POTENTIAL IMPACT OF SATELLITE DATA ON SEA SURFACE TEMPERATURE ANALYSIS 



John C. Wilkerson 

 Project Manager 



Dr. Vincent E. Noble 

 Project Physicist 

 Airborne Remote Sensing Oceanography Project 

 Naval Oceanographic Office 

 Washington, D. C. 20390 



ABSTRACT 



The sea surface temperature field of a 10 x 20 degree area of 

 the North Atlantic is constructed from synoptic temperatures of 

 the ocean surface obtained from High Resolution Infrared Radio- 

 meter (HRIR) data from NIMBUS II meteorological satellite. The 

 computer analysis of these satellite data is compared with the 

 computer analysis of three and one half days of conventional 

 ship data produced for the same time period by the Fleet 

 Numerical Weather Central (FNWC) , Monterey, California, with a 

 manual analysis of the same ship data prepared by the Fleet 

 Weather Central, Norfolk, Virginia, and with an experimental 

 numerical analysis done by NAVOCEANO. Significant features in 

 the surface temperature field, as revealed by the satellite data 

 and confirmed by independent ship and aircraft observations, are 

 absent in the analysis of the conventional ship data. The HRIR 

 radiation temperature data are biased 2.4°C warmer than the 

 "ground truth" Airborne Radiation Thermometer (ART) data, with 

 a residual mean absolute error of 2.0°C. The magnitude of the 

 sea surface temperature gradients associated with the Gulf Stream 

 boundary range from a maximum of 6°C/600 meters from the analog 

 ART data, to 5°C/11 kilometers from the digitized HRIR data, to 

 4°C/110 kilometers from the FNWC ZOOM analysis. The differences 

 between these analyses provide a measure of sea surface tempera- 

 ture analysis improvement potentially available through the 

 increased synoptic data coverage afforded by satellite. 



365 



