不同空間分布情況之雨量資料及堵塞因子對都會區淹水之影響

Abstract

近年來更因氣候變遷及極端氣候事件之影響，使得降雨強度增加，此現象在夏季颱風時也有同樣趨勢，讓淹水災害發生之頻率提高。目前淹水模擬通常是設置單一雨量站來進行模式模擬，此方法雖然模式建置速度較快且簡單，但因單一雨量站可能無法代表一整區之降雨情形；另外，都會區之排水方式通常是由收水器和側溝收水器先收集雨水，若遇到強降雨發生垃圾及落葉堆積造成堵塞現象，使得收水器無法達到設計之收水效果，都有可能造成都會區的積淹水。 本研究為探討不同空間分布情況之雨量資料對於都會區淹水模擬的影響，分別使用雨量站觀測資料以及雷達回波雨量資料進行淹水模擬，比較單一雨量站、徐昇權重法、徐昇分區法、距離權重反比法以及QPESUMS雷達雨量資料，並藉由準確度、偵測率、預兆得分及精確度等方法，評估出何種雨量設置對都會區淹水模擬有較佳之表現。接著設置不同堵塞率以比較何種堵塞率值較接近真實都會區堵塞之情形。 本研究以臺北市士林區街上次分區及後港分區為研究區域，並以兩場短延時事件及一場長延時事件進行模擬與分析，結果顯示，較符合真實淹水事件為距離權重反比法雨量設置，無論是長、短延時事件，其街道淹水範圍之模擬結果皆較貼近實際調查狀況，並設置不同堵塞率得到堵塞因子20%為最接近真實都會區之堵塞率，或許可作為都會區堵塞因子設置之參考。

In recent years, due to climate change and extreme weather events, the rainfall intensity has increased. This phenomenon also has the same trend in summer typhoons, which increases the frequency of flooding disasters. At present, the flooding simulation usually sets up a single rainfall station for model simulation. Although this method is faster and simpler to construct, the single rainfall station may not be able to represent the rainfall situation of a whole area. In addition, the drainage method in the metropolitan area usually collects rainwater from the inlet and the side ditch. If the garbage and the fallen leaves accumulate due to heavy rainfall, the water inlet will not reach the designed water collecting effect, which may cause the metropolitan area the product of flooding. In this study, we investigated the effects of rainfall data from different spatial distributions on flooding simulations in urban areas, using rainfall observation data and radar echo data to simulate flooding. Comparison of single rainfall station, multi-rainfall station Thiessen's Method rainfall setting, multi-rainfall station Thiessen's Method Partition rainfall setting, Inverse Distance Weighted rainfall and radar echo rainfall. Through the methods of Accuracy (ACC), Probability of Detection (POD), Threat Score (TS), Precision or Predictive value (PPV), it is estimated that what kind of rainfall setting has better performance for flooding simulation in the urban area. Then set different clogging rates to compare the clogging rate values closer to the real city blockage, and compare the clogging rates with the Taipei city flooding map. In this study, Jieshang drainage and Hougang drainage basin in Shilin Distrinct, Taipei, is selected as the study area. Two short duration and one long duration rainfall events are used for simulation and analysis. Due to the closeness to real rainfall collection phenomena, the simulated results from Inverse Distance Weighted rainfall setting has a better agreement whether the duration of rainfall event is short or long. And set different clogging rate to get the clogging factor 20% as the closest to the real urban area clogging rate. It is also comparable to the flooding simulation map of Taipei City. The result of setting a 20% clogging rate can make the indicators have better performance.