Generation of Annual Stream Flow for Tungabhadra River Using AR Model

Abstract

The present study was carried out to develop a model for generating the synthetic sequences of stream flows for Munirabad dam on Tungabhaa River. The parameters of autoregressive models were determined by method of moment in generating annual stream flow data. Based on the Akaike Information Criterion (AIC), the AR (1) model was found best for generating annual stream flow data. The parameter estimation of autoregressive models, synthetic generation of annual stream flows and performance evaluation of the models are presented and discussed. The performance of method of moment model in generating the annual stream flow data was tested by comparing the statistical characteristics i. e. , mean, standard deviation, kurtosis and skewness with historical data.