Advanced technologies for forest fire risk assessment and propagation modelling

Abstract

Wildfire management increasingly relies on the integration of advanced technologies for accurate risk assessment and predictive fire spread modeling. Considering the pronounced trend of rising wildfire frequency and intensity, particularly in climate-sensitive Mediterranean regions, timely and precise risk evaluation is crucial for enhancing preventive capacities, optimizing the allocation of operational resources in the field, and protecting critical infrastructure, natural assets, and human lives. Contemporary methodologies enable early detection of fire potential and support decision-making across all phases of fire management from preparedness and initial response to tactical coordination during suppression and post-fire recovery. Special emphasis is placed on the use of  computational models and geospatial tools that facilitate dynamic, real-time assessment of fire behavior, thereby significantly improving operational responsiveness. Fire spread modeling software, such as BehavePlus, FlamMap, and FARSITE, plays a vital role in this process by simulating how fires propagate across complex terrain under varying weather and fuel conditions. These tools support strategic planning, real-time decision-making, and emergency interventions, making them indispensable in wildfire preparedness and management. Fire modeling relies on a combination of static and dynamic geospatial data. Static data such as elevation, slope, and aspect are usually obtained from Digital Elevation Models (DEMs) and remain constant over time. Dynamic data, including wind speed and direction, air humidity, and temperature, are sourced from meteorological services or predictive models like WRF, although such data may sometimes lack high spatial resolution at the local scale. One of the most complex and influential factors is vegetation structure and condition, including canopy type, fuel load, and moisture content. These parameters can be assessed using remote sensing technologies: multispectral satellite imagery (e.g., Sentinel-2 from the Copernicus program) enables the calculation of vegetation indices such as NDVI, while unmanned aerial vehicles (UAVs) equipped with RGB or LiDAR sensors provide high-resolution data on canopy height and density. To estimate the amount and condition of fuel, experts typically combine Canopy Height Models (CHMs), derived from LiDAR or photogrammetry, with vegetation activity data (e.g., NDVI) to model fuel volume and availability. Fuel bulk density values, defined according to vegetation type, are used to convert volumetric estimates into biomass mass, resulting in high-resolution fuel distribution maps expressed in kg/m2 or t/ha. These maps represent a fundamental tool for quantifying fire potential, and their spatial detail and tactical precision are essential for realistic fire spread simulations and effective operational fire planning. Such maps can be further enhanced by incorporating fuel moisture content, either based on field measurements or using moisture indices derived from remote sensing, providing a solid foundation for realistic fire behavior simulations. This comprehensive data processing and modeling system from terrain and weather analysis to fuel assessment based on remote sensing and fire behavior simulation constitutes the central theme of this paper.