The Project description

The Caspian Sea - KaraBogaz Gol ecosystem is a closed basin with its own sea level regime independent of that of the oceans. The Volga river accounts for about 80% for the inflow of water in the sea, the remainder being formed by seven other smaller rivers and inflow from groundwater. The outflow is mainly by evaporation at the sea surface and to the KaraBogaz Gol. The KBG represents the end member where all the water is evaporated. The instrumental water level record of the CS and of the KBG show that they covariate

Fig. 1. Caspian Sea
.
Caspian sea level oscillations may be a hundred time as rapid as eustatic ones. The last short-term sea-level cycle started with a sea-level fall of 3 metres from 1929 to 1977, and a sea-level rise of again 3 m from 1977 onwards, until in 1995 a highstand was reached. Since then smaller oscillations have taken place. The cause of these rapid changes have been much debated, but the most plausible explanation has been given by Rodionov (1994), and is related to the water balance of the sea. Rodionov (1994) showed that Caspian sea-level changes show a statistically significant correlation with secular changes in the discharge of the Volga river. These in turn have been shown to record variations in precipitation over the Volga drainage basin, related to variations in the amount of Atlantic depressions that reach the Russian mainland.
Now the strength of the depression activity in NW Europe has recently been shown to be strongly dependent upon variations in the North Atlantic Oscillation (NAO) (Hurrell, 1996, 2000). The curve depicting the variations in strength of the NAO matches the historic Caspian Sea level change in this century remarkably well (Arpe et al., 2000). Ice accumulation rates in western Greenland also show a decreasing trend from the thirties onwards, down to the seventies, and an increasing trend from the seventies up to the present (Appenzeller et al., 1998). The ice record previous to the last Caspian sea-level cycle shows much more rapid short-term variations than Caspian sea level, but the longer-term average trend is more or less constant in both cases.
These data would indicate that Caspian Sea level could be a good proxy for historical changes in the NAO and might be used to calibrate Global Circulation Models for the Quaternary. On the Holocene time scale, an orbitally forced signal is superimposed upon the decadal-centennial scale NAO oscillation (Kislov & Surkova 1998), but this can be filtered out, as the climate systems reacts in a quasi-linear way on orbital forcing (Crowley & North, 1991). The essential fact remains that Caspian sea-level reflects mainly changes in precipitation across a huge part of the European territory (1.5 mln. sq km) and hence is a valuable additional data source to GCM, as most other proxy data reflect palaeotemperature rather than precipitation.

Conflicting data

Rapid sea-level fluctuations have taken place in the Caspian sea since it became a closed basin about 5.5 Ma ago. The Caspian sea, now at -27 m below oceanic sea level, is known to have had highstands of +50 m and -80? m in the last 100,000 years, and even higher highstands further afield. Svitoch (1991) subdivided the sea-level cycles in five classes based on the order-of-magnitude of their duration, in the orders of magnitude from 105 years to 101 years. The main phases are well known from both onshore and offshore borings and from the occurrence of uplifted marine terraces around the Caspian sea. They have been characterised biostratigraphically mainly on the basis of mollusc biostratigraphy, especially Didacna spp., which has shown a rapid evolution through the Quaternary. However, absolute dating of the Pliocene, Pleistocene and Holocene Caspian sea level changes has been notoriously unsuccessful. Just to cite an example, the Early Khvalyn + 50 highstand is regarded by Rychagov (1977, 1997) to have taken place at 70 ka BP on the base of U/Th datings of marine terraces in Dagestan, but at 15 ka according to Svitoch (1991) on the basis of datings of molluscs along the Lower Volga. This is a very important issue because it determines whether high sea-levels are contemporaneous with the build-up of continental ice sheets, or with deglaciation. More detailed dating problems are given below.

Fig. 2 'Consensus' about Caspian Sea level change in the Pleistocene (Varushchenko et al., 1987).

Pliocene: The Pliocene Productive Series in Azerbaijan, deposited by a Palaeovolga into the South Caspian Basin while the Middle and Northern Caspian were dry, is bracketed by volcanic ash datings between 5.5 and 3.4 Ma, but no absolute ages have been found in up to 8000 m of sediment deposited between those data. Milankovic cyclicity has been postulated in these deposits (Nummedal, 2000) but by lack of dating these cyclicities cannot be put in a chronostratigraphic framework
Early-Middle Pleistocene: Apsheron sediments from Azerbaijan show reversed magnetization (Trubikhin, 1987) and intercalated volcanic ashes give a fission track age of 0.96 Ma (Ganzey, S.S. cited by Mamedov and Alekserov, 1991). Apsheron terraces in Dagestan occur around 300 m absolute height. All four subsequent phases appear to be normally magnetized, and thus are of Brunhes (< 700 ka) age (Rychagov, 1977; Trubikhin, 1987). From the two main Baku marine terraces in Dagestan with absolute heights of 200-220 m TL ages between 400 ?48 ka and 480 ?53 ka have been obtained (Rychagov, 1977), while volcanic ashes enclosed in Baku sediments from Azerbaidzhan have been fission-track dated at 510 ka (Koshkin, 1984, cited by Mamedov and Alekserov, 1991). The main Khazar terraces between 170 and 80 m in the same area give ages between 300 and 145 ka (Rychagov, 1977). Upper Khazar sediments from the North Caspian plain give ages between 100 and 125 ka (Shakhovets and Shlyukov, 1987). While there is little disagreement as to the relative age of these deposits, the absolute height reached by these transgressions and the intervening regressions is still hotly debated (Varushchenko et al., 1987, see Figure).
Last Glacial: In the Khvalyn terraces in Dagestan two major phases have been distinguished. The Early Khvalyn transgression is represented by five successive marine terraces between +50 and 0 m absolute height, and the Late Khvalyn transgression by at least four successive marine terraces between 0 and -20 m absolute height (Rychagov, 1977). The deep Yenotaev regression between the Early and Late Khvalyn transgressions may have reached down to 80m below present sea-level (Maev et al., 1989; Maev 1994). The +50 m Early Khvalyn terrace level is also referred to as called 'maximal transgression'. Deposits of the maximal transgression form the highest level at the surface in the whole North Caspian basin, and there is evidence of the existence of an overflow to the Black sea at +50 m through the Kuma-Manysh depression north of the Caucasus (Menabde and Svitoch, 1990).
There is still total disagreement on the age of the Khvalyn transgression, though all ages are within the range of the Weichselian (Valdaian) glacial stage (isotopic stage 4-2). Early Khvalyn sediments from outcrops in the North Caspian plain between Volgograd and the 0m contour (Fig. 1, 11) give TL ages between 24-26 ka, which are rather close to a 14C age of 34 ka on organic matter from one of the profiles (Shakhovets and Shlyukov, 1987). The Khvalyn marine terraces in Dagestan consistently range in age between 70 ka for the uppermost (oldest) ones to 14.6 ka for the lower (youngest) terrace according to TL datings (Rychagov, 1977). However, these same terraces all gave 14C mollusc ages between 15 and 8 ka, and are indistinguishable by this method. Also U-Th ages are in the range of the 14C ages (Rychagov, 1977). While Rychagov (1977) accepts the TL ages as being the most reliable on the base of their geomorphological consistency, Svitoch (1991) argues for acceptance of the 14C ages on the basis of their proximity in age, their coincidence with the U-Th datings, the fact that they have been obtained by different laboratories and are in the optimal range for this type of datings. In his view, part of the Khvalyn is early Holocene (see also Svitoch et al., 1987, 1993; Kaplin et al., 1993).
There is as yet no final answer to this question. The age of the maximal transgression, 70 ka, 25 ka or 15 ka, is of primary importance to solve the problem of synchronicity of Caspian sea-level changes with global climate change.
This question is also related to the unsolved controversy whether or not proglacial ice-dammed lakes in Siberia and Northern Russia existed during the last deglaciation, and in how far they have led to drainage diversions towards the Caspian Sea (Grosswald, 1980, 2002; Mangerud et al., 2001, Charbit et al., 2002)
Results from the Franco-Russian coring campaign in the Caspian in 1994 (Escudié et al., 1998; Leroy et al., Chalié et al., Jelinovska et al, 1998 suggest that the transition to the Holocene coincided with a regression. In spite of the wealth of biostratigraphical data from cores in a transect through the deepest parts of the Middle and Southern Caspian Sea 14C dating gave conflicting results, so that they cannot be put into a chronostratigraphic framework. As a consequence, little of the palaeoenvironmental work has been published on the project.
Holocene: The most detailed knowledge on Holocene sea-level comes from the Turali-Sulfat section of barriers and incised valley fills along the Dagestan coast described by Rychagov (1977, 1993b). Up to 5 transgressional phases have been described and 14C dated around 8000, 7000, 6000, 3000 and 200 years BP. The maximal absolute height of the Caspian sea reached during these transgressions is around -22 m. We are now studying these same sections again with AMS radiocarbon datings on molluscs and O and Sr isotope studies, indicating a major highstand about 2600 BP (Kroonenberg et al., 2002, Vonhof et al., 2002), which coincides ? with evidence for high rainfall in the upper Volga reaches (Gracheva et al., 2002). The depth of the intervening regressions is much less clear, however. According to Maev et al (1989) the Mangyshlak regression at the start of the Holocene might have reached the -50 to -70 m isobath, and the Derbent regression around 1500 BP at least attained the -34 m isobath. Giralt et al. (2002) showed a 45-years cyclicity from corings in KaraBogaz, in harmony with monitored data on CSL and Volga discharge variations.

Causes of discrepancies

The causes of the problems are related both to technical and sampling problems. Part of the discrepancies might be due to the different dating methods applied, including U-Th, TL and 14C. There is a general scarcity of peat and other plant macrofossil material, so that most existing datings are on molluscs, ostracods etc. Some previously obtained mollusc data may have obtained from reworked or recrystallized specimens.
14C ages might be unreliable because of methane production in submarine mudvolcanoes, presence of carbonate particles (hard water reservoir effect), detrital carbonates, detrital organic matter. The next step is to attempt radiocarbon dating on dinoflagellate or on organic compound produced by dinoflagellates.
In addition, a noticable difference between the present-day 14C activity of sea surface (TDIC) and atmospheric CO2, has been found (detailed in Escudié et al. 1998).
Marine terraces with relatively dated deposits are known to have undergone differential uplift along different portions of Caspian shores, and repeated leveling and GPS studies indicate that even at present some sectors are subsiding while others are uplifting. mainly due to a highly active tectonic environment. That fact adds an additional complexity factor in order to accurately establish the Caspian Sea water level evolution.
Dating deep-sea cores in the centre of the South Caspian might be hampered because of the extreme instability of the slopes leading to frequent turbidity currents. There is a lack of continuous sections so far, and the most promising ones such as Kara Bogaz are difficultly accessible. Without detailed sequences of dates, it remains so far difficult to ascertain the continuity of these sections.,
The IGCP project Dating Caspian Sea Level Change aims to
(1) exchange knowledge and assemble all existing age data in a data base and discuss their quality
(2) consider existing cored and other sampled material for additional dating, including inter-laboratory comparison
(3) consider new sites for coring, sampling and dating in the Caspian region and Volga drainage basin
(4) establish a new Caspian sea level curve in four time scales: historic, Last Glacial and Holocene, Pleistocene and Pliocene.
(5) use CSL curve for validating existing Global Circulation Models.