Résumé
We identified four global fire regimes based on a k-means algorithm using five variables covering the spatial, temporal and magnitude dimensions of fires, derived from 19-year long satellite burned area and active fire products. Additionally, we assessed the relation of fire regimes to forest fuels distribution. The most extensive fire regime (35% of cells having fire activity) was characterized by a long fire season, medium size fire events, small burned area, high intensity and medium variability. The next most extensive fire regime (25.6%) presented a long fire season, large fire events and the highest mean burned area, yet it showed the lowest intensity and the least variability. The third group (22.07%) presented a short fire season, the lowest burned area, with medium-low intensity, the smallest fire patches and large variability. The fourth group (17.3%) showed the largest burned area with large fire patches of moderate intensity and low variability. Fire regimes and fuel types showed a statistically significant relation (CC = 0.58 and CC’ = 0.67, p \textless 0.001), with most fuel types sustaining all fire regimes, although a clear prevalence was observed in some fuel types. Further efforts should be directed towards the standardization of the variables in order to facilitate comparison, analysis and monitoring of fire regimes and evaluate whether fire regimes are effectively changing and the possible drivers.