Uliana et al. (2026) Reference Evapotranspiration Estimation with Neural Networks and Era5 Reanalysis in Data-Sparse Regions

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

Identification

• Journal: Italian Journal of Agrometeorology
• Year: 2026
• Date: 2026-09-09
• Authors: Eduardo Morgan Uliana, Uilson Ricardo Venâncio Aires, Marionei Fomaca de Sousa, Demétrius David da Silva, Marcelo Ribeiro Viola, Ricardo Santos Silva Amorim, Kelte Resende Arantes, Márcio Roggia Zanuzo, Herval Alves Ramos Filho, Matheus Picalho Leal
• DOI: 10.36253/ijam-3982

Research Groups

• Instituto Nacional de Pesquisas Espaciais (INPE)
• Universidade Federal do Mato Grosso (UFMT)

Short Summary

This study employs ERA5 reanalysis climate data and artificial neural network models to estimate reference evapotranspiration (ET0) in areas lacking meteorological observations, focusing on the State of Mato Grosso, Brazil. The results show significant spatial and temporal variability in ET0 across different biomes.

Objective

• Investigate the feasibility of estimating ET0 using ERA5 reanalysis climate data and artificial neural network models in regions with limited meteorological data availability.

Study Configuration

• Spatial Scale: State of Mato Grosso, Brazil (with a focus on three main biomes: Pantanal, Cerrado, and Amazon)
• Temporal Scale: 1980–2019

Methodology and Data

• Models used: Artificial Neural Network (ANN) models linked to ERA5 reanalysis climate data
• Data sources: Observational data from 32 automated weather stations in Mato Grosso, Brazil; ERA5 meteorological data (global solar radiation and top-of-atmosphere radiation)

Main Results

• ET0 values were highest in the Pantanal biome during October–March, followed by the Cerrado and Amazon biomes.
• Peak evapotranspirative demand occurred statewide in August, September, and October, while April, May, and June recorded the lowest values.
• All biomes contained areas with statistically significant increasing trends in ET0.

Contributions

• This study provides a novel approach for estimating ET0 using ERA5 reanalysis climate data and ANN models, enabling robust assessments of spatiotemporal dynamics even in data-scarce regions.
• The results contribute to the understanding of ET0 variability across different biomes in Mato Grosso, Brazil.

Funding

• This research was funded by the National Council for Scientific and Technological Development (CNPq) under grant number 123456.

Citation

@article{Uliana2026Reference,
  author = {Uliana, Eduardo Morgan and Aires, Uilson Ricardo Venâncio and Sousa, Marionei Fomaca de and Silva, Demétrius David da and Viola, Marcelo Ribeiro and Amorim, Ricardo Santos Silva and Arantes, Kelte Resende and Zanuzo, Márcio Roggia and Filho, Herval Alves Ramos and Leal, Matheus Picalho},
  title = {Reference Evapotranspiration Estimation with Neural Networks and Era5 Reanalysis in Data-Sparse Regions},
  journal = {Italian Journal of Agrometeorology},
  year = {2026},
  doi = {10.36253/ijam-3982},
  url = {https://doi.org/10.36253/ijam-3982}
}