Impact of Radar Data Assimilation on the Numerical Simulation of a Severe Storm in Croatia

Abstract

A severe thunderstorm hit the northwestern part of Croatia in the late afternoon and evening of 24 June 2008. This severe event is used as a test case for the Advanced Regional Prediction System (ARPS) high resolution numerical prediction model and for exploring the impact of assimilating conventional and radar data. Radar radial velocity data were assimilated using three-dimensional variational analysis (3DVAR). Radar reflectivity data were used through a cloud analysis procedure where hydrometeors and cloud fields are defined, and adjustments to the in-cloud temperature and moisture fields are made. Results show that without data assimilation, the models were not able to represent the development of the storm nor the proper environment for it. Assimilation of surface observations in the mesoscale outer model provided spatial distribution of convection ingredients that established a proper environment for storm initiation and propagation. Without that, the inner storm-scale model, even with radar data assimilation, is unable to simulate storm development. Using the outer model with assimilation of surface data in combination with an inner model including assimilation of radar data provided the best simulation of storm initiation and development.

Keywords

• radar data assimilation
• severe storm
• numerical weather prediction
• ARPS