Assessment of Precipitation Characters between Ocean and Coast area during Winter Monsoon in Taiwan

Abstract SSM/I has been applied in derivation of liquid water and precipitation for many years. Due to the uncertainty in variance of surface emissivity and characters of precipitation, the comparison of rainfall retrieval with ground truth remains a controversial issue. In order to investigate this issue, data set for precipitation from stratus clouds over ocean area were collected for 5 years. Each data set is collocated with satellite observation and ground truth. Neural network is applied to estimate precipitation. Comparison between retrieved rain rate and surface ground truth was made. Based on the analysis, the wind speed does not appear to have close correlation with precipitation. Mean value of precipitation is consists with ground truth, but most of the heavy precipitations are not captured by the estimation. The best result is obtained by using all 7 channel observations and polarized corrected temperatures to estimate precipitation.