Application of the WRF-LETKF Data Assimilation System over Southern South America: Sensitivity to Model Physics

Improving the initial conditions of short-range numerical weather prediction (NWP) models is one of the main goals of the meteorological community. Development of data assimilation and ensemble forecast systems is essential in any national weather service (NWS). In this sense, the local ensemble transform Kalman filter (LETKF) is a methodology that can satisfy both requirements in an efficient manner. The Weather Research and Forecasting (WRF) Model coupled with the LETKF, developed at the University of Maryland, College Park, have been implemented experimentally at the NWS of Argentina [Servicio Meteorológico Nacional (SMN)], but at a somewhat lower resolution (40 km) than the operational Global Forecast System (GFS) at that time (27 km). The purpose of this work is not to show that the system presented herein is better than the higher-resolution GFS, but that its performance is reasonably comparable, and to provide the basis for a continued improved development of an independent regional data assimilation and forecasting system. The WRF-LETKF system is tested during the spring of 2012, using the prepared or quality controlled data in Binary Universal Form for Representation of Meteorological Data (PREPBUFR) observations from the National Centers for Environmental Prediction (NCEP) and lateral boundary conditions from the GFS. To assess the effect of model error, a single-model LETKF system (LETKF-single) is compared with a multischeme implementation (LETKF-multi), which uses different boundary layer and cumulus convection schemes for the generation of the ensemble of forecasts. The performance of both experiments during the test period shows that the LETKF-multi usually outperforms the LETKF-single, evidencing the advantages of the use of the multischeme approach. Both data assimilation systems are slightly worse than the GFS in terms of the synoptic environment representation, as could be expected given their lower resolution. Results from a case study of a strong convective system suggest that the LETKF-multi improves the location of the most intense area of precipitation with respect to the LETKF-single, although both systems show an underestimation of the total accumulated precipitation. These preliminary results encourage continuing the development of an operational data assimilation system based on WRF-LETKF at the SMN.Fil: Dillon, María Eugenia. Ministerio de Defensa. Secretaria de Planeamiento. Servicio Meteorológico Nacional; Argentina. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales. Departamento de Ciencias de la Atmósfera y los Océanos; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; ArgentinaFil: Garcia Skabar, Yanina. Ministerio de Defensa. Secretaria de Planeamiento. Servicio Meteorológico Nacional; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; ArgentinaFil: Ruiz, Juan Jose. Consejo Nacional de Investigaciones Científicas y Técnicas. Oficina de Coordinación Administrativa Ciudad Universitaria. Centro de Investigaciones del Mar y la Atmósfera. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales. Centro de Investigaciones del Mar y la Atmósfera; ArgentinaFil: Kalnay, Eugenia. University of Maryland; Estados UnidosFil: Collini, Estela Angela. Ministerio de Defensa. Armada Argentina. Servicio de Hidrografía Naval; Argentina. Ministerio de Defensa. Secretaria de Planeamiento. Servicio Meteorológico Nacional; ArgentinaFil: Echevarria, Pablo. Ministerio de Defensa. Secretaria de Planeamiento. Servicio Meteorológico Nacional; ArgentinaFil: Saucedo, Marcos Adolfo. Consejo Nacional de Investigaciones Científicas y Técnicas. Oficina de Coordinación Administrativa Ciudad Universitaria. Centro de Investigaciones del Mar y la Atmósfera. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales. Centro de Investigaciones del Mar y la Atmósfera; ArgentinaFil: Miyoshi, Takemasa. RIKEN Advanced Institute for Computational Science; JapónFil: Kunii, Masaru. RIKEN Advanced Institute for Computational Science; Japó