策略性颱風觀測--共軛模式之颱風駛流敏感向量

Abstract

生成於廣大洋面上的颱風，常對人類的生命財產造成重大威脅。為改善對於可能侵襲陸地之颱風的預報與分析，多年來美國一直持續針對大西洋颶風進行飛機偵察觀測的任務，在太平洋地區，自2003年起，在國科會推動的「追風計畫(DOTSTAR)」中，利用飛機搭載科學家，在觀測資料十分稀少的西北太平洋上空進行投擲「投落送 (dropsonde)」的任務，以取得接近台灣之颱風周圍珍貴的大氣資料。在飛行資源及時間的限制下，如何擬定「策略性觀測計畫」，為整體觀測任務中十分重要的環節，而「敏感區域位置」正是其中最為重要的參考依據。 除了追風計畫目前採用的三種預報敏感性結果之外，本研究延續Wu et al. (2005b) 中以「觀測系統模擬實驗 (OSSE)」探討颱風渦旋初始化的工作，使用MM5共軛模式系統，發展一套創新的方法來尋找影響颱風駛流的敏感區域，並進一步獲得影響整體颱風路徑的敏感區域位置。 本研究選用2004年9月的米雷颱風與6月的敏督利颱風做為展示研究之實驗個案，藉由代表颱風駛流的「反應函數」，配合新定義之「共軛模式推導之颱風駛流敏感向量 (ADSSV)」，可以清楚地表現出在觀測時刻對於模擬未來36小時颱風駛流的敏感位置，再綜合觀測時刻對於各不同「驗證時間」的計算結果，可進一步求得影響整體36小時颱風移動路徑的敏感區域位置。 在研究中並針對颱風初始化問題以及模式中忽略水氣影響的部分進行測試分析，且為了將此敏感結果之計算方式應用於即時觀測之中，亦改變部分實驗流程，並檢驗其結果，使其在不影響科學計算結果之下，能夠更順利地進行實際作業。 此方法將在2005年應用於追風計畫及美國大西洋颶風投落送偵察觀測的實際運作之中。

Since 2003, a field program has been conducted under the name of Dropwindsonde Observations for Typhoon Surveillance near the Taiwan Region (DOTSTAR) (Wu et al. 2005a). For DOTSTAR, the targeted observation is one of the most critical objectives. The important basis to design the observing strategy is to identify the sensitive areas, which would have the critical impact on the numerical forecast results or the fast growth of the forecast error. In addition to various sensitivity products already adopted in DOTSTAR, a new way to identify the sensitive area for the targeted observation of tropical cyclones based on the MM5 adjoint model has been proposed. By appropriately defining the response functions to represent the steering flow at the verifying time, a simple vector, Adjoint-Derived Sensitivity Steering Vector (ADSSV) has been designed to clearly demonstrate the sensitivity locations at the observing time. Typhoons Meari and Mindulle of 2004 have been selected to demonstrate the use of AFSSV. In general, unique sensitive areas 36 h after the observing time are obtained. The impact of the vortex initialization and the dry assumption on the ADSSV has also been studied. In order to use the ADSSV in the field program (such as DOTSTAR), the procedure of the model calculation has been to meet the realtime need. It is shown that the realtime use of ADSSV is feasible, and the results are fairly consistent with our previous findings. The ADSSV will be implemented and examined in the field project, DOTSTAR in 2005, as well as in the surveillance mission of Atlantic hurricanes conducted by HRD. Further analysis of the results will be very important to validate the use of ADSSV.