Improving Numerical Weather Prediction by Maximizing the Use of Assimilated Satellite Data.

An extensive literature review showed a disparity between the large amount of satellite data currently or potentially available for meteorological parameters and the relatively small amount that is actually utilized. The review also documented the various existing data retrieval techniques. Algorithms were developed and modified to retrieve temperature and moisture profiles from satellite infrared and millimeter/microwave radiances. Tools were developed, based on existing Air Force computer codes, to include and test the effect of satellite data in numerical weather prediction. The use of the logarithm of specific humidity as a forecast variable was investigated. The FGGE level II and III data sets existing at AFGL were completed to include all of the special observing periods and to form the basis of data assimilation tests. Finally, an attempt was made to optimize a cloudiness parameterization scheme on the basis of satellite retrievals. A number of recommendations are made for further work in satellite data retrieval, data assimilation and numerical weather prediction. They include the continuation of some of the items presented here, as well as suggestions for new avenues of research.