RIVER RUNOFF ANALYSIS USING RADAR-AMeDAS COMPOSITE PRECIPITATION IN A MIDDLE-SCALE RIVER BASIN AND ITS VERIFICATION BASED ON THE CLASSIFICATION OF RAINFALL CHARACTERISTICS

Abstract

  Runoff analysis using radar-Automated Meteorological Data Acquisition System (AMeDAS) composite precipitation (RAP) data for an distributed hydrological model were conducted in a middle-scale river basin Ayagawa (137.5 km²), Kagawa, Japan. The RAP data were observed spatially, 2.5 km×2.5 km resolution. The result shows that the accuracy of river water discarge was improved by 30-60% compared with the ground-based precipitation data, those were observed at around 17 km interval. Based on a classification of the rainfall characteristics, we clarify that not only a decrease of total amount of precipitation but a weak and specially dispersed precipitation (1∼4 mm/h) in the long events improved the accuracy.