Theme 1-6-Neural computation technique for estimating of hydrologic cycle

Abstract

An artificial neural network (ANN) methodology was employed to estimate the values of crop evapotranspiration (ET) and evaporation from daily values of air temperature. The study investigated two ANN architectures, a radial basis function network and a multi layer feed forward back propagation network for mapping an input output relationship. The ANN models compared favorably with observed values of the processes. At the same time, it represents an improvement upon the prediction accuracy and flexibility over current methods. A statistical analysis of the results suggested that the ANN model could be employed in estimating ET and evaporation using limited weather data with reasonable accuracy.