Sabaghy et al. (2026) Mapping Native Grass Cover with Random Forest Models: Sentinel-2 Versus Sentinel-2 Combined with Sentinel-1 SAR-Derived GLCM Texture Metrics

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

Identification

• Journal: Remote Sensing
• Year: 2026
• Date: 2026-09-13
• Authors: Sabah Sabaghy, Mohammad Abuzar, Steve J. Sinclair, Tony Dugdale, Vanessa Hutchins, Yogendra K. Karna, Jonathan Wilson, Kathryn Sheffield
• DOI: 10.3390/rs18183150

Research Groups

• School of BioSciences, University of Melbourne
• Australian Centre for Biodiversity Analysis, University of Melbourne

Short Summary

This study developed a scalable method using remote sensing data to map the fractional cover-class maps of native C3 and native C4 grasses in remnant native grasslands. The results show moderate accuracy for sparse to moderate grass cover.

Objective

• To develop an evidence-based management approach for grassland conservation by mapping and monitoring native grass species distribution.

Study Configuration

• Spatial Scale: Local scale, focusing on the western outskirts of Melbourne, Victoria, Australia.
• Temporal Scale: Cross-sectional study based on data collected in 2021.

Methodology and Data

• Models used: Random forest machine learning models.
• Data sources:
  • Sentinel-2 optical spectral bands and vegetation indices.
  • Sentinel-1 Synthetic Aperture Radar (SAR)-derived Grey Level Co-occurrence Matrix (GLCM) texture metrics.

Main Results

• The random forest model trained on Sentinel-2 data achieved moderate overall accuracy for native C3 (59.1%) and native C4 (78.1%) grass cover.
• Class-specific metrics showed that lower-cover classes were more reliable than higher-cover classes due to limited training samples.
• Model performance was not improved by the inclusion of Sentinel-1 SAR-derived GLCM texture metrics.

Contributions

• This study provides a scalable method for mapping and monitoring native C3 and C4 grass cover as a component of remnant native grasslands, enabling evidence-based management and biodiversity conservation.

Funding

• Australian Research Council (ARC) Discovery Project DP180100940.

Citation

@article{Sabaghy2026Mapping,
  author = {Sabaghy, Sabah and Abuzar, Mohammad and Sinclair, Steve J. and Dugdale, Tony and Hutchins, Vanessa and Karna, Yogendra K. and Wilson, Jonathan and Sheffield, Kathryn},
  title = {Mapping Native Grass Cover with Random Forest Models: Sentinel-2 Versus Sentinel-2 Combined with Sentinel-1 SAR-Derived GLCM Texture Metrics},
  journal = {Remote Sensing},
  year = {2026},
  doi = {10.3390/rs18183150},
  url = {https://doi.org/10.3390/rs18183150}
}