Analysis and forecast experiments incorporating satellite soundings and cloud and water vapor drift wind information

A system for assimilating conventional meteorological data and satellite-derived data in order to produce four-dimensional gridded data sets of the primary atmospheric variables used for updating limited area forecast models is described. The basic principles of a data assimilation scheme as proposed by Lorenc (1984) are discussed. The design of the system and its incremental assimilation cycles are schematically presented. The assimilation system was tested using radiosonde, buoy, VAS temperature, dew point, gradient wind data, cloud drift, and water vapor motion data. The rms vector errors for the data are analyzed.