THE MAIN GOALS OF THIS PROJECT ARE (1) TO DETERMINE THE OPTIMAL CONFIGURATION OF EXISTING FAST MULTIPLE SCATTERING RADIATIVE TRANSFER (RT) MODELS FOR BOTH RAPID AND ACCURATE CALCULATION OF ALL-SKY MICROWAVE (MW) AND INFRARED (IR) RADIANCES IN THE COMMUNITY RADIATIVE TRANSFER MODEL (CRTM) AND (2) DEMONSTRATE THE IMPACT OF THESE IMPROVEMENTS IN THE NASA GODDARD EARTH OBSERVING SYSTEM MODEL VERSION 5 (GEOS-5) DATA ASSIMILATION SYSTEM (DAS). EXPERIENCE HAS SHOWN THAT THE TWO MULTI-STREAM MODELS CURRENTLY IN THE CRTM WHILE FAST AND ACCURATE STILL LACK THE NECESSARY RAW SPEED FOR USE IN OPERATIONS. THE BASIC ISSUE IS THAT THESE MODELS WHILE FLEXIBLE ARE TYPICALLY SLOWER THAN THEIR 2-STREAM/4-STREAM/ETC. ANALYTIC COUNTERPARTS SO IT IS ALSO IMPORTANT TO CONSIDER THE PERFORMANCE OF LOW-ORDER-STREAM ANALYTIC METHODS FOR POSSIBLE USE IN DATA ASSIMILATION SYSTEMS. THIS PROJECT WILL ASSEMBLE A SMALL NUMBER OF EXTREMELY FAST MODELS (MAINLY ANALYTIC AND ITERATIVE METHODS) AS WELL AS THE CURRENT MODELS IN THE CRTM TO DETERMINE WHICH COMBINATION PROVIDES THE BEST ACCURACY AND SPEED UNDER DIFFERENT CLOUD/PRECIPITATION/ AEROSOL CONDITIONS FOR A VARIETY OF OPERATIONAL MW AND IR SATELLITE SENSORS AND FOR DIFFERENT ACCURACY REQUIREMENTS. HIGH-RESOLUTION NUMERICAL WEATHER PREDICTION (NWP) MODEL OUTPUT WILL BE THE SOURCE OF REALISTIC CLOUD AND AEROSOL PROFILES NEEDED IN THE BENCHMARK TESTS. THE SELECTED MODELS AND THEIR ASSOCIATED TANGENT LINEAR AND ADJOINT FORMS WHICH IN SOME CASES WILL NEED TO BE DEVELOPED WILL BE INTEGRATED INTO THE CRTM AND THOROUGHLY TESTED. FINALLY SEVERAL GLOBAL GEOS-5 DAS EXPERIMENTS WILL BE DONE TO QUANTIFY THE PERFORMANCE OF THE NEW OPTIMIZED CRTM AGAINST THE UNOPTIMIZED CRTM VIA THE COMMUNITY GRIDPOINT STATISTICAL INTERPOLATION (GSI) SYSTEM.