Bo et al. (2026) Spatially explicit estimation of high-resolution irrigation water use across China using earth observation data and deep learning

Identification

• Journal: ISPRS Journal of Photogrammetry and Remote Sensing
• Year: 2026
• Date: 2026-05-07
• Authors: Yong Bo, Xueke Li, Kai Liu, S K Wang, Long Li, Guoxu Li, Hui Li
• DOI: 10.1016/j.isprsjprs.2026.05.002

Research Groups

• State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, China.
• Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters (CICFEMD), Nanjing University of Information Science & Technology, Nanjing, China.
• University of Chinese Academy of Sciences, Beijing, China.
• Department of Earth and Environmental Science, University of Pennsylvania, Philadelphia, PA, USA.

Short Summary

The study developed a physically guided deep learning framework integrating Earth observation data and water balance modeling to estimate high-resolution (500 m) irrigation water use (IWU) across China from 2004 to 2019.

Objective

• To accurately estimate spatially explicit irrigation water use at a high spatial resolution to address water scarcity and promote sustainable agricultural water management in China.

Study Configuration

• Spatial Scale: National (China), with a spatial resolution of 500 m.
• Temporal Scale: 2004–2019.

Methodology and Data

• Models used: Physically guided deep learning framework and water balance modeling.
• Data sources: Multi-source Earth observation (EO) data (including evapotranspiration, precipitation, and surface soil moisture), remotely sensed land surface temperature (LST), and reanalysis-based skin temperature.

Main Results

• Validation Accuracy: The IWU estimates showed robustness with a root mean square error (RMSE) of 25.3 mm/yr at the station scale and 3.9 $\text{km}^3/\text{yr}$ at the national scale.
• Temporal Trends: National IWU peaked around 2013. A positive trend of 2.53 $\text{km}^3/\text{yr}$ was observed between 2004 and 2013, followed by a declining trend of -3.32 $\text{km}^3/\text{yr}$ from 2014 to 2019.
• Drivers of Change: The decline after 2013 is attributed to policy-driven improvements in irrigation conveyance efficiency and stringent water resource management regulations.

Contributions

• Developed a high-resolution (500 m) IWU dataset for China covering a 15-year period.
• Introduced a transferable, physically guided deep learning strategy that leverages water balance paradigms from non-irrigation periods to estimate water loss during irrigation periods.
• Enhanced the physical realism of irrigation monitoring by integrating multi-source EO data and surface cooling effects.

Funding

• Not specified in the provided text.

Citation

@article{Bo2026Spatially,
  author = {Bo, Yong and Li, Xueke and Liu, Kai and Wang, S K and Li, Long and Li, Guoxu and Li, Hui},
  title = {Spatially explicit estimation of high-resolution irrigation water use across China using earth observation data and deep learning},
  journal = {ISPRS Journal of Photogrammetry and Remote Sensing},
  year = {2026},
  doi = {10.1016/j.isprsjprs.2026.05.002},
  url = {https://doi.org/10.1016/j.isprsjprs.2026.05.002}
}