Glacier response to current climate change and future scenarios in the northwestern Italian Alps

We analyze longtime series of annual snout positions of several valley glaciers in the northwestern Italian Alps, together with a high-resolution gridded dataset of temperature and precipitation available for the last 50 years. Glacier snout fluctuations are on average negative during this time span, albeit with a period of glacier advance between about 1970 and 1990. To determine which climatic variables best correlate with glacier snout fluctuations, we consider a large set of seasonal predictors, based on our climatic dataset, and determine the most significant drivers by a stepwise regression technique. This in-depth screening indicates that the average glacier snout fluctuations strongly respond to summer temperature and winter precipitation variations, with a delay of 5 and 10 year, respectively. Snout fluctuations display also a significant (albeit weak) response to concurrent (same year) spring temperature and precipitation conditions. A linear regressive model based on these four climatic variables explains up to 93 % of the variance, which becomes 89 % when only the two delayed variables are taken into account. When employed for out-of-sample projections, the empirical model displays high prediction skill, and it is thus used to estimate the average glacier response to different climate change scenarios (RCP4.5, RCP8.5, A1B), using both global and regional climate models. In all cases, glacier snout fluctuations display a negative trend, and the glaciers of this region display an accelerated retreat, leading to a further regression of the snout position. By 2050, the retreat is estimated to be between about 300 and 400 m with respect to the current position. Glacier regression is more intense for the RCP8.5 and A1B scenarios, as it could be expected from the higher severity of these emission pathways.