Solar Radiation Analysis for Predicting Climate Change Using Deep Learning Techniques

Abstract

Solar radiation, Earth's main energy source, affects surface radiation balance, hydrological cycles, plant photosynthesis, weather, and climatic extremes. A stacking model based on the best of 12 machine learning models predicted and compared daily and monthly sun radiation levels. The results suggest machine learning algorithms use climatic parameters. A trend study of high land surface temperatures and solar radiation showed how solar radiation compounds catastrophic climatic events. GBRT, XG Boost, GPR, and random forest models better predicted daily and monthly sun radiation. The stacking model, which comprises the GBRT, XG Boost, GPR, and random forest models, exceeded the single models in daily solar radiation prediction but did not outperform the XGBoost model in monthly prediction. Stacking and XG Boost models estimate sun radiation best.