The Impact of Radar Data Assimilation for Simulate a Supercell in Brazil Southern

Analysis and studies of supercell in Brazil are usually addressed in a synoptic context because it has no high-resolution measurements. One of the purpose of radar data assimilation is to provide better initial conditions for high resolution explicit convective storm simulations and detect dynamical and thermodynamical structures in convective systems. On 12 July 2016, a supercell was registered by the S-Band Dual-Pol Weather Radar of Cascavel in Southern Brazil. This supercell produced a downburst that blown down a high-voltage transmission tower in the region. So, the purpose of this study is analyzing the impact of radar data assimilation on the WRF model with different assimilation methods to simulate the genesis and the propagation of the supercell.

For this experiments, it has been used the WRF model with 3 km horizontal resolution, initial and boundary conditions from Global Forecast System (GFS) and in the initialization of the model the Digital filter initialization (DFI) is used with option Dolph .For the first experiment has been tested the 3DVAR techniques, from WRF Data Assimilation of National Center for Atmospheric Research (NCAR), with Ensemble physic and dynamic method, 25 ensemble members as tested, and other with 4dvar techniques, also from NCAR, with Ensemble physic and dynamic method, 25 ensemble members, has been used a 6h radar data window, the data come from the radar of Cascavel, Parana-Brazil(S-Band) and the radar of Asunción-Paraguay(S-Band), while in 3DVAR the data was assimilated in warming way.

The preliminary results shows if the WRF model is running with just initial conditions from GFS without assimilation, model was not able to simulate the supercell, yielding a scattered convection. In case of the ensemble in 3dvar and 4dvar the result is better, can reproduce characteristics more similar to those observed in radars data. To gain better insight into the evolution of the storms the radar data assimilation shows essentials in this preliminary results.