Improvements on the modeling of fires: emissions and plume rise processes

Every year, biomass burning is responsible for the emission of many teragrams of carbon released into the atmosphere. Fires emit numerous trace gases, including greenhouse gases and ozone precursors, which can create poor air quality episodes when populations are exposed. Furthermore, biomass burning emits several aerosol particles which can interact with solar radiation, cloud formation and precipitation processes. These impacts can be extrapolated from the local to regional and global scales when the smoke is transported in the higher levels of the troposphere. Chemistry-transport models are used to better understand and predict the transport of smoke. Thanks to increased computational resources available, models are now capable of solving complex chemical and aerosol mechanisms, considering hundreds of species. However, most of the currently available biomass burning inventories include emissions for a limited number of species and often as gridded datasets at resolutions no higher than 0.1°. The latter is specially a problem when conducting high-resolution simulations since fires can be misplaced in the model grid and be subject to different meteorology and subsequent transport. To solve this, we generated a new biomass burning emission inventory named VIIRS-based Fire Emission Inventory (VFEI). VFEI uses the hotspot detections and fire radiative power derived from the Visible Infrared Imaging Radiometer Suite (VIIRS) band-I onboard of the Suomi-NPP satellite. VFEI includes daily emission fluxes for more than 40 species at resolutions of up to 500 m. Another aspect of the modeling of fires, is the vertical distribution of emissions. Among the several approaches that models can consider, we focus on the plume rise model introduced firstly by Freitas et al. However, we use the fire radiative power to calculate the heat flux and convective energy associated to the plume development. In addition, we improved the entrainment treatment on plume rise for a more realistic process. We tested these developments together with VFEI emissions using WRF-Chem for two regions: Southern Africa (September 2016) and North America (Summer 2019). We evaluated our simulations against aerosol optical depth (AOD) observations derived by the Moderate Resolution Imaging Spectroradiometer (MODIS) onboard of the Terra and Aqua satellites and AErosol RObotic NETwork (AERONET) sites. Results show AOD biases below 0.2 for both regions, but higher biases in some fires in North America. In addition, we used airborne data from the ObseRvations of Aerosols above CLouds and their intEractionS (ORACLES; Southern Africa) and the Fire Influence on Regional to Global Environments Experiment and Air Quality (FIREX-AQ; Continental US) field experiments using selected species. Results for both regions show biases for black carbon and carbon monoxide of around 15%, but some fires in North America showed higher biases, probably due to underestimations of fire radiative power due to dense smoke covering.