ESFire30: unravelling four decades of wildfire burned area data in peninsular Spain
Clara Ochoa, Amin Khairoun, José Antonio Rodado Coello, Emilio ChuviecoBackground
Spain has experienced extensive wildfires over the past four decades, yet long-term, high-resolution spatially explicit burned area data covering the whole country are not available.
Aims
This study presents ESFire30, a burned area perimeter dataset for peninsular Spain covering 1985–2023 based on satellite observation, its methods and characteristics, and demonstrates its potential by analysing spatio-temporal trends over this period.
Methods
ESFire30 covers 493,000 km² at 30 m spatial resolution and was developed using machine learning algorithms and visual inspection of Landsat imagery. Minimum burned patch mapped was 5 ha. Two datasets were produced: annual burned area for 1985–2023 and monthly data for 2001–2023 by integrating Landsat and MODIS active fire detection.
Key Results
Results show recurrent fire patterns, with severe years in 1985, 1989 and 1994. The northwest accounted for 31.2% of the total burned area and large fires (>500 ha) for 44.5%. Shrublands were most affected (~50%), followed by forests (~20%), which displayed increasing susceptibility. Fire activity peaked in spring (April–May) and summer (July–September).
Conclusions
The findings reveal significant spatial and temporal variability in wildfire occurrence across Spain.
Implication
ESFire30 provides spatially explicit mapping of fire occurrence across peninsular Spain, facilitating spatial and temporal analysis of fire impacts, while complementing existing databases.