Development of a new method for estimating SCS curve number using TOPMODEL concept of wetness index (case study: Kasilian and Jong watersheds, Iran)

Abstract

In recent years, several empirical and mathematical methods have been developed to estimate runoff, among which the SCS curve number (SCS-CN) method is one of the simplest and most widely used methods. The SCS-CN depends mainly on a CN parameter which corresponds to various soil, land cover, and land management conditions, selected from look-up tables. An application of GIS and RS techniques along with filed investigations made it possible to enhance the method from a lumped one to the level of semi-distributed models in which a specific value can be assigned to each cell in raster maps. The up-to-date procedures require several datasets, field measurements and overlying issues which limits the use of SCS-CN in data-scarce regions. In this research a new method has been developed which estimates the SCS-CN over the catchment with a minimum input dataset and acceptable accuracy and is based on the saturation-excess concept, which is used in the semi-distributed model: TOPMODEL. The proposed method depends on three parameters, including ndrain (soil porosity), $${\bar{\text{z}}}$$ (average distance to watershed water table surface) and m (which controls the effective depth of the saturated soil) and one input dataset, the so-called topographic index. Results showed that the maximum and minimum differences between the basin-averaged CN based on the GIS and RS techniques and the proposed method for Kasilian and Jong watersheds are 12% and 0.3%, respectively. Also, the findings indicated that, of the three parameters of proposed method, the m parameter plays a key role and that by increasing this parameter the basin-averaged CN tends to decrease and vice versa. Because of the dependence on a topographic index, the proposed method is strongly affected by DEM resolution and there are significant differences between low and high-resolution DEMs. However, for a small scale watershed, similar to Kasilian, using DEMs with resolution lower than 100 m considerably decreases the above differences. As an overall conclusion, the proposed method provides acceptable values of SCS-CN which is important for running rainfall-runoff model in a data-limited or data-scarce regions. In addition, creating the gridded map for CN, which is required in most hydrological models, is one of the most important advantages of the proposed method.