農業非點源污染模式應用在蓮華池試驗集水區泥砂量推估

Application of Agricultural Non-Point-Source Pollution Model on Sediment Yield Estimation in Lienhuachi Experimental Watershed

謝宜芸(I-Yun Hsieh)

指導教授 ： 陳明杰

國立臺灣大學/生物資源暨農學院/森林學研究所/碩士(2004年)

https://doi.org/10.6342/NTU.2004.02489

摘要

本研究選擇林業試驗所蓮華池研究中心的4號、5號試驗集水區作為研究對象，蒐集10×10m網格單元的數值地形模型，分析網格的地形特徵，並應用於農業非點源污染模式來模擬特定幾場暴雨產生的逕流量和泥砂產量。
農業非點源污染模式中有多個輸入參數，以實測的逕流量檢核輸入參數的適合度，在參數敏感度分析當中，影響泥砂產量最大的參數為逕流曲線值（CN）值，其次為降雨沖蝕指數（R值），以試誤法決定CN值，使模擬結果的逕流量接近於實測值。在R值方面，考慮研究對象集水區面積較小，故用間隔時間30分鐘的降雨量來計算。
模擬結果顯示，降雨量約在100∼150mm區間時，模擬逕流量最接近實測逕流量，降雨量若少於100mm時，模擬逕流量大都低於實測逕流量，而降雨量在高於150mm的狀況下，模擬逕流量將高於實測逕流量，其中以4號集水區的模擬逕流量與實測逕流值的誤差較小。在洪峰流量部分，模擬洪峰值明顯比觀測值高了許多，其中降雨量約在50∼70mm時，始能在5號集水區模擬出接近實測值的洪峰流量，而至於4號集水區的模擬結果當中，洪峰流量的誤差較大。
在泥砂產量方面，由於蓮華池試驗集水區為林向覆蓋良好的林地，因此在泥砂產量方面數值甚低，選取數場降雨量250mm以下的資料計算結果，總泥砂產量為0∼0.58 tons，4號集水區因為人工林，林相粗糙，地被之枯枝落葉覆蓋較差，而使得泥砂沖蝕量高於林相良好的5號集水區，顯示林相的差異確實會影響土壤沖蝕的能力。

關鍵字

人工林 ； 農業非點源污染模式: 泥砂沖蝕 ； 蓮華池試驗集水區

並列摘要

The object of this study is to estimate sediment yield of Liehwachi forest experimental watershed by Agricultural Non-Point-source Pollution Model (AGNPS model). 10×10m grid unit of digital terrain model (DTM) data of Lienhwachi experimental watershed No. 4 and No. 5 were collected that was used to analyze the topographic characteristics. Also, the analyzed DTM data was applied in AGNPS model to simulate runoff and sediment yield in the condition of specific rainfall occurred.
AGNPS model contains several input parameters. The simulated runoff data by AGNPS model was compared to the observed runoff data, that verify the goodness of input parameters. According to the parameter sensitive analysis, the most influential parameter is runoff curve number (CN), next is rainfall erosivity factor (R). The way of try and error to simulate runoff data is most close to the observed runoff data, that was used to decide CN value. Besides, the R value was calculated by 30 min. interval rainfall intensity.
The results show that when the amount of rainfall is between 100-150mm, the simulated runoff data is most close to the observed runoff data. If the amount of rainfall is less than 100mm, simulated runoff data is lower than the observed runoff data, and rainfall is more than 150mm that is higher than observed data. In the part of peak flow, the simulated data is apparently higher than the observed data. when storm rainfall observed value is about 50-70mm, the simulated value that close to observed value only in watershed No. 5.
In sediment yield estimation aspect, Lienhwachi experimental watershed is covered well by forest, so the value of sediment yield is far lower. Choosing several rainfall event occurred under 250mm, the simulated results of AGNPS model shows that total sediment yield of one rainfall event is 0-0.58 tons. Watershed No. 4 is an artificial forest watershed, canopy and ground cover is not so well as watershed No. 5. Therefore, the simulated sediment yield of watershed No. 4 is higher than watershed No. 5. It indicates that different canopy and ground condition indeed affect the ability of soil erosion.

並列關鍵字

artificial forest ； sediment yield ； Lienhuachi Experimental Watershed ； AGNPS model

參考文獻

13. 周天穎、葉美伶，1997，水里溪集水區檳榔種植對土壤沖蝕之影響及其經濟分析，中華水土保持學報28(2)：87-97。

14. 周天穎、陳瓊珍，1997，德基水庫集水區土壤沖蝕量模擬之研究，中華水土保持學報28(4)：321-332。

17. 張文章、陳清田、陳錦嫣、劉正川，1994，泥岩地區大規模開發對坡地沖蝕之研究，中華水土保持學報25(1)：9-18。

18. 胡蘇澄，1993，蓮華池地區陡坡地栽植檳榔對土壤水文性質與沖蝕性之影響，林業試驗所研究報告季刊8(4)：271-276。

19. 胡蘇澄、盧昭堯、吳益裕、陳台芳，1995，台灣中部蓮華池地區高粘土含量土壤之紋溝間沖蝕率，林業試驗所研究報告季刊10(1):33-40。