Wet snow avalanches on Corsica Island: a long-term perspective from lake sediments

Mountain areas are very sensitive to climate change, which has led to changes in natural hazards that are often linked to disturbances in the cryosphere. In this context, changes in snowfall characteristics and snow cover affect avalanche hazards. Long-term variability can be reconstructed by using historical archives, tree rings and, more rarely, lake sediments. The latter approach is based on the identification of lake sediment consisting of poorly sorted, coarse sediments in a fine matrix, which are often associated with terrestrial organic debris. This sediment is generally brought to the lake within large amounts of wet snow, or via ‘drop stones’ when the ice melts if the avalanche takes place on a frozen surface. Here, we study two high-altitude lakes (Melu and Capitellu) in the Restonica Valley in Corsica, an area where systematic records are lacking, to reconstruct signals related to such large wet snow flows on a millennial scale. The analysis of several sedimentological and geochemical markers enables the characterization of wet avalanche deposits in the two lakes, as well as turbidite-type facies linked to historical earthquakes in Corsica. Age models based on short-lived radionuclides and radiocarbon also make it possible to reconstruct two avalanche chronologies covering 600 and 1750 years in Melu and Capitellu Lakes, respectively, which show similar temporal variations. A comparison with the only long-term chronology available in the Alps (Lake Muzelle, Ecrins) also reveals synchronicity in the secular variability of avalanches, suggesting a common forcing between Corsica and the Alps. Human observations and accident records from recent decades, and snow and weather release conditions reconstructed from a hydrological modeling scheme confirm the ability of the lacustrine avalanche sedimentary method to document local wet snow avalanche activity. The proposed methodology, which is based on paleolimnological studies, may therefore be useful, alone or combined with other avalanche data sources, for tracking changes in avalanche activity and related risks in mountainous areas.