Time-Series Forecasting Using ARIMA and SARIMA Models for Solar NASA POWER Data

Solar energy forecasting is vital, particularly in energy planning for power utilities, as it provides insights for integrating renewable energy into existing power distribution networks. This study investigates the effectiveness of ARIMA and SARIMA models in forecasting solar radiation (ALLSKY_SFC_SW_DWN) using NASA POWER data from 2019 to 2024. The dataset includes solar irradiance and meteorological parameters critical for renewable energy research. Methodologically, both ARIMA and SARIMA models were implemented and evaluated through R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>, MAE, and RMSE metrics, highlighting their capacity to predict temporal trends and seasonal variations. Results demonstrated ARIMA’s superior accuracy with an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> of 0.9343, while SARIMA, incorporating more seasonal factors, achieved an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> of 0.8664. This work not only compares these models’ practical applications but also contributes to solar energy forecasting research by offering insights into their strengths and limitations for decision-making in renewable energy systems.