Abstract
Wildfire emissions can vary substantially between different inventories
due to the assumptions made in the emission creation process, including
the defined vegetation type, fire detection, fuel loading, fraction of
vegetation burned and emissions factors. Here, we focus on the
uncertainty in emission factors and the resulting impact on modeled
composition. We use the Community Atmosphere Model with chemistry
(CAM-chem) to simulate 2014 atmospheric composition and focus on carbon
monoxide (CO), a trace gas emitted from incomplete combustion and also
produced from secondary oxidation of volatile organic compounds (VOCs).
Fire is a major source of atmospheric CO and VOCs. Multiple simulations
are compared, from an ensemble using four fire emission inventories
(CMIP6/GFED4s, QFED2.5, GFAS1.2 and FINN1.5) and a range of sensitivity
tests based on CO and VOC emission factor uncertainties. We compare
model output and evaluate against CO observations from the Measurements
of Pollution in the Troposphere (MOPITT) satellite-based instrument. For
some regions, emission factor uncertainty spans the results found by
using different inventories. Finally, we use modeled ozone (O3) to
investigate how emission factor uncertainty influences the atmospheric
oxidative environment. Overall, accounting for emission factor
uncertainty lends a range of uncertainty to simulated results.