The Estimations of Coefficient of Runoff and Curve Number in Chichiawan Watershed

• Main Author: 楊証中
• Other Authors: 詹仕堅
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/70326949939658660804

碩士 === 國立彰化師範大學 === 地理學系 === 100 === Curve number method which was made by Soil Conservation Service of USA has been a worldwide method for flood estimation, but many scholars have thought that it has some serious problems, including classificational form of antecedent moisture condition and no consideration of characteristics of precipitation. What is above should be explored further in order to increase the correctness of flood estimation. In the thesis, the scope of this area of research is Chichiawan watershed, I use precipitation stations and a stream gauging station from Water Resources Agency, and choose materials of hourly precipitation and hourly stream gauging, which is from 1992 to 2002. Then these materials are carried out under preliminary analysis. In the events of precipitation and runoff of these years, I select the huger storm of one to three in every year, and the total amounts are twenty-one. Then I calculate the value of curve number with the method of base flow separation. In addition, all possible parameters of characteristics of rainfall which affects are calculated, including precipitation, intensity of rainfall and antecedent moisture condition, and coefficient of runoff is figured with the ratio precipitation to runoff volumes. And then, multiple regression analysis is carried out, that is, to analyze the regression between curve number, coefficient of runoff and characteristics of rainfall. After finishing the analysis, in order to reduce the error of the model to the minimum, I use correct estimations, including R2, RMSE and Scatter Diagram. The result of research shows that the effect of antecedent moisture condition is obvious, but curve number should not be separated three grades, and it is a continuous and variable process. Besides, among the characteristics of rainfall, put precipitation, the intensity of rainfall, and the rainfall of antecedent five days into multiple regression analysis, compared with correct estimations, and finally the effects of corelation is great.