RandomFront 2.3: a physical parameterisation of fire spotting for operational fire spread models – implementation in WRF-SFIRE and response analysis with LSFire+
- 1BCAM–Basque Center for Applied Mathematics, Bilbao, Basque Country, Spain
- 2Department of Mathematics, University of the Basque Country UPV/EHU, Bilbao, Basque Country, Spain
- 3Space Research Institute of Russian Academy of Sciences, Moscow, Russia
- 4Institute of Geography, University of Bremen, Bremen, Germany
- 5Department of Atmospheric Chemistry, Max Planck Institute for Chemistry, Mainz, Germany
- 6Ikerbasque–Basque Foundation for Science, Bilbao, Basque Country, Spain
Abstract. Fire spotting is often responsible for dangerous flare-ups in wildfires and causes secondary ignitions isolated from the primary fire zone, which lead to perilous situations. The main aim of the present research is to provide a versatile probabilistic model for fire spotting that is suitable for implementation as a post-processing scheme at each time step in any of the existing operational large-scale wildfire propagation models, without calling for any major changes in the original framework. In particular, a complete physical parameterisation of fire spotting is presented and the corresponding updated model RandomFront 2.3 is implemented in a coupled fire–atmosphere model: WRF-SFIRE. A test case is simulated and discussed. Moreover, the results from different simulations with a simple model based on the level set method, namely LSFire+, highlight the response of the parameterisation to varying fire intensities, wind conditions and different firebrand radii. The contribution of the firebrands to increasing the fire perimeter varies according to different concurrent conditions, and the simulations show results in agreement with the physical processes. Among the many rigorous approaches available in the literature to model firebrand transport and distribution, the approach presented here proves to be simple yet versatile for application to operational large-scale fire spread models.