The presence of sea ice almost doubles the microwave emissivity compared to that of open water, forming the basis for the detection of sea ice and fractional amount. The retrieval of surface emissivity is important to the derivation of numerous other surface products including surface wetness and surface temperatures.

Estimates of surface emissivity require input of simulated microwave channel brightness temperatures for a large variety of surfaces including oceans, sea ice, snow cover of varying types, and wet land. For each surface type, brightness temperatures are computed for a variety of atmospheres at the lower AMSU frequencies (SSM/I 19-, 37-, and 85-GHz channels) and for all of the different viewing angles. Simulated brightness temperatures are then combined with the observed brightness temperatures to retrieve surface emissivity under rain-free conditions. Rain-free pixels are determined a priori by the use of screening procedures and filters.

The accuracy or RMS error in sea ice amount for both SSM/I and AMSU products is approximately 25 percent. Errors in the estimation of surface emissivity represent the single largest contribution to errors in sea ice amount. The errors result largely from brightness temperature contamination by clouds and sea surface winds, as well as uncertainties in the modeled emissivity values for sea ice of varying age.