ASSIMILATION OF SATELLITE PRECIPITABLE WATER IN A METEOROLOGICAL FORECAST MODEL

The lack of local humidity observations over a large portion of the gl obe hinders any improvement of humidity forecasting in meteorological models. However, satellite microwave radiometers routinely provide fie lds of precipitable water content over the oceans, with a horizontal r esolution of a few tens of kilometers. The aim of this paper is theref ore to establish the feasibility of the assimilation of microwave radi ometer precipitable water content in a forecast model. This preliminar y study presents the assimilation method, which is based on optimal in terpolation, and results of analysis and forecast runs over 12 h for o ne case. The Special Sensor Microwave/Imager (SSM/I) radiometer (the s canning microwave radiometer aboard the U.S. Navy Defense Meteorologic al Satellite Program satellite) and the Prevision a Echeance Rapproche e, Integrant des Donnees Observees et Teledetectees (PERIDOT) model (o perational limited-area model of the French Meteorological Office) wer e used to perform this test. We show that assimilation of SSM/I data h as a significant impact on the forecast, since the resulting fields, e specially the humidity field, are more coherent with the next SSM/I da ta and radiosonde observations. The SSM/I assimilation is performed fi rst once only, then throughout a 24-h assimilation cycle. Also, the mo del sensitivity is established to the vertical error in humidity in th e optimal interpolation.