Mapping Future Fire Danger Against Brazil's Landscape Resilience

Fire plays a central role in shaping ecosystem dynamics, biodiversity conservation, and the provision of ecosystem services; however, its role varies markedly among ecosystems. This is particularly critical in Brazil, a country that hosts globally important biomes and underpins vital functions such as climate regulation and the water–energy–food nexus. Recent observational studies indicate that Brazil is already undergoing shifts in the occurrence of extreme heat and drought events, and climate model simulations suggest that these trends will intensify in the future. However, the implications of these shifts for future fire risk patterns remain insufficiently explored, especially within an integrated risk framework that assesses how climate-driven hazard interacts with the heterogeneous resilience of ecosystems across the country.

Here, we ask how likely Brazilian ecosystems are to experience extreme fire danger conditions under future climates, and map how this hazard relates to both historical and projected patterns of landscape resilience. To this end, we perform a nationwide assessment of future fire danger using the Canadian Fire Weather Index (FWI) derived from daily CMIP6-based climate projections retrieved from the CLIMBra dataset, which was developed specifically for Brazil's climate conditions using an observational-based dataset. Employing a novel heatwave-based framework, we identify extreme fire danger events and characterize future changes in their intensity, duration, frequency, and spatial extent. Beyond this climate-based assessment, we contrast these changes from an ecosystem resilience perspective by integrating future fire danger projections with projections of landscape resilience. A Random Forest model, trained on the relationship between land cover and a map of landscape resilience classes, is applied to multiple future land-cover scenarios to estimate concurrent changes in both climate-driven fire danger and landscape resilience. This integrated approach allows us to pinpoint areas where high future fire danger overlaps with low landscape resilience.

Our results project up to approximately 30 additional compound hot-dry days per year by the end of the century across the country. These changes are expected to create a more challenging scenario for fire management, with a widespread increase in extreme fire danger across Brazil. For instance, the spatial extent and number of extreme fire danger days are projected to rise by approximately 69% and 42% on average, respectively, under intermediate-emission scenarios in the first half of the century. This integrated mapping enables us to reveal where projections of intensifying fire weather converge with those of future low landscape resilience, thereby highlighting priority regions and protected areas for targeted action. We believe that our framework will enable the integrated assessment of future fire danger and ecosystem vulnerability. These findings can guide national landscape and territorial policies by helping to prioritize actions in regions facing significant novel fire threats (transformative risk) or intensifying fire regimes (adaptive risk). They underscore the need for proactive fire management and conservation/restoration strategies that explicitly account for both climatic intensification and landscape resilience. Despite inherent uncertainties in climate and land-cover projections, our study provides a critical foundation for supporting more effective environmental planning and decision-making under a changing climate.