Abstract:
Water is one of the important natural resource available to mankind. Proper utilization of this resource requires
assessment and management of the quantity and quality both spatially and temporally. A mathematical model provides
quantitative mathematical description of the processes which includes a collection of mathematical equations
expressing relationships between input and output variables through establishing and estimating the relevant
parameters. The ANN models have been used successfully to model the complex non-linear input-output relationship.
An ANN can be defined as data processing system consisting of a large number of samples. Artificial neural networks
(ANN) have found increasing applications in various aspects of hydrology. The study revealed that a feed-forward
artificial neural network with back propagation algorithm having a single hidden layer with two neurons in the hidden
layer was able to model the rainfall-runoff transformation quite accurately. The correlation coefficient during the training
varies between 0.88 and 0.93 and during testing varies between 0.78 and 0.95 respectively whereas the model
efficiency varies between 73.70% and 85.77% with an overall efficiency of 81.18% during training and between 52.62 %
and 90.01 % with an overall efficiency of 66.71% during testing.