Flood forecasting is an important non-structural solution for reducing flood damages in unprotected flood prone regions (Table 1) Different flood forecasting models are currently in use based upon the availability of data, warning time required and the purpose of the forecast. The present study investigates a recent approach based on Feed-Forward Neural Network (FFNN) to model the runoff process at Jamtara gauging site of Ajay river basin. A univariate FFNN model is developed for flood forecasting using runoff values for six-hours-ahead forecasts. The observed and computed flood hydrographs for six-hours-ahead forecast model are compared and the performance of the developed model is evaluated in term of r.m.s.e. For each flood hydrograph.
