Li et al. (2026) Daily Lake-Surface NDVI Reconstruction Using Multi-Source Machine Learning Under Incomplete Optical Observations

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Remote Sensing
• Year: 2026
• Date: 2026-09-22
• Authors: Hongying Li, Qingyun Yan, Yuanjin Pan, Shuanggen Jin, Weimin Huang
• DOI: 10.3390/rs18193269

Research Groups

• Key Laboratory of Coastal Zone Ecosystem and Environment, China
• Department of Earth System Science, University of California, Irvine

Short Summary

This study presents a multi-source machine learning framework to reconstruct daily lake-surface normalized difference vegetation index (NDVI) under missing optical observations by integrating Cyclone Global Navigation Satellite System (CYGNSS) data, ERA5-Land meteorological variables, and geographic coordinates. The framework achieved an average test R2 of 0.63 and a root mean square error (RMSE) of 0.15 in GNSS-R-covered regions.

Objective

• Investigate the feasibility of using multi-source machine learning to reconstruct daily lake-surface NDVI under missing optical observations

Study Configuration

• Spatial Scale: Lake Taihu and Lake Chaohu, two representative eutrophic lakes in eastern China
• Temporal Scale: Daily time scale for NDVI reconstruction

Methodology and Data

• Models used: CatBoost (version 1.2.10) regression algorithm
• Data sources:
  • Cyclone Global Navigation Satellite System (CYGNSS) observations
  • ERA5-Land meteorological variables
  • Geographic coordinates
  • Fengyun-3F (FY-3F) NDVI

Main Results

• The full-feature model combining GNSS-R observables, meteorological variables, and geographic coordinates achieved an average test R2 of 0.63 and a root mean square error (RMSE) of 0.15 in GNSS-R-covered regions.
• The inclusion of GNSS-R observables provided additional predictive information compared to the model using meteorological variables and geographic coordinates alone.

Contributions

• This study provides a lake-specific empirical approach for maintaining spatially continuous daily NDVI information when optical observations are incomplete.
• The framework can be used as a supplementary tool for monitoring cyanobacterial blooms in eutrophic lakes.

Funding

• National Natural Science Foundation of China (Grant No. 42107019)
• China Postdoctoral Science Foundation (Grant No. 2021M693342)

Citation

@article{Li2026Daily,
  author = {Li, Hongying and Yan, Qingyun and Pan, Yuanjin and Jin, Shuanggen and Huang, Weimin},
  title = {Daily Lake-Surface NDVI Reconstruction Using Multi-Source Machine Learning Under Incomplete Optical Observations},
  journal = {Remote Sensing},
  year = {2026},
  doi = {10.3390/rs18193269},
  url = {https://doi.org/10.3390/rs18193269}
}