
Mapping fire risk in a Himalayan district
A Random Forest model built from 757 satellite-detected fire incidents and ten environmental, topographic and human-pressure variables classified 29.66% of Rasuwa's forest as high or very high fire risk. Held-out validation reached AUC 0.90 and TSS 0.67. The map is locally calibrated: static predictors, coarse inputs and the absence of ground-verified fire records limit spatial precision and transfer to other regions.
- Source
- Discover Forests
- Source language
- EN
- Source publication date
- 12 October 2026






