Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/33281
標題: A Study on Rainfall-Runoff Relation of Qing-Shui River Upstream Watershed with Kinematic Wave Method of HEC-1 Model
應用HEC-1運動波法探討清水溪上游集水區降雨逕流之關係
作者: 林燕旻
Lin, Yen-Min
關鍵字: 運動波法;Kinematic Wave Method;HEC-1水文模式;CN值;臨前水分條件;移動平均法;Hydrology Engineering Center;Curve Number;Antecedent Moisture Condition;Moving Average Method
出版社: 水土保持學系
摘要: 
The study builds the two-D space information of the upstream watershed of Qing-shui river, which is a branch of Zhuo-shui river. The work is done before and after the earthquake with remote sensing method. For getting more precise geographical parameters, we use GIS ArcView and combine HEC-1 with kinematics wave to adjust the parameters of six typhoons, including Mary, Brian, Caitlin, Winnie, Zeb, and Herb. Finally, we use typhoon Amber and Otto to verify the correctness of the parameters, in order to find out the rainfall-runoff relation of the upstream watershed of Qing-shui river. We have the following conclusions:
1. We use five parameters, including length, slope, roughness coefficient for overland flow, roughness coefficient for channel, and CN value, to do the parameter sensitivity analysis. The result shows that CN value is the most sensitive among them. So we do the parameter adjustment of CN value with the first six rainstorms to make errors of three objective functions as small as possible. We do not adopt the CN value of the typhoon Herb, because its CN value is far smaller than the others are. The errors of flood peak flows evaluated from sequences of the other five typhoons or rainstorms are all within 7% and the timings at which flood peaks arrived have no error. As to the error of the total runoff volume, in the case of typhoon Caitlin, the value of the total runoff volume is underestimated because of poor total rainfall. The errors of the other cases are all within 7%. These adjustments are fine. Besides, typhoon Amber and Otto verified the correctness results are fine.
2. Except for the effect of the antecedent rainfall condition, we can find that the CN value used in the adjustment of parameter correction is related to soil and water states in full-rain or poor-rain year after the trend analysis with the dynamic average method of total annual rainfall. If we can know that the rainstorm events are in full-rain or poor-rain year, it will make the parameter verification of the model more easy and to achieve the precise CN magnitude more quickly.
3. In 1999, the 921 earthquake makes the midland of Taiwan quickly and seriously breakdown, resulting in the tremendous change of the plant growing coverage in the watershed of our test domain. So we choose the rainfall data of Zeb typhoon before the earthquake and estimate the runoff in Qing-shui river upstream watershed after the earthquake by CN value estimated with three types of AMC before and after earthquake. The results show that the flow did not increase obviously after the earthquake, because the increased are a of the bare land is 9.5 km2, which only covers 3.7% area of the test domain. Thus its effect on the flow in the study region is not apparent.

本研究旨在以濁水溪之支流清水溪上游之集水區為試區,藉遙測技術完成地震前後之二維空間資料之建立,利用地理資訊系統軟體ArcView精確地粹取集水區之地文參數,並結合HEC-1水文模式中之運動波法,以瑪麗、寶莉、凱特琳、溫妮、瑞伯、賀伯六場颱風暴雨進行參數檢定,並以安珀、奧托兩場颱風暴雨進行模式驗證,以探討清水溪上游集水區降雨-逕流之關係,獲得結論如下:
一、 以漫地流長度、坡度、粗糙係數、渠道粗糙係數和CN值等五個參數進行參數敏感度分析,顯示CN值最敏感 ,故以調整CN值進行前六場暴雨之參數檢定,以使三個目標函數之誤差為最小。結果除了賀伯颱風因其CN 值遠比其他場次之暴雨小,而難以通盤配合故不予列入驗證用之平均外,其餘五場颱風暴雨所推估之洪峰流 量誤差在7 %以內,洪峰到達時刻均無誤差,在總逕流體積誤差方面則凱特琳颱風因其總降雨量偏小,低估了 總逕流體積,其他四場誤差在7 %以內,檢定結果良好。另外安珀與奧托颱風驗證結果也良好。
二、 暴雨的降雨損失受到臨前水分條件的影響最大,以年總降雨量經動平均法行趨向分析之後,發現調整參數所 乘以的CN值乘數的大小與豐水年和枯水年有關,故在調整CN值時若能先進行趨向分析以瞭解各暴雨事件是 處在豐水年或枯水年之水文背景環境的話,將有助於模式之參數檢定更快速獲得精確之CN值。
三、 1999年之921大地震造成台灣中部地區極嚴重之崩塌,使本試區內之局部地表植生覆蓋發生劇烈的改變,故 以地震前之瑞伯颱風雨量資料和地震前後AMC各三類所推估的CN值以推估地震後試區之逕流量。結果顯示 地震前後之流量並無顯著改變,此乃因新增裸露地的面積為9.5 km2,只佔全試區面積254.0Km2的3.7 %而已, 故對試區流量的影響不大。
URI: http://hdl.handle.net/11455/33281
Appears in Collections:水土保持學系

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.