Studies on Mean Areal Rain Rate using Dual-Polarization X-Band Radar over a Small River Basin, Japan

Since National Research Institute for Earth Science and Disaster Prevention (NIED) deployed the first dual-polarization X-band radar around the Tokyo metropolitan area in 2000, there have been more than 30 dual-polarization X-band radars now covering most of the major urban areas of Japan. In the XRAIN project carried out by Ministry of Land, Infrastructure, Transport and Tourism (MLIT), a widespread use of dual-polarization X-band radars has gained significant momentum in Japan. In the past ten years, considerable efforts have been made to obtain better Quantitative Precipitation Estimation (QPE) from radar. Consequently, rainfall estimators using polarimetric parameters from X-band radar have been proved to be in the best harmony with rain gauge measurements without any corrections from surface observations. On the other hand, one key application of operational weather radar is to provide the input of mean areal rain rate to hydrological model for flood forecasting. Most of the common used hydrological models still rely on rain gauges and are usually based on the assumption of uniform rainfall over the catchment. However, even in a single storm, rainfall is highly variable in both space and time. Accurate and short duration rainfall information is required to develop robust hydrological models. Although several methods have been proposed (Ryzhkov et al., 2000; Bringi et al., 2011), few have been done for X-band radar. In this research, the estimate of areal rain rate using dual-polarization X-band radar is compared against a high density network of 30 rain gauges within an area of 20 km2 around the Hayabuchi river basin, Japan. Five different rainfall estimators are considered: using an exponential fit of specific differential phase (KDP) – rain rate (R) relationship with KDP retrieved either from the linear regression of differential phase (ΦDP) or variational method assuming the monotone increasing ΦDP below melting layer; two areal ΦDP algorithms, one is proposed by Ryzhkov et al. while the other by Bringi et al.; and a modified composite method of above two areal ΦDP algorithms.