Machine Learning-Based Prediction of Daily Solar Radiation to Support Renewable Energy Development in Coastal Regions

Aug 3, 2026

Reliable estimation of surface solar radiation is essential for climate analysis and solar energy planning, especially in data-limited coastal regions such as coastal Africa. The study examines long-term variability of surface solar radiation and assesses machine learning model performance using a comprehensive reanalysis dataset covering 1940–2024. Five radiation components are analyzed to quantify atmospheric attenuation from clouds, aerosols, and water vapor. Five ML algorithms are evaluated via train–test split, k-fold cross-validation, and leave-one-out validation. Results show strong multi-decadal variability and that clear-sky radiation consistently exceeds all-sky radiation. Linear regression achieves near-perfect predictive performance (R² ≈ 1.0) with the lowest errors, while ANN suffers from overfitting, and tree/boosting ensembles remain highly accurate, supporting interpretable, efficient solar resource estimation for resilient renewable energy planning.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages13

Reading Time33 min

AccessFree to download

Introduction

Background and motivation

Energy diversification and the role of solar forecasting

Machine learning methods for solar radiation prediction

Methodology and evaluation

Machine Learning–Based Prediction of Daily Solar Radiation to Support Renewable Energy Development in Coastal Regions Mfon D. Umoh, 1,* Udoh F. Evans, 1 Sunday A. Akpan,2 Sunday E. Otene,2 and Akinola B. Olanrewaju 1 1: Directorate of Strategy, Research and Development, Maritime Academy of Nigeria, Oron, Nigeria 2: Department of General Studies, Maritime Academy of Nigeria, Oron, Nigeria Received January 12, 2026; Accepted February 3, 2026; Published March 5, 2026 Reliable estimation of surface solar radiation is essential for climate analysis and solar energy planning, particularly in data-limited regions such as coastal Africa. This study investigates the long-term variability of surface solar radiation and evaluates the performance of machine learning models for its prediction using a comprehensive reanalysis dataset spanning 1940–2024. Five radiation components—net surface solar radiation (SSR), clear-sky net radiation (SSRC), downward surface solar radiation (SSRD), clear-sky downward radiation (SSRDC), and total surface radiation (TSR)—were analyzed to quantify the influence of atmospheric attenuation caused by clouds, aerosols, and water vapor. Five machine learning algorithms—Linear Regression (LR), Gradient Boosting (GB), Random Forest (RF), k-Nearest Neighbours (KNN) , and Artificial Neural Network (ANN)—were implemented and evaluated using train–test split, k-fold cross-validation, and leave-one-out validation. The results reveal strong interannual and multi-decadal variability in solar radiation, with clear-sky radiation consistently exceeding all-sky radiation, confirming the dominant role of atmospheric modulation in the region. Among the tested models, Linear Regression achieved near-perfect predictive performance (R ² ≈ 1.0) with the lowest error statistics, indicating that surface solar radiation over coastal Africa is largely governed by linear radiative processes. Gradient Boosting and Random Forest also demonstrated high accuracy (R² > 0.98), while the Artificial Neural Network showed poor generalization due to overfitting. The findings demonstrate that computationally efficient and physically interpretable machine learning models can reliably estimate long-term solar radiation in coastal Africa. This provides a robust scientific basis for solar resource assessment, photovoltaic system design, and climate resilient renewable energy planning across the region. Keywords: Solar radiation; Machine learning; Linear regression; Ensemble methods; Coastal Africa; Renewable energy prediction Introduction A clean and dependable alternative to fossil fuels, solar energy is employed all around the world. Only about 5 × 104 EJ of the enormous amount of energy that the sun's beams send to Earth is readily harvestable (1 EJ = 1018J) [1] . About 70% of the 342 W/m2 of solar energy that enters the earth's atmosphere may be harvested, with the remaining 30% being scattered or reflected back into space. In addition to other factors like sun orientation, aerosol, temperature, wind speed, direction, and many more, the earth's rotation around the sun causes variations in solar irradiance, making solar energy an unpredictable resource [1] . The energy that the Sun emits and that travels across space to reach Earth is referred to as solar radiation. It is made up of electromagnetic waves, such as infrared (IR) radiation, ultraviolet (UV) rays, and visible light. Only a small portion of the solar energy that the Sun emits in all directions makes it to the Earth's atmosphere. The Earth's surface eventually receives some of the solar energy [2] . The critical need for energy diversification is highlighted by the depletion of traditional energy resources like fossil fuels, oil, and gas, as well as rising pollution and the consequences of climate change. A sustainable answer to these urgent problems is provided by incorporating renewable energy sources into our ene

FAQs

Why is accurate daily solar radiation prediction important in coastal regions?

It supports climate analysis and solar energy planning where observational data are limited, enabling better solar resource assessment and system design.

Which radiation components are analyzed in the study?

Net surface solar radiation (SSR), clear-sky net radiation (SSRC), downward surface solar radiation (SSRD), clear-sky downward radiation (SSRDC), and total surface radiation (TSR).

How do the tested machine learning models perform, and which one is best?

Linear regression delivers near-perfect predictive performance (R² ≈ 1.0) with the lowest errors; gradient boosting and random forest also show high accuracy, while the artificial neural network generalizes poorly due to overfitting.

Aug 3, 2026

Upload by Valentina