評估緩衝透水面對都市洪水模擬之影響

Abstract

低衝擊開發(Low impact development, LID)常作為降低洪災的手段，過去在利用水理模式評估 LID 的減洪效益時，受制於數值模式的限制僅粗略使用土地利用分類，忽略物理機制與管理策略，導致模擬結果與事實偏差大。本研究提出改良式的 LID模擬方法，同時考慮優化土地利用類型以得到更真實的淹水模擬。用雨水管理模型(Storm Water Management Model, SWMM)評估 LID 對區域之影響，缺乏考量二維漫地流，因此本研究採取 3Di 模式(3Di water management, 3Di)來進行評估，此模式有高精度、時空、高程及漫地流的特性。本研究所提出的改良方法為利用中間流結合表面緩衝透水面模擬與不透水面的交互作用，同時導入優化土地利用進行減洪成效評估。模擬結果與傳統方法(中間流、SWMM)和現地實驗比較，顯示改良方法更貼近實驗值。以綠屋頂為例，改良方法、SWMM、中間流的洪峰流量分別為 4.3 x10-5、6.8 x10-5、6.3 x10-5cms，本研究的方法更貼近現地結果的 3.9 x10-5cms。 以中壢區與中原大學為研究區域進行大尺度驗證並討論優化土地利用的差異，用 5m 解析度 DEM 模擬 24 小時三種重現期降雨，結果顯示 10 年重現期時，考慮改良式 LID 與優化土地利用的模擬，洪峰流量會下降 38.3%，流量歷線也相對延遲，20 年重現期時，洪峰流量則會下降 9.7%。對於 LID 較多的區域，優化土地利用對於 LID 更能有效增加滲透及減少逕流來發揮作用。本研究提出的方法增加 LID 模擬的準確性與靈活性，而分散式設置 LID 顯示能有效處理地表逕流，減少排水系統的壓力。此外優化土地利用能準確的模擬水從不透水面傳遞至透水面，於 LID 儲水爾後溢流量排出雨水管之現況。

Low impact development (LID) is frequently employed as a means of reducing flood ing. In the past, when evaluating the flood reduction benefits of LID using hydrological models, the limitations of the numerical model were not fully considered. This resulted in a significant discrepancy between the simulation results and the actual situation. This study proposes an improved LID simulation method that considers optimisation of land use types to obtain a more realistic flooding simulation. The Storm Water Management Model (SWMM) is used to evaluate the impacts of LID on the area, which lacks consideration of two-dimensional diffuse flow. Therefore, this study employs the 3Di water management (3Di) model, which has the characteristics of high accuracy, spatial and temporal, elevation and diffuse flow. The LID in this study comprises green roofs, rain gardens, permeable paving and rain barrels. These elements are employed to simulate the interaction between permeable and impermeable layers by optimising the land use and introducing a buffer permeable layer to simulate the flood mitigation effectiveness. The modified approach in volved the use of intermediate flow in conjunction with the surface buffer permeable layer function for the simulation. The results demonstrated that the modified approach yielded more accurate results than the use of intermediate flow alone, SWMM, and in situ experi ments. In the Case of the green roof, the peak flow rates of the original method, SWMM, and intermediate flow are 4.3, 6.8, and 6.3 (x10-5)cms, respectively. In contrast, the method proposed in this study is closer to the field results of 3.9 x10-5 cms. A comprehensive validation was conducted in Jungli District and Chung Yuan Uni versity as the study area to assess the disparity in land optimisation. The rainfall was sim ulated with a 5-meter resolution digital elevation model (DEM) for three return periods of 24 hours. The results demonstrate that the peak flow rate is reduced by 38.3% and the flow histories are relatively delayed for the 10-year return period, with a reduction of 9.7% for the 20-year return period when considering the modified LID and land optimisation simu lations. The proposed method enhances the precision and adaptability of the LID simula tion. The implementation of decentralised LIDs can effectively address surface runoff and alleviate the pressure on the drainage system. Furthermore, optimised land use can accu rately simulate the phenomenon of water passing from the impermeable layer to the per meable layer, and then being stored in the LIDs before overflowing out of the stormwater drains.