| Popis: |
Forest fires in the Alps can have severe impacts on mountain forests reducing their protection capacity against rock falls and avalanches and increasing flood runoff, mud and debris flows (Moody and Martin 2001, Robichaud et al. 2007). Due to climate change, several studies have shown that the impact of forest fires in the Alpine environment will increase in the coming decades (Elkin et al. 2013, Lorz et al. 2010, Wastl et al. 2012). A key issue in modern forest fire management is the accurate mapping of forest fuels in order to determine spatial fire hazard, plan mitigation efforts, and active fire management (Krasnow et al. 2009). Several surface fuel description systems are currently used by land management agencies in the USA, Europe, Canada and Australia, and most of these systems have the same categories, components and description variables (Sandberg et al. 2001, Scott and Burgan 2005). A generalized description of fuels based upon average fuel properties is called a “fuel model”. A fuel model is a set of fuelbed inputs (e.g. load, bulk density, fuel particle size, heat content and moisture of extinction) used by a specific software for predicting the fire behaviour (e.g. FARSITE). Standard fuel models were developed in the USA by Anderson (1982) and more recently by Scott and Burgan (2005). Standard fuel models that fit the main local vegetation characteristics can be used as input for fire spread modelling and in combination with custom fuel models when available (Duguy et al. 2007, Arca et al. 2009, Jahdi et al. 2015). However, the Standard or custom fuel models have seldom been applied in the Alps. In this study we tested the possibility of defining some custom fuel models for the Eastern Italian Alps, which might allow a more reliable fire behaviour prediction when fire simulator systems are used. The custom fuel models definition was done by means of three steps: In the first step we studied local fire regime and fire behaviour and we tested the hypothesis that the decrease in burned area is related to an improvement in fire-fighting efficiency since the beginning of the 3rd millennium. In the second step fuel properties were measured in the field and analyzed. In the third step we made three fuel model sets based on three different approaches (Forest type association, Prometheus classification, Cluster classification). Then, using FARSITE (Finney 2004), we simulated ten fires that occurred in the Veneto Region from 2003 to 2013. Every fire was simulated using the three custom model sets and the Standard fuel models (Anderson 1982, Scott and Burgan 2005). Lastly, the fuel model set having the higher accuracy was adjusted in order to improve its performance in simulating real fire behaviour. In the Veneto Region, there was a decreasing number of fires per year from 1981 to 2004 and a much more evident decrease in the annual burned area. Fires in both mountain areas and the lowlands usually behave as surface fires and the burned area is seldom larger than ten hectares. We tested the hypothesis that the decrease in burned area is related to an improvement in fire-fighting efficiency since the beginning of the 3rd millennium. The power-law distribution of burned areas seems to confirm that suppression efficiency has been improved because the exponent of the power-law distribution was much higher in the last decade than in the previous two. In mountain areas fuel load paralleled what is reported in the literature for similar forests, but in the lowlands fuel load appeared much higher, probably because those forests are affected by phytosanitary problems that cause a higher amount of deadwood. We found significant differences in fuel load among vegetation types (chestnut, hop hornbeam forests, conifer plantations and shrubland), The most significant difference was litter load (p |
