Secondary ice production in Alpine mixed-phase clouds influencing Po River discharge

The Po River constitutes 28% of the total discharge from all rivers into the Adriatic Sea, considerably regulating the salinity of both the Adriatic and Mediterranean Seas. We propose an integrated high-resolution multi-model system for the Adriatic basin that allows us to represent the physics of the atmosphere (WRF with 6 km resolution), land surface (NOAH-MP), hydrology (WRF-Hydro with 600 m resolution), and their two-way feedbacks at ground level, covering a decadal time range. This system has already been shown to improve the net river discharge within the Adriatic Sea by accounting for over 140 catchments surrounding the basin, working in coupled mode with a mesoscale ocean model (NEMO). Moreover, it enabled us to demonstrate the critical role of decreasing river runoff in shaping the coastal water cycle under a changing climate (Verri et al., 2024).Many climate models underestimate the discharge of the Po River during specific seasons. We primarily attributed this underestimation to the precipitation in the Alps, which significantly affects the river flow. An accurate representation of the primary and secondary ice production (SIP) processes is essential for reducing the discharge bias to capture the observed precipitation statistics. Among the numerous identified SIP processes, most weather and climate models exclusively account for the Hallett-Mossop process, which occurs in mixed-phase clouds. We assess the impact of SIP, including droplet shattering and collisional breakup in the Alpine clouds, besides the Hallett-Mossop process, on the discharge of the Po River. Previous studies have shown that the collisional breakup from ice-ice collisions enhances the predicted ice crystal number concentration by up to three orders of magnitude in wintertime alpine mixed-phase clouds. We compare the WRF-NOAHMP+WRF-Hydro findings with the observed Po River discharge over a decade, from 2001 to 2010. Furthermore, we examine the significance of the seeder-feeder mechanism in alpine clouds, where ice particles sediment from the upper cloud to the lower cloud, accelerating glaciation and enhancing precipitation for the discharge of the Po River. The control experiment, which employs only Hallett-Mossop in the double-moment microphysics scheme, indicates that the Po River discharge is negatively biased by 27.5% after 5 years and 19.3% after 10 years. Moreover, the domain mean cloud fraction during this period is mostly negatively biased compared to MODIS-Terra, with a decadal mean RMSE of 13.9%. We perform sensitivity experiments to assess the decreased Po discharge bias when SIP mechanisms are implemented.