ABSTRACT
GSMaP (Global Satellite Mapping Precipitation) satellite rainfall estimates are evaluated at the hourly time scales and a spatial resolution 0.1° latitude × 0.1°longitude. e reference data came from AMEDAS (Automated Meteorological Data Acquisition System) station network of about 27 rain gauges over Kumamoto Prefecture, Japan. is region has very complex terrain and humidity which is characterized by typhoon. Hence, this area are often hit by ˆash ˆood. e research has been conducted to evaluate hourly GSMaP (i.e., GSMaP_MVK (Moving Vector with Kalman Filter) and GSMaP_NRT (Near Real Time) data with AMEDAS data during ˆood events from 2003 to 2012 and to de“ne the rainfall pattern which causes ˆood. Statistical analysis was used to evaluate the GSMaP data, both qualitative and quantitative. e results indicate that GSMaP_MVK was reasonably good at detecting precipitation events and GSMaP_NRT was inadequate to represent the rainfall AMEDAS data. Long term and short term rainfall patterns were observed over Kumamoto Prefecture before the occurrence of the ˆood.
