Effect of catchment characteristics on the relationship between past discharge and the power law recession coefficient

Abstract

This study concerns the relationship between the power law recession coefficient k (in −dQ/dt = kQα, Q being discharge at the basin outlet) and past average discharge QN (where N is the temporal distance from the center of the selected time span in the past to the recession peak), which serves as a proxy for past storage state of the basin. The strength of the k–QN relationship is characterized by the coefficient of determination R2N, which is expected to indicate the basin’s ability to hold water for N days. The main objective of this study is to examine how R2N value of a basin is related with its physical characteristics. For this purpose, we use streamflow data from 358 basins in the United States and selected 18 physical parameters for each basin. First, we transform the physical parameters into mutually independent principal components. Then we employ multiple linear regression method to construct a model of R2N in terms of the principal components. Furthermore, we employ step-wise multiple linear regression method to identify the dominant catchment characteristics that influence R2N and their directions of influence. Our results indicate that R2N is appreciably related to catchment characteristics. Particularly, it is noteworthy that the coefficient of determination of the relationship between R2N and the catchment characteristics is 0.643 for N = 45. We found that topographical characteristics of a basin are the most dominant factors in controlling the value of R2N. Our results may be suggesting that it is possible to tell about the water holding capacity of a basin by just knowing about a few of its physical characteristics.