本研究已成功建立可提供WRF模擬各式各樣理想化颱風情境之初始場以研究氣旋與各種尺度地形之交互作用所造成氣旋路徑偏轉之影響，並利用多層巢狀網格做高解析度模擬來檢驗不同路徑氣旋之降雨分佈。本研究設計一系列實驗以探討氣旋遭遇地形在上游之路徑偏轉情況，主要分為氣旋參數控制組與地形參數控制組兩組實驗。氣旋參數控制組乃固定一地形，其大小與CRM相當，其山底寬L_x=100km ，山底南北長L_y=300km，改變氣旋之最大暴風風速大小以及最大暴風風速半徑大小，並依駛流場之流速分為U1=4m/s與U2=8m/s兩組。結果顯示駛流風速愈大，則地形造成氣旋上游旋路徑之偏折程度愈小，於地形上之降雨量愈少，但兩組降雨分布相當類似。在地形參數控制組實驗，固定氣旋參數，V_max=20m/s、R_mw=80km，而改變山高與地形南北長。從結果得山脈愈高則造成上游氣旋偏折程度愈大，山脈南北長度愈長則氣旋愈有傾向左偏(南偏)之趨勢。若山高為1500m則造成氣旋左偏之山脈南北長最少需3000㎞，即W1L3H1個案，此對應之無因次參數R_mw/L_y之臨界值為0.027。因此，山脈長為造成氣旋左偏向之必要條件，而山高主要影響其偏折量之大小。分析氣旋96小時之移速，發現地形對於氣旋具有減速作用，且駛流風速愈快則減速作用愈明顯。只考慮氣旋上游移速，地形僅對U1組為加速作用，其餘皆為減速作用，且地形參數控制組氣旋移速之減速量值約占駛流風速之20%。藉由渦度收支之分析，氣旋於上游過山前，主要由渦度平流主導，過山時主要為渦度拉伸作用，此與前人研究相同，而傾斜作用於此時亦有增大。本研究發現，氣旋之不對稱風場對其偏折方向有極大之關係，當氣旋之東側風速較大時，氣旋會傾向右偏折，若氣旋之西側風速較大時，則氣旋會傾向左偏折。此機制亦可解釋地形參數控制組，氣旋於上游出現之反氣旋式路徑，此路徑為Yeh and Elsberry (1993)所發現，但未能解釋之。最後從中挑選W1L1H2個案做山脈斜率對氣旋路徑之延伸探討，可惜結果不甚顯著，表示於此實驗中山脈斜率對於氣旋在上游之偏折並無太大之影響，在下游則受背風渦旋影響，氣旋左偏量值將隨山寬增大而增加。 First of all, this research has successfully established a tool to generate those wanted scenarios of balanced I.C.s for idealized WRF simulations and use WRF high resolution nested domains to figure out the rainfall patterns on the terrain. There are two experiments, one is to control vortex parameters, the other is to control terrain parameters. In the controlling vortex parameters experiment, we use a fixed terrain comparable to CMR with width 100km and length 300km. We find that the speed of steering flow cause the negative effect on the tropical cyclones track deflection and accumulated rainfall amount and the results are the same as previous studies. There is an interesting phenomenon that the patterns of rainfall are in phase between U1 group and U2 group neglecting the amount of rainfall. The results of the controlling terrain parameters experiments show that there is obvious evidence for that the direction of the tropical cyclones track deflection is dominated by the terrain length L_y and the degree of track deflection is dominated by the terrain altitude H. The tropical cyclones will turn left for enough large value of Ly. For example, case W1L3H. (about 3000km with H=1500m, critical value of non-dimensional parameter R_mw/L_y=0.027). Overall, the terrains make the cyclones speed down for 96-hr average translation speed both of two experiments and the faster the steering flow is, the more deceleration the cyclones will be. While only considering the upstream average translation speed of cyclones, U1 group is an exception. The terrain makes the U1 group most cases speed up. The upstream de-acceleration value is about 20% of the steering flow speed. With the vorticity budget analysis, the advection term dominate the voriticity tendency upstream, and the stretching term dominate when the cyclones passing through the mountain with strong blocking. Tilting term also has much contribution to the vorticity tendency while the cyclones across the terrain. The results of vorticity budget analysis are the same as Lin et al (2011). According to the results of terrain steepness experiments, the upstream track deflection is not sensitive to the steepness of the terrain (parallel to the moving direction of cyclones) of the mountain. At last, the direction of cyclones track deflection has a lot of to do with the horizontal asymmetric wind field. The cyclone tends to turn right with the maximum wind appearing at the east side of the cyclone, and tend to turn left with the maximum wind appearing at the west side of the cyclone. This can also be applied to explain the phenomenon of the upstream anti-cyclone track found by Yeh and Elsberry (1993).
