颱風侵襲期間臺灣南北都會區風場特徵分析

Abstract

為了解颱風侵臺期間、南北都會區之風場特性，本研究利用1961~2018年基隆、臺北、臺南及高雄等測站之平均風場資料，進行臺灣南北兩大都會區之颱風期間風場氣候特徵分析。結果顯示，颱風侵臺期間臺北測站主要為偏東風，基隆測站風向較為凌亂；臺南及高雄測站之主要風向則為南北向，而高雄測站之風向又較臺南測站稍微凌亂。受颱風中心位置、測站地理位置和地形等複雜因素之影響，各測站風向及風速之氣候特徵明顯不同。 考量颱風強度和暴風半徑對特定測站風場之影響，本研究利用颱風強度和暴風半徑及測站之平均風場資料，以「回歸分析法」建立個別測站之風力氣候統計分析模式（颱風中心位於每個1度x 1度網格時、均有其特定之回歸方程）。針對梅姬颱風（2016）之分析結果顯示，風力氣候統計分析模式，可大致掌握颱風侵臺期間測站風力之變化；此外，風力氣候統計分析模式明顯優於氣候平均法。針對西行侵臺颱風，評估此分析模式於南北都會區適用性之結果顯示，不同類型路徑颱風的風場結構分布之不同、不同綜觀環境造成颱風結構分布之差異、颱風中尺度對流系統產生之強風區、及颱風於打轉後風場受地形之破壞等，皆可能造成風力氣候統計分析模式產生高估或低估之情形；因此，應用此模式評估颱風期間之風場時，須特別注意上述個別情況之影響。

The observed wind fields (10-minute average) during the typhoon-warning period in 1961-2018 issued by the Central Weather Bureau are analyzed to study the influences of typhoon circulation on the wind fields of two major metropolitan areas in Taiwan. Considerable efforts were made to establish a wind statistical climatology dataset, corresponding to the observed typhoon positions (and typhoon intensity and size), for these two metropolitan areas. Results show that the wind field is strongly affected by typhoon position and Taiwan's topography during the typhoon period. The characteristics of wind climatology are significantly different for different areas. For example, easterly winds occur more often in the north metropolitan area, and northerly or southerly winds, in the south metropolitan area. A statistical-climatology model for the wind field at the meteorological stations in these two metropolitan areas is established using the regression analysis method. The parameters considered in the model include the position, intensity, and size of typhoons. Results show that this model performs better than the climatology model in estimating the wind fields during the period when Typhoon Megi (2016) was affecting Taiwan. Further analyses are done to understand the applicability of the statistical-climatology model in estimating the wind field at the north and south metropolitan areas for the westward-moving typhoons that affected Taiwan. Results show that the differences in the structures of typhoons that travel in different directions or are involved in different environments, the strong wind associated with the mesoscale convective systems of typhoon, and the weakening of the wind field after the looping track of a typhoon may lead the model to overestimate or underestimate the station wind field. Therefore, precaution must be made regarding the possible effects mentioned above when applying this model to estimate the wind field during the typhoon periods.