Fine-Scale Modeling of Individual Exposures to Ambient PM2.5, EC, NOx, CO for the Coronary Artery Disease and Environmental Exposure (CADEE) Study

Air pollution epidemiological studies often use outdoor concentrations from central-site monitors as exposure surrogates, which can induce measurement error. The goal of this study was to improve exposure assessments of ambient fine particulate matter (PM2.5), elemental carbon (EC), nitrogen oxides (NOx), and carbon monoxide (CO) for a repeated measurements study with 15 individuals with coronary artery disease in central North Carolina called the Coronary Artery Disease and Environmental Exposure (CADEE) Study. We developed a fine-scale exposure modeling approach to determine five tiers of individual-level exposure metrics for PM2.5, EC, NOx, CO using outdoor concentrations, on-road vehicle emissions, weather, home building characteristics, time-locations, and time-activities. We linked an urban-scale air quality model, residential air exchange rate model, building infiltration model, global positioning system (GPS)-based microenvironment model, and accelerometer-based inhaled ventilation model to determine residential outdoor concentrations (Cout_home, Tier 1), residential indoor concentrations (Cin_home, Tier 2), personal outdoor concentrations (Cout_personal, Tier 3), exposures (E, Tier 4), and inhaled doses (D, Tier 5). We applied the fine-scale exposure model to determine daily 24-h average PM2.5, EC, NOx, CO exposure metrics (Tiers 1–5) for 720 participant-days across the 25 months of CADEE. Daily modeled metrics showed considerable temporal and home-to-home variability of Cout_home and Cin_home (Tiers 1–2) and person-to-person variability of Cout_personal, E, and D (Tiers 3–5). Our study demonstrates the ability to apply an urban-scale air quality model with an individual-level exposure model to determine multiple tiers of exposure metrics for an epidemiological study, in support of improving health risk assessments.

The AER has two parameters ( s and w ) and five inputs ( leak , in , out , , and V). Parameters s and w were set to literature-reported values based on house-specific information on house height (number of stories) and local wind sheltering (Supplementary Material, Tables S1-S3). Using home addresses, the number of stories and local wind sheltering were determined from satellite and street-level images in Google Earth (version 7.1.7.2606; Google, Mountain View, CA, USA). We used house numbers visible in street-level images to verify the participant homes. The number of stories was verified from online county and real estate databases of property records (Zillow, Seattle, WA, USA; Trulia, San Francisco, CA, USA). To determine V, we multiplied floor area by a ceiling height of 2.44 m (8 ft). The floor area was obtained from the online county and real estate databases.
We determined out and (10 m elevation) from hourly measurements at Raleigh Durham Airport in Morrisville, NC, USA. We calculated 24 h average out and time-matched to the 24 h average PM2.5 measurements. We determined in from daily values from the daily participant questionnaires.
We estimate leak with a literature-reported leakage area model [1,2]. The leak is calculated as leak = [S1] (1) where is the normalized leakage and is the normalization factor. The is predicted from year of construction Ybuilt and floor area Afloor as described by NL=exp (β0+β1Ybuilt +β2Afloor) [S2] (2) where β0, β1, and β2 are regression parameters. The NF is defined as where is the building height. We set to the number of stories multiplied by a story height of 2.5 m and adding a roof height of 0.5 m (Breen et al., 2010). The floor and Ybuilt were obtained from online county and real estate databases of property records as described above.
For the airflow from natural ventilation nat can be calculated as: where nat,wind and nat,stack are the airflows from the wind and stack effects, respectively. The nat,wind is defined as: where is the effectiveness of the openings, and the nat is the area of the inlet openings. Using literature-reported values, we set to 0.3 and nat to one-half of the total area of window and door openings (Breen et al. 2010). The daily participant questionnaires were used to determine number and duration that windows and doors were opened. Window and door opening areas were not collected in CADEE. For windows, we set nat to one-half of the literature-reported value of 619 cm 2 , which is the median daily total window opening area for homes in the same region of central NC as DEPS (Breen et al. 2010). For doors, we set nat to one-half of 3600 cm 2 . The nat,stack is defined as: where D is the discharge coefficient for the openings, g is the gravitational acceleration, Δ NPL is the height from midpoint of lower window opening to the neutral pressure level (NPL) of the building, and { in , out } is the maximum value between in and out . Using literature-reported values, we set D to 0.65, midpoint of lower window opening to 0.91 m, and NPL to one-half of H [1].
For the days with operating window fans, the airflow (Qtotal) was calculated as follows: where Qbal and Qunbal are balanced and unbalanced flow rate respectively, Qleak is the flow from leakage, and Qnat is the flow from natural ventilation. The daily participant questionnaires were used to determine number and duration that window fans were operated. Since whether the window fan system is balanced (i.e. pair of intake and exhaust fan) or unbalanced (i.e. a single intake or exhaust fan) was not recorded, we assume an unbalanced system for all houses with window fan operating (Qbal=0). Qunbal was set at 600 ft 3 /min for each window fan, which is the mid-range value for medium-size window fans (range: 300-900 ft 3 /min) [3].

Sensitivity Analysis
For the sensitivity analysis of time spent in different microenvironments, we determined exposures (E) as defined by E = Fpex Cout (8) where Fpex is the personal exposure factor (dimensionless), and Cout is the outdoor concentration. The Fpex is defined by where fin_home, fin_work, fin_other_bldg, fin_vehicle, fout are the fraction of day spent in indoors at home, work, other buildings; inside vehicles; outdoors; respectively; and Finf_home, Finf_other_bldg, Fin_vehicle, Fout are the infiltration (i.e., attenuation) factors for home, other buildings including work, vehicles, outdoors, respectively. The Finf_home is defined by a steady-state mass-balance infiltration model described by Finf_home = (P AER)/(AER + kr) (9) where P is the penetration coefficient (dimensionless), AER is the air exchange rate (h −1 ), and kr is the indoor removal rate (h −1 ). We set AER to the median value (0.5 h −1 ) measured from homes in the same region of North Carolina as CADEE homes [4]. We used the same parameter as described in the main paper. For PM2.5, P and kr were previously estimated from homes in the same region of NC as CADEE (P = 0.84, kr = 0.21 h −1 ) [4,5]. For EC, NOx, CO, P and kr were obtained from literature-reported values (P = 0.98, 1.00, 1.00; kr = 0.29, 0.5, 0 h −1 ; respectively) [6][7][8]. This yields For the five different microenvironments that participants can spend their time, the infiltration factors (Finf_home, Finf_other_bldg, Finf_vehicle, Fout) vary by a range (max-min) and factor (max/min) of (0.44-1.00; factor of 2.3) for PM2.5, (0.44-1.00; factor of 2.3) for EC, (0.50-1.00; factor of 2.0) for NOx, (1.00-1.00; factor of 1.0) for CO. Therefore, changes in the time spent in ME with substantially different infiltration factors (e.g., indoors versus outdoors) can produce substantial changes in the exposures for PM2.5, EC, NOx., but have little or no effect on exposures to CO.