8A.6 AIR QUALITY MODELING OF PM AND AIR TOXICS AT NEIGHBORHOOD SCALES
Jason Ching*
Atmospheric Sciences Modeling Division
Air Resources Laboratory
National Oceanic and Atmospheric Administration
Research Triangle Park, NC 27711
1 INTRODUCTION
The current interest in fine particles and toxics
pollutants provide an impetus for extending air
quality modeling capability towards improving
exposure modeling and assessments. Human
exposure models require information on
concentration derived from interpolation of
observations taken from monitoring networks.
Causal mechanisms for adverse health from
particuiate matter and other air pollutants are
numerous, but not well understood; however they
provide much of the rationale for the nation's PM
research portfolio (NRC 98, 99). The NRG listed 10
causal hypotheses, each relating to some physical
aspect or speciation of PM, and/or toxic pollutant
species. The distribution of concentration fields for
different PM causal pollutants will be highly complex
at neighborhood scales. However, the number of
locations of samplers of typical networks in urban
areas is generally sparse; also, due to the sheer
myriad of PM and toxic substances, temporal
sampling of physical parameters of PM, speciated
PM and toxic pollutants are limited and varied
varying from sub-hourly to daily or weekly samples,
and/or are surmised as surrogates of the available
measurements. Thus, clearly, the observed
temporal and spatial concentration fields are poorly,
or inadequately resolved for driving exposure
models and conducting health risk assessments.
Currently the EPA emissions based modeling
systems, ModeIs-3 Community Multiscale Air
Quality Modeling System (CMAQ) (Byun and Ching,
1999) is capable of modeling PM 2.5 and PM-10 at
horizontal resolutions of ~36km for regional to 4 km
for urban scale predictions. Urban areas are
sources of large amounts of pollutants that
contribute to significant and inherently subgrid
spatial variability of the concentration fields and to
subsequent exposures. Stationary monitors will be
*On assignment to the National Exposure Research
Laboratory, U.S. Environmental Protection Agency
Corresponding Address: Atmospheric Modeling
Division, NERL, USEPA (MD-80), RTP, NC 27711
email address: ching.jason@epa.gov
unable to characterize this variability. Current
Eulerian-based air quality models' spatial resolution
is coarse and cannot resolve the fine scale details.
The modeling of dispersion of local sources ignores
the regional background. Modeling methodologies
and parameterization techniques for the transport
and dispersion of these local sources in complex
urban canyons are limited. Methods to serve as a
bridge between these modeling and monitoring
approaches to determine concentration variation
arising from the juxtaposition of concentration from
the regional and urban sources are needed.
In this presentation, a framework for extending
the Models-3/CMAQ to be operable at a full range
of scales from regional to the neighborhood scale
for use in exposure modeling is described. As part
of this study, methodologies and approaches
envisioned to develop rational linkages with
ambient and exposure monitors to provide
concentration fields as critical inputs to models of
human exposure (and epidemiologicai studies) are
discussed. This initial study includes refining the
model scales to grid sizes of order 1 km, to develop
the sub-grid scale parameterizations and
subsequently, to deriving functional linkages
between the modeling and the ambient fixed site
and personal exposure monitoring data, and
incorporates in the implementation, physical flow
modeling and visualization, computational fluid
dynamics modeling and statistical techniques.
Specific aspects under study include the
development of functional relationships that
provides a mapping across space and time between
the modeled and monitored fields, considerations of
sensitivities to model grid resolution, and for
different emissions scenarios, for different and full
ranges of averaging time periods from hourly to
annual fields. The effort will include methods for
modeling exposures for a variety of human activity
patterns.
2. PROBLEM DEFINITION
The following discussion provides a conceptual
framework and thus the basis of the requirements
-------
for the study. Given:
Exposure = SUM [Joint (Activity (x,y,P(t)) X
Concentration (x, y, t))] Time
where
Activity
Actual and/or patterns of human activity
including the actual time, t, the time period (P(t))
spent outdoors, the time and pattern of
commuting to and from micro environments,
and the time spent indoors in the various
micro environments. Information needs include
the locations (x, y) where activity occurs.
Concentration:
Ambient spatial (x, y) distribution and temporal
variations (t) of each and all transported primary
and secondary pollutants, and those freshly
dispersed pollutant sources in the urban areas
and in different micro environments. These
fields are influenced by the exchange between
the micro environments and the ambient air.
Time:
Temporal interval for acute (days to months)
and for chronic (months to years) responses for
time intervals, P, for which significant exposure
in given activity takes place (minutes to hours).
Current approaches for modeling ambient
concentration fields at urban to neighborhood
scales for PM and toxics do not yet exist or are
overly simplistic. Urban scale grid models are
unable to depict spatial variation from sources at
subgrid scales. Dispersion models developed to
handle local scale sources do not handle secondary
pollutants. Representation of transport and
dispersion for use in urban air quality models,
especially for PM and Toxics pollutants is a
problem. The modeling of the spatial and temporal
distribution of these trace pollutants will depend
highly on the representation of the transport field
affecting the dispersion of the sources. The mere
introduction of parameterization of building
structures in urban canopy into mesoscale
meteorological models, will increase the drag, and
turbulence causing enhanced horizontal dispersion.
However, the flow and dispersion in street canyons
will in general differ significantly from grid resolved
wind fields for which street canyons are but
roughness elements of an urban canopy (Brown
and Williams, 1998, Brown et al, 1998). They
demonstrate for subgrid features such as buildings
and street canyons that pollutant trapping occurred
in street cavities building up the levels of
concentrations, and some pollutants are transported
upwind of the buildings due to recirculating flows;
additionally, the enhanced vertical dispersion due to
the presence of buildings caused pollutants to be
dispersed further downwind by faster moving winds.
Additionally, the PM and toxic pollutants will
undergo changes in both their physical and
chemical properties during transport and dispersion.
Many toxic pollutants are semivolatile at ambient
conditions and thus can either absorb and/or adsorb
onto ambient particles, thus adding additional
degrees of complication. Exacerbating the modeling
problem is the sheer numbers of toxic pollutants
that will be under consideration. Grouping of HAPS
compounds by toxicity classes, by degree of
reactivity, by volatility, and by the use of surrogates
are modeling approaches that may be used for
initial implementation.
Human exposure depends on time spent in
outdoor and in various micro environments. The
pollutant concentration in micro environments such
as homes, schools, workplace, vehicles, etc., will
depend on both the internal sources and as well as
exchanges between these micro environments and
the ambient air. Personal exposure to ambient
levels of such pollutants will depend not only on the
duration of time spent in each of the various micro
environments, but the time and location of activity,
because pollutant concentration is time and space
variant. The sum of the product of activity and
concentration is the cumulative exposure over some
time period, from relatively short term, acute to
longer term, chronic.
3. Study Approach
The goal of developing neighborhood scale
modeling capability to resolve concentration fields
at neighborhood scales begins with a systems
review, including identifying and reviewing the major
modeling components, followed by implementation
of optional approaches, demonstration and testing
phase, and methodologies for practical operations.
Figure 1 identifies several major components to be
investigated in this study. These components are
discussed below:
3.1 Methodology for Handling Emissions Data:
Local sources may be either, (a) modeled
separately using local dispersion modeling
techniques to provide a basis for determining sub-
grid resolved concentration fields in urban canyons
for further use in exposure assessments or, (b)
incorporated, somehow, as inputs to gridded air
quality simulation models. In the latter case, the
development and testing of methodology(s) for
preparing gridded emission from the sub-grid scale
sources is reviewed.
-------
Tools Data
Concentration Fields
High Resolution Activity
Exposure
Timescale
Detailed Database
—-Emissions
—-Meteorology
•Building data
(Monitoring \
Network /
(Computational \
Fluid /
Dynamics /
1
Extend
Mode!s-3/CMAQ
to Neighborhood
Scales
Exchanges
Ambient-
Micro-Environments
Human
Activity
Physical
Modeling
VFlow
'isualtzation
Figure 1. Design components of Neighborhood scale airquality simulation mode! of PM and Toxic pollutants
3,2 CFD (Computational Fluid Dynamics Model):
Numerical experiments conducted to examine
the dispersion and the concentration fields
associated with emissions in conceptualized street
canyons with varying degrees of complexities and
configurations will provide bases for subsequent
investigations and development of parametric
methodologies fordetaited treatments of the subgrid
variability in gridded air quality models. The
experiments will incorporate increasingly more
complex descriptions of dispersion, chemistry and
deposition. Guidance for more operational
techniques will be an objective.
3.3 Physical Modeling and Flow Visualization:
Physical modeling and flow visualization
experiments will be conducted to provide a basis
for the testing of CFD modeling results, and to
develop and test methodologies for gridding sub-
grid scale emissions and for examining details of
the sub-grid scale transport and dispersion. Early
results of flows over scaled series of 2-D array of
buildings in the USEPA Fluid Modeling Facility's
wind tunnel show changes in the degree of flow
perturbations downwind from the leading edge of a
series of 2-D array of modeled buildings. Such flow
complexities provide a challenging basis for
evaluating CFD models. Carefully designed
experiments will help guide the development of the
parameterizations of dispersion and transport in the
air quality models.
3.4 Modeling and Processes Research:
The Models-3/CMAQ will be set up to operate
with additional nesting at finer grid resolutions to
the current 36, 12 and 4 km set. To achieve this,
this study will involve tasks to prepare emission and
meteorology modeling, input data and science
algorithms at commensurate grid resolution.
Sensitivity and process analyses will be conducted
to investigate and to understand the response and
contribution of different science process modules
-------
and other necessary parameterizations to modeling
at grid resolutions of order 1 km or less.
3.5 Links with monitoring data:
This project will explore and develop practical
methodology that will relate both fixed site and
personal monitoring data to model outputs at four
(and 1,3) km. Numerical and physical modeling will
provide an opportunity to capture the concentration
fields with high temporal and very fine spatial
resolution. Monitoring data provide ground truth
information to check and to evaluate model
predictions, and is typically the basis for driving
exposure models. Candidate approaches such as
Neural Networking and/orothergeospatial-temporal
mapping will be investigated. The resulting
functional fields will greatly enhance the running of
exposure models. An additional spinoff from this
study will be to improve the siting strategy for
deploying monitors.
3.6 Links with Exposure Models:
Develop and subsequently demonstrate
methodology for computing exposures for different
emissions scenarios, including traffic, point/area
sources for different integration time periods from
one hour to annual and for different human
exposure ^situations such as in traffic, outdoor and
indoor exposures, and for different susceptible
populations. The concentration information will
include speciation of PM-2.5 to address health
impact hypotheses such as by total mass, size
distribution, numberdensity (especially for the ultra-
fine particles), and by speciation including unique
properties such as acidity, oxidizing capacity, trace
metals, and toxicity.
4. DISCUSSION AND SUMMARY
This project is expected to develop Eulerian
based air quality modeling methodology(s) and
capability(s) to support human exposure modeling
and investigations testing the various health
hypotheses concerning adverse health effects by
various pollutants (NRC98.99). It has a follow-on
benefit to addressing urban toxics exposure issues.
The project is currently being implemented by a
team of NOAA and EPA scientists and their
collaborators to develop and study methods for
integrating Eulerian models for urban scales with
local scale models (using a combination of CFD and
physical modeling which account for urban canopy
and local emissions sources including traffic as well
as point/area sources of pollutants). The project will
include but not be limited to deriving various
functional linkages between the Models-3/CMAQ
emissions based modeling system concentration
fields of key particulate matter parameters with
ambient fixed site and personal exposure
monitoring data, and to incorporating into the
methodologies, flow visualization, computational
fluid dynamics modeling and statistical techniques.
The project will further develop and derive
functional relationships that provide a mapping
across space and time between the modeled and
monitored fields. The investigation will include a
variety of studies including sensitivities to model
grid resolution, examination of a variety of different
emissions scenarios, and identifying and testing
methods for handling the full range of averaging
time periods from hourly to annual fields. The effort
will include developing methods for modeling
exposures for a variety of human activity patterns.
In this demonstration project, one or more candidate
urban areas will be selected for detailed
investigations. Criteria for selection include the
existence of PM sampling databases and detailed
emissions inventories
Disclaimer: This paper has been reviewed in
accordance with the U.S. Environmental Protection
Agency's peer and administrative review policies
and approved for presentation and publication.
Mention of trade names or commercial products
does not constitute endorsement or
recommendation for use.
References:
Brown, Michael, Cathrin Muller, 1997: The effect of
micro scale urban canyon flow on mesoscale
puff dispersion. 12th Symposium on Boundary
Layers and Turbulence. Vancouver BC,
American Meteorological Society, 463-464.
Brown, Michael J., and Michael D. Williams, 1998:
An urban canopy parameterization for
mesoscale meteorological models. Second
Urban Environment Symposium., 13th
Conference on Biometeorology and
Aerobiology, Albuquerque, NM., American
Meteorological Society,144-147.
Byun, D.W. and Ching, J.K.S., 1999, Editors:
Science Algorithms of the EPA Models-3
Community Multiscale Air Quality (CMAQ)
Modeling System. EPA- 600/R-99/030, ORD,
U.S. Environmental Protection Agency
NRG (National Research Council) 1998: Research
Priorities for Airborne Particulate Matter. I.
Immediate Priorities and a Long- Range
-------
Research Portfolio., Washington DC., National
Academy Press.
NRC (National Research Council) 1999: Research
Priorities for Airborne Particulate Matter. II.
Evaluating Research Progress and Updating
the Portfolio., Washington DC,, National
Academy Press.
-------
NERL-RTP-00663
TECHNICAL REPORT DATA
1. REPORT NO,
1PA/600/A-00/018
. TITLE AND SUBTITLE
Air Quality Modeling of PM and Air Toxics at
Neighborhood Scales
8.PERFORMING ORGANIZATION REPORT NO.
Jason China
9. PERFORMING ORGANIZATION NAME'AND ADDRESS
Same as Block 12
U.S. Environmental Protection Agency
Office of Research and Development
National Exposure Research Laboratory
Research Triangle Park, NC 27711
13.TYPE OF REPORT AND PERIOD COVERED
Extended Abstract, FY-00
14..SPONSORING AGENCY CODE
EPA/600/9
16. ABSTRACT
INTRODUCTION The current interest in fine particles and toxics pollutants provide an impetus for extending air quality modeling
capability towards improving exposure modeling and assessments. Human exposure models require information on concentration
derived from interpolation of observations taken from monitoring networks. Causal mechanisms for adverse health from particulate
matter and other air pollutants are numerous, but not well understood; however it provides much of the rationale for the nation's
PMresearch portfolio (NRC 98, 99). The NRC listed 10 causal hypotheses, each relating to some physical aspect or speciation
of PM, and/or toxic pollutant species. The distribution of concentration fields for different PM causal pollutants will be highly
complex at neighborhood scales. However, the number of locations of samplers of typical networks in urban areas is generally
sparse; also, due to the sheer myriad of PM and toxic substances, temporal sampling of physical parameters of PM, speciated
PM and toxic pollutants are limited and varied varying from sub-hourly to daily or weekly samples, and/or are surmised as
surrogates of the available measurements. Thus, clearly, the observed temporal and spatial concentration fields are poorly, or
inadequately resolved for driving exposure models and conducting health risk assessments. Currently the EPA emissions based
modeling systems, Models-3 Community Multiscale Air Quality Modeling System (CMAQ) (Byun and Ching, 1999) is capable
of modeling PM 2.5 and PM-10 at horizontal resolutions of ~36km for regional to 4 km for urban scale predictions. Urban areas
are sources of large amounts of pollutants that contribute to significant and inherently subgrid spatial variability of the
concentration fields and to subsequent exposures. Stationary monitors will be unable to characterize this variability. Current
Eulerlan-based air quality models' spatial resolution is coarse and cannot resolve the fine scale details. The modeling of dispersion
of local sources ignores the regional background. Modeling methodologies and parameterization techniques for the transport
and dispersion of these local sources in complex urban canyons are limited.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
^IDENTIFIERS/ OPEN ENDED TERMS
o.COSATI
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19, SECURITY CLASS (This Report)
UNCLASSIFIED
21 .NO. OF PAGES
20. SECURITY CLASS (This Page)
UNCLASSIFIED
22. PRICE
------- |