Lake Baikal climatic record between 310 and 50 ky BP: Interplay between diatoms, watershed weathering and orbital forcing

This article was originally published in a journal published byElsevier, and the attached copy is provided by Elsevier for the

author’s benefit and for the benefit of the author’s institution, fornon-commercial research and educational use including without

limitation use in instruction at your institution, sending it to specificcolleagues that you know, and providing a copy to your institution’s

administrator.

All other uses, reproduction and distribution, including withoutlimitation commercial reprints, selling or licensing copies or access,

or posting on open internet sites, your personal or institution’swebsite or repository, are prohibited. For exceptions, permission

may be sought for such use through Elsevier’s permissions site at:

http://www.elsevier.com/locate/permissionusematerial

http://www.elsevier.com/locate/permissionusematerial

Autho

r's

pers

onal

co

py

Lake Baikal climatic record between 310 and 50 ky BP: Interplaybetween diatoms, watershed weathering and orbital forcing

Tomáš Grygar a,⁎, Anna Bláhová a, David Hradil a, Petr Bezdička a, Jaroslav Kadlec b,Petr Schnabl b, George Swann c, Hedi Oberhänsli d

a Institute of Inorganic Chemistry ASCR, 250 68 Řež, Czech Republicb Institute of Geology ASCR, Palaeomagnetic Laboratory, Rozvojová 269, 165 00 Prague 6, Czech Republic

c Department of Geography, University College London, Gower Street, London WC1E 6BT, UKd GeoForschungsZentrum, Potsdam, D-14473 Potsdam, Germany

Received 9 October 2006; received in revised form 24 January 2007; accepted 4 March 2007

Abstract

The environmental record fromLake Baikal, Russia, from 310 to 50 ky BP (MIS 9a toMIS 3) was interpreted using rockmagnetic,UV–Vis spectral, mineralogical, and diatom analyses. The age model was based on a correlation of the diatom and chemicalweathering records and the summer insolation curve at 55°N and checked against an age model based on the proxy of relativepalaeointensity of the Earth's magnetic field. Peaks in chemical weathering within the watershed, inferred from maximumconcentration of magnetic and coloured minerals and mica, the lowest mean Fe oxidation state in silicates and highs in expandableclay minerals correlated with the Northern Hemisphere summer insolation minima at 55°N. Reconstructed changes in weatheringintensity are better correlated to insolation patterns than to global ice volume records. We propose a scheme of yet missingpalaeoenvironmental interpretation of the diatom assemblage, including also some extinct species. Aulacoseira baicalensis andAulacoseira skvortzowii were abundant in the early stages of lake flora recovery immediately after deglaciation and during MIS7e and MIS 5e; periods of more pronounced continental climate and peak chemical weathering. Stephanodiscus formosus var.minor,Cyclotella minuta and Cyclotella ornata dominated in intervals of decreased seasonality and decreased humidity at the end of mostinterglacial/interstadial diatom zones. Stephanodiscus grandis, Stephanodiscus carconeiformis and Stephanodiscus formosus wereubiquitous between MIS 8 and MIS 5, an interval marked by high seasonality, i.e., large differences between winter and summerinsolation, and low humidity revealed by a low hydrolysis of expandable clay minerals in the watershed. Diatom concentrationspeaked in the climatic optima of MIS 7e and MIS 5e and in the short periods marked by shifts to warmer conditions in the uppersections of MIS 5: MIS 5c (103–99 ky BP), MIS 5b (90–88 ky BP), and MIS 5a (84–79 ky BP) in which increased humidity resultedin enhanced hydrolysis of clay minerals. No such short similar climatic optimums were found from MIS 9a to MIS 6. Sharp climatedeteriorations recorded in the diatom and clay mineral records at 107, 94, and 87 ky BP, however, occurred within 1–2 ky of coldextremes in North Atlantic sea surface temperature emphasizing the strong teleconnections between the two localities.© 2007 Elsevier B.V. All rights reserved.

Keywords: Lake sediments; Climate change; Spectroscopy; Magnetic methods

1. Introduction

Lake Baikal represents a unique archive of EastEurasian continental climate. In this region the climate

Palaeogeography, Palaeoclimatology, Palaeoecology 250 (2007) 50–67www.elsevier.com/locate/palaeo

⁎ Corresponding author.E-mail address: [email protected] (T. Grygar).

0031-0182/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.palaeo.2007.03.001

Autho

r's

pers

onal

co

py

reflects both global climatic changes in addition to morelocalised continental features (Todd and Mackay, 2003).On the one hand there is evidence of strong palaeotelec-onnections between Lake Baikal and the North Atlanticregion, such as the response to Heinrich events recordedin Lake Baikal (Prokopenko et al., 2001c; Swann et al.,2005). On the other, the continental character of EastEurasia significantly amplifies certain climatic fluctua-tions which are rather weakly represented in marinesediment records. These are similar to the differencebetween the continental pollen records from Europe andrecords from the Atlantic Ocean with the former showingsubstantially shorter interglacials relative to thecorresponding marine isotopic stages (Tzedakis et al.,2004). The palaeoclimatic record from Lake Baikal hasbeen intensively studied since the 1990s, particularly inthe frame of two international initiatives, BDP andCONTINENT. The proxies developed and used withinthese initiatives for palaeoclimatic and palaeoenviron-mental reconstructions havemainly been based on diatom(Prokopenko et al., 2001a; Khursevich et al., 2001; Rioualand Mackay, 2005) and pollen analyses (Tarasov et al.,2005; Granoszewski et al., 2005). At the same timethough, elemental (Chebykin et al., 2004; Goldberg et al.,2005) and mineral (Sakai et al., 2005; Grygar et al., 2005)analyses have also been used to reconstruct palaeoclimateconditions in and around Lake Baikal.

The interpretation of Lake Baikal detritus clay mineralrecords has developed from assuming that neoformationof expandable clay minerals occurred during interglacials(Yuretich et al., 1999; Horiuchi et al., 2000) to anassumption that less straightforward clay mineral altera-tions occurred by pedogenesis (Fagel et al., 2003) to ahypothesis of complete dissolution of clay minerals ininterglacials (Sakai et al., 2005). In addition, there is also apossibility that changes in claymineralogymay have beenaffected by the environmentally controlled source arearather then changing neoformation (Grygar et al., 2005).Developing a clear understanding for the interpretation ofclay mineral assemblages is essential due to the fact thatclay minerals, and especially expandable clay minerals,have a relatively well defined pattern in recordingpalaeoclimatic changes. In contrast to this, concentrationsof other silicate minerals are often invariant to changes inthe climate. For example Fe oxide minerals, otherwise themost sensitive palaeoenvironmental indicator, are mostlikely affected by post-depositional diagenesis. In ourrecent works (Grygar et al., 2005, 2006), we used CationExchange Capacity (CEC), a novel proxy recording theconcentration of expandable clay minerals in thesediment, to determine the palaeoclimate and palaeoen-vironmental changes in and around Lake Baikal. A

significant advantage of CEC is that it can be obtained bysimple and selective chemical analyses which are muchless time consuming and more reliable than previouslyused empirical methods based on X-ray diffractionanalysis.

Diatoms in Lake Baikal have been commonly used toreconstruct palaeoclimatic and palaeoenvironmentalchanges over glacial-interglacial cycles (e.g., Khursevichet al., 2001) as well as within interglacials (e.g., Rioual andMackay, 2005) and glacials (e.g., Swann et al., 2005). Suchwork is based on the often distinct environmentalcharacteristics of individual taxa, the strong first-orderrelationship which exists between insolation patterns anddiatom abundance (Prokopenko et al., 2001a) and thediversity, appearance and extinction of individual species(Khursevich et al., 2001). Despite this, the interpretation ofdiatom records is not always clear and is often subject touncertainties or assumptions. For example, the palaeoen-vironmental interpretation of individual diatoms species isoften limited by the unknown ecological characteristics ofboth endemic and extinct diatom taxa. This is particularlytrue prior to the Holocene with many of the dominantspecies from the Pleistocene now extinct (Khursevichet al., 2001). A further feature of the Lake Baikal diatomrecord is the high extent of diatom dissolution, which canoccur both in the water column and at the surfacesediment-interface (Ryves et al., 2003). As such, there isnot always a straightforward link between the diatoms inthe sediment and past levels of lake productivity, a problemwhich can be further confused by the differentialpreservation of individual species (Battarbee et al.,2005). In addition it is likely that the relative dissolutionof diatoms was greater during glacials, either due to higherbacterial/biological action (Swann and Mackay, 2006) ordue to the lower concentrations of dissolved silica in thelake following suppressed chemical weathering in thewatershed. Consequently, due to the relative problemsassociated with both biological and non-biologicalproxies, understanding and interpreting the sedimentrecord from Lake Baikal in terms of past environmentaland climatic changes may be best achieved through acombined biogenic and non-biogenic approach (Horiuchiet al., 2000; Chebykin et al., 2004; Grygar et al., 2006).

The aim of this study is to describe the palaeoenviron-mental changes in the Lake Baikal watershed betweenMIS8 and MIS 4 by combining rock magnetic, mineralogical,and diatom based proxies. This work is a continuation ofour previous work describing the course of the last glacialcycle (Grygar et al., 2006). The methodologies used withinthis current study have been described and verified in ourprevious reports (Grygar et al., 2005, 2006) on the uppersection of the core studied in this report. In this report

51T. Grygar et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 250 (2007) 50–67

Autho

r's

pers

onal

co

py

special attention is paid to the interpretation of rockmagnetic measurements and the environmental interpreta-tion of these mineralogical proxies related to Fe oxides andsilicates including clay minerals. From this, we attempt tofind an unequivocal proxy detailing changes in chemicalweathering within the Lake Baikal watershed. Diatomassemblage records are subsequently compared to theseinorganic proxies in order to establish a better understand-ing of the relationship between the different parameters soas to provide further insights into the palaeoclimate changeswhich occurred in this region over glacial–interglacialcycles. Finally, our reconstruction of climatic changes in theLake Baikal region betweenMIS 5 andMIS 8 is comparedto changes recorded in the North Atlantic region in order todistinguish climatic differences and similarities between thetwo geographical localities.

2. Methods

2.1. Sampling and magnetic measurements

Hydraulic (piston) core VER98-1-13 was obtainedfrom the Academician Ridge, Lake Baikal, at 53.561°N,108.011°E. The upper part of the resulting section wasdescribed in previous reports (Grygar et al., 2005, 2006).The core was sampled using plastic cubic boxes with aninner volume of 6.7 cm3 (Natsuhara Giken Co., Japan)producing a continuous series of samples with a meanvertical distance of 2.2 cm between the box centres.Magnetic measurements were done after sedimentsampling in a naturally wet state. Low field volumemagnetic susceptibility (MS) was measured using theKappabridge KLY-3S (sensitivity of 1.2·10-8 SI units).All samples were demagnetised in 5–6 steps with amaximum alternating field of 100 mT by the LDA-3device. The natural remanent magnetisation (NRM)components were measured with the JR-6A or JR-5Aspinner magnetometers after each demagnetisation stepin order to gain the primary components of the NRM.The relative magnetic field palaeointensity values werecalculated as the ratio of NRM value at 20 mT in thedemagnetisation field to the anhysteretic remanentmagnetisation (ARM) value gained at 20 mT in thedemagnetisation field combined with a 0.05 mTconstantbiasing field. The ARM was imparted to the samplesusing a AMU-1A device and measured on the spinnermagnetometers. After magnetic measurements, half ofthe sample was left to dry and then ground in an agatemortar to produce samples for further chemical andphysical analyses. The plastic boxes containing theremainder of the sample were stored in a refrigeratorbefore being used for the diatom analysis.

2.2. X-ray diffraction analysis

Conventional powder X-ray diffraction (P-XRD) wasperformed to characterize the mineralogical composi-tion of the sediment. A portion of each sample was airdried at 50 °C and powdered before analysis with aSiemens D5005 diffractometer (Bruker). Diffractogramswere measured in the 2Θ range 2–70° (CuKα) with 20 scounting at 0.02° steps, resulting in a total measuringtime of almost 19 h. The X-ray tube was operatedcontinuously to minimize short-term intensity variationswith the actual primary beam intensity checked byregular (monthly) measurements of a reference corun-dum specimen (SRM 1976, NIST). The diffractionpatterns were then normalized to the mean integralintensity of the (104) line of the reference. The variationsof selected well-crystallineminerals were estimated fromvariations of integral intensities at selected diffractionlines. The content of biogenic silica was estimated fromthe height of the extremely broad diffraction (shoulderwith width FWHM ∼5° in 2Θ scale) centred at 2Θ 22°,as was also done by Fagel et al. (2003).

High-temperature X-ray diffraction (HT-XRD) wasperformed following the methodology outlined in ourprevious report (Grygar et al., 2005) to identify majorclay minerals. Samples in a naturally wet state weresuspended in a minimum volume mixture of water andethanol (1:4) and the suspension was poured onto aheated support (Pt-foil) in an Anton Paar HTK16 high-temperature chamber. The diffraction patterns wereacquired with a X'Pert PRO diffractometer (PANaly-tical) with CoKα radiation and X'Celerator multichan-nel detector. Analyses were performed at 25–300 °Cwith 5 °C steps in the 2Θ range 4–40° with 0.017° stepsresulting in a total measuring time of almost 23 h.

2.3. Cation-exchange capacity (CEC)

CEC, calculated using the Cu-trien method (adoptedfrom the original methodology of Meier and Kahr,1999), is a very convenient quantitative method for theanalysis of total expandable clay structures. The methodhas recently been successfully tested by Ammann et al.(2005). We have previously successfully used the Cu-trien method to analyse the upper part of the VER98-1-13 section (Grygar et al., 2005). 250 mg of sample (airdried at 50 °C and then powdered) was re-suspended in5 ml of water before the addition of 5 ml of 9 mMsolution of Cu(trien)SO4 (trien=1,4,7,10-tetraazade-cane). The suspension was then stirred for 10 min andfiltered into 50 ml flasks. Under these conditions, with∼50% of the Cu-trien consumed for the exchange, Cu2+

52 T. Grygar et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 250 (2007) 50–67

Autho

r's

pers

onal

co

py

ions replace exchangeable cations in the expandablestructures. The difference in concentration of Cu2+

(ΔCu2+) and the concentration of evolved Ca2+ andMg2+ were determined by atomic absorption oremission spectroscopy using AAS3 (Carl-Zeiss Jena,Germany). The mean error given by a possibleadsorption of Cu2+ or dissolution of calcium ormagnesium salts expressed as a molar non-equivalencyof ions involved in the reaction (ΔCu2+–Ca2+–Mg2+)/ΔCu2+ was 3%, i.e., the systematic error for CECdetermination was insignificant.

2.4. Diffuse reflectance spectroscopy (DRS)

The UV–Vis (electron) spectra of dried and groundsamples were measured using a Perkin Elmer Lambda35 spectrometer equipped with an integrating sphere(Labsphere). The interpretation of the electron spectra ofFe-bearing clay minerals and oxides has been describedin previous studies (Grygar et al., 2003; Hradil et al.,2004) and was recently applied to Lake Baikalsediments (Grygar et al., 2006). In this report we usedtwo characteristics based on reflectance (%): RUV,reflectance in UV–vis region at 270 nm (charge-transferbands of total Fe3+), and RVis, mean reflectance invisible-light region (400–700 nm, total lightness). Forstatistical analyses, RUV and RVis were recalculated toabsorbance using the Kubelka–Munk formula. Addi-

tionally another parameter, B/C, was obtained based onthe amount of absorption calculated from the Kubelka–Munk formula (Grygar et al., 2006) and records theabsorption ratio of Fe2+–Fe3+ (blue chromophor ofaliovalent Fe minerals, 13900 cm−1, ∼720 nm) to Fe3+

absorption (green chromophor of ferric compounds,16000 cm−1, 625 nm). As detailed within Grygar et al.(2006), the B/C ratio is a good measure of Fe oxidationstate within detritus minerals.

2.5. Diatom analysis

Diatom slide preparation, counting and identificationwere performed using a method for Lake Baikalsediments previously described in Mackay et al. (1998)and used in our previous work (Grygar et al., 2006).Sediment samples were dried at 50 °C, weighted withoutgrinding and suspended in test tubes with a definedaddition of polyvinylbenzene microspheres to calculateddiatom concentrations (Battarbee and Kneen, 1982). Theresulting suspension was evaporated on microscopeslides and mounted using Naphrax (Brunel Miroscopes,UK). For counting an optical microscope with oilimmerse objective at ×1000 magnification was used.The dominant diatom flora through the analysed periodof Stephanodiscus grandis (Khurs. and Log.), Stepha-nodiscus carconeiformis (Khurs. and Log.), Stephano-discus formosus (Khurs. and Log.), Stephanodiscus

Fig. 1. Comparison of magnetic and spectral-based proxies. Arrows show increasing chemical weathering intensity.


Autho

r's

pers

onal

co

py

formosus var. minor (Khurs. and Log.), Stephanodiscusflabellatus (Khurs. and Log.), Aulacoseira baicalensis(Meyer) Simonsen, Aulacoseira skwortzowii (Edlund,Stoermer & Taylor) (syn. Aulacoseira islandica var.helvetica), Cyclotella minuta (Skv.) Antipova,, Cyclo-tella ornata (Skv.) Flower, Cyclotella baicalensis(Meyer) Skv., Cyclotella operculata (Agardh) Kützingand Cyclotella krammeri Håkansson were identified.Benthic taxa were grouped together.

3. Results

3.1. Rock magnetic logs, UV–Vis spectral propertiesand their mineralogical interpretation

It is well known that the concentration of magneticminerals decreased in sediments in Lake Baikal duringwarmer and/or more humid periods, with minima inmagnetic susceptibility (MS) and anhysteretic rema-nence magnetisation (ARM) minima corresponding tointerglacials or interstadials (Peck et al., 1994; Demoryet al., 2005; Grygar et al., 2005, 2006). The most likelyexplanation for this is enhanced chemical weathering ofFe-bearing minerals in more humid climates. Fehydroxy-oxides and Fe bearing silicates such as darkmicas and amphiboles, and ferrimagnetic magnetite, areresponsible for the MS of the sediments. On the otherhand, ARM reflects only the concentration of ferrimag-

netics. MS (reflecting the sum of ferro-, para-, anddiamagnetic properties of the sediments) and ARM(reflecting mainly ferrimagnetic mineral properties) donot display the same patterns or changes over theanalysed interval (Fig. 1), showing that in some parts ofthe section the minima of ferrimagnetics are broaderthan the minima influenced by higher concentration ofthe paramagnetics. This difference probably indicatesthat ferrimagnetic minerals are more easily weathered inthe Lake Baikal watershed than paramagnetic mineralsand/or that some extra paramagnetics are formed by theweathering of ferrimagnetics. As such, periods withlowered ferrimagnetics and high paramagnetics likelyrepresent periods of moderate chemical weathering.

The diffuse reflectance electron spectra (DRS)produced three proxies, each defined in the methodol-ogy and shown in Figs. 1 and 2. RVis is the meanreflectivity in the Vis region, i.e., the lighter sampleshave higher values of RVis. The main coloured mineralsin sediments from the Academician Ridge are amphi-boles and dark micas (Grygar et al., 2005, 2006), both ofwhich are very sensitive to chemical weathering in ahumid climate. There is a good correlation between RVisand total diatoms or biogenic silica (Table 1), showingthat white and highly light-scattering SiO2 can contrib-ute to the pattern of high RVis in humid/warm climates.RUV is a mean reflectivity at 270 nm, which is inverselyproportional to the total Fe (FeTOT as obtained by

Fig. 2. Comparison of total diatoms (light triangles marks insolation maxima with age in ky BP), CEC and B/C (dark triangles mark insolation minimawith age in ky BP).


Autho

r's

pers

onal

co

py

chemical analysis) in the sediments; FeTOT decreased ininterglacials or interstadials (Grygar et al., 2006). B/Cratios have been shown to be a useful proxy for theoxidation state of Fe in alumosilicates, assuming thatonly a small fraction of Fe is bound in Fe3+ oxides(Grygar et al., 2006). This indicator is very sensitive tochemical weathering: if Fe2+ bearing micas areweathered to expandable clay minerals only a smallfraction of Fe is mobilized while the majority is oxidizedto Fe3+ and retained in the structure of the alumosili-cates. This process is accompanied by a decrease in theB/C ratio. Generally, if all Fe-based proxies change “inphase”, i.e., if RVis increased and RUV, B/C, MS andARM simultaneously decreased, chemical weatheringwas highly intensive and vice versa.

3.2. Age model

In our previous work (Grygar et al., 2006), the upperpart of the core section was dated based on a correlationof the relative palaeointensity variations in the VER98-1-13 sediment core to relative palaeointensity variationscalculated from ODP site 984 (Channell, 1999). Thesame dating approach was used for other recent LakeBaikal studies within the framework of the CONTI-NENT project, as detailed by Demory et al. (2005) andused, amongst others, by Swann et al. (2005), Gran-oszewski et al. (2005) and Tarasov et al. (2005). Morethan 30 points were found for reliable correlation of bothrelative palaeointensity records. However, the magneticassemblage in the Lake Baikal and the ODP site 984records could not be affected with exactly comparablepost-depositional processes. Despite this uncertaintyreversal excursions, such as the Blake (ca. 120 ka) andthe Iceland Basin (ca. 180 ka) events detected in theVER98-1-13 core section, are robust control points in therelative palaeointensity dating of the lake sediments.

An independent age model was created by correlat-ing palaeoenvironmental proxies to the NorthernHemisphere summer insolation, calculated accordingto the Berger solution (Berger, 1978) (Fig. 2, left panel).The advantage of such an approach is that peaks ininsolation often coincide with sedimentary peaks indiatom populations and other palaeoproductivity prox-ies both in interglacials and glacials. This is illustratedby the same method also being used to derive an agemodel for the long diatom record from Lake Baikalcores BDP96-2 (Prokopenko et al., 2001a) and BDP-69-1 and-2 (Prokopenko et al., 2006). The resulting agemodel for core VER98-1-13 is in good agreement withother cores from the Academician Ridge in which theage model was obtained by comparing some palaeo-productivity proxy to a marine δ18O record (Peck et al.,1994; Fagel et al., 2003; Sakai et al., 2005).

To obtain an orbitally tuned age model during theBrunhess, Prokopenko et al. (2001a) tied NorthernHemisphere summer insolation maxima and the onset ofthe main interglacial (interstadial) diatom booms. Thetiming of the onset of interglacial and interstadial diatompeaks has been confirmed by Prokopenko et al. (2001b)and Morley et al. (2005) with the first diatom peaks inLake Baikal appearing at 15–13 ky BP with the maindiatom boom developing soon after the Younger Dryas at∼10 ky BP, while the Northern Hemisphere summerinsolation maximum occurred at 11 ky BP. According tothe detailed study of the Kazantsevo interglacial (RioualandMackay, 2005), based on a palaeomagnetic agemodelwhich is explicitly independent of orbital forcing, thefastest increase in the size of the interglacial diatomassemblages occurred at 128–127 ky BP, coinciding withthe Northern Hemisphere summer insolation maximum at128 ky BP. A feature of the Lake Baikal record is thedramatic switches between cold and warm periods evenwhen Northern Hemisphere insolation is relatively

Table 1Squares of regression coefficients of correlation between selected proxies

Variable FeTOT Diatom biovolume Diatom number CEC B/C RVis RUV XRD amorphous ARM

MS 0.51 0.26 0.26 0.11 0.04 0.38⁎ 0.36⁎ 0.65 0.29ARM 0.31 0.11 0.10 0.14 0.13 0.30⁎ 0.24⁎ 0.19XRD amorphous 0.35 0.60 0.52 0.24 0.20 0.56⁎⁎ 0.66⁎⁎

RUV 0.60⁎ 0.43⁎⁎ 0.50⁎⁎ 0.49⁎ 0.00⁎ 0.69⁎

RVis 0.48⁎ 0.31⁎⁎ 0.16⁎⁎ 0.37⁎ 0.00⁎

B/C 0.01 0.15 0.14 0.00CEC 0.41 0.21 0.38Diatom number 0.40 0.50Diatom biovolume 0.34

Critical values of R2 at the 95% probability level are ca. 0.06 in all correlations with FeTOT and XRD amorphous component and ca. 0.04 for othercorrelations.⁎ reflectance recalculated to Kubelka–Munk absorbances, ⁎⁎ original reflectance in %.


Autho

r's

pers

onal

co

py

smooth. This is likely attributable to the strong climaticteleconnections which exist between the North Atlanticregion and Central Asia/Lake Baikal (Prokopenko et al.,2001c; Todd and Mackay, 2003; Swann et al., 2005). Forexample, the Northern Hemisphere summer insolationmaxima at 105 ky BP (onset of MIS 5c) and at 84 ky BP(onset of MIS 5a) in the North Atlantic were preceded byabrupt climate changes (McManus et al., 1994; Chapmanand Shackleton, 1999; Shackleton et al., 2002). Similarly,the onset of the MIS 7e, MIS 7c and MIS 5e warmintervals was preceded by large iceberg discharges anddecreases in sea surface temperature in the North Atlantic(McManus et al., 1994, 1999; Hiscott et al., 2001;Kandiano and Bauch, 2003) that accelerated the onset of awarm mode in the global ocean circulation. Due to thestrong atmospheric teleconnections between the two sites,these fast switches in the North Atlantic permit relativelyeasy identification of the onset of warm periods in theBaikal sedimentary record.

In an analogy to the relationship between the diatomrecord and insolation maxima, we suppose that theremust be a correlation between chemical weatheringminima and summer insolation minima. We tied thechemical weathering minima, identified as maxima ofARM,MS, or B/C, to the NH summer insolation minima(Fig. 2, right panel). The resulting age model is shown inFig. 3 together with the age model developed for theupper part of the core (roughlyMIS 6 toMIS 3) using thevariations of the Earth's magnetic field (Grygar et al.,2006). The total mean difference of these models was ca.1 ky with a mean standard deviation of ca. 3 ky. Because

the magnetic measurements and sampling for furtheranalyses resulted in each sample comprising on averagea 2.2 cm sample interval, each data point represents atime interval of approximately 0.5 ky. This “averaging”is consistent with the uncertainty of our age model whichis on the order of a few ky. The orbitally tuned age modelwas used to re-plot the experimental results in Figs. 4and 5. The timing of peaks in ARM,MS and B/Cwas notalways perfectly synchronous, but all occur within arange of about 3 ky, indicating that their environmentalsignal is not explicitly related to the same “recording”mechanism. These chemical weathering minima werealways found in periods free of diatom valves exceptduring the insolation minimum at 255 ky BP in themiddle of MIS 8.

3.3. Analysis of clay minerals

The use of Cation Exchange Capacity (CEC), a mea-sure of the concentration of expandable clay minerals, asan environmental proxy was previously discussed byGrygar et al. (2005). The Cu-trien method is very specificto the target compounds, i.e., smectites, vermiculite andinter-stratified clay minerals with expandable components(Meier andKahr, 1999; Ammann et al., 2005). In addition,it is one of the simplest chemical analytical methodsapplicable to large numbers of samples. CEC is directlyproportional to the total concentration of expandable claystructures and does not need empirical calibration, asrequired by P-XRD analysis of clay minerals (Yuretichet al., 1999; Fagel et al., 2003). The pattern of decreasedclay mineral content and expandable clay minerals in awarm/humid climate was found by Sakai et al. (2005) andGrygar et al. (2005), respectively. This pattern is alsoobvious from the comparison of environmental proxies inFigs. 2 and 5. This pattern has only two reasonableexplanations: either hydrolysis (dissolution) of expand-able clay minerals is occurring under humid and warmclimates in interglacials and major interstadials or someenvironmentally controlled shift is occurring in thesediment source area. In any case, CEC minima are lo-cated in periods of increased diatom concentration, al-though not all diatom maxima are coincident with CECminima.

The qualitative analysis of the claymineral assemblagewas performed using high-temperature XRD (HT-XRD).The results of HT-XRDwere similar to our previous studycovering the last glacial cycle (Grygar et al., 2005). Inperiods with enhanced chemical weathering and highdiatom productivity, the relative percentage of chloriteincreased and micas decreased, showing that the lateris preferentially consumed by chemical weathering. In

Fig. 3. Age model obtained by orbital tuning (solid line) and itscomparison to a palaeointensity age model developed for the upperpart of the core (dashed line). Palaeointensity age model taken fromGrygar et al. (2006).


Autho

r's

pers

onal

co

py

stadials and glacials, a well-ordered expandable claymineral is present, most probably vermiculite: an inter-mediate stage of mica weathering. As follows from Fig. 5,the CEC pattern resembles variations in mica contentestimated by P-XRD. This confirms the inter-layering ofillite–smectite and vermiculite (themajor expandable clayminerals found by Fagel et al., 2003; Grygar et al., 2005)with the parent mica.

3.4. Diatom analysis

Two types of information can be obtained from diatomanalysis, the concentration of individual diatom species,

which can be recalculated to biovolume measurementsusing the reference volumes given in Morley et al. (2005),Rioual and Mackay (2005), Swann and Mackay (2006)and Rioual (personal communication), and a record ofchanges in species diversity over time. Diatom frustulesincrease the content of P-XRD amorphous componentsand strongly scatter UVand Vis light thus increasing RVisandRUV. In Table 1, diatom biovolumeswere compared tothe mineralogical proxies of the opal content. Theregression coefficients of correlations between diatomsand the non-biogenic proxies improve in zones dominatedby smaller sized Cyclotella and Aulacoseira species.Changes in total diatom concentrations have a similar

Fig. 4. Interrelation of total diatom concentrations, concentration of ferromagnetics (ARM), UV reflectance and total content of amorphouscomponents (from XRD). Grey rectangles indicate diatom-rich periods.


Autho

r's

pers

onal

co

py

pattern to other proxies including magnetic susceptibility,RUV, and the amount of biogenic opal estimated byP-XRD(Fig. 4 and Table 1).

Several important features in the diatom diversity are,mostly, in agreement with previous reports (Fig. 6). Ingeneral glacial diatom assemblages contain onlyminimal numbers of diatom valves except for MIS 8.In the upper part of MIS 8 the glacial assemblages are

almost solely composed of S. grandis, S. carconeifor-mis, S. formosus, and S. formosus var. minor (also re-ported by Khursevich et al., 2001) and in MIS 6 andTermination II mostly by C. minuta and C. ornata (alsoreported by Edlund and Stoermer, 2000; Khursevichet al., 2001; Rioual and Mackay, 2005). As expected,more diverse and larger diatom populations were foundin interglacials and interstadials due to the increased

Fig. 5. Proxies of lake productivity (concentration of total diatoms), chemical weathering in the watershed (B/C and mica variations) and the CECproxy of humidity compared to summer insolation at 55°N and eccentricity in the Earth's orbit. Grey rectangles indicate diatom rich periods.


Autho

r's

pers

onal

co

pyFig.6.T

hesuccession

patternsof

diatom

scomparedto

proxiesof

chem

icalweatheringin

thewatershed.G

reyrectangles

representm

ajor

diatom

peaks(upper

MIS

8andMIS

6diatom

peaksareno

tshaded).


Autho

r's

pers

onal

co

py

solar insolation and subsequent reduction on seasonalice cover and snow thickness over the lake.

At the beginning of MIS 8 and in particular at theonset of MIS 7 an obvious diatom succession starts withpeaks in A. baicalensis and A. skvortzowii, which areimmediately followed by more stable periods dominatedby S. grandis and S. carconeiformis. Similar changesoccur during the peaks in diatom concentrations duringMIS 5. However, while the abundance and presence ofAulacoseira taxa vary throughout the analysed interval,large-celled Stephanodiscus taxa are present throughout.The ecological requirements of these extinct Stephano-discus taxa remain unknown, though an importantrequirement is believed to be deep water mixing andturbulence in the water column (Edlund and Stoermer,2000; Rioual and Mackay, 2005). In addition theirpresence, even in glacial aged samples, may indicate awide tolerance range and an ability to adapt even todifferent, more inhospitable, conditions. Alternatively, itmay indicate the robustness of these taxa to dissolutionin the water column relative to other Lake Baikal taxa.Both A. baicalensis and A. skvortzowii are cold watertaxa requiring clear winter ice permitting deep watermixing in late winter and early spring. According toRioual and Mackay (2005) both taxa have a similarecology developing under the ice during the springmonths. However, blooms of A. skvortzowii start todevelop in near shore waters before extending intopelagic/offshore locations during the spring monthswhile A. baicalensis primarily occurs in offshore waters(Mackay et al., 2000; Richardson et al., 2000).

The seasonal distribution of solar radiation duringinsolation peaks inMIS 8,MIS 7 andMIS 5 was differentto today with increased eccentricity leading to decreasedwinter insolation and increased summer insolation,producing faster spring warming (and probably alsomuch faster autumn cooling). These seasonal continentalclimatic contrasts gradually reduced fromMIS 8 toMIS 5.The concentrations of extant small-sized C. minuta,medium-sized C. ornata, and extinct small-sized S.formosus var. minor peak at the end of periods of highdiatom abundance (Fig. 6), i.e., in periods approaching theNorthern Hemisphere summer insolation minima. S.formosus var. minor, in addition to C. minuta and C.ornata, also peaked in some stadials, such as 208–200 kyBP (MIS 7b) and 94–90 ky BP (MIS 5b). Highconcentrations of the large S. grandis and S. carconei-formis were followed by peaks in S. formosus var. minorin the lower part of MIS 8 and the upper part of MIS 7,with a mean lag of ∼3 ky, and by peaks of C. minuta inthe uppermost sections of MIS 7 and by both C. minutaand C. ornata in the lowermost peaks of MIS 5 with a

mean lag of ∼4 ky. Similar succession trends have alsobeen found in sediments from the Academician Ridgeduring the climatic optima corresponding to MIS 5c andMIS 5a (Chebykin et al., 2004, Grygar et al., 2006),MIS 7and the Kazantsevo (Khursevich et al., 2001), and atContinent Ridge during the Kazantsevo interglacial(Rioual and Mackay, 2005).

The prevalence of C. minuta and C. ornata at the endof diatom peaks close to Northern Hemisphere summerinsolation minima and in a period of rather stableNorthern Hemisphere summer insolation maximum inlate MIS 6, together with the abundance of these taxa inLakeBaikal in theHolocene and today, suggests that thesetaxa can adapt to the decreased inter-seasonal insolationcontrasts which may have been more unfavourable forlarger Stephanodiscus taxa. In the early Kazantsevo A.skvortzowii prevailed, while in the latter parts of theinterglacial A. baicalensis appeared. This Aulacoseirasuccession was also found by Edlund and Stoermer(2000) at Buguldeika Saddle, Khursevich et al. (2001) attheAcademicianRidge and byRioual andMackay (2005)at Continent Ridge, suggesting that it reflects a reductionin snow/ice cover throughout the North Basin. Followingthe early Kazantsevo, A. baicalensis then prevailedthroughout the upper part of MIS 5.

Today, the Holocene seasonal insolation contrasts arethe lowest of the last 300 ky, and accordingly the currentLake Baikal diatom flora is of a similar low diversity asthat experienced duringMIS 13-MIS 9e (Khursevich et al.,2001). The extinction of S. grandis and S. carconeiformisoccurred at some point in the last glacial after MIS 5,perhaps in response to the changes in seasonal insolationdescribed above. This kind of insolation forcing in diatomspecies development is in agreement with the longPleistocene record of Khursevich et al. (2001) whofound that major diatom variations/extinctions werealigned to variations in the Earth's orbit eccentricity.

All peaks in diatom concentrations occurred inperiods of enhanced chemical weathering of primaryminerals as indicated by spectral, mineralogical androck magnetic analyses. Consequently, to exclude apossible direct inter-relationship of diatom and miner-alogical records in further analysis, it is important toestimate which mineralogical or rock-magnetic para-meters are not explicitly related to the lake bioproduc-tivity by post-depositional processes, in other words,which mineral parameters are likely to reflect onlychemical weathering in watershed. The concentration offerrimagnetic particles, evaluated by ARM, is not safefrom this point of view. Post-depositional reductivedissolution of magnetite by organic matter can decreaseARM in periods of increased lake productivity. FeTOT,


Autho

r's

pers

onal

co

py

RVis and RUV reflectance was also excluded from furtherinterpretation as they are strongly related to the con-centration of diatom frustules within the sediment.

4. Discussion

4.1. Interpretation of the diatom record

Chemical weathering within the Lake Baikal water-shed is an essential source of silicic acid for the watercolumn. The oxidation of Fe2+ in alumosilicates to Fe3+

is explicitly related to the oxidation and leaching ofexcess cations during mica transformation to vermicu-lite and/or smectite and their possible further hydrolysis(dissolution). B/C, however, is not correlated to thediatom proxies and total Fe content (Table 1). In severalzones the B/C parameter decreases before the onset oflarge diatom concentrations, as if certain levels ofsilicate weathering were required to supply sufficientsilicic acid to allow the subsequent diatom boom,although after a few ky lag. That lag has recently beenrelated to the seasonal shifts of the perihelion andperhaps different humidity and temperature optima inthe interglacials (Prokopenko et al., 2006). This trend isobvious in MIS 6 from 160 ky BP and in MIS 8 from280 ky BP. Although silicic acid is generally not agrowth-limiting factor for modern diatom populations,this may have altered in the past, especially in and soonafter dry and cold, glacial or stadial periods. However, itremains unclear at present whether the increase indiatoms at the beginning of interglacials was in responseto levels in silicic acid and other nutrients rising above acritical threshold following increased watershed weath-ering in addition to insolation related change or, ifdiatom increases occurred only in response to insolationforced changes in ice/snow cover over the lake.

Diatom concentrations in Lake Baikal have beenmost commonly directly related to changes in insolationas discussed by Prokopenko et al. (2001a), Khursevichet al. (2001), and most recently Prokopenko et al.(2006). As Khursevich et al. (2001) noticed, the diatomassemblage from MIS 9a to MIS 5 was highly stablewith no dramatic taxa extinctions. This interval ismarked by a well defined insolation pattern with in-terglacials and interstadials falling into periods of highseasonal differences caused by high eccentricity, i.e.,resulting in summer insolation above and winter in-solation below present levels. This increased seasonal-ity, or climate continentality would have enhanced thevertical mixing of the water column during spring andearly summer, believed to be pre-requisites for thegrowth of large-cell taxa S. grandis, S. carconeiformis,

and C. baicalensis (in MIS 5e), and enabled fast trans-portation of nutrients into the photic zone. Edlund andStoermer (2000) and Rioual and Mackay (2005) sup-posed that the extant large cell diatoms required clearwinter ice cover and deep water mixing. Since S. grandisand S. carconeiformis were present throughout theanalysed section including the MIS 8 glacial, we assumethat they are able to prevail even in environmentally poorcondition when turbulence may not have been high. Thisis reiterated by their presence even in glacial aged sam-ples. Alternatively, it is also possible that these extincttaxa are more resistant to dissolution than other taxa.

Concentrations ofC. minuta,C. ornata, and S. formosusvar. minor peaked in periods with decreased seasonlity,i.e., low Northern Hemisphere summer insolation min-ima (Fig. 6). Their maxima after the peak for the largerStephanodiscus taxa can imply a decrease in the deepmixing of the water but may also reflect a reducedsupply of nutrients due to climatic worsening in the laterpart of the interglacial or interstadial. CEC decreased inhumid environment, with peaks in Aulacoseira taxa inMIS 5 coincident with humidity maxima (CECminima)at 105–100 and 85–78 ky BP (Fig. 6). Conversely,peaks in Cyclotella taxa at 99–96 and 76–72 ky BPcoincided with high CEC values, i.e., a probably drierclimate but with chemical weathering still relativelyhigh as indicated by low amounts of magnetic minerals,FeTOT, and relatively low level of Fe2+/Fe3+ in silicates.The same pattern of high CEC and low Fe2+/Fe3+ is alsofound in MIS 7a during a shallow local maximum ofC. minuta. The characteristic succession of Cyclotellaspecies in the Kazantsevo interglacial was also accom-panied by the same CEC and Fe2+/Fe3+ pattern with lessabundant concentrations of expandable clay minerals inthe early stage of MIS 5e at 128 ky BP coincident withlow numbers of C. minuta. After 123–122 ky BP, how-ever, the amount of C. minuta and C. ornata started toincrease with a peak between 121 and 117 ky BP.

4.2. The duration and stability of the Kazantsevointerglacial (MIS 5e)

The course of MIS 5 and other climatically warmperiods are shown in Fig. 6. From at least 140 ky BP theindices of chemical weathering reveal a stepwise in-crease toward a maxima at 128–126 ky BP. Oscil-lating sedimentary conditions (MS), sharp spikes ofmica content (XRD), oscillations of Fe2+/Fe3+ contentand clay mineralogy between 136 and 129 ky BP andthe residues of the late glacial MIS 6 assemblage ofC. minuta and ornata accompanied deglaciation. At128 ky BP, the lower Kazantsevo stage started with a


Autho

r's

pers

onal

co

py

sharp increase in A. skvortzowii and the omnipresentStephanodiscus species including S. formosus var. mi-nor and S. grandis. Between 128 and 122 ky BP, diatomassemblage were probably the most diverse betweenMIS 8 and MIS 5 with assemblages including typicalMIS 5e species such as S. flabellatus and C. krammeri,stable concentrations of A. skvortzowii and severalfurther Cyclotella species. During this lower Kazantsevostage mica, vermiculite and total expandable clayminerals were at their lowest levels, probably due tothe high humidity, which would have led to intensehydrolysis of alumosilicates. At 122 ky BP there was aminima in the occurrence of all major diatom species,after which S. flabellatus and C. krammeri vanishedfrom the diatom assemblage while C. baicalensis andC. ornata appeared in significant concentrations andS. formosus increased with respect to concentrations inthe lower Kazantsevo. In this upper Kazantsevo zonebetween 122 and 117 ky BP, the amount of expandableclay minerals grew continuously while Fe2+/Fe3+ in sil-icates remained at low, interglacial, values. This apparentdiscrepancy is explained by increased levels of vermic-ulite (HT-XRD), an expandable clay mineral formedby mild oxidative weathering of dark micas that hadprobably been further hydrolysed (dissolved) in thelower Kazantsevo. C. minuta and C. ornata peaked justbefore the end of the diatom zone at ∼117 ky BP,reflecting the shift of the bioproductivity maximum fromspring to late summer or autumn in accordance with thelowered inter-seasonal insolation contrast. This mayperhaps also reflect a limited input of nutrients to thelake from 116 ky BP caused by a decrease in chemicalweathering as glacial conditions became established.

The duration and stability of the last interglacial haveinitiated a vivid discussion in the last decade leading tosimilar questions about the palaeoclimatic changesrecorded in the Lake Baikal record. The Kazantsevointerglacial in the sediment record analysed here lastedfrom 128 to about 117 ky BP. Similar ages are alsoreported by Rioual and Mackay (2005), Tarasov et al.(2005), Granoszewski et al. (2005) and Grygar et al.(2006). Such duration and timing are almost synchronouswith estimates from marine δ18O records (Shackletonet al., 2002) and North Atlantic sea surface temperaturerecords (McManus et al., 1999; Kandiano and Bauch,2003). The Kazantsevo as recorded in Lake Baikal henceended simultaneously with an important short, but verydramatic, climate deterioration in Central Europe at118 ky BP (Sirocko et al., 2005), which is close to theNorthern Hemisphere summer insolation minimum at116 ky BP. Elsewhere in Europe, the Eemian or MIS 5eoptimum ended between 120 and 110 ky BP depending

on the latitude of the analysed site with the climaticdeterioration earlier in Northern Europe and later inSouthern Europe (Goni et al., 2005). The sharp MIS 5dminima in sea surface temperature in the North Atlantic(cold events C24 and C23) occurred much later in theNorth Atlantic between 108 and 102 ky BP (Chapmanand Shackleton, 1999; Shackleton et al., 2002) continuingheat transportation long after the insolation minimum(McManus et al., 2002).

A dramatic decrease in diatom opal was found byProkopenko et al. (2001a) at 122 ky BP and was in-terpreted as a dramatic mid-Kazantsevo cooling. Contrari-ly, a rather small-scale climatic worsening at 120 ky BPfollowed by very moderate cooling was inferred fromdiatom analyses by Rioual and Mackay (2005). Irrespec-tive of the magnitude of the changes and its precise dating,the change was accompanied by a shift in the diatomcommunity to rarer taxa. A decrease in diatom concentra-tions and biovolume between DAZ4 and DAZ5 at 120 kyBP in Rioual andMackay (2005) shows similarities to ourrecord with decreases in S. flabellatus and increases inconcentrations of S. formosus, S. formosus var. minor andC. minuta. Khursevich et al. (2001) found similar changesat the Academician Ridge in their stratigraphical denota-tion of diatom zones with A. skvortzowii in LDAZ 6(lower) completely replaced by A. baicalensis in LDAZ 5(upper) while S. flabellatus was exclusively found inLDAZ 6, C. baicalensis exclusively in LDAZ 5 and withchanges in C. minuta and C. ornata again similar to ourrecord. S. grandis, S. carconeiformis and S. formosuswere present throughout the MIS 5e interval. As such theclimate deterioration in the last part of the Kazantsevomust be evaluated from other taxa such asAulacoseira andCyclotella species and non-biogenic proxies which appearto be more responsive to climatic changes. From this,trends indicating drier conditions were found at the end ofMIS 5 by Grygar et al. (2006). Chebykin et al. (2004)assumed that the deficiency of nutrients in the lakewas dueto low riverine input, inferred fromU-series isotopes, led tothe decrease in diatoms at the end of MIS 5. Morespecifically, they found two successions from A. baica-lensis via S. grandis to C. minuta accompanied by asubstantial decrease of the riverine input of U to the lakefrom 100 to ∼93 ky BP (corresponding to MIS 5c) andfrom ∼83 to 74.5 ky BP (MIS 5a).

The Kazantsevo (128–117 ky BP) was the climaticoptimum of the studied core section. The diversity of thediatom assemblage peaked during this period with thehydrolysis of alumosilicates also at its most intense,probably due to the high humidity also occurring in thisinterval. Recently Edlund (2006) noted that the diatomdiversity, otherwise generally very low in Lake Baikal


Autho

r's

pers

onal

co

py

over the last 500 ky, was at its highest in the Kazantsevo,MIS 3 and in the Holocene. This may be due to theintensive chemical weathering in the watershed over thisinterval, resulting in the high influx of nutrients into thelake. A typical feature of the Kazantsevo diatomassemblage is the occurrence of peak diatom concentra-tions for most taxa relative to other periods in the analysedinterval. In addition, new species emerged including S.flabellatus, A. baicalensis, and C. baicalensis. Further-more, MIS 5 was the only period betweenMIS 8 andMIS4 where relatively high concentrations of Aulacoseira,C. minuta and C. ornata prevailed alongside the moredominant S. grandis and S. carconeiformis, indicating thespread of bioproductivity to all seasons and the presenceof early ice-free dates in addition to reduced spring andautumn ice/snow cover. Change in the diatom assem-blages in the upper part of theKazantsevo can therefore beattributed to decreasing seasonality, due to orbital forcing,combined with a stepwise decrease in the intensity of thechemical weathering as conditions switched towards thesubsequent local glaciation (117–107 ky BP) and theexpansion of ice cover/thickness over the lake.

4.3. Chemical weathering minima

Extremes in B/C, log(ARM) and MS were found atabout 280, 230, 185, 115, and 65 ky BP, i.e., mostly inperiods when summer insolation at 55°N dropped below460 W/m (Figs. 5 and 7) and when these minima wereclose to low obliquity. These weathering minima did notcorrespond to similar peaks in SPECMAP, as shown inFig. 7 where grey rectangles, indicating the chemicalweathering minima, are always coincident with diatombarren intervals. The mean marine δ18O values in theseintervals were not always extremely high with the firstsigns of cooling Lake Baikal during MIS 8 and MIS 6(zones I and III, respectively) occurring before the endof the MIS 7e andMIS 5e warm intervals as indicated bythe marine δ18O record (zones II and IV, respectively).In addition the decrease in diatoms and increases inchemical weathering in Lake Baikal at the end of MIS5e (zone V) appear to have occurred rapidly with animmediate switch from interstadial or interglacial toglacial conditions. Such fast switches are most likelycaused by the development of a local glaciation in aform of river valley glacier in the Lake Baikal catch-ment. This sensitivity to cooling during decreasedsummer NH insolation is similar to the response ofhigh-latitude N Atlantic region, and North Europe icesheet growth soon after interglacial climatic optima.

During the mild glacial MIS 8, XRD and CEC showvalues indicative of “cold-regime” conditions in spite of

moderate chemical weathering (B/C, ARM, MS) andtwo diatom maxima. The lower diatom maxima wasmarked at its onset by increases in Aulacoseira sp.before finishing with assemblages being dominated byS. grandis. The second peak is dominated by S. gran-dis. The lower diatom peak ends at ∼290 ky BP whenNorth Atlantic sea surface temperatures were almost ashigh as during MIS 5c and MIS 5a (Kandiano andBauch, 2003) while in Lake Baikal both lake produc-tivity and indices of chemical weathering were similar tothose experienced in MIS 7e. In the Northern Hemi-sphere, MIS 8 was less severe than the MIS 6 and MIS 2glacials, both in terms of global ice volume and NorthAtlantic sea surface temperatures (McManus et al.,1999; Hiscott et al., 2001; Kandiano and Bauch, 2003).In Lake Baikal the intensity of chemical weatheringincreased toward the end of these glacials. Again, lakeproductivity followed changes in insolation rather thanglobal ice volume or North Atlantic temperatures,although in Lake Baikal during MIS 6 there was asmall diatom peak dominated by the warmer water taxaC. minuta and C. ornata and smaller concentrations ofS. grandis at about 150 ky BP when the North Atlanticunderwent the lowest sea surface temperatures of thepenultimate glaciation (Kandiano and Bauch, 2003).

4.4. Comparison of MIS 7 and MIS 5

Three warm periods were identified in MIS 7 (Figs. 4and 5). Although the vast majority of the diatom species inMIS 7 are today extinct, some conclusion can be drawn.The diatom assemblage fromMIS 7c to MIS 7a resemblesthat of the upper peak in diatom concentrations during theMIS 8 glacial while mineralogical proxies indicate muchweaker chemical weathering than in the correspondingMIS 5 sub-stages. An almost identical content of biogenicSiO2 inMIS 7 andMIS5was reported byProkopenko et al.(2001a). Similarly, in our record the content of P-XRDamorphous sediments in the three diatom peaks in MIS 7was almost as high as in the MIS 5e peak. However, thehydrolysis of expandable clay minerals was much weaker,particularly in MIS 7a, indicating most likely a less humidenvironment. The diatom record and indices of chemicalweathering clearly indicated that MIS 7e was the climaticoptimum of MIS 7. There were no sharp climaticoscillations in the upper part of MIS 7 such as thosefound in the upper parts ofMIS 5, indicating that the end ofMIS 7 was a fairly stable period.

Such an evaluation of theMIS 7 climate in LakeBaikaland its watershed is similar to European continentalsequences (Reille et al., 2000; Tzedakis et al., 2004) andalso to other records from Lake Baikal (Prokopenko et al.,


Autho

r's

pers

onal

co

py

2001a; Khursevich et al., 2001). The MIS 7e climaticoptimum was rather short and ended with a very harshcold period, MIS 7d, both in the European continent andin the North Atlantic (Reille et al., 2000; Kandiano andBauch, 2003; Tzedakis et al., 2004). The stadial MIS 7bwasmuch less pronounced thanMIS 5b at our site, similarto the pollen and diatom records fromAcademician Ridge(Sakai et al., 2005). The stadial corresponding to MIS 7bin Europe was also weakly pronounced in pollen record

from Central France (Reille et al., 2000) while a veryshallow minimum in North Atlantic sea surface temper-aturewas found inMIS 7bwith thewhole of theMIS 7c toMIS 7a interval characterised by a rather uniform warmperiod (Kandiano and Bauch, 2003). North Atlantic seasurface temperatures, though, were significantly lower inMIS 7 than in MIS 5e (Kandiano and Bauch, 2003).

In West European continental sequences, the MIS 7eoptimum was much shorter than MIS 5e (Tzedakis et al.,

Fig. 7. Correlation of minima (grey rectangles) in chemical weathering (expressed as magnetic susceptibility in the bottom panel) and diatomconcentrations with summer insolation pattern at 55°N and a δ18O composition of planktonic foraminifera from ODP site 980 in the North Atlanticcore (top panel, from McManus et al., 1999).


Autho

r's

pers

onal

co

py

2004). A similar pattern is also present in Lake Baikal forMIS 7e and MIS 5e (Figs. 4 and 5). Although the diatompeak of MIS 7e lasted about 11 ky, it started with arelatively long, ∼4 ky, long period dominated by Aula-coseira taxa and without significant hydrolysis ofexpandable clay minerals. This was followed by a∼5 ky minor peak in Aulacoseira taxa coeval with amajor peak inS.grandis and otherStephanodiscus species.The latter ∼5 ky long period should hence be consideredthe climatic optimum of the MIS 7 interval. The relativeshortness of the climatic optimum in Lake Baikal duringMIS 7emay be due to two intense iceberg discharges fromthe Laurentide ice sheets (Hiscott et al., 2001) resulting inextremely low sea surface temperatures in the North EastAtlantic Ocean region in MIS 7e compared to otherMIS 7warm intervals in theNorthernHemisphere. An alternativeexplanation could be that the MIS 7e insolation maximumwas not as pronounced with the low eccentricity of theEarth's orbit resulting in the onset of an interglacial whichwas not as intensive as the corresponding onset of MIS 5e.

5. Conclusion

Orbital forcing is very strongly pronounced in the LakeBaikal region due to its extreme continentality, highlatitude and weak marine influence. Orbital forcing,namely the summer Northern Hemisphere insolation at55°N, can be tightly related to diatom peaks (summerinsolation maxima) and chemical weathering minima inthe watershed (summer insolation minima, especially inperiods of low obliquity) and was used to construct an agemodel, confirmed by the palaeomagnetic dating. Neitherthe total number of diatom valves, total diatom biovo-lumes nor the amount of biogenic opal can be used aloneto interpret the palaeoclimate/palaeoenvironment. Instead,conditions can be evaluated from combining mineralproxies of chemical weathering and taxonomic diatomanalysis to the species level. Diatoms were very abundantnot only in interglacials and interstadials, but also in theinsolation maxima of the MIS 8 glacial. Increased diatomdiversity across all species indicates the extension of lakeproductivity and climatically more favourable environ-mental conditions during the climatic optima of MIS 7e,MIS 5e and some shorter periods fromMIS 5c to MIS 5a.Two “inorganic” proxies of watershed chemical weather-ing not explicitly affected or controlled by lake produc-tivity are the oxidation state of Fe in silicates and oxides,measured by diffuse reflectance spectroscopy, and thecontent of expandable clay minerals, obtained bydetermining the cation exchange capacity.

Although periods of the harshest glacial climate didnot coincide with δ18O maxima in marine foraminifera

records, sharp climatic changes in the North Atlanticregion between the middle of MIS 5e and MIS 5a arewell expressed in the Lake Baikal sediment record. Theextent and magnitude of local glaciation in the Baikalwatershed, however, do not correspond well with theaverage extent of global glaciation. The combination ofdiatom analysis and weathering indices indicates thatwithin the interval from MIS 9a to MIS 4, humidity washighest in MIS 5e and in the short periods between MIS5c and MIS 5a and moderate in MIS 7e, MIS 8e and theearly MIS 8 interstadial. In contrast, the least humidperiods were the interstadials MIS 8c, MIS 7c and MIS7a. Unusually the MIS 8 glacial was very moderate com-pared to the early half of MIS 6, the climatically harshestperiod in the Lake Baikal region. Furthermore, in contrastto MIS 8 which was interrupted by two interstadials ofrelatively intensive chemical weathering and increases inlake productivity, MIS 6 was interrupted by only a single,very weak, increase in lake productivity.

Acknowledgement

The cores were retrieved from RV Vereschaginoperated by the Limnological Institute RAS SB Irkutsk,Russia, with the coring technology provided by the Prof.Meischner Group from the University Göttingen, Ger-many. The chemical, mineralogical and X-ray powderdiffraction analyses were partly funded by the GrantAgency of the Czech Republic (IAA3032401) and kind-ly performed by Jana Dörflová, Petr Vorm and AntonínPetřina (Institute of Inorganic Chemistry ASCR). Theauthors thank Marie-France Loutre (Universite catholi-que de Louvain, Belgium) for kindly providing the Earthorbital parameters.

References

Ammann, L., Bergaya, F., Lagaly, G., 2005. Determination of thecation exchange capacity of clays with copper complexes revisited.Clay Minerals 40, 441–453.

Battarbee, R.W., Kneen, M.J., 1982. The use of electronically countedmicrospheres in absolute diatom analysis. Limnology and Oceanog-raphy 27, 184–188.

Battarbee, R.W.,Mackay, A.W., Jewson,D.H., Ryves, D.B., Sturm,M.,2005. Differential dissolution of Lake Baikal diatoms: correctionfactors and implications for palaeoclimatic reconstruction. Globaland Planetary Change 46, 75–86.

Berger, A., 1978. Long-term variations of daily insolation and Quater-nary climatic changes. Journal of the Atmospheric Sciences 35,2362–2367.

Channell, J.E.T., 1999. Geomagnetic paleointensity and direction-al secular variation at Ocean Drilling Program (ODP) site 984(Bjorn Drift) since 500 ka: comparisons with ODP site 983(Gardar drift). Journal of Geophysical Research, B: Solid Earth104, 22,937–22,951.


Autho

r's

pers

onal

co

py

Chapman, M.R., Shackleton, N.J., 1999. Global ice-volume fluctua-tions, North Atlantic ice-rafting events, and deep-ocean circulationchanges between 130 and 70 ka. Geology 27, 795–798.

Chebykin, E.P., Edgington, D.N., Goldberg, E.L., Phedorin, M.A.,Kulikova, N.S., Zheleznyakova, T.O., Vorob Vyova, S.S., Khlystov,O.M., Levina, O.V., Ziborova, G.A., Grachev,M.A., 2004.Uranium-series isotopes as proxies of Late Pleistocene climate andgeochronometers in bottom sediments of Lake Baikal. Geologia iGeofizika 45, 539–556.

Demory, F., Nowaczyk, N.R., Witt, A., Oberhansli, H., 2005. High-resolution magneto stratigraphy of late quaternary sediments fromLake Baikal, Siberia: timing of intracontinental paleoclimaticresponses. Global and Planetary Change 46, 167–186.

Edlund, M.B., 2006. Persistent low diatom plankton diversity withinthe otherwise highly diverse Lake Baikal ecosystem. NovaHedwigia 130, 339–355 (suppl.).

Edlund, M.B., Stoermer, E.F., 2000. A 200,000-year, high-resolutionrecord of diatom productivity and community makeup from LakeBaikal shows high correspondence to the marine oxygen-iso-tope record of climate change. Limnology and Oceanography 45,948–962.

Fagel, N., Boski, T., Likhoshway, L., Oberhansli, H., 2003. LateQuaternary clay mineral record in Central Siberia Lake Baikal(Academician Ridge, Siberia). Palaeogeography, Palaeoclimatology,Palaeoecology 193, 159–179.

Goldberg, E.L., Chebykin, E.P., Vorobyova, S.S., Grachev, M.A.,2005. Uranium signals of paleoclimate humidity recorded insediments of Lake Baikal. Doklady Earth Sciences 400, 52–56.

Goni, M.F.S, Loutre, M.F., Crucifx, M., Peyron, O., Santos, L.,Duprat, J., Malaize, B., Turon, J.L., Peypouquet, J.P., 2005.Increasing vegetation and climate gradient in Western Europe overthe Last Glacial Inception (122–110 ka): data-model comparison.Earth and Planetary Science Express 231, 111–130.

Granoszewski, W., Demske, D., Nita, M., Heuman, G., Andreev, A.A.,2005. Vegetation and climate variability during the Last Intergla-cial evidenced in the pollen record from Lake Baikal. Global andPlanetary Change 46, 187–198.

Grygar, T., Dědeček, J., Kruiver, P., Dekkers, M.J., Bezdička, P.,Schneeweiss, O., 2003. Iron oxide mineralogy in Late Miocene redbeds from La Gloria, Spain: rock-magnetic, voltammetric and visspectroscopy analyses. Catena 53, 115–132.

Grygar, T., Bezdička, P., Hradil, D., Hrušková, M., Novotná, K.,Kadlec, J., Pruner, P., Oberhänsli, H., 2005. Characterization ofexpandable clay minerals in Lake Baikal sediments by thermaldehydration and cation exchange. Clays and Clay Minerals 53,389–400.

Grygar, T., Kadlec, J., Pruner, P., Swann, G., Bezdička, P., Hradil, D., Lang,K., Novotna, K., Oberhänsli, H., 2006. Paleoenvironmental record inLake Baikal sediments: environmental changes in the last 160 ky.Palaeogeography, Palaeoclimatology, Palaeoecology 237, 240–254.

Hiscott, R.N., Aksu, A.E., Mudie, P.J., Parsons, D.F., 2001. A340,000 year record of ice rafting, palaeoclimatic fluctuations, andshelf-crossing glacial advances in the southwestern Labrador Sea.Global and Planetary Change 28, 227–240.

Horiuchi, K., Minoura, K., Hoshino, K., Oda, T., Nakamura, T.,Kawai, T., 2000. Palaeoenvironmental history of Lake Baikalduring the last 23000 years. Palaeogeography, Palaeoclimatology,Palaeoecology 157, 95–108.

Hradil, D., Grygar, T., Hrušková, M., Bezdička, P., Lang, K.,Schneeweiss, O., Chvátal, M., 2004. Green earth pigment fromKadan region, Czech Republic: use of rare Fe-rich smectite. Claysand Clay Minerals 52, 767–778.

Kandiano, E.S., Bauch, H.A., 2003. Surface ocean temperatures in thenorth-east Atlantic during the last 500000 years: evidence fromforaminiferal census data. Terra Nova 15, 265–271.

Khursevich, G.K., Karabanov, E.B., Prokopenko, A.A., Williams, D.F.,Kuzmin, M.I., Fedenya, S.A., Gvozdkov, A.A., 2001. Insolationregime in Siberia as a major factor controlling diatom production inLake Baikal during the past 800,000 years. Quaternary International80–81, 47–58.

Mackay, A.W., Flower, R.J., Kuzmina, A.E., Granina, L.Z., Rose, N.L.,Appleby, P.G, . Boyle, J.F., Battarbee, R.W., 1998. Diatomsuccession trends in recent sediments from Lake Baikal and theirrelation to atmospheric pollution and to climate change. Philosoph-ical Transactions of theRoyal Society of London. B 353, 1011–1055.

Mackay, A.W., Battarbee, R.W., Flower, R.J., Jewson, D.H., Lees, J.A.,Ryves, D.B., Sturm, M., 2000. The deposition and accumulation ofendemic planktonic diatoms in the sediments of Lake Baikal and anevaluation of their potential role in climate reconstruction duringthe Holocene. Terra Nova 9, 34–48.

McManus, J.F., Bond, G.C., Broecker, W.S., Johnsen, S., Labeyrie, L.,Higgins, S., 1994. High-resolution climate records from the North-Atlantic during the last interglacial. Nature 371, 326–329.

McManus, J.F., Oppo, D.W., Cullen, J.L., 1999. A 0.5-million-yearrecord of millennial-scale climatic variability in the North Atlantic.Science 283, 971–975.

McManus, J.F., Oppo, D.W., Keigwin, L.D., Cullen, J.L., Bond, G.C.,2002. Thermohaline circulation and prolonged interglacial warmthin the North Atlantic. Quaternary Research 58, 17–21.

Meier, L.P., Kahr, G., 1999. Determination of the cation exchangecapacity (CEC) of clay minerals using the complexes of copper(II)ion with triethylenetetramine and tetraethylenepentamine. Claysand Clay Minerals 47, 386–388.

Morley, D.W., Leng, M.J., Mackay, A.W., Sloane, H.J., 2005. Lateglacial and Holocene environmental change in the Lake Baikalregion documented by oxygen isotopes from diatom silica. Globaland Planetary Change 46, 221–233.

Peck, J.A., King, J.W., Colman, S.M., Kravchinsky, V.A., 1994. A rock-magnetic record from lake Baikal, Siberia—evidence for LateQuaternary climate change. Earth and Planetary Science Letters 122,221–238.

Prokopenko, A.A., Karabanov, E.B., Williams, D.F., Kuzmin, M.I.,Shackleton, N.J., Crowhurst, S.J., Peck, J.A., Gvozdkov, A.N., King,J.W., 2001a. Biogenic silica record of the Lake Baikal response toclimatic forcing during the Brunhes. Quaternary Research 55,123–132.

Prokopenko, A.A., Karabanov, E.B., Williams, D.F., Kuzmin, M.I.,Khursevich, G.K., Gvozdkov, A.N., 2001b. The detailed record ofclimatic events during the past 75,000 yrs BP from the Lake Baikaldrill core BDP-93-2. Quaternary International 80–81, 59–68.

Prokopenko, A.A.,Williams, D.F., Karabanov, E.B., Khursevich, G.K.,2001c. Continental response to Heinrich events and Bond cycles insedimentary record of Lake Baikal, Siberia. Global and PlanetaryChange 28, 217–226.

Prokopenko, A.A., Hinnov, L.A., Williams, D.F., Kuzmin, M.I.,2006. Orbital forcing of continental climate during thePleistocene: a complete astronomically tuned climatic recordfrom Lake Baikal, SE Siberia. Quaternary Science Reviews 25,3431–3457.

Reille, M., De Beaulieu, J.L., Svobodová, H., Andrieu-Ponel, V.,Goeury, C., 2000. Pollen analytical biostratigraphy of the last fiveclimatic cycles from a long continental sequence from the Velayregion (Massif Central, France). Journal of Quaternary Science 15,665–685.


Autho

r's

pers

onal

co

py

Richardson, T.L., Gibson, C.E., Heaney, S.I., 2000. Temperature,growth and seasonal succession of phytoplankton in Lake Baikal,Siberia. Freshwater Biology 44, 431–440.

Rioual, P., Mackay, A.W., 2005. A diatom record of centennialresolution for the Kazantsevo Interglacial stage in Lake Baikal(Siberia). Global and Planetary Change 46, 199–219.

Ryves, D.B., Jewson, D.H., Sturm, M., Battarbee, R.W., Flower, R.J.,Mackay, A.W., Granin, N.G., 2003. Quantitative and qualitativerelationships between planktonic diatom communities and diatomassemblages in sedimentingmaterial and surface sediments in LakeBaikal, Siberia. Limnology and Oceanography 48, 1643–1661.

Sakai, T., Minoura, K., Soma, M., Tani, Y., Tanaka, A., Nara, F., Itoh,N., Kawai, T., 2005. Influence of climate fluctuation on clayformation in the Baikal drainage basin. Journal of Paleolimnology3, 105–121.

Shackleton, N.J., Chapman, M., Sanchez-Goni, M.F., Pailler, D.,Lancelot, Y., 2002. The classic marine isotope substage 5e.Quaternary Research 58, 14–16.

Sirocko, F., Seelos, K., Schaber, K., Rein, B., Dreher, F., Diehl, M.,Lehne, R., Jager, K., Krbetschek, M., Degering, D., 2005. A lateEemian aridity pulse in central Europe during the last glacialinception. Nature 436, 833–836.

Swann, G.E.A., Mackay, A.W., 2006. Potential limitations of biogenicsilica as an indicator of abrupt climate change in Lake Baikal,Russia. Journal of Paleolimnology 36, 81–89.

Swann, G.E.A., Mackay, A.W., Leng, M.J., Demory, F., 2005.Climatic change in Central Asia during MIS 3/2: a case studyusing biological responses from Lake Baikal. Global and PlanetaryChange 46, 235–253.

Tarasov, P., Granoszewski, W., Bezrukova, E., Brewer, S., Nita, M.,Abzaeva, A., Oberhänsli, H., 2005. Quantitative reconstruction ofthe last interglacial vegetation and climate based on the pollenrecord from Lake Baikal, Russia. Climate Dynamics 25, 625–637.

Todd, M.C., Mackay, A.W., 2003. Large scale climatic controls onLake Baikal ice cover. Journal of Climate 16, 3186–3199.

Tzedakis, P.C., Roucoux, K.H., de Abreu, L., Shackleton, N.J., 2004.The duration of forest stages in southern Europe and interglacialclimate variability. Science 306, 2231–2235.

Yuretich, R., Melles, M., Sarata, B., Grobe, H., 1999. Clay minerals inthe sediments of Lake Baikal: a useful climate proxy. Journal ofSedimentary Research 69, 588–596.


Lake Baikal climatic record between 310 and 50 ky BP: Interplay between diatoms, watershed weathering and orbital forcing

Documents