Zbynek Sokol, Institute of Atmospheric Physics CAS

Nowcasting of hailstorms by the COSMO NWP model

Pesice Petr


The contribution deals with the evaluation of the possibility of using the current numerical models for nowcasting of large hail (diameter ≥ 2.5 cm). We apply the COSMO NWP model with a two moment cloud microphysics with six hydrometeors including hail. The model is run at a horizontal resolution of 1.1 km and with the assimilation of radar reflectivity. The model forecasts are evaluated for several events with large hail determined by direct observations and derived using a recognition algorithm based on radar and sounding data. The accuracy of model forecasts was evaluated subjectively by eye, by Fractions Skill Score and by bias. Distances between locations of forecasted and observed hail were also analysed and evaluated. We will demonstrate that (i) forecasts up to 90 minutes brings additional information, which can be used to improve nowcasting of hailstorms, however for longer lead times the usefulness of model forecasts is problematic because of large errors and significant underestimation of forecasted hail, (ii) when the model forecasts large hail then there is a 50% chance that large hail will be observed within approximately 30 km.

File available here