納莉颱風時期二維雨滴譜儀之雨滴粒徑分及雷達降水估計

The 2d-video Distrometer Raindrop Size Distribution and Radar Rainfall Estimate in Typhoon Nari (2001) Case

張偉裕(Wei-Yu Chang) ； 陳台琦(Tai-Chi Chen) ； 林沛練(Pay-Liam Lin)

《大氣科學》 31卷3期 (2003/09) Pp. 261-284

摘要

雨滴粒徑分布 (Raindrop Size Distribution)依降雨的情況有不同的特性，而且雨滴粒徑分布可決定雲中液態水含量(W)，回波強度(Z)，降雨率(R)等降雨積分參數，因此分析雨滴粒徑分布的特性十分重要。Marshall＆Palmer(1984)針對層狀降水提出雨滴粒徑分布的型態為指數型態分布：N(D)N｡exp(-ΛD)，但與實際觀測資料相比對皆有描述不佳的情形發生，Uibrich＆Atlas(1984)提出另一個修正的雨滴粒徑分布型態：Gamma Distribution，其假設雨滴粒徑分布合(fitting)到Gamma Distribution 近似分布的三個係數N｡、m、Λ，來描述降雨率增減時雨滴粒徑分布的變化。對2001年納莉(Nari)颱風侵台期間中12小時的資料，依不同降雨率、回波強度，探討N｡、m、Λ的分布，並計算Z=AR(superscript b)關係式中的係數A，b。
研究發現N｡、m、Λ的大小與降雨量成反比，但降雨率越小，標準差越大;b則和降雨量成正比(約1.2到1.4之間)，A的分布則和N｡、m、Λ相似(約200到400之間)。且五分山雷達觀測之回波強度Z(subscript radar)與雨滴譜儀計算之回波強度Ｚ(subscript radar)比較，發現五分山雷達觀測之回波強度Z(subscript radar)似乎系統性的低估約3.3dBZ。故將回波修正3.0Dbz後的回波強度，帶入以回波強度區分之兩修正Z-R關係式，發現其計算的一小時做計降雨量與地面觀測降雨量，有不錯的一致性。

關鍵字

雨滴譜儀 ； 雨滴粒徑分布 ； Z-R關係式 ； 修正回波

並列摘要

The drop size distribution (DSD) has great variation in different types of rainfall condition. The DSD can determine the different moment of rainfall integral parameters such as liquid water content (W),rainfall rate(R), reflectivity factor(Z),etc. Ulbrich(1981) suggest the DSD should be Gamma distribution:N(D)= N｡D(superscript m)exp(-ΛD). Through the calculation of different moments of DSD, the coefficient No, m, Λ of Gamma distribution were calculated. The relations of Z-R were also derived through the coefficient No, m, Λ.
In Nari case, we had analyzed 12hrs date of distromter. The analysis of N0, m, N｡, Λindicated the smaller m, Λ are associated with larger rainfall rate. In the same time, the smaller A was associated with the heavier rainfall rate, and greater b associated with the lighter rainfall rate. In the statistics of these parameters in 12 hours, the coefficient m, Λ, b and A had the same tendency.
We also have applied the distrometer derived Z-R relations to the radarreflectivity. Before this calculation, the correlation between the distrometer derived reflectivity and the radar observed reflectivity was checked. During the heavy rain period, radar reflectivity was about 3.0 dBZ less than the distrometer.
The smoothing in horizontal and the difference of reflectivity between 1.75km and surface may both cause the underestimate. During the weak rainfall period there was very small difference between distrometer and radar. The area rainfall accumulation are quite satisfactory by using the distrometer derived Z-R relation and corrected radar reflectivity.

並列關鍵字

Distrometer ； Raindrop size distribution ； Z-R relations ； Radar reflectivity