Wang et al. (2025) Study on the Spatiotemporal Variation of Vegetation Characteristics in the Three River Source Region Based on the CatBoost Model

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

Identification

• Journal: Remote Sensing
• Year: 2025
• Date: 2025-12-13
• Authors: Jun Wang, Siqiong Luo, Hongrui Ren, Xufeng Wang, Jingyuan Wang, Zisheng Zhao
• DOI: 10.3390/rs17244024

Research Groups

Not specified in the provided text.

Short Summary

This study developed a machine learning-based framework to integrate MODIS and GIMMS NDVI data, reconstructing a 1 km monthly NDVI dataset for the Three River Source Region (TRSR) from 1982 to 2014, which revealed an overall increasing trend in vegetation greenness with significant spatial heterogeneity.

Objective

• To propose and validate a machine learning-based downscaling framework for integrating MODIS and GIMMS NDVI data to reconstruct a long-term, high-resolution NDVI dataset.
• To systematically analyze the spatiotemporal variation characteristics of vegetation in the Three River Source Region (TRSR) from 1982 to 2014 using the reconstructed dataset.

Study Configuration

• Spatial Scale: Three River Source Region (TRSR), with a spatial resolution of 1 km.
• Temporal Scale: Monthly data from 1982 to 2014.

Methodology and Data

• Models used: Random Forest (RF), LightGBM, and CatBoost. CatBoost was identified as the optimal algorithm for spatiotemporal data fusion (R² = 0.9014, RMSE = 0.0674, MAE = 0.0445).
• Data sources: Moderate-resolution Imaging Spectroradiometer (MODIS) Normalized Difference Vegetation Index (NDVI) and Global Inventory Monitoring and Modeling System (GIMMS) NDVI data.

Main Results

• The constructed long-series 1 km monthly GIMMS-MODIS NDVI dataset for 1982-2014 showed high consistency with MODIS NDVI data.
• From 1982 to 2014, the NDVI in the TRSR exhibited a significant increasing trend, with an average growth rate of 0.0020 per decade (p < 0.05).
• NDVI showed obvious spatial heterogeneity, characterized by a decreasing gradient from southeast to northwest.
• The Yellow River source exhibited the most evident vegetation recovery, the Yangtze River Source area showed moderate improvement, whereas the Lancang River Source area displayed little noticeable change.
• Broad-leaved forests experienced the most significant growth, while cultivated vegetation displayed a marked tendency toward degradation.

Contributions

• Development and validation of a novel machine learning-based downscaling framework for spatiotemporal NDVI data fusion, outperforming conventional statistical approaches.
• Generation of a high-accuracy, long-term (1982-2014), high-resolution (1 km monthly) NDVI product specifically for the Three River Source Region (TRSR).
• Provides a robust methodological foundation for advancing vegetation dynamics research in other high-altitude regions facing similar data limitations.

Funding

Not specified in the provided text.

Citation

@article{Wang2025Study,
  author = {Wang, Jun and Luo, Siqiong and Ren, Hongrui and Wang, Xufeng and Wang, Jingyuan and Zhao, Zisheng},
  title = {Study on the Spatiotemporal Variation of Vegetation Characteristics in the Three River Source Region Based on the CatBoost Model},
  journal = {Remote Sensing},
  year = {2025},
  doi = {10.3390/rs17244024},
  url = {https://doi.org/10.3390/rs17244024}
}