Author(s): Petr Vohnicky; Eleonora Dallan; Francesco Marra; Nadav Peleg; Marco Borga
Linked Author(s):
Keywords: Future hydrology; Hydrological modelling; Stochastic weather generator
Abstract: Climate change is intensifying precipitation extremes globally, yet the hydrological consequences for flood risk and seasonal runoff remain uncertain due to complex interactions between precipitation, snow dynamics, soil moisture, and catchment characteristics. Impact assessment requires long meteorological time series that capture both natural variability and climate trends. Convective-permitting climate models provide physically realistic sub-daily precipitation, but are computationally limited to decadal simulations. Stochastic weather generators offer long synthetic records but typically lack explicit climate change integration. We address this gap by reparametrizing the AWE-GEN-2d weather generator for future conditions using factors of change from six convective-permitting simulations under RCP8.5. The framework is applied to alpine catchments in Italy's Upper Adige River Basin. Runoff is simulated using the semi-distributed model ICHYMOD, with snow processes estimated by TOPMELT snowpack model. Hydrological simulations driven by generated data reproduce observed runoff patterns with minor seasonal deviations. Future projections reveal decreased snow water equivalent and intensified summer extreme precipitation, leading to increased winter runoff from reduced snow storage and elevated flash-flood risk in small, fast-responding catchments during summer. This integrated approach highlights the value of combining convective-permitting projections with stochastic weather generators to assess future runoff, extreme flows, and snow contributions in high-elevation catchments.
Year: 2026