Constructing the downscale precipitation using ANN model over the Kshipra river basin, Madhya Pradesh

Abstract

The present study is focused on simulating the impact of climate change on the behavior of precipitation of Kshipra river basin in Madhya Pradesh, India. Artificial neural network (ANN) model was used to construct of the downscale precipitation scenario. A General Circulation Model (GCM) viz. Hadley Centre Coupled Model, version 3 (HadCM3), from Hadley Center, UK has been used for the study. In Model, monthly weather data on the basis of rapid economic growth under A1B scenario (A balanced emphasis on all energy sources) were considered. The four predictor variables which are used in ANN model formulation are screened from a set of 26 predictors based on correlation analysis of observed precipitation. The basic ANN architecture was optimized for training of the model by first selecting the training algorithm and then varying the number of neurons in the hidden layer. Twelve different training algorithms have been used. Further, the model was evaluated by varying the number of neurons from 1 to 30 in the hidden layer. The performance of model was evaluated in terms of the correlation coefficient (R), mean square error (MSE), root mean square error (RMSE) and mean absolute error (MAE). The results of model revealed that the predicted precipitation and observed precipitation are better correlated (R=0.911 and 0.853 during training and validation runs) with back propagation variable learning rate “traingdx” algorithm. © 2016, Association of Agrometeorologists. All rights reserved.