Modelling Fire Behavior to Assess Community Exposure in Europe: Combining Open Data and Geospatial Analysis
Abstract
Predicting where the next large-scale wildfire event will occur can help fire management agencies better prepare for taking preventive actions and improving suppression efficiency. Wildfire simulations can be useful in estimating the spread and behavior of potential future fires by several available algorithms. The uncertainty of ignition location and weather data influencing fire propagation requires a stochastic approach integrated with fire simulations. In addition, scarcity of required spatial data in different fire-prone European regions limits the creation of fire simulation outputs. In this study we provide a framework for processing and creating spatial layers and descriptive data from open-access international and national databases for use in Monte Carlo fire simulations with the Minimum Travel Time fire spread algorithm, targeted to assess cross-boundary wildfire propagation and community exposure for a large-scale case study area (Macedonia, Greece). We simulated over 300,000 fires, each independently modelled with constant weather conditions from a randomly chosen simulation scenario derived from historical weather data. Simulations generated fire perimeters and raster estimates of annual burn probability and conditional flame length. Results were used to estimate community exposure by intersecting simulated fire perimeters with community polygons. We found potential ignitions can grow large enough to reach communities across 27% of the study area and identified the top-50 most exposed communities and the sources of their exposure. The proposed framework can guide efforts in European regions to prioritize fuel management activities in order to reduce wildfire risk.
