Introducing VFEI: A VIIRS-based Fire Emission Inventory for High-Resolution Model Grids

We developed a new open biomass burning emission inventory using VIIRS I-band (375 m resolution at nadir) hotspots detections and the most recent compilations of emission factors available. We named this inventory as VIIRS-based Fire Emission Inventory (VFEI). VFEI provides daily global emission fluxes at ~500 m resolution for 46 species of gases and aerosols. We compared VFEI against other 4 major biomass burning emission inventories: the Quick-Fire Emission Dataset (QFED), the Global Fire Assimilation System (GFAS), the Global Fire Emissions Dataset (GFED) and the Fire Inventory from NCAR (FINN). Results showed that globally, VFEI and these other four inventories produce comparable total carbon emissions (~2000 Tg per year). However, differences of up to a factor of 7 appear at the regional level, probably due to uncertainties on emission factors in different biomes and lack of observational data in remote areas. We further conducted simulations using the Weather Research and Forecast model with Chemistry (WRF-Chem) using VFEI emissions. Two domains and periods were selected for simulations: (i) Southern Africa during September 2016 to overlap with the ObseRvations of Aerosols above CLouds and their intEractionS (ORACLES) field experiment, and (ii) North America during July-August 2019 to make use of the airborne data of the Fire Influence on Regional to Global Environments Experiment and Air Quality (FIREX-AQ) campaign. Results showed that when comparing against MODIS Multi-Angle Implementation of Atmospheric Correction (MAIAC) AOD retrievals, the model in both domains perform with low biases (±0.05) and high correlations (R>0.76). This is further proved by comparing against in-situ data from the AErosol RObotic NETwork (AERONET) sites. When comparing against airborne data from the ORACLES and FIREX-AQ campaigns, the model represents well the concentrations of carbon monoxide and black carbon, with correlations (R) of 0.77 in Southern Africa. Nonetheless, performance in the North American simulation showed higher biases because the model failed to reproduce the correct timing (diurnal cycle) and transport of smoke plumes. This highlights the necessity for using geostationary satellites to constrain the diurnal cycle of fires.

The figure below shows an example of the impact of high-resolution biomass burning emissions in fine model grids. An example of how a large fire is mapped from a coarse inventory (~10 km; top row) into a 3 km model grid (second row). The benefit of the high-resolution of VFEI emissions is shown in the third row and how VFEI is mapped into the same 3 km model grid (bottom row). It is evident that using VFEI allows for a more representative mapping of the fire.