梅姬颱風（2010）之海氣交互作用－觀測與海氣耦合模式整合分析

Abstract

梅姬颱風（2010）行經菲律賓海及南中國海時，於兩個海域出現不同的海洋特徵，且在颱風強度發展上具有顯著差異。為了瞭解海洋回饋對於颱風強度的影響，本研究利用含有完整物理過程的WRF（Weather Research and Forecasting）大氣模式及3D-PWP（Price-Weller-Pinkel）海洋模式之海氣耦合模式，模擬具有兩種顯著海洋特徵差異的梅姬颱風。梅姬颱風為ITOP（Impact of Typhoons on the Ocean in the Pacific）實驗的關鍵個案，具有大量飛機及船舶進行聯合觀測所得的大氣海洋觀測資料。為了驗證模式模擬結果，我們利用ITOP觀測資料、衛星觀測資料進行比對，亦透過EnKF（Ensemble Kalman Filter）方法同化ITOP期間觀測資料，驅動美國海軍提供之EASNFS（Eastern Asian Seas ocean Nowcast/Forecast System）系統得到海洋分析場，進而針對模擬結果及海洋分析場進行整合與比較分析。另外透過不同海洋動力過程之敏感性實驗，包含垂直混合、水平平流、垂直平流及壓力梯度力作用，我們亦深入了解颱風引發cold wake之形成機制。 控制組實驗結果顯示，不論有無考慮海洋冷卻效應，皆能充分掌握梅姬颱風之路徑，且颱風強度變化趨勢於菲律賓海也與觀測結果一致；然而，於南中國海期間出現顯著海表面溫度冷卻現象，有無考慮海洋冷卻效應模擬之颱風強度變化呈現明顯差異，其中以考慮三維海洋模式之實驗模擬結果與衛星及ITOP觀測最為接近。穩定邊界層分析結果顯示，受到颱風引發之海表面溫度冷卻效應影響，颱風邊界層呈現高度減少及穩定度增加的現象，並進而造成颱風內流角度於顯著海洋冷卻區域增加。 敏感性實驗結果顯示，各項海洋動力過程所造成海洋結構改變之貢獻程度有所差異。藉由颱風強勁風應力形成紊流，引發垂直混合作用，使下層冷海水捲入表層使海表面溫度冷卻；透過垂直平流作用，深層冷海水被帶至上層，使整體海洋結構向上抬升，造成海洋混合層厚度減少不利於颱風強度發展；水平平流作用及壓力梯度力作用皆為趨緩海洋冷卻之水平梯度，壓力梯度力雖然影響程度較小，但可以對較深之海洋有所影響。此研究以線性觀點討論不同動力過程之間複雜的非線性交互作用可能會造成誤差，因此未來仍需進一步的檢驗。 整體而言，本研究評估梅姬颱風與海洋之交互作用並利用完整資料進行整合與比較分析，包括海氣耦合模式模擬、衛星觀測資料、EASNFS海洋分析場以及飛機與船舶於ITOP期間之觀測資料。為了探討颱風結構受到海洋冷卻效應之影響，進一步檢驗穩定邊界層之形成以及內流角度增加的現象。另外本研究也釐清各項動力過程對於颱風引發海洋結構變化之影響，我們認為，考慮完整海洋動力過程有助於在實際個案中解釋上層海洋熱力結構對於颱風強度的影響。

When Typhoon Megi (2010) passed through the Philippine Sea and the South China Sea, distinct patterns of oceanic response and significant differences in tropical cyclone (TC) intensity were observed in the two basins. In this study, the dynamics of oceanic responses to Megi is investigated by using a full-physics coupled model based on the WRF (Weather Research and Forecasting) model and 3DPWP (Price-Weller-Pinkel) ocean model. The unprecedented atmospheric and oceanic data used in this study were obtained from aircrafts and research vessels during ITOP (Impact of Typhoons on the Ocean in the Pacific, 2010). The model results are compared with the observation from satellites, the in-situ measurement during ITOP, and the ocean analysis field from the NRL Eastern Asian Seas Ocean Nowcast/Forecast System (EASNFS) with the atmospheric forcing from our analysis and with ITOP data assimilated. The sensitivity experiments with different processes of ocean dynamics, including ocean current shear-induced entrainment, horizontal advection, vertical advection, and pressure gradient, are performed to understand the mechanisms of the TC-induced cold wake formation. The control experiments reasonably well capture Megi’s track. The simulated intensities are close to observations in the Philippine Sea while being diverse in the South China Sea, for which the closest result is taken by a three-dimensional coupled model simulation because of the significant SST cooling. The different oceanic responses in these two basins show that it is valuable to include the three-dimensional dynamic processes in the ocean especially for TCs with slow translation speed and under unfavorable upper ocean thermal structure. In the analysis of a stable boundary layer, the depth and stability of the TC boundary layer are relatively limited and enhanced by the TC-induced SST cooling respectively, and thus the inflow angle would be larger over the cold wake region. From the sensitivity experiments, potential mechanisms of oceanic response to different dynamic processes are evaluated. Strong surface wind stress associated with the TC can generate turbulence in the upper ocean, thus reducing SST through the vertical mixing of cooler water from the deeper ocean. On the other hand, the upper ocean temperature profile would be shifted upward and provides an unfavorable condition for TC intensification due to the effect of the vertical advection. In addition, the horizontal ocean temperature gradient under and around the TC center is reduced due to horizontal advection and pressure gradient terms. Nevertheless, the interpretation of the individual impact of these dynamical processes is not explicit because the nonlinear effects remain extant in the experimental approach of this study. In all, this study provides a comprehensive comparison among atmosphere-ocean coupled model simulations, satellite data, EASNFS ocean analysis, and aircraft and vessel observations during ITOP. The changes of the atmospheric and oceanic structures in correspond to the TC-ocean interactions (e.g., the formation of the cold wake and stable boundary layer, and the enhanced inflow angle) are examined with care. Also, this work shed some lights on the impact of each physical process on the TC-induced variations in the ocean thermal structure.