IMPROVED SPATIAL RESOLUTION SEA SURFACE TEMPERATURE FIELDS FROM AMSR-E WITH THE EXCEPTION OF A RELATIVELY SHORT GAP BETWEEN THE FAILURE OF AMSR-E (4 OCTOBER 2011) AND THE BEGINNING OF THE AMSR2 OPERATIONAL PERIOD (2 JULY 2012), THE SUITE OF AMSR INSTRUMENTS HAVE PROVIDED A NEARLY CONTINUOUS TIME SERIES OF DAYTIME (1330 LOCAL SUN TIME (LST)) AND NIGHTTIME (0130 LST) SEA SURFACE TEMPERATURE (SST) FIELDS FROM 2 JUNE 2002 TO PRESENT. BECAUSE THESE RADIOMETERS SAMPLE IN THE MICROWAVE, COMPLETE GLOBAL COVERAGE IS ACHIEVED DAILY WITH THE EXCEPTION OF REGIONS OF HEAVY RAINFALL, REGIONS OF PERSISTENT RADIO FREQUENCY INTERFERENCE (RFI) AND RELATIVELY SMALL DIAMOND SHAPED REGIONS AT LOW LATITUDES RESULTING FROM ORBITAL GAPS, GAPS WHICH ARE GENERALLY FILLED THE FOLLOWING DAY. THIS GLOBAL DAILY SST RECORD PROVIDES ESSENTIAL INFORMATION ON OCEAN PHENOMENA AND IS ESPECIALLY VALUABLE IN CLOUD CONTAMINATED REGIONS WHERE INFRARED SST RETRIEVALS ARE DEGRADED OR UNATTAINABLE. BUT MANY OF THESE PERSISTENTLY CLOUDY REGIONS ARE CHARACTERIZED BY HIGH TEMPORAL AND SPATIAL VARIABILITY, WHICH IS NOT CAPTURED IN THE CURRENT SUITE OF MICROWAVE DATA PRODUCTS. WE PROPOSE TAKING ADVANTAGE OF THE OVERSAMPLED MICROWAVE DATA FOR THE AMSR INSTRUMENTS TO CREATE A SUBSTANTIALLY HIGHER RESOLUTION MICROWAVE SST DATA RECORD. SO HOW DOES THE RETRIEVAL ALGORITHM WORK? SST ARE RETRIEVED SIMULTANEOUSLY WITH SURFACE WIND SPEED, ATMOSPHERIC WATER VAPOR, CLOUD LIQUID WATER, AND RAIN RATE FROM AMSR-ES 5 LOWEST FREQUENCY SPECTRAL BANDS. (ALTHOUGH RETRIEVAL FOR AMSR2 IS SIMILAR, WE FOCUS ON AMSR-E HERE FOR BREVITY.) THE FOOTPRINTS OF THESE SPECTRAL BANDS RANGE FROM 14 8 KM AT 36.5 GHZ TO 70 40 KM AT 6.93 GHZ, WITH THE COARSER BANDS BEING THE MOST SENSITIVE TO SST. BECAUSE THE RETRIEVAL ALGORITHM USES ALL SPECTRAL BANDS SIMULTANEOUSLY, THE REMOTE SENSING SYSTEMS (RSS) ALGORITHM, WHICH WILL BE USED IN THIS PROJECT, BEGINS BY RESAMPLING THE MEASURED BRIGHTNESS TEMPERATURES FOR EACH SPECTRAL BAND TO A NOMINAL 56KM FOOTPRINT SIMILAR IN AREA TO THAT OF THE COARSEST (6.93 GHZ) BAND. THE RESAMPLED BRIGHTNESS TEMPERATURES ARE THEN USED TO DETERMINE THE PARAMETER VALUES. THIS RESULTS IN AN AMSR-E/AMSR2 TIME SERIES THAT IS VERY VALUABLE AS A MEASURE OF DAILY GLOBAL SST AS WELL AS TO DELINEATE LARGE SCALE OCEAN PHENOMENA. HOWEVER, THE DATA AS PRESENTLY PROCESSED DO NOT PROVIDE INFORMATION AT SCALES SMALLER THAN O(50 KM) DESPITE THE FACT THAT EACH SPECTRAL BAND IS SAMPLED ON A 10 10 KM GRID; I.E., THE 56KM SST FOOTPRINTS ARE SIGNIFICANTLY OVERSAMPLED. A BAYESIAN ALGORITHM WILL BE DEVELOPED TO DECONVOLVE THE OVERSAMPLED DATA TO AN APPROXIMATE 10KM PRODUCT. MODIS SST AVERAGED UP TO THE 10 KM REGIONS OF OVERLAP WILL BE USED WHERE CLEAR TO CONSTRAIN THE ALGORITHM. SST FIELDS FROM OTHER SATELLITES WILL BE USED IN A STATISTICAL SENSE TO CONSTRAIN THE BAYESIAN SOLUTION AS WELL. THIS ALGORITHM WILL THEN BE APPLIED TO ONE YEAR OF AMSR-E DATA AND THE SPATIAL CHARACTERISTICS OF THE RESULTING TIME SERIES WILL BE ANALYZED BOTH SPECTRALLY AND VIA SST GRADIENT FIELDS. NOTE THAT RSS ALSO PRODUCES AN APPROXIMATELY 25KM AMSR-E SST PRODUCT, WHICH IS ALSO HEAVILY USED, BUT THESE FIELDS ARE NOT OBTAINED BY DECONVOLVING AN OVERSAMPLED DATASET. RATHER THEY ARE OBTAINED BY INTERPOLATING BETWEEN SAMPLES TAKEN AT A COARSER RESOLUTION; THE FOOTPRINT OF PIXELS ON THE 25KM GRID ARE APPROXIMATELY 56KM.