APPLICATION OF NONLINEAR MULTI-CHANNEL ALGORITHMS FOR ESTIMATING SEA SURFACE TEMPERATURE WITH NOAA 14 AVHRR DATA

NOAA global operational NOAA/AVHRR Nonlinear Sea Surface Temperature (NLSST) retrieval algorithms were used to generate Global Area Coverage (GAC) sea surface temperature (SST) measurements in the global ocean in 1998. The accuracy of SST retrieved from daytime split window NLSST algorithm and nighttime triple window NLSST algorithm for NOAA 14 AVHRR data was investigated in this study. A matchup dataset of drifting buoys and NOAA 14 satellite measurements in the global ocean was generated to validate these operational split window and triple window algorithms. For NOAA 14 in 1998, we had 14095 and 22643 satellite and buoy matchups that matched within 25 km and 4 hours for daytime and nighttime, respectively. The satellite derived SST had a bias of less than 0.1℃ and standard deviation of about 0.5℃. This study also showed that the NLSST algorithm provided the same order of SST accuracy in different time of the year and under a wide range of satellite zenith angle and water vapor represented by the channel 4 and 5 brightness temperature difference. Therefore, NLSST algorithms are usually independent of season, geographic location, or atmospheric moisture content. Comparison between the low resolution AVHRR GAC data accuracy and high resolution Local Area Coverage (LAC) data accuracy is also discussed.

NOAA global operational NOAA/AVHRR Nonlinear Sea Surface Temperature (NLSST) retrieval algorithms were used to generate Global Area Coverage (GAC) sea surface temperature (SST) measurements in the global ocean in 1998. The accuracy of SST retrieved from daytime split window NLSST algorithm and nighttime triple window NLSST algorithm for NOAA-14 AVHRR data was investigated in this study. A matchup dataset of drifting buoys and NOAA-14 satellite measurements in the global ocean was generated to validate these operational split window and triple window algorithms For NOAA-14 in 1998, we had 14095 and 22643 satellite and buoy matchups that matched within 25 km and 4 hours for daytime, respectively. The satellite derived SST had a bias of less than 0.1°C and standard deviation of about 0.5°C. This study also showed that the NLSST algorithm provided the same order of SST accuracy in different time of the year and under a wide range of satellite zenith angle and water vapor represented by the channel 4 and 5 brightness temperature difference. Therefore NLSST algorithms are usually independent of season, geographic location, or atmospheric moisture content. Comparison between the low resolution AVHRR GAC data accuracy and high resolution Local Area Coverage (LAC) data accuracy is also discussed. Key words sea surface temperature (SST) - NOAA/AVHRR - nonlinear SST algorithm Project 49576281 supported by NSFC, and also supported by NOAA Coastwatch Program.