雷達定量降水估計(Quantitative precipitation estimation, QPE)產品往往容易低估，由於波束加寬、部分波束遮蔽(Partial beam blockage, PBB)影響和垂直高度上的雲物理變化等。在雷達波束採樣到最低高度下方，經由暖雨過程發展，所導致明顯的低層降雨增強的情況，尤為顯著低估降雨的誤差來源。本研究使用降雨垂直剖面(Vertical profile of rain, VPR)修正方法評估S波段雙偏極化的五分山雷達(RCWF)在2017年至2021年間各類降水事件的降雨估計表現，包括梅雨鋒面、颱風及冬季冷鋒鋒面類型。此外，藉由自洽(Self-consistency)方法，改善弱降水時推估比差異相位差(Specific differential phase, K_DP)參數的不確定性。實際個案分析顯示相較使用沿路徑計算K_DP的QPE結果，基於自洽法的降雨估計在中小雨區能減緩K_DP參數高估的情況，至少5%的改進。而VPR方法的應用，依不同的降雨類型和雷達變數而異，在層狀降水事件期間，對PBB效應最嚴重的區域具有顯著的改善，以R(Z,Z_DR )關係式改進最為明顯，使得標準化平均偏差(NMB)降低至少20%。另一方面，颱風期間的強降水事件則對K_DP相關的降雨估計結果皆有負面的影響。;The radar-based quantitative precipitation estimation (QPE) products are well known that tend to underestimate due to the beam broadening effect, partial beam blockage (PBB) effect, vertical microphysical evolution, and sampling discrepancy. The underestimation of the QPE is most pronounced for the precipitation with pronounced low-level enhancement due to warm- rain processes. This study assessed the vertical profile of rain (VPR) correction for the radar-based QPE of S-band polarimetric radar (RCWF) at various types of precipitation events from 2017 to 2021. In addition, the retrieved K_DP from self-consistency method, where K_DP^* (K_DP^(**)) was based on reflectivity(Z) (along with differential reflectivity (Z_DR)) were investigated to reduce the uncertainties of deriving specific differential phase (K_DP) in light rain. The results of the real case study indicate that the K_DP^*(/K_DP^(**))-based QPE shows improvement in normalized mean bias by at least 5 % at low and moderate rain rates compared to the range-derivative K_DP-based QPE. Moreover, the impact of the VPR method varies with different rainfall types and radar variables. The improvements are most pronounced in the regions where the PBB effect is most severe. The VP-R(Z,Z_DR) shows significant improvement during the stratiform precipitation event with the normalized mean bias decreasing by at least 20%. On the other hand, the heavy precipitation event during the typhoon has a negative impact on all K_DP-based rainfall estimations.
