CHARACTERIZING EVOLUTION OF BIOMASS BURNING SMOKE USING DUAL-VIEW GEOSTATIONARY SATELLITE RETRIEVALS

Abstract

Biomass burning has long been as a key source of global aerosol and greenhouse gas emissions. Emissions from biomass burning have significant impacts on air quality, human health, and radiative budgets. In recent years, increases in the frequency and severity of wildfires have drawn more attention to the need for improved characterization of smoke and the complex processes which govern its emission and evolution during transport in the atmosphere. Included in these are factors related to burning conditions, such as efficiency of combustion and fuel type, as well as factors related to atmospheric processing, such as photo-enhancement and photo-bleaching of absorption and coagulation, all of which influence the microphysical and optical properties of smoke. Observations from geostationary satellites, such as GOES-East and GOES-West, provide a unique opportunity to study these factors, as they can view a biomass burning event at high temporal resolution and track changes in the source and the smoke plume over time. Unfortunately, current GOES operational products use a limited set of aerosol models and rely on single-satellite observations, which often leads to smoke being classified broadly as a generic aerosol type. This limits our ability to characterize smoke properties accurately. A new retrieval method has been developed to address these limitations. It expands the range of candidate aerosol models to capture the full diversity of smoke, from fresh emissions to particles transported downwind that may have undergone complex interactions with the environment. Using a Bayesian framework, the method integrates reflectance data from both GOES-East and GOES-West and provides rigorous uncertainty estimates. It retrieves the two most plausible fine modes and their contributions to total optical depth, along with coarse mode and total absorption aerosol optical depth. We analyzed an August 2021 wildfire event near Kelowna, British Columbia, Canada, which provided a rare opportunity in which smoke plumes passed directly over a nearby Aerosol Robotic Network (AERONET) site for retrieval evaluation. Our retrievals generally agree with coincident AERONET aerosol optical depth (AOD) and absorption AOD (AAOD) to within their respective uncertainties. For most of the period, the plume is dominated by mixtures of fine-mode brown carbon and non-absorbing aerosols. Coarse-mode aerosols contribute only about 10% of total AOD, increasing to ~20% late in the retrieval window. To investigate two key sources of uncertainty in smoke processes, namely aging during transport and temporal changes in fire activity, we examined trends in retrieved aerosol properties using trajectory analysis. In the plume core, the retrieved optical and microphysical properties show patterns consistent with findings from in situ field campaigns and laboratory studies. After about an hour of aging, the single-scattering albedo (SSA) increases from 0.92 to 0.96 and the fine-mode effective radius doubles, indicating that fresh smoke near the source is dominated by small, strongly absorbing particles, whereas aging produces larger, less absorbing particles, reflecting both microphysical growth and chemical processing during transport. In addition, smoke closest to the fire hotspot shows a decrease in SSA from 0.95 to 0.90 from mid-morning to early afternoon, consistent with intensified midday combustion producing more strongly absorbing carbonaceous aerosol. Correspondingly, the contribution from brown carbon aerosols increases from ~30% of total AOD in mid-morning to ~70% in early afternoon. Although these results are based on a single case study, the good agreement with AERONET provides confidence in the fidelity of the new aerosol retrievals. Future work will apply this retrieval method and accompanying trajectory analysis to a broader set of biomass-burning events, particularly those with supporting in situ measurements, to improve understanding of smoke and fire evolution and to provide quantitative properties for parameterization development.