雙偏極化雷達參數 (水平回波)、 (差異反射率)、 (差異相位差)、 (比差異相位差)及 (相關係數)可以經由隸屬函數(Membership function)來求得空中的水象粒子資訊(PID)。因台灣缺乏飛機觀測資料，故吾人使用May and Keenan(2003)實驗求得之各類水象粒子的偏極化參數範圍來改善隸屬函數，並以此與WRF模式模擬結果進行比對分析，同時可了解現今WRF模式在雲粒子分類的情況。 研究個案選取2005年05月12日的梅雨鋒面豪大雨事件，分海上對流性降水、陸上對流性降水及層狀降水三種情況討論。模擬結果顯示，使用具有淺對流調節功能的BMJ積雲參數化法表現最佳，模擬的亮帶高度與真實亮帶高度相符，約在4~5 km處。 另外，模擬的水象粒子與雷達觀測的位置分佈與型態也相符。 計算水象粒子混合比方面，Lin et al.、WSM6、Thompson et al.三種雲微物理參數法與雷達估計比較大多低估。 綜合表現屬WSM6 scheme與用雷達資料計算出的混合比最接近。 總體而言，WRF模式在模擬降水回波及雲微物理的型態方面表現不錯，但其量值及位置與觀測的時空分佈有差量。且因模式的水象粒子分類較雷達少，比較時必須將偏極化雷達分類出的數個類別合併後，再與模式比較。 模式與經過調整隸屬函數後所得的水象粒子比對結果良好。將來加入飛機觀測資料，更可改善隸屬函數，得到更準確的水象粒子資訊，進一步透過模式做降水機制的研究。 The NCU C-band dual-polarization (C-Pol) radar parameters include the reflectivity (Zhh), the differential reflectivity (Zdr), the specific differential phase (Kdp), and the zero lag cross-correlation of horizontal and vertical wave (Rhv) can classify hydrometeor particles by using fuzzy logic method. Since the lack of airborne observations in Taiwan, we use the Ranges of polarimetric variables and temperature for various hydrometeor species delivered by P.T May and T.D. Keenan (2003) to correct the membership functions which were applied to S-Pol radar originally. Furthermore, we compare the microphysical classification of particles with WRF model and we can learn more about the difference between the different microphysical schemes in WRF model. The three research cases are Mei-yu frontal rainfall events on May 12, 2005. One is the convective precipitation at sea, another is the convective precipitation at land, and the third is the stratiform precipitation around Taipei city. The model result shows encouraging features from the Betts-Miller-Janjic cumulus parameterization run. The height of bright band from observation and simulation are very close, at 4 to 5 km. The pattern and location of hydrometeor particles are also quite similar. Three microphysical schemes outputs were also analyzed in every thirty minutes. The order magnitudes of the mixing ratio of all the hydrometeor species by WSM6 scheme are compatible with radar retrieved quantities.
