Flood Hydrograph Analysis Through Employing Physical Attributes Using Two and Multiple Variables Regression Factor and Cluster Analysis

Abstract

Since direct experimental evidence is not available, this must be verified through a modeling approach, provided<br />adequate data be available. Many statistical methods are used to study the relation between independent and<br />dependent variables.This research was carried out at the western part of Jazmurian basin tlocated in the southeast of<br />Iran. In this paperused ten physical characteristics such as area (A), perimeter (Pr), average elevation of basin (av.e),<br />average slope (av.s), gravelious coefficient (G), length of main stream (L), pure slope of main stream (P), length of<br />output to one point equivalent center of basin (Lc), Time of concentration (Tc) and lag time( Tl) as independent<br />variables and nine hydrograph component such as Qp, Q25, Q50, Q75, Tp, T25, T50, T75 and Tb as dependent<br />variables.We investigate flood hydrograph through the physical attributes using two and multiple variables regression<br />factor and cluster analysis.With the data of twelve hydrometric stations. Normality test was done using Kolmograph-<br />Smironov. After using four mentioned methods and with the use of modeling, the relations between dependent and<br />independent variables weres defined. The evaluation of hydrologic model behavior and performance is commonly<br />made and reported through comparisons of simulated and observed variables. Frequently, comparisons are made<br />between simulated and measured stream flow at the catchments outlet. Significant models have correlation coefficient<br />bigger than 0.325 at 0.01 significant level and higher than 0.250 at 0.05 significant levels. Three criteria such as root<br />mean square error (RMSE), relative error (RE) and coefficient of efficiency (CE) were used for selecting the ultimate<br />models. The results revealed that with the use of physical characteristics of the basin we can determine the synthetic<br />hydrograph. The results also showed that the two- variable models have higher efficiency in estimating the discharge<br />variables of the simulated hydrographs. After the cluster analysis for group in which are more station s, it results in<br />more significance of the model than one whose group included less stations.