Résumé
The study of past climate variability from ice core investigations has been largely developed both in polar areas over the past decades and, more recently, in tropical regions, specifically along the South American Andes between 0° and 20°S. However a large gap still remains at mid-latitudes in the Southern Hemisphere. In this framework, a 15.3-m long shallow firn core has been extracted in March 2005 from the summit plateau of Monte San Valentín (3747 m, 46°35′S, 73°19′W) in the Northern Patagonia Icefield to test its potential for paleoclimate and paleoenvironmental reconstructions. The firn temperature is −11.9°C at 10-m depth allowing to expect well preserved both chemical and isotopic signals, unperturbed by water percolation. The dating of the core, on the basis of a multi-proxy approach combining annual layer counting and radionuclide measurements, shows that past environment and climate can be reconstructed back to the mid-1960s. A mean annual snow accumulation rate of 36 ± 3 cm year−1 (i.e., 19 ± 2 g cm−2 year−1) is inferred, with a snow density varying between 0.35 and 0.6 g cm−3, which is much lower than accumulation rates previously reported in Patagonia at lower elevations. Here, we present and discuss high-resolution profiles of the isotopic composition of the snow and selected chemical markers. These data provide original information on environmental conditions prevailing over Southern Patagonia in terms of air masses trajectories and origins and biogeochemical reservoirs. Our main conclusion is that the San Valentín site is not only influenced by air masses originating from the southern Pacific and directly transported by the prevailing west winds but also by inputs from South American continental sources from the E–NE, sometimes mixed with circumpolar aged air masses, the relative influence of these two very distinct source areas changing at the interannual timescale. Thus this site should offer a wealth of information regarding (South) Pacific, Argentinian NE–E areas and Antarctic climate variability.