Validation of NUBF correction method for DPR

Abstract:

Dual-frequency Precipitation Radar (DPR) is currently operational, and we are aiming to realize a high observation accuracy than the TRMM / PR. DPR algorithm has left a lot of research. Non-uniform Beam Filiing (NUBF) among them seems to be an important issue in order to improve the observation accuracy. According to Iguchi et.al (2009), the NUBF correction which has been introduced into TRMM / PR algorithms, Path Integrated Attenuation (PIA) is then assumed to be gamma distribution, and calculates the correction parameters. In this study, the validity of the NUBF correction using the gamma distribution was verified using ground based radar data. In addition, it was also verification of NUBF correction method using the footprint of KaPR that have been added in the DPR algorithm.

Verification data used the high resolution ground based radar data than the TRMM / PR and DPR. This is to obtain detailed precipitation echo data in the observation footprint. Using this data, we calculated the coefficient of variation s_n of PIA in the observation footprint. The value of s_n calculated by the observation footprint was true value in this study. Estimation method of s_n is two-fold. One method of calculating the s_n is considered as a fractal shapes the 3 × 3 pixels including the observation footprint contained in TRMM / PR Algorithm, the other is the footprint of KaPR closer to the observation footprint (2 × 2 pixels). Result of the verification, the NUBF correction method TRMM / PR Algorithm (3 × 3 pixels), it was found to be overestimated as compared to s_n of the true value. Moreover, the s_n estimation using the footprint (2 × 2 pixels) of KaPR, the difference between the true values becomes smaller. Furthermore, as compared with the case of using the correction method TRMM / PR algorithm, s_n estimate was found to be better.

The results, in NUBF correction techniques on DPR, we propose the use of footprint KaPR (2 × 2 pixels).