Error Characteristics of the Atmospheric Correction Algorithms Used in Retrieval of Sea Surface Temperatures from Infrared Satellite Measurements: Global and Regional Aspects

Abstract

A database of cotemporal, collocated satellite and in situ observations is used to examine the association between the brightness temperature differences measured by the thermal infrared channels (T45) of the Advanced Very High Resolution Radiometer (AVHRR) and water vapor estimates (ω) derived from the Special Sensor Microwave Imager (SSM/I). This channel difference is used to estimate the atmospheric correction (due mostly to water vapor absorption) in sea surface temperature (SST) algorithms. The association between T45 and ω is found to be greatest for tropical latitudes; for mid- and high latitudes, the association is best during summer. However, the association tends to decrease toward mid- and higher latitudes during other periods. SST residual errors (satellite − buoy) show a negative mean in the Tropics, suggesting undercorrection for water vapor attenuation in the Tropics. This underestimation is explicitly shown for SST residuals in the high-water-vapor regimes of the Arabian Sea. In mid- and high latitudes, the variability of atmospheric water vapor and air–sea temperature difference contributes to the weaker association between T45 and ω and results in positive mean SST residual errors. A differential form of SST algorithm that incorporates the use of a “first-guess estimate” that correlates with SST is observed to give the least residual error.