在建築環境風場的考量上，行道樹之設置對行道風場有重要之減風效果。一般在氣流通過離散樹木之流場分析，於初步之設計階段多求諸於風洞試驗，其除了需付出相對較高之人力與時間外，技術上的困難(如縮尺之限制)亦是常見之問題。相對地，應用數值模擬則不失為解析此類問題之另一研究工具。因此，本研究擬建立一套能準確預測流體經過離散闊葉樹之流場數值模式，沿用前人提出適用於群樹之控制方程式，進而建議出樹木相應枝葉參數之變化關係，以提供行道風場在初步設計規劃階段風場分析之便捷方法。 本研究之主要探討對象為榕樹，其為闊葉樹。研究工作中除了數值計算外，亦包括了風洞試驗，以作為數值參數率定與數值模擬結果驗證之依據。風洞試驗係採用Cobra Probe量測樹體模型後方截面之平均速度及均方根速度剖面；於流場之模擬係採用微可壓縮流之動力計算方法，並加入源項以反映出枝葉實質之物理效應，再以試誤法將數值計算及風洞試驗結果作比對，尋求最佳之率定值，繼而改變樹體模型之厚度及枝葉體積比，以獲得樹體特性參數之變化關係。 研究結果顯示，整體而言樹體特性參數隨枝葉體積比增加而增加；又隨樹體模型厚度減小而有明顯變化。研究中繼依據自矩形樹體模型情況獲得之樹體特性率定結果，再進行風流經三維獨立樹木、雙株樹木之數值模擬計算，其結果與風洞試驗之下游風速剖面量測結果相比有不錯之吻合度。 Planting trees in an open area can generally slow down the near-ground wind speed thus becomes an important way in pedestrian wind planning. At the preliminary design stage, to analyze the flow past discrete trees, wind tunnel model experiments are mostly employed. Besides the amount of human labor and time required in the execution of experimental work, however, technical difficulties (such as the scale effect) are generally encountered. In contrast, the adoption of numerical simulations can be another alternative for the flow analysis. Accordingly, the study is to establish a numerical model, capable of correctly predicting flows past discrete broadleaf trees that can be handily adopted in the governing equations proposed by previous researchers to explore a relationship to properly describe the physical effect of trees so as to provide a handy tool for pedestrian wind analysis during the preliminary design for local wind environments. The study concentrates on the problem of the banyan trees which subject to broadleaf plants. In addition to numerical computations, wind tunnel measurements were also performed to guide and confirm the numerical simulations. In the wind tunnel experiments, Cobra Probe was adopted to measure the downstream cross-sectional mean and root-mean-square velocity profiles. Besides, the variation of the tree parameter was obtained as a function of the tree body density and width. A weakly-compressible-flow method is used in the numerical simulations by adding a source term in the momentum equations to reflect the physical effect of trees. Furthermore, by using a try-and-error method to compare with the measurement profiles corresponding to a prescribed thickness and volume density of trees, the tree parameter that lead to the best agreement between the measurement and calculated profiles was then obtained. The results showed that basically the tree parameter increased with increases of both the tree body density and varied significantly with the tree body widths. Moreover, based on the calibrated relationship of the tree factor, numerical computations were further performed to simulate flow past an isolated tree and dual trees. The predicted downstream wind speed profiles showed good agreements compared to the measurement results.