摘要
本文使用Penn State/NCAR的第五代中尺度靜力／非靜力的數值模式（MM5），對侵台颱風路徑變化進行36小時的數值模擬。實驗個案選取發生於1990年9月6日1200 UTC的中度颱風黛特（Dot），其中心路徑於侵台的前後期間近乎直行。使用不同積雲參數化的測試結果顯示，颱風中心路徑會受到若干影響，但並不明顯。靜力與非靜力的測試比較，則發現非靜力於10公里水平解析度模擬並無顯著效應。模擬的颱風路徑與觀測值比較相當合理，顯示中尺度模式MM5有相當潛力可以用於研究颱風的路徑。但另一方面，颱風受到地形影響所生成的副中心強度仍未及觀測值，推測原因可能是由於目前MM5模式對颱風環流並無中尺度特別加強的初始化，因此無法克服颱風環流強度不足的問題。由MM5的實際個案模擬，清楚展現侵台颱風副中心生成的過程。原低層中心在山前受阻，其渦度逐漸被颱風繞流及強烈過山氣流平流至下游處，產生新低壓或渦旋中心。此山後副中心可繼續向上發展，或由過山之高層中心向下沿伸而成一高、低層中心偶合的渦旋。此與過去的理想化個案研究文獻所提出的解釋是相當一致的。
並列摘要
The fifth generation of Penn state/NCAR hydrostatic/nonhydrostatic numerical model was employed to simulate the track evolution of typhoons impinging Taiwan. The Dot typhoon occurring on 1200UTC 6 September 1990 was selected to be the simulation case which exhibits a nearly straight movement during the event. It was found that the typhoon center track can be influenced by different cumulus parameterizations, but not significantly. Comparisons between hydrostatic and nonhydrostatic test runs in horizontal resolution of 10 km tend to indicate that nonhydrostatic effects do not have great influence on the track. The model for typhoon track is in agreement with the observations, indicative of the capability of the MM5 model for typhoon track simulation. On the other hand, the secondary-low center caused by terrain effects is weaker than the observed, possibly due to the simple model initialization scheme for typhoon circulations to be reasonably imposed. The real case simulation by MM5 clearly exhibits the formation processes of the secondary-low center for the impinging typhoon. The vorticity of the original low-level vortex center blocked upstream of the mountain is transported to the downstream side by the splitting flow and intense downslope wind, resulting in the formation of the new low or vortex center. The secondary-low center can develop upward or reform with the upper-level typhoon center as a new center as the latter passes over the mountain, in consistence with the explanations of several idealized case simulations in the past.