Modeling the health benefits by transitioning from conventional to zero-emissions appliances standards in the United States (Wellcome-ZEAS) Results: Air Quality

This UNC dataset contains monthly and pseudo-annual (4-month) mean Community Multiscale Air Quality Model (CMAQ) v5.4+ outputs for the continental U.S. (12US1, 12 km). It includes (1) base pollutant concentrations (ACONC) and (2) first order Decoupled Direct Method (DDM-3D) sensitivities (ASENS) of PM2.5 (ATOTIJ), O3 (including MDA8: O3_8hr), and NO2 with respect to precursor emissions from the building sector, stratified into six fuel-based groups (G1–G6) defined below. The four modeled months (January, April, July, October 2021) are averaged to produce a pseudo-annual mean. Related model and methods references CMAQ v5.4+ (Wyat Appel et al., 2018) CB6r5ae7 chemical mechanism (Luecken et al., 2019) - CB6r5 gas-phase chemical mechanism coupled with the AERO7 aerosol module DDM-3D (Dunker, 1984; Dunker et al., 2002; Koo et al., 2007; Napelenok et al., 2006; Napelenok et al., 2008) Domain, resolution, and coverage Modeling domain: 12US1 (CONUS) Horizontal resolution: 12 km × 12 km Grid size: 455 × 299 cells (CONUS coverage) Time coverage: January, April, July, October 2021 (monthly means) Pseudo-annual coverage: average of the four months above Group definitions (building sector fuel groups) The 2021 Emissions Modeling Platform (EMP), developed based on the 2020 National Emissions Inventory (NEI) released in spring 2023, was used in this study. The inventory includes updates to better represent 2021 emissions conditions. We have processed the building sector emissions from 2021 Emission Modeling Platform hosted by CMAS Datawarehouse (https://registry.opendata.aws/cmas-data-warehouse) through Sparse Matrix Operator Kernal Emissions (SMOKE) v 5.2 (https://zenodo.org/records/17671730#libraryItemId=18432936) for 6 sub-groups for the building sector as listed below: G1: Commercial/Institutional Oil (oil + fossil fuel) G2: Commercial/Institutional Gas (natural gas) G3: Commercial/Institutional Coal G4: Residential Oil (oil + fossil fuel) G5: Residential Gas (natural gas) G6: Residential Coal Contents (what’s included) 1) Base concentrations (ACONC, monthly + pseudo-annual) NetCDF files containing monthly mean concentrations (and pseudo-annual mean).File names: MONTHLY_Average_ACONC_Wellcome_v54_gcc_12US1_202101.nc MONTHLY_Average_ACONC_Wellcome_v54_gcc_12US1_202104.nc MONTHLY_Average_ACONC_Wellcome_v54_gcc_12US1_202107.nc MONTHLY_Average_ACONC_Wellcome_v54_gcc_12US1_202110.nc ANNUAL_Average_ACONC_Wellcome_v54_gcc_12US1_2021.nc (pseudo-annual) ACONC variables:O3_8hr, O3, NO2, ATOTIJ, NH3, ASO4IJ, ANO3IJ, ANH4IJ, SO2, NO, NOX, CO, APOCIJ, AECIJNotes: O3_8hr = monthly mean of MDA8 O₃ (daily max of 8-hr running mean) NO2 = monthly mean NO₂ ATOTIJ = monthly mean PM2.5 2) DDM sensitivities (ASENS, monthly + pseudo-annual; by group) NetCDF files containing monthly mean DDM sensitivities of PM2.5, O3, and NO2 to precursor emissions for each fuel group (G1–G6), plus pseudo-annual mean sensitivities. File names: MONTHLY_Average_ASENS_Wellcome_G[G]_12US1_v54_DDM3D_2021[MM].nc Where: · Group Index [G] = 1, 2, 3, 4, 5, 6 · Month Index [MM] = 01, 04, 07, 10 (January, April, July, October) ANNUAL_Average_ASENS_Wellcome_G[G]_12US1_v54_DDM3D_2021.nc Where: · Group Index [G] = 1, 2, 3, 4, 5, 6 ASENS variables (Sensitivity variables): O3: O3_NOX_8hr, O3_VOC_8hr, O3_NOX, O3_VOC NO2: NO2_NOX PM2.5: ATOTIJ_NOX, ATOTIJ_VOC, ATOTIJ_PPM, ATOTIJ_SO2, ATOTIJ_NH3 Interpretation: Each variable represents the first-order sensitivity of the listed pollutant metric to the listed precursor emissions, for a specific group and month (or pseudo-annual mean). 3) MD1_Ave_NO2 “daily-maximum aligned NO₂ sensitivity” (monthly + pseudo-annual, by group) These NetCDF files represent a NO₂ sensitivity metric aligned with the daily maximum NO₂ hour at each grid cell: For each day and grid cell, identify the hour when NO₂ is maximum Extract the sensitivity at that hour/grid cell (paired consistently with time/bulk) Average over days to obtain a monthly mean (and pseudo-annual mean) File names: MONTHLY_Average_MD1_Ave_NO2_Wellcome_G[G]_v54_gcc_12US1_2021[MM].nc Where: Group Index [G] = 1, 2, 3, 4, 5, 6 · Month Index [MM] = 01, 04, 07, 10 (January, April, July, October) ANNUAL_Average_MD1_Ave_NO2_Wellcome_G[G]_v54_gcc_12US1_2021.nc Sensitivity variable: NO2_DMAX_G[G](Group Index) Where: · Group Index [G] = 1, 2, 3, 4, 5, 6 Processing summary CMAQ v5.4+ CB6r5 was instrumented with DDM-3D to compute grid-specific first-order sensitivities. A total of 24 DDM simulations were performed for 2021: 6 groups × 4 months (Jan/Apr/Jul/Oct). For each month: computed monthly mean ACONC and monthly mean ASENS. Pseudo-annual means are computed as the average of the four monthly means (Jan/Apr/Jul/Oct). Folder structure (After untar of the tar files) ACONC/ MONTHLY_Average_ACONC_*.nc ANNUAL_Average_ACONC_*.nc ASENS/ MONTHLY_Average_ASENS_*_G[1..6]_*.nc ANNUAL_Average_ASENS_*_G[1..6]_*.nc MONTHLY_Average_MD1_Ave_NO2_*_G[1..6]_*.nc ANNUAL_Average_MD1_Ave_NO2_*_G[1..6]_*.nc File format and software Format: netCDF Typical tools to read: VERDI, Panoply, ncdump/ncview, Python (xarray, netCDF4), R (ncdf4) Model: CMAQ v5.4+ with DDM-3D Keywords CMAQ; DDM-3D; air quality modeling; building emissions; PM2.5; ozone; nitrogen dioxide; sensitivities; CB6r5; CONUS; 12US1; pseudo-annual; source apportionment; emissions perturbation; netCDF Notes “Annual” files represent a pseudo-annual mean (average of Jan/Apr/Jul/Oct 2021), not a full 12-month mean. Variable units and grid/projection metadata are stored in-file netCDF attributes using the I/O API convention.