We used simulation modelling to analyse spatial variation in wildfire exposure relative to key social and economic features on the island of Sardinia, Italy. Sardinia contains a high density of urban interfaces, recreational values and highly valued agricultural areas that are increasingly being threatened by severe wildfires. Historical fire data and wildfire simulations were used to estimate burn probabilities, flame length and fire size. We examined how these risk factors varied among and within highly valued features located on the island. Estimates of burn probability excluding non-burnable fuels, ranged from 0–1.92 × 10–3, with a mean value of 6.48 × 10–5. Spatial patterns in modelled outputs were strongly related to fuel loadings, although topographic and other influences were apparent. Wide variation was observed among the land parcels for all the key values, providing a quantitative approach to inform wildfire risk management activities.

In the last two decades, several models were developed to provide temporal and spatial variations of fire spread and behaviour. The most common models (i.e. BEHAVE and FARSITE) are based on Rothermel's original fire spread equation and describe fire spread and behaviour taking into account the influences of fuels, terrain and weather conditions. The use of FARSITE on areas different from those where the simulator was originally developed requires a local calibration to produce reliable results. This is particularly true for Mediterranean ecosystems, where plant communities are characterised by high specific and structural heterogeneity and complexity. To perform FARSITE calibration, an appropriate fuel model or the development of a specific custom fuel model is needed. In this study, FARSITE was employed to simulate three fire events in Mediterranean areas using different fuel models and meteorological input data, and the accuracy of results was analysed. A custom fuel model designed and developed for shrubland vegetation (maquis) provided realistic values of rate of spread, when compared with estimated values obtained using standard fuel models. Our results confirm that the use of both wind field data and appropriate custom fuel models are crucial to obtain reasonable simulations of wildfire events occurring on Mediterranean vegetation during the drought season.

We analyzed wildland fire occurrence and size data from Sardinia, Italy, and Corsica, France, to examine spatiotemporal patterns of fire occurrence in relation to weather, land use, anthropogenic features, and time of year. Fires on these islands are largely human caused and can be attributed to negligence, agro‐pastoral land use, and arson. Of particular interest was the predictive value of a fire weather index (FWI) that is widely used by fire managers to alert suppression crews. We found that an increase in the FWI from 30 to 60 produced on average an approximate eightfold increase in the odds of a large fire, regardless of the time of year during the fire season or land use type. Total area burned per fire season was positively correlated with the number of days with FWI > 40 over the period studied. Strong interactions between time of year and land type were also observed for both the probability of ignition and large fire. For example, the estimated odds of a large fire on agricultural lands in southern Sardinia was approximately 10 times larger than the forest and shrubland land type for areas close to roads, early (May) in the fire season. Conversely, toward the end of the fire season (September), we estimated the odds of a large fire in these same areas at about half the value estimated for the forest land classes. Of the explanatory variables analyzed, only FWI had an effect on the probability of a large fire (P < 0.1). The results of the study can be used in several ways including the following: (1) allocating fire detection and suppression resources to specific locations during the fire season; (2) prioritizing fuel breaks along specific road segments that have high predicted ignition rates; (3) refining the current fire danger indices; and (4) parameterizing wildfire simulation models to test how changing land use and climate change may affect spatial patterns in burn probability and intensity. Copyright © 2014 John Wiley & Sons, Ltd.