DEVELOPMENT OF RAINFALL- RUNOFF MODEL INCORPORATING STORM DURATION

Abstract

Estimation of runoff is required for planning, design and management of the water resources projects. With this end in view a number of models varying from simple empirical relations to most complex physically based models have been developed to simulate the complex process of runoff generation from rainfall. The Natural Resource Conservation Service — Curve Number (NRCS-CN) model, formerly named and still popular as Soil Conservation Service Curve Number (SCS-CN) model is one of them. Besides being fairly accurate in runoff prediction, the versatility of SCS-CN applications lies in the fact that the method is simple, easy to understand and apply, stable, and capable of incorporating several watershed runoff producing characteristics, viz, soil type, land cover and practice, hydrologic condition, and antecedent moisture condition (AMC).The model is very useful for ungauged watersheds'. The method computes volume of surface runoff for a given rainfall event using parameter - curve number, which is derived from watershed characteristics and 5-day antecedent rainfall. In spite of its widespread and lasting success, several significant issues in the conventional SCS-CN model need to be studied and investigated further. The following issues such as (1) implementation of antecedent moisture condition (AMC) procedure; (2) consideration of 'mean' or `median' CN as a representative CN of a watershed in the model; (3) fixation of initial abstraction ratio (A) at 0.2; (4) potential maximum retention parameter usage; and (5) effect of storm intensity or duration in the runoff estimation, fall in this category. Based on these identified issues, the present study is aimed at enhancing the model predictions by improving the hydrologic algorithms contained within them and making the model a simple, realistic and logical one in its structure. Besides this the performance of suggested models vis-a-vis already developed models has been evaluated by applying them to 82 small watersheds of the United States of America (U.S.A.). As demonstrated in, the study, the new proposed models are very useful for. reliable prediction of runoff from rainfall in small watersheds.