Avilez et al. (2026) A SAFER Model for Accurately Estimating Actual Evapotranspiration in Sugarcane Using Sentinel-2

Identification

• Journal: Sugar Tech
• Year: 2026
• Date: 2026-01-06
• Authors: Alberto Mario Arroyo Avilez, Julian David Ome Narváez, Julián Fernando Mateus-Rodríguez
• DOI: 10.1007/s12355-025-01706-y

Research Groups

• Colombian Sugarcane Research Center (Centro de Investigación de la Caña de Azúcar de Colombia), Cenicaña Experiment Station, Cali, Florida, Colombia

Short Summary

This study successfully applied the SAFER (Simple Algorithm for Evapotranspiration Retrieving) model with Sentinel-2 imagery to accurately estimate actual evapotranspiration (ETa) and crop coefficients (Kc) in sugarcane crops in the Cauca River Valley, Colombia, demonstrating its utility for optimizing irrigation and water management.

Objective

• To estimate actual evapotranspiration (ETa) in sugarcane crops in the Cauca River Valley, Colombia, using the SAFER model and Sentinel-2 images processed with the 'agriwater' R package.
• To calculate reference evapotranspiration (ETo) from local meteorological stations to determine crop coefficients (Kc) as the ETa/ETo ratio.
• To optimize water use and promote efficient water resource management in sugarcane cultivation in semi-arid regions.

Study Configuration

• Spatial Scale: Sugarcane fields across five municipalities (El Cerrito, Ginebra, Guacarí, Buga, Yotoco) in the Cauca River Valley, Colombia, covering an estimated 241,000 hectares.
• Temporal Scale: One sugarcane growing cycle, from January 2022 to June 2024 (2.5 years).

Methodology and Data

• Models used:
  • SAFER (Simple Algorithm for Evapotranspiration Retrieving) model for actual evapotranspiration (ETa) estimation.
  • Penman–Monteith method (FAO 56) for reference evapotranspiration (ETo) calculation.
  • Residual model for surface temperature estimation (combining Sentinel-2 bands with air temperature).
  • Inverse Distance Weighting (IDW) for spatial interpolation of meteorological data.
  • Third-degree polynomial adjustment for trend analysis of ETa and Kc.
• Data sources:
  • Satellite images: 23 Sentinel-2 images (January 2022 - June 2024), providing multispectral data (13 bands) with spatial resolutions of 10 meters, 20 meters, and 60 meters, and a temporal resolution of 5 days.
  • Meteorological data: Hourly data from automated weather stations managed by Cenicaña, including mean air temperature, global solar radiation, wind speed, and relative humidity.
  • Agronomic data: Cédula Cañera database for determining crop age and phenological stages.
  • Software: R with the 'agriwater' package for data processing and SAFER model implementation.

Main Results

• The SAFER model estimated an average actual evapotranspiration (ETa) of 4.04 millimeters per day (mm/day) over the sugarcane growing cycle.
• ETa values ranged from a minimum of 0.1 mm/day at the initial growth stage to a maximum of 8.1 mm/day during the middle growth stage (around 270 days).
• Potential evapotranspiration (ETp) had an average of 4.5 mm/day, with a minimum of 0.14 mm/day and a maximum of 6.1 mm/day, consistently higher than ETa.
• A strong linear relationship was found between ETa and ETp, with a coefficient of determination (R²) of 0.88 and a Root Mean Square Error (RMSE) of 0.56.
• Crop coefficients (Kc) increased from an average of 0.6 in the initial stage to 1.1 (with a maximum of 1.25) in the middle stage, then decreased to 0.8 in the final stage, consistent with crop phenology.
• Spatial distribution maps of ETa revealed significant heterogeneity across sugarcane plots, indicating varying water demands influenced by factors like crop age and management.

Contributions

• Successfully validated the SAFER model using Sentinel-2 imagery for accurate estimation of actual evapotranspiration (ETa) and crop coefficients (Kc) in sugarcane crops in the Cauca River Valley, Colombia.
• Demonstrated the operational versatility of the SAFER model, which does not require "anchor pixels" or thermal bands, making it suitable for regional-scale water management studies.
• Provided a comprehensive spatial and temporal analysis of sugarcane water dynamics over a full growing cycle, offering critical insights for optimizing irrigation practices.
• Integrated diverse data sources (high-resolution satellite imagery, local meteorological data, and agronomic records) with the 'agriwater' R package, establishing a robust methodology for efficient water resource management in high-demand crops.
• Highlighted the potential of remote sensing-based models to address challenges of climate change and agricultural sustainability by improving water use efficiency in semi-arid regions.

Funding

• Colombian Sugarcane Research Center (Cenicaña)
• Sugar mills
• Farmers

Citation

@article{Avilez2026SAFER,
  author = {Avilez, Alberto Mario Arroyo and Narváez, Julian David Ome and Mateus-Rodríguez, Julián Fernando},
  title = {A SAFER Model for Accurately Estimating Actual Evapotranspiration in Sugarcane Using Sentinel-2},
  journal = {Sugar Tech},
  year = {2026},
  doi = {10.1007/s12355-025-01706-y},
  url = {https://doi.org/10.1007/s12355-025-01706-y}
}