The Mediterranean basin is widely recognized as a climate change hotspot, with projections consistently indicating an intensification of heat and drought conditions in the coming decades. Within this context, Italy represents a paradigmatic Mediterranean country, experiencing frequent and often extensive wildfires, particularly in its southern and insular regions, where drought, high temperatures, complex topography, and dense human presence interact. Here we investigate to what extent monthly burned area at provincial (NUTS3) scale can be modeled using only environmental predictors, including climate and land use variables. We compile a dataset of ground-based burned area polygons for 107 Italian provinces over 2007–2021, normalize burned area by provincial area, and use as explanatory variables monthly climate variables from ERA5 and aggregated CORINE land-cover classes. We train a random forest regression model in a leave-one-year-out cross-validation framework to predict monthly burned area anomalies and to rank predictors using permutation importance and partial dependence analysis. The model achieves statistically significant correlations between observed and predicted anomalies in most provinces, with an average correlation around 0.6. At the national scale, the most influential predictor, after the seasonal cycle, is evaporation, followed by total cloud cover and total precipitation, while land-cover variables have only secondary importance. Our results highlight the central role of evaporative demand in modulating burned area in Italy, implying that projected increases in evaporative demand under global warming could further promote drier conditions in which fuels are more often ready to burn, possibly leading to more common and persistent extreme wildfire seasons.

Evaporative demand as a key driver of fires in Italy

Gabriele Vissio
Primo
;
Alice Baronetti;Antonello Provenzale;Mara Baudena
Co-ultimo
;
2026

Abstract

The Mediterranean basin is widely recognized as a climate change hotspot, with projections consistently indicating an intensification of heat and drought conditions in the coming decades. Within this context, Italy represents a paradigmatic Mediterranean country, experiencing frequent and often extensive wildfires, particularly in its southern and insular regions, where drought, high temperatures, complex topography, and dense human presence interact. Here we investigate to what extent monthly burned area at provincial (NUTS3) scale can be modeled using only environmental predictors, including climate and land use variables. We compile a dataset of ground-based burned area polygons for 107 Italian provinces over 2007–2021, normalize burned area by provincial area, and use as explanatory variables monthly climate variables from ERA5 and aggregated CORINE land-cover classes. We train a random forest regression model in a leave-one-year-out cross-validation framework to predict monthly burned area anomalies and to rank predictors using permutation importance and partial dependence analysis. The model achieves statistically significant correlations between observed and predicted anomalies in most provinces, with an average correlation around 0.6. At the national scale, the most influential predictor, after the seasonal cycle, is evaporation, followed by total cloud cover and total precipitation, while land-cover variables have only secondary importance. Our results highlight the central role of evaporative demand in modulating burned area in Italy, implying that projected increases in evaporative demand under global warming could further promote drier conditions in which fuels are more often ready to burn, possibly leading to more common and persistent extreme wildfire seasons.
2026
Istituto di Scienze dell'Atmosfera e del Clima - ISAC - Sede Secondaria Torino
Istituto di Geoscienze e Georisorse - IGG - Sede Pisa
climate drivers,
evaporation,
fires,
land cover,
Mediterranean Italy,
random forest regression
File in questo prodotto:
File Dimensione Formato  
Vissio2026.pdf

accesso aperto

Descrizione: Evaporative demand as a key driver of fires in Italy
Tipologia: Versione Editoriale (PDF)
Licenza: Creative commons
Dimensione 6.99 MB
Formato Adobe PDF
6.99 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.14243/608689
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? 0
social impact