Uliana et al. (2025) Estimation of Global Solar Radiation in Unmonitored Areas of Brazil Using ERA5 Reanalysis and Artificial Neural Networks

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Atmosphere
• Year: 2025
• Date: 2025-11-19
• Authors: Eduardo Morgan Uliana, Juliana de Abreu Araujo, Márcio Roggia Zanuzo, Alvaro Henrique Guedes Araujo, Marionei Fomaca de Sousa Junior, Uilson Ricardo Venâncio Aires, Herval Alves Ramos Filho
• DOI: 10.3390/atmos16111306

Research Groups

Not explicitly mentioned in the provided text.

Short Summary

This study developed and trained an artificial neural network (ANN) model using ERA5 reanalysis data to accurately estimate and map global radiation (GR) distribution, demonstrating its efficacy in areas with sparse sensor networks.

Objective

• Train an artificial neural network (ANN) model to estimate global radiation (GR) based on ERA5 reanalysis data.
• Map the spatial distribution of GR in the study area (Mato Grosso, Brazil).

Study Configuration

• Spatial Scale: Mato Grosso, Brazil
• Temporal Scale: Not explicitly mentioned, but historical time series were generated.

Methodology and Data

• Models used: Artificial Neural Network (ANN)
• Data sources:
  • Global radiation (GR) observations from 32 automatic weather stations of the Brazilian National Institute of Meteorology (for model training).
  • ERA5-ECMWF reanalysis data (for model input: air temperature, precipitation data).
  • Top-of-atmosphere solar radiation (R0) calculated from latitude and day of the year (for model input).

Main Results

• The calibrated ANN model achieved high accuracy in estimating global radiation (GR), with Nash–Sutcliffe and Kling–Gupta efficiency indices exceeding 0.99.
• The model successfully generated historical time series and spatial distribution maps of GR for the study area.
• ERA5 reanalysis data, when used as input for an ANN, enables precise and accurate GR estimations, even in locations without ground-based meteorological stations.

Contributions

• Development of a highly accurate artificial neural network (ANN) model for global radiation (GR) estimation utilizing readily available ERA5 reanalysis data.
• Provision of a robust methodology to generate GR historical time series and spatial distribution maps in data-sparse regions.
• Demonstration of ERA5 data's effectiveness as a primary input for ANN-based GR estimation, offering a valuable solution for areas lacking ground-based meteorological stations.

Funding

Not explicitly mentioned in the provided text.

Citation

@article{Uliana2025Estimation,
  author = {Uliana, Eduardo Morgan and Araujo, Juliana de Abreu and Zanuzo, Márcio Roggia and Araujo, Alvaro Henrique Guedes and Junior, Marionei Fomaca de Sousa and Aires, Uilson Ricardo Venâncio and Filho, Herval Alves Ramos},
  title = {Estimation of Global Solar Radiation in Unmonitored Areas of Brazil Using ERA5 Reanalysis and Artificial Neural Networks},
  journal = {Atmosphere},
  year = {2025},
  doi = {10.3390/atmos16111306},
  url = {https://doi.org/10.3390/atmos16111306}
}