Parameter sensitivity of the LETKF-WRF system for assimilation of radar observations in a case of deep convection in Argentina

JUAN RUIZ; CELESTE SAULO; PAULA MALDONADO

Bergen

Workshop; 12th International EnKF Workshop; 2017

Nansen Environmental and Remote Sensing Center (NERSC)

Very short-term weather forecasts are useful for identifying areas of high-impact weatherevents associated with deep convection that can significantly affect people?s lives. One wayto generate adequate forecasts is by using high-spatial-resolution numerical modelsinitialized with high temporal and horizontal resolution observations, as those provided byweather radars. Thus, a data assimilation system capable of assimilating radar data isessential to improve very short-term weather forecasts. The aim of this work is to comparedifferent approaches for the estimation of the error covariance matrix by using the LocalEnsemble Transform Kalman Filter (LETKF) coupled with the Weather Research andForecasting (WRF) model for a case of deep convection over central Argentina. Severalobserving system simulation experiments were performed by assimilating synthetic radarobservations, including reflectivity and Doppler velocity. In these experiments we studied thesensitivity to i) the specification of the initial and boundary perturbations, ii) the type andmagnitude of the multiplicative inflation and iii) the localization scale. The best configurationof the data assimilation system was used to assimilate real radar data from the Anguil radarsite (La Pampa, Argentina) for an event of deep organized convection. Results obtained willbe presented at the workshop.