United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-79-026
December 1979
Application of
Photochemical Models
Volume II
Applicability of Selected
Models for Addressing
Ozone Control
Strategy Issues
-------
APPLICATION OF PHOTOCHEMICAL MODELS
Volume II
Applicability of Selected Models for Addressing
Ozone Control Strategy Issues
prepared by
Association of Bay Area Governments
Hotel Claremont
Berkeley, California 94705
in association with
Bay Area Air Quality Management District
San Francisco, California
Lawrence Livermore Laboratory
Livermore, California
Systems Applications, Inc.
San Rafael, California
prepared for
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
EPA Project Officer: John Summerhays
Contract No. 68-02-3046,
Final Report, December 1979
-------
PREFACE
This document is one of four volumes intended to provide information
relevant to the application of photochemical models in the development
of State Implementation Plans. The reports are particularly directed
toward agencies and individuals responsible for preparation of
non-attainment plans and SIP revisions for ozone. The four volumes are
titled as follows:
Application of Photochemical Models
Volume I - The Use of Photochemical Models in Urban Ozone
Studies
Volume II - Applicability of Selected Models for Addressing
Ozone Control Strategy Issues
Volume III - Recent Sensitivity Tests and Other Applications
of the LIRAQ Model
Volume IV - A Comparison of the SAI Airshed Model and the
LIRAQ Model
This work is to a large extent based on the photochemical modeling
experience gained in the San Francisco Bay Area in support of the 1979
Bay Area Air Quality Plan, the following individuals made significant
contributions to this work:
Association of Bay Area Governments - Ronald Y. Wada
(Project Manager)
- M. Jane Wong
- Eugene Y. Leong
Bay Area Air Quality Management District - Lewis H. Robinson
- Rob E. DeMandel
- Tom E. Perardi
- Michael Y. Kim
Lawrence Livermore Laboratory - William H. Duewer
Systems Applications, Inc. - Steven D. Reynolds
- Larry E. Reid
The authors wish to express their appreciation to John Summerhays, EPA
Project Officer in the Source Receptor Analysis Branch of OAQPS, for his
thoughtful review and conments on earlier drafts of this report.
-------
TABLE OF CONTENTS
Page
INTRODUCTION 1
BACKGROUND 1
ALTERNATIVE APPROACHES TO OZONE MODELING 1
OZONE CONTROL STRATEGY ISSUES 2
Hydrocarbon vs NOx Control 2
Assessing Attainment of the Ozone Standard
and a Short-Term N02 Standard 3
Control of Low Level vs Elevated Emissions 3
Selective Control of Hydrocarbon Compounds 4
Long Range Transport . . . - 4
Effects of Spatial and Temporal Variation
of Ozone Precursor Emissions 6
SPECIFIC MODEL ASSESSMENTS 6
ROLLBACK 9
MODEL FEATURES 9
APPLICABILITY OF ROLLBACK TO OZONE CONTROL STRATEGY
ISSUES 12
STANDARD EKMA (EMPIRICAL KINETIC MODELING APPROACH) 13
MODEL FEATURES 13
APPLICABILITY OF STANDARD EKMA MODEL TO OZONE CONTROL
STRATEGY ISSUES 16
CITY-SPECIFIC EKMA USING OZIPP (OZONE JSOPLETH PLOTTING
PACKAGE) 7 17
MODEL FEATURES 17
APPLICABILITY OF CITY-SPECIFIC EKMA MODEL TO OZONE
STRATEGY CONTROL ISSUES 19
HANNA ATDL PHOTOCHEMICAL MODEL (ATMOSPHERIC TURBULENCE AND
DIFFUSION LABORATORY) . . . ." ~ 21
MODEL FEATURES 21
APPLICABILITY OF ATDL MODEL TO OZONE CONTROL STRATEGY
ISSUES 23
-------
TABLE OF CONTENTS
Page
ELSTAR (ENVIRONMENTAL LAGRANGIAN SIMULATOR OF TRANSPORT
AND ATMOSPHERIC REACTIONS) 25
MODEL FEATURES 25
APPLICABILITY OF ELSTART TO OZONE CONTROL STRATEGY
ISSUES 27
TRACE (TRAJECTORY ATMOSPHERIC CHEMISTRY AND EMISSIONS) 29
MODEL FEATURES 29
APPLICABILITY OF TRACE TO OZONE CONTROL STRATEGY
ISSUES 31
LIRAQ-2 (LIVERMORE REGIONAL AIR QUALITY MODEL, VERSION 2) 33
MODEL FEATURES * 33
APPLICABILITY OF LIRAQ-2 TO OZONE CONTROL STRATEGY
ISSUES 35
SAI AIRSHED 37
MODEL FEATURES 37
APPLICABILITY OF AIRSHED TO OZONE CONTROL STRATEGY
ISSUES 39
IMPACT (INTEGRATED WDEL FOR PLUMES AND ATMOSPHERICS IN
COMPLEY TERRAIN) ~ 41
MODEL FEATURES 41
APPLICABILITY OF IMPACT TO OZONE CONTROL ISSUES 43
MADCAP (MODEL OF ADVECTION, DIFFUSION AND CHEMISTRY OF AIR
POLLUTION)
MODEL FEATURES 45
APPLICABILITY OF MADCAP TO OZONE CONTROL ISSUES 47
RECOMENDATIONS FOR FUTURE WORK 49
FURTHER PERFORMANCE EVALUATION AND SENSITIVITY ANALYSIS .... 49
IMPROVING USER SERVICES 50
REFERENCES 51
-------
INTRODUCTION
BACKGROUND
Over the past several years a variety of photochemical models have been
developed to describe the relationship between ozone precursor emissions
and subsequent ozone production in the atmosphere. These model
development efforts have been reviewed on nunerous occasions. (See, for
example, Systems Applications, Inc., 1976, U.S. EPA, 1977, and Argonne
National Laboratory, 1977.) A primary objective in the development of
these models has been to evaluate the effectiveness of alternative
emission control strategies in reducing ambient ozone levels. At
present, six ozone control strategy issues can be identified as being
potentially significant for any region with an ozone problem. The
purpose of this model assessment is to evaluate the applicability of
selected models to the control strategy issues identified.
Since no quantitative, side-by-side comparison tests of ozone model
performance have been made to datej' judgments concerning the adequacy of
treatment of a particular issue by a given model have been limited to
aspects of the model's basic theoretical structure. For example, no
judgment is made regarding the adequacy of a 31-step chemical kinetics
mechanism versus a 51-step or 76-step mechanism—they all replicate smog
chamber data and whether one is "more valid" than another is arguable at
this point. Conversely, if a key variable is ignored or assumed to
behave in a way which is known to be untrue under certain circumstances,
then a judgment is made regarding the applicability of the model to such
circumstances. The evaluation is presented on two levels: First, the
general features and limitations of each type of modeling approach are
reviewed with respect to the issues; and, second, more detailed aspects
and applications of specific models are presented.
ALTERNATE APPROACHES TO OZONE MODELING
The variety of ozone models currently available may be categorized as
rollback models, empirical models or deterministic models. Rollback
models are the simplest, easiest and cheapest models available. While
there are various forms of rollback models, they are all based on the
presumption that for any given region a reduction in hydrocarbon
emissions will produce a proportional reduction in maximum ozone levels.
The complexities of the ozone formation and transport process are
ignored, and the models are difficult in practice to validate due to
lack of reliable historical emissions inventory and air quality data.
Empirical models use existing air quality monitoring data and/or data
from laboratory smog chamber studies to deduce the sensitivity of
maximum ambient ozone concentrations to changes in ambient hydrocarbon
and NOx concentrations. Theoretically, empirical models could, with the
appropriate input data, be formulated to address any question of
interest. Due to the practical limitations of existing data, however,
application of existing empirical models to the evaluation of
alternative control strategies requires an assumption that the ambient
1
-------
hydrocarbon and NOx concentrations used are directly proportional to
regional emissions for each pollutant. In addition, since monitoring
data are used in the construction of the model, there is no simple means
by which the model's performance can be evaluated.
Deterministic models attempt to mathematically describe the behavior of
pollutants in the atmosphere. They may be further subdivided into three
major classes: Gaussian, Lagrangian, and Eulerian. Gaussian models are
relatively simple and presume both spatially uniform and steady-state
atmospheric conditions. As a result they cannot be validly applied to
describe the time-varying chemical and meteorological processes involved
in ozone formation. Lagrangian models "trace" a parcel of air as it is
carried by the wind across a given area. Typically, Lagrangian models
focus on the chemistry of ozone formation within an "air parcel" while
portraying meteorological transport and diffusion processes in
simplistic terms. Finally, Eulerian models define a fixed grid over a
given area. Pollutant mass is passed from one grid cell to another
according to wind, diffusion, and chemical reaction parameters specified
for each cell at each time step. Eulerian models are generally the most
sophisticated, require the most input data, are the most expensive to
run, and provide the most output information.
Because deterministic models are based on mathematical representations
of physical and chemical processes, an evaluation of model performance
is possible for a given region by comparing model simulations of
historical conditions with the ozone levels recorded under those
conditions. While such a comparison does not necessarily confirm a
model's capability to predict future situations, it at least provides an
intermediate point at which model performance may be independently
assessed.
OZONE CONTROL STRATEGY ISSUES
Available models may be applied to six ozone control strategy issues.
The general characteristics of the modeling approaches previously
summarized lead to certain general conclusions regarding their
applicability in each case. The control strategy issues identified for
this evaluation are described below, as are general conclusions
regarding model applicability in each case.
Hydrocarbon vs. NOx Control
Ever since the mechanism for photochemical ozone formation was first
• identified, it has been known that the primary ingredients were
hydrocarbons, oxides of nitrogen, and sunlight. The most obvious
targets for control are anthropogenic sources of hydrocarbons and NOx;
however, the relative emission reduction that should be applied to each
precursor to achieve the ozone standard in most regions has never been
adequately estimated. Recent model application efforts have begun to
address this problem, but results have yet to be adequately verified.
Thus, the first issue of significance in ozone modeling is: can
available models be applied to provide estimates of the degree of
2
-------
hydrocarbon and NOx control required to achieve the ozone standard in a
given region?
Most of the modeling approaches in some way account for the roles of
hydrocarbons and NOx in ozone formation. The rollback models do not
include NOx and therefore cannot assess the effect of changes in NOx
emissions on ambient ozone levels.
Assessing Attainment of the Ozone Standard and a Short-Term N02 Standard
Many of the models currently available incorporate a role for NOx in
ozone formation. This feature may or may not be sufficient to allow the
same models to be applied to assess whether a given region will attain a
short-term (e.g., 1-to 3- hour average) N02 standard. The setting of
such a standard is called for in the 1977 Clean Air Act Amendments if
the medical and scientific data support the need for such a health-based
standard. The question is: can existing models be applied to
simultaneously assess attainment of both ozone and N02 ambient air
quality standards in a given region?
Since attainment of a short-term N02 standard has not been an issue of
concern in the past, none of the existing models has ever been applied
to such a question; therefore, there is little or no experience to
suggest whether existing models could successfully address the issue.
On a theoretical basis, there is no reason why each of the modeling
approaches could not be redirected to address the issue of attaining a
short-term N02 standard, with each approach retaining a set of
advantages and disadvantages similar to those previously described.
However, the atmospheric conditions which result in a buildup of high
N02 levels differ from those which result in high ozone levels.
Previous studies of model performance have focused on ozone-episode
days, and there is a need to carry out studies on N02-episode days in
order to assess model performance under such conditions (i.e.,
additional evaluation and/or field verification would be required before
existing models could be accepted as satisfactorily addressing a
short-term N02 standard).
Control of Low-Level vs. Elevated Emissions
Elevated stacks have long been used to reduce the maximum ground-level
concentrations of such primary pollutants as particulate matter and
sulfur dioxide. Whether elevated emissions of hydrocarbons or nitrogen
oxides have a lesser impact on ground-level ozone concentrations than
low-level emissions has not been established. If there is a difference
* in effect, then this.could alter the relative cost-effectiveness of
emission control for elevated vs. low-level sources. The issue from a
modeling perspective is: can existing models assess the relative effect
of comparable emission reductions from elevated vs. low-level emissions
in a given region?
For a model to distinguish the difference between low-level and elevated
emissions 1t must have spatial resolution in the vertical dimension. In
addition, the model should include some treatment of plume rise and
3
-------
other small-scale influences on plume dispersion and chemistry.
Rollback or empirical models, which deal with aggregated regional
emission totals, cannot address this issue. While the Lagrangian and
Eulerian models explicitly address spatial resolution in the horizontal
dimensions, not all incorporate the vertical resolution which would
allow them to address this issue. In addition, regardless of the
theoretical possibilities, proper validation of any model for this
purpose ideally would involve the collection of meteorological and air
quality data at varying levels above the ground—existing ground-level
monitoring data which is commonly collected would provide only partial
verification of a model to address this issue.
Selective Control of Hydrocarbon Compounds
Laboratory (smog chamber) studies of oxidant formation have shown that
different hydrocarbon compounds react at different rates and produce
varying levels of maximum ozone. Early efforts at hydrocarbon control
(e.g., Los Angeles' Rule 66) focused on those compounds which were known
to be highly "reactive". More recent studies seem to indicate that
while the rate of ozone production may differ, the cumulative ozone
production potential of most hydrocarbon compounds is roughly the same
(Dimitriades, 1977). The issue is: can existing models be applied to
determine the viability of a strategy that would selectively control
emissions df specific classes of hydrocarbon compounds to attain the
ozone standard in a given region?
For a mode! to distinguish the effect of controlling different
hydrocarbon compounds 1t must take separate account of each class of
compounds of Interest. Lagrangian and Eulerian models Include a
chemical kinetics mechanism that may incorporate from one to several
classes of hydrocarbon species. An emission inventory developed for
input to such a model must be disaggregated Into the species
classifications by which the model makes its computations. Thus, the
selectivity in the types of species control which may be assessed by a
given model is restricted by the definitions of the species
classifications used in the model. For example, if a model treats only
one class of hydrocarbon compounds in its computations (e.g., total
hydrocarbons or non-methane hydrocarbons) then that model cannot
distinguish the difference between a control program which would remove
a given quantity of a "highly reactive" compound and one which would
remove a similar quantity of a less-reactive compound.
Long-Range Transport
There is an ample body of evidence indicating that ozone and ozone
precursors can be and are transported over long distances beyond
conventional "air basin" boundaries. The impact of emissions 1n one
region on ozone levels in another has been an issue of concern in many
areas across the nation. For this evaluation the general issue is: can
existing models be applied to assess long-range transport of ozone and
its precursors? This issue can be divided into two related issues: (1)
can existing models be applied to assess the effects of long-range
transport of pollutants from "upwind" sources on ozone levels in a given
4
-------
study area?; and (2) can existing models be applied to assess the
influence of emissions in the study area on levels of ozone and
precursors in other areas downwind?
The impact of ozone transported into a study area can be considered in
all the models evaluated in this report. The version of rollback
sanctioned by EPA considers transport by relating a city's emissions
specifically to the city's contribution to its ozone problem. A
similiar process is performed in using the standard version of EKMA.
The other models consider transport as a direct input into the model.
Thus the remainder of this report will focus on the impact of a study
area on the pollutant transport downwind.
The issue of influence of a study area on transport long distances
downwind is complicated by the lack of a precise definition of
"long-range". From a modeling standpoint, long-range could be defined
as that distance over which the validity of the assumptions inherent in
existing models becomes questionable. This distance will vary depending
on the particular atmospheric conditions being modeled and the model
being used. Rollback and available empirical models do not incorporate
spatial resolution at all, hence they cannot be applied to address this
issue. Lagrangian models are limited by the distance over which their
assumptions of no horizontal diffusion or wind shear are valid. Both
Lagrangian and Eulerian models may be limited by how well their chemical
kinetics can simulate what occurs under nighttime conditions.
In each of these cases, there is little or no experience as to whether
models in their existing forms can validly simulate the processes that
result in long-range transport of ozone and ozone precursors. While it
is conceivable that Lagrangian and Eulerian models could be modified to
provide results over longer transport distances than is currently the
case, the validity of the application depends to a large extent on the
particular transport mechanism of significance over a given area. For
example, a commonly suggested mechanism for long range transport
involves pollutants that become trapped in stable atmospheric layers
aloft, are transported long distances, and that then are transported to
the surface at a location far removed from the original source area. A
mechanism of this complexity cannot be addressed by existing models
without further data acquisition and performance evaluation.
In general, available models are intended for application to urban areas
with horizontal distance scales of roughly 10 to 100 kilometers. If the
particular transport mechanism of significance in a given area can be
identified, and if monitoring data are available to characterize the
" "upwind" boundary conditions for the study area, then existing models
could be applied to assess the impact of pollutants transported into a
given region. In general, however, calculation of effects over distance
scales significantly greater than 100 kilometers would be best addressed
using a model specifically designed for such purposes. For example, EPA
is currently funding the development of a mesoscale photochemical model
to be applied to the Northeastern United States.
R
W
-------
Effects of Spatial and Temporal Variation of Ozone Precursor Emissions
The complex process of ozone formation varies in space and time.
Differing spatial and temporal emission patterns within a given region
could possibly produce different ozone levels. Thus the issue is: can
existing ozone models account for and be applied to assess the effects
of alternative spatial and temporal patterns of precursor emissions?
The model requirements to address this issue are self-evident; only
existing Lagrangian and Eulerian models can be applied to assess the
effects of spatial and temporal emission variations. Of these, Eulerian
models provide a more comprehensive format for investigation of this
issue, since both input and output data cover the entire modeling region
rather than selected trajectories.
SPECIFIC MODEL ASSESSMENTS
This section contains brief descriptions of selected ozone models for
which versions are currently in the public domain. In many cases the
models undergo periodic update and modification, so it is advisable to
check with the developer of a given model to obtain details of any
recent changes. The descriptions presented below are based on current
versions of the models.
The specific models reviewed are as follows:
o Linear Rollback
o EKMA, developed by EPA
o City-specific EKMA using Ozone Isopleth Plotting Package
(OZIPP) developed by EPA
o Hanna ATDL, developed by Atmospheric Turbulence and Diffusion
Laboratory, NOAA
o ELSTAR, developed by Environmental Research and Technology,
Inc.
o TRACE, developed by Pacific Environmental Services, Inc.
o LIRAQ, developed by the Lawrence Livermore Laboratory
o SAI Urban Airshed Model, developed by Systems Applications,
Inc.
o IMPACT, developed by Science Applications, Inc.
o MADCAP, developed by Science Applications, Inc.
This sampling of available ozone models is not intended to be
all-inclusive; rather, it is intended to be representative of the most
* widely-applied or most generally-accessible models available at this
time. There have been numerous attempts over the past several years to
develop ozone models, and the previous reviews cited can be consulted
for appropriate references. There may be other perfectly adequate
models in existence or currently under development which were excluded
from this assessment; conversely, inclusion of a particular model in
this assessment does not necessarily constitute endorsement of its use
for any specific application.
6
-------
Each model description includes a brief history of the model's
development and previous applications, basic approach, variables
considered, input data requirements, computer requirements (based on
actual operating experience) and references for further documentation.
Following each description is a discussion of the model's applicability
to each of the ozone control strategy issues previously discussed. The
results of the evaluation are summarized in simplified form in Table 1.
This table is intended to provide summary information which supplements
the more detailed discussion to follow, and should not be interpreted as
providing a relative "ranking" of the models listed.
7
-------
TABLE 1. SUMMARY OF MODEL ASSESSMENTS
Ozone Control Strategy Issues
Model
Rollback
EKMA, standard
EJCMA, city-speelfic
(OZIPP)
HANNA ATDL
E/S
Box
ELSTAR
TRACE
LIRAQ
SA1 Airshed
IMPACT
MADCAP
Type of Hodtl
E/S - Mp1r1cal/tt*t1st1cal
LT « laaranglan Trajectory
EG - £ulerlan 6r1d
Issue Ratings
1. Demonstrated applicability (previous verification performed)
2. Theoretically applicable. United by data availability (I.e.,
no previous verification performed)
J. Theoretical limitations (applicable only in limited
clrcuestances)
4. Not applicable
Data Requirements Ratings
a aggregated amissions data, no meteorological data
b spatially, temporally, and species disaggregated missions
data, meteorological data along trajectory corridors
c spatially, temporally, and species disaggregated amissions
data, meteorological data throughout urban study area
Cost Ratings
A Inexpensive (little or no computer requirement)
B moderate cost (moderate computer requirement, cost per run
ranges from $10 to $100)
C expensive (substantial computer requirements, cost per run
ranges from $100 to $1,000)
Output Ratings
A extensive output produced (e.g., hourly concentration fields
for entire modeling region)
B moderate output produced (e.g., concentrations along
specified trajectories within the modeling region)
C limited output produced (e.g., percent emission reduction
required, or expected ozone maxlain)
8
-------
ROLLBACK
MODEL FEATURES
Model Type
Statistical
Historical Data
The concept of regional emissions rollback (deNevers, 1975 and
Roth, 1976) had been practiced for many years prior to its sanction
for use in the preparation of State Implementation Plans mandated
by the Clean Air Act of 1970. During the 1970s rollback was used
extensively for this purpose (Cheng, 1975 and DeMandel, 1978); many
variations of the rollback technique have been devised but only
linear rollback with corrections for background and transport has
been accepted by the EPA.
Description
In rollback models the emissions/air quality relationship is
simplistically displayed on a two-dimensional graph as shown in
Figure 1. Using one of these rollback curves and knowing the
highest or second highest one-hour average oxidant measurement
during a given year, one could infer the percentage control of
hydrocarbon emissions needed to achieve the ozone standard.
The technical basis of these models suffers from major
deficiencies. Among these are the following:
- The role of NOx emissions in the formation of ozone is
ignored.
- Varying photochemical reactivities for different organic
compounds are ignored.
- The air quality effect of control strategies which result
in non-uniform emission reductions over a given region
cannot be evaluated (i.e., emission reductions must be
assumed to occur uniformly).
. Spatial Resolution
Horizontal - none
Vertical - none
Temporal Resolution
None
9
-------
LINEAR ROLLBACK
O ,O0 JO /2- -te ¦/# & -2Z 26 .70 -2K7 -3f ^
MAXIMUM 1-HOUR PHOTOCHEMICAL OXIDANT CONCENTRATION (ppm)
Figure 1. Comparison of Hydrocarbon Rollback Models
-------
Meteorology
None
Chemistry
None
Input Data Requirements
Meteorological Variables - None
Emissions Data - annual, aggregated regional non-methane
hydrocarbon inventory.
Air Quality Data - second highest ozone level in some selected base
year.
Key Output
Percent reduction of regional non-methane hydrocarbon emissions
required to achieve the ozone standard.
Approximate Computer Costs
No computer necessary.
11
-------
APPLICABILITY OF ROLLBACK TO OZONE CONTROL STRATEGY ISSUES
Hydrocarbon vs NOx Control
Rollback techniques are not applicable to this issue because the
role of NOx emissions in the formation of ozone is ignored.
Assessing Attainment of the Ozone Standard and a Short-Term N02 Standard
The principle of rollback can be applied to estimate the control
requirements necessary to meet either an ozone standard via
hydrocarbon emission reduction or an N02 standard via reductions in
NOx emissions. But because rollback does not acknowledge the
chemical Interaction of hydrocarbons and NOx, it cannot be used to
weigh this issue.
Control of Low-Level vs Elevated Emissions
Rollback 1s applied to regionally aggregated emission inventories
which are not defined by sources. It cannot be used to address
this Issue.
Selective Control of Hydrocarbon Compounds
The rollback technique does not make any provisions for various
degrees of hydrocarbon reactivity and thus cannot be used to
address this Issue.
Long Range Transport
Rollback does not incorporate a transport mechanism and thus cannot
be used to address this Issue.
Effects of Spatial and Temporal Variation of Ozone Precursor Emissions
Annual, regional emission inventories are used in rollback
techniques. The model cannot be used to address this Issue.
12
-------
STANDARD EKMA (EMPIRICAL KINETIC MODELING APPROACH)
MODEL FEATURES
Model Type
Empirical
Historical Data
EKMA was released by EPA in November, 1977. Since then, several
studies have been performed to compare EKMA results with other
techniques (deNevers, 1978 and DeMandel, 1978) and numerous states
have used the model in the development of their State
Implementation Plans.
Description
EKMA was developed by the EPA as a more advanced ozone prediction
relationship than ROLLBACK. EKMA is a series of isopleths that
express the relationship between early morning downtown NOx and
hydrocarbon levels and the peak ozone levels downwind of a city
center under a given set of meteorological, emissions and air
quality conditions (See Figure 2) (U.S. EPA, 1977). Theoretically,
given a sufficiently large data base of observed early morning NOx
and NMHC levels with the corresponding maximum afternoon ozone
levels at a selected site and with a single meteorological
scenario, a similar set of isopleths can be drawn for any site.
In EKMA, the reduction in a city's design ozone concentration
caused by a given emissions reduction is assumed to be equal to the
reduction that can be derived through the use of the plotted
curves. This equality is assumed regardless of what conditions led
to the city's design ozone concentration. A key output of EKMA is
the effect of given HC and NOx emission reductions on expected
maximum ozone levels.
For general applications, the EPA has developed a standard set of
EKMA isopleths based on a predefined set of conditions (U.S. EPA,
1978). For more specific applications, a computer program is
available to generate city-specific curves (U.S. EPA, 1978) which
more closely match those conditions under which the design ozone
concentration actually occurred. In both cases, the isopleths are
derived using a kinetics model which is based essentially on a
Lagrangian concept. The differences between a kinetics model and a
Lagrangian model are 1) the formulations used to consider emissions
and mixing height are different, and 2) the kinetics model is
customarily applied with cruder assumptions than are used with
Lagrangian models.
13
-------
O.B
O.t
/.0
O.f
0.20
'6 /0
NMHC (ppwt)
-------
Spatial Resolution
Horizontal - none
Vertical - none
Temporal Resolution
None
Meteorology
None - A predefined meteorological scenario is assumed.
Chemistry
None - A predefined mix of hydrocarbon species and reactions is
assumed.
Input Data Requirements
¦v
Meteorological Variables - none
Emissions Data - annual, aggregated non-methane hydrocarbon and NOx
inventory.
Air Quality Data - second highest ozone level in some selected base
year and the corresponding level of morning NMHC and NOx.
. Key Output
Percent reduction of non-methane hydrocarbons and/or NOx to achieve
the ozone standard.
Approximate Computer Costs
No computer necessary.
15
-------
APPLICABILITY OF STANDARD EKMA MODEL TO OZONE CONTROL STRATEGY ISSUES
Hydrocarbon vs NOx Control
Because EKMA relates hydrocarbon and NOx levels to ozone
concentrations, it can be used to address this issue. The
isopleths are intended to be used to evaluate the effect of
relative changes in regional non-methane hydrocarbon and NOx
emissions on peak ozone levels. However, EKMA was not designed to
predict absolute ozone values as a function of hydrocarbon and NOx
levels and should not be thus applied.
Assessing Attainment of the Ozone Standard and a Short-Term N02 Standard
In theory EKMA can be used to generate a set of N02 isopleths
because a chemical kinetic mechanism is used to express the HC/NOx
relationship. These isopleths could then be used to evaluate
whether achieving an N02 standard is counterproductive to reducing
ozone levels or vice versa. The EPA 1s now developing a set of N02
Isopleths.
Control of Low-Level vs Elevated Emissions
EKMA employs a regionally-aggregated emissions inventory which does
not distinguish between elevated and low-level sources. Therefore,
the standard EKMA, cannot be used to address this issue.
Selective Control of Hydrocarbon Compounds
Because EKMA utilizes a predetermined mix of non-methane
hydrocarbon compounds 1t cannot be used to address this issue.
Long-Range Transport
The standard EKMA isopleths are based on (and thus limited by) a
ten-hour simulation along a hypothetical trajectory. The short
simulation time is generally insufficient to treat long-range
transport issues.
Effect of Spatial and Temporal Variation of Ozone Precursor Emissions
A predefined set of spatial and temporal emissions variations are
assumed for EKMA standard method. Thus it is not applicable to
this issue.
16
-------
CITY-SPECIFIC EKMA USING OZIPP (OZONE ISOPLETH PLOTTING PACKAGE)
MODEL FEATURES
Model Type
Lagrangian Trajectory
Historical Data
See standard EKMA
Description
The city-specific EKMA (U.S. EPA, 1978) is patterned after a
trajectory model. It is most applicable to cities with simple,
uniform meteorology and emissions distribution, that can be
characterized by a single trajectory. For this model, the
trajectory is predetermined. With the city-specific EKMA it is
possible to generate a set of ozone isopleths which more closely
matches those conditions under which the design ozone concentration
actually occurred.
Spatial Resolution
Horizontal - the precursor emissions can be apportioned along the
predetermined trajectory path.
Vertical - none
Temporal Variation
Hourly precursor emissions are input to the city-specific EKMA.
Peak ozone levels, because of the present meteorology, are assumed
to occur in the early afternoon hours and are reported hourly.
Meteorology
A single scenario of a sunny day is assumed. A predetermined
trajectory precludes the need for developing a wind field.
Chemistry
76-step reaction.mechanism
three hydrocarbon classes: propylene, butane, and adehydes
Input Data Requirements
Meteorological Variables - mixing height, latitude, date.
Emissions Data - distribution of total precursor emissions along
the trajectory on an hourly basis for NOx, N02, propylene, butane,
and aldehydes.
17
-------
Air Quality Data - initial concentrations for all input species at
the city center at the start of the simulation and for precursors
and ozone along the trajectory.
Key Outputs
Set of isopleths showing the relationship between NOx and
hydrocarbons for different resultant oxidant levels.
Approximate Computer Requirements
To generate a complete set of city specific isopleths (ten-hour
simulations) requires seven minutes on a Univac 1110 along with 38K
bytes of core. A single trajectory requires ten seconds of
computation time.
18
-------
APPLICABILITY OF THE CITY-SPECIFIC EKMA MODEL (OZIPP), TO OZONE CONTROL
STRATEGY ISSUES
Hydrocarbon vs NOx Control
EKMA essentially attempts to relate the early morning precursor
(hydrocarbons and NOx) concentrations to peak ozone concentration.
Each precursor level can be adjusted independently of the other.
Therefore, the city-specific EKMA can be used to address this
issue.
The isopleths are designed to be used to evaluate the effect of
relative changes in hydrocarbon and/or NOx emissions on the
expected peak ozone levels. EKMA was not designed to estimate the
absolute ozone values as a function of hydrocarbon and NOx levels
and should not be thus applied.
Assessing Attainment of the Ozone Standard and a Short-Term N02 Standard
In theory EKMA can be used £o generate a set of N02 isopleths
because a chemical kinetic mechanism is used to express the HC/NOx
relationship. These isopleths would then be used to evaluate
whether achieving an N02 standard is counterproductive to reducing
ozone levels or vice versa. The EPA is now developing a set of N02
isopleths.
Control of Low-Level vs Elevated Emissions
The city-specific EKMA can not be used to address this issue. It is
a single layer model and thus can not differentiate between
emission sources at varying heights.
Selective Control of Hydrocarbon Compounds
Hydrocarbons are divided into three classes, propylene, butane, and
aldehydes, and the amount of each emission can be varied to study
their effect on ozone formation. Differences in the effectiveness
of control programs directed at selected species within any one of
these classes cannot be readily assessed.
Long-Range Transport
EKMA is limited to 10-hour simulations along a predetermined
trajectory. Long-range transport implies multi-day simulation
capabilities which EKMA does not possess. It is not capable of
addressing this Issue because it was not so designed.
Effect of Spatial and Temporal Variation of Ozone Precursor Emissions
The city-specific EKMA can be used to address the question of
spatial and temporal variability along the trajectory path. The
emissions input to the model are spatially resolved along the
trajectory path and temporally resolved on an hourly basis.
19
-------
HANNA ATDL PHOTOCHEMICAL MODEL (ATMOSPHERIC TURBULENCE AND
DIFFUSION LABORATORY)
MODEL FEATURES
Model Type
Box
Historical Data
The Hanna ATDL model was released in 1973. No updates have been
issued since. Model application has consisted of comparing its
results against observations and predictions of more complex
models.
Description
The ATDL is a stationary box model (Hanna, July, 1973 and Hanna,
October, 1973). A volume of-air is assumed to be the reaction
vessel. Within this box, emissions are assumed to be mixed
instantaneously. A gross representation of ventilation is
simulated by a constant unidirectional wind. Fixed mixing heights
are used for each stability classification. Concentrations of
unreactive compounds are assumed to be directly proportional to
each species' emission strength and inversely proportional to the
wind, which is moving a quantity of pollutants out of the box per
unit time.
To represent photochemical reactions, it is assumed that for any
mechanism, a related set of dimensionless equations can be written
relating the precursors and ozone levels. By analyzing the
dimensionless quantities, the relative importance of the chemical
reaction, meteorology and source strength to that species can be
eval uated.
The ATDL technique of dimensionless equation analysis was applied
to the five-equation photochemistry simulation developed by
Friedlander and Seinfeld (Friedlander, 1969). This reaction
mechanism consisted of three simple reactions for N02, NO and
reactive hydrocarbons, plus two proportional equations, one of
which is for ozone. By making the equations dimensionless and
making various assumptions, such as the emissions ratio of NO to
N02, the dimensionless quantities are solvable by hand. Knowing
these, the equilibrium concentration of the five pollutant species
can be obtained.
The disadvantages of this model are that it uses a simple advection
scheme, eliminates dispersion and further simplifies an already
simple chemical kinetics mechanism. Its advantage is that the
various equations can be solved by hand without the use of
computer.
21
-------
Spatial Resolution
Horizontal - single area cover of about 100 km2
Vertical - the height of the single layer box is a function of
atmospheric stability. It is essentially the height of the mixed
layer.
Temporal Resolution
One hour
Meteorology
A unidirectional wind is used to ventilate the reaction "box". The
purpose of the wind is to limit the residence and hence reaction
time of the precursors.
Chemistry
The ATDL model has used the Friedlander-Seinfeld chemistry
(Friedlander, 1969) which consists of three reactions for N02t NO
and reactive hydrocarbons and two proportional relationships, one
of which 1s for ozone calculations.
Input Data Requirements
Meteorological Variables - wind speed
Emissions Data - hourly emissions data for the box area for each of
the species involved 1n the chemical reactions
A1r Quality Data - none
Key Output
Predictions of ozone, NO, N02 and reactive hydrocarbon levels for
the study area.
22
-------
APPLICABILITY OF THE HANNA ATDL PHOTOCHEMICAL MODEL TO OZONE CONTROL
STRATEGY ISSUES
Hydrocarbon vs NOx Control
The Hanna ATDL model uses a simplification of the 1969
Friedlander-Seinfeld chemistry to predict ambient concentrations of
N02, NO and reactive hydrocarbons as a function of each species'
source strength. Ozone is calculated as the ratio of the predicted
N02 to NO levels times a constant based on observed data. Thus to
have valid ozone results, the N02 and NO predictions must be
accurate.
In one study (Hanna, July, 1973), the model was shown to predict NO
and reactive hydrocarbon values reasonably well. However, the N02
results showed little agreement with the observed levels. This is
consistent with the discrepancy between estimated and observed
ozone concentrations because of the dependency of the estimated
ozone level on the estimated N02/N0 ratio.
Because the model (based on the Friedlander-Seinfeld) chemistry has
not produced accurate N02 and thus ozone predictions, it cannot be
used at this time to address this issue even though the ATDL model
has the capability to accept varying hydrocarbon and NOx emissions.
Assessing Attainment of the Ozone Standard and a Short-Term N02 Standard
The model has not performed well in predicting either N02 or ozone
levels, and thus should not be used for this purpose without
further development and/or validation testing.
Control of Low-Level vs Elevated Emissions
The ATDL is a single-layer, well-mixed box model. Emissions are
assumed to immediately disperse uniformly within the confines of
the study area. Thus no distinction is made between low-level and
elevated sources. The ATDL model is not applicable to this issue.
Selective Control of Hydrocarbon Compounds
A single hydrocarbon class is used for the photochemical
simulations. The model is not applicable to this issue.
Long-Range Transport
The model is restricted to predictions within a stationary square
area. It cannot be used to address this issue.
Effects of Spatial and Temporal Variation of Ozone Precursor Emissions
The model accepts hourly aggregated NO, N02 and reactive
hydrocarbon emissions. It can address questions of temporal
variations by asuming different emission rates for each
calculation. Because the Hanna ADTL model is a well-mixed box
model, the spatial variation of emissions cannot be evaluated.
-------
ELSTAR
(ENVIRONMENTAL LAGRANGIAN SIMULATOR OF TRANSPORT
AND ATMOSPHERIC REACTIONS)
MODEL FEATURES
Model Type
Lagrangian Trajectory
Historical Data
ELSTAR is the name of the photochemical trajectory model delivered
to the Coordinating Research Council. A model similar to ELSTAR
has been delivered to EPA who plans to release it in the near
future (original release date March, 1979) [Lloyd, December, 1978],
This set of computer codes supersedes the DIFKIN model which was
developed in 1973 (Martinez, 1973).
The DIFKIN version has been used in many applications, particularly
for highway impact studies (Allen, 1976). ELSTAR has been applied
to the St. Louis area (Lloyd, December, 1978).
Spatial Resolution
Horizontal - variable, user-specified
Vertical - the distance between the surface and stable layer can be
divided into a number of equal height layers of any desired depth.
Ten layers have been used.
Temporal Resolution
Hourly time steps for input and output.
Meteorology
ELSTAR uses a mass-consistent wind field along a trajectory defined
by the flow of the wind and bounded in the vertical by the ground
and mixing height. Within the wind field, vertical mixing between
the layers is assumed to be due solely to turbulence.
* Chemistry
45-step reaction mechanism
Five hydrocarbon classes: paraffins, olefins, aromatics, higher
aldehydes and formaldehyde.
25
-------
Input Data Requirements
Meteorological Variables - wind speed and direction, temperature,
humidity, diffusivity coefficients as a function of height,
inversion height field, longitude and latitude and date.
Emissions Data - an hourly gridded emission inventory for the study
area. Emissions of the following seven compounds must be input:
NOx, CO, paraffins, olefins, aromatics, formaldehyde and higher
aldehydes.
Air Quality Data - concentrations for all input species plus ozone
at the start of the trajectory. Model verification requires
concentrations of all input species and ozone along the trajectory
path.
Key Outputs
Concentrations of ozone and other user-selected species along the
trajectory path at hourly intervals for each vertical layer. One
set of outputs is produced per trajectory.
Approximate Computer Requirements
A twelve-hour simulation requires 40 to 60 seconds of computer time
on a CDC 7600 per trajectory. The number of trajectories required
to fully characterize an area 1s a function of the complexity of
the terrain and the heterogeneity of the spatial distribution of
the emission sources in the study area. The observed pollutant
concentration fields would indicate the spatial variability of the
area and the number of trajectories necessary to fully define the
study zone. In areas where a great number of trajectories are
required, use of an Eulerian model might be more economical.
26
-------
APPLICABILITY OF ELSTAR TO OZONE CONTROL STRATEGY ISSUES
Hydrocarbon vs NOx Control
ELSTAR has a photochemical simulation module which relates
hydrocarbon and NOx emissions to resultant ozone concentrations.
It can be used to address this issue. DIFKIN, the predecessor
model, has been used to weigh the merits of varying degrees of
hydrocarbon and/or NOx emission control on ozone formation (Allen,
1976).
Assessing Attainment of the Ozone Standard and a Short-Term N02 Standard
ELSTAR can predict hourly N02 and ozone values along the selected
trajectory. It therefore can be used to evaluate the emission
reduction necessary to meet both standards. These control
strategies are often detrimental to the other pollutant. For the
St. Louis study (see ref. Lloyd, et. al., 1978) the model's N02
prediction capability was a major evaluation criterion.
Control of Low-Level vs Elevated Emission
ELSTAR distinguishes between low-level and elevated emission
sources in its calculations. Emissions are dispersed into that
vertical grid cell which corresponds to the effective plume rise of
the individual source. Thus, ELSTAR 1s applicable to evaluating
the effectiveness of control strategies aimed at either low level
or elevated emission sources, or combinations of the two.
Selective Control of Hydrocarbon Compounds
ELSTAR can be used to address this issue because hydrocarbons are
disaggregated Into five classes before being input to the
photochemistry module. Emissions of each class can be
independently varied, simulating selective hydrocarbon control
schemes, to study their effect on ozone formation.
Long-Ranqe Transport
Because ELSTAR is a trajectory model, the effects of horizontal
turbulence and shear forces in all three dimensions are assumed to
be negligible. In other words, the air parcel is assumed to remain
intact as it traverses along the trajectory. For short distances,
this can be a valid assumption. However, as the travel time and
distance become greater, the effect of horizontal shear becomes
more pronounced. One way to reduce the effect of this force is to
go to a coarser grid scale. This would have the effect of damping
the peak concentration because the polluants would be averaged over
a larger area. Another problem area is how a trajectory model
would handle transport of pollutants 1n an inversion layer, which
is an area of high wind shear.
27
-------
Before ELSTAR can be used to address this issue, further field
validation would be needed.
Effects of Spatial and Temporal Variation of Ozone Precursor Emissions
ELSTAR can be used to evaluate control strategies based on this
issue because the model retains the temporal and spatial
characteristies of emissions in its calculation. However, these
variations only affect the ozone results if the trajectories pass
directly over the source(s) being altered.
28
-------
TRACE
(TRAJECTORY ATMOSPHERIC CHEMISTRY AND EMISSIONS)
MODEL FEATURES
Model Type
Lagranglan trajectory
Historical Data
TRACE, released in November, 1977 (Drfvas, March, 1977 and Drivas,
November, 1977) is the latest version of the REM I (Wayne, 1973)
and REM II (Drivas, February, 1977) trajectory models first
developed in 1973. TRACE and its predecessors have been used 1n
the preparation of many environmental impact reports for large
point sources such as offshore oil terminals.
Spatial Resolution
Horizontal - variable, user-specified
Vertical - none, a single well-mixed layer 1s assumed
Temporal Resolution
One-hour Intervals for Input and output
Meteorology
A mass-consistent flow field defined In the vertical by the ground
and the Inversion base represent the trajectory of the air parcel.
Chemlstry
Uses the EKMA chemistry (76-step reaction mechanism).
Three hydrocarbon classes: propylene, butane, and aldehydes.
Input Data Requirements
Meteorological Variables - wind speed and direction, temperature,
humidity, inversion height and date.
Emissions Data - a grldded hourly emission Inventory for all
emissions disaggregated into NO, N02 and the three hydrocarbon
classes.
Air Quality Data - concentrations for all input species and ozone
at the start of the trajectory. For model verification,
concentrations of all input species and ozone along the trajectory
path are needed.
29
-------
Key Outputs
Tabulation of predicted ozone and user-selected species
concentrations on an hourly basis along the trajectory. One set of
outputs is produced per trajectory.
Approximate Computer Requirements
A 12-hour simulation of a single trajectory requires about 3
minutes of calculation time on an IBM 370/158.
30
-------
APPLICABILITY OF TRACE TO OZONE CONTROL STRATEGY ISSUES
Hydrocarbon vs NOx Control
TRACE uses the EKMA chemistry module for photochemical simulations.
This chemistry is suited to evaluating the ozone reduction due to
decreases in hydrocarbon and/or NOx emissions; thus, TRACE is
applicable to this issue.
Assessing Attainment of the Ozone Standard and a Short-Term N02 Standard
TRACE can be used to predict both the N02 and ozone concentrations
resulting from hydrocarbon and NOx emissions. Because of this, it
can be used to evaluate the tradeoffs between reductions of
hydrocarbon or NOx emissions necessary to achieving both standards.
Control of Low-Level vs Elevated Emission
TRACE is a single-layer trajectory model. It resembles a reaction
chamber which is being blown' along a path defined by the wind.
Emissions entering the air parcel are assumed to become uniformly
and instantaneously mixed. No distiction is made between low-level
and elevated sources. Therefore the model cannot be used to
address this Issue.
Selective Control of Hydrocarbon Compounds
The TRACE chemical kinetic mechanism is the same as that used in
the c1ty-spec1fic EKMA. It distinguishes three classes of
hydrocarbons, butane, propylene, and aldehydes. Each can be varied
to determine its effect on oxidant formation along a selected
trajectory and for a given meteorological scenario, and selective
hydrocarbon control schemes can be simulated.
Long-Range Transport
TRACE is a single-layer trajectory model based on the assumption
that a single well-mixed column of air traveling along a trajectory
will stay intact. In actuality, this column of air 1s acted upon
in all three dimensions by wind shear and diffusion. This
assumption, together with TRACE'S inability to simulate transport
in the inversion field because 1t is limited to a single layer,
severely limits the model for addressing this issue.
Effects of Spatial and Temporal Variation of Ozone Precursor Emissions
TRACE can be used to evaluate spatial and temporal variations in
emissions along its trajectory path.
31
-------
LIRAQ-2
(LIVERMORE REGIONAL AIR QUALITY MODEL, VERSION 2)
MODEL FEATURES
Model Type
Eulerian grid
Historical Data
LIRAQ-1, a nonphotochemical version, was published 1n 1972. This
was supplemented in 1975 by a photochemical version called LIRAQ-2
(MacCracken, 1975, MacCracken, 1978, and Bass, 1977). The only
difference in the two is that the latter has a photochemical
simulation module. In 1977, the LIRAQ-2 version was updated to
include sulfate chemistry simulation. Both LIRAQ-1 and 2 were used
to develop the Air Quality Maintenance Plan (ABAG, 1977 and Duewer,
1978) and the 1979 Non-Attainment Area Plan for the San Francisco
Bay Area. It is presently being applied to the St. Louis area.
Spatial Resolution
Horizontal - variable, user-specified. In previous applications,
one to five kilometer square grids arranged In a 20 x 20 square
array have been used. The largest areal coverage studied was 100
km x 100 km using a five kilometer grid.
Vertical - single layer from the surface to the top of the mixed
layer with an imposed, fitted concentration profile.
Temporal Resolution
Variable time step, graphical output available at 30-m1nute
Intervals.
Meteoroloqy
LIRAQ uses a mass-consistent flow field specifying depth of mixed
layer, wind speed and direction for each grid cell; this field 1s
interpolated for each hour from fields produced at three-hour
intervals. An additional code, MASCON, 1s available to derive flow
fields from hourly-averaged wind and mixing depth data.
Chemistry
48-step reaction mechanism
three tydrocarbon classes: alkanes, alkenes (olefins), aldehydes
33
-------
Input Data Requirements
Meteorological Variables - wind speed, direction, temperature,
humidity, inversion height, topography, latitude, longitude, date.
Emissions Data - gridded hourly emissions inventory for area and
point sources within the study area disaggregated into seven
species: alkanes, alkenes, aldehydes, nitric oxide (NO), carbon
monoxide (CO), sulfur dioxide (S02) and particulates. The first
four species are used in the photochemical simulation.
Air Quality Data - model operation requires ambient concentrations
of all emitted species plus ozone at the boundaries all day and for
the entire study area for the first hour simulated. Verification
further requires a thorough sample of ozone, N02 and other species
concentrations in several cells for each hour of the prototype day
being simulated.
Key Outputs
Tabular data provided for- ozone and a variety of species.
Concentration fields for all species calculated by LIRAQ can be
output as isopleth maps.
Hour-by-hour plots of pollutant concentrations for oxidant, N02,
and each input species at user-specified grid locations.
Approximate Computer Requirements
A 24-hour simulation with a 20 x 20 x 1 grid requires 60 minutes of
machine time and 1 million bytes on a CDC 7600.
34
-------
APPLICABILITY OF LIRAQ-2 TO OZONE CONTROL STRATEGY ISSUES
Hydrocarbon vs NOx Control
Because LIRAQ-2 does have an algorithm to simulate the
photochemical reactions between hydrocarbons and NOx to form ozone,
it can be used to address this issue.
Assessing Attainment of the Ozone Standard and a Short-Term N02 Standard
LIRAQ is theoretically applicable to addressing the attainment of
both an ozone and an N02 standard because of its photochemistry
module. It can evaluate the tradeoffs between an optimal ozone and
an optimal N02 control strategy to meet both standards.
However, before such a decision can be made with LIRAQ, some
refinement of the N02 predictions must be made. In previous
experience with the model its simulation of N02 levels on high
ozone days (as opposed to high N02 days) is generally not as good
as the ozone simulation.
Control of Low-level vs Elevated Emission
In its emission inventory, LIRAQ does require a separate file for
low-level and elevated emission sources though it is basically a
single-layer model. Elevated sources are assumed to affect only
the vertical average concentration if below the inversion layer.
The amount of the emissions used in the calculations are dependent
on the height of the Inversion base (Duewer, 1978). These sources
are not included in the calculation of the vertical profile which
1s used to determine the ozone concentration within a cell.
Therefore, LIRAQ is not applicable to this issue.
Selective Control of Hydrocarbon Compounds
The LIRAQ chemical kinetic mechanism distinguishes three classes of
hydrocarbon source emissions. Variations in emissions of these
three classes can be tested for their effect on ozone formation.
LIRAQ's applicability to this issue is limited to the types of
variations that can be simulated with the three hydrocarbon
classes.
Long-Range Transport
LIRAQ could be modified to evaluate transport of pollutants into
adjacent regions by increasing the grid size to large distances,
e.g., ten km. However, 1f there is a significant possibility that
the transport mechanism involves some vertical transport component
rather than simple horizontal flow, or transport in the inversion
layer which is subjected to significant wind shear (a simulation
capability that LIRAQ lacks) then LIRAQ could not be applied.
35
-------
Effects of Spatial and Temporal Variation of Ozone Precursor Emissions
Since LIRAQ is a deterministic model which retains the spatial and
temporal characteristics of emissions in its calculations, it is
well-suited to evaluating the effects of variation of these
parameters on ozone formation within the area covered by the grid.
36
-------
SAI AIRSHED
MODEL FEATURES
Model Type
Eulerian grid
Historical
The SAI AIRSHED model has been in continuous development since the
early 70's. In 1973, it was first used for a study of the Los
Angeles area (Reynolds, 1973). Since that study many changes have
been made to the model including a new photochemistry module.
These modifications were all documented for the present version
(Reynolds, 1978 and Reynolds, 1979). AIRSHED has been applied to
many areas including Los Angeles, Denver, Sacramento, and St.
Louis.
Spatial Resolution
Horizontal - variable, user-sped fled. AIRSHED has been run with 1
to 10 km square grids arranged rectangularly to cover the study
area. The largest study area was Los Angeles with a grid of
roughly 300 x 150 km.
Vertical - up to 10 vertical cells of these usually at least one is
within the mixed layer and at least one 1s in the Inversion.
Temporal Resolution
User-specified resolution of Input and output.
Meteorology
Mass-consistent three-dimensional wind field Including the depth of
the mixed layer, wind shear between layers and cells and wind speed
and direction 1n three dimensions for each grid cell for each
layer. At present, one of the wind field input preparation
algorithms is intended for use in areas where thermal,
topographical and frictional effects on wind flow are not
pronounced. Another code more applicable to a complicated terrain
• is now being evaluated; in addition, a straightforward inverse
distance weighted interpolation scheme is also available.
Chemistry
31-step reaction mechanism
Four hydrocarbon classes: single-bonded carbon atoms, very
reactive double-bonded carbon atoms, moderately reactive
double-bonded atoms and carbonyl-bonded carbon atoms. These types
37
-------
of bonds are characteristic of paraffins, olefins, aromatics and
aldehydes, respectively.
Input Requirements
Meteorological Variables - wind speed and direction at the surface
and for each vertical layer; inversion height field; insolation
observations; temperature, pressure, relative humidity and
topography; latitude and longitude and date.
Emissions Data - a gridded hourly emission inventory for all area
and point sources located within the study area. Emission rates
for the following species must be included for the photochemical
simulation: NOx, CO, paraffins, olefins, aromatics and aldehydes.
Air Quality Data - model operation requires ambient concentrations
of all emitted species plus ozone at the boundaries all day and for
the entire study area for the first hour simulated. Verification
further requires a thorough sample of ozone, N02 and other species
concentrations in several cells for each hour of the prototype day
being simulated.
Key Outputs
Tabular concentrations of ozone and each input species at hourly
intervals (option available through the developer for producing
isopleth plots of each species).
Approximate Computer Requirements
A 14-hour simulation with a 25 x 25 x 4 grid requires 20 minutes of
machine time and nearly 1 million bytes on a CDC 7600.
38
-------
of bonds are characteristic of paraffins, olefins, aromatics and
aldehydes, respectively.
Input Requirements
Meteorological Variables - wind speed and direction at the surface
and for each vertical layer; inversion height field; insolation
observations; temperature, pressure, relative humidity and
topography; latitude and longitude and date.
Emissions Data - a gridded hourly emission inventory for all area
and point sources located within the study area. Emission rates
for the following species must be included for the photochemical
simulation: NOx, CO, paraffins, olefins, aromatics and aldehydes.
Air Quality Data - model operation requires ambient concentrations
of all emitted species plus ozone at the boundaries all day and for
the entire study area for the first hour simulated. Verification
further requires a thorough sample of ozone, N02 and other species
concentrations in several cells for each hour of the prototype day
being simulated.
Key Outputs
Tabular concentrations of ozone and each input species at hourly
intervals (option available through the developer for producing
isopleth plots of each species).
Approximate Computer Requirements
A 14-hour simulation with a 25 x 25 x 4 grid requires 20 minutes of
machine time and nearly 1 million bytes on a CDC 7600.
38
-------
APPLICABILITY OF AIRSHED TO OZONE CONTROL STRATEGY ISSUES
Hydrocarbon vs NOx Control
With its photochemical simulation capabilities, AIRSHED can be used
to assess the validity of control schemes aimed at reducing
hydrocarbons and/or NOx to achieve the ozone standard. Therefore,
AIRSHED is applicable to this issue. It was used in Denver to
assess various HC/NOx control schemes (Anderson, 1977).
This issue has been specifically addressed in several studies: in
the Los Angeles region, four schemes for controlling vehicular HC
and/or NOx were evaluated; in Denver, various HC and/or NOx control
schemes were examined; and various HC/NOx control schemes based on
demographic changes for a number of areas were studied.
Assessing Attainment of the Ozone Standard for a Short-Term N02 Standard
AIRSHED can predict N02 leveTs in addition to ozone so that it can
evaluate the control strategies necessary for achieving a
short-term N02 standard and an ozone standard. The model has been
used to predict hourly N02 levels in Los Angeles (Reynolds, 1973),
Denver (Anderson, 1977) and St. Louis. The model performance for
these N02 predictions has generally been found to be somewhat
poorer than for ozone. However, N02 concentrations on these study
days were generally relatively low. There is a need to further
establish the model's N02 performance characteristics.
Control of Low-Level vs Elevated Emissions
AIRSHED is applicable to this issue because 1t is a multi-layer
model in the vertical, and it can distinguish between low-level and
elevated emissions. Emissions are added to the layer corresponding
to their total effective plume rise. If this height exceeds the
top of the modeling region, the pollutants are excluded from
subsequent AIRSHED calculactions. For ground-level sources,
emissions are added to the-lowest layer. The model has been used to
evaluate the impact of emissions from major elevated point sources
on regional air quality.
Selective Control of Hydrocarbon Compounds
AIRSHED separates hydrocarbons into four classes according to the
reactivity of the carbon bond. The distribution of the actual
hydrocarbons can be changed to reflect control strategies. This
would result in a change in each hydrocarbon category that contains
the actual emissions. AIRSHED can be used to answer the question
of hydrocarbon reactivity control.
39
-------
Long-Range Transport
AIRSHED has been developed with enough generalities and includes
treatment of wind shear and transport in the inversion layer so
that it can theoretically be used to evaluate the transport of
pollutants from one air basin into an adjacent region by expanding
the grid size to ten km, for example. (The largest region to which
the model has been applied is 150 to 300 km. on a side.) With the
expanded scale, the adjacent region would become a part of the
study area. Before this model or any model can be applied to a
long-range transport analysis, it must be verified.
Effects of Spatial and Temporal Variation of Ozone Precursor Emissions
Because AIRSHED is a deterministic Eulerian model, its emissions
must be defined with spatial and temporal characteristics. Thus,
it is well suited to studying the effects of varying the diurnal
emission cycle and of varying the spatial distribution of
precursors on ozone formation within the study area. AIRSHED has
been used for this purpose in^the Los Angeles (Anderson, 1977 ) and
Denver (Reynolds, 1977) areas.
40
-------
IMPACT
(INTEGRATED MODEL FOR PLUMES AND ATMOSPHERICS
IN COMPLEX TERRAIN)
MODEL FEATURES
Model Type
Eulerian grid
Historical Data
IMPACT (Fabrick, 1977 and Ranzieri, 1979) was released in March
1977. It is an update of the DEPICT model (Sklarew, 1976). Since
its release, the model has been used for a study in the Sacramento
area (Amar, 1978).
Spatial Resolution
Horizontal - variable resolution, user-specified for each
application
Vertical - 50 to 200 meters per layer. The number of layers is
limited by computational requirements.
Temporal Resolution
One hour Interval for input and output.
Meteorology
IMPACT uses a mass-consistent flow field specifying depth of mixed
layer, wind speed and direction at the surface and for the vertical
layer over each horizontal grid cell for each hour of simulation.
Chemi stry
Two chemical kinetic mechanisms are provided as user options:
Mechanism # 1 - Five species including 1 hydrocarbon class
(limited to 12-hour daytime simulation)
Mechanism # 2 - Fourteen species including four hydrocarbon
classes: olefins, aromatics, paraffins, and aldehydes
Input Data Requirements
Meteorological Variables - wind speed and direction, topography,
inversion height field, longitude and latitude, insolation,
temperature, pressure, relative humidity, date.
41
-------
Emissions Data - a gridded, hourly emission inventory for all area
and point sources in the study zone. For the simpler chemical
kinetic mechanism, the inventory should include three species:
N02, NO, and HC (reactive hydrocarbons). For the #2 chemistry, the
following inputs should be provided: N02, NO, olefins, aromatics,
paraffins, aldehydes.
Air Quality Data - concentrations for all input species and ozone
along the study area boundary and for as many grid cells as are
available at the start of the simulation run should be provided.
For model verification, these air quality data should be available
for every hour of the simulation day.
Key Outputs
Two-dimensional array of hourly-average concentrations of ozone and
other user-specified species for each vertical level at each
prediction point.
Contour plot of species concentrations for each hour.
Approximate Computer Requirements
A 12-hour simulation of a 25 x 25 x 10 grid requires about 1
million bytes of memory and one hour of computation on an IBM
370/168.
42
-------
APPLICABILITY OF IMPACT TO OZONE CONTROL STRATEGY ISSUES
Hydrocarbon vs NOx Control
The model IMPACT can be used to address this issue because it can
simulate the HC/NOx relationship in ozone formation. The model has
been applied to evaluate the effects of concurrent hydrocarbon and
NOx emissions reductions.
Assessing Attainment of the Ozone Standard and a Short-Term N02 Standard
The model can be used to generate concurrent N02 and ozone levels.
Therefore, it can be used to address this topic.
Control of Low-Level vs Elevated Emissions
The model accepts both area and point sources. Area sources are
considered to be low-level emissions while point sources are
considered to be elevated. Several codes are available as options
in IMPACT to determine total plume rise from these sources. There
are no major differences among the options which are provided to
make the model more general. The point source emissions are added
to the vertical grid cell coinciding with the final plume height of
the source. Thus, IMPACT is well-formulated to address this issue.
Selective Control of Hydrocarbon Compounds
The simpler chemical kinetic mechanism involves only a single
hydrocarbon class and thus cannot be used to address this issue.
The more complex chemical kinetic mechanism utilizes four classes
of hydrocarbons (olefins, aromatics, paraffins and aldehydes) to
represent non-methane hydrocarbon emissions. The quantity of each
of these species can be varied to simulate the effects of selective
hydrocarbon control schemes on ozone formation.
Long-Range Transport
IMPACT could theoretically be used to address long-range transport
by modifying the wind generation module to verify its capability to
transport pollutants in an Inversion layer. However, as for all the
models in this study, further field verification would be needed
before IMPACT could be applied.
Effects of Spatial and Temporal Variation of Ozone Precursor Emissions
IMPACT is a deterministic grid-based model. Emissions must be
fixed within the study area and a temporal distribution allotted to
each. Thus, the model can be used to evalute this issue.
43
-------
MADCAP (MODEL OF ADVECTION, DIFFUSION AND CHEMISTRY
OF AIR POLLUTION)
MODEL FEATURES
Model Type
Eulerian grid
Historical Data
MADCAP was developed in 1978. The San Diego area is using the
model in their non-attainment planning process (Sklarew, 1979).
Spatial Resolution
Horizontal - variable, user-specified. The largest area studied is
a 37 x 51 grid of 2 km cells for the San Diego area.
Vertical - five layers divfded at the following elevations: 50,
75, 125, 250 and 500 meters.
Temporal Resolution
One hour resolution of input and output.
Meteorology
MADCAP uses a mass-consistent three dimensional flow field
specifying depth of mixed layer, wind speed and direction at the
surface and for the vertical layer over each horizontal grid cell
for each hour of simulation. This 1s the same wind field module as
In IMPACT.
Chemi stry
(Uses the LIRAQ-2 chemistry)
48-step reaction mechanism
Three hydrocarbon classes: alkanes, alkenes (olefins), aldehydes
» Input Data Requirements
Meteorological Variables - wind speed and direction; inversion
height field, insolation, temperature, pressure, relative humidity,
topography, site location and date.
Emissions Data - a gridded hourly emission inventory of all area
and point sources in the study region. Seven species must be
included: alkanes, alkenes, aldehydes, nitric oxide, carbon
monoxide, sulfur dioxide and particulates. (For hydrocarbon
45
-------
sources, MADCAP has a list of 17 source types arid their relative
hydrocarbon splits.) NOx is divided into NO and N02 according to
weighting factors for three source types.
Air Quality Data - concentrations for all input species and ozone
at the boundaries of the study area and for as many grid cells as
are available at the start of the simulation day are needed. For
model verification these values should be available for every hour
of the run.
Key Outputs
Tabular printouts of hourly ozone and other user-specified species
concentrations at selected sites. Isopleth maps for oxidant and
each desired species can also be generated at user option.
Hour-by-hour plots of pollutant concentrations for ozone and each
input species at user-specified locations.
Approximate Computer Requirement
For an 11-hour simulation using a 37 x 51 x 5 grid of 2 km
resolution in the horizontal, MADCAP requires about 1 hour of
computation time on a Univac 1100/80. Core requirement is about
185,000 words of a 4 byte-memory (740,000 bytes). This cost does
not include the emissions inventory and wind field generation
subroutines whose output is fed into MADCAP.
46
-------
APPLICABILITY OF MADCAP TO OZONE CONTROL STRATEGY ISSUES
Hydrocarbon vs NOx Control
MADCAP uses the LIRAQ-2 photochemical kinetic mechanism to simulate
ozone production. This mechanism is capable of accepting varying
HC and NOx emission levels and predicting the resultant ozone
concentration. Thus, MADCAP can address this issue.
Assessing Attainment of the Ozone Standard and a Short-Term N02 Standard
MADCAP can predict both ozone and N02 concentrations so that it can
be used to discern whether an optimal ozone reduction scheme is
counterproductive to attaining an N02 standard or vice-versa. Seme
further refinement might be needed before the model can be used for
such an assessment because in the San Diego study, some problems in
accuracy were encountered (Sklarew, 1979). It is conjectured that
much of the error is due to the NOx emission split into NO and N02.
Control of Low-Level vs Elevated Emissions
Five vertical layers are built into MADCAP. Emissions can be added
into any of the layers. The bottom layer is comprised of low-level
sources while point sources are added to that vertical layer
corresponding to their total effective plume height. If this
height exceeds the maximum study height of 500 m, the emissions are
excluded from subsequent calculations. Thus the model does
distinguish between sources of different height and can be used to
address this issue.
Selective Control of Hydrocarbon Compounds
Hydrocarbons are divided into three categories, based on their
reactivity: olefins, aliphatic aldehydes, and alkanes. Each of
these hydrocarbon classes can be varied to study the effect of
different emissions ratios on ozone formation. MADCAP can be
applied to this issue.
Long-Range Transport
Before using MADCAP for long-range transport, two things need to be
done: 1) the wind field module would need to be verified as to its
capability to simulate pollutant transport in the inversion layer;
and 2) the model needs to be verified with actual data (as do all
the models in this report.)
Effects of Spatial and Temporal Variation of Ozone Precusor Emissions
Being a deterministic model, MADCAP retains spatial and temporal
emission characteristics in its calculation. Thus, MADCAP can be
used to address variations in these parameters.
47
-------
RECOMMENDATIONS FOR FUTURE WORK
After undergoing several years of development and specialized
applications, ozone models are increasingly being used to support air
quality decision-making across the nation. A central role is envisioned
for these models in the development of State Implementation Plan
revisions for ozone. With this in mind, the immediate needs for further
work consist of what is necessary to convert existing models from
research tools to widely-disseminated, user-oriented, decision-making
tools. These needs may be divided into two categories: (1) further
performance evaluation and sensitivity analysis; and (2) improving user
services.
FURTHER PERFORMANCE EVALUATION AND SENSITIVITY ANALYSIS
A number of unresolved questions regarding the performance of
photochemical models can be identified. The most obvious is that for
many of the control strategy issues posed in this report, no "track
record" of model applications is available. The development of such
track records is crucial so that model users interested in investigating
one or more of these issues will have a base of experience by which data
requirements, model applicability, experimental design and model
performance may be judged.
The existing methodology by which model performance 1s evaluated
involves the comparison of model-simulated ozone distributions to
measured ozone distributions under a variety of prototypical
meteorological conditions. Stated differently, the models' performance
in response to varying meteorological conditions 1s compared with what
1s actually measured^ Conversely, in long-range planning applications,
such as will be required for the preparation of State Implementation
Plan revisions, it 1s the models' performance in response to varying
hydrocarbon and NOx emissions levels and distributions that is
important. Demonstration of "good" performance under varying
meterologlcal conditions does not necessarily imply good performance
under varying emissions conditions. Therefore, a different approach to
model verification should be devised which would be responsive to this
major area of model application.
A third recommendation related to performance evaluation is that a set
of performance standards or guidelines should be developed to aid model
* users in assessing the adequacy of a model's performance in a given
region before the model 1s used in a decision-making context.
Currently, there is no objective measure to assess the adequacy of a
model's performance in a given situation. Without a uniform set of
performance standards it becomes the responsibility of each user to
judge the acceptability of a model for any given application.
Finally, an important unknown at this time is the relationship between
model performance and the quantity and quality of the meteorological,
emissions, and air quality input data. Since resources for model
49
-------
applications are limited, potential model users will be concerned with
the cost of input data collection and preparation. At the present time
very little information is available on how the quantity and quality of
input data for different variables affects model performance.
IMPROVING USER SERVICES
Since the number of potential photochemical model users is expected to
increase substantially over the next few years, the demand for various
services to users will increase. Such services may include:
o Periodic update and maintenance of model codes to include the
most recent data (e.g., rate constants for chemical kinetics
mechanisms)
o Code optimization for large, complex models to reduce their
resource requirements
o Preparation of more detailed user-oriented handbooks on models
to reduce the level of expertise required to use them
o Provision of low - or no-cost services to users for designing
and troubleshooting model application projects.
50
-------
REFERENCES
INTRODUCTION
Argonne National Laboratory, "Report to U.S. EPA of the Specialist
Conference on the EPA Modeling Guidelines," Chicago, Illinois, February
22-24, 1977.
Dimitriades B., ed., "International Conference of Photochemical Oxidant
Pollution and Its Control," Proceedings, Vol. 1, EPA-600/3-77 001a,
Environmental Sciences Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, January,
1977.
Roth, P.M., Reynolds, S.D., Anderson, 6.E., Pollack, R.I., Yocke M.A.,
and Killus J.P., "An Evaluation of Methodologies for Assessing the
Impact of Oxidant Control Strategies," prepared for the American
Petroleum Institute by Systems Applications, Inc. (SAI), San Rafael,
California, SAI Report No. EF76-112, August, 1976.
U.S. Environmental Protection Agency, "Uses, Limitations and Technical
Basis of Procedures for Quantifying Relationships Between Photochemical
Oxidants and Precursors," EPA-450/2-77-021a, Office of Air Quality
Planning and Standards, U.S. EPA, Research Triangle Park, North
Carolina, November, 1977.
ROLLBACK
Chang, T.Y., Weinstock, B., "Generalized Rollback Modeling for Urban Air
Pollution Control," Journal of the Air Pollution Control Association.
1975, Vol. 25., pp. 1033-1037.
DeMandel, R.E., Robinson, L.M., Fong, J.S.L., and Wada, R. "Comparisons
of EPA Rollback, Empirical/Kinetic and Physico-chemical Oxidant
Prediction Relationships in the San Francisco Bay Area" Journal of the
Air Pollution Control Association, April, 1979, Vol. 29. no.4. pp
352-357.
deNevers, N., and Morris, J.R., "Rollback Modeling: Basic and
Modified," Journal of the Air Pollution Control Association. 1975. Vol.
25, pp. 943^*7^ — 1
Roth P.M., Reynolds, S.D., Anderson G.E., Pollack R.I., Yocke M.A., and
Killus, J.P., "An Evaluation of Methodologies for Assessing the Impact
of Oxidant Control Strategies," prepared for the American Petroleum
Institute by Systems Applications, Inc. (SAI), San Rafael, California,
SAI Report No. EF76-112, August, 1976.
51
-------
IMPACT
Amar, P.K., and Ranzieri, A.J., "Validation of a Dispersion Model for
the Transport of Atmospheric Tracer in a Complex Terrain" presented at
the 71st Annual Meeting of the Air Pollution Control Association,
Houston, Texas, June 25-30, 1978.
Fabrick, A., Sklarew, R.C., and Wilson, J.G., "Point Source Model
Evaluation and Development Study," prepared for the California Air
Resources Board and the California Energy Resources Conservation and
Development Commission by Science Applications, Inc. (SAI), Westlake
Village, California, SAI Contract No. A5-058087, March, 1977.
Fabrick, A., Sklarew, R.C., Wilson, J.G. , and Taft, J., "Point Source
Model Evaluation and Development Study, Appendix C - User Guide to
IMPACT," prepared for California Air Resources Board and the California
Energy Resources Conservation and Development Commission by Science
Applications, Inc. (SAI), Westlake Village California, SAI Contract No.
A5-058087, March, 1977.
Ranzieri, A., Manager of the Modeling Section, California Air Resources
Board, Sacramento, California, personal communication, January, 1979.
Sklarew, R.C. and Wilson J.G., "Applications of DEPICT to the Gansfield,
Navajo, and Ormond Beach Air Quality Data Bases," Science Applications,
Inc., Westlake Village, California, July, 1976.
EKMA ant* OZIPP
DeMandel, R.E., Robinson L.H., Fong, J.S.L, and Wada, R. "Comparison of
EPA Rollback, Empirical Kinetic and Physicochemical Oxidant Prediction
Relationships in the San Francisco Bay Area," Journal of the Air
Pollution Control Association, April, 1979, Vol.29, no.4, pp 352-357.
deNevers, N. and Morris, J.R. "Rollback Modeling: Basic and Modified,"
Journal of the Air Pollution Control Association, 1975. Vol. 25. dp
§43-947.
U.S. Environmental Protection Agency, "Uses, Limitations and Technical
Basis of Procedures for Quantifying Relationships Between Photochemical
Oxidants and Precursors," EPA-450/2-77-021a, Office of Air Quality
Planning and Standards, U.S. EPA, Research Triangle Park, North
Carolina, November, 1977.
U.S. Environmental Protection Agency "User's Manual for Kinetics Model
and Ozone Isopleth Plotting Package," EPA-600/8-78-014a, Environmental
Sciences Research Laboratory U.S. E.P.A., Research Triangle Park, North
Carolina, July, 1978.
52
-------
HANNA ATDL
Friedlander S.K., and Seinfeld O.H., "A Dynamic Model of Photochemical
Smog," Environmental Science and Technology, November, 1969, Vol. 3, pp.
1175-llST:
Hanna S., "Application of a Simple Model of Photochemical Smog,"
presented at the 3rd Clean Air Congress of the International Union of
Air Pollution Prevention Associations, Dusseldorf, Germany, October
8-12, 1973.
Hanna S., "A Simple Dispersion Model for the Analysis of Chemically
Reactive Pollutants." Atmospheric Environment, July, 1973, Vol. 7, pp.
803-817.
ELSTAR
Allen P.D., et.al., "Transportation Systems and Regional Air Quality - A
DIFKIN Sensitivity Analysis," prepared by Office of Transportation
Laboratory, California Department of Transportation, Sacramento
California, Report No. CA-D0T-TL-7169-2-76-27, April, 1976.
F. Lurmann, et. al., A Lagrangian Photochemical A1r Quality Simulation
Model - Adaptation to the St. Louis - RAPS Data Base. EPA-w)0/8-79-01J> a
and b, June 1979.
Lloyd, A.C., Deputy Manager, Environmental Analysis Division,
Environmental Research and Technology, Santa Barbara, California,
personal communlcation, December, 1978.
Martinez, J.R., "User's Guide to D1ffus1on/K1net1cs (DIFKIN) Code,"
prepared for the U.S. Environmental Protection Agency, Office of A1r
Quality Planning and Standards, Research Triangle Park, North Carolina
by Environmental Research and Technology, Santa Barbara, California,
EPA-R4-012b, October, 1972.
TRACE
Drivas P.J., "TRACE (Trajectory Atmospheric Chemistry and Emissions)
User's Guide", Pacific Environmental Services Inc. (PES), Santa Monica,
California, PES Report No. TPT016, November, 1977.
Drivas, P.J. "Photochemical Modeling of Proposed Freeway Alternatives
• in Phoenix," prepared for Aerovlronment Inc. under contract to the City
of Phoenix, Arizona by Pacific Environmental Services, Inc., Santa
Monica, California, February, 1977.
Drivas, P.J. and Wayne, L.G. "Validation of an Improved Photochemical
Air Quality Simulation Model." Pacific Environmental Services, Inc.
(PES), Santa Monica, California, PES Report No. TP-014, March, 1977.
53
-------
.Wayne L.6., Kokin, A. and Weisburd, M. 1., "Controlled Evaluation of the
Reactive Environmental Simulation Model (REM) " Volume V: Final Report,
prepared for U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Research Triangle Park, North Carolina, by
Pacific Environmental Services Inc., Santa Monica, California,
EPA-R4-73-013A, 1973.
LIRAQ-2
Association of Bay Area Governments, Bay Area Air Pollution Control
District, and Metropolitan Transportation Commission, "Draft Air Quality
Maintenance Plan," Environmental Management Plan for the San Francisco
Bay Region, December, 1977.
Bass, A.S., et. al., "The Livermore Regional Air Quality Model (LIRAQ):
A Technical Review and Market Analysis," prepared for the National
Science Foundation, Washington D.C.., by Environmental Research and
Technology (ERT), Report No. P-2348-1, February, 1977.
Duewer, W.H., et. al;" The Livermore Regional Air Quality Model: II.
Verification and Sample Application in the San Francisco Bay Area,"
Journal of Applied Meteorology, March, 1978, Vol. 17, pp. 273-311-
MacCracken, M.C. and Sauter, G.D., eds., "Development of an Air
Pollution Model for the San Francisco Bay Area," prepared for U.S.
Energy Research and Development Administration by Lawrence Livermore
Laboratory, University of California, Livermore, California, Report No.
UCRL51983, December, 1975.
MacCrac^n, M.C., et. al., "The Livermore Regional Air Quality Model:
I. Concept and Development," Journal of Applied Meteorology, March.
1978, Vol. 17, pp. 254-272.
airshed"
Anderson, G.E., Hayes, S.R., Hillyer, M.J., Killus, J.P., and Mundkur
P.V., "Air Quality in the Denver Metropolitan Region 1974-2000,"
prepared for the U.S. Environmental Protection Agency, Region VIII,
Denver, Colorado, by Systems Applications Inc. (SAI), San Rafael
California, EPA-908/1-77-002, May, 1977.
Reynolds, S.D., Liu, M.K., Hecht, T.A., Roth, P.M., and Seinfeid, J.H.,
"Further Development and Evaluation of a Simulation Model for Estimating
Ground-Level Concentrations of Photochemical Pollutants," Final Report
for the U.S. Environmental Protection Agency, Office of Research and
*¦ Development, Research Triangle Park, North Carolina, by Systems
Application Inc. (SAI), San Rafael, California, SAI Report No. R73-19,
February, 1973.
Reynolds, S.D., Reid, L.E. and Tesche, T.W., "An Introduction to the SAI
AIRSHED Model and Its Usage." Systems Applications, Inc. (SAI), San
Rafael, California, SAI Report No. EF78-53R, December, 1978.
54
-------
Reynolds, S.D. et.al., "Photochemical Modeling of Transportation Control
Strategies, Volume I. Model Development, Performance Evaluation and
Strategy Assessement," Systems Applications, Inc. (SAI), San Rafael,
California, SAI Report No. EF79-37, March, 1979.
MADCAP
Sklarew, R.C., "Verification of the MADCAP Model of Photochemical Air
Pollution in the San Diego Air Basin," presented at the Annual American
Meteorological Society Meeting, Reno, Nevada, January, 1979.
Sklarew, R.C., Joncich, M.A., and Tran, K.T., "An Operational Air
Quality Modeling System for Photochemical Smog in the San Diego Air
Basin," presented at the Annual American Meteorological Society Meeting,
Reno, Nevada, January, 1979.
55
-------
TECHNICAL REPORT DATA
1Please read Jnwvctions on the reverse before completing)
1. REPORT NO. 2.
ERA-450/4-79-026
3. RECIPIENT'S ACCESSl01s* NO.
4. TITSEAND SUBTITLE
Application of Photochemical Models
Volume II: Applicability of Selected Models for
Addressing Ozone Control Strategy Issues
5. REPORT DATE
December 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Wada, Ronald Y., et al.
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Association of Bay Area Governments
Hotel Claremont
Berkeley, California 94705
10. PROGRAM ELEMENT NO.
2AA635
11. CONTRACT/GRANT NO.
68-02-3046
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This document is an assessment of the applicability of a number of selected photo-
chemical models to a variety of ozone control strategy issues. The issues include:
hydrocarbon vs. NO^ control; applicability to NO2 strategy assessment; low-level
vs. elevated emissions; selective control of hydrocarbon compounds; long-range
transport; and spatial and temporal variation of ozone precursor emissions. The
specific models reviewed are as follows: linear rollback; EKMA; OZIPP; Hanna ATDL;
ELSTAR; TRACE; LIRAQ; SAI Urban Airshed Model; IMPACT; and MADCAP. Data require-
ments, costs, and output produced by each model are also compared.
17. ^ KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b. 1DENTIF1ERS/OPEN ENDED TERMS
c. COSATI Field/Group
Photochemical Modeling
SIP Development
18. DISTRIBUTION STATEMENT
RELEASE UNLIMITED
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (Thispage)
. Unclassified
22. PRICE
EPA Fo-m 2220-1 (9-73)
------- |