CLOUDS ARE RECOGNIZED AS THE MAIN MODULATOR OF EARTH'S RADIATION BUDGET AS THE ATMOSPHERIC CONSTITUENT CARRYING THE LARGEST UNCERTAINTY IN FUTURE CLIMATE PROJECTIONS. THE GOAL OF THE PROPOSED PROJECT IS TO CREATE A NEW FRAMEWORK FOR EVALUATING CLOUDS AND THEIR RADIATIVE EFFECTS IN ATMOSPHERIC GLOBAL CLIMATE MODELS (AGCMS) USING SPACE-BASED CLOUD AND RADIATIVE FLUX OBSERVATIONS. THE PROPOSED EFFORT WILL EXTEND AND IMPROVE EXISTING METHODS BUT WILL ALSO CREATE NEW EVALUATION PARADIGMS. OUR PROPOSED EVALUATION METHODS WILL BE APPLIED ON GSFC'S OWN AGCM GEOS-5 TO WHICH WE HAVE FULL ACCESS THAT ALLOWS US TO MODIFY THE SOURCE CODE AND RUN OUR OWN EXPERIMENTS AND ALSO ON THE CLOUD OUTPUT OF CMIP6'S ARCHIVE WHERE SOME LIMITATIONS ARE ANTICIPATED BUT WHICH OFFERS THE ADVANTAGE OF SIMULTANEOUS EVALUATION AND COMPARISON OF MULTIPLE MODELS. WE PROPOSE TO: (A) USE ISCCP AND MODIS GLOBAL CLOUD REGIMES CAPTURING THE MAIN MODES OF CLOUD TOP PRESSURE AND CLOUD OPTICAL THICKNESS COVARIATIONS ACROSS THE GLOBE TO EVALUATE THE JOINT HISTOGRAMS OF THESE TWO CLOUD VARIABLES AS GENERATED BY CMIP6 AGCMS RUNNING COSP'S ISCCP AND MODIS SIMULATORS. WHENEVER JOINT HISTOGRAM OUTPUT IS NOT AVAILABLE WE WILL GENERATE IT OURSELVES USING CLOUD PROPERTY PROFILES AS INPUT TO AN OFFLINE VERSION OF COSP. (B) EXAMINE THE SENSITIVITY OF RESULTS OBTAINED AS ABOVE TO THE CLOUD GENERATOR USED. THIS ENTAILS REPLACING THE CLOUD GENERATOR CURRENTLY USED WITH THE ONE WE HAVE PAIRED IN GEOS-5 WITH THE RRTMG RADIATIVE TRANSFER CODE. THIS GENERATOR HAS ADDITIONAL CAPABILITIES SUCH AS INHOMOGENEOUS LAYERS OF CLOUD CONDENSATE AND PARAMETERS (THE VALUES OF WHICH ARE GUIDED ACTIVE CLOUD REMOTE SENSING) THAT CONTROL HOW CLOUD FRACTION AND DISTRIBUTIONS OF CONDENSATE OVERLAP IN THE VERTICAL. AGAIN THE OFFLINE VERSION OF COSP WILL BE USED. (C) ADD A NEW DIMENSION TO THE CLOUD EVALUATION FRAMEWORK THAT FOCUSES ON CLOUD VERTICAL STRUCTURE (CVS). AS A REFERENCE DATASET WE WILL USE ONE RECENTLY CREATED FROM CLOUDSAT'S 2B-CLDCLASS-LIDAR PRODUCTS THAT ASSIGNS OBSERVED CLOUD VERTICAL CONFIGURATIONS TO NINE SIMPLIFIED CVS CLASSES BASED ON CLOUD CO-OCCURRENCE IN THREE STANDARD ATMOSPHERIC LAYERS. AGAIN SENSITIVITY TO CLOUD GENERATOR CHOICE WILL BE EXAMINED. (D) COMPARE THE RADIATIVE EFFECTS OF CLOUDS WITH OBSERVATIONS ACCORDING TO THEIR CVS AN EFFORT THAT IS ONLY FEASIBLE FOR GEOS-5. WE WILL COMPOSITE THE CLOUD RADIATIVE EFFECT BY CVS CLASS AND WILL COMPARE WITH AN EXISTING REFERENCE DATASET BASED ON CLOUDSAT'S 2BFLXHR- LIDAR PRODUCT. THIS WILL ENABLE EXAMINATION OF WHETHER THE CONTRIBUTION OF THE VARIOUS CVS CLASSES TO THE EARTH'S RADIATION BUDGET IN GEOS-5 AGREES WITH SATELLITE OBSERVATIONS. AT THE CONCLUSION OF THE PROJECT OUR KNOWLEDGE OF CLOUD REPRESENTATION IN AGCMS WILL HAVE SUBSTANTIALLY IMPROVED AND NEW PATHWAYS WILL BE AVAILABLE FOR GLOBAL MODELERS TO IMPLEMENT IN THEIR OWN EVALUATION OF MODELS THEY DEVELOP AND ARE RESPONSIBLE FOR.