Error Analysis of the Rain Rate from TRMM Precipitation Radar Using the Ground Based Observations over Korea

摘要

Preliminary error analysis results of the near surface rain rate (NSR) measured by Precipitation Radar (PR) onboard Tropical Rainfall Measuring Mission (TRMM) satellite are introduced in this paper. Statistical errors are analyzed by comparison of rain rates measured by TRMM PR and ground-based Instruments, i.e. Automatic Weather System (AWS) over Korean Peninsula. The ground and satellite measurements are fundamentally different time and spatial resolution. AWSs sample the rainfall continuously in time but at a separate point, while TRMM PR samples an area average over the field of view but a snapshot in time. Therefore the comparison of two data with different temporal and spatial resolution include much errors. The purpose of this study is to investigate the errors according to the field of view of averaged grid, the number of AWS stations and the threshold of rain rate. However, in this paper we will introduce only the variations of errors by the number of AWS in a FOV and by the averaging time of rainrate, in advance. Results show that the RMSEs are decreased inversely by the number of AWS rain gauges, but correlation coefficients (CC) are increased as the number of AWS rain gauges are more. In addition, the CC according to the number of AWS tend to increase as the averaging time is longer and shows opposite relation with RMSE. The PR agrees well to the AWSs with the highest CC of 0.79 and the smallest RMSE of 2.285. The RMSE is smallest when 10 rain gauges in a grid averaged for 30 minutes. However the PR Seems to estimate rain rate lower than AWSs.