanks are below 1ng
and therefore have a no effect on the isotopic data.
The sample-spike mixture is recovered in about 1µl 2.5 N HCl and than loaded on previously
washed and outgassed single Re filament (µm) in combination with 1µl of a Ta 2O5-activator
solution (Birck 1986). The activator solution stabilises the signal intensity resulting in precise
Ca isotope measurements. Finally the filament is slowly heated to until a red glow is visible.
2.3 TIMS measurement protocol
All Ca isotope measurements were carried out on a modified single cup AVCO mass
spectrometer equipped with a Thermolinear ion source. It was operated in positive ion
mode with a 7.6 kV acceleration voltage and a 1011 Ω resistor setting for the Keithley 624
LNFPA electrometer. Ca isotopes are measured successively in peak jumping mode in
descending sequence (masses 48, 44, 43, 40). During data acquisition 40K+ and 87Sr2+ were
continuously monitored on mass 41 (40K/41K = 1.7384 * 10-3) and 43.5 to trace possible
isobaric interferences. No interferences have been observed at measuring temperatures.
Table 1: Instrumental set-up for Ca isotope analysis using TIMS
Property
Flight tube vacuum
Source vacuum
vacuum
Setting
2-6 *10-9 mbar
3-5 * 10-7 mbar
3-4 * 10-3 mbar
Accelerating voltage
Amplifier resistor
7.6 kV
1011 Ω
Baseline
Each block
Delay time: 3-6s
Integration time 5s
Peak centering
Each block
Data collection
Blocks per run: ≥ 10
Scans per block: 11
Integration time: 5-6s
To start the measurement, the filament is heated stepwise to approximately 1.95A
(corresponding to a temperature of about 1500°C). After the first heating step the signal is
focussed, peak centering is performed and the peak shape is controlled. Heating is continued
up to 2.1 ± 0.1A. Data acquisition is started when the intensity of the ion beam on mass 40
reaches ≥ 800mV. In general, measurements run at ion beam intensities on mass 40 of around
IV
Appendix A.I: PhD-thesis Dorothee Hippler, Bern (2004)
2-4V. Before and after each block (summarising 11 scans) the baseline is recorded (≈ 5s
integration time). At least 10 blocks are measured. Measurement runs are normally stopped
when the uncertainty (2σ standard deviation) of the block averages δ44/40Ca value is lower
than 0.20‰.
Until today most of the published work on Ca isotopes has been obtained by using a singlecollector TIMS technique. However, other measurement techniques have also been tested:
e.g. (1) multicollector TIMS (Heuser et al. 2002), (2) hot-plasma multicollector ICP-MS
(Halicz et al. 1999), (3) ICP-MS with hexapole collision cell (Boulyga & Becker 2001), (4)
cool-plasma multicollector ICP-MS (Fietzke et al. 2003).
References
Birck J. L. (1986)
Precision K-Rb-Sr isotope analysis: Application to Rb-Sr chronology. Chemical geology, 56,
73-83.
Boulyga S. F. and Becker J. S. (2001)
ICP-MS with hexapole collision cell for isotope ratio measurements of Ca, Fe, and Se.
Fresenius J. Anal. Chem. 370, 618-623.
Compston W. and Oversby V. M. (1969)
Lead isotope analysis using a double spike. Journal of Geophysical Research, 74(17), 43384348.
Eugster O., Tera F. and Wasserburg, G. J. (1969)
Isotopic analyses of barium in meteorids and in terrestrial samples. J. Geophys.Research, 74,
3897-3908.
Fietzke J., Eisenhauer A., Gussone N., Bock B., Liebetrau V., Nägler Th. F., Spero H. J.,
Bijma J. and Dullo C. (2004)
Direct measurement of 44Ca/40Ca ratios by MC-ICP-MS using the cool plasma technique.
Chemical Geology, 206, 11-20.
Gussone N. (2003)
Ca isotope fractionation in calcium carbonates: Calibration and application. PhD-thesis,
Christian-Albrechts-Universität Kiel, Germany, 135p.
Halicz L., Galy A., Belshaw N. S. and O'Nions R. K. (1999)
High-precision measurements of calcium isotopes in carbonates and related materials by
multiple collector inductively coupled plasma mass-spectrometry (MC-ICP-MS). Journal of
Analytical Atomic Spectrometry, 14, 1835-1838.
V
Analytical protocol for the determination of Ca isotopic composition
Heumann K. G., Lieser K. H., et al. (1970)
Difficulties in measuring the isotopic abundances of calcium with a mass spectrometer.
Revent Dev. Mass spectrometry, 457-459.
Heumann K. G. and Lieser K. H. (1972)
Untersuchung von Calciumisotopieeffekten bei heterogenen Austauschgleichgewichten. Z.
Naturforsch. 27b (2), 126-133.
Heumann K. G., Klöppel H. and Sigl G. (1982)
Inversion der Calcium-Isotopenseparation an einem Ionenaustauscher durch Veänderung der
LiCl-Elektrolytkonzentration. Z. Naturforsch. 37b , 786-787.
Heuser A., Eisenhauer A., Gussone N., Bock B., Hansen B. T. and Nägler T. F. (2002)
Measurements of calcium isotopes (δ44Ca) using a multicollector TIMS technique.
International Journal of Mass Spectrometry, 220, 385-397.
Hippler D., Villa I. M., Nägler T. F. and Kramers J. D. (2004)
A ghost haunts mass spectrometry: real isotope fractionation or analytical paradox?
Geochimica et Cosmochimica Acta, 68 (11), A215.
Hofmann A. (1971)
Fractionation corrections for mixed isotope spikes of Sr, K, and Pb. Earth Planet. Sci. Lett,
10, 397-402.
Johnson C. M. and Beard B. L. (1999)
Correction of instrumentally produced mass fractionation during isotopic analysis of Fe by
thermal ionisation mass spectrometry. Int. J. Mass Spectrometry, 193, 87-99.
Nägler T. F. and Villa I. M. (2000)
In pursuit of the 40K branching ratios: K-Ca and 39Ar-40Ar dating of gem silicates. Chemical
Geology, 169, 5-16.
Nägler T. F., Eisenhauer A., Müller A., Hemleben C. and Kramers J. (2000)
The δ44Ca-temperature calibration on fossil and cultured Globigerinoides sacculifer: New tool
for reconstruction of past sea surface temperatures. Geochemistry Geophysics Geosystems, 1,
2000GC000091.
Russell W. A., Papanastassiou D. A. and Tombrello T. A. (1978)
Ca isotope fractionation on the earth and other solar system materials. Geochimica et
Cosmochimica Acta, 42, 1075-1090.
Siebert C., Nägler Th. F. and Kramers J. D. (2000)
Determination of molybdenum isotope fractionation by double-spike multicollector
inductively coupled plasma mass spectrometry. G3, 2, 2000GC0000124.
VI
Appendix A.II: PhD-thesis Dorothee Hippler, Bern (2004)
A.II Publications and Conference Abstracts
Publications
D. Hippler, K. F. Darling, A. Eisenhauer, T. F. Nägler (to be submitted)
Genetic diversity and implications for Ca isotope thermometry in polar oceans.
D. Hippler, A. Eisenhauer, T. F. Nägler (subm.)
Tropical SST history inferred from Ca isotope thermometry over the last 140ka. Geochimica
Cosmochimica Acta.
A. Eisenhauer, T. F. Nägler, P. Stille, J. Kramers, N. Gussone, B. Bock, J. Fietzke, D. Hippler, A. D.
Schmitt (2004)
Proposal for an international agreement on Ca notation resulting from discussions at workshops on
stable isotope measurements held in Davos (Goldschmidt 2002) and Nice (EGS-AGU-EUG 2003),
Geostandard Newsletters, 28.
A. Immenhauser, T. F. Nägler, T. Steuber, D. Hippler (in press)
A critical assessment of mollusk 18O/16O, Mg/Ca, 44Ca/40Ca ratios as proxies for Cretaceous seawater
temperature seasonality, Paleo3.
D. Hippler, A.-D. Schmitt, N. Gussone, A. Heuser, P. Stille, A. Eisenhauer, T. F. Nägler (2003)
Ca isotopic composition of various standards and seawater, Geostandard Newsletters, 27, 13-19.
RV Sonne Cruise Report SO164 RASTA (2003)
Rapid climate changes in the western tropical Atlantic - Assessment of the biogenous and sedimentary
record - Balboa-Balboa, May 22-June 28, 2002; Geomar Report 109, 151 pages.
Conference Abstracts
D. Hippler, K. F. Darling, A. Eisenhauer, T. F. Nägler (2004)
Ca isotopes in high-latitude marine settings: Testing the limits of reliable SST estimates based on N.
pachyderma (sin.). Abstract 2nd Swiss GeoScience Meeting 2004, Lausanne.
R. Kozdon, A. Eisenhauer, D. Hippler, C. Millo, G. Bartoli, U. Pflaumann, N. Gussone, M. Weinelt,
M. Sarnthein (2004)
δ44/40Ca in North Atlantic N. pachyderma (sin.): Potentials and limits at the “cold end”. Abstarct Int.
Conference on Paleoceanography 2004, Biarritz, B2-9.
VII
Publications and conference abstracts
D. Hippler, I. M. Villa, T. F. Nägler, J. Kramers (2004)
A ghost haunts mass spectrometry: real isotope fractionation or analytical paradox? Abstracts of the
13th Annual V. M. Goldschmidt Conference 2004, Copenhagen, Geochimica et Cosmochimica Acta,
68 (11), A215.
* Conference abstract is attached because the content is not covered by one of the manuscripts.
R. Kozdon, A. Eisenhauer, M. Sarnthein, M. Weinelt, J. Fieztke, N. Gussone, D. Hippler, B. Bock, C.
Millo (2004)
δ44/40Ca in the planktonic foraminifer N. pachyderma (sin.): A new proxy for the reconstruction of past
sea surface temperatures, Abstracts of the 13th Annual V. M. Goldschmidt Conference 2004,
Copenhagen, Geochimica et Cosmochimica Acta, 68 (11).
R. Kozdon, A. Eisenhauer, M. Sarnthein, M. Weinelt, D. Hippler, N. Gussone, C. Millo, J. Simstich
(2004)
Calcium isotope fractionation in the planktonic foraminifer N. pachyderma (sin.) as a new and sensitive
proxy for the reconstruction of sea surface temperatures in high northern latitudes, EGU 2004 Nice,
Abstract EGU04-A-03641.
D. Hippler, A. Müller, T. F. Nägler, A. Eisenhauer (2003)
Temperature changes in tropical oceans over the last 140.000 years based on δ44/40Ca, Mg/Ca and δ18O:
Warming predates melting at glacial-interglacial transitions, Eos. Trans. AGU, 64(46), Fall Meet.
Suppl., Abstract PP32A-0277.
D. Hippler, K. Darling, A. Eisenhauer, T. F. Nägler (2003)
δ44/40Ca in planktonic foraminifera: A new proxy for palaeo-sea surface temperatures in high- and lowlatitude oceans, Berichte der Deutschen Mineralogischen Gesellschaft, Beih. Z. Eur. J. Mineral. Vol.
15, No.1, p. 83.
D. Hippler, K. Darling, T. F. Nägler, A. Eisenhauer (2003)
δ44/40Ca in planktonic foraminifera G. sacculifer and N. pachyderma (sin.): A new proxy for palaeo-sea
surface temperatres in high- and low-latitude oceans, Abstract, The Micropalaeontological Society’s
Foraminiferal Group Spring Meeting 2003, GEOMAR, Kiel.
D. Hippler, T. F. Nägler, A. Eisenhauer (2003)
A 150ka record of sea surface temperature changes in the tropical Atlantic based on δ44/40Ca in
planktonic foraminifera G. sacculifer, EGS-AGU-EUG Joint Assembly, EAE03-A-02208, CL271WE2P-1403.
A. Eisenhauer, N. Gussone, A. Heuser, T. F. Nägler, D. Hippler, H. Spero, D. Lea, J. Bijma, U.
Reibesell, G. Langer (2003)
Biogechemical control on Ca-isotope fractionation, EGS-AGU-EUG Joint Assembly, EAE03-A14683, VGP32-1TU3O-005.
VIII
Appendix A.II: PhD-thesis Dorothee Hippler, Bern (2004)
D. Hippler, N. Gussone, K. Darling, A. Eisenhauer, T. F. Nägler (2002)
δ44Ca in N. pachyderma (left): A promising tool for SST-reconstructions in high-latidude oceans, Eos.
Trans. AGU, 83(47), Fall Meet. Suppl., Abstract PP51A-0305.
T. F. Nägler, D. Hippler, C. Siebert, J. D. Kramers (2002)
Paleoproxies: Heavy Stable Isotope Perspectives. Eos. Trans. AGU, 83(47), Fall Meet. Suppl., Abstract
PP51A-0263.
D. Hippler, N. Gussone, K. Darling, A. Eisenhauer, T. F. Nägler (2002)
δ44Ca in N. pachy (left): A new SST-proxy in polar regions, Abstracts of the 12th Annual V. M.
Goldschmidt Conference 2002, Davos, Geochimica et Cosmochimica Acta 66, 15A, A331.
A. Immenhauser, T. F. Nägler, T. Steuber, D. Hippler (2002)
Seasonal variations in a “greenhouse” Earth: Creteceous coastal sea surface temperatures inferred from
18O/16O, Mg/Ca, 44Ca/40Ca ratios, Abstracts of the 12th Annual V. M. Goldschmidt Conference
2002, Davos, Geochimica et Cosmochimica Acta 66, 15A, A354.
IX
Publications and conference abstracts
A ghost haunts mass spectrometry: Real isotope fractionation or
analytical paradox?
D. Hippler, I. M. Villa, T. F. Nägler and J. D. Kramers
Institute of Geological Sciences, University of Bern, Switzerland
International Goldschmidt Conference 2004, Copenhagen, Denmark
Abstract - Ca isotope palaeothermometry based on planktonic foraminifera N. pachyderma (sin.) and
G. sacculifer has been recently establish, the latter by comparison with other reliable sae surface
temperature proxies and by the systematic analysis of cultured individuals [1]. This result has been
challenged by [2], who dispute the regular fractionation trend of G. sacculifer. For analysis of Ca
isotopes we use high-precision thermal ionisation mass spectrometry. Ionisation (and therefore
fractionation) depends critically on the purity of the Ca fraction. Moreover, ailphatic fragments with 3
and 4 C atoms produce quasi-isobaric interferences in the Ca mass range. To avoid bias by organic
molecules, we completely removed them (see below) and obtained reproducible results.
A similar problem is encountered in U-Th dating of speleothems, where microbial colonisation can
give rise to organic contamination. In the case of U-Th, there is no isobaric interference, but organic
molecules adversely affect the desolvating nebulizer, and the ionisation of U and particularly Th in the
plasma source is likely impaired as well. U and Th isotope ratios of organic contaminated samples are
inaccurate and irreproducible. We speculate that unresolved organic interferences may affect the
fractionation of any element in plasma source mass spectrometry.
To eliminate organic molecules, oxidation is the method of choice. Dissolution of samples in nitric acid
can oxidise some organic molecules, but by far not all, as the oxidation potential of the nitrate anion is
at most 0.957V, while biomass contains numerous molecular species which require >1V for oxidation.
Hydrogen peroxide has an oxidation potential of 1.776V for the following reaction in acid medium:
H2O2 +2H+ + 2e- → 2H2O.
Use of hydrogen peroxide + nitric acid mixture for sample dissolution (and, if needed, also on the
column eluate) has resulted in elimination of all organic interference in most samples. The mass ratio
of the oxidising mixture is not decisive, as long as (1) hydrogen peroxide is in excess of organic
molecules, and (2) a low pH is maintained throughout.
References:
[1] Nägler et al. (2000), G3, 2000GC000091.
[2] Sime et al. (2003), Eos Trans. AGU, 64(46), Fall Meet. Suppl., Abstract PP11A-0203
X
Curriculum Vitae
Personal details
Dorothee Hippler
Date of birth: 23.09.1974
Place of birth: Esslingen am Neckar
Nationality: German
Education
Doctorate studies
April 2001 to November 2004
University of Bern,
Institute of Geological Sciences
Isotope Geology Group
University and scholar education
October 1996 – February 2001
Certificate: Diplom in Geological Sciences
Albert-Ludwigs-Universität Freiburg i. Br. (D)
Institute of Geological Sciences
September 1999 – November 2000
University final degree work
Albert-Ludwigs-Universität Freiburg i. Br. (D)
and University of Basel (CH)
October 1994 – October 1996
Certificate: Vordiplom (Geology)
Albert-Ludwigs-Universität Freiburg i. Br. (D)
Institute of Geological Sciences
Juli 1994 - September 2004
Max-Planck Gymnasium, Nürtingen (D)
Certificate: Abitur
XI