Climate-driven rainfall forecasting for small-scale hydropower potential in Dhaka using SARIMAX modeling

Abstract

Bangladesh is one of the most populous countries in the world and faces significant challenges in meeting its energy demands due to rapid population growth and economic development. Considering this, even a small amount of power can be significant in fulfilling part of the demands. The purpose of this paper was to investigate the feasibility of forecasting hydropower generation in Dhaka, Bangladesh, using predicted precipitation trends. Daily rainfall readings from 1981 to 2025 were accessed and used to determine monthly and annual forecasts for the period 2026-2045. The forecasting model (SARIMAX), which incorporates 2-meter air temperature and specific humidity as exogenous variables, was used, and its performance was compared with historical data. SARIMAX exhibited reliable data, closely matching the actual situation, especially regarding seasonal changes. This model can be suitable for Dhaka as a data-informed approach to renewable energy planning. The forecasted data indicates that Dhaka receives a substantial amount of rain annually, nearly 2,564 mm. The power generation results suggest that the yearly rainfall in Dhaka could generate approximately 0.117 MW, 0.234 MW, and 0.351 MW of continuous power under 1 m, 2 m, and 3 m head conditions, respectively, with an average runoff coefficient of 0.60 and 80% system efficiency. This highlights the importance of using climate forecasting in sustainable energy strategies, especially in rapidly urbanizing and climate-vulnerable regions like Dhaka.