Abstract :
Rainfall runoff modeling using neural network technique has been attempted. Neural network gives good non linear relationship and mimics the physical behavior of the real case example: Neural network architecture with input layer, single hidden layer and an output layer was selected. Daily lag time precipitations and daily stream flow are considered as inputs. Different scenario of input combinations for runoff as output was tried. The response of output for input variation was analyzed. It was observed that mean square error show minimum value for a combination of three day lag precipitation and two day lag runoff, further variation increases the mean square error. Nethravathi basin originating from Western Ghats and flowing towards the west reaching the Arabian Sea was selected as study area.
Keyword :
Neural Network, Sensitivity, Rainfall-Runoff.