Copenhagen Business School (CBS) and the Technical University of Denmark (DTU) establish a strategic collaboration on TREEADS project, in terms of research and innovation. Joint research efforts into preliminary findings and the project’s approach to risk analysis and integrated wildfire management provide a probabilistic framework for wildfire risks and their economic consequences.
Increasing extreme wildfires have significantly disrupted ecosystems, carbon storage and human settlements. A changing climate in recent years may play a substantial role in the increasing frequency, intensity, and duration of wildfires. Consequently, for preparedness and climate change risk analyses, indicator-based approaches that consider mainly environmental drivers are typically used as means of representing wildfire risks. This includes the broadly used Fire Weather Index (FWI), which is part of European Forest Fire Information System (EFFIS).
Understanding the TREEADS activity
Integrating several TREEADS innovations, Technical University of Denmark (DTU) supported by the Copenhagen Business School (CBS) have piloted a generic approach for eliciting quantitative assessments of wildfire risk. The ambition is to introduce an improved tool for underpinning economic risk assessment aligned with existing needs for cost-benefit analyses and climate change adaptation, including insurance applications.
TREEADS Pilot implementations involve three (3) main components:
- a stochastic wildfire ignition model considering human and environmental factors;
- fully probabilistic special fire hazard mapping based on a large ensemble of fire spread simulations that integrates our wildfire ignition model and is based on 30-year hydrometeorological data sets;
- economic risk assessment (mapping and sectoral) considering ecosystems, buildings, selected critical infrastructure and tourism.
CBS and DTU research methodology for local wildfire ignition modelling:
- logistic regression model;
- generalised additive model;
- machine learning model.
Copenhagen Business School and Technical University of Denmark considered all three forms, in order to derive the most generic stochastic model formulation in a pan-European perspective. A variety of explanatory variables related to human activities were tested, including nighttime lights (2012-2021), road density, and power line density.
The preliminary findings suggest that the marginal probability of fire ignition initially increases with the presence of economic activity. This probability reaches a peak and subsequently declines as economic activity intensifies, indicating more urbanised areas.
Moreover, at the stage of running a simulated logistic regression to estimate the probability of ignition, specifically from historical ignition point data in Crete (Greece), while a high-resolution FWI, calculated from ERA5 reanalysis, was used to represent the fire weather conditions. The simulation results indicated that an increased number of days with moderate fire danger within a year, correlates with a heightened probability of fire ignition at any point within a grid.
TREEADS demonstrates the utility of novel probabilistic and spatially explicit assessment of the wildfire risk under past, present and future conditions. The application of stochastic ignition models using a generic methodology, facilitate robust identification of critical ignition points and support local fire spread simulations for preparation and emergency management.
Overall, this fruitful collaboration aims to provide a methodology that will help practitioners and stakeholders across different pilot areas to better identify “risk hotspots”, and to estimate the current impact and economic risks related to wildfires using a more realistic methodology.
