Descrizione 1
Emiliano Agrillo, Federico Filipponi, Roberto Inghilesi, Alessandro Mercatini, Alice Pezzarossa, Nazario Tartaglione
The indicator analyzes the extent of forest areas affected by wildfires at the national, regional, and protected area scales between 2018 and 2023. The distribution and the surface of burned forested areas show significant annual variability. The 2023 ranks as the worst year after 2021, which was an exceptional year in terms of wildfire-affected forest ecosystems within the analyzed time series.
In 2023, the most affected regions were Sicily (10,080 hectares) and Calabria (2,987 hectares), which together accounted for 83% of the total forest nationwide area affected by wildfires. The most impacted forest classes are the evergreen forests (holm oak forests and Mediterranean scrub), followed to a much lesser extent by deciduous mixed forests (oaks and beech forests) and coniferous forests (pine, fir and larch forests).
Although the time series covers a short period, a linear regression analysis of the national forest cover affected by wildfires indicates an increasing trend. A statistically significant increasing trend has also been calculated for the regions of Sicily, Calabria, Apulia, and Aosta Valley over the considered period. A positive trend has also been observed for Regional Nature Reserves and those included in the Special Protection Areas (SPAs) and Special Areas of Conservation (SACs) within the Natura 2000 Network.
The indicator (related to status and changes) represents the extent of forest ecosystems affected annually by large wildfires since January 1, 2018. Data are categorized by forest ecosystems type and level of aggregation (national, regional, and protected natural areas).
The indicator is based on data provided by the European Forest Fire Information System (EFFIS), which, since 2018, has been able to detect wildfires as small as 1 hectare thanks to the new generation of MSI multispectral sensors onboard the Sentinel-2 satellite. The system achieves an accuracy of approximately 95% in identifying burned areas across Europe.
To evaluate the impact of wildfires on various forest ecosystems, the burned area data from EFFIS is overlaid onto the forest ecosystem map from ISPRA using the ECM-F4 machine-learning model. The forest classifications refer to the EUNIS nomenclature (updated in 2021) at Levels II and III:
- T1– Deciduous broadleaf forests (oaks and beech forests)
- T2 – Evergreen broadleaf mixed forests (holm oak forests and Mediterranean scrub)
- T3 – Coniferous forests (fir and pine forests)
- T34 - Temperate subalpine forests (larch forests)
The unclassified forest ecosystems from the ECM-F4 model are also considered and labeled with TNC (Trees Not Classified).
The aim is toTo provide useful information for assessing the status of forest ecosystems affected by large wildfires, based on the annual extent of burned areas.
Law No. 353 of November 21, 2000 (Framework Law on Wildfires).
Law No. 155 of November 8, 2021, whose first article introduces "Urgent measures for strengthening coordination, technological updates, and enhancing operational capacity in forecasting, prevention, and active firefighting against wildfires.
Law No. 132 of June 28, 2016, establishing the National Network System for Environmental Protection and regulating the Institute for Environmental Protection and Research (see Article 11).
MASE Directive (UDCM prot. No. 67, February 22, 2024) concerning the functions and duties assigned to the Institute for Environmental Protection and Research for the 2024-2026 three-year period, Article 2 "priority action lines," Eleventh paragraph "Geographic Information Systems and Cartography for Environmental Protection." Directive link.
SNPA Council Resolution. No. 108/21 of May 18, 202, on the (Doc. - Operational framework, areas of intervention, and specialized competencies for environmental emergencies.
State-Regions Conference Resolution of December 17, 2009, link.
Descrizione 2
Agrillo E, Filipponi F, Pezzarossa A, Casella L, Smiraglia D, Orasi A, Attorre F, Taramelli A. (2021). Earth Observation and Biodiversity Big Data for Forest Habitat Types Classification and Mapping. Remote Sensing. 2021; 13(7):123.
E. Agrillo, F. Filipponi, R. Inghilesi, A. Mercatini, and A. Pezzarossa. Monitoraggio dei cambiamenti degli ecosistemi forestali in Italia. Technical report, Istituto Superiore per la Protezione e la Ricerca Ambientale, 2022.
Barbosa, P., Kucera, J., Strobl, P., Vogt, P., Camia, A., & San-Miguel-Ayanz, J. S. (2006). European Forest Fire Information System (EFFIS) Rapid damage assessment: Appraisal of burnt area maps in southern Europe using MODIS data (2003-2005). Forest Ecology and Management, 234(1), S218.
Barredo, J. I., Brailescu, C., Teller, A., Sabatini, F. M., Mauri, A., & Janouskova, K. (2021). Mapping and assessment of primary and old-growth forests in Europe. Publications Office of the European Union, Luxembourg.
Blasi, C., Burrascano, S., Maturani, A., & Sabatini, F. M. (2010). Foreste vetuste in Italia. Ministero dell’Ambiente e della Tutela del Territorio e del Mare: Roma, Italy.
Blasi, C., Bovio, G., Corona, P. M., Marchetti, M., & Maturani, A. (2004). Incendi e complessità ecosistemica. Dalla pianificazione forestale al recupero ambientale. https://www.mite.gov.it/sites/default/files/archivio/allegati/aib/incen…
Cervellini, M., Zannini, P., Di Musciano, M., Fattorini, S., Jiménez-Alfaro, B., Rocchini, D., ... & Chiarucci, A. (2020). A grid-based map for the Biogeographical Regions of Europe. Biodiversity Data Journal, 8.
Chytrý, M. & co-authors (2020). EUNIS Habitat Classification: Expert system, characteristic species combinations and distribution maps of European habitats. Applied Vegetation Science, volume 23, issue 4, pp 648-675
European Commission, Joint Research Centre, San-Miguel-Ayanz, J., Durrant, T., Boca, R. et al., Advance report on forest fires in Europe, Middle East and North Africa 2022, Publications Office of the European Union, 2023, https://data.europa.eu/doi/10.2760/091540
Schaminée, J.H.J., Chytrý, M., Hennekens, S.M., Janssen, J.A.M., Knollová, I., Marceno, C., Tichý, L. & data providers (2020). Support to EUNIS habitat classification revision IV. Report for the EEA - Framework Contract: EEA/NSS/17/002/Lot 1, specific contract No 3417/B2019/EEA6
Inventario Nazionale delle Foreste e dei serbatoi forestali di Carbonio, 2015
https://www.mase.gov.it/pagina/elenco-ufficiale-delle-aree-naturali-protette-0
https://www.mase.gov.it/pagina/rete-natura-2000
https://ec.europa.eu/environment/nature/natura2000/biogeog_regions/
https://www.isprambiente.gov.it/it/attivita/Crisi-Emergenze-ambientali-e-Danno/centro-operativo-per-la-sorveglianza-ambientale/ecosistemi-ed-incendi-boschivi-in-italia
The data employed in the following indicator may not fully align with other national databases specifically designed for storing wildfires data, such as the Comando Carabinieri Tutela Forestale - Nucleo Informativo Antincendio Boschivo, especially concerning small-scale fires. Furthermore, the EFFIS (European Forest Fire Information System) system, which serves as the data source for wildfire areas, has been shown to successfully identify 95% of the areas burned annually across Europe (https://effis.jrc.ec.europa.eu/about-effis/technical-background/rapid-damage-assessment). This accuracy is due to the high-resolution data provided by Sentinel-2 MSI (used since 2018).
Under certain conditions, satellite-derived products may have data omissions in wildfire detection. These omissions can result from factors such as cloud cover, fire size and duration, wildfire type (crown or undergrowth fire), lack of synchronization between the fire event and the satellite’s orbital pass, and the physiognomic characteristics of the affected ecosystem.
Despite these limitations, the primary strength of the EFFIS product lies in its ability to rapidly provide data for analysis, an essential capability for assessing urgently wildfire impacts.
For a complete assessment of the impact of wildfires on the health of the affected forest ecosystems, it would be useful to obtain data on forest biomass loss and fire severity of each event.
To improve information on affected national forest and woodland cover, an EUNIS legend will be applied in the future, at greater thematic detail, in all categories in the indicator to date.
To improve information regarding the health of forests affected by fire, in the future given the availability of new products provided by the European Earth Monitoring System from satellites (Copernicus LMS), it will be possible to define the biomass affected for each individual fire event.
Qualificazione dati
Data on burned areas are freely available for download at the link: https://effis.jrc.ec.europa.eu/applications/data-and-services.
Data on the national distribution of burned areas related to major forest categories are available at the link: https://groupware.sinanet.isprambiente.it/prodotti-operativi-di-sorveglianza-ambientale/library/disturbance-agents/burnt-areas-italian-terrestrial-ecosystem.
Data on the national distribution of forest categories (Ecosystems Classification Model Forest4 ECM-F4) are available at the link: https://groupware.sinanet.isprambiente.it/prodotti-operativi-di-sorveglianza-ambientale/library/ecosystems-classification-model
EUNIS version 2021 legend data are available at the link:
https://www.eea.europa.eu/data-and-maps/data/eunis-habitat-classification-1
Data on spatial administrative units can be accessed from the Istat portal at the link:
https://www.istat.it/it/archivio/222527
Data on protected areas are accessible from the EEA portal at the links:
https://www.eea.europa.eu/data-and-maps/data/nationally-designated-areas-national-cdda-17 https://www.eea.europa.eu/data-and-maps/data/natura-13
Data on Ramsar areas can be accessed from the national geoportal at the link:
http://wms.pcn.minambiente.it/ogc?map=/ms_ogc/wfs/RAMSAR.map
Data on biogeographic regions can be accessed at the link:
https://zenodo.org/record/3934970#.Y00m8y8QM6l
Data on old-growth forests can be accessed at the link: https://publications.
National, Regional
2018-2023
Qualificazione indicatore
Standard statistical methods (mean and linear regression) were used to look at the values of annual time series for the total forest areas affected by large fires at the national, regional, and protected area levels. This analysis was carried out in order to identify possible trends and to observe the evolution and behaviour of the phenomenon in various regions and at different spatial scales.
An analysis of the ISTAT time series on total forest area burned since 2000 highlights 2007, 2017, and 2021 as outlier years due to the extent of the affected areas. Given that the ratio between total area and burned forest area remains relatively stable, 2021 was excluded from the trend analysis in this indicator.
The time series analysis also reveals the stochastic nature of the phenomenon, which significantly influences trend patterns. To ensure robust statistical interpretation, efforts were made to assess the significance of the identified trends. In cases where no significant trend was detected, the evaluation of the system's status was evaluated by comparing the current value of the indicator to the average value over the period analyzed.
The methodology used to construct the indicator starting from 2018 is based on the assessment of the areas affected by wildfire by distinguishing the different types of forest ecosystems. For a quantitative assessment of the burned forest surface, the Forest Ecosystem Classification Model—Forest4 (ECM-F4) created by ISPRA was used to overlay the polygons that of the burned areas by EFFIS onto the distributions of EUNIS forest types. These forest types were defined following the EUNIS (European Nature Information System—last update 2021) nomenclature at Levels II and III: T1-deciduous broadleaf mixed forests (oaks and beech forests); T2-evergreen broadleaf forests (oak and Mediterranean scrub); T3-conifer forests (fir and pine forests); T34-temperate subalpine forests (larch forests). Forests not classified by the ECM-F4 model are also considered, designated under the TNC code, which represents the category for non-classified forest ecosystems The EFFIS polygons and , together with other spatial datasets used, were projected onto a single decimal metric coordinate system to calculate a zonal statistics on the different thematic areas. The annual values of the total surface area affected by wildfire in the different forest types were calculated at the national and regional scales, in the protected natural areas of the EUAP network (Official List of Protected Areas) established by the State-Regions Conference Resolution of December 17, 2009 (National Parks; Regional Parks; State Reserves; Regional Reserves) and in the sites belonging to the Natura 2000 network: Special Areas of Conservation (SAC) and Special Protection Areas (SPA). Additionally, the effects of wildfires were also considered in the wetlands site of Ramsar Convention, in the national old-growth forests network, and in the biogeographical regions (Alpine, Continental, and Mediterranean).
Using spatial analysis tools, the intersected polygons of each forest type classified by ECM-F4 zones was calculated. Zonal statistics were processed using R, and QGIS. Within each category of protected natural areas, burned areas were estimated separately, accounting the overlaps between the EUAP and Natura 2000 network boundaries.
To ensure accurate calculations, EFFIS wildfire areas were overlaid with thematic polygons (e.g., regions, protected areas) using a thematic overlay approach. The total burned area was then derived by summing the affected areas within each category. This method prevented double-counting of wildfire-impacted areas, ensuring precise estimation of the total burned surface.
In 2023, 15,698 ha of forest area were burned at the national scale, a value far exceeding the average of the considered period. While the total area of forest affected by large wildfires in 2023 was significantly lower than in 2021, it was higher than in 2022 (14,797 ha). Therefore, the trend of forest area affected by wildfires annually in Italy can be considered negative for 2023.
At the regional scale, in Abruzzo, Basilicata, Calabria, Liguria, Molise, Puglia, Sicily, and Aosta Valley, the status is negative, as the 2023 value exceeds the average for the considered period. For all other regions, the area affected by fire in 2023 is lower than the average value.
Overall, the indicator status is therefore considered poor at the national scale.
The linear regression analysis of the national forest cover data affected by wildfire for the years 2018, 2019, 2020, 2022, and 2023 shows an increasing trend corresponding to 3,268 ha per year.
The increase in the national area affected by fire annually since 2018 is clearly visible in Figure 1, as is the exceptional relevance of 2021. At the regional scale, the trend shows significant differences in the evaluation.
A statistically significant increasing trend was calculated for the regions of Sicily, Calabria, Puglia, and Aosta Valley. In the regional analyses, the data for 2021 was also excluded from the regression analysis, as it was considered an exceptional event. Regarding protected areas, the positive trend was also observed for the Regional Nature Reserves (confirming the 2022 data analysis), and those included in the Special Protection Areas and Special Areas of Conservation of the Natura 2000 network.
The trend analyses for the remaining Italian regions and other categories of protected areas (National Parks, State Nature Reserves, Regional Nature Parks, Ramsar Areas, and Other Protected Natural Areas) exhibited low statistical significance.
At the national level, the average value for the area affected by wildfire annually between 2018 and 2023 is 10,634 ha.
Dati
Table 1: Area of the main forest categories affected by fire at the regional level
ISPRA elaboration on ISPRA and JRC data
TNC: burned area of unclassified forests; T1: burned area of deciduous broadleaf forests; T2: burned area of evergreen broadleaf forests ; T3: burned area of coniferous forests ; T34: burned area of sub-alpine temperate forests; TOT: sum of TCN, T1, T2, T3 and T34.
Table 2: Area of the main forest categories affected by fire in the different categories of protected areas and old-growth forests (2018)
ISPRA elaboration on ISPRA and JRC data.
TNC: burned area of unclassified forests; T1: burned area of deciduous broadleaf forests; T2: burned area of evergreen broadleaf forests ; T3: burned area of coniferous forests ; T34: burned area of sub-alpine temperate forests; TOT: sum of TCN, T1, T2, T3 and T34.
The calculations made for individual protected area categories are made without considering overlaps between them. So it is possible to make sums within the different categories but not between them; alternatively, there would be a risk of overestimating burned areas, particularly between EUAP and Natura 2000 categories.
Table 3: Area of the main forest categories affected by fire in the different categories of protected areas and old-growth forests (2019)
ISPRA elaboration on ISPRA and JRC data.
TNC: burned area of unclassified forests; T1: burned area of deciduous broadleaf forests; T2: burned area of evergreen broadleaf forests ; T3: burned area of coniferous forests ; T34: burned area of sub-alpine temperate forests; TOT: sum of TCN, T1, T2, T3 and T34
The calculations made for individual protected area categories are made without considering overlaps between them. So it is possible to make sums within the different categories but not between them; alternatively, there would be a risk of overestimating burned areas, particularly between EUAP and Natura 2000 categories.
Table 4: Area of the main forest categories affected by fire in the different categories of protected areas and old-growth forests (2020)
ISPRA elaboration on ISPRA and JRC data.
TNC: burned area of unclassified forests; T1: burned area of deciduous broadleaf forests; T2: burned area of evergreen broadleaf forests ; T3: burned area of coniferous forests ; T34: burned area of sub-alpine temperate forests; TOT: sum of TCN, T1, T2, T3 and T34
The calculations made for individual protected area categories are made without considering overlaps between them. So it is possible to make sums within the different categories but not between them; alternatively, there would be a risk of overestimating burned areas, particularly between EUAP and Natura 2000 categories.
Table 5: Area of the main forest categories affected by fire in the different categories of protected areas and old-growth forests (2021)
ISPRA elaboration on ISPRA and JRC data.
TNC: burned area of unclassified forests; T1: burned area of deciduous broadleaf forests; T2: burned area of evergreen broadleaf forests ; T3: burned area of coniferous forests ; T34: burned area of sub-alpine temperate forests; TOT: sum of TCN, T1, T2, T3 and T34
The calculations made for individual protected area categories are made without considering overlaps between them. So it is possible to make sums within the different categories but not between them; alternatively, there would be a risk of overestimating burned areas, particularly between EUAP and Natura 2000 categories.
Table 6: Area of the main forest categories affected by fire in the different categories of protected areas and old-growth forests (2022)
ISPRA elaboration on ISPRA and JRC data.
TNC: burned area of unclassified forests; T1: burned area of deciduous broadleaf forests; T2: burned area of evergreen broadleaf forests ; T3 burned area of coniferous forests ; T34 burned area of sub-alpine temperate forests
The calculations made for individual protected area categories are made without considering overlaps between them. So it is possible to make sums within the different categories but not between them; alternatively, there would be a risk of overestimating burned areas, particularly between EUAP and Natura 2000 categories.
Table 7: Area of the main forest categories affected by fire in the different categories of protected areas and old-growth forests (2023)
ISPRA elaboration on ISPRA and JRC data.
TNC: burned area of unclassified forests; T1: burned area of deciduous broadleaf forests; T2: burned area of evergreen broadleaf forests ; T3 burned area of coniferous forests ; T34 burned area of sub-alpine temperate forests
The calculations made for individual protected area categories are made without considering overlaps between them. So it is possible to make sums within the different categories but not between them; alternatively, there would be a risk of overestimating burned areas, particularly between EUAP and Natura 2000 categories.
Figure 1: Area of major forest categories affected by fire nationwide
ISPRA elaboration on ISPRA and JRC data.
TNC: unclassified forests; T1: deciduous broadleaf forests; T2: evergreen broadleaf forests; T3 coniferous forests; T34 sub-alpine temperate forests
Figure 2: Regional forest area covered by fire (2023)
ISPRA elaboration on ISPRA and JRC data.
Figure 3: National forest area affected by fire by biogeographic regions.
ISPRA elaboration on ISPRA and JRC data.
Forest area covered by fire by biogeographic regions in percent (histogram columns) and forest area covered by fire in Italy in hectares (black line)
Figure 4: National area of major forest categories affected by fire in protected natural areas (EUAP)
ISPRA elaboration on ISPRA and JRC data
T: unclassified forests; T1: deciduous broadleaf forests; T2: evergreen broadleaf forests; T3 coniferous forests; T34 sub-alpine temperate forests
The EUAPs considered refer to the official list of protected areas, see Resolution of the State-Region Conference of December 17, 2009. Only terrestrial EUAPs were considered
In the reference time series, 2021 recorded the largest burned forest area at the national scale (Figure 1), followed by 2023 in terms of total affected surface area.
From 2018 to 2023, the forest types most affected was evergreen broadleaves forest (T2 – predominantly holm oak forests and Mediterranean scrub), followed by deciduous mixed broadleaves forest (T1 – predominantly oak forests) and conifers (T3 – predominantly pine forests). This result does not align with the relative abundance estimates of forest cover in the national territory, which show deciduous mixed broadleaf forests as the most widespread in Italy (source: National Forest Inventory and Carbon Stock of Forests, 2015).
The region with the most significant increasing trend of wildfire affected surface is Sicily, with an annual rate of 2,044 ha.
Aosta Valley is the region with the lowest growth trend, with a value of about 10 ha per year.
As shown in Figure 2, the regions with the largest forest areas affected by wildfire in 2023 are Sicily (10,080 ha) and Calabria (2,987 ha). Together, these two regions account for 83% of the total national forest area affected by fire. In both regions, especially in Sicily, the burned areas were predominantly evergreen broadleaf forests (T2), mainly consisting of holm oaks and Mediterranean scrub. Sicily alone represents 64% of the entire forest area affected by fire in Italy in 2023.
In 2023, no forest area were affected by wildfires in the region of Veneto, Friuli-Venezia Giulia, Emilia-Romagna, Umbria, and Marche (Table 1).
As seen in Figure 3, over the time series period, the biogeographical region with the largest forest area affected by wildfire is consistently the Mediterranean region. The year with the largest forest area burned in the Alpine biogeographical region was 2019, while the year with the largest area burned in the Continental biogeographical region was 2022. In 2023, 96% of the forest areas affected by wildfire were in the Mediterranean biogeographical region.
Regarding the national protected area system, the analysis of the time series shows that the Special Areas of Conservation category exhibits the largest increasing trend, with a value of 1,064 ha per year (Tables 2 to 7).