Exploring Uncertainties in Model-based Estimates of Air Pollution Impacts from Deforestation Fires in Northern South America

Abstract Description: Deforestation and biomass burning in the Amazonian region of northern South America substantially impact regional air pollution. Accurate characterizations of fire emissions are essential to improve assessment of the impact of forest fires on air quality and identify potential co-benefits of deforestation policies. Satellite-based estimates have been widely used to predict fire-related emissions and their air quality impacts. However, large uncertainties exist in these fire emission estimates across different fire inventories due to differences in the methods used to calculate burn areas, fuel types, fire intensity, and emission factors. In addition, spatial resolution of the model and plume rise treatments can be significant sources of uncertainty in model predictions of pollutant concentrations. In this study, we evaluate uncertainties associated with four commonly used fire inventories: FINNv1.5, FINNv2.5, GFEDv4.1s, and GFEDv5, focusing on their effects on pollutant concentration predictions through WRF-Chem simulations. We apply varying levels of resolutions in WRF-Chem simulations to explore the uncertainty associated with model resolution and attempt to identify appropriate grid size for smoke simulations. We also explore different plume rise approaches in WRF-Chem to weight associated uncertainties. With this study, we aim to evaluate major sources of uncertainty in model-based estimates of air pollution from deforestation activities and more accurately quantify the potential benefits of reducing fires in northern South America.