Validering av Prometheus för modellering av brandspridning och effekten av brandgator i Sverige : En fallstudie från Ljusdalsbränderna 2018

Wildfires are an increasing concern in Sweden due to climate change-induced droughts and growing human activity, with the 2018 Ljusdal fires marking the most severe wildfires in recent Swedish history. The scale and cost of suppression efforts raised interest in passively reducing fire spread through fuel-treatment methods such as firebreaks. This study examines how these firebreaks could affect fire spread in boreal forests through comparing the 2018 fire spread without and with this intervention. Because of uncertainty in reclassifying Swedish vegetation to fuel types compatible with Prometheus software, six fuel maps were developed, differing primarily in how wetlands, roads, and clearcuts were treated. Simulations were run using 2018 conditions, and the predicted fire perimeters were validated against Copernicus satellite data using the Jaccard index. While no method perfectly matched the observed fire, a custom fuel map (Scenario 6) achieved the highest Jaccard score (0.29), suggesting an improved accuracy when adapting literature-based logic to local conditions. In a second phase, existing roads were widened in the model to serve as 40 m and 90 m firebreaks. Both widths substantially reduced fire spread, but unexpectedly, 40 m firebreaks performed equally well or better than 90 m ones. Sensitivity analyses suggest that placement and rerouting dynamics may be more influential than width alone. The results highlight the importance of locally adapted fuel classifications and show Prometheus’ potential for scenario-based fire mitigation, if combined with ecological and contextual nuance.