Utilization of Specific Attenuation for Rainfall Estimation in Southern China

This study uses rain gauge observations to assess the performance of different radar estimators R(ZH), R(KDP) and R(A) in estimating precipitation based on the observations of an S-band polarimetric radar over southern China during a typical convective storm and an extremely severe typhoon, i. e., Typhoon Manghkut. These radar estimators were derived from observations of a local autonomous particle size and velocity (Parsivel) unit (APU) disdrometer. A key parameter, alpha (α), which is the ratio of specific attenuation A to specific differential phase KDP with three fixed values (α=0.015 dB deg-1, α=0.0185 dB deg-1 and α=0.03 dB deg-1) was examined to test the sensitivity of the R(A) rain retrievals. The results show that: (1) All radar estimators can capture the spatio-temporal patterns of two precipitation events, R(A) with α=0.0185 dB deg-1 is well correlated with gauge measurement via higher Pearson's correlation coefficient (CC) of 0.87, lower relative bias (RB) of 16%, and lower root mean square error (RMSE) of 17.09 mm in the convective storm while it underestimates the typhoon event with RB of 35%; (2) R(A) with α=0.03 dB deg-1 shows the best statistical scores with the highest CC (0.92), lowest RB (7%) and RMSE (25.74mm) corresponding to Typhoon Manghkut; (3) R(A) estimates are more efficient in mitigating the impact of partial beam blockage. The results indicate that α is remarkably influenced by the variation of drop size distribution. Thus, more work is needed to establish an automated and optimized α for the R(A) relation during different rainfall events over different regions.