4 Discussion

↓79

Lake Baikal is known for a list of superlatives such as the World’s oldest, deepest and largest (by water volume held) lake, with a high number of endemic species. Shortly resumed, Lake Baikal represents an extreme in the global spectrum of freshwater types (Flower 1998). This lake is therefore object of major interest to biologists and paleoenvironmental scientists alike (Flower 1998). The understanding of the biological record contained in Baikal’s sediments is central to understanding not only past climate variability in the region but also impacts of recent environmental change and pollution (Flower 1998).

↓80

The main objective of the present thesis was to study the phytoplankton pigments as a tool for tracking environmental changes in Lake Baikal. Three main tasks were performed for this purpose: (1) the recent phytoplankton and photosynthetic pigments were analysed to improve the knowledge on their occurrence and formation in the euphotic zone; (2) the transfer of the pigments through the water column and the incorporation into the sediment were analysed to investigate degradation processes in the aphotic, but oxic water column and oxic or oxidised surface sediment; and finally (3) sedimentary pigments were analysed to prove their potential to record environmental changes that happened long ago.

4.1  Recent phytoplankton pigments in Lake Baikal as markers for community structure and environmental changes

Phytoplankton responds quickly to environmental changes and thus can be very useful for assessing changes in aquatic ecosystems. This is especially the case for Lake Baikal where monitoring has been carried out for more than 60 years (Kozhov 1963, Galazii 1993, Kozhova and Izmest’eva 1998, Popovskaya 2000, Goldman and Jassby 2001 and references therein). However, up until now, few multiparameter studies have been conducted in parallel of regional, vertical and seasonal variability. Also, studies conducted in different decades have had to be used to estimate the varying importance of the autotrophic picoplankton (APP; cf. Kozhova and Izmest’eva 1998, Popovskaya and Belykh 2004). Phytoplankton pigments other than Chla were not yet investigated. Here, a pigment-based approach complemented with traditional microscopic counts was used to determine phytoplankton spatial and seasonal distribution patterns over a three year-long period (2001-2003).

4.1.1  Regional phytoplankton and pigment distribution and driving factors

Autotrophic picoplankton: Former studies already indicated the importance of cyanobacterial and eukaryotic APP in the South (Votintsev et al. 1972, Nagata et al. 1994) and Centre (Nakano et al. 2003) as well as at near-shore or river delta stations (Boraas et al. 1991), but only few studies compared the APP distribution in the different regions (cf. Popovskaya and Belykh 2004). Most of these former studies used light and a few used epifluorescence microscopy (cf. Popovskaya and Belykh 2004). The broadest study on APP was recently performed by Belykh and Sorokovikova (2003) using scanning electron microscopy. They reported an APP abundance that varied by 2 to 10 times in different parts of the lake. Based on light-microscopic long-term data from spring 1964 to 1990, Popovskaya and Belykh (2004) reported dominance of cyanbacterial APP (Synechocystis limnetica) in the North, which was not confirmed in our study in summers 2001-2003 as the APP contribution to total biovolume was 3-6 times lower in the North than in the South and Selenga Delta. However, Popovskaya and Belykh (2004) reported furthermore that during late summer 2000, using scanning electron microscopy highest numbers of cyanobacterial APP were registered in the South.

↓81

The marker pigment zeaxanthin allowed the contribution of cyanobacterial APP to the total Chla to be calculated and the results showed that the contribution to total Chla was also 4-6 times lower in the North than in the South and Selenga Delta. Based on the contribution to total Chla data and environmental data, canonical correlation analysis (CCA) could be calculated to define the environmental conditions that most probably impact the phytoplankton composition. This CCA showed that latitude is negatively correlated to the cyanobacterial APP contribution. Both the APP contribution to total biovolume and the contribution to total Chla contrasted, therefore, suggestions of Popovskaya (2000) that APP is generally most prominent in the North.

In most marine and freshwater systems the contribution of APP to the total biovolume and production was inversely correlated to increasing trophy (Stockner 1991, Callieri and Stockner 2002). This general concept might not be adopted for Lake Baikal, assuming higher trophy in the deltas and oligotrophic conditions in the pelagic regions of the lake (Callender and Granina 1997, Genkai-Kato et al. 2002). The CCA indicated that water depth (and thus the distance to the shore, which may be indicative for the trophic state of the regions) was of minor influence for the cyanobacterial APP contribution. In growth experiments with Synechocystis limnetica,it was concluded that temperature is the major driving force (Richardson et al. 2000). We now can specify that stratification was the dominant factor for increasing cyanobacterial APP contribution.

Nano- and microphytoplankton: The nano- and microphytoplankton biovolume varied between the basins and also between years, but not significantly. A general increase of the biovolume from North to South, as reviewed by Goldman and Jassby (2001), could not be confirmed in 2001 and 2002. It could also not be confirmed that the regions with shallow waters and with river water input, such as the Selenga Delta showed highest biovolumes (even including the APP), as mentioned in former comparative studies (Bondarenko et al. 1996, Popovskaya 2000). Therefore, if eutrophication occurs in the Selenga Delta due to anthropogenic impacts (Popovskaya 2000) it might affect the phytoplankton composition rather than the total biovolume.

↓82

Kozhova (1987) noted for river-estuary regions enrichment by the flora of the tributaries. These species might not persist as a diversity increase could not be found in the delta. However, one sign of changing phytoplankton community was the mass development of the N2-fixing Aulosira sp., which indicated a possible N-limitation in the Selenga Delta. Low N/P ratios (14) were found in the Selenga Delta in preliminary nutrient measurements in July 2003, whereas the ratios in the open basins varied between 25 and 36 (V. Straškrábová and J. Borovec, HBI Ceske Budejovice, Czech Republic, personal comm.). Increasing P-load with the tributaries (Callender and Granina 1997) was supposed to create the N-limitation (Goldman and Jassby 2001).

Phytoplankton pigments: As has been pointed out by Kozhova et al. (1985), it is impossible to delineate a region within the pelagic basins characterised from year to year by constant higher or lower Chla. In fact, in the North for example, each year showed significantly different Chla concentrations. While lowest Chla concentrations in 2001 were found in the Centre, they were found in the North in 2002 and 2003. Nevertheless, the combined data set (July 2001+2002+2003) showed significant differences for Chla and the sum of carotenoids, as well as for several marker pigments. For example, significantly higher Chla concentrations in the South compared to the Centre and North were found.

Besides variations among the open basins, significantly higher Chla concentrations were found at the river inflows. Nutrient enrichment could be assumed to trigger Chla increase, but the CCA suggested instead a higher correlation with stratification. Stratification was enhanced in the delta because of the warm river water inflow and because the shallow water zone warmed up faster than the deep water basins. The role of eutrophication might, therefore, be secondary up to now for total phytoplankton abundance. The development of the N2-fixing Aulosira sp. (as aforementioned) indicated, however, enhanced P-loading from the Selenga River and thus a possible change in species composition might be expected if nutrient loading further increases.

↓83

Marker pigment analysis showed significant changes of the phytoplankton community within the open basins that could not be stated as statistically significant using the limited capacity of phytoplankton counts, showing the benefit of fast techniques for large sample sets. For example, marker pigments of the Chlorophyta, cyanobacteria and Eustigmatophyceae decreased significantly from South to North, while those of Bacillariophyceae+Chrysophyceae remained relatively constant among the three open basins. Marker pigments are also potential indicators of varying environmental conditions in regions where total biovolume, Chla and sum of carotenoids did not show significant differences. For example, Chla and total carotenoids were both high in the Selenga Delta and Barguzin Bay suggesting similar environmental conditions. Nevertheless, the marker pigments of Bacillariophyceae+Chrysophyceae fucoxanthin and Chlc were significantly higher in the Barguzin Bay than in the Selenga Delta, whereas the marker pigments of cyanobacterial APP zeaxanthin and ß-carotene were significantly lower. Therefore, different environmental conditions are likely at the two sites. Discriminance analysis enhanced the assumption of particular phytoplankton composition at some sites such as Barguzin Bay, Academician Ridge, and Maloe More.

Estimation of chemotaxonomic group contribution: In a previous, preliminary study based on the data set gathered in July 2001, it was shown that the share of selected chemotaxonomic groups could be estimated in Lake Baikal using conversion factors determined by multiple linear regression and by the CHEMTAX matrix factorisation program (Fietz and Nicklisch 2004, Appendix A). New factors calculated from the whole data set 2001–2003 could add a fifth group, i.e. Bacillariophyceae+Chrysophyceae, Cryptophyta, Chlorophyta, Eustigmatophyceae and cyanobacterial APP. The presence of Eustigmatophyceae was proven in another study (Fietz et al. submitted, Appendix B).

The estimation of the relative compositions and distributions of the distinct chemotaxonomic groups done in the preliminary study (Fietz and Nicklisch 2004, Appendix A) could be confirmed now with the more extensive dataset. However, although the factors calculated in the present study for Chlorophyta (Chlb, lutein) and Eustigmatophyceae (violaxanthin) did not differ significantly from those reported for 2001 (Fietz and Nicklisch 2004, Appendix A), the factors for Bacillariophyceae+ Chrysophyceae (fucoxanthin, Chlc) as well as for cyanobacterial APP (zeaxanthin) were significantly lower than those reported for 2001. Hence, the Bacillariophyceae+ Chrysophyceae community and the cyanobacterial APP changed in 2002 and 2003 towards cells with higher amounts of accessory pigments (fucoxanthin and Chlc) per Chla. Most of the ratios fitted also well to those of isolated Baikalian strains (Fietz and Nicklisch 2004, Appendix A).

↓84

High-resolution analysis of phytoplankton community structure changes using fluorometry: Horizontal fluorescence transects confirmed the great variability of phytoplankton abundance (given here by the total Chla concentration). Previous continuous fluorescence records were performed in spring 1990 (Granin et al. 1991). Nonetheless, the probe measured Chla only and the authors were therefore unable to relate Chla maxima to specific phytoplankton groups. The Moldaenke FluoroProbe used in this study detected the emitted fluorescence at five wavelengths and allowed calculating the contribution of preselected phytoplankton groups to the total Chla. However, information on dominant phytoplankton groups and their pigment characteristics was required to calibrate the FluoroProbe accurately. Today calibration has been performed for three dominant groups (A. Nicklisch, HU Berlin, Germany): Bacillariophyceae+ Chrysophyceae+Pyrrophyta (with similar fluorescence characteristics), Chlorophyta and phycoerythrin-containing cyanobacteria.

The respective transects enhanced the assumption of great heterogeneity in the phytoplankton composition. For instance, within a transect from Ushkanin Islands along the coast of Svyatoi Nos Peninsula into the Barguzin Bay, the contribution of Chlorophyta increased strongly, while those of Bacillariophyceae+Chrysophyceae+ Pyrrophyta and cyanobacteria decreased (Fig. 49). Another transect from the port of Listvianka (west coast of South basin) to the North of Selenga Delta showed how distinct phytoplankton groups were responsible for small scale (few kilometres wide) peaks (Fig. 49). This latter transect also indicated the changing phytoplankton composition from dominances of Chlorophyta and Bacillariophyceae+Chrysophyceae+ Pyrrophyta to cyanobacteria in the Selenga Delta region (Fig. 49), as has been assumed from discrete water sample analyses (A. Nicklisch, HU Berlin, Germany, pers. comm.). These profiles indicated the potential of in situ fluorescence measurements that, in regular monitoring, may complement laboratory studies of microscopic counts and HPLC-aided pigment analyses.

Fig. 49. Horizontal fluorescence transects showing total Chla (black), contribution of Bacillariophyceae+Chrysophyceae+Pyrrophyta (brown), Chlorophyta (green) and cyanobacteria (blue). Data provided by A. Nicklisch (HU Berlin, Germany) from unpublished data sets.

4.1.2 Pigments as markers for vertical changes during stratification and homothermy

↓85

Regional and daily varying wind, insolation, and stratification have been shown to strongly influence the phytoplankton vertical distribution (Bondarenko et al. 1996, Kartushinsky 1997). Below the ice, the homogeneous layer was due to convective flow-fields (Kelley 1997, Granin et al. 1999). This layer, which may vary between 10 m (Popovskaya 2000) and 50 m (Zavoruyev et al. 1992), was restricted to the upper 25 m in 2001-2003 (Fig. 24). The euphotic zone under the clear, snow-free ice in Lake Baikal in 2003 was 1.7·Secchi depth (Straškrábová et al. 2005) and therefore a 15-35 m thick euphotic zone can be assumed. The convective layer under the ice was then not deeper than the euphotic zone. In spring 2001, 2002 and 2003 Chla concentrations up to 0.5 nmol L-1 reached 250 m and concentrations over 2 nmol L-1 reached 100 m. Thus, cells were obviously transported out of the zone of maximal productivity (euphotic zone).

In summer, the phytoplankton is concentrated in the upper 25 m but even during stratification Chla was found down to 100 m in 2001 and 2002 (Fig. 24). Genkai-Kato et al. (2003) found that cells collected during mixing as well as during stratification in the deep water of Lake Baikal (500 m) were able to photosynthesis when exposed to surface levels of irradiance. They suggested that live Bacillariophyceae, remnants from the spring community, sank out to greater depths during stratification, which was supported by unpublished taxon depth profiles of our long-term monitoring.

However, even during homothermy the Baikalian pelagial is not homogenous. The South is ice-free many weeks before the North and thus is stratified earlier. Warm water inflows from rivers, such as the Selenga and Barguzin, also enhanced the stratification locally (Fig. 26). Primary productivity and biomass increased strongly at stratified stations in summer 1990 (Goldman et al. 1996). According to the present data, the higher production was due to Chlorophyta and APP, both highly correlated to stratification.

↓86

As aforementioned, the Moldaenke FluoroProbe used in this thesis allowed estimating the contribution of selected phytoplankton groups to the total Chla. These profiles clearly showed that in the South the Chla maxima were formed mainly by Chlorophyta, while those in the Selenga Delta were formed mainly by cyanobacteria (Fig. 50; A. Nicklisch, HU Berlin, Germany, pers. comm.). At Academician Ridge the Chla maximum was formed in the upper layer by cyanobacteria, while the lower part was formed by Chlorophyta and Bacillariophyceae+Chrysophyceae+Pyrrophyta (Fig. 50; A. Nicklisch, pers. comm.). According to Granin et al. (1991) early summer warming of the water induce two phytoplankton maxima: one near the photic zone and the other in greater depths, due to the formation of two “convective cells” each with high internal turbulent exchange. The phytoplankton in the upper “convective cell” will develop, while the phytoplankton population in the deeper “convective cell” will decline, getting light insufficiency. The authors assumed that smaller phytoplankton is likely to form the upper maximum, while larger ones (such as Aulacoseira) are more likely to form the deeper one. With further warming the larger ones will decline. Such convective cells may have developed at Academician Ridge. However, because calibration is not yet accomplished discussion about would be rather speculative. Nonetheless the preliminary high-resolution depth profiles of phytoplankton community structure highlight the usefulness of multiwavelength fluorometric measurements in such large and deep lakes, where dramatic changes of the regional and vertical phytoplankton distribution occur.

Taken together, the present data of regional differences in the depth profiles confirmed Granin’s (Granin et al. 1991) conclusion that mesodifferences of the hydrological-climatic conditions determine the meso-inhomogeneities in the spatial distribution of water temperature, depth of convection development at a certain time, chlorophyll content, phytoplankton productivity and distribution of the physical transparency of the water.

Fig. 50. FluoroProbe depth profiles (lines) showing total Chla as well ascontributions of selected phytoplankton groups to the total Chla. Measured (by HPLC) total Chla and contributions are given as well as the estimated total Chla based on marker pigments (see Appendix A and chapter 3.1 for marker pigment based calculations). Letters in quotations refer to Fig. 18. Data provided by A. Nicklisch (HU Berlin, Germany; from unpublished data set).

4.1.3 Seasonal dynamics of phytoplankton and pigments and driving factors

↓87

The bacillariophycean spring development began beneath the ice. The spring bloom 2002 was founded on Aulacoseira baicalensis and Stephanodiscus meyerii. Thus, although the spring peak 2002 did not reach the biovolume of real “Melosira”-years, it was based on two formerly called “Melosira” species (Kozhova and Kobanova 2000). The dominance of large Bacillariophyceae in spring is not conform with the generally agreed PEG-model for freshwater lakes, which would assume a spring crop of small, fast growing algae such as Cryptophyta and small centric Bacillariophyceae (Sommer et al. 1986). Such dominance of large Bacillariophyceae was found in only five out of 18 compared lakes, and four of these five were stratifying, temperate, eutrophic lakes or reservoirs not deeper than 34 m (references in Sommer et al. 1986). The CCA illustrated that Bacillariophyceae+Chrysophyceae in Lake Baikal were negatively correlated to temperature for seasonal succession as has been suggested from culture experiments with Aulacoseira baicalensis (Richardson et al. 2000). In summer A. baicalensis was found in greater depths only. This species sank out through the summer stratified water column to cooler waters, probably to avoid hotter summer temperatures. Nonetheless, in this study we cannot definitely determine whether A. baicalensis was limited by temperature or by competition with other phytoplankton groups.

The decline of biovolume and Chla (the ´clear-water´ phase) in June might be due to intense grazing. Consistent with the PEG-model, edible Cryptophyta became dominant, as well as small Chrysophyceae. In summer the ratios of pigments vs. biovolume of the Bacillariophyceae and Chlorophyta decreased indicating a shift towards smaller, pigment-rich cells with increasing temperatures. During summer, functional groups indicated summer stratification in oligo- to mesotrophic conditions (E+X, Reynolds 2002). Contrary to the PEG-model, which predicts growth of non-edible algae, an explosive growth of edible algae, such as eukaryotic and cyanobacterial APP took place in Lake Baikal.

The importance of APP varied with the season with a maximum in summer and a minimum under the ice in former years (Moskalenko 1971, Goman 1973, Belykh and Sorokovikova 2003, Popovskaya and Belykh 2004). APP epifluorescence and light microscopic estimations supported these trends for 2002-2003. Belykh and Sorokovikova (2003) noted a small APP peak in April of each year from 1997 to 2000, which could in fact be found again in 2003. According to the CCA the seasonal cyanobacterial APP development might be triggered by temperature and stratification. Light limitation due to mixing (>100 m) much below the euphotic zone (<40 m) might depress the APP formation at the time of maximal homothermy after ice-break up and in autumn. However, during inverse stratification under the ice APP contributed surprisingly large amounts to the total Chla (c. 9 %) as well as to the total primary production (up to 40 %, Straškrábová et al. 2005).

↓88

High fucoxanthin/biovolume and Chlc/biocolume ratios (Fig. 30) indicated a summer development of small pigment-rich Bacillariophyceae or Chrysophyceae cells besides the development of picoplanktonic cells. This summer development of small cosmopolitans at Bolshye Koti enhanced the suggestion of the shift within the phytoplankton community attributed to global warming (Popovskaya 1991, -2000, Mackay et al. 1998, Bondarenko 1999). However, the summer communities in the open basins were still dominated by endemics (Cyclotella baicalensis, ornata, and minuta) and therefore, we may claim that fortunately, until now, the warming or eutrophication possibly affects the nearshore regions, such as Bolshye Koti (where the seasonal monitoring was conducted), but that the open basins still remain unaffected, due to the huge water masses.

Those APP were replaced by large Bacillariophyceae only towards autumn. A regular autumn maximum has been described for Lake Baikal (Popovskaya 2000), which was found in 2001 and 2003 (Fig. 30A). Its absence in 2002 might be due to strong winds, which were noted in this year and which induced a transport into the deeper, aphotic layers. Consistent with the PEG-model was a reduction of light energy input, which caused a decline of the total biovolume in early winter.

The maximum of the nano- and microphytoplankton biovolume found in February/March 2003, when ice cover was 0.8 m thick and almost free from snow, could be due to convection under the ice when solar radiation warmed the near-surface water. Then voluminous, non-motile Bacillariophyceae can be maintained days or even months near the surface providing cells withenough light for growth (Kelley 1997, Granin et al. 1999). Those under-ice blooms are often related to Aulacoseira, Gymnodinium or recently Nitzschia acicularis (Popovskaya 2000, Bondarenko 1999), but in 2002 the dominant species were Asterionella formosa and Synedra acus.

↓89

Aulacoseira baicalensis, the classic example of a diatom able to bloom under spring clear ice (Kozhova and Izmest’eva 1998), was surprisingly insignificant before ice-break up in 2003. That could be due to the snow cover and the consequent stopping of light penetration, which could have inhibited the building of the convective layer, essential for its maintenance in the euphotic zone. A. formosadominated this site for the first time since the beginning of the long-term monitoring at Bolshye Koti. According to the regional distribution A. formosa was localised at Bolshye Koti, as it wasn’t found in abundance elsewhere. It may be that this species is an opportunistic taxa filling a niche where available but never dominating the whole lake. Its mass development at Bolshye Koti was probably a result of multiple asexual reproductions (Kobanova and Izmest’eva 2003).

4.1.4 Remote sensing

Optical remote sensing analyses base on the light leaving the water surface (differential absorption and backscatter of irradiance inside the water), corrected for atmospheric and surface effects (e.g. caused by air molecules, aerosols, and the lake surface itself; Heim et al. 2005). Jeffrey et al. (1997) predicted that because pigments are the only biological parameter measurable from space, pigments will serve as basic or proxy parameter for global mapping of components of the ocean carbon cycle. However, the Ocean Colour research is well developed for the World’s oceans, but is still in development for coastal zones and today, investigations on large oligotrophic inland water bodies have rarely been carried out (e.g. EEGLE – Episodic Events-Great Lakes Experiment Understanding the Historical Magnitude of Spring Turbidity Plumes in Southern Lake Michigan, and KITES – Keweenaw Interdisciplinary Transport Experiment in Superior) in the North American Great Lakes (e.g. Bergmann et al. 2004).For Lake Baikal, there was another chance to conduct an optical remote sensing study by using the interdisciplinary CONTINENT field data set, of which the most important has been the pigment data set presented in this thesis (Heim et al. 2005; Heim, in prep.).

Semovski (1999) first conducted preliminary remote sensing studies of Lake Baikal’s water constituents with AVHRR satellite data in the visible wavelength range, which did, however, not result in an applicable Chla algorithm. Within the CONTINENT framework global Chla algorithms, and Chla algorithms for oligotrophic systems (e.g. Iluz et al. 2003) were evaluated using the here presented pigment data set (cf. Heim et al. op. cit., Heim op. cit.). In addition, blue-green ratio Chla algorithms were calculated by linear regression using the Chla pigment data sets 2001 and 2002 (cf. Heim et al. op. cit., Heim op. cit.). However, the most robust performing algorithm for the SeaWiFS acquisitions in 2001 and 2002 was found with the Ocean Colour “OC2” Chla algorithm (O’Reilly et al. 1998) that represents a blue-green ratio algorithm based on a global pigment data set (830 cases); by evaluating the present pigment data set, accuracies below 35 % for pelagic waters covering all bio-optical provinces were achieved(cf. Heim et al. op. cit., Heim op. cit.) meeting NASA quality standards (O’Reilly et al. 1998).

↓90

The remotely sensed data provided continuous records of the Chla concentration in the upper photic zone (as a measure of the productivity) at nearly each point of the cloud free parts of satellite image (cf. Heim et al. op. cit., Heim op. cit.). The images provided important information of the formation of eddies and patches, as well as of the differential formation of phytoplankton abundance and community structure within the different basins (Fig. 51). For example, the North is shown to be more oligotrophic than the South in summers 2001 and 2002 (Fig. 51), but small, pigment-rich eddies were also revealed in the North, which can easily be overlooked even with hundreds of water samples in a 600 km long lake.

Fig. 51. Quasi-true colour SeaWifs for 6 August 2001 (Heim pers. comm.). The blue colour indicate more oligotrophic water; turquoise colouring indicate cyanobacterial picoplankton-rich water; yellow and brownish colours indicate terrigenous and more productive regions.

Based on these satellite images the differential seasonal phytoplankton variation, represented by the total Chla, could be established (Fig. 52). The results of which showed a much higher summer maximum in the South than in the North and Centre for example. This difference should be kept in mind when seasonal data from the Bolshye Koti site are extrapolated to other areas of the lake. However, it has to be admitted here, that these data are still preliminary, because in regions, which are optically influenced by terrigeneous matter, e.g. in the Selenga Delta, there is considerable Chla overestimation (3 to 5 fold) due to additional absorption processes compared to the measured Chla in July.These areas had to be masked and were excluded from the remote sensing Chla data set. Nonetheless, available SeaWiFS satellite images combined the regional and the seasonal monitoring, which was not possible by direct sampling or measurements due to the extreme size of the lake. These results prove the usefulness of phytoplankton pigment detection from space for regular monitoring in large lakes.

↓91

Fig. 52. SeaWiFS Chla time series (June-September 2001and 2002). Diamonds represent the ultra-oligotriophic gyre in which the northern mooring was situated; rectangles represent the Continent Ridge coring site; triangles represent Academician Ridge; crosses represent the Central basin; stars represent the Posolski area (Selenga Delta); circles represent the central South basin in which the southern mooring was situated. (B. Heim pers. comm.)

4.1.5 Does a pigment-based approach accurately monitor phytoplankton community and environmental changes in Lake Baikal’s euphotic zone?

Using a combined methods approach of pigment-based analyses and measurements complemented with traditional microscopic counts, this investigation of regional, vertical, and seasonal distribution patterns has provided a broad overview of the recent phytoplankton community structure of Lake Baikal. A for freshwater exceptional broad range of natural changes of the total phytoplankton abundance and composition could be proven. While the total biovolume did not show significant changes, total Chla, which is closely related to primary production, revealed impacts of the river inflow and shallow strait regions. Marker pigment changes revealed differences between regions, where biovolumes and Chla were not significantly different, e.g. between Selenga Delta and Barguzin Bay. Thus, phytoplankton pigments reflected the variability of primary productivity (measured as total Chla) and phytoplankton community composition (via marker pigments) in Lake Baikal’s euphotic zone.

Factorisation and ordination of the broad pigment sample set and environmental variables provided further insights into the driving forces. Temperature and stratification were shown to have major impact on the composition of the phytoplankton community structure. Different stratification regimes within a lake are rather unusual and set Lake Baikal apart from other freshwater systems. We can expect that a possible long-term warming that affect the peculiar stratification regime, would lead to significant changes in the phytoplankton group and species composition towards smaller, pigment-rich cells such as small diatoms up to picocyanobacteria at the expense of the large endemic diatom flora that prevailed up until now.

↓92

Recently, palaeoecological analysis of preserved markers such as diatom valves or photosynthetic pigments is increasingly used to monitor environmental change in response to climate and human activities. Insights into the driving forces will aid interpretation of sediment formation in this ancient lake.

4.2  Phytoplankton pigment transfer through the water column and preservation within the surface sediment

The aim of the second task within this thesis was to determine transfer functions for the organic matter, especially lipophilic photosynthetic pigments, to determine the degradation processes within the water column and surface sediment and to infer how the main phytoplankton groups were represented in the deposited material.

4.2.1  Fluxes in Lake Baikal compared to marine and freshwater systems

The mass fluxes of dry matter as well as of Chla, pheopigments a and carotenoids in Lake Baikal corresponded well to those found in different oceanic regions (Welschmeyer and Lorenzen 1985a, -b, Landry et al. 1995, Barlow et al. 1995, Nodder and Gall 1998). Fluxes in the moderatively productive marine Dabob Bay, in contrast, were much higher (Welschmeyer and Lorenzen 1985a). Organic carbon fluxes at the bottom of Lake Baikal (0.9-1.4 km) were within the highest ranges or higher than those reported for deep moorings (0.5-4 km) in open ocean areas (Ittekott 1996, Lampitt and Antia 1997). Due to its extreme depth and extension of the euphotic zone, organic matter fluxes in Lake Baikal should be compared with marine rather than with freshwater systems. Nonetheless, mass and pigment fluxes corresponded also to different oligo- to mesotrophic lakes; eutrophic lakes, in contrast, showed higher rates, but the depths of those lakes varied from 3 to 36 m only (Baines and Pace 1994, Poister et al. 1999, Hurley and Armstrong 1991). Little is known about the pigment or organic matter flux in deep lakes, because the study is limited to few lakes, such as Lake Tanganyika (~1450 m) or Caspian Sea (~1000 m), whereby the latter is not a freshwater system. Lake Baikal differs also from lakes such as Lake Tanganyika by its oxic bottom water and surface sediment, its heterogeneous stratification and low temperatures (<15° C in the pelagial). Anyway, no transfer functions are known for organic material or photosynthetic pigments in these comparable deep lakes.

4.2.2 Degradation processes in the water column of the South basin

↓93

The permanent temperature of 3-4° C below 250 m and the higher temperatures of 12-15° C during summer stratification in the epilimnion (Kozhova and Izmest’eva 1998) bring Lake Baikal closer to oceanic systems. The low temperatures might depress the degradation rates relative to shallower and warmer aquatic systems, because most degradation processes (photooxidation, grazing etc.) are temperature dependent (Leavitt 1993). On the other hand, Lake Baikal is oxygenated down to the maximal depth due to lake overturn, convections and deep-water currents (Weiss et al. 1991) and enhanced oxidation might occur across the whole water column.

Fig. 53. Schematic diagram of degradation in the water column from the euphotic zone to the sediment surface.

In southern Lake Baikal, the most prominent degradation of the settling material occurs within the upper 250 m water column, which is the wind-mixed depth during overturn, where particles are suspended (Müller et al. 2005). The degradation below 250 m was very low for most pigments. The biphasic character of the flux curves (Appendix C - Tab. 2. Appendix C - Tab. 3, Fig. 32, Fig. 33) clearly highlights that the degradation is divided into a stronger and weaker degradation phases. Only 24 % of the trapped Chla at 40 m settled deeper than 250 m. These low rates indicated that the initially settling Chla was transformed into pheopigments or colourless compounds. Possible causes could be (1) photooxidation due to extended residence in the euphotic zone during mixing, (2) death of living or moribund settling cells in the dark during stratification or (3) zooplankton grazing and further bacterial destruction.

↓94

Photooxidation due to extended residence in the euphotic zone by wind induced turbulence might explain the losses between the suspended and the settled matter, but it can only affect the uppermost trap, as the euphotic zone is limited to approximately 40 m (Kozhova and Ismest’eva 1998). Thus, the loss of Chla between 40 and 250 m should be caused by the death of settling cells in the dark or grazing. The loss of chlorophyllide a (which may represent moribund cells) may be caused by autolysis, bacterial destruction or also grazing. The much stronger decrease of pheophorbide a between 40 and 250 m almost certainly results from mesozooplankton faecal pellets being destroyed.

Death of living or moribund cells in the dark might be less common in Lake Baikal than in other aquatic systems. For the Dabob Bay and Central Pacific Gyres (Welschmeyer and Lorenzen 1985a, -b) as well as for a marine convergence zone (Head and Horne 1993) an over 10-fold higher average pheopigment a than Chla flux was found. As Chla is associated to the flux of intact cells, sinking of intact cells is an insignificant loss term in those areas. In southern Lake Baikal, in contrast, the pheopigment a flux was only three times the Chla flux, and it can therefore be assumed that sinking of living cells is important. The importance of the sinking of living cells can be explained by the high inherent sinking rates of “heavy” Bacillariophyceae as bacillariophycean sinking rates of 60-100 m d-1 have been reported in Lake Baikal (Ryves et al. 2003).

Pigment destruction by zooplankton grazing is a complex issue: rates of transformation and degradation depend on gut passage time, and hence of edible cell concentration and even on food quality. It has been shown that micro- and mesozooplankton can degrade Chla in part or even entirely into non-fluorescent breakdown products (Klein et al. 1986, Burkill et al. 1987, Barlow et al. 1988, Head and Harris 1992). Grazer size implies differences of faecal packaging and sinking rates. Faecal pellets of mesozooplankton, such as cladocerans, have high sinking rates, whereas faecal debris of microzooplankton, such as protozoa, has negligible sinking rates (Welschmeyer and Lorenzen 1985a). High sinking rates tend to prevent photooxidation, while low sinking rates lead to long permanence within the euphotic zone and therefore to a conversion into colourless products. Pheopigments found in the traps may originate from mesozooplankton rather than from microzooplankton, since they were found in the traps but not in the water samples, and hence they may originate from fast sinking faecal pellets (Welschmeyer and Lorenzen 1985a).

4.2.3 Composition of the settling material in the South basin

↓95

The relative contribution of the various accessory pigments can be used to infer various algal groups as sources of sinking material. The contributions of dominant phytoplankton groups could be determined with appropriate factors for the respective marker pigments. However, these calculations showed several limits, when applied to the traps.

On the one hand, the total Chla concentrations in the sediment traps could not be accurately estimated based on the factors determined for the water samples. It has been suggested that the overestimation of Chla resulted from a stronger degradation of carotenoids or Chlb than Chla. Carotenoids are more stable than chlorophylls in the presence of light and oxygen (Leavitt and Findlay 1994) and in the presence of grazers (Strom et al. 1998), but in oxic surface sediments of Lake Baikal carotenoids were more susceptible to decomposition than chlorophylls (Soma et al. 2001a). Further, fucoxanthin, the predominant carotenoid, easily decomposes and only zeaxanthin, which contributes only a small amount to the total Chla, is known to be stable (Leavitt and Findlay 1994). The differential chlorophyll and carotenoid degradations might therefore not be applied from one aquatic system to another.

On the other hand, Bacillariophyceae+Chrysophyceae dominated the sinking material at 90 % and Chlorophyta and cyanobacterial picoplankton made only minor contributions. However, Chlorophyta and the cyanobacterial picoplankton were also important contributors to the standing crop at least in summer (cf. Fietz and Nicklisch 2004, Appendix A, and chapter 3.1) and a dominance of 90 % is also not common during most of the year and hence, even the rough comparison with long-term studies (Kozhova and Ismest’eva 1998, Popovskaya 2000, chapter 3.1.3), indicate that a 90 % dominance of Bacillariophyceae+Chrysophyceae marks a group-specific sedimentation. This overestimation may be caused by the fact that fucoxanthin and Chlc originated sometimes from very large Bacillariophyceae (cf. Fietz and Nicklisch 2004, Appendix A, and chapter 3.1) which have very fast sinking rates (60-100 m d-1, Ryves et al. 2003)so that its pigments suffered less from light induced degradation processes than pigments of the smaller Chlorophyta and the cyanobacterial picoplankton. Selective grazing might also be an important factor, as it can also be assumed that autotrophic picoplankton is preferentially grazed by microzooplankton and mesozooplankton, whereas large siliceous Bacillariophyceae suffered less from zooplankton grazing (Hurley and Armstrong 1990).

↓96

Hence, the record of the phytoplankton standing crop by trapped pigments in southern Lake Baikal was group-specific. Some carotenoids are more preserved than others, and they cannot record whole algal assemblages, unlike differential degradation rates be taken into account. This has been found also for several shallow lakes in surface sediment studies (Leavitt and Carpenter 1990, Leavitt and Findlay 1994, Leavitt 1993). Similar problems are also well known from established methods such as bacillariophycean valve based analyses. Battarbee et al. (2005) also warn that differential dissolution of diatom species occurs mainly at the water to sediment interface.

4.2.4 Comparison of the sedimentation and degradation between South and North

Since the southern and northern mooring sites were located about 400 km away from each other and were characterised by different morphometry, ice cover, length of vegetation period, temperatures (Shimaraev et al. 1994) and productivity (Kozhov 1963, Kozhova and Izmest’eva 1998), differences of the pigment fluxes and of the pigment degradation were expected.

The Chlas fluxes recorded in the top traps were greater in the South than in the North in the first year, but lower in the second year. The fact that the first South mooring was deployed during 16 months, instead of 12 months, because of the inaccessibility of the site (thin ice in 2002), and, therefore, recorded two spring diatom peaks (M. Sturm, EAWAG, Switzerland, unpublished results from sequential traps), could have caused the higher fluxes therein. However, also in the North the flux was twice as high in the first year than in the second. Thus, the interannual variability of pigment deposition in the upper traps overlay the regional impact.

↓97

In contrast to the top traps, significantly lower Chlas fluxes into the bottom traps in the North than in the South were found. Within same traps much lower sterol and fatty acid fluxes were found at the lake bottom in the North than in the South as well (Russell and Rosell-Mélé 2005). As there was no definite regional impact found for Chla or Chlas fluxes in the upper traps, the regional impact might act during settling below 40-50 m water depth.

The pigment flux in the South decreased in two phases (two-exponential regression model) with a strong degradation in the upper hundred meters and a much weaker below, while the decrease in the North occurred with a constant degradation rate throughout the whole water column (one-exponential regression model). Processes within the upper wind-induced mixed layer, as they were described in Fietz et al. (2005, chapter 4.2.2) were obviously particularly important in the South, but not in the North. An important cause for the regional variances can be higher summer temperatures in the upper layer of the South compared to the North (Shimaraev et al. 1994), because in the South, ice break-up, warming and stratification begins earlier than in the North, where the water column is homogeneous during most time of the year and increases rarely above 10° C (Shimaraev et al. 1994). These higher temperatures may result in higher degradation rates within the upper layer in the South (Cuddington and Leavitt 1999).

Chlb and its degradation products showed similar degradation to TOC at both sites, while Chlc was degraded significantly stronger in the North than in the South (decay slope significantly higher in the North, Appendix C - Tab. 6). Fucoxanthin was the major carotenoid in the top traps at both sites. This fucoxanthin likely resulted from large Bacillariophyceae, which rapidly sink out of the euphotic zone, being then protected against photooxidation (Fietz et al. 2005, chapter 4.2.3). However, fucoxanthin is known to be a very labile pigment (Repeta and Gagosian 1984, Leavitt and Findlay 1994) and it was in fact subject to stronger degradation compared to other carotenoids within the water column. However, despite the strong degradation, the contribution of fucoxanthin to total carotenoids (and also the estimated contribution of Bacillariophyceae+ Chrysophyceae to total Chla, Fietz et al. 2005, chapter 4.2.3) was high at the lake bottom in the South, due to the very high initial fucoxanthin fluxes trapped in the 40 m traps, while other carotenoids, such as lutein, ß-carotene and diadinoxanthin, dominated at the lake bottom in the North (cf. Fig. 37).

↓98

Taken together, the main contribution to the initial settling material was formed by heavy, non-edible Bacillariophyceae in both basins. However, the degradation processes during sedimentation varied between both basins. While major degradation occurred in the uppermost wind-induced mixed layer in the South, strong degradation was found down to the lake bottom in the North. As a result higher losses of organic matter and in particular of photosynthetic pigments were found in the North compared to the South. Additionally, pigment-specific degradations were found to differ between both sites. The regionally varying extent and manner of the degradation must be considered for further interpretation of fossil biologic proxies recovered from different sites within a lake.

4.2.5 Degradation within the surface sediment

Oxic water-sediment interface: Another extensive degradation before permanent burial occurred at the water to sediment interface. Most pigments occurred in much lower concentrations in the core top, than even in the deepest sediment traps (Tab. 9). Several previous studies in other freshwater sediments indicated that fossil pigment concentration directly reflects the standing crop abundance in the euphotic zone, while others indicated that up to 99 % of the autochthonous pigments are lost during sinking (see review in Leavitt 1993). In this study we found a loss of up to 75 % during sinking towards the bottom trap but up to 98 % loss after settling at the sediment surface. Similar high losses were found for the major lipid biomarker classes (Russell and Rosell-Mélé 2005) and for diatom valves (Battarbee et al. 2005), thus the figures here are likely to be a reliable dimension for the losses in a deep oxygenated lake.

The oxygen penetration into the sediment varies from 5 mm in the Selenga Delta to 50 mm in the North basin and Academician Ridge (Martin et al. 1993, Mackay et al. 1998, Granina et al. 2000, Vologina et al. 2000). Hence, buried pigments in Lake Baikal remain up to a few decades in an oxygen-saturated zone. The anoxic sediment trap material, in contrast, remained relatively protected from oxidation after burial into the traps. Thus, strong degradation affected those pigments or pigment degradation products that reached the sediment surface. Thereby, the degradation of chlorophyll, pheophorbide and pheophytin was much stronger than that of organic carbon. Typical Chla/TOC ratios in phytoplankton are 6-28 µmol g-1 (Sterner and Elser 2002) and whereas the Chlas/TOC ratios found in the top traps were within that range, the ratios found in the surface sediment were much lower except in the eastern South (Tab. 10).

↓99

The Chlas/DM and Chlas/TOC ratios varied about one order of magnitude among the distinct regions (South, Selenga Delta, North and Barguzin Bay). Similar high variation was already found for Chlas/DM in a previous study in Lake Baikal (Soma et al. 2001a). In both studies the Chlas/DM ratios were much higher in the South and Selenga Delta than in the North. Therefore, it may correspond to the significantly higher Chla concentration indicating higher primary production in the euphotic zone of the South and Selenga Delta compared to the North in summer (cf. chapter 3.1.1). Seasonal monitoring in the high latitude is required for further comparisons. However, one would expect also higher ratios in the Barguzin Bay, where the river inflow caused higher Chla concentrations in the euphotic zone, but the Chlas/DM ratio in the core top was much lower than in the South and Selenga Delta. Dilution with terrigeneous anorganic material is likely as the TOC content was also low. Referring pigments as units per TOC avoid the problem of dilution with terrigeneous anorganic material in studies of fossil pigments.

The differential degradation patterns at the investigated regions (South, North, Selenga Delta and Barguzin Bay) were also obvious from the share of Chla degradation products to total Chlas. For instance, only less than 10 % of the original Chla that reached the lake bottom traps was preserved in the surface sediment in the South, but less than 5 % in the North. Chla was best preserved in the northern Selenga Delta. Also, the share of the stable pheophytin to total Chlas was greater and that of the labile pheophorbide was lower in the North compared to the South. Pyropheophytin a, in contrast, that occurred only in traces in the trap material was well preserved at both sites. A conversion from pheophytin a to pyropheophytin a, where pheophytin a looses its methylcarboxylated group, could also be possible (cf. Fig. 3). Finally, the differential degradation was also obvious from share of carotenoids to total carotenoids: the labile fucoxanthin and diadinoxanthin (Leavitt and Findlay 1994, Leavitt and Hodgson 2001), both from Bacillariophyceae+ Chrysophyceae, were dominant carotenoids in the core tops from South, Selenga Delta and Barguzin Bay, whereas the only detected carotenoids in the North, in contrast, were the stable lutein and canthaxanthin.

Oxidised layer: Further degradation occurred below the water to sediment interface within the oxidised layer, which can be up to 30 cm thick in Lake Baikal (Granina et al. 2000). The redox boundary is indicated by Mn and Fe accumulation, forming crusts in some regions of Lake Baikal (Vologina et al. 2003, Granina et al. 2004). Thus, buried pigments remain several centuries within the oxidised environment. Cuddington and Leavitt (1999) predicted that rates of pigment deposition were inversely related to the thickness of the oxic conditions, because the degradation within the surface sediment results predominantly from chemical oxidation and enzymatic lysis by microbial activity (Fig. 53). In Lake Baikal both processes were suggested to have higher effect in the North, as the decay was significantly higher there than in the South or Selenga Delta.

↓100

Sharp decreases of the total Chlas and carotenoids were found within these oxidised layers in all short cores (cf. Fig. 40). A sharp decrease of chlorophylls within the upper 20 cm was also found in 12 Italian lakes (Guillizoni et al. 1983). It has been attributed to recent eutrophication. Although eutrophication has been suggested from diatom valve studies for the past 150 years in Lake Baikal (Mackay et al. 1998), it is not sufficient to explain the strong increase of Chla (from the historical point of view) within the upper 10 cm in all cores. This sharp increase of Chla with time or decrease with depth certainly resulted from degradation processes within the sediment rather than from production changes in the euphotic zone.

One indication of strong degradation was the high pheophytin/pheophorbide (lipophilic/water soluble) ratios in the surface sediments. Another indicator was the strong decay of the Chlas/TOC ratios with depth of the surface sediment. Diagenetic processes, such as canthaxanthin formation also indicated that changes resulted from processes within the surface sediment rather than from processes in the euphotic zone.

Hence, other than diatom valves (Mackay et al. 1998), the photosynthetic pigments may not be suited to reconstruct recent, i.e. during past 150 years, short-term variations in Lake Baikal. That does, however, not exclude that photosynthetic pigments can be useful tools for reconstructing long-term relative variations (in contrast to absolute abundance) before the time spanned by the oxidised layer.

4.2.6 Do recently buried pigments reflect the phytoplankton standing crop?

↓101

The record of the phytoplankton standing crop by trapped pigments in Lake Baikal was group-specific. Some carotenoids were more preserved than others, and they cannot record whole algal assemblages, unlike differential degradation rates be taken into account. The main contribution to the settling material was formed by heavy, non-edible Bacillariophyceae. Strong and variable degradation processes controlled the sedimentation of small, light and edible phytoplankton. Basically, in the South these processes took place within the upper 250 m of the water column, while the processes extended over the whole water column in the North. The pigment preservation was much better in the South than in the North, although the water column is deeper. The sedimentation out of the euphotic zone can be projected backward using the preliminary regression models given in the present study.

A second degradation took place within the oxydised surface sediment. Rough estimates were given to calculate initial Chla concentrations from sedimentary pigments below the oxidised layers. Nonetheless, oxygen penetration may have varied in the past, and therefore the Chlas/TOC ratios or Chlas/DM ratios are potentially unsuitable to determine total phytoplankton abundance but preferably to estimate relative changes. That has been concluded already for smaller or shallower lakes (Leavitt 1993), but seems to be particularly important in the deep, oxidised rift system of Lake Baikal.

In as much as the paleoecological analysis of preserved markers such as photosynthetic pigments is increasingly used to monitor environmental change in response to climate and human activities, the complexity and variability of the degradation, revealed in this study, should improve our understanding of the limits of such retrospective analyses. Because Lake Baikal is unusual in terms of size and depth, it represents an interesting end-member in investigations of pigment biogeochemistry. The conclusions are difficult to apply to other mainly shallower freshwater systems, but can considerably contribute to the understanding on the manner in which organic molecules are incorporated into the sediments in cold, deep, oxygenated lakes and in marine systems.

4.3 Reconstruction of past phytoplankton variations

↓102

The aim of this part of the thesis was to determine the potential of fossil photosynthetic pigments for the reconstruction of past relative abundance and community structure of phytoplankton in Lake Baikal. The studies on the vertical transfer and preservation of photosynthetic pigments in Lake Baikal showed, that the decomposition within the water column and the surface sediment is strong and that pigments were likely unsuitable to reconstruct changes of the past hundreds of years (chapter 4.2). However, the present study showed that major paleoclimatological changes of the past 130,000 years were assessed and we may presume that the Chlas/TOC ratio in Lake Baikal is a useful indicator for past relative changes of the phytoplankton abundance. This was demonstrated in all three Holocene cores (´Vidrino´, ´Posolski´ and ´Continent Ridge´) and in the deep core section from the Continent Ridge that encompassed the last glacial and interglacial periods.

4.3.1  Holocene

Regional differences of the Holocene record: Comparison between the Holocene cores from the three coring sites (´Vidrino´, ´Posolski´ and ´Continent Ridge´, located in the South basin, Selenga Delta and North basin, respectively) indicated lowest productivity (lower average Chlas/TOC ratio), and thereby lower temperatures, but frequent climate oscillations (higher coefficient of variation, cf. Tab. 12) at Continent Ridge. Temporal changes varied also regionally, and today we are not able to discern whether these differences resulted from different phytoplankton production or pigment preservation. Oxygen isotope analyses (Morley et al. 2005) have indicated south-northwards shifts of the weather systems influences, for example of the North Atlantic Westerlies and Siberian High, during Holocene. Such shifts may explain differential records in the North and South and by their impact on the catchments area also differential records in the Selenga Delta.

The C/N ratios (< 12 atomic ratio) proved the essentially autochthonous nature of the organic compounds buried at these three sites (even at river inflow) that has also been demonstrated for a series of other sites previously (Qiu et al. 1993). Typical phytoplankton C/N ratios in suspended and settling material are around 6 (Redfield ratio), while those of higher plant leaves are much higher (c. 45, Hutchinson 1975, Sterner and Elser 2002) and those of littoral macrophytes about 15 (Likens 1985).

↓103

Holocene climate records: Before 9 kyr BP (Boreal), the very low amounts of TOC, and very low ratios of Chlas/TOC and Chlbs/TOC, found in the Continent Ridge core, may be a late consequence of a cooling known as the Younger Dryas in European sites, which interrupted the warming in the Lake Baikal region (Colman 1995). Grey scale density in thin sections as well as pollen and sporomorph analyses (conducted on parallel kasten cores, Boës et al. 2005, Demske et al. 2005), indicated the Younger Dryas was represented at c. 12 kyr BP. This event has also been imprinted in the sedimentary photosynthetic pigment record of a core in the South basin (Tani et al. 2002).

During the Boreal, the TOC content and the Chlas/TOC ratios increased markedly. At c. 9 kyr BP, the Chlas/TOC ratios indicated a maximum of phytoplankton abundance, which may even have exceeded the current abundance. Pollen analyses indicated that high temperatures and favourable moisture favoured an optimum development of dark-coniferous taiga at Continent Ridge at that time (Demske et al. op. cit.). A maximum of diatom valve abundance was also found at that time in the Selenga Delta, which was related to high temperature and humidity in western Siberia (Karabanov et al. 2000a).

The strong Chlas/TOC decline between 8-7 kyr BP corresponded to a mid-Holocene cooling event reported from pollen analyses (Demske et al. op. cit.). A Chlas/TOC maximum occurred between 6.5-5.7 kyr BP during the following Atlantic, which was, however, less expressed than the Boreal maximum. At that time, an Atlantic maximum was also found for diatoms (Karabanov et al. op. cit.) and pollen (Horiuchi et al. 2000). Maximum distribution of Scots pine forests marked the Holocene thermal optimum at that time (Demske et al. op. cit.), and, therefore, although from phytoplankton records (pigments + diatom valves) the Holocene optimum occurred during Boreal, the pollen record indicated the Holocene optimum during Atlantic.

↓104

Subsequently, a cooling period at the Atlantic-Subboreal transition (5.7-4.5 kyr BP) was indicated by TOC content and Chlas/TOC ratio, diatom abundance (Karabanov et al. op. cit.) and pollen data (Demske et al. op. cit.). During that time, the southern cores showed increasing temperatures (increased Chlas/TOC ratio), probably indicating the aforementioned shifts of the major weather systems. From c. 4.5 to 3 kyr BP the Chlas/TOC ratios increased, indicating a Subboreal optimum, which was also indicated by grey scale density and pollen data (Boës et al. op. cit., Demske et al. op. cit.). This period has previously also been shown to be one of high diatom productivity and biogenic silica accumulation (Qiu et al. 1993, Karabanov et al. op. cit.). However, the widely accepted opinion that the warmest period of the Holocene in the Baikal region was the Subboreal (Karabanov et al. op. cit.) could not be confirmed here for the Continent Ridge site, as the Boreal optimum was higher than that of the Subboreal. The cores at Vidrino and Posolski did not reach the Boreal period and hence no comparison with the Subboreal optimum could be made.

Subsequently, between 3 and 1 kyr BP a cooling period may be assumed at all three sites as we recorded lowest TOC content, Chlas/TOC and Chlbs/TOC ratios, and carotenoids/Chlas ratios during that time. In a previous pigment study (Tani et al. 2001) lowest total Chlas and carotenoid concentrations were also found during approximately the same period in a core from the South basin. From 1 kyr BP to present an increase of the Chlas/TOC ratio indicated an increase of the phytoplankton productivity, but as aforementioned (chapter 4.2.5) this increase was related to strong degradation within the oxidised layer rather than to productivity changes.

Variation of phytoplankton composition during Holocene: During the Boreal Chlas/TOC maximum, the Chlbs/Chlas ratios slightly increased, while Chlc was not detected; we assume, therefore, that during this Boreal optimum Chlorophyta occurred in higher amounts than Bacillariophyceae. Since today Chlorophyta occur mainly during early summer and in the South basin, while Bacillariophyceae occur mainly in the cooler waters (Kozhova and Izmest’eva 1998; chapter 3.1), the Boreal increase of the Chlbs/Chlas strengthened the assumption of a Holocene temperature optimum at that time. During the subsequent Chlas/TOC minimum (c. 8-7 kyr BP), the Chlbs/Chlas ratio was also high, however, so that evidently some cold-adapted Chlorophyta occurred during cold periods of the Holocene. Varying Chlbs/lutein ratios throughout the cores also indicated changes of the chlorophycean species composition.

↓105

At the time of the Atlantic temperature optimum (c. 6.5 to 5.5 kyr BP) the Bacillariophyceae started enhanced growth before the Chlorophyta (as the Chlc/Chlas peaked at the early optimum (c. 6.2 kyr BP) and the Chlbs/Chlas thereafter (c. 5.8 kyr BP)). Nonetheless, Chlc peaks were not correlated to those of diatom valve abundance (Rioual and Mackay 2005), which could be due high Chla/Chlc ratios in Bacillariophyceae, as has been suggested in previous studies (cf. Squier et al. 2002) or to strong degradation of Chlc. Chlc is, alike pheophorbide, missing the phytol ester group and is therefore more water soluble than Chla, Chlb or its pheophytins. Moreover, Chlc cannot degrade to pheophytins and therefore no degradation products were detected.

Pigments from cyanobacteria were not detected in the cores, because cyanobacteria either do not or only negligibly settle into Lake Baikal’s sediment due to their small size, lightness and edibility (cf. chapter 4.2). Thus, unfortunately, tracking the development of the cyanobacterial picoplankton cannot be accomplished directly by tracking the marker pigments in Lake Baikal. Fossil canthaxanthin has previously been used to track cyanobacterial development in freshwater lakes, including Lake Baikal (Soma et al 1996, Tani et al. 2002). However, canthaxanthin could not be detected or was present only in trace amounts in more than 300 water samples from Lake Baikal (although the presence of cyanobacterial picoplankton was demonstrated by other marker pigments and epifluorescence counts therein) and in isolated Baikalian cyanobacterial strains (cf. chapter 3.1 and Fietz et al., submitted, Appendix B). We assume that the canthaxanthin detected in Lake Baikal’s sediment comes not directly from phytoplankton (especially cyanobacteria), but from crustaceans that feed on phytoplankton. The crustacean carapace contains protein-bound astaxanthin. Within the sediment these carapaces are destroyed and astaxanthin is liberated. Under reducing conditions, astaxanthin is then transformed to canthaxanthin. With that assumption, the canthaxanthin detected in Lake Baikal’s sediment indicated enhanced zooplankton abundance.

Most of the detected carotenoid degradation products were not attributable to their parent pigments. The use of HPLC-MS (connexion of HPLC and mass spectrometry) based methods will presumably provide further information on the degradation products in future studies conducted in similar oligotrophic and deep lakes with oxic hypolimnia. Mass spectrometry identifies the nature of molecular modifications because it confirms the molar mass of a pigment or degradation product as well as characteristic fragmentation patterns (Leavitt and Hodgson 2001) and has been successfully used in a few marine and freshwater systems (e.g. Squier et al. 2002). However, mass spectrometric techniques require pure compounds in high concentrations for the analysis of pigment characteristics (Leavitt and Hodgson 2001). For the Lake Baikal sediment samples detailed structural information about phospholipid mixtures were gathered by the use of HPLC-ESI-MS-MS that allowed deeper insights into the compositions of microbial communities and the influence of environmental conditions (Zink and Mangelsdorf 2004). The use of similar techniques for the sedimentary pigments of these Baikal sediment samples can be hypothesised from preliminary measurements (unpublished data).

4.3.2 Last Interglacial (Kazantsevo)

↓106

Kazantsevo climate record: The TOC content and the Chla/TOC ratio were much higher during the Kazantsevo Interglacial compared to the preceding (Tazovsky, equivalent to European Saalian) and subsequent (Zirianski, equivalent to European Weichselian and MIS 5d; Karabanov et al. 1998) Glaciations. Hence, although strong degradation my have occurred during cold and warm periods, the relative changes in the TOC content and the Chla/TOC ratio delivered important information about paleoclimatic changes.

The strong increase in the Chlas/TOC ratio we observed at c. 128 kyr BP was in good agreement with the simultaneous increase of the total sporomorph concentrations (particularly of Pteridinium aquilinium spores), indicative of a considerable rise in temperature (Granoszwski et al. 2005). Study of diatom valves also indicated a strong increase in productivity at c. 128 kyr BP, as biovolume accumulation rate increased by an order of magnitude (Rioual and Mackay 2005). At that time insolation was maximal in the Baikal region (Prokopenko et al. 2002). According to pollen, diatoms, and biogenic silica the Kazantsevo Interglacial lasted for 11-12 kyr BP (Edlund and Stoermer 2000, Prokopenko et al. 2002, Granoszewski et al. op. cit., Rioual and Mackay op. cit.). Assuming a sedimentation rate of 6.4 cm kyr-1 (Demory et al. 2005), the TOC and chlorophylls in our study indicated approximately the same time span. Studies of the corresponding European Eemian pointed to duration of 11 to 13 kyr (Tzedakis et al. 2003).

The Chlas/TOC ratio was highest between c. 126 and 121 kyr BP. This period corresponded to the maximum diatom abundance (Rioual and Mackay op. cit.). The pollen fossils indicated a short-lived rise of temperature and moisture at c. 126.4 kyr BP (Granoszewski et al. op. cit.), and the TOC content we recorded was highest around that date. However, neither the Chlas/TOC ratio (Fig. 8) nor the diatom abundance (Rioual and Mackay op. cit.) exhibited a peak at that time. A change of phytoplankton composition towards cells with low pigment content, and diatoms with high dissolution, was likely.

↓107

During the Kazantsevo, the Chlas accumulation rate and the Chlas/TOC ratio indicated short cooling events at c. 125.5, 123.5 and 122 kyr BP and temperature optima in between at c. 124.5, 122.5, and 121 kyr BP. These short-lived events did not all correspond to the pollen and diatom variations (Granoszewski et al. op. cit., Rioual and Mackay op. cit.). This apparent mismatch could be due to the fact that the measurements were done on parallel cores and, although they were dated and correlated accurately, short peaks were possibly attributed to slightly varying dates. Overall, pollen, diatom and pigment data indicated a suite of strong, short-lived, oscillations of the weather conditions during the Kazantsevo Interglacial. Phytoplankton abundance was halved or doubled within centennial time scales only.

Within the late Kazantsevo, a first drop of the TOC, Chlas accumulation rate and Chlas/TOC occurred at c. 119.5 kyr BP, which was even more expressed by the diatom abundance (Rioual and Mackay op. cit.). This drop has also been found in previous studies on biogenic silica and diatom valve abundance in other regions of Lake Baikal (Karabanov et al. 2000b, Prokopenko et al. 2002).

Lithology and photosynthetic pigments indicated an abrupt end of the Kazantsevo Interglacial that closely followed a sharp decrease of the insolation towards 116 kyr BP (Prokopenko et al. 2002). At c. 116 kyr BP the insolation turned and increased again during the Early Zirianski (Prokopenko et al. 2002). However, even during the Early Zirianski Glaciation, the Chlas/TOC ratio we observed pointed to considerable oscillations of the paleoclimatic conditions, which were not observed in the TOC content or lithologically nor were they revealed within the diatom and pollen studies (Prokopenko et al. 2002, Granoszewski et al. op. cit., Rioual and Mackay op. cit.). The short-term changes may have been caused by fast responses of phytoplankton other than diatoms, such as Chlorophyta.

↓108

Phytoplankton composition: Three Chlbs/Chlas maxima were found during the Interglacial: around 126, 123, and 118 kyr BP, indicative of three temperature optima. The Chlbs/Chlas peaks during the glacial periods may be explained by the occurrence of so-called “snow algae”, which today are formed by Chlorophyta (e.g. Chlamydomonas, Hoham and Duval 2001). These algae could have developed on the ice and snow that likely covered the lake during most of the year during the cold periods. The Chlbs/Chlas ratios during the Early Zirianski exceeded 20 % (Fig. 46). Chlb/Chla ratios within the euphotic zone as well as in isolated Chlorophyta were found to be 30 to 40 % (Fietz and Nicklisch 2004, Appendix A; Fietz et al., submitted, Appendix B). Hence, the Chlas during the Early Zirianski Glaciation was formed nearly completely by Chlorophyta. Different Chlorophyta species obviously developed under very different climatic conditions.

Occurrence of Chla derivatives: Tani et al. (2002) suggested from a core spanning c. 24 kyr BP, that pyropheophytin and SCE are the dominant Chla derivatives in Lake Baikal sediments whose ages exceed 10-15 kyr BP. In our study, pyropheophytin and even SCEs, which may be of greatest stability (Prowse and Maxwell 1991, Talbot et al. 1999), were not the dominant Chla derivatives during the Kazantsevo Interglacial; pheophorbide a, in contrast, contributed up to 50 % of the total Chlas during the Kazantsevo Interglacial. The high contribution of pheophorbide a to the total Chlas indicated high grazing activity, because pheophorbide is a degradation product of chlorophyll following passage through zooplankton gut. Enhanced zooplankton growth strengthened the assumption of warm conditions during that period. In a 2.8 million years old sediment SCEs were the only preserved chlorophyll derivatives and were thought to be proxies of global planetary changes in Lake Baikal (Soma et al. 2001b). They were also thought to be formed by passage through zooplankton guts (Soma et al. 2001b). However, as SCEs are only trace components within the surface sediments (unpublished), a conversion from pheophorbide after burial should be considered too.

Interestingly, the distinct forms of pheophorbide, pheophytin, and pyropheophytin showed greater variability during the Holocene than during the Kazantsevo Interglacial (Fig. 9). Different forms resulted from either different phytoplankton species or from differential sedimentation. Thus, it might be assumed that either phytoplankton compositions or the sedimentation conditions were more variable during the Holocene than during the Kazantsevo Interglacial. Further analysis and experiments on the origin of the distinct degradation forms will provide further insights into past environmental conditions and degradation processes.

↓109

Perylene: Perylene occurred during the Holocene and Kazantsevo Interglacial without correlation to chlorophylls or carotenoids. Sillimann et al. (1998) did also not found significant correlation in Lake Ontario, but Soma et al. (1996) and Tani et al. (2001), in contrast, found significant correlations between perylene and photosynthetic pigments in cores from the South basin of Lake Baikal. Perylene has been found in lakes and seas evenly in recent (except oxic surface) and ancient sediments (Aizenshtat 1973, Louda and Baker 1984, Silliman et al. 1998, Jiang et al. 2000, and references therein). The organic precursor of perylene was suggested to be phytoplankton or its photosynthetic pigments (Aizenshtat 1973, Laflamme and Hites 1978), but many studies indicated also a potential terrestrial precursor even in aquatic sediments (Jiang et al. 2000, and references therein), and in situ formation from non-specific precursors is although likely. All studies agreed, however, that only anoxic conditions allow the formation of perylene and that therefore perylene is an indicator of depositional conditions rather than of organic matter (Silliman et al. 1998). With that assumption, changing redox conditions may be tracked in Lake Baikal. However, as far as neither the precursor, nor the formation processes, nor the preservation are definite, interpretations based on perylene variations are rather speculative. Nonetheless, the differential occurrence of perylene within the regions and periods in Lake Baikal may incite further detailed studies.

Reconstruction of the past phytoplankton abundance: Pollen data indicated that in Lake Baikal more favourable climatic conditions prevailed during the Kazantsevo Interglacial than during the Holocene (Granoszewski et al. 2005), which was confirmed for phytoplankton by the significantly higher mean Chlas/TOC ratios we observed during the Kazantsevo (Tab. 13). However, the predominance of large-celled diatoms during the Kazantsevo maximum dampens this assumption, as these indicate deep-mixing events and clear ice cover that preferentially occur at low temperatures (Edlund and Stoermer 2000). An alternative to higher temperatures is that higher productivity (higher mean Chlas/TOC ratios) resulted from a higher nutrient availability. The significantly lower C/N-ratios (Tab. 13) found during the Kazantsevo compared to the Holocene indicated higher nitrogen availability (increase of N) or lower productivity (decrease of C) during the Kazantsevo. Lower productivity was unlikely because the Chlas/TOC ratios (as aforementioned) were significantly higher; thus, higher nitrogen availability was likely. Phosphorus-enrichment in Lake Baikal do not serve as a direct marker for paleoproductivity, but is controlled by porewater chemistry and sedimentation rates (Fagel et al. 2005). However, the sedimentary phosphorus-content (measured as P2O5) was 35 % higher during several thousands years before the onset of the Kazantsevo compared to several thousands years before the onset of the Holocene (Fagel, pers. comm. from unpublished data on the parallel core CON01-603-2). We therefore assume that possible limiting nutrients (nitrogen and phosphorus) were more available for the phytoplankton development during the Kazantsevo than for the phytoplankton development during the Holocene. That could explain the long lasting high Chlas/TOC ratios found during the Kazantsevo, which are unlikely to be caused by temperature changes only. Another possible factor is that increasing productivity reduced the thickness of the oxidised layer at the sediment surface and thus reduced the degradation and increasde the preservation.

Transfer models for the degradation within the surface sediment (oxidised layers) and within the water column have been calculated based on a series of short cores and two series of sediment trap moorings previously (cf. chapter 3.2). The degradation within the surface sediment (oxidised layer) at Continent Ridge was exponential: y=F*exp(-x/3.69), whereby y was the Chlas/TOC ratio (in µmol g-1) at the depth x (in cm) and F the estimated ratio at the sediment surface (in µmol g-1)(Appendix C - Tab. 7). The maximal Chlas/TOC ratio at the Kazantsevo optimum was 0.38 µmol g-1. The degradation below the oxidised layer can be neglected, and therefore, x is set as the approximate thickness of the oxidised layer. Assuming an at least 5 cm thick oxidised zone at the Kazantsevo maximum, a Chlas/TOC ratio of 1.5 µmol g-1 resulted [F=0.38/exp(-5/3.69)]. The recent ratio in the core top at Continent Ridge is 0.73 µmol g-1. Assuming thicker oxidised layer even increased the reconstructed Chlas/TOC ratio. Hence, the reconstructed Chlas/TOC ratio for the Kazantsevo maximum was twice as high as that preserved recently.

↓110

The results from sediment trap moorings were used to reconstruct the Chlas/TOC ratio of the euphotic zone during the Kazantsevo (cf. chapter 3.2). Strong degradation occurred at the water to sediment interface: from the Chlas/TOC that reached the lake bottom only c. 13.5 % were preserved in the surface sediment of the North, and therefore, the reconstructed ratio in the bottom trap during Kazantsevo optimum was 11 µmol g-1. Furthermore, the decrease of the Chlas/TOC ratio within the water column (from top trap at 50 m to bottom trap at c. 900 m) in the North basin was linear: y=F-0.019*x (Appendix C - Tab. 6). Therefore, in 50 m water depth, which is the lower boundary of the euphotic zone in Lake Baikal (Kozov 1963, Kozhova and Izmest’eva 1998), a Chlas/TOC ratio of 27 µmol g-1 was estimated at the Kazantsevo optimum [F=11+0.019*850].

Thus, the reconstructed Chlas/TOC ratio for the Kazantsevo maximum was higher than the Chlas/TOC ratio determined in the 50 trap of the North mooring (2002-2003), which was 16 µmol g-1 (cf. chapter 3.2.2). However, analyses of ocean colour satellite data (using appropriate bio-optical modelling), indicated that the Continent Ridge site differed from the site of the mooring in phytoplankton abundance and terrigeneous input (Heim et al. 2005). Hence, interpretation must be drawn carefully. However, apparently the reconstructed Chlas/TOC ratio, which indicated the phytoplankton productivity, suggested more favourable conditions during the Kazantsevo optimum than today.

4.3.3 Do fossil pigments track past phytoplankton community structure and environmental changes?

It could be shown that fossil Chlas serve in Lake Baikal as reliable indicators for global climatic changes. Moreover, fossil Chlbs tracked the development of Chlorophyta. However, other pigments, such as Chlc or carotenoids, were not suitable to track changes of the phytoplankton community structure as their degradation products were not detected or could not be related definitely to their parent pigments. Yet a few major trends can be stated from sedimentary phytoplankton pigments for Lake Baikal: (1) Higher Chlas/TOC ratios indicated higher phytoplankton production and thereby higher water temperatures during the Kazantsevo Interglacial compared to the glacial periods. Strong climate oscillations occurred during the Interglacial and phytoplankton was halved or doubled within centennial time scales. (2) Highest phytoplankton production during the Holocene occurred at c. 9 kyr BP at the time of climate amelioration following the Younger Dryas (Boreal). (3) Short maxima occurred during the late Atlantic and at the Subboreal/Subatlantic transition. The comparison of the sedimentary pigments with published diatom records indicated that phytoplankton other than diatoms contributed to the buried organic material.


© 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 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated:
06.10.2006