GIS and Remote Sensing-Based Forest Fire Risk Assessment Using a Multi-Criteria Fire Risk Index in the Potohar Region, Pakistan

Abstract

Forest fires are becoming increasingly frequent and severe worldwide due to climate change, prolonged dry periods, and expanding human activities, posing substantial threats to forest ecosystems, biodiversity, and socio-economic resources. In Pakistan, the Potohar region, particularly the Margalla Hills National Park, has experienced recurrent wildfire incidents that highlight the need for reliable fire-risk assessment tools. This study presents a Geographic Information System (GIS) and remote sensing-based approach for assessing forest fire risk by integrating multiple environmental and anthropogenic factors into a Fire Risk Index (FRI). Seven variables, including vegetation condition derived from the Normalized Difference Vegetation Index (NDVI), land surface temperature (LST), slope, aspect, elevation, distance from roads, and burned area identified using the Normalized Burn Ratio (NBR), were extracted from Landsat 8/9 imagery, ASTER Digital Elevation Model (DEM), and ancillary spatial datasets. A weighted multi-criteria overlay analysis was performed in ArcGIS to generate a spatially explicit forest fire risk map of the Potohar region. The resulting map identified the Margalla Hills and surrounding forested landscapes as the most fire-prone areas during the pre-monsoon season, primarily due to elevated temperatures, dense vegetation, steep terrain, and proximity to human activities. Burn severity analysis revealed approximately 8.09 km² of forest area affected during the June 2022 wildfire event, while land surface temperature increased by nearly 8°C in the burned region following the fire. The predictive performance of the proposed model was evaluated using Kernel Density Estimation (KDE), which demonstrated good agreement between historical fire occurrences and high-risk zones. The developed framework provides an efficient and cost-effective decision-support tool for identifying vulnerable areas, improving wildfire preparedness, prioritizing resource allocation, and supporting sustainable forest management in Pakistan. The methodology is transferable to other fire-prone landscapes with similar ecological and climatic characteristics.