Wildfires are a significant ecological process influenced by climate, fuel availability, ignition sources, and human activities. Fires modify climate through emissions and impact the biosphere via biomass burning, potentially leading to deforestation. While beneficial in some ecosystems (e.g., savannas), wildfires pose environmental hazards, particularly during extreme events with substantial economic, social, and ecological consequences. The impact of landscape fire smoke on human health is also a major concern. Existing research indicates that climate factors, particularly fuel availability (often measured by net primary production, NPP) and precipitation, exert a greater influence on global biomass burning than human activities across broad spatiotemporal scales. However, regional-scale drivers are more diverse, and human activity can be the dominant factor in specific areas. Given the strong climate-fire relationship, climate change is expected to alter the spatial distribution of fire activity. Studies suggest increases in fire season severity and wildfire potential, and a possible shift towards a temperature-dominated global fire regime. This study aims to demonstrate that simple climate indicators can reproduce and explain the current global pattern of fire-prone regions and to utilize these indicators to predict future changes in fire activity. The study focuses primarily on the climate-fire relationship, minimizing the influence of interannual variability, ignition elements, and human factors (unless climate-related). The central hypothesis is that, at broad spatial and decadal scales, fire occurrence is highly probable wherever favorable climatic fire conditions exist.

Numerous studies have investigated the effects of climatic and human factors on global fire activity. These studies show that while human activity plays a role, climate variables such as temperature and precipitation are the primary drivers of fire activity on large spatial and temporal scales. The variability in global interannual fire response to climate has also been analyzed. However, the precise magnitude and geographical distribution of future changes in fire activity remain uncertain. Previous work has pointed to potential increases in fire season severity and wildfire potential, suggesting a future dominated by temperature rather than precipitation or human influences. This study builds upon these previous efforts by focusing on simple climate indicators to model current and future fire-prone areas.

The study uses the Global Fire Emissions Database (GFED4) to obtain global monthly burned area data from 1996–2016 at 0.25° spatial resolution. Climate data are from WFDE5, which utilizes the WATCH Forcing Data methodology applied to ERA5 reanalysis products. These data are downscaled to 0.25° resolution. To identify fire-prone regions, the authors compare the global distribution of temperature and precipitation indicators with GFED4 burned area data. They categorize regions into four classes based on Köppen-Geiger climate classification (Tropical, Arid, Temperate, Boreal) and establish thresholds defining fire-prone months. Each class is characterized by the primary climatic driver of fire activity (low precipitation, high temperature, or a combination). Thresholds are determined by comparing the probability distribution of climatic variables at locations with high fire activity versus those with low activity. Fire-prone years (FPY) are defined as years with at least one month meeting the thresholds, and potential fire season length (PFSL) is the average number of fire-prone months during FPY. For future fire-climate classification, the authors apply these thresholds to future climate projections from CMIP5 GCMs under the RCP8.5 scenario (worst-case climate change scenario). Two approaches were considered: one assuming quick vegetation adaptation and one without. Given uncertainty about rapid adaptation, the analysis focuses on projected changes in fire-climate classification variables, maintaining the general division of Tropical, Arid, Temperate, and Boreal regions as in present conditions. The study recognizes potential limitations, acknowledging that climate-fire relationships may not be stationary. The method excludes regions where cropland land cover exceeds 90% due to the likely influence of agricultural practices on fire activity. The thresholds are automatically selected for each distribution to maximize the area between the density functions of fire points and non-fire points. The selection process prioritizes variables and thresholds based on area difference and classified fire-point percentage criteria.

The study reveals a strong correlation between observed fire locations and specific climate conditions. Over 70% of the land area is well classified as either fire-prone or fireless based on the climate classification. The four fire-climate classes identified (Tropical dry season, Arid fuel-limited, Temperate dry and hot season, and Boreal hot season) effectively capture the observed spatial patterns of fire activity. Future projections under the RCP8.5 scenario indicate a substantial increase in fire-prone areas. Boreal forest areas are projected to experience the most significant expansion (47% increase, and 111.5% increase for the most frequent categories), extending northward and southward. Temperate regions also show considerable expansion (24.5% increase), particularly in Southern China and Southern Europe. Tropical regions exhibit fewer spatial changes (6.3% contraction), with notable differences in South America. Arid regions show a smaller expansion (5%), with considerable uncertainty near desert areas. Analysis of potential fire season length (PFSL) and fire-prone years (FPY) reveals further key findings. Boreal regions are projected to experience a lengthening of the potential fire season. While some areas of Eastern Asia might see a slight shortening due to increased precipitation, the overall trend is a lengthening driven by amplified Arctic warming. Increased FPY is projected particularly in northerly latitudes. Temperate regions also display PFSL lengthening, especially in Southern Europe due to combined precipitation decline and warming. The Western US, already experiencing increasing fire season length and large fires, may also see further increases. Tropical regions show less change in PFSL, with some areas experiencing shortening (Northern African savanna) and others lengthening (Southern Africa). Variations in precipitation patterns potentially linked to ENSO changes might drive these changes. Overall, the study projects a significant increase in fire risk worldwide, particularly in Boreal and Temperate regions.

The findings demonstrate a significant link between climate change and future wildfire activity, especially in high-latitude regions. The projected increase in fire-prone areas and fire season length highlights the urgent need for enhanced wildfire management strategies. The study's focus on climate as a primary driver provides valuable insights into large-scale fire patterns. The uneven impact on different climate zones underscores the need for region-specific adaptation and mitigation measures. While the study simplifies the complexity of wildfire dynamics by focusing on climate, it offers robust projections of spatial and temporal changes in global fire risk. The limitations regarding vegetation adaptation and the uncertainties associated with future precipitation patterns should be considered when interpreting the results. The study's results align with many previous studies pointing toward increasing fire severity and fire season length in various regions of the world. The increased risk in boreal regions is particularly alarming given the potential release of significant amounts of stored soil carbon into the atmosphere, potentially creating a positive feedback loop accelerating warming.

This study provides compelling evidence for a substantial expansion of global fire-prone areas by the end of the 21st century, driven mainly by rising temperatures in Boreal and Temperate zones. The projections highlight the necessity of implementing effective wildfire prevention and management strategies globally. The findings underscore the significance of incorporating climate change projections into wildfire risk assessment and preparedness plans. Future research could focus on improving the understanding of vegetation adaptation to changing climate conditions and the interaction between climate change, human activities, and wildfire dynamics at finer spatial and temporal scales.

The study primarily focuses on the climate-fire relationship and simplifies the complex interplay of factors contributing to wildfires. The analysis does not explicitly account for changes in vegetation composition and structure, human-induced ignitions, or the influence of extreme weather events beyond monthly averages. The assumption of non-stationary climate-fire relationships introduces uncertainty in projections, particularly concerning precipitation variability and its influence on fuel conditions. The reliance on CMIP5 GCMs and RCP8.5 scenario also contributes to inherent uncertainties in the future projections. The resolution of GFED4 data, while suitable for large-scale analysis, might not capture regional variations in fire activity.