Abstract:
To use sufficiently fine data interval to catch the complete temporal variations of observed flows is impossible and impractical. It is important to see that the number of parameters of the model is not so large that modelling the process becomes cumbersome. Utilising easy accessibility to computational facilities like PC's and the acceptability of concepts of physics to describe the land phase of hydrological cycle in space and time daily rainfall - runoff modelling is being made easy and simple. Also for most of the catchments, daily data records are available for both rainfall and flow.
In this study, a simple 5-parameter model, based on the concept of probability distributed method is applied to simulate the daily runoff over a 9-year period of 1989 to 1997 at Tammavaram on the Gundlakamma River in Prakasam district of Andhra Pradesh. A Fortran 77 program is used to undertake automatic optimization of the model to simulate the observed flows using an appropriate objective function. From the calibration and validation of the modelling study, it is found that the model, though a five-parameter one, could respond properly to the rainfall and resulted in a reasonable efficiency of 72.14% in calibration and 68.25% in validation.
