Impacts of the 2023 Canadian Fires on US Air Quality Simulated by NOAA UFS-AQM with Aerosol Data Assimilation

The National Air Quality Forecasting Capability (NAQFC) is expected to undergo a major upgrade using NOAA’s Unified Forecast System (UFS) framework, incorporating a regional online air quality model (AQM). This model includes the Finite Volume Cubed-Sphere Dynamic Core (FV3), the Common Community Physics Package (CCPP), and the chemical mechanism from the Environmental Protection Agency (EPA)’s Community Multiscale Air Quality (CMAQ) model. Anthropogenic emissions are based on the National Emission Inventory (NEI) Collaborative with a base year of 2016. Fire emissions are derived from satellite-retrieved fire radiation power (FRP), referred to as Regional Hourly Advanced Baseline Imager (ABI) and Visible Infrared Imaging Radiometer Suite (VIIRS) Emissions, or RAVE. To enhance the initial conditions of the UFS-AQM, a chemical data assimilation (DA) system has been developed under the Joint Effort for Data-Assimilation Integration (JEDI). This 3D-Var system incorporates surface in-situ data assimilation using AIRNow data and AOD assimilation using VIIRS AOD (both S-NPP and NOAA-20). The corresponding B-matrix, which includes horizontal length scales and background error standard deviation, has been developed using the Hollingsworth-Lönnberg (H-L) (1986) method. This JEDI-based chemical DA system has been applied to constrain the UFS-AQM over an expanded North American domain at 13 km resolution, with a case study involving the Alberta wildfire event in May 2023. The assimilation of in-situ AIRNow data consistently improves the model's predictions. However, due to the relatively sparse coverage of AIRNow stations over the wildfire source regions in Canada, the effectiveness of in-situ assimilation is limited. In contrast, the VIIRS AOD DA shows a stronger impact in those areas. One issue of the AOD DA is that its increments may not always align with the AIRNow DA increments due to uncertainties related to aerosol types, sizes, vertical distributions, forward operators, and retrieval issues. Despite these challenges, in most scenarios, especially during intense fire events, AOD assimilations continue to enhance predictions as the extensive spatial coverage of satellite data over remote regions compensates for the lack of dense in-situ measurements.