Alavi et al. (2025) High-Resolution Crop Evapotranspiration Estimation Using the Automated OPTRAM-ETc Method

Identification

• Journal: Earth Systems and Environment
• Year: 2025
• Date: 2025-10-09
• Authors: Mohammad Alavi, Atefeh Nouraki, Saeid Homayouni, Mohammad Albaji, Mona Golabi, Abd Ali Naseri, Paul Célicourt
• DOI: 10.1007/s41748-025-00852-3

Research Groups

• Department of Irrigation and Drainage, Faculty of Water and Environmental Engineering, Shahid Chamran University of Ahvaz, Ahvaz, Iran
• Centre Eau Terre Environnement, Institut national de la recherche scientifique (INRS), Quebec, QC, Canada

Short Summary

This study proposes an automated Optical Trapezoid Model for crop evapotranspiration (OPTRAM-ETc) using high-resolution Sentinel-2 data to overcome limitations of thermal-optical models. It demonstrates robust field-scale ETc estimation in sugarcane cultivation through novel wet/dry edge parameterization and multi-year temporal data fusion, providing actionable insights for water management in water-scarce regions.

Objective

• Development of an innovative approach to determine dry and wet edges in the OPTRAM-ETc model using high-resolution Sentinel-2 imagery to estimate ETc at the field scale.
• Evaluation of the OPTRAM-ETc model’s performance using five Vegetation Indices (VIs): Fractional Vegetation Cover (FVC), Soil-Adjusted Vegetation Index (SAVI), Enhanced Vegetation Index (EVI), Red Edge Normalized Difference Vegetation Index (RENDVI), and Modified Beer-Lambert Law (MBLL).
• Development of the OPTRAM-ETc model through temporal data fusion.

Study Configuration

• Spatial Scale: Sugarcane cultivation systems in Khuzestan, Iran, covering an experimental section of 225 hectares (250 m × 1000 m fields). High-resolution Sentinel-2 data (10-20 m) was used to generate field-scale ETc maps.
• Temporal Scale: Three consecutive growing seasons (2018–2021), utilizing Sentinel-2 imagery from December 2018 to October 2021.

Methodology and Data

• Models used:
  • Optical Trapezoid Model for crop evapotranspiration (OPTRAM-ETc)
  • Locally calibrated FAO-56 Penman–Monteith model (LCETc) as ground-based reference.
  • Interval-Based Regression and Filtering Method (IRF) for automated wet/dry edge determination.
  • Density-Enhanced Interval-Based Regression (D-IRF) for automated wet/dry edge determination.
• Data sources:
  • Satellite: Copernicus Sentinel-2 MSI Level-2A surface reflectance data (Google Earth Engine), Shuttle Radar Topography Mission (SRTM) dataset (approximately 30 m resolution).
  • Observation: Site-specific meteorological variables (air temperature, wind speed, relative humidity, net radiation) from automated weather stations (2018-2021). Localized biometric data (sugarcane height).
  • Derived: Shortwave-Infrared Transformed Reflectance (STR), Normalized Difference Vegetation Index (NDVI), Fractional Vegetation Cover (FVC), Soil-Adjusted Vegetation Index (SAVI), Enhanced Vegetation Index (EVI), Red Edge Normalized Difference Vegetation Index (RENDVI), Modified Beer-Lambert Law (MBLL), Normalized Difference Water Index (NDWI).

Main Results

• The proposed OPTRAM-ETc approach, using both IRF and D-IRF methods, demonstrated robust performance in estimating ETc, achieving R² values between 0.87 and 0.89 and Root Mean Square Error (RMSE) values of 1.64–1.99 mm d⁻¹ across the three crop years.
• The Density-Enhanced Interval-Based Regression (D-IRF) method significantly improved accuracy, reducing RMSE by 17% and Mean Absolute Error (MAE) by 22% compared to the IRF method, particularly for moderate to high ETc rates.
• The choice of vegetation index (FVC, SAVI, EVI, RENDVI, MBLL) did not significantly impact the accuracy of ETc estimation, confirming the model’s generalizability. OPTRAM-MBLL showed slightly superior performance (R² = 0.89, RMSE = 1.65 mm d⁻¹, MAE = 1.22 mm d⁻¹).
• Incorporating multi-year temporal fusion (2-year and 3-year) further enhanced model performance, with the 2-year fusion model achieving an R² of 0.91 and RMSE of 1.40 mm d⁻¹ for the 2021 crop year, demonstrating improved temporal consistency and robustness.
• The high-resolution ETc outputs enable spatially explicit irrigation scheduling, improve early-season water stress detection, and support long-term monitoring in agricultural systems.

Contributions

• Introduction of a novel, fully automated approach (IRF and D-IRF methods) for determining wet and dry edges in the OPTRAM-ETc model, minimizing user intervention and eliminating the need for site-specific calibration or ground-based auxiliary measurements.
• First comprehensive evaluation of the OPTRAM-ETc model's performance across a broad spectrum of five different vegetation indices, confirming the model's generalizability regardless of the chosen VI.
• Pioneering application of temporal data fusion (using two and three years of data) within the OPTRAM-ETc framework to enhance accuracy and reduce uncertainties stemming from long-term temporal changes in surface conditions, leading to more dependable ETc estimates over extended periods.
• Development of a high-resolution, field-scale ETc estimation method that relies solely on optical remote sensing data (Sentinel-2), removing the dependency on thermal data or individual date image calibration.

Funding

• MITACS Globalink Program (projects IT37025 and IT36900) provided financial assistance for research internships.

Citation

@article{Alavi2025HighResolution,
  author = {Alavi, Mohammad and Nouraki, Atefeh and Homayouni, Saeid and Albaji, Mohammad and Golabi, Mona and Naseri, Abd Ali and Célicourt, Paul},
  title = {High-Resolution Crop Evapotranspiration Estimation Using the Automated OPTRAM-ETc Method},
  journal = {Earth Systems and Environment},
  year = {2025},
  doi = {10.1007/s41748-025-00852-3},
  url = {https://doi.org/10.1007/s41748-025-00852-3}
}