Regionalization of Parameters of Rainfall-Runoff Model for Large Ungauged Catchments

Abstract

Predicting accurately inflows into the reservoirs in order to operate the dam Gates to safely release the floods in Indian rivers has been a difficult task for the project engineers. Generally, the flood water is released after the reservoir level reaches the danger zone, causing man-made flood problem in the downstream of the reservoir. One of the major constraints in developing forecast model is the scarcity of climatic data and raingauge network. With these limitations, the SCS-CN method developed by USDA is suitable for predicting inflow into the reservoirs by using the single model parameter “the Curve Number” and rainfall as input to the model. In this study, the hydro-meteorological data observed in Wainganga sub-basin of the Godavari river basin in Madhya Pradesh has been used to develop of the SCS-CN based rainfall-runoff model. In the SCS runoff model, the total rainfall is separated into three components, viz. the initial abstraction, actual retention after runoff begins and actual runoff. The SCS model computes the runoff per storm basis. In India, generally the rainfall is recorded on daily basis, i.e. during last 24 hours. However, in reality there may be multiple storms of shorter durations in a day or any storm may continue for more than one day during the monsoon months. The observed rainfall and runoff data have been used for development and validation of the flow model under different watershed conditions, AMC-I, AMC-II and AMC-III. The Direct Runoff computed from the model has been used for comparison with the observed flow to check the efficiency of the Rainfall-Runoff model. The results show good correlation between the observed runoff and model runoff.