Adaptive solar irradiance forecasting in arid regions: Enhancing accuracy with localized atmospheric adjustments

Abstract

This study refines the Bird model for forecasting solar radiation on oriented and inclined surfaces in Southern Tunisia, known for its high solar potential. Integrating the model into an Excel tool simplifies radiative transfer equations, enhancing solar forecast precision and accessibility. Over a decade of meteorological data enriches the model with atmospheric parameters like aerosol optical depth, water vapor content, and ozone thickness, extending its applicability across unmonitored regions. The improved model achieves coefficient of Determination (R²) values of 0.989 for Global Horizontal Irradiance (GHI), 0.969 for Direct Horizontal Irradiance (DHI), and 0.974 for Diffuse Horizontal Irradiance (DifHI), with reduced Root Mean Square Error (RMSE) values of 24.66 W/m² for GHI, 36.59 W/m² for DHI, and 12.66 W/m² for DifHI. Optimal panel inclinations are 40° to 60° in winter, 10° to 30° in summer, and 30° to 40° in spring and autumn for maximum solar energy capture. These findings enhance the understanding of solar dynamics in arid regions and support strategic solar energy deployment, offering a scalable framework for similar environments globally and promoting sustainable energy development.