Fuel types and potential fire behavior in Sardinia and Corsica islands: a pilot study

Wildland fires represent a serious threat to forests and wooded areas of Mediterranean Basin. Regarding the last ten years, Spain, Portugal, Italy, Greece and France recorded an annual average of about 50,000 forest fires and about 470,000 burned hectares (European Communities, 2009). Cover, type, humidity status, and biomass and necromass load of vegetation are critical variables in affecting wildland fire occurrence. In particular, fuel physical characteristics such as loading (weight per unit area), size (particle diameter), and bulk density (weight per unit volume) of the live and dead biomass contribute to the spread, intensity, and severity of wildland fire. So that, the availability of accurate fuel data at different spatial and temporal scales is essential for fire management applications, from fire behavior prediction to fire effects simulation to ecosystem simulation modeling. One of the goals of Proterina-C project is to evaluate the fire danger in Mediterranean areas and characterize the vegetation parameters involved in the combustion process. In this context, the objectives of this work are i) to identify and describe the different fuel types mainly affected by fire occurrence in Sardinia and Corsica Islands and ii) clusterize the selected fuel types in function of their potential fire behavior. In the first part of the work, the available time series of fire event perimeters and the land use map data were crossed and analysed in order to identify the main land use types affected by fires. Field sampling sites were then randomly identified on selected vegetation types and the following variables were collected: live and dead fuel load, depth of the fuel layer, plant cover. Dead and live fuel load were inventoried following the standardized classes (1h, 10h, 100h) of the USDA National Fire Danger Rating System. In the second part of the work the potential fire behavior for every experimental site was then calculated by BEHAVE fire behavior prediction system (Andrews, 1989), using as input data the collected fuel variables. Fire behavior was simulated by setting different weather scenarios representing the most frequent summer meteorological conditions. The simulation outputs (fireline intensity, rate of spread, flame length) were then used to perform a cluster analysis in order to group the different fuel types based on their potential fire behavior. The results of this analysis can be used to produce fire behavior fuel maps that are important tools in locating and rating fuel treatments, evaluating fire hazard and risk for land management planning, and aiding in environmental assessments and fire danger programs modeling.