Uncertainty in the representation of snowpack over the Alps in CMIP5 global climate models

Global Climate Models (GCM) still have too coarse spatial resolution to reproduce the small-scale variability of snowpack in orographically complex areas, nevertheless they are a unique tool to simulate past and future climate. The quantification of the uncertainties in GCM simulations is essential to critically analyze future climate change projections. In this work, we evaluate the representation of the snowpack over the Greater Alpine Region in the state-of art Global Climate Models participating in the Coupled Models Intercomparison Project phase 5 (CMIP5). We compare the spatial distribution, annual cycle and temporal variability of snow depth represented by the GCMs to the major reanalysis products, i.e. Climate Forecast System Reanalysis, Modern EraRetrospective analysis for Research and Applications, ERA-Interim/Land, and 20th Century Reanalysis, highlighting common features and discrepancies. We explore the projected changes in snow depth for the XXI century in RCP4.5 and RCP8.5 scenarios. Our analysis shows that the spread among both the global climate models and the reanalyses is remarkable, indicating high uncertainty in the snow depth estimation. Nevertheless, GCMs are in overall agreement in indicating a strong decrease of the snow depth in the Alpine region especially in the second half of the XXI century.