Study of the variations in the Antarctic snow accumulation and its links with climate change in the Southern Hemisphere over the past centuries

The Antarctic Ice Sheet (AIS) is the biggest reservoir of fresh water on Earth. During the 20th century, the AIS has gained mass at its surface through an enhanced snow accumulation. However, our understanding of the Antarctic snow accumulation variability is still limited, which directly impacts on our confidence in the future contribution of the AIS to the sea-level rise. The main objective of this thesis is to highlight the physical mechanisms behind snow accumulation variability by investigating the links between snow accumulation and other variables characterizing the surface climate at high southern latitudes. The present work is based on various types of data – instrumental observations, paleoclimate records and climate model simulations –, focuses on different time periods – from the last millennium to the 21st century – and uses diverse methodologies, such as data assimilation. Our results show that the variability in snowfall and temperature are strongly related over Antarctica, and therefore that the classical Clausius-Clapeyron relationship explains relatively well the multi-decadal, regional to continental-scale snow accumulation changes. Our work also highlights that data assimilation is particularly appropriate for reconstructing the surface climate changes in the Antarctic over the past centuries as the observational network is sparse and includes various types of records and because of the non-stationarity of the link between the proxy records and reconstructed variables. With this technique, we have been able to provide robust past estimates of temperature and snow accumulation but also of variables that cannot be directly obtained from the records themselves such as the atmospheric circulation and even the sea-ice coverage.