Defining global syndromes of fire and the relationship of these to biomes, climate and human activity

e understanding of fire is limited largely to the annual extent of burning as detected by satellites. This is problematic because fire is multi-dimensional, and focus on individual metrics belies both the complexity and importance of fire within the Earth system. In an applied sense, the lack of a unified understanding of fire impedes estimation of GHG emissions or prediction of future fire regimes as a consequence of changing patterns of climate and land use. To address this we identified five key characteristics of fire regimes: size, frequency, intensity, season and extent. We combined new global datasets with existing datasets to examine cross-correlations among characteristics. We demonstrate that only certain combinations of fire characteristics are possible and this likely reflects fundamental energetic constraints derived from interactions between under-lying fuel types, climate and rates of re-growth post-fire. For example, very intense fires can only occur infrequently because a system requires a lengthy period to develop sufficient fuel to burn. Further, very cool fires only occur infrequently because fuels are not available to burn. Following, we applied a clustering algorithm to these data to determine whether we could identify syndromes of fire regimes. Pyromes, as global syndromes of fire are conceptually analogous to biomes (global syndromes of vegetation) where the extent of each pyrome is determined solely as a product of the fire characteristics themselves. A point of difference to biomes being that no one has previously attempted to quantify the global range of fire syndromes. We identified five pyromes, four of which we believe represent distinctions between crown, litter and grass-fuelled fires. The relationship of pyromes to biomes and climate are not deterministic as different biomes and climates may be represented within a single pyrome. There were, however, striking correlations between particular pyromes and biomes. For example frequent, intense, large fires were associated with grasslands and tropical savanna in 75% of instances. Rare intense, large fires were dominantly associated with boreal forests. Crucially, we identified a fifth pyrome that we consider to represent human-engineered modifications to fire characteristics. This pyrome, characterised by infrequent, cool, small fires that can occur throughout the year, occurs within all biomes, and was dominant in regions of extensive land transformation. Our research presents a conceptual framework that may help develop capacity to predict future fire, as our analysis suggests that pathways of change in future fire regimes are unlikely to be unilaterally responsive to climate in a deterministic way.