SEA LEVEL AND PALAEOCLIMATIC CHANGES IN THE SOUTH AND MIDDLE CASPIAN SEA REGION SINCE THE LATEGLACIAL FROM PALYNOLOGICAL ANALYSES OF MARINE SEDIMENT CORES

A review of pollen, spores, non-pollen palynomorphs and dinocyst analyses made in the last two decades is proposed here. Building on spare palynological analyses before 1990, a series of new projects have allowed taking cores in the deeper parts of the Caspian Sea, hence providing access to low-stand sediment. However, still nowadays no complete record exists for the Holocene. The first steps towards quantification of the palynological spectra have been taken. Some of the most urgent problems to solve are the uncertainties related to radiocarbon dating, which are especially acute in the Caspian Sea.


INTRODUCTION
The Caspian Sea has known many small and large-scale changes of its water level: c. 160 m in the last glacial-interglacial cycle and > 3 m in the last century [Kroonenberg et al. 2000]. In the latter period, these changes have had a dramatic impact on socio-economical activities around the sea [Kazancı et al., 2004;Leroy et al., 2010]. To reconstruct past sea level changes in the Caspian Sea (CS) and past climates of the region, the traditional approach so far has been to look at outcrops, to analyse their sediment and micro/macrofossil contents and to obtain radiocarbon dates on bivalve shells. Low stands are not recorded with this method otherwise than by a hiatus. The CS level variability is dominated by the variability of precipitation over the Volga River basin. At a longer timescale it is not impossible that other drivers of the water level played a role such as anthropogenic and tectonic ones.
Recently marine cores have been obtained in the shallow and more rarely in the deeper central and southern basins of the Caspian Sea ( Fig. 1). Their multidisciplinary analyses covering both low and high stands holds the key to understanding firstly when sea level changes occurred, which is a step before understanding why they occur and secondly how climate changed, how fast and what were its drivers.

DRIVERS OF CASPIAN SEA LEVELS
In summer 2010, extreme temperatures well above 30 °C have affected Moscow for nearly two months. As a direct result of this and combined drought, extensive wildfires occurred in the Volga region. Global Climate Models have suggested that drought over the Volga basin would occur when ENSO is in La Niña phase [Arpe et al. 2000], and this is what occurred in 2010. Nowadays the Volga River brings 80-85 % of the river water to the CS. However a few centuries ago, the Uzboi River (now defunct) brought water from the Amu-Darya [Létolle, 2000] a river whose source is in the Pamir and Tien-Shan and therefore its water is derived from the melting of monsoon-fed glaciers. Therefore the Caspian Sea water levels may be influenced both by climate of northern Europe and by climate over the western Himalayas.

PROXIES
Besides pollen (for example the former work of Abramova [1980] and Vronsky [1980] and the current work cited here) and non-pollen palynomorphs [Mudie et al., in press] a new proxy is being developed in the Caspian region, which is dinocysts. These small prokaryote organisms have many endemic forms in the Caspian region and it is only recently that their taxonomy has been firmly established [Marret et al., 2004] allowing now different scientists to use the same names and compare their data. Various forms, species and genera are related to different environments such as water salinity, water temperature, and nutrient content [Mertens et al., 2009]. Therefore this method is a proxy for sealevel changes. Plates 1 and 2 show some forms characteristic of the Caspian Sea and the Karabogaz Gol.

SURFACE SAMPLES
Surface samples are essential to interpret past changes, as they are a stepping stone to quantification by linking microfossil assemblages to environmental and climatic conditions (analogues). A collection of surface samples form useful training sets   Karpytchev [1993]. This poor precision needs to be resolved. The best material to date would be remains of terrestrial plants, which are however quite rare in marine cores.
For the more recent times, i.e. the last 150 years, the radionuclid method is the best, either 210 Pb alone or in combination with 137 Cs. The combination of radiocarbon and radionuclid methods however still leaves a gap between AD 1750, the most recent reliable radiocarbon ages due to a subsequent plateau, and AD 1860, the oldest age obtained by radionuclids. Pollen, spores and dinoflagellate cysts have been analysed on these sediment cores . The pollen and spores assemblages indicate fluctuations between steppe and desert. In addition some outstanding zones display a bias introduced by strong river inflow. The dinocyst assemblages change between slightly brackish (abundance of Pyxidinopsis psilata and Spiniferites cruciformis) and more brackish (dominance of Impagidinium caspienense) conditions.

THE LAST FEW CENTURIES
During the second part of the Holocene, important flow modifications of the Uzboy River and the Volga River as well as salinity changes of the Caspian Sea, causing sealevel fluctuations, have been reconstructed. A major change is suggested at ca 4 cal. ka BP with the end of a high level phase in the south basin (core CP14). Amongst other hypotheses, this could be caused by the end of a late and abundant flow of the Uzboy River, carrying to the Caspian Sea either meltwater from higher Eurasian latitudes or water from the Amu-Darya and the western Himalayas. A similar, later clear phase of water inflow has also been observed from 2,1 to 1,7 cal. ka BP in the south basin and probably also in the north of the middle basin.

THE EARLY HOLOCENE AND LATEGLACIAL
A further two cores from the same cruise of 1994 are being analysed for the pollen and dinocyst content. These Kullenberg cores are each 10 m long [Chali'@е et al., 1997]. Core GS05 from the south basin (museum number SR01GS9405) was taken in a slightly different coring station than core CP 14, i.e. in a more southerly location, but the two cores seem to overlap for a millennium. Core and core GS18 from the middle basin (museum number SR01GS9418) comes from the same station than core CP 18 . However preliminary dating on ostracod shells suggests that no overlap occur between the pilot and the Kullenberg cores due to severe losses at the top of the Kullenberg cores during corer penetration. The dinocyst assemblages of the middle basin core show a late change from slightly brackish water to more brackish water (as in the present) only at 4 cal. ka and not at the transition to the Holocene. The dinocyst assemblages of the southern core change at 9,5 cal. ka BP, but from the present day values of salinity (brackish) to a lower salinity. This period of lower salinity correspond to that seen at the base of core CP14, which terminates at c. 4 cal. ka BP. Therefore the two basins did not have the same water level history giving a possible role to the Apsheron sill.

CONCLUSIONS
In the absence of a complete palynological record for the Holocene, much remains to be done in the Caspian Sea. In the near future a transfer function for pollen and dinocysts should be developed at the scale of the whole sea. Palaeoclimatic records from continuous fine-grained marine cores covering a whole climatic cycle with robust age-depth model are cruelly needed [Cordova et al., 2009].