Use of the Information Content in Satellite Measurements for an Efficient Interface to Data Assimilation

Abstract

Abstract Satellite observations are the most assimilated data type by operational meteorological centers. Spaceborne instruments can make measurements all over the globe and provide observations for assimilation even where the coverage of other data is poor. It is therefore most important that such observations, which are only indirectly related to the state of the atmosphere, are assimilated as optimally as possible. In this study, a detailed characterization of both retrievals and observed radiances for assimilation is provided, along with an error analysis. A method for assimilating remote sounding data while preserving its information content is presented. The main features of the technique are as follows: (i) the retrieval–forecast error cross covariance is removed even when the retrieval is severely constrained by a priori information, (ii) the radiative transfer calculations for radiance assimilation are done offline, and (iii) the number of assimilated quantities per observation is reduced to the number of effective degrees of freedom in the observation.