Assessing the impact of wind field on fire spread and behaviour under climate change scenarios

Fire prevention and fire management require knowledge and evaluation of the combined effects of different factors affecting the fire environment. In the absence of extreme fire behavior (torching, spotting, crowning, and fire whirls), the effect of fire environment on fire spread and behavior can be adequately simulated using different models, mainly based on semi-physical approaches. In addition, the effects on fire behavior can be integrated at various scales using spatially and temporally explicit fire spread and behavior simulators, e.g. the FARSITE fire area simulator. Criticisms of fire simulators frequently concern the need of high resolution environmental data, in particular data of environmental variables that quickly change in space and time, e.g. the wind speed and direction. Consequently, drawbacks arise when we attempt to evaluate the regional and local impact of the expected climatic changes on fire behavior and fire severity. Applications of downscaling techniques are needed to generate probability distributions of wind speed and directions. The aim of this work is (1) to evaluate the accuracy of wind field maps obtained by dynamical downscaling of a global climate model, (2) to compare the wind field maps with wind data obtained by computational fluid dynamics models, and (3) to analyze the sensitivity of fire simulators to the inaccuracy of wind data. The results show differences in fire spread and fire severity when different approaches are used. The study also showed the capabilities of wind field maps from computational fluid dynamic models to evaluate the potential impact of climate changes on fire behavior and risk in Mediterranean areas.