Local ensemble transform Kalman filter experiments using radar observations: a case study over central Argentina

Very short-term numerical weather forecasts are useful for identifying potential development areas of high impact weather events associated with deep convections (e.g. convective systems, strong winds, tornados, etc.) that can significantly affect people and their activities. One way to generate adequate forecasts is to use high spatial resolution numerical models (with horizontal resolutions O(1 km) or higher), which must be initialized from the initial conditions containing information at the meso-gamma scale. The convective initiation is strongly influenced by the areas of low-level convergence and divergence, and this environmental information can be included in the initial conditions by using Doppler radar observations. For this reason, the development of radar data assimilation systems in the areas frequently affected by high impact weather events associated with deep convections should be a priority. The aim of this work is to present the advances in the development and implementation of a radar data assimilation system that takes advantage of the developing radar network called “SINARAME” in Argentina. The system can assimilate reflectivity and Doppler velocity and is based on the Local Ensemble Transform Kalman Filter coupled with the Weather Research and Forecasting (WRF) model. The system is run at a horizontal resolution of 2 km and with 60 ensemble members over a 240x240 km domain. To test the system's performance, an extreme event on January 11th 2010 which is characterized by the occurrence of deep organized convections over central Argentina is investigated. Observations are taken at the Anguil radar site with a temporal frequency of 10 minutes. A quality control based on polarimetric variables is implemented to remove ground clutters while preserving clear air radial velocity data. Results obtained with the LETKF-WRF radar assimilation system as well as a preliminary evaluation of the very short range, high resolution, ensemble forecasts will be presented at the conference.