The Infulence of Different Interpolation Method to Objective Analysis on 3-Dimensional Wind Field in Taiwan

本研究以光化學軌跡模式(Trajectory Photochemical Air Quality Model,TPAQM)中氣象模組在風場模擬上為架構，對產生第一步風場中客觀分析使用的Barnes逐步訂正法作測試與改良，希望經過後續地形繞流及變分分析產生之風場，能夠忠實反映實際的狀況。 在地面客觀分析上可分三部分做測試，第一部份以不同的初始場，第二部分以不同的掃描次數，第三部分以剔除殘差較大測站的方式，分別進行客觀分析並探討測站殘差的收歛情形，接著經過後續分析產生最終風場後，對風場圖及性能參數做一比較。在探空客觀分析上則以不同的掃描半徑作客觀分析，同樣經過後續分析產生最終風場後，對風場圖及性能參數做一比較。 結果顯示地面客觀分析中，第一部分在使用不同初始場進行客觀分析的情況下，在風場及性能的表現上無明顯不同。第二部分以不同的掃描次數進行客觀分析後，發現若掃描次數太多在極偏遠地區會發生模擬值發散的現象，而在測站密集的區域掃描次數增加對風場的性能影響不大，對測站稀疏的區域在性能上則有改善的跡象。第三部分剔除殘差較大測站進行客觀分析後，測站密集的區域在性能上有提升的效果，對測站稀疏區域則各有好壞。在探空客觀分析上，較大的掃描半徑可將較遠的NCEP資料影響變大，但風場有單一化的現象，對於海拔高度500 m以上的測站，在性能上影響不大。The focus of this research is to rely on trajectory photochemical air quality model (TPAQM) as the base modeling analysis to obtain an unbiased prediction on wind distribution from gradual correction according to Barnes modeling. Different genealogical conditions were experimented and modeled. Three directions of analyses were performed: 1.) investigate different initial conditions. 2.) different scan numbers. 3.) error analyses to obtain unbiased investigation and prediction. The results were compared and fitted to adjust relevant parameters to model potential air flow distribution analyses. Different model scan radiuses were analyzed for unbiased prediction. The results indicated that under different initial genealogical conditions, wind field figures and characteristics were not much different. Secondly, different numbers of scan analysis indicated that too many repetition at suburb area would result in scattered wind distribution. Thirdly, discarding erroneous prediction, better model prediction fit was obtained. In wind filed distribution analysis, larger scanning radius would impact more largely on NCEP data