ESTIMATIVA OPERACIONAL DA UMIDADE DO SOLO PARA INICIAÇÃO DE MODELOS DE PREVISÃO NUMÉRICA DA ATMOSFERA PARTE I: DESCRIÇÃO DA METODOLOGIA E VALIDAÇÃO

OPERATIONAL SOIL MOISTURE ESTIMATE FOR INITIALIZATION OF NUMERICAL WEATHER FORECAST MODELS PART I: MODEL DESCRIPTION AND VALIDATION This work describes an operational numerical model to estimate soil moisture for initialization in weather forecast models. This project aims at improving the quality of the weather forecast model outputs, especially in precipitation prediction and the representation of the planetary boundary layer. The soil moisture estimation is obtained by providing satellite precipitation estimates to a simple hydrological model. This technique considers current vegetation and soil textural class maps, which describe both bio-physiologic and soil properties in the model domain. The final product is the spatial structure in eight soil layers of the ratio between liquid water volume and soil volume. The modelled soil moisture is able to coherently reproduce the expected patterns throughout the seasons over South America, with satisfactory agreement on time variability. Results also show that the horizontally heterogeneous soil moisture field initialization in the BRAMS simulation improves the weather forecast quality.