United States Environmental Protection Agency	Office of Research and Development

National Exposure Research Laboratory
FY02 Research Abstract

Government Performance Results Act (GPRA) Goal 1
APM22

Significant Research Findings:

A Modeling Analysis of Exposure of a Susceptible Subpopulation

to PM of Ambient Origin

Understanding human exposures to environmental pollutants is critical for the
U.S. Environmental Protection Agency (EPA) to assess and manage human
health risks. For particulate matter (PM), air pollution epidemiology studies
have found statistically significant associations between PM concentrations
and acute and chronic health outcomes. These epidemiology studies use PM
measurements from stationary air monitoring sites within a community as a
surrogate for personal exposure levels for the population. To better
understand the relationship between personal exposures and stationary
community measurements for PM, National Exposure Research Laboratory
(NERL) has developed a research program to improve the data available for
personal exposure to PM and to develop exposure models that predict
population exposures to PM. Measurements of personal exposure to PM
include contributions from both PM of ambient origin and indoor sources of
PM. Modeling analyses are needed to estimate personal exposure to PM of
ambient origin from the personal exposure measurement data. In addition,
certain populations, such as the elderly or asthmatics, may be more at risk for
health effects due to PM exposure. This effort addresses Research Area 1 of
the National Research Council's Research Priorities for Airborne Particulate
Matter.

Research	NERL is currently refining and evaluating a population exposure model for

Approach	PM, called the Stochastic Human Exposure and Dose Simulation (SHEDS-

PM) model. The SHEDS-PM model estimates the population distribution of
PM exposures by randomly sampling from distributions of ambient PM
concentrations and exposure factors to estimate the distribution of PM
exposures in various locations that people spend time in (such as in homes,
in vehicles, outdoors, etc.). SHEDS-PM model output also includes
estimates of the contribution from both PM of ambient origin and indoor
sources of PM. An application of the SHEDS-PM model was implemented
using measurement data from the NERL RTP PM Panel Study. The study
included daily measurements of PM2 5 personal exposure, as well as indoor
residential, outdoor residential, and ambient (central site) PM2 5

Scientific
Problem and
Policy Issues

National Exposure Research Laboratory — November 2002


-------
concentrations, during a seven-day study period for each participant within
each of the four seasons from June 2000 - May 2001. Daily measurements of
ambient PM2 5 concentrations and residential air exchange rates and human
activity pattern data obtained from daily diaries for each participant were
used as input to the model. Additional model inputs, such as demographic
data and residential characteristics, were matched to the NERL RTP PM
Panel Study population.

Results and	The distribution of daily average total PM2 5 exposures predicted by the

Implications SHEDS-PM model compared well with the measurements of personal PM2 5
exposures from the RTP PM Panel Study. The median total PM2 5 exposure
(50th percentile) for the SHEDS-PM predictions was 20 |ig/m3 compared to
18 |ig/m3 for the personal PM25 exposure measurements. The median
personal exposure to ambient PM2 5 predicted by the model was 9 |ig/m3 for
the study population and ranged from 2-33 |ig/m\

The model results were compared to estimates of ambient PM2 5 exposures
derived from analysis of 24-hr integrated personal PM2 5 measurement data
collected during the study. This comparison was used to evaluate the
SHEDS-PM model prediction of population exposures to ambient PM2 5 and
indicated that the model results were generally consistent with the estimates
derived from measurement data.

This research effort is part of a ongoing research program to develop a
scientifically robust exposure modeling system to analyze the relationship
between PM sources, ambient air PM concentrations, and personal exposures
and dose. The development of an improved, user-friendly PM exposure
modeling system will give EPA better capability to set National Ambient Air
Quality Standards (NAAQS) for PM. Case-study application of the
modeling system will identify the major factors influencing PM exposures,
including the relative contributions of outdoor and indoor PM concentrations
to personal exposure and dose, as well as the effects of human activities,
housing characteristics and seasonal factors (i.e. meteorology). These
applications will also help NERL's human exposure measurement program
prioritize and direct studies toward the data needed for reducing uncertainty
in the model predictions. The greater understanding of personal exposure to
PM obtained through application of this source-to-dose modeling system will
also be valuable for improving the power of epidemiological studies of the
impact of PM pollution on mortality and morbidity.

This research project included collaboration between NERL's Human
Exposure Measurements and Exposure Modeling Programs. The NERL RTP
PM Panel Study was conducted through Contract #68-D-99-012 to the
Research Triangle Institute and Assistance Agreement #CR-828186-01-0 to
Shaw University. Assistance with the development of the SHEDS model

Research
Collaboration
and Publications

National Exposure Research Laboratory — November 2002


-------
program and application to the NERL RTP PM Panel Study was provided by
ManTech Environmental Technology, Inc. under Contract #68-D-00-206.

This research has recently been presented at an international conference:
Burke, J., A. Rea, J. Suggs, R. Williams, J. Xue, H. Ozkaynak, "Ambient Particulate Matter
Exposures: A Comparison of SHEDS-PM Exposure Model Predictions and
Estimates Derived from Measurements Collected During NERL's RTP PM
Panel Study." Presented at the 12th Conference of the International Society of
Exposure Analysis (ISEA), Vancouver, BC, August 11-15, 2002.

Future Research Future research will include further refinements and evaluation of the

SHEDS-PM exposure model. In addition, the program's focus will transition
to the various components of PM associated with health effects. A refined
SHEDS-PM model will be used to predict exposure and dose (the
magnitude, frequency and duration of) to various chemical species and/or
size components of PM for the general population and susceptible sub-
populations. Activities will focus on enhancing the model to include recent
scientific advancements and developing new model inputs and algorithms as
needed for case study applications.

Questions and inquiries on NERL's human exposure and dose modeling
research can be directed to:

Haluk Ozkaynak, Ph.D.

US EPA Office of Research and Development
National Exposure Research Laboratory
Ariel Rios Building (MC-8601-D)

1200 Pennsylvania Avenue, NW
Washington, DC 20460

Phone: 202/564-1531
E-mail: ozkaynak.haluk@epa.gov

Contacts for

Additional

Information

National Exposure Research Laboratory — November 2002


-------