Reservoir water availability potential assessment using SARIMA modeling of monthly local rainfall data

Abstract

Abstract The uncertainty of water availability in rainfed reservoirs is one of the impacts of climate change which has become a serious environmental problem in reservoir management. The accuracy of local rainfall prediction is required in planning a sustainable reservoir management strategy. This study aims to construct a monthly local rainfall forecasting model to assess the potential water availability in the reservoir. The SARIMA (Seasonal Autoregressive Integrated Moving Average) forecasting techniques are used in this study to predict local rainfall based on daily rainfall data for the 2007-2019 period at monitoring stations in Gembong Reservoir, Pati, Indonesia. The results showed that the SARIMA model is a good forecasting model according to the MASE (Mean Absolute Scaled Error) criteria. Forecasting using the SARIMA model shows an increase in the accumulated average local rainfall, so there is a potential for water abundance in the reservoir. Reducing sedimentation in reservoirs and maintenance of sustainable reservoir infrastructure is one of the mitigation strategies that need to be prioritized to increase the carrying capacity of the reservoir.