LAUSR

Abstract On 20 June 2016, a convective cell, which developed on the Polish-Slovakian border, under favorable kinematic and thermodynamic conditions, evolved into a supercell. It moved for 500 km, heading northeast, and remained isolated during its entire lifecycle. The storm was responsible for producing multiple large hail (up to 7.5 cm) and severe wind gust events. At around 1700 UTC (+/− 5 min), a tornado was reported near Sulow, and after moving for 21 km it dissipated near Skawinek (Lubelskie Voivodeship). Based on the damage survey, the peak intensity of the tornado was estimated at F1/T3 in the Fujita/TORRO scale. In this paper, we perform two simulations using the WRF model to evaluate the possibilities of predicting this event. The first simulation is a WRF base run, while the second version uses well established data assimilation techniques applied to the model domain with high spatial resolution. The surface observational data, together with the atmospheric soundings, are assimilated in the second model run. Three one-way nested domains were defined, with spatial resolutions of 9 km, 3 km and 1 km. Observations from meteorological stations, severe weather reports, and radar reflectivity from the POLRAD network are compared with the simulations to assess their quality. An additional parameter of an updraft helicity track is used to analyze the lifecycle of a simulated supercell. Results indicate a slight overestimation of temperature, dew point and wind speed. Our findings confirm that downscaling to convection-permitting resolution makes it possible to successfully predict tornadic supercells. Data assimilation, applied to the second model run, improved the model's performance. The model's errors, calculated by comparing the results with meteorological observation, were reduced if compared to the base run. Additionally, the number of convective cells and storm tracks was appropriate.