Satellite Sea Surface Temperature Measurements an Introduction

Abstract

One of the goals of the session on sea surface temperature measurements at this conference was to discuss the long standing problems that have prevented satellite infrared observations from achieving agreement with ship and other in-situ results to better than 1° C average r.m.s. temperature error. Some errors are definitely due to atmospheric contamination of the satellite data, and the problems of correcting these for current satellites were discussed by several authors, but several more general points were also raised. Sea surface temperatures as measured by thermometers, bucket or intake, may be “accurate” insofar as the calibration of the thermometer is concerned, but what do such so-called ground truth measurements really represent? The question of comparison data sets received much discussion, with general agreement that one must be very careful not to quote “error” in the context that one measurement represents the “truth” against which another measurement or inferred value is compared. Comparison data sets, especially for SST, needs much more investigation. Part of the residual differences, not errors, between different measurement sets will consist of real differences in physical processes, or in time and space sampling of the different measurements.