FOREST ECOSYSTEM AREAS AFFECTED BY WILDFIRES: STATUS AND CHANGES

national, regional and protected natural areas levels were analyzed with standard statistical methods (average and linear regression), to evaluate both the existence of trends and the variability and general characteristics of the phenomenon at different spatial scales. 

From the analysis of the time series of forest surfaces affected by fire starting from 2018, 2021 emerges as an anomalous value (outlier) for the extension of impacted forest surfaces. For this reason, in the indicator presented here, 2021 is excluded from the statistical analyzes carried out for the trend analysis. From the examination of the series, the stochastic component of the phenomenon emerges, which conditions the analysis. We therefore attempted to first verify the significance of the results with regards to the trends. In cases where the presence of a significant trend was not found, the assessment of the state of the system was discussed in terms of comparison of the current value of the indicator with the average value in the period considered. 

The methodology used to build the indicator starting from 2018 is based on the evaluation of the surfaces affected by fire, distinguishing the different classes of forest ecosystems present. To obtain a quantitative evaluation of the burned forest covers, the polygons corresponding to the burned areas identified by EFFIS were considered together with the distributions of the forest categories obtained from the ECM-F4 forest ecosystem classification system created by ISPRA. These forest classes were defined following the EUNIS nomenclature (European Nature Information System - last update 2021) at level II and III: T1-deciduous broad-leaved forests (e. g. Oak and beech forests); T2-evergreen broad-leaved forests (e. g. Holm oaks and tall scrub); T3-coniferous forests (e. g. Fir forests, pine forests and plantations); T34-sub-alpine temperate forests (e. g. Larch forests). Forests not classified by the ECM-F4 model with the code TNC (class relating to non-classified forest areas) are also considered. The EFFIS polygons, together with the other spatialized datasets used, were reprojected into a single decimal metric coordinate system, in order to calculate statistics relating to the different themes. The annual values of the total surface area affected by fires in the various classes were calculated at a national and regional scale, in the protected natural areas of the EUAP network (Official List of Protected Areas) established by resolution of the State-Regions Conference of 17 December 2009 (National Parks; Regional Parks; State Reserves; Regional Reserves) and in the sites belonging to the Natura 2000 network: Special Conservation Areas (SAC) and Protection Zones Special (SPA)). Furthermore, the effects of forest fires were also considered in humid areas under the Ramsar Convention, and in biogeographical regions (Alpine, continental and Mediterranean). With specific geometric-spatial analysis tools on the QGIS platform it was possible to calculate the surface area of ​​each forest class mapped by the ECM-F4 product for the identified intersection areas. The spatial-statistical calculations were carried out with the aid of R, Python and QGis software.  For protected natural areas, estimates of burned surfaces were carried out separately within the individual typologies. In fact, as is known, there are overlaps in the founding boundaries due to different administrative acts underlying the definition of the EUAP network and the Natura 2000 network, with consequent planimetric intersections both within the individual categories of the networks and between the two systems. In this case the EFFIS polygons of the areas affected by fire were intersected ( overlay  thematic) with the polygons of the different themes considered and the union of all the areas belonging to the same typology was then evaluated in order to avoid counting the same surface twice.