Improved Attenuation-Based Radar Precipitation Estimation Considering The Azimuthal Variabilities Of Microphysical Properties
Hao Huang,Kun Zhao,Haonan Chen,Dongming Hu,Peiling Fu,Qing Lin,Zhengwei Yang
DOI: https://doi.org/10.1175/JHM-D-19-0265.1
2020-01-01
Journal of Hydrometeorology
Abstract:The attenuation-based rainfall estimator is less sensitive to the variability of raindrop size distributions (DSDs) than conventional radar rainfall estimators. For the attenuation-based quantitative precipitation estimation (QPE), the key is to accurately estimate the horizontal specific attenuation A(H), which requires a good estimate of the ray-averaged ratio between A(H) and specific differential phase K-DP, also known as the coefficient alpha. In this study, a variational approach is proposed to optimize the coefficient alpha for better estimates of A(H) and rainfall. The performance of the variational approach is illustrated using observations from an S-band operational weather radar with rigorous quality control in south China, by comparing against the alpha optimization approach using a slope of differential reflectivity Z(DR) dependence on horizontal reflectivity factor Z(H). Similar to the Z(DR)-slope approach, the variational approach can obtain the optimized alpha consistent with the DSD properties of precipitation on a sweep-to-sweep basis. The attenuation-based hourly rainfall estimates using the sweep-averaged a values from these two approaches show comparable accuracy when verified against the gauge measurements. One advantage of the variational approach is its feasibility to optimize alpha for each radar ray, which mitigates the impact of the azimuthal DSD variabilities on rainfall estimation. It is found that, based on the optimized alpha for radar rays, the hourly rainfall amounts derived from the variational approach are consistent with gauge measurements, showing lower bias (1.0%), higher correlation coefficient (0.92), and lower root-mean-square error (2.35 mm) than the results based on the sweep-averaged alpha.