颱風於接近陸地時，由於與地形及環境流場的交互作用，模式初始化時必須充分考慮颱風不對稱量。本計畫擬利用三維及四維變分方法同化目前可資利用之遙測資料（包括衛星近地風場，GPS ground-based及掩星資料等），選定近年來重要歷史颱風（如莫拉克、凡那比及梅姬等）及未來颱風，探討資料同化對近台颱風模擬（強度、路徑及局地降水）之影響，同時使用位渦反演將颱風準平衡流場及輻散風場（含垂直速度）求出，提供完整的三維環流結構（風場，壓力場及溫度場）進行虛擬資料同化(BDA)之比較研究。本計畫為期三年，除探討多種資料同化對侵台颱風模擬之影響，主要亦將同化雷達回波及徑向風觀測，以了解PV-retrieved vortex及BDA vortex結合多時間、高密度之局部真實觀測對近台颱風模擬之影響。第一年（100年度）著重於使用WRF及WRF 3DVAR進行個案同化模擬及敏感度實驗，同時進行位渦反演分析。為探討海溫變化對颱風演變之影響，在WRF模擬中亦將加入一維海洋混合層作用。第二年除延續第一年工作，亦將進行WRF cycling 3DVAR及MM5 4DVAR來同化雷達觀測資料，探討颱風渦旋spin-up及非對稱化等問題。最後一年，除評估上述颱風渦旋初始化方法及資料同化之綜合成效，由最佳同化模擬中，進一步分析颱風侵台前後動力過程、地形作用及造成超大豪雨的可能機制。 Model initialization should consider the effect of asymmetric structures associated with a typhoon near landfall due to the interaction with the topography and environmental flow. This project will use both 3DVAR and 4DVAR to assimilate available observation data including satellite near-surface wind speed, GPS ground-based and radio occultation observations to investigate the impact of these data assimilation on the intensity, track and local intense rainfall of typhoons nearing Taiwan, with the focus on recent typhoons (e.g., Morakot, Fanapi, and Megi) and future typhoons. Along this line, the PV-inversion retrieved vortex will be used to provide a three-dimensional complete circulation (wind, pressure and temperature) from the quasi-balanced flow and divergence field (including vertical velocity) for typhoon initialization in comparison to bogussing data assimilation (BDA). The project lasts for three years and has an inclination to the combined effects of radar radial velocity and reflectivity as assimilated with the BDA vortex or PV-retrieved vortex. The primary interest will be in the impact of the multiple time levels and high spatial density radar data on the simulation of the typhoon nearing landfall. The first year will be focused on assimilation and simulation using WRF and WRF 3DVAR, in addition to PV-inversion works for selected typhoons, with sensitivities to the changes in sea surface temperature on the typhoon evolution through a 1-D ocean mixed-layer model. The second year continues the works of the first year and will attempt to assimilate the radar data using WRF cycling 3DVAR and MM5 4DVAR to investigate the influential role of the vortex spin-up and asymmetry. The last year would try to assess the performance of a combination of vortex initialization and data assimilation. Based on the best simulation among the model experiments, the dynamical processes, the topographical effects and possible mechanisms for super-intense rainfall associated with selected typhoons will be analyzed in detail. 研究期間：10008 ~ 10107
