Holocene hydroclimate of the Volga Basin recorded in speleothems from the Central and Southern Ural Mountains, Russia

Abstract

<p>Hydroclimatic variability over the Volga River watershed (western Russia) strongly influences agricultural production, forest and grassland ecology, Caspian Sea level, and associated economic stability. Climate model forecasts of these variables suggest long-term increases in crop shortfalls and over-basin evaporation in the Volga and Caspian regions, respectively, but these projections currently lack validation from high-quality paleoclimate data. We present decadal-scale geochemical proxy data (δ<sup>13</sup>C, Mg/Ca, Sr/Ca) from <sup>230</sup>Th-dated stalagmites retrieved from four caves along a 640-km north-south gradient in the Ural Mountains, which collectively cover the entire Holocene (11.7 ka to present). Orbital trends in δ<sup>13</sup>C and Mg/Ca are broadly coherent between Ural speleothems and suggest that following Early Holocene aridity, warm-season precipitation was paced by summer insolation and has gradually declined since ~8 ka, consistent with model hindcasting. Centennial-scale variability, which is exceptionally well replicated between the Southern (Kinderlinskaya Cave) and Central (Geologov-3 Cave) Ural sites, supports a dynamic link between Volga hydroclimate and sea-surface temperature in the northeastern North Atlantic and Barents Sea. Important discrepancies exist, however, with data from the southernmost cave sites, which may be related to past migration of a strong latitudinal precipitation gradient that characterizes the modern basin and approximates the limit of North Atlantic influence. Finally, we conduct model-data comparisons and review our composite dataset in light of pollen- and soil-based proxies from across the Volga region, as well as reconstructions of Caspian Sea level. Our results provide a temporally well-constrained and spatially coherent portrait of Holocene hydroclimate for the Volga River watershed, thus constituting an important metric for future modeling studies.</p>