Abstract Description: On February 7, 2024, EPA promulgated a new 9.0 µg/m3 annual PM2.5 National Ambient Air Quality Standard (NAAQS). EPA will likely designate PM2.5 nonattainment areas in February 2026 with attainment demonstration SIPs likely due in Fall 2027. PM2.5 consists of primary emitted PM2.5 species from sources such as biomass burning, cooking, combustion, fugitive dust, and brake and tire wear as well as secondary formed PM2.5 sulfate, ammonium nitrate, and secondary organic aerosol (SOA) that are formed in the atmosphere from primary emitted gaseous species of SO2, NOX and VOCs, respectively. To model these secondary species requires a photochemical grid model (PGM) with gas-phase and aqueous-phase chemical mechanisms and an aerosol thermodynamic module. However, many of the monitoring sites that fail to attain the new 2024 PM2.5 NAAQS are roadway sites and/or impacted by primary PM2.5 emissions from local sources. In theory, a PGM could use a very fine grid resolution (10s of meters) to resolve the local source impacts at nearby monitors. However, in practice PGM modeling using an extremely fine grid resolution to address local source impacts as well as including urban and regional sources would be prohibitive for addressing the annual PM2.5 NAAQS. This paper discusses the development of a hybrid PGM-plume modeling approach for address near-source as well as far-field PM2.5 impacts.
The Comprehensive Air-quality Model with extensions (CAMx) is used with a multiscale two-way nested grid structure (e.g., 12/4/1.33/0.44 km) with the coarsest grid (e.g., 12 km) resolution domain covering the U.S. and the finest grid (e.g., 444 m) including the monitor and local source(s). The local source primary PM emissions are tagged for treatment by the CAMx Particulate Source Apportionment Technology (PSAT). After the CAMx annual run, PSAT is used to remove the primary PM contributions from the local sources from the CAMx output. A plume (e.g., AERMOD) or puff (e.g., CALPUFF) model is then used to model the local source primary PM emission impacts at the monitoring site and other locations. The CAMx PGM and plume/puff model results are then combined to estimate future year PM2.5 concentrations and the hybrid modeling system can be used to evaluate control measures on the urban/region scale or from the local sources. This paper demonstrates the hybrid PGM-plume modeling approach for a PM2.5 SIP that modeled the Monongahela River Valley in Allegheny County, Pennsylvania where large coke and steel operations have impacts on local monitoring sites.
