Extreme rainfall event in Crimea: Cloud-resolving modeling and radar observations

We present a case study on extreme rainfall event in Crimea in September 2018. The event was caused by extratropical cyclone forming above the Black Sea. The cyclone approached the Crimean Mountains from the south, producing over 100 mm of rainfall in Yalta on September 6 and causing a flash flood. In the mountains, about 140 mm of rainfall was reported. 

To study this extreme event, we use the WRF model v.4.0.1 forced by the boundary conditions from ECMWF operational analysis with the spatial resolution of approximately 10 × 10 km. The model was run for 8 days of September 1 – 8, and 5 microphysical schemes were tested (WDM6, Morrison, Milbrandt, NSSL, and Thompson). Other model parameters were set identical to CONUS configuration suite. The simulations were done for two one-way nested convective-resolving domains with spatial resolution of 2.7× 2.7 km and 0.9 × 0.9 km. The simulations were verified using the meteorological radar observations in Simferopol airport and GPM measurements.

All of the microphysical schemes substantially underestimate the amount of rainfall reaching the ground compared to observations. However, several schemes (Milbrandt, Morrison, and WDM6) do add value to the forecasts, producing significantly larger amount of rainfall compared to the driving model that almost completely missed it on the local scale. WDM6 performs best to capture the proper location of the squall line and to reproduce the rainfall orographic enhancement in the mountains. The amount of rainfall in the child domain was also slightly larger compared to the parent one. Despite the rainfall underestimation, we also show that the simulated reflectivity patterns are in good agreement with observations, although the convective cores are wider and less intense compared to the observed by the radar.