Sensitivity experiments to design a regional assimilation system combining the LETKF and the WRF model

MARCOS SAUCEDO; JUAN RUIZ; CELESTE SAULO

Montreal

Congreso; World Weather Open Science Conference; 2014

World Meteorological Organization

This work is motivated by the necessity of developing a regional forecast and data assimilation system adequate to be run operationally by the Argentine Weather Service. To accomplish this goal our strategy is based on the combination of the Local Ensemble Transform Kalman Filter (LETKF) implemented in the Weather Research and Forecasting model (WRF). A series of sensitivity experiments using OSSE approach have been designed to analyze the error growth and evaluate the optimum combination of model and boundary settings. The main purpose is to produce an adequate ensemble spread taking into account the different error sources. This set of experiments has been done with 40 ensemble members, adapting covariance inflation and relatively low resolution (aprox. 100 km horizontal grid spacing) given limited computational resources available. Experiments include: perfect and imperfect model with perfect and imperfect boundary conditions (i.e. 5 different cases). Perfect model / Perfect boundary refer to the same model configuration and boundary conditions used to generate the ?true state?. The imperfect model is obtained changing model parameterizations with respect to the ?true state? run, while the imperfect boundary conditions are provided by introducing errors in the lateral boundary conditions. Results show that, in general, adaptive inflation makes a good job in adjusting the ensemble spread, still, there is a need to implement additional approaches like using diverse parameterizations and boundaries perturbations to better represent the different sources of error.