Autoregressive time series modeling for prediction of rainfall, runoff and silt load in Jharkhand state of central India

Abstract

India is characterized by marked difference in climate and catchment properties across the country. The study was autoregressive time series model, for prediction of annual rainfall, runoff and silt load in karkara nala watershed Barakar catchment Jharkhand. The area of the watershed is 1751 ha and the data of 9 years from 1993 to 2001 were used to develop the model. The goodness of fit and adequacy of models were tested by Box-Pierce Portmonteau test, Akaike Information Criterion and by comparison of measured and predicted data correlogram. The graphical representation between measured and predicted correlogram. Where in the cases of rainfall, runoff and silt load there is a very close agreement between them. The lower values of error indicate the adoptability of the model for forecasting of rainfall, runoff and silt load. The comparison of different developed models, statistical characteristics and graphical representations, Autoregressive model AR (1) is proposed for generation of silt load and AR (2) is proposed for generation of rainfall and runoff in karkara nala watershed Jharkhand, Clearly shows that the developed model can be use efficiently for the prediction of rainfall, runoff and silt load for the Damodar catchment which is benefit for the farmers and research workers for water harvesting, ground water recharge, flood control and development of their water management strategies.