Enhanced future vegetation growth with elevated carbon dioxide concentrations could increase fire activity

Wildfires are a significant Earth system process influencing ecosystem composition and the atmosphere. Increased fire frequency, size, and extended fire seasons have been observed in various regions, such as the western United States, over the past decade. Projected future increases in fire activity are attributed to factors including intensified drought, more frequent heatwaves, and changes in fire suppression strategies. Anthropogenic climate change is expected to worsen fire weather conditions globally, impacting areas like the western US, Australia, the Mediterranean, and even the Amazon under low climate mitigation scenarios. In addition to these climate factors, rising atmospheric CO2 concentrations are associated with increased carbon uptake and storage by terrestrial biomes, known as the CO2 fertilization effect. This effect has been linked to increased terrestrial biosphere activity and global "greening." However, the impact of this fertilization effect on wildfires remains uncertain. Some studies suggest that CO2 fertilization increases fire occurrence by boosting fuel load, while others propose it may mitigate fire severity by increasing live fuel moisture content. These responses vary depending on fire regimes; in fuel-limited regimes, fire is more sensitive to fuel load changes, whereas in flammability-limited regimes, it's more responsive to fuel moisture changes. Fire regimes themselves can also shift in response to climate change. This research utilizes seven state-of-the-art Earth system models from CMIP6, each with varying fire activity representations, to quantify the impact of an idealized CO2 increase on fire carbon emissions. The study aims to assess projected wildfire changes under idealized CO2 increases in current models and to determine the relative significance of physical climate impacts (e.g., warming and drying) versus vegetation dynamics (e.g., CO2 fertilization and enhanced vegetation growth).

The literature surrounding the impacts of climate change and rising CO2 on wildfire activity presents a complex picture. Studies have shown a clear link between increased temperatures, drought conditions, and the intensification of fire weather, leading to more frequent and severe wildfires in various regions. The role of CO2 fertilization, however, is less clear-cut. While enhanced vegetation growth due to CO2 fertilization could lead to increased fuel loads, potentially fueling more intense fires, it might also increase live fuel moisture, thereby mitigating fire severity. The interplay of these factors is dependent on the specific fire regime, with fuel-limited regimes being more sensitive to fuel load and flammability-limited regimes being more sensitive to fuel moisture. The existing literature highlights the need for comprehensive modeling studies that can account for these complex interactions to better understand and predict future wildfire trends.

This study employed seven state-of-the-art Earth system models (ESMs) from the Coupled Model Intercomparison Project version 6 (CMIP6) to investigate the impact of idealized increases in atmospheric CO2 concentrations on fire activity. All models included varying degrees of complexity in their representation of fire processes. The primary variable of interest was "fire carbon emissions" (fFire) obtained from the CMIP6 database. The models were used to simulate the effects of a 1% annual increase in CO2 concentrations, allowing for the observation of changes over a 140-year period. To disentangle the relative contributions of biogeochemical and radiative effects of CO2, two additional sets of simulations were conducted: one where only the biogeochemical effects of CO2 were considered (1% per year CO2-bgc), and another where only radiative effects were considered (1% per year CO2-rad). These separate simulations allowed the researchers to isolate the influence of vegetation dynamics (CO2 fertilization) from the direct effects of warming and drying. The model's ability to simulate fire carbon emissions was first evaluated using historical simulations (2002-2014) extended to 2021 using SSP5-8.5. The model outputs were compared with two satellite-based observational datasets: the Global Fire Emissions Database version 4.1 (GFED4.1s) and the Fire Inventory from NCAR version 2.5 (FINNv2.5). Regional and global comparisons were made to assess model performance in capturing the magnitude, interannual variability, and seasonality of fire carbon emissions. Statistical analysis, including pooled t-tests and binomial tests, were used to assess the statistical significance of the model results and the consistency of model responses across different regions and simulations. In addition to fire carbon emissions, the study also analyzed changes in net primary productivity (NPP), leaf area index (LAI), and other climate variables such as near-surface air temperature, precipitation, soil moisture, and relative humidity. Spatial and temporal correlation analyses were conducted to explore the relationships between fire activity and climate and vegetation changes across different simulations. The sensitivity of the fFire response to changes in vegetation was also investigated using regression analysis. The study considered various model representations of fire modules, including those with dynamic vegetation and those with prescribed vegetation types.

The multi-model mean (MMM) results across seven ESMs reveal a robust increase in fire carbon emissions (fFire) in response to a 1% annual increase in atmospheric CO2 concentrations. Globally, the MMM percent increase in fFire was 127.7 ± 79.2% over a 40-year period centered on CO2 doubling (66.4 ± 38.8% for a 40-year period). All models showed an increase in fFire. When the outlier model, GFDL-ESM4, was excluded, this value reduced to 80.9 ± 20.4%. This significant increase was largely attributed to biogeochemical mechanisms, specifically the CO2 fertilization effect, and the resulting enhanced vegetation growth which increased fuel loads. The radiative impacts of CO2, including warming and drying, resulted in a negligible change in fFire (1.7 ± 9.4%). Regional analyses showed substantial increases in fFire across most land areas, particularly in North America, Europe, Asia, and parts of the Southern Hemisphere. The US saw a MMM percent increase of 229.1 ± 106.2%. The analysis of 1% per year CO2-bgc (biogeochemical effects only) and 1% per year CO2-rad (radiative effects only) simulations further demonstrated that the increase in fFire was primarily due to the biogeochemical effects of CO2 on vegetation. The 1% per year CO2-bgc simulations showed similar large increases in fFire while the 1% per year CO2-rad simulations showed minimal change or even decreases, except in some high-latitude boreal regions. The spatial patterns of the NPP response closely mirrored those of the fFire response across all three simulations, indicating that biomass production (fuel availability) is a major driver of changes in fire carbon emissions. Significant correlations were found between the NPP and fFire responses. While many models found these correlations, there were some exceptions. Furthermore, there is some lag correlation between vegetation changes and subsequent increases in fFire. The analysis also revealed significant climate responses to the CO2 increase, including warming, changes in precipitation, drying, and shifts in surface energy fluxes. There are also some interactions between vegetation and climate changes that could result in a two-way feedback loop further increasing or decreasing the intensity of wildfires. While some climate changes might moderate the fire increase (e.g., increased soil moisture), others could amplify it (e.g., decreased precipitation). The study explored additional model findings, including the spatially varying vegetation response under 1% per year CO2-rad, with increases at high northern latitudes and decreases in many tropical regions. This points to the potential for warming and drying to reduce biomass in tropical regions and increase biomass at higher latitudes. Finally, a comparison with results from future emission scenarios (SSP3-7.0) showed that while the magnitude of the fFire response was smaller, the general pattern of increased fire carbon emissions under elevated CO2 remained consistent.

The study's findings directly address the research question by demonstrating a robust and significant increase in fire carbon emissions under idealized increases in atmospheric CO2. The results highlight the primary role of the CO2 fertilization effect and enhanced vegetation growth in driving this increase, overshadowing the relatively negligible contribution from direct climate impacts (warming and drying). This has important implications for understanding future wildfire trends and their interaction with the carbon cycle. The close relationship between changes in vegetation productivity and fire activity underscores the need for integrated approaches to managing fire risk and climate change mitigation. The research suggests that policy efforts should not only focus on mitigating the direct effects of climate change but also consider the significant influence of vegetation dynamics on fire activity. The results also highlight the complexity of the climate-vegetation-fire system and the limitations of current climate models in fully capturing the intricate interactions among these components. The large model spread for some regions indicates uncertainties, and the study's focus on idealized CO2 increases may not fully represent the complex interactions found in realistic future scenarios. Future research should focus on improving model representation of vegetation dynamics and fire processes in order to decrease the model uncertainty associated with these projections and reduce the potential for bias.

This research demonstrates a robust increase in fire carbon emissions under idealized increases in atmospheric CO2 concentrations, primarily driven by the CO2 fertilization effect and subsequent enhanced vegetation growth. The findings emphasize the crucial role of vegetation dynamics in influencing future wildfire activity and highlight the need to incorporate these ecological drivers into climate change mitigation and fire risk management policies. Future research should focus on improving the representation of vegetation dynamics, fire processes, and the interactions between climate and ecological factors in Earth system models to reduce uncertainties and improve the accuracy of projections.

The study acknowledges several limitations. First, the use of idealized CO2 increase scenarios may not fully capture the complexity of real-world climate change, which involves multiple interacting factors. Second, the large model spread in regional fFire responses points to uncertainties in model representation of fire processes. Third, most models lack dynamic vegetation, potentially limiting their ability to capture the full range of climate-vegetation-fire feedbacks. Finally, the study primarily focuses on fire carbon emissions and does not analyze other aspects of fire activity such as burned area or fire intensity, which would improve understanding.