Abstract ID: 3.12425 | Accepted as Talk | Talk/Oral | TBA | TBA

Mountain regions are critical “water towers” for downstream ecosystems and human communities, yet they are increasingly affected by meteorological droughts due to climate change. The Alpine region is experiencing amplified warming, reduced snow accumulation, and shifts in precipitation patterns, all of which impact water availability both locally and downstream. In this context, seasonal predictions can provide early warning of extreme seasons. However, it is essential to i) assess the skill of state-of-the-art forecast systems in predicting meteorological anomalies, and ii) develop methods to model the snow-hydrological response to the predicted climate anomalies.

The PRIN-2022 SPHERE (Seasonal Prediction of water availability: enHancing watER sEcurity from high mountains to plains) project is developing a forecasting chain based on Copernicus seasonal forecast systems, integrating predictions of meteorological variables, downscaling methods and snow-hydrological models to simulate snow water equivalent (SWE), snow depth, and river discharge in the Alpine region at the kilometer resolution. This study evaluates the skill of three leading seasonal forecast systems—ECMWF System 5, Météo-France System 6, and CMCC SPS3—in predicting temperature and precipitation anomalies (input of the forecast chain) and it explores how the skill propagates through the modeling chain to SWE seasonal forecasts.

We develop a flexible analysis tool to assess skills of meteo-snow-hydrological seasonal forecasts using four key metrics: anomaly correlation coefficient (ACC), Brier score (BS), area under the ROC curve (AUC), and continuous ranked probability score (CRPS). The evaluation is conducted for winter and summer seasons over the period 1993–2014, comparing forecasts to ERA5 reanalysis as a reference. Particular focus will be given to SWE forecasts as a novel component, examining the skill propagation along the forecast chain.

This research contributes to advancing the use of seasonal forecasts for drought risk assessment in mountain regions, offering insights into how multi-model seasonal prediction can support hydrological forecasting and climate adaptation strategies.