Lake Baikal is located in the southeastern Siberia between 51° and 56° North latitude and between 104° and 110° East longitude (Fig. 7); it lies at 486 m elevation and is surrounded by mountain ridges that rise northward to the altitudes of 2500 - 3000 m (Galazii 1993). Lake Baikal is 23 million years old and is the oldest lake of the world (Kozhov 1963). It measures more than 600 km long by 80 km wide and the deepest point is over 1.6 km (Kozhov 1963). It holds 23,000 km³ of water and 20 % of the World´s fresh surface water. Lake Baikal has an exceptional clarity, which allows 40-50 m of visibility (Kozhova and Imest´eva 1998).
Lake Baikal can be divided into three basins (Fig. 7). The North basin is separated from the Central basin by the Academician Ridge, and the Central basin is separated from the South basin by the more than 20 km wide Selenga Delta. All basins exhibit an asymmetric, half-graben morphology, with steep margins on the northwestern side and more gradual, though still faulted margins on the southeastern side (Colman et al. 2003, Fig. 8).
|Fig. 7. Map of Lake Baikal locating relevant regions for the present study and impressions of the respective surrounding regions.|
|Fig. 8. Half-graben morphology from West to East: Maloe More, Olkhon Island, Central basin; from Kozhova and Izmest’eva 1998.|
While approximately 365 inflows enter the Baikal, from which the river Selenga is the largest one, only one effluent is known, the river Angara at the western shore of the South basin (Fig. 7). The lake´s catchment area encompasses c. 540.000 km2 of forest and desert environments. Most of the watershed area is surfaced with rocks so that water inflow has little mineral or chemical content. The chemical composition is constant through the water column. Budgets of nitrogen and phosphorus compounds have been poorly investigated reporting contradictory results (Granina 1997). Recent studies on nutrient availability indicated, that, although Lake Baikal is a freshwater lake, the nutrient status of the pelagial is ocean-like, and the only nutrient-rich regions of the lake are the deltas of the main river inflows (Genkai-Kato 2002).
Lake Baikal is a dimictic lake. However, even during homothermy the wind-induced overturn is limited to the upper 250-300 m; the water masses below the wind-induced mixed layer are homothermal (c. 4° C; Granin et al. 1991, Shimaraev et al. 1994). Deep water ventilations occur regularly (Weiss et al. 1991, Shimaraev et al. 1994). The lake freezes by December or January, depending from latitude. Ice thickness often exceeds one meter and holds mostly only a thin snow cover. Melting starts from April to May.
The climate is continental. The winters are influenced by the Siberian high pressure system, which brings cold Arctic air into the region (clockwise circulation). In spring the central pressure of the Siberian High decreases and the air masses shift west towards Europe (Bradbury et al. 1994). During this transitional period the Siberian High becomes weaker, but more variable in the Lake Baikal region and this variability affect timing and rate of ice thaw and consequently the extent and rate of the characteristic spring phytoplankton development (Bradbury et al. 1994). In summer the low-pressure system of the Asiatic Low extends towards the Baikal region and brings moisture from the Indian Ocean (counterclockwise winds; Bradbury et al. 1994).
CONTINENT summer cruises:Sampling for the study of regional distribution was conducted in July 2001, 2002 and 2003 during the CONTINENT cruises CON 01-4, CON 01-5, CON 02-8, CON 03-9 with the research vessel “Vereshchagin” (Fig. 9). In 2001 samples for pigment analyses were taken from 0.5, 5, 10 and 30 m water depth. Samples for phytoplankton counting were taken from 5, 10 and 30 m water depth. In 2002 samples for pigment and phytoplankton analyses were taken from 5, 10 and 30 m and/or in the deep Chla maxima determined with a submersible fluorimeter (FluoroProbe, bbe Moldaenke GmbH, Kiel, Germany). In 2003 only samples for pigment analyses were taken (same depths as in 2002). Due to the differences in sampling depth, samples were grouped into “0.5 m”, “5 m”, “10 – 20 m” and “20 – 30 m” and “45 – 85 m”. In 2001 temperature was directly measured in the samples. Temperature in 2002 and 2003 was provided by the fluorimeter and by CTD (Conductivity, Temperature, and Depth) profiles (R. Gnatovsky and N. Granin, Limnological Institute of Irkutsk, Russia).
|Fig. 9. Research vessel “Vereshchagin” (left) and on board of the research vessel “M.M. Kozhov” (Bolshye Koti in the background) (right).|
Seasonal monitoring at Bolshye Koti: For the study of seasonal dynamics, weekly sampling from May 2002 to June 2003 was carried out at the long-term sampling site of the Scientific Research Institute of Biology (SRIB, State University Irkutsk, Russia), located 2.8 km offshore from Bolshye Koti (51°54´ N 105° 04´E, Fig. 7, Fig. 9). Water depth at that site was 800 m. Samples from this latter station were taken with the research vessel “M.M. Kozhov” in summer or from the ice in winter and processed at the Biological Station Bolshye Koti (G. Kobanova, L. Kraschuk, E. Pislegina, and L. Izmest’eva, SRIB).
Samples for autotrophic picoplankton (APP, 0.2 – 3 µm; 50 mL) were preserved with formaldehyde (0.7 % final concentration) and filtered through black Nuclepore© polycarbonate filters (0.2 µm pore size). The filter was placed on a microscope slide, quickly dried and covered with a drop of fluorescence-free immersion-oil and a coverslip. Once frozen, preparations were stable for months. APP were counted at 1000-x magnification using a Zeiss Axioskop epifluorescence microscope equipped with filters for green (546 nm excitation filter, 580 nm splitter and 590 nm barrier filter) and blue (450-490 nm excitation filter, 510 nm splitter and 520 nm barrier filter) excitation. Eukaryotic APP fluoresced deep red (>665 nm) when excited with blue or green light, whereas cyanobacterial APP fluoresced light-red (<665 nm) when excited with green light (Phycobilins). Phycoerythrin and phycocyanin containing cyanobacteria were distinguished by their respective yellow or extreme weak emission at blue light excitation, but this difference was not definite in all stored preparations. Cell counts were converted to biovolume according to their size and geometric form (Fietz and Nicklisch 2004, Appendix A). During the seasonal monitoring APP was counted with a light microscope, whereby colonies were easily identified, but single cells could be overlooked.
In both years of phytoplankton counts (2001 and 2002) samples (1-2 L) were enriched by filtering through Nuclepore© polycarbonate filters (2 µm pore size), fixed with some drops of Lugol´s solution and stored at room temperature. Counting and identification was done according to the settling technique (Utermöhl 1958). In 2001 the algae were classified in accordance with Ettl et al. (1986) (H. Täuscher, Leibniz Institute of Freshwater Ecology and Inland Fisheries, Berlin, Germany). In 2002 the taxonomic composition of algae was established more in detail (G. Kobanova, SRIB) in accordance with “The Keys of Freshwater Algae of the USSR” (see Kozhova and Izmest’eva 1998, p. 325 for references), with monographs (references in Kozhova and Izmest’eva 1998, Bourrelly 1957), and additional keys (references in Kozhova and Izmest’eva 1998, Topachevsky and Masyuk 1984, Wasser et al. 1989, Gleser et al. 1992), and supplemented by articles (references in Kozhova and Izmest’eva 1998,Edlund et al. 1996). Gymnodinium coeruleum (Pyrrophyta) is often cited within Lake Baikal phytoplankton assemblages (Kozhova 1987, Kozhova and Izmest’eva 1998, Genkai-Kato et al. 2003), but has been omitted from the phytoplankton counts in the present study because it did not contain chloroplasts and, therefore, was counted as protozoan. All species were tentatively grouped into functional associations according to the scheme proposed by Reynolds et al. (2002). We consulted with its authors over the classification of the species we have encountered.
In this study we also used chemotaxonomic groups, which may be classes or families, according to the respective pigment compositions. In that way we used “Bacillariophyceae+Chrysophyceae”, because both families contain the marker pigments fucoxanthin and Chlc but other families of their class “Heterokontophyta” contain other marker pigments. Also, we used “Chlorophyta”, because all phytoplankton families of this class contain the same pigment composition, but we used “cyanobacterial picoplankton“ that clearly dominate the Baikalian cyanobacteria, because of their marker pigments zeaxanthin and caloxanthin, not prominent in filamentous cyanobacteria. Eustigmatophyceae were considered to be a common member of the Baikalian eukaryotic picoplankton, although they were not considered in previous phytoplankton studies (Kozhov 1963, Kozhova and Izmest’eva 1998, Popovskaya 2000), because their presence was suggested from peculiar pigment ratios (Fietz and Nicklisch 2004, Appendix A); furthermore, three new strains of Nannochloropsis limnetica (Eustigmatophyceae) were isolated from Lake Baikal recently (Fietz et al., submitted, Appendix B).
Duplicate samples for HPLC-aided pigment determination (1–2.5 L), were filtered through Whatman GF/F-filters with 25 mm diameter, put in 2 mL reaction vessels, immediately freeze-dried and stored frozen in the dark. Chlorophylls, carotenoids, and their derivatives were extracted in 2001 with 1 mL of a mixture of acetone, methanol and water (80:15:5 by volume, Leavitt et al. 1989) and in 2002 and 2003 with 1 mL of dimethylformamide under dim light at 4° C. No significant difference was found between both solvents. The extraction was done by vibration shaking at a frequency of 2000 min-1 with a supplement of glass beads (0.75-1 mm) over 1.5 h. An ionpairing reagent solution (15 g L-1 tetrabutyl ammonium acetate and 77 g L-1 ammonium acetate) was added 10:1. The extract was centrifuged for 20 min at 4° C at 2500·g in a cooled centrifuge (Biofuge Fresco Heraeus Instruments, Hanau, Germany). The separation, identification and quantification of pigments were performed according to Woitke et al. (1994) with a Waters HPLC system described by Fietz and Nicklisch (2004, Appendix A).
HPLC-aided pigment analysis was provided for the three summer cruises as well as for the intensive monitoring from May 2002 to June 2003. Additional weekly Chla data were provided from January 2001 to December 2003 by the long-term monitoring programme conducted by the SRIB. For these analyses water samples were taken at the aforementioned Bolshye Koti station, filtered through 0.7 mm pore size Nuclepore® polycarbonate filters, dried in cold, dark conditions and stored frozen. Extraction was done with 96 % acetone. Extracts were centrifuged and the absorbance of the supernatant was measured with a spectrophotometer at 750, 665, 645 and 630 nm. Chla was calculated according to guidelines given by the SCOR-UNESCO workgroup (1966).
During the CONTINENT summer cruises 2002 and 2003 as well as during the regular August cruise 2003 of the SRIB a series of fluorescence depth profiles were taken across the lake with a submersible fluorimeter (FluoroProbe, bbe Moldaenke, Kiel, Germany, Fig. 10) provided by the GFZ Potsdam (H. Oberhänsli). The FluoroProbe uses five excitation wavelengths and allows discriminating the contribution of the dominant groups (http://www.chlorophyll.de/; Beutler et al. 2001, -2002a, -2002b). Because the calibration at factory was not satisfying for Lake Baikal, the FluoroProbe was newly calibrated based on the HPLC and microscopic results (A. Nicklisch, Humboldt University, Berlin, Germany). Additionally to the depth profiles, which were taken at high resolution (c. 1 m) down to a depth of 120 m, a series of horizontal transects (resolution of 100 - 1000 m) were taken during the same cruises. The FluoroProbe was then connected to a drinking water pump on board, which pumped continuously water from c. 3 m water depth into the FluoroProbe.
|Fig. 10. FluoroProbe vertical measurements.|
Four moorings were deployed within the project (M. Sturm, EAWAG, Switzerland): During 16 months from March 12, 2001 to July 5, 2002, a sediment trap mooring, comprising 15 integrating traps (Fig. 11), was deployed in the centre of the South basin (51°42 N / 105°01 E), where water depth reached 1400 m. The traps were deployed at 40, 100, 255, 350, 445, 540, 635, 730, 825, 922, 1015, 1113, 1210, 1305, and 1396 m (cf. Müller et al. 2005). The same mooring string was deployed at the same site during the ensuing 12 months from July 6, 2002 to July 5, 2003. Another sediment trap mooring, comprising 9 integrating traps was deployed in the centre of the North basin (54°27 N/ 109°04 E) from July 9, 2001 to July 8, 2002 at 50, 255, 335, 445, 555, 720, 775, 885, and 903 m. Due to technical disturbances material from the uppermost two traps were lost in 2001-2002. The same mooring string was deployed at the same site during the ensuing 12 months from July 9, 2002 to July 9, 2003.
The integrating traps were two acrylic cylinders with each an active area of 65 cm2 and an aspect ratio of 1:8 (EAWAG-130, Ohlendorf and Sturm 2001). After recovery of the traps, the overlaying water was siphoned off and the sampling cups were covered with aluminium foil to avoid photo-degradation of the pigments. Duplicate subsamples of the suspended sediment trap material were filtered within two hours upon recovery through GF/F-filters (Whatman, Kent, UK) and immediately frozen. The filters were freeze-dried within 24 h upon recovery and stored frozen at -20° C in the dark until analysis.
|Fig. 11. Sediment trap deployment (left) and top traps after recovery (right)|
Chlorophylls, carotenoids, and their derivatives were extracted with 1.25 mL of dimethylformamide under dim light at 4° C. The extraction was done by vibration shaking with a frequency of 2000 min-1 over 3 hours. 125 µL of the aforementioned ion-pairing reagent solution (cf. chapter 2.2.3) were added. The extract was centrifuged for 20 min at 4° C at 5000·g in a cooled centrifuge (Biofuge Fresco, Heraeus Instruments, Hanau, Germany) and the supernatant was transferred in vials for HPLC-analysis. The separation, identification and quantification of pigments of all samples were done similarly to the water samples. The eluting peaks were monitored at 440 nm using a Waters 996 photodiode array detector and at 410/670 nm (excitation/detection wavelength) using a Waters 474 fluorescence detector.
Unialgal cultures, acidified cultures, cultures of Dunaliella tertiolecta (with high chlorophyllase activity), standards, and literature data were used for identification of chlorophylls and various degradation products. The distinction between Chla degradation products and Chlb degradation products was performed using the ratio of the fluorescence at 410/670 vs. 430/650 nm. According to Soma et al. (2001b) and Soma et al. (2003), the concentration of SCEa and b was determined assuming that their fluorescence at 410/670 nm and 430/650 nm (excitation/detection wavelength) was identical to their respective pheophorbides.
The terms Chlas or Chlbs mean the sum of all respective chlorophylls, epi- and allomers and degradation products. Chlc degradation products were not identified definitely. The same was true for carotenoids. Most carotenoid degradation products could not definitely be attributed to their parent carotenoid and thus, the description of the sedimentary carotenoids in the following will be limited to the intact carotenoids.
The total organic carbon (TOC) and total nitrogen (TN) contents of the sediment trap samples 2001-2002 were determined with a EURO-EA® CNS-analyser at EAWAG (M. Sturm, Switzerland). The TOC and TN contents of moorings 2002-2003 were determined with a Vario EL CHNOS elemental analyser (Elementar Analysesysteme GmbH, Germany) after acidification with 0.2 N HCL.
The main CONTINENT coring sites were chosen after intensive side scan sonar studies above preselected regions. High-resolution seismic data (M. De Batist and F. Charlet, Renard Centre of Marine Geology, Belgium), long core description (N. Fagel and F. Hauregard, University of Liège, Belgium) and physical property measurements (H. Oberhänsli, J. Klump, F. Démory, and P. Sorel, GFZ Potsdam, Germany) were performed to characterise the sedimentary environment of these main coring sites.
The three coring sites were located on elevated plateaus: Vidrino Shoulder, Posolski Bank, and Continent Ridge (Fig. 12). The Vidrino Shoulder, in the eastern part of the South basin, is composed of a series of elevated ridges, perpendicular to the coast, and separated by deeply incised channels (Charlet et al. 2005). The main material is medium to fine-grained sediment (Charlet et al. op. cit.). The core was taken on a flat crest of one of the ridges, where seismic data suggest a stable depositional environment (Charlet et al. op. cit.). The upper meter of the sediment was laminated and was composed of diatom-rich biogenic sediments. The uppermost 10 cm were oxidised. There were no visible signs of bioturbation. According to Martin et al. (2005) bioturbation was low in the three investigated recent sediments (Vidrino, Posolski, and Continent Ridge), but a possible disturbance should not be dismissed. The effect of biological mixing by oligochaeta was less important in the deepest stations (Continent Ridge and Vidrino) than in the shallower regions (Posolski).
|Fig. 12. Map locating the coring sites with lithological descriptions of the analysed segments.|
The Posolski Bank is a shallow plateau within the Selenga Delta Accommodation Zone. The core was taken from the central part that is characterised by undisturbed sedimentation and predominance of fine-grained sediments (Charlet et al. op. cit.). The upper 25 cm of the sediment is homogeneous and composed of biogenic muds. Below, the sediment is composed of a mixture of biogenic and terrigenic muds (Charlet et al. op. cit.). The sedimentation is predominantly hemipelagic; the terrigeneous input is low (< 30 %) at Posolski, despite the input from the Selenga River at that site (Charlet et al. op. cit.).
The Continent Ridge is located in the North of the Svyatoi Nos Peninsula near the eastern coast of the North basin. Within the Continent Ridge, the core was taken at a flat crest with featureless morphology, thought to represent undisturbed sedimentation (despite the tectonic activity of the rift system; Charlet et al. op. cit.). Fine-grained sediments prevailed (Charlet et al. op. cit.). The upper 10 cm were oxidised and were dominated by diatoms. Below, the sediment was composed of a mixture of biogenic and terrigenic muds (Charlet et al. op. cit.). The sedimentation is, like at Posolski, predominantly hemipelagic and the terrigeneous input is < 30 %. The sediment of the Kazantsevo Interglacial, which was taken from the Continent Ridge site only, was laminated, biogenic and diatom-rich (Charlet et al. op. cit.). The transitions between glacial and interglacial phases were clearly marked by the transition from grey to brownish (diatom-rich) colour of the sediment (Fig. 13).
|Fig. 13. Core CON01-603-3, segment 458-590 cm: shift from grey terrigenic material (glacial periods) towards brownish diatom-rich material (warm period).|
Short cores: During the coring campaigns in July 2001 and July 2002, a series of short (gravity) cores (60 – 70 cm long, covering all or parts of the Holocene) were taken using a gravity corer (EAWAG-63) with a PVC-liner (Ø 63 mm) (M. Sturm, EAWAG, Switzerland; E. Vologina, Institute of Earth Crust, State University Irkutsk, Russia). Cores were collected with intact water to sediment interfaces. For pigment analyses cores from ´Vidrino´, ´Posolski´, and ´Continent Ridge´ coring sites were sliced on ship under dim light in contiguous samples 0.5 cm thick for the upper 20 cm and 1 cm thick below.
Six additional short cores taken from the South basin, Selenga Delta, Barguzin Bay, and North basin, were dedicated to the study of the pigment degradation processes within the oxidised surface sediment. From these additional cores only the oxidised layers (up to 20 cm) were sliced into contiguous 1 cm thick samples. All short cores slices were freeze-dried on ship within 24 h upon slicing and stored frozen at -20° C until analysis.
Piston cores: Several piston cores were taken during the CONTINENT cruise in July 2001 using aluminium liners (Ø 120 mm; Fig. 14; D. Meischner, University of Göttingen, Germany). The piston cores had length of approximately 11 m. The piston cores were sealed, eliminating at a maximum headspace for gases, stored in the liner at ambient temperature during transport and at 4° C in the laboratory at GFZ Potsdam (Germany). For pigment analysis samples from piston core number CON01-603-3 (Continent Ridge) were taken from the centre of the liner, which should not have been affected by light or oxygen. From this piston core the segments W0006 and W0005 (corresponding to cm 428 – 578), covering the Kazantsevo and its transitions to the glacial periods, were sliced at each centimetre. All subsamples (0.4-1.5 g dry weight) were immediately frozen, freeze-dried within 7 days and stored frozen at –20° C until analysis.
|Fig. 14. Recovery of the c. 11 m long piston cores. Fotos from J. Klump (http://continent.gfz-potsdam.de/html/gallery).|
The analyses of pigments (HPLC) and TOC and TN in the sediment samples were performed according to the methods described in chapter 2.3.2 and 2.3.3 for sediment trap samples.
Surface sediment: The recent mass accumulation rates of the short cores were determined based on excess-activity of 210Pb, measured with a CANBERRA well-type γ-detector (M. Sturm, EAWAG, Switzerland). These rates are valid for the uppermost layer (approximately 150 years), but are difficult to be extrapolated into longer-term sedimentation rates, valid for depositional time scales of thousands years.
Holocene: According to AMS radiocarbon dating of pollen, the average sedimentation during the Holocene was 17.29 ± 0.39 cm kyr-1 at Vidrino, 11.6 ± 2.3 cm kyr-1 at Posolski, and 6.86 ± 0.21 cm kyr-1 at Continent Ride (Piotrowska et al. 2004). The 14C sedimentation rates were also confirmed by determinations of the magnetostratigraphy on parallel piston cores (Demory et al. 2005). Subsequent dates (gathered by AMS radiocarbon dating or magnetostratigraphy) are quoted as kyr before present (kyr BP) and refer to calibrated ages. However, as dating in Lake Baikal is a complex issue for a number of reasons (cf. Colman et al. 1996, Karabanov et al. 2000a) dating of the most recent Holocene periods should be considered as indicative rather than exact.
|Fig. 15. Density along core CON01-603-2 used for dating, for pollen and for diatom analyses and along core CON01-603-3 used for pigment analyses. Both cores were taken from the same site (Continent Ridge) in the North basin. The highlighted rectangle marks the segment where photosynthetic pigments were analysed. Pollen and diatom analyses showed that the Kazantsevo Interglacial in core CON01-603-2 was between 6.10 m to 7.20 m (Granoszewski et al. 2005, Rioual and Mackay 2005). According to the density, this section corresponded to the section between 4.80 m and 5.60 m in core CON01-603-3.|
Last Interglacial (Kazantsevo): The Kazantsevo segment has been determined within the 11 m long piston core based on the dating by magnetostratigraphy (Demory et al. 2005), on the lithology (Charlet et al. 2005), as well as on the high-resolution pollen and diatom records (Granoszewski et al. 2005, Rioual and Mackay 2005). These analyses have been carried out on a parallel core (CON01-603-2), whereas for pigment analyses core CON01-603-3 was used, because analyses of photosynthetic pigments required that the sediment had not previously been exposed to light or oxygen. Both cores were taken from the same site (Continent Ridge). High-resolution density analysis of both cores allowed locating the Kazantsevo Interglacial within the core CON01-603-3 (Fig. 15). Density was measured with a GEOTEK Multi-Sensor Core Logger at an interval of 0.5 cm with a counting time of 5 sec (P. Sorel, GFZ Potsdam, Germany). According to the diatom analysis the Kazantsevo spanned from 127.5 to 117 kyr BP (Rioual and Mackay 2005). According to the pollen analysis, the Kazantsevo spanned from 129 to 117.4 kyr BP (Granoszewski et al. 2005).
Variance analyses, Spearman-Rho and Pearson correlations, linear regressions, principal components analysis (PCA) and discriminance analysis were calculated with SPSS© (SPSS Inc., Chicago, IL, USA) statistical package. Canonical correlation analysis (CCA) was calculated with Statistica© (StatSoft Inc., Tulsa, OK, USA). Nearest neighbouring interpolation was performed using ArcMap© (ESRI Geoinformatik GmbH, Kranzberg, Germany). The Simpson index and Shannon-Wiener index were calculated with BioDap© (New Brunswick, Canada, cf. Magurran 2003) using cell abundances. Curve fittings were performed with TableCurve 2D® (Systat Software Inc., Point Richmond, CA, USA).
|© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.|
|DiML DTD Version 4.0||Zertifizierter Dokumentenserver|
der Humboldt-Universität zu Berlin