The Research Triangle Park particulate matter panel study: modeling ambient source contribution to personal and residential PM mass concentrations

The Research Triangle Park (RTP) Particulate Matter (PM) Panel Study represented a 1-year investigation of personal, residential and ambient PM mass concentrations across distances as large as 70 km in central North Carolina. One of the primary goals of this effort was to estimate ambient PM 2.5 contributions to personal and indoor residential PM mass concentrations. Analyses indicated that data from the two distinct non-smoking subject populations totaling 38 individuals and 37 residences could be pooled. This resulted in nearly 800 data points for each variable. A total of 55 measurements believed to have been potentially influenced by personal or residential exposure to passive environmental tobacco smoke were not included in the analysis database. Variables to be examined included C ig (concentration of indoor generated PM), E ig (personal exposure to indoor generated PM), F inf (ambient PM infiltration factor), and F pex (personal exposure to PM of ambient origin factor). Daily air exchange rates (AER) were measured and statistical modeling to derive estimates of particle penetration ( P) and particle deposition ( k) factors was performed. Seasonality, cohort grouping, participant or combinations of these variables were determined not to be significant influences in estimating group infiltration factors. The mean (±std) mixed model slope estimates were AER=0.72±0.63, P=0.72±0.21, k=0.42±0.19, and F inf=0.45±0.21. These variables were then used in a number of mixed effects models having varying features of single, random or fixed intercepts and/or slopes to determine the most appropriate means of estimating ambient source contributions to personal and residential settings. A mixed model slope for F pex (±SE) was 0.47±0.07 using the model with the highest degree of fit.