翡翠水庫隨著水庫蓄水範圍使用管理辦法等法規的長期實施,點源污染已逐步控制,非點源污染問題的重要性則相對提高。尤其北宜高速公路通車後對於水庫水質的影響,值得持續關注。集水區內非點源污染的傳輸相當複雜,若欲藉由大規模的監測,以獲得合理的評估,在實際上並不可行。故集水區水質模式在經過實測值檢定與驗證後,即成為分析不同土地管理措施下對水文暨水質衝擊的有效工具。 本研究旨在檢驗SWAT(Soil and Water Assessment Tool)模式於模擬翡翠水庫集水區非點源污染的適用性。SWAT模式為集水區尺度、連續時間模擬的模式。此模式整合了CREAMS、GLEAMS和EPIC等美國農業研究中心所研發的模式,以評估集水區在不同管理作業下對水文、泥砂、營養鹽和農藥承載的影響。 研究共收集2002至2006年等資料,以坪林站的流量、泥砂量和營養鹽資料進行模式的檢定和驗證。其中前三年的資料執行檢定,並以2005至2006年的資料驗證結果。月流量經檢定和驗證後,效率係數分別為0.55和0.73,顯示SWAT對水文過程模擬的能力相當不錯;另外5年累積的模擬泥砂總量與實測值的相對誤差只有-0.57%,顯示長期的泥砂承載模擬也十分合理。SWAT對總磷的模擬也非常合理,其月預測值檢定與驗證後的效率係數分別為0.47和0.65;硝酸態氮效率係數稍低,但在統計上仍可被接受,然而模式則明顯低估有機氮與氨氮的承載。綜合驗證的結果,可知SWAT具有長期模擬混合土地使用集水區水文暨水質反應的能力。檢定和驗證後的SWAT模式可進一步標定集水區中非點源承載較嚴重的關鍵區域,再藉由設置適當的BMPs,並應用總量管制(TMDL)的方法,期能有效地控制集水區的非點源污染,以維護翡翠水庫的水質。
During the past decades, point source has been reduced effectively through regulation that the reservoir administration has enforced, therefore, nonpoint source pollution is proved to be the major contribution to the eutrophication problem. The difficulty in assessing and identifying nonpoint source pollution resides in the complex nature of nonpoint source. Moreover, continuous water quality monitoring is expensive and spatially impractical in mixed land use watersheds. Accordingly, the validated watershed-scale water quality models are helpful in assessing water quality impacts which result from different management. This paper aims to calibrate and validate the capability of SWAT(Soil and Water Assessment Tool) model in predicting the nonpoint source pollution generated from the Feitsui Reservoir watershed. SWAT is a watershed-scale, continuous time, physically based hydrologic/water quality model with spatially explicit parameterization. Data collected for major model inputs include landuse and soil distribution maps which were generated by GIS tool as well as weather data from a series of stations within watershed. Flow, sediment and nutrient data in Pinglin station were used to calibrate(2002-2004) and validate(2005-2006) the model’s performance. Preliminary results in flow showed a strong agreement between monthly measured and simulated data with efficiency coefficients greater than 0.5. The validated SWAT model then will be a useful tool to evaluate the effects of appropriate BMPs implementation to control nonpoint source pollution effectively.