In this study, the sensitivity test was carried out with the satellite brightness temperature (TBs) with respect to COMs cloud data and radar rainfall estimates. Then the comparison of standard resolution (14kmⅹ14km) and high resolution (7kmⅹ7km) of TBs (21V, 85V) was performed by computing each probability density function of various seasons under no-rain conditions. Finally we proposes Kalman filter methods which are applied to the TBs in order to reduce the effects (or noises) of geographical and geophysical variability of satellite brightness temperature data according to the results of the first two tests. The newly proposed method is applied to the rain/no-rain classification (RNC), and is compared with the previous RNC methods by ROC (Receiver Operating Characteristic) analysis. In the results, spring and fall seasons show relatively low accuracy for rain/no-rain classification, and so it is needed to adjust the coefficients of scattering index (SI) over land around the Korean peninsula. The brightness temperature data with the Kalman filter method provides the increased accuracy for the classification.

Sung-Dae Yang is a Senor Researcher of Global Environment System Research Division, Korea National Institute of Meteorological Research(NIMS). He obtained his Ph.D at Iowa State University in 2004, then he went to Florida State University for postdoc fellowship. From 2007, he has been working at NIMS. His research Interests mainly include: Numerical Analysis, Scientific Computing, Control Problem and Optimization, etc.