降雨空間變異性對非點源污染推估之影響

Abstract

集水區非點源污染推估一般是基於一個代表性降雨量推估逕流量及沖刷下來的非點源污染量，所以降雨量是影響非點源污染推估的重要參數之一。然而降雨量會隨著空間而不同，並非是一個固定的量，故以單一代表性雨量為基準，可能會造成非點源污染量推估上的誤差，進而影響後續防治工作之規劃及決策。因而有必要分析集水區雨量在空間之空間變異性對推估非點源污染量可能造成之影響。本研究以Thiessen法、K-nearest法與Inverse Distance三種權重法推估雨量之空間分佈上之變異性，分別依上述方法分析雨量之空間變化，及比較所推估非點源污染量之差異性，而非點源污染是採用單一場次版AGNPS模式模擬其流佈與污染量。 本研究以德基水庫集水區為案例區，採用集水區中十個雨量站降雨資料，並將集水區分為26與56子集水區下，分析雨量之空間變異，進而評估雨量空間變異在推估非點源污染上之影響。結果顯示，由於Thiessen法只以最近一個雨量站為依據，而其他方法是以數個雨量站值的權重平均，故Thiessen法所得結果變異性較大。唯依三個方法結果，發現不少暴雨事件在同一場次降雨下，整個德基集水區降雨量隨著空間不同而有明顯之差異，這些變異導致後續非點源污染量推估值差異大，故不適合以一個單一雨量代表整個集水區降雨值，雨量空間變異性使各子集水區推估流入水庫總磷污染貢獻量之趨勢有所改變，令子集水區污染量之排序改變，進乃影響管制優先順序之決策。故應考量雨量空間變異性，較符合實際情形，且較能改善決策品質。 降雨之時間變異性亦會影響非點源污染之推估，故本研究亦以逐月、溼乾兩季、四季之分季方式，探討時間變異性對非點源污染的影響，雖然此部分對推估非點源污染量甚為重要，唯因時間變異性成果較不顯著，故暫列入附錄中，以供後續進一步研究時參考。

Rainfall is one of the most important departments counting NPSP because of the extent of NPSP generated from catchment areas surrounding a reservoir is normally estimated based on a design rainfall. However, rainfall generally varies in different space. This spatial change may significantly influence the estimation of runoff volume and associated NPSP arising within a reservoir watershed. The NPSP management, reservoir operation, and water treatment operation strategies developed based on the estimation will also be altered. This study therefore analyzes the effects of spatial variation in precipitation on the estimation of pollution levels from non-point sources in a reservoir watershed. We use three common methods Thiessen polygun, the inverse distance and K-nearest-neighborhood to analysis the difference of rainfall spatial variation and differs the differences between the NPSP. The NPSP is counting form one event AGNPS model. The Derchi reservoir watershed located in central Taiwan, R.O.C. is the area studied. Precipitation data monitored at 10 gauge stations and 26 and 56 sub-watersheds are used to assess the spatial rainfall variations and NPSP within the watershed. According to the results obtained in this study, because of according to the nearest rain gauge, the Thiessen method have the largest spatial variation. Many rainfall events have obvious spatial variations effecting the variations of NPSP. It’s not good using one value rainfall to count the NPSP. Spatial variations of rainfall effect the sort of NPSP value arising form each sub-watershed and will effect the decision of watershed controls . Because of getting real NPSP and having the right watershed control, we should consider the rainfall spatial variation. This time series also influence the estimation of NPSP arising within a reservoir watershed. This study therefore analyzes the effects of monthly and seasonal variation in precipitation on the estimation of pollution levels from non-point sources in a reservoir watershed. Because of the result is not good, we put it in appendixes.