Improvement of rain rate estimation using the GMS-5 IR data and verification

Hourly rain rates are estimated operationally in Korea Meteorological Administration(KMA) since last year by the regression method (A_REG), which uses the GMS-5 IR and Automatic Weather Systems(AWS) gauges data over Korean Peninsula. As the results, A_REG method estimates rain rate better than other retrieval methods such as probability matching method and look up table method do. However, this equation can't represent on the whole area including oceans because a number of rain gauges are limited over Korean Peninsula. In this study the A_REG method has been improved by using the radar rain rate over Japan instead of AWS over Korea Peninsula. As the statistical verification, the estimated rain rates by R_REG method are improved than existing method (A_REG). The mean errors of R_REG nearly don't appear though A_REG show positive bias. Root mean square errors of R_REG are 2-3 times smaller than A_REG. The correlation coefficients of two methods are similar.