颱風模式初始場不確定性對路徑預測影響之初步研究

A Preliminary Study on the Sensitivity of the Typhoon Track Model Forecasts to the Initial Fields

葉天降(Tien-Chiang Yeh) ； 黃康寧(Kaug-Ning Huang) ； 陳得松(Der-Song Chen)

《大氣科學》 27卷3期 (1999/09) Pp. 215-233

摘要

在颱風路徑模式預測裡，由於觀測資料之不足，初始場常需透過人為之處理，因此存在明顯之不確定性。本文主要以中央氣象局颱風路徑預報模式為研究對象，針對1996年Cam（凱姆）、Gloria（葛樂禮）、Dale（戴兒）及Herb（賀伯）四颱風個案，以改變人為加入渦旋之大小、強弱、最大風速所在半徑、中心位置等颱風初始場參數及不同大尺度場、不同模式垂直加熱分布等13種不同處理，透過模擬以分析研究模式預測路徑對初始場改變之敏感性。結果顯示，這些初始場參數之改變都對模式颱風預測路徑有相當明顯之影響，24/48小時預測中心位置差異平均達55/68公里，其中在環境場有比較明顯之水平或垂直風切時，模式之預測路徑對初始場參數之改變更敏感。此外也顯示，若在加入虛擬渦旋前，所用之濾波方法未能有效去除初始場所含不完整之颱風渦旋時，可能使模式颱風路徑預測誤差增大，此時引用類似DeMaria（1987）之方法修正駛流，似能減少預測誤差。

關鍵字

颱風路徑預報 ； 路徑預測誤差對模式初始場之敏感性

並列摘要

In operational environment, it is not easy to analyze the typhoon structure correctly from the current observation networks. Typhoon track forecast models usually require a bogussing procedure in the model initialization for simulate the typhoon circulation. This study examines the sensitivity of the model track forecasts to the model initial fields. Cases selected are typhoons Cam (00UTC 22 May), Gloria (00UTC 24 July), Dale (12UTC 7 November), and Herb (12UTC 29 July) of 1996. The model used is the operational typhoon track forecast model of Central Weather Bureau. Different initial fields are obtained by changing the parameters of the size, the strength, the radius of the maximum wind, and the location of the bogussing vortex. Different large scale environments are also examined. The simulation results show that the track forecast does sensitively depend upon the small change of the initial field. The mean distances between the forecast centers of the control forecasts and the forecast centers of the simulations with various specifications of the bogussing vortex are 55/68 km in 24/48 hours forecast. Among those cases, larger distances between forecast centers are found for the cases of typhoons in the environment with large horizontal or vertical wind shears. The results also show that some of the cases with large model track forecast errors are due to the incomplete removal of the model analyzed typhoon structure before adding the bogussing vortex. For those cases, by changing the vortex specifications in the bogussing procedure could not improve the forecast result effective. Applying the approach of modifying the environmental flow such as that proposed by DeMaria (1987) can, sometimes, produce a better forecast.

並列關鍵字

Typhoon Track Forecast ； Sensitivity of the Track Forecast to the Model Initial Field