CLOUD PROPERTY DETERMINATION FROM MULTISPECTRAL DMSP OLS

THE PROPOSED STUDY WILL DETERMINE CLOUD PROPERTIES FROM SIMULATED MULTISPECTRAL DMSP OLS DATA BY COMBINING A SCENE MODEL APPROACH WITH A VARIETY OF CLOUD PARAMETER RETRIEVAL METHODS. THE MULTISPECTRAL OLS DATA WILL BE CREATED BY RESAMPLING LANDSAT THEMATIC MAPPER SPATIALLY AND SPECTRALLY. THE PROPOSED SENSOR HAS SEVERAL SPECTRAL BANDS WHICH, IN PRINCIPLE, ALLOW CLOUD AMOUNT AND HEIGHT TO BE RETRIEVED. CLOUD PARTICLE SIZE, CLOUD PHASE, AND CLOUD THICKNESS MAY ALSO BE RETRIEVABLE. THE PROPOSED SCENE MODEL WILL ACCOUNT FOR PARTIAL BEAM FILLING AND IDENTIFY HOMOGENEOUS AREAS IN THE SCENE, THEREBY ALLOWING EFFICIENT COMPOSITING FOR CLOUD PARAMETER RETRIEVALS. WE PLAN TO UTILIZE A VARIETY OF CLOUD PARAMETER RETRIEVAL METHODS INCLUDING BISPECTRAL, THRESHOLD, AND STATISTICAL METHODS. AS WE DEVELOP AND APPLY PARTICULAR ALGORITHMS, WE WILL PERFORM A VARIETY OF RADIATIVE TRANSFER CALCULATIONS TO OPTIMIZE THESE METHODS.