Liu et al. (2026) Cropland evapotranspiration based on Sentinel-2 shortwave infrared data and ensemble Kalman filter

Identification

• Journal: Agricultural and Forest Meteorology
• Year: 2026
• Date: 2026-01-27
• Authors: Lu Liu, Yunjun Yao, Qingxin Tang, Xueyi Zhang, Yufu Li, Joshua B. Fisher, Jiquan Chen, Jia Xu, Xiaotong Zhang, Ruiyang Yu, Zijing Xie, Jing Ning, Jiahui Fan, Luna Zhang
• DOI: 10.1016/j.agrformet.2026.111035

Research Groups

• State Key Laboratory of Remote Sensing and Digital Earth, Faculty of Geographical Science, Beijing Normal University, Beijing, China
• Beijing Engineering Research Center for Global Land Remote Sensing Products, Faculty of Geographical Science, Beijing Normal University, Beijing, China
• School of Geography and Environment, Liaocheng University, Liaocheng, China
• Key Laboratory for Meteorological Disaster Monitoring and Early Warning and Risk Management of Characteristic Agriculture in Arid Regions, CMA, Yinchuan, China
• Jincheng Meteorological Administration, Jincheng, China
• Schmid College of Science and Technology, Chapman University, Orange, CA, USA
• Department of Geography, Environment, and Spatial Sciences, Michigan State University, East Lansing, MI, USA
• Department of Infrastructure Engineering, Faculty of Engineering & IT, University of Melbourne, Melbourne, VIC, Australia

Short Summary

This study introduces a novel STR–EnKF–PT model to accurately simulate daily cropland latent heat of evapotranspiration (LE) at a 20-meter resolution using Sentinel-2 data. The model demonstrated superior performance and robust probabilistic forecasting capabilities when validated against ground observations from eddy covariance sites across the United States.

Objective

• To develop and evaluate a novel shortwave infrared-transformed reflectance (STR)-and ensemble Kalman filter (EnKF)-based Priestley–Taylor (STR–EnKF–PT) model for simulating daily cropland latent heat of evapotranspiration (LE) at a fine spatial resolution using Sentinel-2 data.

Study Configuration

• Spatial Scale: Cropland LE at 20-meter resolution across six different regions in the United States.
• Temporal Scale: Daily LE estimation, assessed over a two-year period from 2019 through 2020.

Methodology and Data

• Models used: STR–EnKF–PT model (integrating Shortwave infrared-transformed reflectance, Ensemble Kalman filter, and Priestley–Taylor model).
• Data sources: Sentinel-2 shortwave infrared data; Ground observations from 10 eddy covariance (EC) sites in the United States.

Main Results

• The STR–EnKF–PT model outperformed competing methods at four validation sites.
• Validation metrics showed a coefficient of determination (R²) of 0.54–0.84 (at 99 % confidence level), a root-mean-square error (RMSE) of 26.1–38.0 W/m², a Kling–Gupta efficiency (KGE) of 0.69–0.91, and a bias of -16.3–9.4 W/m².
• The ensemble system exhibited robust probabilistic forecasting with a mean reliability score of 0.0010, a mean continuous ranked probability score (CRPS) of 29.91 W/m², and appropriate spread-error relationships with a mean ratio of 1.239.
• The model accurately depicted spatial patterns of cropland LE at 20-meter resolution, effectively distinguishing the LE status of cultivated land.

Contributions

• Introduces the innovative use of soil moisture (SM) constraints derived from shortwave infrared-transformed reflectance (STR) for high-resolution (20-meter) cropland LE estimation.
• Significantly enhances estimation accuracy and enables reliable probabilistic forecasting through the incorporation of the Ensemble Kalman Filter (EnKF).
• Provides a practical approach with significant implications for efficient cropland irrigation water utilization and agricultural risk management.

Funding

• The provided paper text does not contain specific funding information.

Citation

@article{Liu2026Cropland,
  author = {Liu, Lu and Yao, Yunjun and Tang, Qingxin and Zhang, Xueyi and Li, Yufu and Fisher, Joshua B. and Chen, Jiquan and Xu, Jia and Zhang, Xiaotong and Yu, Ruiyang and Xie, Zijing and Ning, Jing and Fan, Jiahui and Zhang, Luna},
  title = {Cropland evapotranspiration based on Sentinel-2 shortwave infrared data and ensemble Kalman filter},
  journal = {Agricultural and Forest Meteorology},
  year = {2026},
  doi = {10.1016/j.agrformet.2026.111035},
  url = {https://doi.org/10.1016/j.agrformet.2026.111035}
}