近年來，因颱風造成的災害越來越嚴重。颱風挾帶的豪大雨和衍生的土石流，常導致人員和財產的重大損失。因此颱風定量降水預報為目前防災最重要的課題之一。 Kidder等(2005)利用SSM/I、TMI及AMSU等被動式微波資料估算熱帶氣旋降雨，參考官方預報颱風路徑資料，利用平移方法預測24小時後豪大雨的潛勢預報(Topical Rainfall Potential，TRaP)，此技術提供一種快速又實用的方法估算定量降雨，應用此方法預測降雨比數值模式的預測更接近於實際觀測(Ferraro等，2005)，但是TRaP仍然有許多待改進的空間，例如:TRaP方法中，假設颱風降雨的空間分布不會改變，忽略了颱風會旋轉且強度會改變的事實。 TRaP是美國官方目前使用預報颶風24小時累積降雨量的方法。本篇研究希望改進TRaP方法的假設，考慮颱風旋轉及強度變化的影響。本研究使用SSM/I衛星觀測資料反演降雨率，並用同步衛星紅外線資料估算颱風的旋轉角度及強度變化，預測颱風未來6小時內的累積降雨量，最後和島嶼測站的降雨資料相比較。研究結果發現考慮颱風的旋轉及強度變化對於改進降雨的估算能提升準確度，只考慮颱風的平移時，其累積降雨和測站真實觀測的相關係數為0.79，若同時考慮平移、旋轉與強度變化相關係數可增加為0.96。 In recent year, the damages caused by typhoons are more and more serious.The heavy rainfall from typhoons often causes inland flooding and mudslides that threaten lives and property which affect the livelihood of people. One of the most important areas of typhoon research focuses on the quantitative prediction of a typhoon’s precipitation. The forecasting from American agencies use the TRaP method to predict a hurricane’s accumulated rainfall for the next 24 hours. This technique is considered fast and practical. The main purpose of this study is to also apply the TRaP method, but take into account the influence from a typhoon’s rotation and intensity variation. This research uses SSM/I data, GOES-9 data and MTSAT data to estimate the instant rainfall brought forth by typhoons between 2003~2005 around Taiwan. The accumulated rainfall is projected every six hours, and compared with the actual rainfall data recorded from the island’s weather stations. The study’s results indicate that by also considering the rotation and intensity of a typhoon, it will further improve the accuracy of the rainfall estimates. The correlation coefficient of accumulated rainfall between station and the method only concern about advection is 0.79.If we consider the advection, rotation and intensity variation of typhoon, the correlation coefficient can increase to 0.96.
