a,6 NEIGHBORHOOD SCALE MODELING OF PM« AND AIR TOXICS CONCENTRATION DISTRIBUTIONS
TO DRIVE HUMAN EXPOSURE MODELS
Jason Ching', Avraham Lacser2, Tanya L, Ode', Jerry Herwehe3, and Daewon Byuri4
'Atmospheric Sciences Modeling Division, NOAA ARL, Research Triangle Park, North Carolina
On assignment to the National Exposure Research Laboratory, U.S. EPA
'Israel Institute for Biological Research, Ness Zksna, Israel
3 Atmospheric Turbulence and Diffusion Division, NOAA ARl, Oak Ridge, Tennessee
¦"Department of Geosctences, University of Houston, Houston, Texas
1,	INTRODUCTION
Air quality (AQ) simulation models provide a basis
for implementing the National Ambient Air Quality
Standards (NAAQS) and are a tool for performing risk-
based assessments and for developing environmental
management strategies. Fine particulate matter (PM1.5).
its constituents and size and number distribution, as well
as airborne toxic pollutants ("air toxics") have
characteristically different degrees of spaliat and
temporal variability especially in urban areas and in
different geographical-climatic regimes. In this study,
we explore the specific role of AQ models as a means to
drive human exposure models (Burke et at. 2001) and to
address situations in which pollutants exhibit high
spatial and temporal variability. We seek a capability to
capture the resolved-scale concentration fields and to
provide measures of sub-grid-scale variability in
concentration distributions that impact human
exposures. This modeling approach is meant to
enhance and complement the more limited data from
central site monitoring networks to provide the
concentration fields at high temporal and spatial
resolutions. By providing further information on
concentration variability at sub-grid scales, we complete
the requirements needed for exposure assessments.
The various elements of this modeling approach and
some of their specific modeling issues are described
below.
2.	APPROACHES AND SCOPE
Our goal is to simulate the chemical and physical
attributes of PM25 and air toxics such that the output
fields can be used to address both the traditional air
quality management needs and to perform modeling of
human exposure to various air pollutants. For this
effort, we begin with the Models-3/Cammunity
Multiscale Air Quality (CMAQ) modeling system. CMAQ
has a 'one-atmosphere", multiple-pollutant capability
(Byun and Ching, 1999) and thus is well suited to
handle the photochemistry and other important
atmospheric processes that impact the fate and
transport of air pollutants. We anticipate that different
pollutants mil have characteristic spatial scales that the
CMAQ multiscale capability will capture. Strategies to
reduce human exposures can then ba developed using
emission projections.
*Corresponding author address: Jason Ching,
AMD/NERL/USEPA, (MD:D243-03). RTP, NC 27711;
email: ching.Jason@epa.gov
Next, human population exposure assessments will
require an urban modeling emphasis, CMAQ and its
pre-processing models will be modified with appropriate
methodologies to handle urban morphological features
and vegetation coverage. The added complexity of
dispersion of pollution sources distributed within urban
areas introduces additional complexity to modeling the
concentration variability, an important consideration for
exposure assessments. Finally, linkage to human
exposure models will require information and modeling
of human activity ol the population and of pollution
exchange between ambient and the various
microenvironments {Ching et al. 2000). Modeling
linkages lo human exposure modeling wilt not be
discussed here.
3. IMPLEMENTATION
3,1 Resolved-Scale Modeling
The simulation of the complex dispersion and
transport at the neighborhood scales (-1 km ho-lzontal
grid spacing) is an important first step. We developed
and implemented an urban canopy parameterization for
the Pennsylvania Stale University, National Center for
Atmospheric Research (PSU'NCAR) Mesoscale Model
(MM5) (Grell et al. 1994), which is used to provide
meteorological fields for the chemistry-transport model
in CMAQ. The urban canopy parameterization (Lacser
and Otte. 2002) combines urban building morphologies
with urban land use categories to produce
meteorological fields that include the heterogeneous
effects of the urban areas at that scale. This urban
canopy parameterization is based on the drag approach
(e.g.. Brown 2000), and it is applied to grid cells in MM5
that have a non-zero fraction of urban land use. The
horizontal components of the momentum equations and
the turbulent kinetic energy (TKE) equation were
modified to account for the area average effect of the
sub-grid urban elements. The parameterizations
produced significant differences in the mean and
turbulent fields within the urban canopy, especially in
areas characterized by high density of tall buildings,
MM5 with the urban canopy parameterization simulated
vertical profiles of wind speed and TKE that were highly
consistent with results from wind tunnel studies. These
results illustrate the importance of accounting for the
urban morphological structures in modeling the flow in
urban areas. It provides a basis for more accurately
resolving the magnitude and spatial details of the
modeled air quality fields, especially for pollutant

-------
species that may have fine spatial gradients that are
important for human exposure assessment.
3.2 Sub-Crid-Scale Modeling
A property of CMAQ (and most such grid models) is
that the local, near-source concentration distributions
due lo dispersion of point, line, and area sources are not
explicilly treated; rather, emissions from such sources
are immediately dispersed throughout the grid volume.
Thus, near-source concentrations will be higher than
from the resolved-scale predictions, and consequently,
exposure assessments will significantly underestimate
the potential levels of human exposures in the near
source areas. The design problem Is to seek a model
resolution that assures the Inclusion of the important
contributing atmospheric processes and also provides a
means for estimating the near-source concentration
distributions. For practical considerations, we address
the latter requirement by developing concentration
probability distribution functions, PDFs, for the sub-grid-
scale concentration distributions. This system of
resolved-scale and sub-grid-scale concentration
distributions provides a more robust set of modeling
outputs for human population assessments. We identify
and propose methodologies for handling two
contributing sources of this variability, including: (1)
dispersion of point and area sources from street canyon
flows using a combination of computational fluid
dynamics and wind tunnel modeling techniques, and (2)
direct coupling between turbulent motions and reactive
pollutants when chemical reaction time scales are on
the order of the turbulent eddy time scale (Herwehe
2000) In the latter case, a coupled targe-eddy
simulation photochemical model (LESchem) (Herwehe
2000) is used lo address sub-grid-scale variability for
poilutanls with different chemical reactivity rates and
turbulent transport time scales. Using this modeling
approach, the chemistry and turbulent transport fields
are solved as a coupled system, which allows bi-
directional feedback between the chemistry and
dynamics during the simulation.
4. RESULTS
Preliminary results for a case study for Philadelphia
indicate that the extent of the resolved-scale spatial
variability varies with each pollutant species, and the
grid-resolved variability does not necessarily increase
monotonically with increased grid resolution (Ching
et al. 2000). This means that the grid resolution
selected for use in exposure modeling may need to be
ascertained by numerical experiments.
Analysis of the modeled time series of the hourly
mean and standard deviation of PMis and the number
density of particles in the Aitken nuclei mode shows
large differences in the diurnal trend, as well as the
degree and characteristics of spatial variability between
these two parameters that show a sensitivity response
lo model grid size. Both the number and mass of the
Aitken mode particles (including the ultra-fine particles)
are believed lo contribute to health risk.
Results to date for the sub-grid modeling focus on
the turbulence-induced concentration fluctuations.
Results indicate a wide range of sub-grid chemical
variability dependent on pollutant species (e.g., large
variability for formaldehyde and acetaldehyde; relatively
small variability for carbon monoxide) at the surface as
well as aloft in the mixed layer due to the degree of
photochemical reactivity in atmospheric mixtures and
various trace species chemical lifetimes.
Disclaimer. The information in this manuscript has been
prepared under funding by the United States
Environmental Protection Agency. It has been
subjected to Agency review and approved for
publication. Mention of trade names or commercial
products does not constitute endorsement or
recommendation for use.
5. REFERENCES
Brown, M. J,, 2000; Urban parameterizations for
mesoscale meteorological models. Mesoscale
Atmospheric Models- Z. Boybcyi, ed. Wessex
Press
Burke, J. M., M J. Zufall. and H A. Ozkaynak. 2001: A
population exposure model for particulate matter:
case study results for PM2.5 in Philadelphia, PA.
Journal o! Exposure Analysis and Environmental
Epidemiology, 11(6), 470-489.
Byun, D. W,, and J. K. S. Ching, 1999; Science
algorithms of the EPA Models-3 Community
Multiscale Air Quality (CMAQ) Modeling System.
EPA-600/R-99/030, U.S. EPA.
Ching, J„ A. Lacser. D. Byun, and W. Benjey, 2000: Air
quality modeling of PM and toxics at neighborhood
scales to improve human exposure assessments.
Third Sym. on the Urban Environment, Davis, CA
Amer. Meteor. Soc. P96-97.
Grell, G. A., J. Dudhia, and D. R. Stauffer, 1994: A
description of the fifth-generation Penn State/NCAR
mesoscale model (MM5). NCAR Tech. Note
NCAR/TN-398+STR, 138 pp.
Herwehe J. A . 2000: A numerical study of the effects of
large eddies on trace gas measurements and
photochemistry in the conveclive boundary layer.
Ph.D. Dissertation, Department of Atmospheric
Sciences. University of Alabama in Huntsville.
242 pp
Lacser, A., and T. L. OUe. 2002: Implementation of an
urban canopy parameterization in MM5. Fourth
Sym. on the Urban Environment, Norfolk, VA,
Amer. Meteor. Soc.

-------
TECHNICAL REPORT DATA
1. REPORT NO.
2.
3 .
4. TITLE AND SUBTITLE
Neighborhood Scale Modleing of PM25 and Air Toxics
Concenration Distributions to Drive Human Exposure
Models
5.REPORT DATE
6.PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Jason Ching, Avraham Lacser, Tanya Otte, Jerry Herwehe, and Daewon Byun
8.PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
National Exposure Research Laboratory - RTP, NC
Office of Research and Development
U.S. linvironmental Protection Agency
Research Triangle Park, NC 27711
IO.PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Same as 9.
13.TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Air quality (AQ) simulation models provide a basis for implementing the National Ambient Air Quality Standards (NAAQS)
and are a tool for performing risk-based assessments and for developing environmental management strategies. Fine
particulate matter (PM2 5), its constituents and size and number distribution, as well as airborne toxic pollutants ("air toxics")
have characteristically different degrees of spatial and temporal variability especially in urban areas and in different
geographical-climatic regimes. In this study, we explore the specific role of AQ models as a means to drive human exposure
models (Burke et al. 2001) and to address situations in which pollutants exhibit high spatial and temporal variability. We
seek a capability to capture the resolved-scale concentration fields and to provide measures of sub-grid-scale variability in
concentration distributions that impact human exposures. This modeling approach is meant to enhance and complement the
more limited data from central site monitoring networks to provide concentration fields at high temporal and spatial
resolutions. By providing further information on concentration variability at sub-grid scales, we complete the requirements
needed for exposure assessments. The various elements of this modeling approach and some of their specific modeling
issues are described below.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/ OPEN ENDED TERMS
c.COSATI



18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
Unclassified
21.NO. OF PAGES
20. SECURITY CLASS (This Page)
Unclassified
22. PRICE

-------