Breathing Zone Exposure Assessment - Aerosols: Measurement, Dosimetry, and Health Effects - page 33

beneits (Sarnat et al., 2007). Rodes et al. (2010) observed that accurate assessments for the most
exposed required BZ methodologies (compared with ixed-location surrogates) in a general popula-
tion study in metro Detroit, MI. Clearly, if understanding the levels deining the most exposed is an
important goal for the exposure study, BZ level metrics must be given serious consideration.
An early study by Sherwood (1966) of radioactive aerosol concentrations in a nuclear materials
laboratory showed that collecting data from workers using personal exposure monitors (PEMs) in
the BZ provided levels of beta activity that exceeded single-location, room-average, microenviron-
mental* exposure measurement (MEM) values by a mean ratio of 7.7 (mean BZ-to-room ratio). The
primary explanation for the substantially elevated personal levels was the workers' closer proximity
to the beta sources than the ixed-location, room monitor. In a similar study for both alpha and beta
activity, Stevens (1969) similarly reported much higher BZ PEM levels compared with concurrent
MEM levels, ranging from ratios of 2 to 3 when workers were near scattered, multiple point sources
in the same room, and ratios from 5 to 15 when they were in close proximity to a single-room source.
Again the rationale was suggested to be driven primarily by the composite point source-to-worker
proximity during the 8 h exposure interval. Parker et al. (1990) studied the release of polydisperse
0.5 μm particles into a test room and reported that personal exposure measurements at the lapel of a
manikin were 5-10 times higher 0.5 m from a point source than room average samplers a few meters
away. They also reported that real-time measurements at the lapel and at the mouth, separated by
only 0.3 m, showed very poor correlation for the concentration luctuations. The potential for point
sources to provide nonuniform concentrations in occupational and nonoccupational indoor settings
(respectively) was discussed by Nicas (1996) and Furtaw et al. (1996), who provided models to esti-
mate the inluence of the room ventilation system on the room concentrations. Rodes et al. (1991)
showed that the ratios of BZ to ixed-location indoor measurements for residential settings differed
signiicantly from those in workplace environments, attributed primarily to the stronger workplace
localized sources. However, they reported that the residential 90th percentile BZ exposures were
still typically as much as four times higher than ixed metrics at the cohort median levels.
The limited ability of some particle sizes to readily penetrate to the indoor environment (e.g.,
Thornburg et al., 2001) can play a key role in deining the representativeness of outdoor sampling
methods to estimate exposures. Without utilizing personal-level exposure characterization, the times
spent outdoors and indoors can only be estimated from outdoor monitoring, or exposure models that
combine the two. Burke et al. (2001) developed the U.S. EPA CHAD exposure model to establish the
exposure distributions for PM 2.5 from ambient monitoring data for a cohort in Philadelphia, PA, and
observed that while the model reasonably estimated the interquartile range (IQR), the uncertainties
became substantial and excess at the 90th percentile. Accurate deinition of the most exposed would
require BZ exposure assessments. Koistinen et al. (2004) reported that data from ambient monitor-
ing locations signiicantly overestimate the contributions of outdoor trafic and long-range transport
aerosol compared with personal exposures, and underestimate the contributions from indoor sources.
2.2.1.2 Characterizing the “Most Exposed”
Rodes et al. (1991) observed that activity pattern information during the integration period, in addition
to source proximity, was critical to understanding nonoccupational personal exposures for aerosols.
They suggested that the PEM-to-MEM ratios for residential exposure settings can be signiicantly
different than occupational, due to typically weaker and more dispersed residential point sources
and signiicant periods that include minimal or no sources. They reported median PEM-to-MEM
ratios for nonoccupational aerosol exposure studies ranging from approximately 1.5 to 2.0 (much
lower than typical occupational ratios), but still reported that the 90th or 95th percentile—“most
exposed”—portions of the residential study populations exhibited ratios exceeding 4.0. Wallace
et al. (2006), using a personal real-time aerosol nephelometer, identiied a wide variety of sources
* Microenvironment is deined here to mean a localized, contained volume that generally deines the concentration—most
often approximately bounded by the perimeter of the room when indoors.