應用HEC-1水文模式探討福山一號集水區之降雨－逕流關係

Abstract

本研究以林業試驗所福山研究中心一號試驗集水區為試驗地，應用HEC-1水文模式，選用模式提供之運動波法，為逕流歷線演算的方法。目的在於探討福山一號集水區之降雨－逕流關係與模式對小面積集水區之適用性。在2002年至2006年的觀測降雨量及水位資料中，挑選8場單峰型態降雨事件為模式參數檢定基準，並以11場近似單峰降雨型態降雨事件，以及雷馬遜颱風、辛樂克颱風、敏督利颱風、艾利颱風、納坦颱風、碧利斯颱風、寶發颱風與珊珊颱風等8場颱風期間降雨事件，進行模式驗證。 根據觀測資料所作之降雨－逕流歷線圖，發現幾乎在所有的降雨事件中，降雨幾乎立即反應成為渠道逕流輸出；換言之，如福山一號集水區這樣植被覆蓋良好、人為干擾甚低的集水區，對洪峰逕流量與渠道逕流輸出量的減緩效果，應不能被過分強調。 經由模式模擬結果發現，模擬結果僅能顯現降雨事件之降雨逕流歷線特徵，而利用洪峰流量誤差百分比、洪峰到達時間誤差及總逕流體積誤差百分比三個目標函數驗證模式之結果，顯示此27場降雨事件之洪峰流量誤差百分比均異常偏高（平均值：1292.5%），雖然對於8場單峰型態降雨事件及11場近似單峰型態降雨事件，其洪峰到達時間誤差與總逕流體積誤差百分比的驗證結果，顯示模式模擬結果尚稱良好；但8場颱風期間降雨事件中，有3場的洪峰到達時間誤差顯示洪峰提早到達，而有過半場次之總逕流體積誤差百分比超過100%。 由於模式模擬結果之目標函數驗證結果不佳，故HEC-1水文模式不適合於福山一號集水區之降雨－逕流關係的模擬。

This study was to probe the rainfall-runoff relationships of FuShan Experimental Watershed No. 1 (FS1) of Taiwan Forestry Research Institute (TFRI). HEC-1 model based on the Kinematic wave method was used to calculate runoff hydrograph, that compare to the observed runoff hydrograph, to examine whether HEC-1 would be applicable to small watershed or not. Among the collected rainfall-runoff data at FS1 during 2002 to 2006, 8 rainfall-runoff cases with single-peaked hyetograph were chosen to calibrate the input parameters of HEC-1 model. Also, another 11 rainfall-runoff cases with similarly single-peaked hyetograph and 8 typhoon cases, namely Rammasun, Sinlaku, Mindulle, Aere, Nock-Ten, Bilis, Bopha, and ShanShan, were used to verify the reliability of the input parameters. From the collected rainfall-runoff data analysis, we found that the rainfall would immediately turn into runoff in most of the hydrological cases. In other words, the hydrological function of water conservation at FS1 should not be over emphasized considering the rapid change from rainfall to runoff in this area. The simulation results done based on HEC-1 model only demonstrated the characteristics of rainfall-runoff hydrograph of all the hydrological cases. However, if we use 3 target functions, Error of Peak Discharge (EQP), Error of Time to Peak (ETP), and Error of Total Runoff Volume (EV), to evaluate the simulation results, it shows that the values of EQP of all the 27 rainfall-runoff cases are exceptionally high (Avg. 1292.5%). Although the values of ETP and EV did not show negative results for the 8 rainfall-runoff cases with single-peaked hyetograph and the 11 rainfall-runoff cases with similarly single-peaked hyetograph, still 3 out of the 8 typhoon cases indicate an early timing of peak discharge. Besides, more than half of the 8 typhoon cases appear that the total runoff volume are over 100%. According to the results of the 3 target functions are quite convincing, we could conclude that HEC-1 model is not a proper model to simulate the rainfall-runoff relationships of FS1.