Investigation of spatiotemporal variability in South American wildfire emissions and its impacts on CO concentrations

Wildfires in South America are a significant concern, causing high emissions and deforestation rates. They affect air quality, radiation balance, and sensitive ecosystems like the Amazon rainforest. Wildfires are expected to intensify with future land use and climate changes, making it crucial to enhance decision-making tools. Models of atmospheric composition, combined with wildfire emissions inventories, support decision-making by simulating events and their impacts on air quality. There are currently a range of wildfire/biomass burning emission inventories, which all use different approaches. This can lead to substantial differences in estimated emissions and thus impacts on atmospheric composition estimation.  This study aims to assess four inventories (2004-2022) in South America: Global Fire Emissions Database (GFED), Fire INventory from NCAR (FINN), Global Fire Assimilation System (GFAS) and Brazilian Biomass Burning Emission Model (3BEM-FRP), focussing on carbon monoxide (CO) given its relatively large emission and complementary satellite missions retrieving atmospheric CO. Our results analyse the temporal consistency in the emission seasonal cycles from the inventories and quantify the spatial agreement/differences between them. We also exploit the Measurements Of Pollution In The Troposphere (MOPITT) retrieved CO to assess the links between emission inventory tendencies with that of the atmospheric temporal evolution. Finally, we use an offline version of the INteractive Fire and Emission algoRithm for Natural envirOnments (INFERNO) model, within the Joint UK Land Environment Simulator (JULES) framework to investigate simulated skill of emissions of CO against the observational constraints above as INFERNO is the fire model of choice in the UK Earth System Model (UKESM).