[A-CG06_29AM1] Satellite Earth Environment Observation

[ACG06-12] 二周波降水レーダ用の減衰補正手法の開発

*瀬戸 心太¹、井口 俊夫² (1.長崎大学　大学院工学研究科、2.情報通信研究機構)

キーワード：二周波降水レーダ, 全球降水観測計画, 減衰補正

A new attenuation correction method is developed for the Dual-frequency Precipitation Radar (DPR) on the core satellite of the Global Precipitation Measurement (GPM) mission. Hitschfeld and Bordan's attenuation correction method (HB method) assumes relation between the specific attenuation k and the effective radar reflectivity factor Z_e (k-Ze relation) as k=aZ_e^. The new method is based on HB method, but k-Ze relation is modified as k=eaZ_e^ by using dual-frequency ratio of Z_e (DFR) and surface reference technique (SRT). Therefore, the new method is called HB-DFR-SRT method (H-D-S method in short). While the authors' previous attenuation correction method called HB-DFR method (H-D method in short) results in underestimation of precipitation rates for heavy precipitation, H-D-S method and its improved version try to correct the negative bias by means of SRT. When only single-frequency measurement is available, H-D-S method can be easily switched to HB-SRT method (H-S method in short), which is similar to the attenuation correction method used in the TRMM/PR standard algorithm. The attenuation correction methods are tested with a simple synthetic dataset of DPR. As long as SRT gives the prefect estimates of path integrated attenuation (PIA) and the parameters of k-Ze relation (a and b) are given properly so that e could be vertically constant, H-S method is much better than the dual-frequency methods. In reality, SRT has error and we cannot give the parameters of k-Ze relation properly so that e should be vertically variable. Tests with SRT error and vertical variation of e show that H-D method is better than H-S method for weak precipitation but H-S method is better than H-D method for heavy precipitation. It is because SRT is unreliable for weak precipitation and DFR is unreliable for heavy precipitation. H-D-S method shows not the best but stable results for both weak and heavy precipitation, and it may work well for medium precipitation. Quantitative evaluation should be done with real measurement dataset of DPR.