The influence of geoengineering on evolving fire activities under climate change

Cheng-En Yang and Joshua Fu, Civil and Environmental Engineering, University of Tennessee, Knoxville, TN, Forrest Hoffman, Computational Earth Sciences Group, Oak Ridge National Laboratory, Oak Ridge, TN

Background/Question/Methods
Wildfires have become more intense and frequent in recent years as a result of climate change. Fire activities greatly influence terrestrial ecosystems by perturbing the carbon cycle and thus enhancing climate change regionally. With rising surface temperatures projected by Earth system models, stronger water deficits and lower soil moisture suggest elongated drought periods that increase the probability of fires in the future. To mitigate such adverse impacts, various climate intervention strategies are proposed to offset further warming brought by elevated anthropogenic emissions, resulting in reduced fire activities in such a geoengineered world. However, as vegetation productivity increases, accelerated by CO2 fertilization in a cooler climate, biomass that serves as available fuel load could become higher in a geoengineered climate than a non-geoengineered climate, causing a more severe burning activity once a fire occurs. Therefore, it is important to investigate fire activities to understand their impacts on ecosystems and predict changes in fire risks to human health and property loss. In this study, we analyze fire-related variables from the Geoengineering Large Ensemble Project (GLENS) model outputs to estimate how fire activities would evolve and their potential impacts on carbon cycle under different climate scenarios.
Results/Conclusions
The projected global burned area due to fires under the Representative Concentration Pathway 8.5 (RCP8.5) scenario increases (1.2×108 ha/yr) during 2020–2097 while that under the stratospheric sulfate aerosol injection (SAI) geoengineering scenario decreases (−4.1×103 ha/yr). Among all ecoregions, the ecosystems impacted most by climate geoengineering are croplands for the burned area (reduced from 3.8×107 ha/yr to −7.6×10−3 ha/yr), sparsely vegetated regions for fuel load (reduced from 3.8×10−4 PgC/yr to 3.4×10−4 PgC/yr), and mixed forests regions for fire-induced carbon loss (increased from 3.7×10−2 PgC/yr to 4.5 PgC/yr). Deciduous needleleaf forests have the least changes for the burned area (reduced from 3.3×107 ha/yr to −1.2×10−2 ha/yr), fuel load (reduced from 3.8×10−4 PgC/yr to 3.4×10−4 PgC/yr), and fire-induced carbon loss (increased from 3.7×10−2 PgC/yr to 4.5 PgC/yr) from RCP8.5 due to SAI. In addition, the ratios of fire-induced carbon loss to burned area show an increasing trend under SAI but a decreasing trend in RCP8.5 globally, implying higher fire severity (more complete carbon combustion) when a fire occurs under SAI in GLENS.