The Uncertainty of Background Brightness Temperatures in Microwave Land Rainfall Algorithm

The rain/no-rain classification of the land algorithm for pixels of the Tropical Measuring Mission Microwave Imager (TMI) was evaluated using the dense network of ground rain gauges in place across the Korean peninsula. It was found that raining events almost completely overlapped with non-raining events in the two-dimensional space of the 21- and 85-GHz brightness temperature (TB) axes. The Empirical Orthogonal Function (EOF) analysis of the TMI TBs indicates that the reasonable separation of raining/non-raining events is not possible from the TMI frequencies alone.

Also, the validity of using the global constant background TB (270 K at 85-GHz) in the land algorithm was examined in order to address the importance of background TBs under clear-sky conditions. It is suggested that more realistic temporal and spatial background TBs be used as a reference to reduce large uncertainties implied by large monthly and diurnal fluctuations in observed background TBs, and thus obtain more accurate scattering amounts from ice particles in clouds over land.