Abstract:
The primary objective of this study is to develop a time series model to predict the rainfall for Musi Reservoir Rain gauge station for the future years. Water is an important resource for the development of any community and especially for a developing country like India it should be used judiciously. Rainfall is the major means by which fresh water is made available. Effective and efficient operation of water resource systems frequently demands the prediction of hydrological sequences like rainfall. Hence forecasting models are necessary to predict the future rainfall based on past and current information. Rainfall prediction will tell us whether the coming years will be a good, bad or average rainfall years. Accordingly contingency measures can be planned well in advance, tide over difficulties in augmenting water for the use of the teeming population. In this study forecasting model namely Box-Jenkins ARMA model is used to predict the rainfall. Rainfall data of rain gauge station located at Musi Reservoir Project for past 37 years, (1968-2004) was obtained. Auto correlogram and partial auto correlogram drawn for the series indicate that model can be ARMA (2, 2). Auto Regressive Moving Average model was fitted to the data and checked its validity by comparing fitted values with actual rainfall values. Forecast was made for next 6 years (2005-2010). Forecast values are well fitted with actual values. Thus prediction of rainfall is necessary for better planning of water resources and for optimum water management and utilization.
