R et al. (2026) Forest encroachment prediction using multi-temporal satellite data and machine learning: a case study of Bandipur National Park, India

Identification

• Journal: Scientific Reports
• Year: 2026
• Date: 2026-09-16
• Authors: Pushpa B. R, H. R. Chaitanya, Chandhana U. Shankar, R Sudarshan
• DOI: 10.1038/s41598-026-70597-0

Research Groups

• Department of Remote Sensing and GIS, Indian Institute of Technology (IIT) Roorkee
• Centre for Earth Sciences, Indian Institute of Science (IISc), Bangalore
• Bandipur National Park authorities

Short Summary

This study presents an artificial intelligence-based framework for forest encroachment prediction in Bandipur National Park, India, using multi-temporal Sentinel-2 satellite imagery integrated with topographic, land-cover, and anthropogenic variables. The proposed framework achieved high classification accuracy of 94.17% using the Random Forest model.

Objective

• To develop an AI-based framework for forest encroachment prediction in Bandipur National Park, India.
• To evaluate the performance of different machine learning models (Random Forest, Multi-Layer Perceptron, and CNN–LSTM) on multi-temporal Sentinel-2 satellite imagery integrated with topographic, land-cover, and anthropogenic variables.

Study Configuration

• Spatial Scale: Local scale (Bandipur National Park)
• Temporal Scale: Long-term (2016-2025)

Methodology and Data

• Models used:
  • Random Forest (RF)
  • Multi-Layer Perceptron (MLP)
  • Convolutional Neural Network–Long Short-Term Memory (CNN–LSTM)
• Data sources:
  • Sentinel-2 satellite imagery (2016-2025)
  • SRTM DEM
  • ESA WorldCover v200
  • Hansen Global Forest Change v1.8 dataset
  • OpenStreetMap road and settlement layers

Main Results

• The Random Forest model achieved the highest classification accuracy of 94.17% for forest encroachment prediction.
• The CNN–LSTM model showed relatively lower performance compared to RF and MLP models.

Contributions

• This study contributes to the development of an AI-based framework for forest encroachment prediction in Bandipur National Park, India.
• The proposed framework integrates multi-temporal Sentinel-2 satellite imagery with topographic, land-cover, and anthropogenic variables to improve classification accuracy.
• The results demonstrate the potential of machine learning models (RF and MLP) in predicting forest encroachment.

Funding

• This research was funded by the Department of Science and Technology (DST), Government of India (Grant No. DST/ICPS/RESOLV/2020).
• Additional support was provided by the Indian Institute of Science (IISc) and the Indian Institute of Technology (IIT) Roorkee.

Citation

@article{R2026Forest,
  author = {R, Pushpa B. and Chaitanya, H. R. and Shankar, Chandhana U. and Sudarshan, R},
  title = {Forest encroachment prediction using multi-temporal satellite data and machine learning: a case study of Bandipur National Park, India},
  journal = {Scientific Reports},
  year = {2026},
  doi = {10.1038/s41598-026-70597-0},
  url = {https://doi.org/10.1038/s41598-026-70597-0}
}