AFFRI-Austrian Forest Fire Research Initiative

Austrian forests do not fulfil the characteristics of fire prone ecosystems, nor were they heavily fire-impacted so far. Due to the debate on likely climate change it is hypothesized, that the risk of forest fires as potential disturbance agent will increase in coming years and decades here too. In this context the study design of this Austrian Forest Fire Research Initiative (AFFRI) encompasses two major objectives (i) to identify Forest Fire "hot spots" in Austria in dependence of vegetation, climate and location, and (ii) to develop a fire-vegetation simulator for Austrian conditions. It is planned to find out where and why there exist potential forest fire "hot spots" in Austria, and if there is an increase of the probability for expansion under the consideration of the effects of global warming. This includes the consideration of fire weather options and topo-climatological aspects, as well as the classification of forest fuels for both, the more fire-sensitive forest types and for more fire-endangered railway sections, especially in the Alps and its foothills. In order to describe forest behaviour we will do a comparative assessment of two contrasting fire modelling approaches. The hybrid 3D-patch model PICUS v1.41, which has been particularly developed for Austrian forest conditions will be extended by a fire risk and fire spread module and compared with Fire-BGC, a model successfully tested for North-American conditions. We aim at capturing the driving factors for fire behaviour in mountain forests in order to evaluate the fire-vegetation simulator with selected fire cases. In that context a scenario analysis will be done to explore the interaction of management and a changing climate according to (i) the effects of different management regimes on fire behaviour and (ii) the effects of climate change scenarios on fire behaviour.

In order to characterise recent forest fires in Austria, a wildfire database has been established for the period from 1993 to 2012 with in total 3700 fires. The date and location of the fire, the size of the burned area, fire causes as well as number of fire brigades involved were recorded. Spring and summer were the main fire seasons for forest fires in Austria. Major fire hotspots were identified in the South of Lower Austria, in Carinthia and Tyrol. In total, anthropogenic causes made up for the major part of forest fires in Austria (85%), lightning-caused forest fires have had a share of 15% (making up to 40% in the summer months). Additionally the relationship between socio-economic factors and the spatial occurrence of forest fires has been studied. It was found that railroad, forest road and hiking trail density as well as agricultural and forestry developments may contribute significantly to fire ignition. To predict wildfire occurrence in different eco-regions with regard to fire weather conditions an appropriate Index was selected. Based on a comparison of the index values of fire days and non-fire days it was shown that in the summer season the Canadian Build-up index, the Keetch Byram Drought Index as well as the mean daily temperature have the best performance. In the winter season the German M68dwd is the best performing index. It was demonstrated that the non-parametric methods are robust to differences in index value frequency distribution and may allow more valid comparisons of fire danger indices. Forest types were classified according to their characteristics in fire behaviour. Based on field sampling and fuel analysis, a fuel map for Austria and fuel models for the more fire-prone pine forests have been developed. The average fuel load for the different classes ranged between 1,1 and 9,1 t/ha. The fuel models were used as input for fire behaviour simulations which allowed estimating rate of spread and fire intensities. The application of physiological modelling to wildfire hazard analyses has highlighted the importance of biophysical factors in forest fire ignition. Under extreme summer conditions, fire hazard is more closely associated with long term dry periods, which reduce the moisture content of heavier fuels. Spring fires are associated with a high mass of labile carbon in the forest litter, as this collects over the winter. Short periods of very dry air, might rapidly dry surface fuels. With both modelling approaches PICUS and BGC it was shown that the effects of different fire severities can be monitored. As fire intensity and fire behaviour are highly dependent from stand structure, it was shown that models that represent that structure have benefits. Model runs with BIOME-BGC applied to Austria over the past half-century show for the summer season a reduction in days of low fire danger in general, and an increase in extreme fire days in most regions. It was possible to confirm this trend under climate change conditions using a localized and bias corrected regional climate model. The results indicate that the number of extreme fire days will increase in the eastern parts of Austria in general and at lower altitudes in alpine valleys.