Forecasting hydropower production in Tanzania using the SARIMA model

Abstract

There is a growing demand for energy consumption showing that adequate energy supplies are essential for economic growth. Predicting energy supplies is crucial and thus accurate predictions help minimise the growing energy demand-production gap and enable power plant managers to easily and promptly detect any anomalies or failures in electricity production by analysing deviations from the predicted trends. The main objective of this study was to forecast hydropower production in Tanzania using secondary univariate time series monthly data for the past 22 years (2002-2023). A total of 264 data points were used for the prediction using the seasonal Autoregressive Integrated Moving Average model (SARIMA) following Box and Jenkins methodology’s ability for handling seasonal data. The results show that there will be no substantial decline in hydropower production (KWh) until December 2025. The forecasts show that the hydropower generated, overall will not exceed 227,250,650 KWh but will be at the peak in May 2025 and start to decrease towards December 2025 with not more than a 40% decrease in every month. The forecast will not only help the power plant managers but also policymakers to devise mechanisms that will ensure the gap between energy demand and production is balanced for the welfare of the country’s development.