Spatiotemporal comparison of highly-resolved emissions and concentrations of carbon dioxide and criteria pollutants in Salt Lake City, Utah for health and policy applications: a chronic obstructive pulmonary disease case study

This study addresses the need for robust highly-resolved emissions and concentration data required for planning purposes and policy development aimed at managing pollutant sources. Adverse health effects resulting from urban pollution exposure are dependent on proximity to emission sources and atmospheric mixing, necessitating models with high spatial and temporal resolution. This study compares the increased risk of health complications when patients are exposed to short term high-levels of air pollution vs. longer term exposure to lower levels of air pollution. As urban emission sources co-emit carbon dioxide (CO₂) and criteria pollutants (CAPs), efforts to reduce specific pollutants would synergistically reduce others. We present emissions inventories and modeled concentrations for CO₂ and CAPs: carbon monoxide (CO), lead (Pb), nitrogen oxides (NO_x), particulate matter (PM_2.5 and PM₁₀), and sulfur oxides (SO_x) for Salt Lake County, Utah. We compare the resulting concentrations against stationary and mobile measurement data and present a systematic quantification of uncertainties. The emissions inventory for CO₂ is based on the Hestia emissions data inventory that resolves emissions at an hourly, building and road link resolution as well as hourly gridded emissions with a 0.002^o x 0.002^o spatial resolution. Two methods for deriving criteria pollutant emission inventories were compared. One was constructed using methods similar to Hestia but downscales total emissions based on the 2011 National Emissions Inventory (NEI). The other used Emission Modeling Clearinghouse spatial and temporal surrogates to downscale the NEI data from annual and county-level resolution to hourly and 0.002^o x 0.002^o grid cells. The gridded emissions from both criteria pollutant methods were compared against the Hestia CO₂ gridded data to characterize spatial similarities and differences between them. Correlations were calculated at multiple scales of aggregation. The CALPUFF dispersion model was used to transport emissions and estimate air pollutant concentrations at an hourly 0.002^o x 0.002^o resolution. The resulting concentrations were spatially compared in the same manner as the emissions. Modeled results were compared against stationary measurements and from equipment mounted atop a light rail car in the Salt Lake City area. The comparison between both approaches to emissions estimation and resulting concentrations highlights spatial locations and hours of high variability and uncertainty. We used the electronic medical record of an integrated hospital system based in Utah, Intermountain Healthcare, to identify a cohort of patients with Chronic Obstructive Pulmonary Disease (COPD) who were seen between 2009-2014. We determined patient demographics as well as comorbidity data and healthcare utilization. To determine the approximate air pollution dose and time exposure, we used each patient's home and work address in conjunction with the transported emissions. Multivariate analysis adjusting for patient demographics, comorbidities and severity of COPD was performed to determine association between air pollution exposure and the risk of hospitalization or emergency department (ED) visit for COPD exacerbation and an equivalency estimate for air pollution exposure was developed. We noted associations with air pollution levels for each pollutant and hospitalizations and ED visits for COPD and other patient comorbidities. We also present an equivalency estimate for dose of air pollution exposure and health outcomes. This analysis compares the increased risk of health complications when patients are exposed to short term high-levels of air pollution vs. longer term exposure to lower levels of air pollution. These findings highlight pollutant emissions and exposures spatial and temporal heterogeneity and associated health effects.