Monthly Inflow Forecasting of Three Multi-Purpose Reservoirs

Abstract

The need for inflow discharge forecasts is the first step in the process of integrating water management. To overcome this problem, a discharge forecasting analysis system is needed. This paper adopts a seasonal autoregressive integrated moving average forecasting analysis model, SARIMA. This method was chosen and then applied to the inflow discharge data of the Wonorejo Reservoir to obtain the best model. Determination of the best model through forecasting performance measures using the minimum Mean Square Error (MSE). The best model has an MSE of 11.79 on discharge data for 18 years from 2003 to 2020. The best forecast model is then evaluated on the Bendo Reservoir and Sampean Reservoir. The difference between this paper and others is that one model is used for three different multi-purpose reservoirs and obtains feasible results for each reservoir. Therefore, the authors conclude that the forecasting results of the SARIMA (1,0,0)(0,1,1) 12 model can be applied to Wonorejo Reservoir, Bendo Reservoir, and Sampean Reservoir in East Java Province, Indonesia. The best model from the analysis process is that in the Wonorejo Reservoir, the inflow prediction is satisfactory for the next five years, the Sampean Reservoir for the next four years, and the Bendo Reservoir is the best forecast for the next three years. The results of this forecasting model can be used to analyze the optimization of multi-purpose reservoir management and reduce the risk of reservoir water shortages. Further research can be carried out to achieve extreme values in inflow discharge forecasting.