Mi et al. (2026) Mapping wide-area land subsidence from groundwater use in the North China plain by machine learning-based InSAR adjustment

Identification

• Journal: Remote Sensing of Environment
• Year: 2026
• Date: 2026-01-06
• Authors: Jiang Mi, Zhou Wu, X. Wang, Lin Bai, Zhiwei Li, Zhong Lu
• DOI: 10.1016/j.rse.2025.115226

Research Groups

• School of Geosciences and Info-Physics, Central South University, Changsha 410083, PR China
• School of Geospatial Engineering and Science, Sun Yat-sen University, Zhuhai 519082, PR China
• School of Geological Engineering and Geomatics, Chang'an University, Xi'an 710054, PR China
• School of Environment and Spatial Informatics, China University of Mining and Technology, Xuzhou, 221116, PR China

Short Summary

This study introduces a novel machine learning-based InSAR adjustment method to accurately map wide-area land subsidence across the North China Plain (NCP) from 2014 to 2022, revealing significant subsidence in central and coastal plains and quantifying associated groundwater depletion, with an observed alleviation after 2021.

Objective

• To propose a machine learning-based adjustment approach for routinely mapping wide-area land subsidence.
• To fully evaluate land subsidence and associated groundwater depletion across the entire North China Plain from the end of 2014 to 2022, assessing the long-term impact of water resource management policies.

Study Configuration

• Spatial Scale: North China Plain (NCP), covering approximately 56,882 km² experiencing subsidence.
• Temporal Scale: End of 2014 to 2022 (approximately 8 years).

Methodology and Data

• Models used: Machine learning-based adjustment approach (adaptive model selection for each SAR frame), plate motion model.
• Data sources: Synthetic Aperture Radar (SAR) interferometry (InSAR) data (Sentinel-1), Global Navigation Satellite System (GNSS) data.

Main Results

• The proposed method improved subsidence velocity accuracy from 3.8-17.5 mm/yr to 2.0 mm/yr, validated by independent GNSS data.
• Approximately 56,882 km² of the NCP experienced land subsidence greater than 20 mm/yr.
• The central alluvial and coastal plains were identified as primary subsidence areas, with a maximum cumulative subsidence reaching 2 m.
• The average subsidence velocity peaked in 2018 at 38.5 mm/yr.
• Land subsidence has shown alleviation after 2021.
• Total groundwater loss from the confined aquifer in the NCP was estimated at 24.9 billion m³ between the end of 2014 and 2022.
• Of this total, 20.2 billion m³ (81%) occurred from October 2014 to the end of 2020, while 4.7 billion m³ (19%) occurred from January 2021 to December 2022, indicating a reduced rate of loss.

Contributions

• Develops a novel machine learning-based InSAR adjustment approach that adaptively selects the optimal model for each SAR frame, effectively minimizing varying long-wavelength errors for improved wide-area subsidence mapping.
• Integrates GNSS data and a plate motion model to mitigate incidence angle-related InSAR measurement differences in overlap regions between consecutive tracks.
• Provides the first comprehensive, wide-area, and long-term evaluation of land subsidence and associated groundwater depletion across the entire North China Plain, offering critical insights into the effectiveness of water resource management policies.
• Offers new evidence supporting China's groundwater management practices in addressing land subsidence in the NCP.

Funding

• Not specified in the provided text.

Citation

@article{Mi2026Mapping,
  author = {Mi, Jiang and Wu, Zhou and Wang, X. and Bai, Lin and Li, Zhiwei and Lu, Zhong},
  title = {Mapping wide-area land subsidence from groundwater use in the North China plain by machine learning-based InSAR adjustment},
  journal = {Remote Sensing of Environment},
  year = {2026},
  doi = {10.1016/j.rse.2025.115226},
  url = {https://doi.org/10.1016/j.rse.2025.115226}
}