EPA-450/4-81-005C
APPLICATION OF THE EMPIRICAL
KINETIC MODELING APPROACH
TO URBAN AREAS
Volume III: Philadelphia
By
Systems Applications, Inc.
101 Lucas Valley Road
San Rafael, California 94903
EPA Contract No: 68-02-3376
EPA Project Officer: Gerald L Gipson
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office Of Air And Radiation
Office Of Air Quality Planning And Standards
Research Triangle Park, North Carolina 27711
October 1985
-------
This report has been reviewed by the Office Of Air Quality Planning And Standards, U.S. Environmental
Protection Agency, and approved for publication as received from the contractor. Approval does not signify
that the contents necessarily reflect the views and policies of the Agency, neither does mention of trade
names or commercial products constitute endorsement or recommendation for use.
EPA-450/4-81-005C
-------
CONTENTS
Di sclaimer
Abst ract
1. Introduction ,
Background and objectives !!!!!!!!!!!!!!!!!!! 1
The Empirical Kinetic Modeling Approach (EKMA)....!""! 2
An overview of EKMA comparison evaluations !.' 3
2. Model Appl ications !!!!!!!" 5
Description of the models used in this study..!!!!!""] 5
Description of the Philadelphia modeling region....!!!!! n
Characterization of the Philadelphia modeling days....!! 11
Inputs used in the Urban Airshed Model
simulations ?3
Inputs used in the Systems Applications
Trajectory Model simulations 30
Inputs used in the Level II OZIPM/CBM
model simulations >s> 4g
Inputs used in the Level III EKMA model simulations!!!!! 49
3. Comparison of Model Results 55
4. Summary, Conclusions, and Recommendations !!!!!!!!!!!!! 64
Summary , fi.
Conclusions ^ 64
Recommendations !!!!! 66
References !!!!!!
5. Append ix !!!!!!!!! 71
85068r3
-------
-------
FIGURES
1 Philadelphia airshed modeling region ............................ 12
2 Geographical location of the Philadelphia airshed
modeling region .................................. ,~
3 Synoptic situation, 0700 EST, July 13, 1979 ..................... 21
4 Physical boundaries used in the 13 July 1979 airshed
model simulation ............... or
Boundary specifications for 19 July 1979 simulation ............. 31
6 Mixing height profile for urban and rural cells for the
13 July 1979 simulation ...........
7 Mixing height profiles for urban and rural cells on
19 July 1979
85068r2
34
36
8 Airshed model surface winds for 13 July 1979 37
9 Airshed model surface winds for 19 July 1979 41
10 Trajectory path for 13 July 1979 47
11 Trajectory path for 19 July 1979 48
IV
-------
-------
TABLES
1-A Maximum predictions/observations for 13 July 1979
for ozone (pphm) ............................................... 15
1-B Performance measures for the UAM simulation of the
13 July 1979 episode in Philadelphia..... ...................... 16
2-A Maximum predictions/observations for 19 July 1979
for ozone (pphm)
2-B Performance measures for the UAM simulation of the
19 July 1979 episode in Philadelphia ........................... 19
3 Initial conditions for 13 July 1979 ............................ 24
4 Initial conditions for 19 July 1979 ............................ 25
5 Background concentration values for 13 July at the top
of the modeling region— as initial concentrations above
the mixing height, and for all levels of all boundaries
except the levels below the mixing height on the southeast
bounda ry ............................................ 27
6 Southeast boundary conditions for cells below the mixing
height for the simulation of 13 July 1979 ...................... 28
7 Boundary conditions used for the northeast and east
boundaries below the mixing height estimated from data
collected at the Van Hiseville, New Jersey monitor ............. 32
8 Urban and rural mixing height values used in the
DIFFBREAK file for 13 July 1979 ................................ 33
9 Urban and rural mixing height values used in the
DIFFBREAK file for 19 July 1979 ................................ 35
85068T2
-------
-------
10 Background concentration values for 19 July used at the
top of the modeling region (TOPCONC), initial conditions,
and for all boundaries except the levels below the mixing
height on the northeast and east boundaries 44
11 Total daily emissions by source type (mole) in the 1979
Philadelphia inventory 45
12 Initial conditions and emissions rates used in the
Level II CBM/OZIPM calculations for 13 July 1979 50
13 Initial conditions and emissions rates used in the
Level II CBM/OZIPM calculations for 19 July 1979 51
14 Volatile organic compound (VOC) and nitrogen oxides
(NOX) emissions by county for 1980 53
15 Initial conditions used in the Level III EKMA
calculations for 13 July 1979 53
16 Initial conditions used in the Level III EKMA
calculations for 19 July 1979 54
17 Model ozone results (pphm) for Philadelphia on
13 July at Roxy Water Pump monitor 56
18 Model ozone results (pphm) for Philadelphia on
19 July at the Downington monitor 57
19 Percent reduction in predicted ozone between scenarios
at Roxy Water Pump monitor for 13 July 5q
20 Percent reduction in predicted ozone between scenarios
at Downington monitor for 19 July 60
21 Hydrocarbon emission reductions required to meet the
NAAQS for ozone from the sensitivity simulations of
13 and 19 July
85068T3 1 V1-
-------
-------
1 INTRODUCTION
BACKGROUND AND OBJECTIVES
For several years the Office of Research and Development of the U.S.
Environmental Protection Agency (EPA) has sponsored a coordinated program
consisting of fundamental chemical research, smog chamber experimentation,
and model development. The objective of this program is to develop mathe-
matical models capable of simulating the dynamics of the chemical reac-
tions, and the dispersion of the gaseous pollutants, found in the lower
troposphere. These atmospheric models will be used both to gain an under-
standing of air pollution and to predict the probable outcome of air pol-
lution control scenarios. To this end, the EPA Office of Air Quality
Planning and Standards (OAQPS) has embarked on an extensive program of
model comparison and evaluation.
The specific purpose of the OAQPS program is to assess the suita-
bility of available photochemical oxidant modeling approaches for use by
states as planning tools in ozone air quality planning and management.
The program includes comparisons and evaluations of various approaches for
typical United States cities. As part of the program, large, detailed
data bases are being assembled for several example cities.
The primary objective of this report is to evaluate the Empirical
Kinetics Modeling Approach (EKMA) through application of the photochemical
box model (OZIPM) that forms its basis. Because some models are grid-
based and thus designed to reproduce spatial and temporal variations with-
in the metropolitan airshed, detailed information files are required. The
EPA has prepared highly detailed data concerning precursor emissions,
meteorology, and ambient air quality for the cities of Tulsa and Philadel-
phia to test various air quality simulation models such as OZIPM and the
Urban Airshed Model (UAM). In addition, data bases are already available
for St. Louis, Los Angeles, San Francisco, and Sacramento. The availa-
bility of detailed data and the results of parallel studies using
sophisticated models offer the following advantages for an in-depth study
of EKMA:
85068P3 2
-------
(1) The simple application of the OZIPM box model, using either the
actual available data inputs or the corresponding data simulated
by other models, should generate results that can be compared
both with ambient air data and with the results of the other
models. The timing of results regarding the ozone maxima and
precursor decay should also be comparable.
(2) Tests of the sensitivity of the results to variations in the
input data can demonstrate important data requirements for suc-
cessful application of OZIPM. This information should be useful
in determining its cost effectiveness for future uses. The
results of the sensitivity study will also be useful for improv-
ing OZIPM in case the present version does not adequately simu-
late observed ambient data or the simulations of more complex
models.
(3) The application of the EKMA procedures to air quality planning
demonstrates its utility compared with that of methods based
either on simpler models or on more complex models (such as the
UAM).
(4) The potential effectiveness of OZIPM model improvements can be
assessed using the results of the complex models as a standard.
THE EMPIRICAL KINETIC MODELING APPROACH
The EPA has developed a method for estimating the emission controls
needed to meet the National Ambient Air Quality Standard (NAAQS) for ozone
(03) concentrations in urban areas. This method, known as the Empirical
Kinetic Modeling Approach (EKMA), uses an isopleth diagram of ozone con-
centrations that is related to hydrocarbon (HC) and nitrogen oxide (NO )
precursor levels. Each point on the isopleth diagram represents the maxi-
mum one-hour ozone level reached as a result of the specific HC and NO
combination described by the abscissa and ordinate values of the isopleth
diagram. The ozone value generated from each HC and NOX combination is
obtained by means of a computer-based trajectory model—the Ozone Isopleth
Plotting Package (OZIPP) (Whitten and Hogo, 1978) and OZIPM for alternate
chemical mechanisms (EPA, 1984). The OZIPM trajectory model contains many
simplifying assumptions and options that are designed to strike a balance
between factors such as the state of knowledge, data availability, com-
puter size, predictive accuracy, and overall cost.
85068P2
-------
AN OVERVIEW OF EKMA COMPARISON EVALUATIONS
Four studies have been performed by Systems Applications to evaluate
the EKMA through comparison of the results of OZIPM trajectory model with
those of other models. Three of these four studies were sponsored by the
EPA; the Philadelphia study is the third of the three. The first study
compared the UAM and the Systems Applications Trajectory models with vari-
ous versions of the box model used in EKMA for the Los Angeles area (Whit-
ten and Hogo, 1981). In the second study, the LIRAQ model formed the
basis for the San Francisco area comparison, and the Systems Applications
models were used for the Sacramento area (Whitten et al., 1981). The UAM
was also used in the study of the Tulsa area and for this study of the
Philadelphia area.
A comparison evaluation is performed at several levels to provide a
spectrum of information. The independent application of the EKMA-OZIPM in
areas where other models have been applied merely validates the common
assumption that different models can sometimes produce different
results. In addition, this study attempts to provide information explain-
ing why specific features of the different models lead to different
results. For instance, a specific feature of grid models (in contrast to
trajectory models in general) is their ability to simulate some wind shear
effects. Grid and trajectory models have been found to produce different
results when wind shear effects are significant; the reverse is also true
(similar results can occur when wind shear is minimal). Thus, specific
model differences can be associated with differences in results. Given
the previous example, a prospective user should apply the trajectory model
with a measure of discretion in situations known to have a high wind
shear.
Different levels of evaluation are possible within the EKMA frame-
work. The control scenario estimations generated by EKMA isopleth dia-
grams are the end result of a series of steps involving model assumptions
and adjustments. The basis of the EKMA is a trajectory model, which is
quite simple from the standpoint of dispersion. This trajectory model can
be applied in the absolute sense as an atmospheric model on its own. The
OZIPM computer code that contains this model also has an array of options,
each of which will be geared to some specific standardized setting if the*
user does not exercise this option.
The isopleth diagram is used in a relative manner, so that if all the
options and inputs fail to generate a simulated ozone value that agrees
with observations, the final discrepancy is eliminated by using the obser-
ved value. Some discrepancy is almost always expected because, except for
the chemistry employed, the EKMA-OZIPM is a rather simplistic approach.
85068T3 2
-------
The use of the observed ozone value rather than the simulated value is
intended to account for some of the simplifications inherent in the EKMA-
OZIPM.
The levels of evaluation in this study begin with the specific fea-
tures that define each basic model; other levels involve input data and
options. At each level, the present study attempts to make comparisons
with other models, observations, and alternate methodologies. However,
the evaluation of atmospheric models does not necessarily fit a specific,
well-organized framework. Problems that are unique to some areas are not
always apparent ahead of time. Thus, the study of the Philadelphia area
reported here is only part of an overall evaluation of the EKMA. The
following sections present the specific results and analysis of the
Philadelphia study as they relate to the EKMA evaluation.
85068p3
-------
SECTION 2
MODEL APPLICATIONS
This section describes the photochemical air quality models used in
this study and the inputs associated with each model. Descriptions of the
Philadelphia study region and the modeling days used in that study are
also presented. As noted, the purpose of this study is to compare the
ozone predictions of air quality models of different complexity (in
particular, to compare EKMA calculations with current UAM applications)
These models can be classified on the basis of their complexity according
to the levels of analysis described in the 1979 Federal Register (65669-
65670) on ozone modeling:
Level I: Sophisticated dispersion models such as the Urban Airshed
(UAM) and Systems Applications' Trajectory models.
Level II: Box models such as OZIPM/CBM.
Level III: The city-specific EKMA.
Level IV: The standard EKMA.
The UAM is a three-dimensional grid model that mathematically simu-
lates the physical and chemical processes responsible for photochemical
smog. One of the advantages offered by a three-dimensional grid model is
the ability to consider spatial and temporal effects of control strate-
gies, whereas models of less complexity consider temporal effects only
The structure of the UAM consists of an array of cells, the total volume
of which represents an urban area. The horizontal dimensions of each cell
are constant, but the vertical thickness of the cells vary according to
the depth of the mixed layer and top of the boundary region throughout the
simulation. The UAM can simulate the advection of pollutant species
through the modeling region, the diffusion of material from cell to cell,
the injection of primary source emissions into the modeled region, and the
chemical transformations of reactive species into intermediate and
secondary products.
85068r3 3
-------
The UAM was applied to Philadelphia for two days in 1979—13 July and
19 July—to compare model predictions with station observations. The
model was then used to evaluate ozone control requirements for the Phila-
delphia area. The Inputs to the UAM served as a data base for the series
of less complex models. Descriptions of each model are presented in the
following sections followed by a summary of UAM inputs (described in
detail in the Appendix). Inputs to the less complex models are also pre-
sented.
DESCRIPTION OF THE MODELS USED IN THIS STUDY
The following models were used 1n this study:
The Urban Airshed Model (UAM)—a three-dimensional grid model that
uses CBM-II chemistry.
The Systems Applications Trajectory Model—In terms of inputs, com-
putations, and results at points along a selected trajectory, this
model differs from the UAM only in the elimination of horizontal
dispersion and vertical winds.
Level II box model (OZIPM/CBM)—uses the same meteorological inputs
as the Systems Applications Trajectory Model, but also uses both the
CBM-II and the CBM-III chemistry.
The level III city-specific OZIPM model--uses Carbon-Bond chemistry
as described by EPA (1984) and the standard EKMA chemistry as
described by EPA (1981).
In this section, the principal differences among these models are
highlighted.
Comparison of the Urban Airshed and Systems Application Trajectory Models
These two models were designed to complement each other; therefore,
many algorithms are identical for both models (several subroutines of the
computer codes are actually shared). Furthermore, the models use the same
Input files. The trajectory model begins on a square of the grid model at
a specific time. The path of the trajectory is then determined solely by
the wind speed and direction in the lowest level, or some average of the
Detailed discussions of this comparison are available elsewhere (see
Reynolds et al., 1979).
85068T3 3
-------
lowest two levels, of the grid-model input files. Because the other
levels in the grid model often advect with velocities that are different
from the lowest level, the two models often generate different results as
the simulations progress in time. Nevertheless, the trajectory model
provides an inexpensive means of evaluating parts of the grid model at
specific locations and times.
A special feature of the trajectory model is its ability to be opera-
ted backward in time. This feature was used to generate the trajectories
used in this study. The time and location of either the observed or the
simulated ozone maximum were used to start the backward trajectory to 0500
EOT. Thus, the forward trajectories always arrived at the desired spot at
the proper time for comparison with either observations or grid-model
results. Changes in emissions often lead to changes in the time and loca-
tion of maximum ozone concentrations. Small changes in times of maximum
ozone concentration at some specific monitoring site can be associated
with rather large differences in the location of the origin of, and in the
actual pathway taken by, the air parcel producing the specific ozone maxi-
mum. Therefore, when compared with a grid model, any trajectory model
used as a tool for control strategies possesses the inherent shortcoming
of being incapable of accounting for these changes.
Comparison of the Systems Applications Trajectory Model
and the OZIPM/CBM ModeT"
The following similarities and differences exist between the Systems
Applications' Trajectory model and the level II OZIPM/CBM:
Chemistry. Both of these models employ CBM-II chemistry, which is
also used in the UAM. However, the photolytic constants are computed
differently in the OZIPM/CBM than in the UAM and Trajectory models.
In order to compare different models, the same data base is needed
and the chemistry used in the models must be as similar as possible
so that any differences in the results can be attributed to indi-
vidual model characteristics and not the differences in the chemis-
try. The N02 photolysis rates normally used in the OZIPM/CBM are
somewhat lower than those used in the UAM and Trajectory models,
which are based on actual measurements. Therefore, for this study
the OZIPM/CBM values were modified for N02 photolysis using a con-
stant factor so that all models agreed at 1300 EOT.
In the UAM and Trajectory models, the rates of all other photolytic
reactions vary with N0? photolysis, with a fixed ratio for each,
whereas in the OZIPM/CBM the photolysis rates vary independently.
85068T3 3
-------
Aldehyde photolysis was adjusted in the OZIPM/CBM to provide agree-
ment between the models at 1100 EOT. Actually, these differences in
photolysis rates with zenith angle constitute an advantage in
sophistication for the OZIPM/CBM.
The use of fixed ratios to N02 photolysis in the Systems Applications
models, in place of the variable ratios used in the OZIPM/CBM model
can lead to minor differences in chemistry. The ratios are at a
maximum at solar noon and a minimum at sunrise and sunset (Whitten et
al., 1980). The photolysis rates for aldehydes change more with
solar zenith angle than the N02 photolysis rate does because alde-
hydes photo!yze at the short wavelength end of the solar spectrum,
whereas N02 photolyzes nearer the long wavelength end of the ultra-
violet spectrum. The short ultraviolet wavelengths are more affected
by the ozone in the stratosphere than are the longer wavelengths. At
high zenith angles (morning and evening), the solar rays must pass
through a thicker layer of ozone. Therefore, the ratio of aldehyde
photolysis to N02 photolysis varies throughout the day, reaching a
maximum at solar noon. For this particular study, the OZIPM/CBM
aldehyde photolysis values were multiplied by a constant value
throughout each day so that the models agree at 1100 EOT. Therefore,
in the early morning the photolysis rate of the aldehydes in the
OZIPM/CBM would be somewhat less than that used in the Systems Appli-
cations models.
For this particular study, we chose to modify the OZIPM/CBM to make
it as much like the DAM and Trajectory models as possible. By remov-
ing all possible potential variables between models, we hoped to be
able to better compare the fundamentals underlying these models. The
CBM version of OZIPM has normally been applied in order to assist
users of the Airshed Model to estimate potential control scenarios.
For such applications the photolysis adjustments discussed here are
normal.
Reactivity. The Systems Applications Trajectory and Airshed models
are operated with reactivity splits that vary with each source cate-
gory. However, the OZIPM/CBM does not have this feature; instead,
reactivity is assumed to be constant throughout the day. The OZIPM
computer code can accept four different hydrocarbon reactivities: one
for background; one for the initial conditions; one for the aged
hydrocarbons aloft; and one for the emissions that occur throughout
the day. Hence, the only differences in reactivity between models
were in the temporal variations aloft and in the variable emissions
85068T3 3
-------
used In the Systems Applications models compared with the fixed reac-
tivities employed in the emissions for the OZIPM models. The reac-
tivities used in the OZIPM/CBM were determined from averages of emis-
sions used in the Trajectory model; for averaging the hydrocarbons
aloft, a weighting factor was used each hour that was equal to the
positive change in mixing height.
In a sense, the N0n/N0x ratio also acts as a reactivity effect. The
OZIPM/CBM uses a fixed ratio of N02/NOX (10 percent) for all NOX
emissions, whereas the Systems Applications models allow this ratio
to vary according to their source category (the average ratio is
typically about 5 percent). At this time, it has not been attempted
to change the OZIPM/CBM or to fix the ratio in the Systems Applica-
tions models, because it is felt that the difference is trivial. For
NOX aloft, the OZIPM/CBM uses pure N02, which is close to typical
values found in the Trajectory Model simulations because the N02/NOX
ratio is almost always greater than 0.85; the values aloft are often
greater than 0.99.
Integration Scheme. The OZIPM/CBM uses a Gear-type method without
any steady-state approximations, whereas the Trajectory Model uses
the same Crank-Nicholson finite-differencing scheme and steady-state
approximations that are used in the Airshed Model. However, our
previous tests using smog chamber experiments suggest that the dif-
ferences introduced by the numerical integration schemes should be
minimal if the steady-state species are carefully chosen.
Vertical Layers. The UAM and Trajectory models share the same
vertical layers and algorithms for numerical integration, but addi-
tional vertical contributions are introduced into the UAM for con-
vergent or divergent winds. The OZIPM/CBM is limited to providing
calculations of the mixed layer as a whole, but fixed concentrations
in the layer aloft can be introduced as the mixing height rises. The
Systems Applications models account for mixing within the mixed
layer, mixing within the inversion layer, and eddy diffusion among
all layers. The OZIPM/CBM assumes instantaneous mixing within the
entire mixed layer, and no eddy diffusion is considered across the
boundary defined by the mixing height.
In our comparison studies, the effectiveness of multilayers and eddy
diffusion was one of the central issues to be evaluated in this
region of our model series. A problem to avoid, which does confound
the comparison, is the entrainment of pollutants from aloft. The
OZIPM/CBM requires fixed concentrations aloft, so these must be
chosen from some "proper" average of the time-dependent concentra-
tions computed in the Systems Applications Trajectory Model. With
85068T2 3
-------
the average value used in this study, the concentrations entrained
when the mixing height rises most significantly were weighted most
heavily. The concentration entrained was weighted using the ratio of
the change in the mixing height per hour to the total change in the
mixing height per day. A more complex weighting scheme might involve
the time when the greatest effect on the chemistry occurs; however,
this effect was not evaluated. Concentrations aloft are not con-
sidered for the time before the inversion rises or after it has
reached some maximum. As the mixing lid drops late in the afternoon,
the only process affecting the surface layer from concentrations
aloft is eddy diffusion. Eddy diffusion is treated in the UAM and
Trajectory models, but is neglected in the OZIPM/CBM.
Mixing height. All algorithms used in the OZIPM/CBM, as well as
those used in the UAM and Trajectory models, can use identical mixing
heights with values that vary linearly in time between specified
values at each hour.
Temperature. The temperature varies in the UAM and Trajectory models
vertically, but the OZIPM/CBM currently employs a fixed temperature
for each hour. This difference can be significant. Temperature
affects both vertical diffusion and chemistry in the Systems
Applications models, but it can affect only chemistry in the
OZIPM/CBM because vertical mixing is assumed to occur instantaneously
within the mixed layer.
Emissions. Both the UAM and the Systems Applications Trajectory
models allow for emissions into all layers. The OZIPM/CBM, however,
treats all emissions in the mixing layer alike and does not directly
address emissions above the mixing layer. In the OZIPM/CBM, emis-
sions in the layer aloft must be accounted for in the determination
of the fixed concentrations aloft. Normally, the OZIPM computer code
treats emissions as a multiple of the initial conditions. However,
since the sum of the emissions in the lower layers of the Systems
Applications' Trajectory model was used for the emissions in the
mixing layer in the OZIPM/CBM, the OZIPM code was modified to treat
the emissions in an absolute manner. This modification was imple-
mented by using the transported surface layer inputs as the initial
concentrations and was done primarily for ease of comparison. The
normal initial concentrations (the CALC mode in OZIPM) were specified
as 1.0 for both HC and NOX, but these values were not added as
initial concentrations. They were used only for emission fraction
development because emissions are expressed relative to these values
in the OZIPM code.
Emissions for the Systems Applications Trajectory models are given as
total moles emitted into each cell during each hour. To arrive at
85068T3 3
-------
values appropriate for the OZIPM/CBM, we used the following conver-
sion procedure: the mixing volumes of the Trajectory Model and the
OZIPM/CBM were assumed to be equal below the mixing height, such that
where A is the area of the horizontal cross-section of the air parcel
defined in Trajectory Model simulations and Zmix is the mixing
height. If the volume is expressed in cubic centimeters, the moles
emitted in one hour (the number available from the Trajectory Model
computer output) can be divided by this volume to give moles /cnr
units. This, in turn, can readily be converted to the OZIPM/CBM
units of concentration (ppm) by using a conversion factor of 2.445 x
10 ppm moles'1 cm , which is technically correct for a perfect gas
at 298 K and 1 atmosphere of pressure. The numbers calculated at
this point are almost directly applicable to the OZIPM/CBM model.
First, the use of 1.0 for the initial concentrations of HC and NOX
in the CALC mode implies that each one-hour emission number would be
the concentration added during the hour if the mixing height were the
same as the initial value ZQ. However, the computer codes are writ-
ten so that, at each instant in time, the emission factor is multi-
plied by Z0/Zmix. Therefore, the Zmix drops out and the actual num-
bers (En) used are
En = A x 2.445 x 1010/Z0
where A and ZQ are expressed in centimeters.
DESCRIPTION OF THE PHILADELPHIA MODELING REGION
The Philadelphia metropolitan area is not an isolated "airshed" like
the Tulsa region; rather, it is nested in the expansive Northeast Urban
Corridor. The local ambient ozone levels recorded on a particular day in
Philadelphia may be partly the result of the influx of ozone precursors
from the New York metropolitan area to the northeast or from the Balti-
more/Washington D.C. urban area to the southwest. The Philadelphia
area is shown in Figures 1 and 2. Much of the detailed UAM modeling
descriptions and results are presented by Braverman and Haney (1985).
This section and the appendix summarize the UAM study for the Philadelphia
area.
CHARACTERIZATION OF THE PHILADELPHIA MODELING DAYS
From the Philadelphia oxidant data study performed during the summer
of 1979, two days were chosen for the validation of the UAM. This section
85068T3 3
-------
NORTH
0 10 20 30 40 50
KILOMETERS
CHESTER
MiImington.
0
DOWNTOWN PHILflDELPHlfl RREfl
5 10
KILOMETERS
RMS Lab.
Rncore
Bristol
Br ipant me
5 Cercden
6 Chester
7 Cleymont
6 Conshohocken
9 Defense Suppcrt
10 Downington
11 Franklin Injittuie
12 Island Rd. fiirport Ci-c'e
13 Lunberton
H Nornstown Rrrrory
15 Northeast Rirprrt
16 Roxy Hater Purap
17 5E Sewage Plant
IB South Broad
19 Sunn it Bridge
20 SH Corner BrrecYBut ler
21 Trenton
22 Van Htsevi I le
23 Vine I end
FIGURE 1. Philadelphia airshed modeling region.
(Bold outline is emission region; lighter lines are
county boundaries.)
12
-------
fHILBDELfHIfi
R1R5HED
FIGURE 2. Geographical location of the Philadelphia airshe;
modeling region.
B3033f
13
-------
describes the meteorological and air quality conditions associated with
the 13 and 19 July 1979 study days, considering, in particular, the
general transport characteristics of each day and the resulting pollutant
patterns observed over the monitoring network. The maximum observed
ozone concentrations and the predicted values for the two days is pre-
sented in Tables 1 and 2. Tables 1A and 2A present a detailed comparison
of the maximum measured and calculated ozone concentrations for each of
the 22 monitoring stations. We note that throughout this section any
reference to time refers to time at the beginning of the hour (e.g., 0500
EOT represents 0500-0600 EOT).
Characterization of the 13 July 1979 Ozone Episode
The highest and most widespread ozone concentrations measured during
the summer of 1979 occurred on Friday, 13 July. These high concentrations
were the result of a buildup of precursors resulting from near-stagnant
conditions on the previous day, and generally weak and variable winds pre-
vailing until noon of the 13th, when a stronger southerly flow was
established. The synoptic pattern showed high pressure throughout the day
dominating the surface and upper-level (2000 m) flow fields (see Figure
3). This high-pressure system was part of the seasonally semipermanent
Bermuda high-pressure cell centered east of Florida. The high pressure
weakened through the course of the day as a trough (the remnants of Hurri-
cane Bob) over the Ohio Valley moved slowly eastward bringing precipita-
tion to western Pennsylvania.
Light west-northwesterly surface winds were present throughout the
region from noon on 12 July until the early-morning hours of 13 July when
the winds became calm or very light with a northerly direction. This wind
flow pattern was responsible for transporting ozone and precursors to the
southeast of the region during this period. Surface wind measurements
showed these calm-to-very-light winds during the early-morning hours of
13 July with northerly directions increasing slightly in speed and shift-
ing to a general southerly direction by late morning (1000 EST). This
shift in surface wind direction resulted in a recirculation of material
that had been transported to the southeast during the previous 24 hours:
(1) Influx of ozone and precursor from the southeast; (2) a buildup of
regional ozone and precursors from near stagnant conditions on the pre-
vious day (12 July); (3) the day's emissions; (4) generally light winds;
(5) high region-wide temperatures; and (6) mostly clear skies were the
primary conditions leading to high ozone concentrations.
Evidence of this recirculation of ozone and precursor material and
the possible existence of a reservoir of ozone aloft from the previous day
is shown in the observed ozone concentration data at upwind monitors. For
85068r2 3
14
-------
o.
o.
o
IVI
o
OC
O
cr>
r^
cr>
CO
oo
z
o
a:
LU
to
CO
o
oo
z
o
at
3
<0
u. a t-
O UJI <
»-=><*
a: or H
=>«-iZ
O o x ui
z LU < o
*£ Z
a. o
u
or <
U. LU 3 OL
o >r t-
e: ^ z
Q£ LU X LU
3 -t i
CM
110 ^^
u. o a <
O LU LU O£
O u a: Z
»- « LU LU
K a co u
< LU CD Z
oi ec o O
a. u
o
•—i
o
ct:
a.
x
•a:
ur <
»-=>«:
O X LU
LU < (J
O
o
CO
a h-
Lur <
CO X UJ
00 < U
or z
o
u
to
z
o
z a. «/»»-• < a. _i tf
UJfluZ< O.QQ
i£ = Z « Z —
O 00 O O
BO -> CX:ZXLU(J>« H»-
«O ZLU O O tfl z_jo oc:»
-JOCHLU >- r x z«i<: zuj«
« r
15
-------
TABLE 1-B. PERFORMANCE MEASURES FOR THE UAM SIMULATION OF THE
13 JULY 1979 EPISODE IN PHILADELPHIA
Performance Attribute
Performance Measure
Values
Accuracy of the airshed Ratio of predicted to
peak prediction measured station peaks
Time difference between
predicted and measured
station peaks
Station Peaks
Predicted
Station peaks
Systematic bias
Gross error
Measured
Mean Deviation
Normalized
Average
Std. dev.
Nonnormalized
Average
Std. dev.
Bounds at the 90
percent confidence
level
Mean absolute deviation
Normalized
Average
Std. dev.
Nonnormalized
Average
Std. dev.
0.940
-2 hours
19.3 pphm
(Roxy Water)
20.5 pphm
(Conshohocken)
•0.086
0.180
•0.998 pphm
2.491 pphm
•2.790 and 0.793 pphm
0.148
0.130
2.021 pphm
1.718 pphm
85068T *»
16
-------
TABLE 1-B (Concluded)
Performance Attribute
Performance Measure
Values
All 03 Concentrations > 5 pphm
Systematic bias Mean deviation
Normalized
Average
Std. dev.
Nonnormalized
Average
Std. dev.
Gross error
Temporal correlation
Spatial alignment
Mean absolute deviation
Normalized
Average
Std. dev.
Nonnormalized
Average
Std. dev.
Temporal correlation
coefficients
Each station
All-station average
Spatial correlation
coefficients
Each hour
All-hour average
•0.150
0.361
•1.660 pphm
2.885 pphm
0.295
0.255
2.670 pphm
1.981 pphm
0.273 to 0.956
0.737
0.041 to 0.735
0.456
85068p
17
-------
o.
Q.
UJ
z
o
M
O
CC.
O
o>
t—I
>-
_l
ID
•-D
o:
o
CO
o
a:
UJ
oo
CQ
o
•^
co
O
LU
CtL
CL.
I
CVJ
LU
_J
CO
<0
i i
U.O
QCOZ H-
3«-iZ
O Q x ui
O
u
u. LU ^ ee.
O> ZH-
OooXZ
xo o
00
ui ui O
CP CJ ^
of
0X1-
O X UI
UJ < O
o
u
aiZH-
UJ •-• Z
to x uj
oo < u
0±x
O
o
en
O
t-4
K
C/)
z a. to « <« a. _j cs
LUQ- z< < a. 00
ikiOZ"* z a: »- o of oo
CKZXUJ. - vt H- X z
ui < zu. X _iXot oc x =>Xuj
18
-------
TABLE 2-B. PERFORMANCE MEASURES FOR THE UAM SIMULATION OF THE
19 JULY 1979 EPISODE IN PHILADELPHIA
Performance Attribute
Performance Measure
Values
Accuracy of the airshed Ratio of predicted to
peak prediction measured station peaks
Time difference between
predicted and measured
station peaks
Station Peaks
Predicted
Measured
Station peaks
Systematic bias
Gross error
Mean deviation
Normalized
Average
Nonnormalized
Average
Std. dev.
Bounds at the 90
percent confidence
level
Mean absolute deviation
Normalized
Average
Std. dev.
Nonnormalized
Average
Std. dev.
0.836
1 hour
14.2 pphm
(Downington)
17.0 pphm
(Roxy Water)
•0.021
0.275
0.205
0.179
2.293 pphm
1.930 pphm
8506er
19
-------
TABLE 2-B (Concluded)
Performance Attribute
Performance Measure
Values
All O concentrations > 5 pphm
Systematic
bias
Gross error
Mean Deviation
Normalized
Average 0.005
Std. dev. 0.398
Nonnormalized
Average 0.024 pphm
Std. dev. 3.163 pphm
Mean absolute deviation
Normalized
Average
Std. dev.
0.288
0.273
Temporal correlation
Non normalized
Average
Std. dev.
Temporal correlation
coefficients
2.400 pphm
2.051 pphm
Spatial alignment
Each station
All-station average
Spatial correlation
coefficients
•0.914 to 0.970
0.525
Each hour
All-hour average
•0.421 to 0.643
0.202
85068r
20
-------
FIGURE 3. Synoptic situation, 0700 EST, July 13, 1979.
(Source: Allard et al., 1981)
033,
21
-------
the morning of 13 July, stations to the south and east were upwind. Exam-
ination of the ozone concentration during the initial onset of mixing is
one way of providing information on ozone levels aloft. At 1000 EST,
observed ozone concentrations were 8.4 pphm at Brigantine, New Jersey; 9.7
pphm at Lumber-ton, New Jersey; and 12.3 pphm at Ancora, New Jersey. Ano-
ther upwind monitor (Vineland, New Jersey) could not provide ozone data
for 1000 EST; however, an hour earlier (900 EST) this monitor recorded an
ozone concentration of 10.2 pphm. For this time period (onset of mixing),
these ozone concentrations were among the highest observed during the
summer of 1979. This indicates that a large reservoir of ozone existed
aloft and was available for mixing down to the surface.
No upper-air measurements were available in Philadelphia on this day;
however, radiosonde data from New York City (JFK Airport) and Washington
D.C. (Dulles International Airport) showed very light westerly winds at
2000 m throughout the day.
Because surface wind flow patterns established a southerly direction
by late morning, peak ozone concentrations occurred north of the high
urban emission source region. The highest ozone concentration recorded on
this day was a value of 20.5 pphm at Conshohocken at 1600 EST. The second
highest value of 20 pphm occurred earlier in the afternoon at the Roxy
Water Pump monitor at 1400 EST. Thirteen monitors recorded ozone concen-
trations greater than 12 pphm.
Characterization of the 19 July 1979 Ozone Episode
The second-highest peak ozone readings in the Philadelphia Air
Quality Control Region (AQCR) during the 1979 summer oxidant data enhance-
ment study occurred on Thursday, 19 July (Allard et al., 1981). On this
day, the Philadelphia urban plume was transported to the west of the cen-
tral city and precursor transport from the New York/New Jersey urban area
was the apparent cause of a substantial portion of the ozone concentra-
tions measured in the Philadelphia area.
A broad high-pressure area extending from the midwest through the
northeast to Nova Scotia influenced surface wind patterns. The eastern
core of this high-pressure area was situated to the north of Philadelphia
in the early morning hours, bringing light northerly winds through the
airshed. This high-pressure ridge moved eastward throughout the day. As
a result, the surface winds veered to an easterly direction around noon
and then to a southeasterly flow that persisted the rest of the day.
Winds aloft showed a persistent west-southwesterly flow during the entire
day. Surface trajectory analysis indicated that air parcels arriving in
the urban center of Philadelphia at 1200 EST originated to the northeast
in the New Jersey/New York urban area.
85068T2 3
22
-------
The peak ozone concentration recorded on 19 July 1979 was 17.0 pphm
at the Roxy Water Pump monitoring station 10 km northeast of central
Philadelphia. The same value was also recorded at Conshohocken. An ozone
reading of 16.0 pphm was also recorded at Downington, 45 km due west of
downtown Philadelphia. Six monitors recorded ozone concentrations equal
to or greater than 12 pphm and an additional four monitors recorded ozone
concentrations greater than or equal to 10 pphm.
INPUTS USED IN THE URBAN AIRSHED MODEL SIMULATIONS
Detailed descriptions of the meteorological, emissions, and boundary
concentration inputs used in the UAM are presented in the appendix to this
document. In this section, we summarize some of the inputs used for each
of the two modeling days.
The study region used in the Urban Airshed Model simulations was
defined during the development of the emission inventory by EPA (1982).
The modeling region itself is 12,500 km2 in area and comprises five
counties in New Jersey, five counties in Pennsylvania, and one county in
Delaware. The emission grid system used for the emission inventory con-
sists of 502 cells, 5 km on a side. Surrounding the modeling region
(Figure 1) is a row of cells representing the boundary cells and contain-
ing the boundary conditions (concentrations) used in the UAM simula-
tions. Initial conditions for all species were specified using all avail-
able monitoring data in the Philadelphia region and these data are shown
in Tables 3 and 4 for 0000 EST on 13 and 19 July.
The physical boundaries used in the 13 July simulation are presented
in Figure 4. Because of the stagnation characteristics of this episode,
much of the large airshed region was "blocked off" and not included in the
simulation. No cells containing major emission sources were excluded by
this procedure. Estimated background values for all species for 13 July
were designated for all boundaries except the Southeast boundary (Table
5). Concentrations of NO, N0~, CO, and hydrocarbons were specified on the
basis of work by Kill us (1982) to reflect concentrations of a typical
urban atmosphere. Because an urban plume from Philadelphia was trans-
ported to the east late on 12 July and then recirculated back by a south-
easterly flow on the 13 July, there 1s a different set of concentrations
for the Southeast boundary grid cells below the mixing height (Table 6).
The concentrations were made to duplicate the inflow of aged air parcels
from the previous day's emissions.
Because of wind flow through the airshed on 19 July and the need to
limit simulation costs, certain unnecessary grid cells were eliminated.
85068T3 3 23
-------
TABLE 3. INITIAL CONDITIONS FOR 13 JULY 1979 (CONCENTRATIONS IN PPM)
Station
AMS Lab.
Ancora
Brigantine
Bristol
Camden
Chester
Claymont
Conshohocken
Defense Support
Downington
Franklin Inst.
Island Rd. Airp. Cir.
Lunberton
Norristown Armory
Northeast Airp.
ftoxy Water Pump
SE Sewage Plant
South Broad
Summit Bridge
SW Corner Broad/Butler
Trenton
Van Hiseville
Vineland
Easting
(m)
491600
511800
546000
511000
491700
469000
461500
474500
483800
436000
485200
480300
518000
473500
499000
479500
487200
486100
441000
487000
520000
559000
498200
Northing
(m)
4428500
4392400
4377506
4440000
4419000
4410000
4406400
4435600
4418300
4426000
4422800
4414800
4423000
4440000
4436000
4433100
4417300
4421600
4376000
4428000
4452000
4439500
4371200
NO
0.005
—
—
—
0.022
—
--
--
—
0.002
0.010
--
0.006
—
0.035
—
—
0.020
0.000
—
—
0.001
0
-
-
-
0
0
0
-
-
0
0
-
0
0
0
-
-
0
0
-.
-
0
N02 CO
.075 1.5
-
-
.
.105 2.4
.030 —
.045 —
.
.
.012 0.2
.065 —
-
.031 1.2
.075 —
.060 —
-
-
.080 3.5
.003 0.2
i* tm ••
.006 0.2
0.024 --
0
0
0
0
0
0
-
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
°3
.035
.051
.037
.000
.004
.058
-
.000
.020
.073
.035
.010
.010
.000
.000
.045
.010
.035
.046
.035
.009
.005
.054
RHC
1.05
—
—
—
—
2.10
—
—
—
0.00
—
—
0.20
0.39
—
--
—
0.30
0.05
—
~
--
—
Carbon-Bond Fraction
RHC Component (%
PAR
OLE
ETH
ARO
CARB
as Carbon)
74.0
2.8
4.1
13.2
5.9
83033TH 6
24
-------
TABLE 4. INITIAL CONDITIONS FOR 19 JULY 1979 (CONCENTRATIONS IN PPM)
Station
AMS Lab.
Ancora
Brigantine
Bristol
Camden
Chester
Claymont
Conshohocken
Defense Support
Downington
Franklin Inst.
Island Rd. Airp. Cir.
Lumberton
Norristown Armory
Northeast Airp.
Roxy Water Pump
SE Sewage Plant
South Broad
Summit Bridge
SW Corner Broad/Butler
Trenton
Van Hiseville
Vineland
Easting
(m)
491600
511800
546000
511000
491700
469000
461500
474500
483800
436000
485200
480300
518000
473500
499000
479500
487200
486100
441000
487000
520000
559000
498200
RHC
Northing
(m)
4428500
4392400
4337506
4440000
4419000
4410000
4406400
4435600
4418300
4426000
4422800
4414800
4423000
4440000
4436000
4433100
4417300
4421600
437600U
4418000
4452000
4439500
4371200
NO
0.000
--
--
--
0.011
--
--
--
--
0.
0.
—
0.
—
0.
--
—
0.
0.
--
—
0.
--
006
030
000
000
020
000
000
N02 CO
0.
--
--
0.
0.
0.
0.
--
—
0.
0.
—
0.
--
0.
--
--
0.
0.
--
—
0.
0.
025 0.5
--
—
039 —
051 2.5
024 —
045 —
--
--
015 0.2
050 0.5
--
028 0.0
—
015 —
—
—
045 1.5
009 0.2
—
,
006 0.2
025 —
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
—
0.
0.
0.
0.
--
0.
03
015
015
017
012
004
026
020
000
000
024
010
015
008
030
010
005
010
009
002
RHC
0.35
--
--
--
—
0.95
—
--
--
0.05
--
—
0.40
0.29
—
--
--
0.25
0.10
--
--
0.021 —
0.
010
--
Carbon-Bond Fraction
Component (%
PAR
OLE
ETH
ARO
CARB
as
74.
2.
4.
13.
5.
Carbon)
0
8
1
2
9
J033P
3 8
25
-------
0
HORJH
10
90
30
10
DQUN
SUM
T7T
TKEN KQBB
HOW
CON*so
SHEWS
LUMB
£>£/£?§»*
JSLA
cues
CLfiY
RNtQ
6*15
0
Southeast Boundary':
F'T '•IW'XW -+^ I « I i .1 1 I I f I I t I i « i j , i
10
20
30
SOUTH
FIGURE 4. Physical boundaries used in the 13 July 1979 Airshed Mode1
simulation (shaded area is not included in the simulation).
63033r
26
-------
TABLE 5. BACKGROUND CONCENTRATION VALUES
FOR 13 JULY AT THE TOP OF THE MODELING REGION—
AS INITIAL CONCENTRATIONS ABOVE THE MIXING HEIGHT,
AND FOR ALL LEVELS OF ALL BOUNDARIES EXCEPT THE
LEVELS BELOW THE MIXING HEIGHT ON THE SOUTHEAST
BOUNDARY
Species
Concentration
(ppm)
NO
N02
03
CO
ETH
OLE
PAR
CARB
ARO
PAN
BZA
0.001
0.002
0.08
0.2
0.001
0.0004
0.040
0.010
0.0008
0.000025
0.00001
3033P3 8
27
-------
o
z ^
"* 5-
x f-
5-8:
^ 2
re
^ oo
?£
_j |—1
d 5
* £
2*
d w
d y
« I
« s
^ <*
^2
0 ^
s =
0 X
z ^
Ss
£ u_
< °
?*
= 2
§5
Si
1 1 1 _
I ">
= £
S^
ac:
o
. LL.
10 1-
Uj2
Is
—
UJ
UJ
_J
0
o:
««
Q.
CO
o
CVJ
0
o
0) t-
E 0)
•I- 4J
H- C
OOOOO oo^t«*CNJCV40OC\jr^or^C\!Osj
OOOOOOOOOOOCDOOOOOC3
oooooooocooocoooco
oooooooooooooooooo
vovovovovovovoi/)
oooooooooooooooooo
ro^r»-voooo«— »mcvjcsj-^«— lO^ro^co^ oo
un^-roroc\ji-HCsjrr>r^o«-«c\jPO^-»o^to vo
oooooooooooooooooo
o>o%^^J>a>a^r>.^-cMCvjcvjeMC\jcNjc>jcvj c\j
oooooooooooooooooo
ooooooooooooooooo o
oooooooooooooooooo
,_t,_4f-Hf_4f-^,— li— lf-4<-H<— It— It— If— «,-«,— (,-Hf— 1 f— 1
oooooooooooooooooo
oooooooooooooooooo
oooooooooooooooooo
oooooooooooooooooo
oooooooooooooooooo
^-icvjro^j-mvop^oo^o^^eNjro^-'^voi^ oo
OOOOOOOOOt— •«— it— ii— i »-H t— i i— i f-i «— i
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
oooooooooooooooooo
ooooooooooooooooo o
o»— iCMCo^tnvor^oo^o^^cvjco^LOvor^
OOOOOOOOOO«-*«— 1«-^«— 1«— if— ii— i r- •
23
-------
Q
UJ
Q
o
o
CM CVI CM CM CM CM
PO CO PO PO PO PO
oooooo
Q- OOOOOO
oooooo
oooooo
o o o o o
CO OOOOOO
OO 00 00 00 30 00
o oooooo
3Z OOOOOO
oooooo
oooooo
oooooo
oooooo
oooooo
oooooo
_ oooooo
o oooooo
000*000
VD VO VO VO VD VO
oooooo
at oooooo
a. oooooo
00 ID ^ VO «—I O
ID VO ID ID ID ID
co oooooo
o ••••••
oooooo
_ CM CM CM CM CM CM
o oooooo
z: oooooo
oooooo
•-^ «—< «-H «-H f-M ,—1
oooooo
o oooooo
oooooo
0)
oooooo
oooooo
Cy» O »-< CM PO ^>
«—I CM CM CM CM CM
oooooo
oooooo
00 O> O t-H CM PO
«-• «-• CM CM CM CM
29
-------
The physical boundaries used in the simulation are shown in Figure 5.
Background values were designated for all boundaries except the East and
Northeast boundaries. Estimates of boundary conditions for the East and
Northeast boundaries below the mixing height are presented in Table 7.
Observed NO and N02 for the 0500-0600 EST were used as input boundary
values for 0000-0600. Since hydrocarbons were not measured at the Van
Hiseville monitor, estimates of the influx of total reactive hydrocarbons
across the Northeast and East boundaries below the mixing height were
specified by multiplying the hourly NOX concentrations at the Van Hi Se-
ville Monitor by the Philadelphia surface emission inventory hydrocarbon/
NOX ratio of 6.
Mixing height profiles for 13 July were estimated using temperature
soundings of JFK and Dulles airports since no soundings were available for
Philadelphia on this day. Mixing heights for 13 July are shown in Table 8
and Figure 6. Mixing height profiles were developed for 19 July using
available radiosonde observations and sodar data for the Philadelphia
area. Mixing heights for 19 July are shown in Table 9 and Figure 7.
Three-dimensional wind fields for 13 July and 19 July were generated
from station measurements and aloft winds obtained from radiosonde data.
Wind direction on the morning of 13 July was very light and from the
north. Around 1000 EST, however, a shift occurred and winds with higher
speeds came from the southeast for the rest of the day. Winds on the
morning of 19 July were predominantly from the north. Throughout the day
a 180° shift occurred; at noon the wind direction was predominantly from
the east; during the evening it came from the south-southeast. The grid-
ded, smoothed surface vectors for selected hours are presented in Figure 8
for 13 July and in Figure 9 for 19 July.
Except for the ozone concentration, the estimated background concen-
trations above the mixing height are the same for both days. On the basis
of examination of upwind monitoring data at the time of mixing, the value
specified for ozone was reduced to 0.06 ppm. Estimated background concen-
trations are listed in Table 10. The gridded and elevated point source
emission inventory was prepared for EPA in 1981 by Engineering Science,
Inc. (EPA, 1982). Daily emission values for total NOX and total hydrocar-
bon are presented in Table 11.
INPUTS USED IN THE SYSTEMS APPLICATIONS TRAJECTORY MODEL SIMULATIONS
The meteorological and emissions files used in the Systems Applica-
tions Trajectory Model are the same as those used by the UAM. By running
the Systems Applications Trajectory Model in a backward mode, a path to
the maximum ozone observed can be determined for the day of interest.
85068P3 3 30
-------
30 r
30
Northeast Boundary
i:?;j^
East
Boundary
FIGURE 5. Boundary specifications for 19 July 1979 simulation
(Bold line is emissions region.)
63033r
31
-------
•x.
•— cc
X 0
— H-
z
LU O
Z Z
r—
>-
3 LU
O CO
_i ac
LU LU
CC ""D
LU LU
T BOUNDARI
SEVILLE, N
«« z
LU
•z.
c <:
z >
LU
t— Z
00 H-
LU \—
z <:
h-
CC O
O LU
Z •—
CJ
LU LU
Z -J
H- _J
O
O
u_ «-^ +^
00 t-
>- LU 4->
< »— OJ
Z O C
=D •— O
CC Z ^-'
•
!•«•*
LU
_j
»—
z
<
a.
CD
O
Z
H-
LU
LU
_J
O
CC
a.
ro
O
CVJ
0
O
h-
co
LU
L.
O
z
o
o
o
0
cvj
o
o
o
o
o
LO
CO
o
o
o
CVJ
vo
o
o
o
LO
CO
o
«— 1
o
o
CO
o
o
o
0
o
0
o
1
o
o
o
0
OOOOOOOO«-*»—«'— ii— 1«— (OOOO
ooooooooooooooooo
ooooooooooooooooo
oooooooocoooooooo
CVJCVJCVJCVJCVJCOCOCVJ.— »OOOOOOOO
ooooooooooooooooo
ooooooooooooooooo
r^. r-» cvj 00
VOVOVOVOVOCVJOOUD COVO«—«LO CVJ VO VO
OOOOO*— •CVJCT^«*«— • O O O O O O O
ooooooooooooooooo
ooooooooooooooooo
LO
LO lOLOLOOOOCTNCVJ
LOLOLOLOLOCVJLOCOf"-^«— IVOCO^COCVJLO
COCOCOCOCOOOCOO^>COr<-VOOOLO«*'LOVO
CT^ CT^ O^ O^ CT> C3 ^^ CT^ ^^ '•^ CD CD CD CD CD CD CD
OOOOO«— 1«— lOOOOOOOOOO
ooooooooooooooooo
ooooooooooooooooo
LO LO
LO tor^coLO^^''— 'LO
CVJCVJCVJCVJCVJCfiLO LO«-^OOLOOCrtCTNVO^
vOvovovOvOP^r^vOCOOOOOOOOO
ooooooooooooooooo
ooooooooooooooooo
ooooooooooooooooo
LOLOLOLOLOO>OCVJVO^fCVICVJCOCVJ«-*.— iCVJ
cocococococo^-co.— 100000000
ooooooooooooooooo
CD CD CD CD CD CD CD ^^ LO 00 r^ ^^ P1^ f"^ vo LO LO
ooooooooooooooooo
ooooooooooooooooo
COCOCOCOCO^-<^-«!j-CVlOOOOOOOO
ooooooooooooooooo
ooooooooooooooooo
ooooooooooooooooo
ooooooooooooooooo
ooooooooooooooooo
ooooooooooooooooo
o o o o o o o o »-* «— « t— i i™* «-^ i— » «— i f— i «— A
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
ooooooooooooooooo
ooooooooooooooooo
o
0
o
0
0
0
o
o
o
o
o
LO
00
00
0
o
o
o
LO
o
vO
o
o
o
o
0
o
0
o
o
o
0
^^
o
0
o
o
f— 1
1
o
0
o
o
o
0
o
o
o
CO
a>
o
Q
o
o
LO
00
o
o
o
o
LO
VO
o
o
o
o
o
o
0
o
o
o
0
o
0
o
o
CVJ
1
o
0
*— 1
0
o
o
o
o
o
o
CVJ
o
o
o
o
LO
o
LO
o
o
o
o
LO
CO
0
0
0
o
00
o
o
o
o
o
o
o
o
o
o
o
CVJ
1
o
o
o
o
o
o
o
0
o
o
VO
o
o
o
o
LO
«— 1
VO
o
o
o
o
CVJ
o
o
o
o
CVJ
o
o
o
o
o
o
0
£\J
^^
o
o
o
o
CVJ
1
o
o
o
o
0
0
0
o
o
0
LO
T-4
o
0
o
LO
LO
CO
0
o
o
LO
CVJ
o
o
o
o
00
o
o
o
o
0
o
0
0
LO
o
o
o
o
CVJ
1
o
o
o
o
o
o
0
o
o
o
LO
o
o
o
4-J
LO -f
LO —
co a.
i— • CO
o
o >,
O 0
4->
c
LO OJ
CVJ >
o •—
o
O «tJ
O £
a.
'aJ
*3- -0
00 TO
00 r-
^^ •*••
O J=
o
o
o
o
o
0
o
flrt
o
o
o
o
CVJ
1
o
o
«—
id Fractioi
Carbon)
O CO
CO 'O
O ^-"
5
tn
Oi
O
-------
TABLE 8. URBAN AND RURAL
MIXING HEIGHT VALUES USED
IN THE DIFFBREAK FILE FOR
13 JULY 1979
Time
(EST)
0000
0030
0100
0130
0200
0230
0300
0330
0400
0430
0500
0530
0600
0630
0700
0730
0800
0830
0900
0930
1000
1030
1100
1130
1200
1230
1300
1330
1400
1430
1500
1530
1600
1630
1700
1730
1800
1830
1900
1930
2000
2030
2100
2130
2200
2230
2300
2330
2400
Urban
(m)
250
250
250
250
250
250
250
250
250
250
250
250
250
270
295
375
450
680
925
1160
1200
1330
1480
1480
1480
1500
1530
1530
1530
1475
1420
1365
1310
1220
1130
1035
950
855
770
675
590
500
410
370
330
290
250
250
250
Rural
(m)
100
100
100
100
100
100
100
100
100
100
100
100
100
135
150
250
350
620
925
1160
1200
1330
1480
1480
1480
1500
1530
1530
1530
1475
1420
1365
1310
1220
1130
1035
950
730
525
320
100
100
100
100
100
100
100
100
100
33
-------
Legend:
•. Urban
... Ruro I
0
12
Time (hours)
IB
FIGURE 6. . Mixing height profile for urban and
rural cells for the 13 July 1979 simulation.
B3033r
34
-------
TABLE 9. URBAN AND RURAL
MIXING HEIGHT VALUES USED
IN THE DIFFBREAK FILE FOR
19 JULY 1979
Time
(EST)
0000
0030
0100
0130
0200
0230
0300
0330
0400
0430
0500
0530
0600
0630
0700
0730
0800
0830
0900
0930
1000
1030
1100
1130
1200
1230
1300
1330
1400
1430
1500
1530
1600
1630
1700
1730
1800
1830
1900
1930
2000
2030
2100
2130
2200
2230
2300
2330
2400
Urban
(m)
190
190
190
190
190
190
190
190
190
190
190
190
190
190
190
240
350
460
600
740
890
1060
1240
1420
1530
1530
1480
1410
1340
1270
1200
1120
1020
940
850
750
660
570
480
410
340
300
250
250
250
250
250
250
250
Rural
(m)
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
260
440
490
540
560
580
630
880
1320
1430
1430
1380
1310
1240
1170
1100
1020
920
840
750
480
160
100
100
100
100
100
100
100
100
100
100
100
100
35
! B
-------
1700
1600
1500
1400
1300
•j 1200
Q) 1100
Q)
1000
900
.*>
.c
•? 800
Q)
^ 700
CD
•- 600
x
£ 500
400
300
20S
100h
' i i i i I i i i i i I i i i i i I i i
Legend:
_ Urbon
... Ruro I
'''''''''''II
12
Time (hours]
IB
FIGURE 7. Mixing height profiles for urban and rural
cells on 19 July 1979.
83033,
36
-------
I
J I
• s
NIND SPEED (M/S)
5 30
'•/i/iVi/i/i/i/l/i/i/i/i
t t ft t t t ft t
f} f 111 f 11
f 1111 r f 11
15 20
(a) 0 - 100 EST
FIGURE .8. Airshed model surface winds for 13 July 1979.
37
-------
I i (
I I
15
• 5
HIND SPEED IH/S)
20 25
25
15
10
4-14 4 4 441 ***"«'*** /
32
I 4 4
* *
\ vvv
V J f V
\ Y Y ^ jg - r
\**"**f4***\\V V~5" ^
K- > ' * < t ^ X\i^^
^\^xf t • t * % % x^^NN^
??v\^v** x ^^^^^^N ^'V^
^•^^^S\\ ^\\V\\^XV ^^'*^'^>*'^^'N.^^2
^^^^\\\\ {>J>?\\^X^^^-^--Hr^^^^^I^
^^^^\\\VV>f >? \\^>^----~H***>^^
*^^ " ^ V \ V \ \ ^X^^^* ** "" * * "* "* "*'^8^*^* -S. *
* * *
^
J*
^L
10 15 20
(b) 600 - 700 EST
35
FIGURE 8 (continued)
63033r
33
-------
0
10
15 20 25
(c) 1200 - 1300 EST
30
FIGURE 8 (continued)
)33,
39
-------
0
10
15 20
(d) 1800 - 1900 EST
25
30
FIGURE 8 (concluded)
83033r
40
-------
m
(a) 400 - 500 EST
FIGURE 9. Airshed model surface winds for 19 July 1979
63033r
41
-------
I I I I I I
• 5
HIND SPEED (M/S)
20 25
30
10
15 20
(b) 1200 - 1300 EST
25
30
35
FIGURE 9 (continued).
63C33r
42
-------
lilt
1 r r i' r r r r r r r r r r r r r r r r r r r i
0
15 20 25
(c) 1800 - 1900 EST
30
35
FIGURE 9 (concluded).
B3033r
43
-------
TABLE 10. BACKGROUND CONCENTRATION VALUES
FOR 19 JULY USED AT THE TOP OF THE MODELING
REGION (TOPCONC), AS INITIAL CONDITIONS, AND
FOR ALL BOUNDARIES EXCEPT THE LEVELS BELOW
THE MIXING HEIGHT ON THE NORTHEAST AND EAST
BOUNDARIES
Species
NO
N02
°3
CO
ETH
OLE
PAR
CARB
ARO
PAN
BZA
Concentration
(ppm)
0.001
0.002
0.06*
0.2
0.001
0.0004
0.040
0.010
0.0008
0.000025
0.00001
A value of 0.05 ppm was used below
the mixing height for all boundaries
except the Northeast and East
boundaries.
85068 4
44
-------
o
or
or
o
o en co vo oo
vo cr* co cvj ,-<
CT^ r***« oo o^ in
• A A A A
o> c t-H m oo
co **H cy^ o^ co
i—• in ^f *j- 10
o
in c if) r^
vo oo in cr>
r-* cc in vo
A A A A
o co o> r^
f*. »—i vo co
in oo vo co
00
A
vo
CO
A
«*
3:
o.
CO
CJ
oc
C^ CO CJ
CVJ ^ LO
vo
^t
oo
•
tr>
t-t in vo
*± co CM
oo vc cvj
vc »-H
OJ i— I
CM
I
r-l CO
in
in
oo
vo
«3-
VC
O
in
00
CO
00
^t ^f o
cr vo in
*-H vo c
LU
d.
LU
O
or
=5
O
oo
CO
00
z:
c
»— i
00
00
UJ
>-
cv
vo
oo
O
in
cvj
oo
CVJ
•
in
oo
co
OC
m
CM
•
in
o>
cvj
oo
c
00
ir>
CO
cvj
oo
vo
•o
+J
o
o
cvj
in
in
vo o
oc
ro
in
o
•
cvj
in
o
cvj
oo
co
oo
oc
vo
•
vo
^»-
vo
•>
vo
C
CVJ
in
CVJ
CVJ
oc ,-«
CO
CVJ
CO
cvj
CO
00
A
CO
vo
o f—
«o
4->
o
to
£
a
oo
c
o
JD
t.
U
O
C-
^^
^
(O
f
Q.
'o)
TJ
^
o
fl»*
Q.
OO
C 00 •-• CVJ C^
^- cvj ^j- co in
r^. *— i
cc
or tu 3: c or
O «J -r-
* c a> ^5
i— «r- C_ O
uj ar «c z
45
-------
This path 1s usually based on the wind fields in the first vertical layer
only. As discussed earlier, the use of the wind fields in only the first
layer can lead to differences in the maximum ozone predicted by the Sys-
tems Applications Trajectory model and the UAM on days with wind shear.
For 13 July, the two highest observed ozone values occurred at both Con-
shohocken and Roxy Water Pump monitor. However, since the Roxy Water Pump
monitor also had the highest predicted ozone value, it was used to deter-
mine the trajectory path to the ozone maximum for 13 July. The ozone
value at Downington was used to determine the trajectory path to the ozone
maximum for 19 July. Downington was chosen because in addition to having
the highest predicted and a high observed ozone concentration, it was also
estimated to be primarily influenced by the Philadelphia plume and not by
the boundary conditions. The trajectory paths for the two modeling days
are shown in Figures 10 and 11. Because the wind flow between the two
lowest levels was important after 1100 EST on 13 July the average of the
two levels was used to determine a trajectory path.
INPUTS USED IN THE LEVEL II OZIPM/CBM MODEL SIMULATIONS
Using the results of the Systems Applications Trajectory Model, emis-
sions and meteorological conditions are constructed for a Level II type of
EKMA model with Carbon-Bond chemistry (OZIPM/CBM). The Trajectory Model
prints out the emissions and mixing heights along the trajectory path at
hourly intervals. These mixing heights can be used as direct inputs to
the OZIPM computer code. The emission rates from the Trajectory Model are
in units of moles/hr for each of the Carbon-Bond species and are converted
to ppmC/hr for volatile organic compounds (VOC) by the following equation:
Emissions Rate (ppmC/hr) = [24450/(ZQ* 40002)][Emissions Rate (moles/hr)]
where Zg is the initial mixing height.
Initial NOX and VOC precursors for OZIPM are determined from the
initial conditions using the average of the layers below the mixing height
in the Trajectory Model. NOX and VOC precursors aloft are determined from
the amount of precursors that entered the mixed layer from aloft in the
Trajectory Model simulations. As a result, the OZIPM NOX and VOC aloft
precursors are often higher than measured values. The chemistry used in
the Trajectory Model will produce secondary organic products (such as
carbonyls) and ozone in the layer aloft. Thus, if we were to compare
assumed aloft morning conditions in the CBM/OZIPM with actual morning
measurements, concentrations may be higher for the OZIPM models. This
85068T3 3
46
-------
to
rj••' '''•••
rOODC
SOUTH
(a) Averaged trajectory
to
10
SO
(b) Level I trajectory
65066
0000
20
SOU 7*
(c) Level II trajectory
FIGURE 10. Trajectory paths for 13 July 1979.
47
-------
30
JO
NQF.Th
10
20
30
30
1600
0100
Stfffff
BKIS
VINE
\ \ i i i i _± > > • t i i
20
30
SOUTH
FIGURE 11. Trajectory path for 19 July 1979.
85068
48
-------
method was used because the species aloft continuously react In the S«
^
conditions for
INPUTS USED IN THE LEVEL III EKMA MODEL SIMULATIONS
i
Plan (SIP) for Philadelphia AQCR fDER 19831 Al? f« • P lementatlon
85068P3 3
49
-------
TABLE 12. INITIAL CONDITIONS AND EMISSIONS RATES USED IN THE
LEVEL II CBM/OZIPM CALCULATIONS FOR 13 JULY 1979
(a) Initial Conditions
Species Surface Aloft
NOX
NMOC
Oo
0.0253 ppm
0.7184 ppmC
0.0412 ppm
0.0004 ppm
0.0539 ppmC
0.0775 ppm
N02/NOX = .9922
Hydrocarbon reactivity
Surface
OLE =
PAR =
ARO =
ETH =
CARB =
0.0277
0.7391
0.1311
0.0409
0.0612
carbon fraction
Aloft
0.0060
0.7148
0.0762
0.0311
0.1719
(b) Emissions Rates and Mixing Heights
Species
Time
(CDT)
0500-0600
0600-0700
0700-0800
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
1500
VOC
(ppmC/hr)
0.0331
0.0802
0.4684
0.5557
0.3221
0.1800
0.1669
0.1528
0.1114
0.0802
NOX
(ppm/hr)
0.285
0.359
0.806
0.1110
0.0593
0.0318
0.0255
0.0224
0.0248
0.0180
Mixing Height
at the Beginning
of Each Hour
(m)
250.0
250.0
295.0
450.0
925.0
1200.0
1480.0
1480.0
1530.0
1530.0
1420.0
85068 <*
50
-------
TABLE 13. INITIAL CONDITIONS AND EMISSIONS RATES USED IN THE
CBM/OZIPM CALCULATIONS FOR 19 JULY 1979
(a) Initial Conditions
Species
NOX
NMOC
Surface
0.0073 ppm
0.1484 ppmC
0.0281 ppm
Aloft
0.0008 ppm
0.0534 ppmC
0.0611 ppm
N02/NOX = 0.988
Hydrocarbon reactivity
Surface
OLE
PAR
ARO
ETH
CARB
0.0267
0.7358
0.1275
0.0404
0.0697
carbon fraction
Aloft
0.0069
0.7173
0.0725
0.0323
0.1710
(b) Emissions Rates and Mixing Heights
Species
Mixing Height
at the Beginning
Time
(CDT)
0500-0600
0600-0700
0700-0800
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400
VOC
(ppmC/hr)
0.0284
0.0384
0.1349
0.1559
0.2214
0.9076
0.5303
0.0615
0.0394
NOX
(ppm/hr)
0.1082
0.0306
0.0294
0.0278
0.0429
0.0628
0.0443
0.0087
0.0109
of Each Hour
(m)
188
164
190
350
600
890
1094
1430
1437
1332
85068
-------
TABLE 14. VOLATILE ORGANIC COMPOUND
(VOC) AND NITROGEN OXIDES (NOX)EMISSIONS
BY COUNTY FOR 1980.
VOC NOX
County (kg/day) (kg/day)
Bucks 94,730 51,846
Chester 48,522 29,231
Delaware 115,193 77,268
Montgomery 81,967 47,094
Philadelphia 156,169 100,084
Burlington 41,400 55,700
Camden 54,900 54,100
Gloucester 87,900 52,600
Mercer 43,100 66,400
Salem 32,300 38,100
New Castle 103,400 140,523
85068 «* 52
-------
TABLE 15. INITIAL CONDITIONS USED IN THE LEVEL III EKMA
CALCULATIONS FOR 13 JULY 1979
(a) Initial Conditions
Species
NOX
NMOC
Oo
Surface
0.0 ppm
0.0 ppmC
0.004 ppm
Aloft
0.0 ppm
0.0 ppmC
0.105 ppm
N02/NOX = 0.25.
Hydrocarbon reactivity: standard OZIPP conditions
Morning mixing height: 250 m at 0800 CDT.
Afternoon mixing height 1232 m at 1500 CDT.
NMOC/NOX at 0600 to 0900 CDT = 6.9.
Design 03 = 0.20 ppm at Roxborough (1400 CDT).
(b) Emissions Fractions
Time
(CDT)
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
NMOC
0.120
0.120
0.120
0.120
0.120
0.160
0.160
0.160
0.160
0.160
85068
-------
TABLE 16.
(a) Initial Conditions
Species
NOX
NMOC
Surface
0.0 ppm
0.0 ppmC
0.03 ppm
Aloft
0.0 ppm
0.0 ppmC
0.032 ppm
N02/NOX = 0.25.
Hydrocarbon reactivity: standard OZIPP conditions,
Morning mixing height: 250 m at 0800 COT.
Afternoon mixing height: 1132 m at 1400 COT.
NMOC/NOX at 0600 to 0900 CDT = 6.9.
Design 03 = 0.17 ppm at Roxborough (1400 CDT).
(b) Emissions Fractions
Time
(CDT)
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
NMOC
0.550
0.550
0.550
0.550
0.550
0.550
0.700
0.700
0.700
0.700
0.700
0.700
85068 *»
54
-------
SECTION 3
COMPARISON OF MODEL RESULTS
This section presents a comparison of the absolute predictions for
each of the models discussed in the previous sections. Results of the
four modelS--UAM; Systems Applications Trajectory; Level II OZIPM (with
CBM-II and CBM-III chemistry) and Level III EKMA (with CBM-II and Dodge
chemistry)—are compared for 13 and 19 July 1979 in Tables 17 and 18. In
these tables, HC stands for a hydrocarbon reduction; BKHC indicates a
reduced background concentration from 0.0576 ppmC or higher to 0.033 ppmC
hydrocarbons; and BK03 indicates a reduced background ozone from 0.06 ppm
to 0.04 ppm, on 19 July and from 0.08 ppm to 0.04 ppm on 13 July.
A wind field with considerable shear between the two layers simulated
within the mixed layer was used in the UAM simulations for 13 July day.
An effect of such shear combined with the rapid mixing occurring during
midday is to create considerable horizontal dispersion of emitted
pollutants. Another effect is to create a pseudoslowing of the trajectory
path determined from the vector average of the winds in these two lowest
layers of the UAM. Figure lOa shows the vector-averaged path created from
the paths defined by the winds in layers 1 and 2. Figures lOb and c show
the paths to the same final point defined by using either layer 1 or 2,
respectively. Since a thorough mixing occurs between these two layers on
a time scale of only a few minutes, any emissions occurring in the vast
area between the two paths shown in Figures lOb and c will influence the
air simulated to arrive at the final common point at 1300 hours. Hence,
the emissions along the path shown in Figure lOa are also considerably
dispersed away from the common point shown.
The agreement between the UAM and Systems Applications' Trajectory
Models based on emissions along the averaged path shown in Figure lOa is
apparently fortuitous. Table 17 shows that the base case versions agree
fairly well, but shows a consistent divergence between the models as the
severity of controls 1s simulated.
A brief Investigation into the effect of the precursor pathway of the
13 July day was conducted. The Trajectory Model simulates a moving column
of air with a horizontal size identical to one grid square of the UAM.
85068r3 7
55
-------
g
1— 1
g
ar
0.
^
O.
o£
LU
H-
<
3
X
o
or:
h-
'"I
— • *•
-D
CO
r— 1
O
^^
1— 1
z
Q-
_l
LU
O
^j
•~^
^•^
Q.
g
I-1-
E
Q.
QL
CO
1 |
— ^
CO
1 1 1
rr
1 1 1
O
O
i
LU
8
i>!
LU
-J
CO
Q)
_J
^
<
z:
LU
>— •
1—1
^"^
r-.
O)
>
OJ
_l
^
z:
LU
t— t
' '
'cv
a;
_j
N^
_
g
h-
o
LU
-3
"^
P^
_.
tO CM VO
i— 1 r- 1 r-l
o vo to o
• • • •
O^ VO ^f rH
rH r-l r-l t— t
CM r»>» co co co co rM to CM r^ ^t~ vo
f*>» rH f*^ ^^ tO CD VO CO tO O^ tO CM
rH r-l rH rH rH
C0r^«- CMrHCOVO tOCOrHO,
rH rH rH rH rH rH
rH 10
CT> lO
O
O CO CO tO tO O^ CO O O O^ O^ f"»- *^
CM rHrHrH rHrH rHrH
o o o co coo
a, z z z o o z
"O I/) CO CO CO CO CO CO
> °^3CZ "i° 000
CO 0) O CO
8. CM o
CO CO ^t CM
rH
CD VO CO rH
tO CD ^- CM
rH
LO VO CO CO
^f CD tO ^^
rH
CO CO CO
o o o
0 0 0
CO CO CO
CO • • •
oooo
• z z z
o
^ LO O LO
CO CM LO t^
o
4->
o
^
Q.
Q. •
VO d.
r^ a.
to
0 «5f
o •
o
o o
c <*-
0 I
**^ c ^^
4-> O CX.
*J 4J CO
tn to o
c. •
CT> 4-> O
t^ T! —
t- o o
o c t-
*J O «*-
• *r- O
4) C QJ
3 O -O C
r- E C 0
CD O O
r: c.
•o •»-> a> -o
d) -i£ CZ
o *o o
ex 3 J3
CV XJ o
EC_ II 3
h_ || ^
a» -o
z: <^ o 01
<: jo z L.
=> o ^
CO CO
O) CO O
XT r: •« 2^:
4-> 4-> C CO
O
V) V) -r- ••
f- -r- 4_> (/)
O C
O) > O
c o
CO ^ T3
O O fC .C
U
a; a; o o
«/> co c_ E
^O ^O ^5 QL
£> £i >> CX
^y cu oo
C C N CO
O O 0
O O Z O
* 4-C0")
56
-------
o
1—
t—t
E
*-
g
i—
CD
SE
»-H
Z
2E
0
UJ
1C
[I
^^
1
• J
en
«-H
z
o
^^
»— t
O-
_J
UJ
o
^c
_J
t— (
•JLk
fl-
oe:
o
U_
^"~*
E
JC
ex
ex
"*~^
CO
^~
— J
ID
CO
UJ
D£
UJ
Z
o
g
III
UJ
8
z;
CO
t— 1
UJ
CO
1—
•*-
^r
UJ
^^
HH
r—
CU
>•
0)
_J
UJ
CD
8
o
CO
1
^
CD
O
*
<
UJ
*"•*
^•4
1-4
r—
CU
>
cu
— 1
UJ
CD
8
o
CO
1
3E
CD
O
cu
CO
1
CD
O
CNJ
^
CD
O
^N»
g
1—
•-"
t> o o o o
cu cu o co o
V) *«***4 3= O 3=
jo O +J
(O •!-
4_> 4_> cu o o
T3 ^J O^ ^D
CU ^ C
•4-> 4-> <_> 3
u 3 JO
r— T3
CU T3 CU TJ
^ eu t. eu
o > o
E c- u 3
cu -o
2E co O CU
C CD
o
(^ (^ *^* • •*
•f— ^_ 1 * (/)
u c
CU CU 3 O
3 3 T3 JO
> o
c o
CU
-------
Hence the only emissions that affect the Trajectory Model are those along
the pathway within the horizontal area of one grid square. As Figures lOb
and c imply, the wind shear in the 13 July UAM simulations mixes emissions
from over a rather wide area.
Two simulations for 13 July were run using the Systems Applications
Trajectory Model along pathways parallel to the one shown in Figure lOa
but displaced one grid cell to either side. The center path result was
0.191 ppm ozone at 1300, as reported in Table 17; the two adjacent path-
ways produced similar and higher (near 0.25 ppm) results, respectively.
Hence the choice of pathway resulted in fortuitous agreement with the UAM
results. The choice of pathway can result in a fairly high diversity in
emissions and the UAM predictions contained a large horizontal mixing
component due to the wind shear.
In general, the UAM shows much less sensitivity to hydrocarbon con-
trols than do the Systems Applications Trajectory Model and Level II OZIPM
models used for the 13 July cases as presented in Table 19. The cause of
the lower sensitivity may be related to the wind shear because the dis-
crepancy shown for the 19 July cases (Table 20) is much less. Further-
more, there was good agreement in all other comparisons of trajectory
models matched to UAM input (e.g., Whitten and Hogo, 1981; Killus and
Whitten, 1983). The study of Killus and Whitten (1983) is especially
noteworthy because the UAM was modified to use a significantly different
chemical mechanism. Results of parallel control strategy simulations
using the UAM and Level II OZIPM models agreed very closely when the
chemistries were identical, even though the two chemical mechanisms pre-
dicted quite different responses to precursor controls. Hence, the wind
shear in the 13 July simulation appears to influence the control strategy
response. Unfortunately sufficient time and funding were not available
for this study to confirm and elucidate the effect of wind shear on con-
trol predictions. Until this effect is studied more completely, the UAM
results must be considered suspect since wind shear in the presence of
rapid vertical mixing over long periods of time is difficult to justify
physically and no data or physical evidence are available to support such
wind shear on 13 July in the Philadelphia area.
Tables 19 and 20 show a difference in Level II OZIPM results for the
CBM-II and CBM-III chemical mechanisms. Contract limitations did not
allow investigation of these differences; however, it is noteworthy that
on 19 July, the CBM-II mechanism is somewhat more sensitive to controls or
background changes, whereas on 13 July, the two chemical mechanisms show
similar sensitivity to hydrocarbon control.
Tables 17-20 also show predicted ozone and ozone reductions for the
Level III (city-specific) EKMA using either the CBM-III or the standard
85068r2 /
r 58
-------
or
o
t—
•— i
o
z:
o.
2!:
o.
or
LU
1—
!X
X
c
or
r~
^
L/">
C
or
uj
o
CO
LU
3
h-
LU
CO
c
c
c
LU
O
o
LU
or
»— i
"3Z.
£^
0 -It
E >-
LU __l
S
, •*
t __
^c i* >
LU ^^
or or
LU O
o>
^^
LU
_J
CO
•a:
•4-
^^
si
LU
i— i
*""*
f—
QJ
^^
.
or
C
•-. o
c »— i iiii iiii iiii
immm iiii 1111 iiii
t— « CVJ ^J-
iccr^cc iiii iiii iiii
«— t CO VC
1 ro •?" c^-1 1111 iiii 1111
t— i CVJ *tf-
cvccm cocLfic cvj^-o^in o^ftccr
comr- ^tvccc «-H«K*-VDCC cvj^-r>.cc
icvjcr>r^ cy>i-HVC«-H ^»>ccou1; incr. ^tcy1.
cri f^coccvj inoom
vvj^m f— i *± in ic cotnio cvjmr*«.r^
o o o co co co
=c 3: n: o o o
«/> CC CO CO CO CO CC OOO
OOOO OOO OOO O°S 5^ ^
s^1^ ^^^^ ^^i^i^ £S*vS
cocvjinr^ cocvjinr^ cocvjinr*^ cocvjinr^
o
E
cx
ex •
E
1C CX
1*^ CX
in
c **
o •
c
o o •
c u- 'oi
0 E >
•»- C CX O CX r-
4-9 4_> co o N
c c c
•^ 0< E
C. O O O>
O C C. t/>
• » ^ i* JM
^— ' ^^ f*« fv
• •»- o o
A) C QJ
3 O T3 C O;
f— E C O t«
f^7 •? ^i ^T3
> 0^ O O JD
^^ C 4->
4-> 4-» O 3
O «C «C O O)
f- -C 1. >
"C? ^U C^ *F*
Ol 3 X7 -* 4->
CX «O O XJ tt> X3
O > o O;
E t- M 3 to
Qj X? V)
o *t a.
CC CO X
4-> 4-> C CO to
0 C t/>
> 00
C O 3
o^ o o c
t/> tn L. E > CX l_
OJ O» CO CX
C C « CO
00 0 i—
Nl N O • f—
c o 3: c «t
*
59
J-
00
o
ID
00
-------
CC
o
1—
z
z
o
f—
CD
»— i
Z
*^
c
O
1—
<
CO
c
1— «
cc
^c
z
LU
O
CO
Z
LU
LU
3
1—
LU
CC
LU
Z
C
f^J
c
o
LU
J—
C_3
i— t
Q
LU
OH
O_
Z
t— i
Z
c
p
o *
:r> *
Q >•
UJ _J
cc ^3
*"3
i—
z o>
UJ fH
<_)
^** ^^
u» c
o- u.
•
0
CVJ
LU
_J
CO
<
t—
«C LU
z: CD
SX Q
LU O
0
»— i
o) *?
> 3C
a* cc
-J O
* LU
i
a* co
_J CJ
CO
z:
a* co
_J O
>•
CC
c
t—
»— i O
CT> CO
t-H ^" r^~
i r*^ co r*>
CVJ CO
i vo f*^ cr
«-H CVJ
i r- ^ oc
»— t CO ^~
1 1 1 1
1 ^^ O^ VO
I-H CVJ CO
^u
V)
o o o o
ac z ac
at
(^ ^^ ^^ ^^
3
*D
>
^j
a>
u
•^
^u
at
o.
,_
a*
T?
o
E
2T
to
"^"
at
3
at
to
u
a*
to
(Q
^3
at
c
o
ISI
O
^
• f—
c at
o >
f- at
*o
4_> at
to c
o
cr> M
c o
L. at
O tO
f— O
c
o at
E to
ai J3
p"
4-> O
^j
at i-
^^ *j
r— (Q
ai
L.
at T3
> a»
L. CO
at co
co at
JD C.
o ex
X
at a*
t** •
•*-> c a*
o c.
CO i- 0
C 3
Oi O T3
to ^ at
ro C. t_
O (O
O 4->
a* o c
to t. a>
(Q TUU ^j
J3 >» «.
.JT a*
at ex
C II
o •—
M O i—
o 3: «s:
+-CO1 •>
*
ao
a?
60
-------
Dodge chemistry. Two sets of results are presented in Tables 17 through
20 for each city-specific EKMA model (with CBM-III or Dodge chemistry).
The first set compares city-specific EKMA predictions based on ozone
design value (base case) similar to the UAM-predicted ozone. This set
allows comparison of the city-specific EKMA with more sophisticated models
even though the inputs differ. The second set of city-specific EKMA
results is based on the observed ozone at the monitoring station. This
set of results would have been obtained under regulatory analysis of the
two modeled days and provides a comparison of the UAM and city-specific
EKMA as regulatory tools. In these comparisons, the Level II OZIPM can
also be called city-specific EKMA in the sense that alternate procedures,
as outlined in the city-specific EKMA guidelines (EPA, 1981), are applied
for the two modeled days.
Tables 18 and 19 present absolute ozone predictions for the city-
specific EKMA compared to those of more sophisticated models for 13 July
and 19 July 1979. For the city-specific EKMA simulations using CBM-III,
the nominal default precursor background was assumed as transport condi-
tions (EPA, 1984). Since city-specific guidelines were applied for these
comparisons, no attempt was made to simulate the scenarios in which back-
ground conditions were altered along with the emission changes.
Tables 19 and 20 show the percent reduction in ozone predicted by the
city-specific EKMA models. From Tables 19 and 20 we see that the city-
specific EKMA is less sensitive to hydrocarbon reductions for both modeled
days compared to the UAM and trajectory models when the CBM-III is used.
When the Dodge chemistry is used in the city-specific EKMA, a greater
sensitivity to hydrocarbon controls is shown for both modeled days.
Estimates of the VOC reductions needed to lower peak ozone to the
level of the National Ambient Air Quality Standard (NAAQS) of 12 pphm are
shown in Table 21. For the July 13 day (considering no change in boundary
conditions), the UAM VOC estimate is 45 percent. The control estimates of
the Level II models are lower than those of the UAM (when there are no
changes in the boundary conditions) for both carbon-bond mechanisms. We
see from Table 21 that both CBM-2 and CBM-3 give similar control estimates
for the different sensitivity scenarios. The Level II models (CBM-2 and
CBM-3) predict lower control estimates for the case in which background 03
is reduced concurrent with emission reductions. For cases in which back-
ground HC is reduced concurrent with emission reductions and when both HC
and 0)3 are reduced concurrent with emission reductions, the Level II
models predict control estimates similar to those of the UAM. With
the Level III approach, the control estimates obtained with the Dodge
mechanism are slightly higher than those of the UAM (50 percent versus
45 percent), while the CBM-3 estimates are higher yet (70 percent and 69
percent versus 45 percent). Note also that there is little difference
85068r3 7
-------
between results obtained using the predicted UAM base case peak ozone or
the observed peak in the Level III analyses on this day because they are
very similar in value (19.3 pphm versus 20.0 pphm, as shown in Table 17).
The results are more mixed for July 19, 1979, however. The UAM estimate
is 25 percent. The Level II estimates differ significantly, depending
upon the mechanism that is used (24 percent and 42 percent for the CBM-2
and CBM-3, respectively). Further investigations of the Level II simula-
tions forQ19 July show that the chemistry of excited singlet state oxygen
atoms (0 ) formed from ozone photolysis (which is accounted for in CBM-3,
but not in CBM-2) can lead to very different control estimates. The O'D
chemistry can have a major effect when NOX concentrations are high during
the late afternoon hours of the simulation, as in the case of the 19 July
simulations. For the Level III analyses, the control estimate obtained
with the Dodge mechanism is slightly lower than the UAM estimate when the
same base ozone value is used (19 percent versus 25 percent), but slightly
higher when the observed ozone value is used as the base. The CBM-3
mechanism, when used in the Level III analysis, produces control estimates
that are higher than either the UAM estimates or the Level Ill/Dodge esti-
mates, regardless of the base case ozone.
A comparison of the model estimates presented in Table 21 provides
some insight into the results that might be obtained if the results were
used in a regulatory analysis. Several caveats are in order however.
First, since only two days were analyzed, it is very difficult to discern
any general patterns. The inclusion of more days is precluded by the lack
of UAM simulations. Second, the VOC reductions shown in Table 21 corre-
spond to peaks measured at single monitoring sites (i.e., the Roxy Water
Pump for 13 July and Downington for 19 July). The UAM estimates of VOC
reduction needed to lower peak ozone in the entire region to 12 pphm are
higher (Braverman and Haney, 1985). Also, the limited number of days
analyzed with the UAM make it difficult to assess definitively the differ-
ences that might actually occur in a true regulatory analysis. For
example, the State Implementation Plan (SIP) for ozone (DER, 1983) esti-
mated that 44 percent reduction in VOC would be required to achieve the
standard. This estimate was obtained from a detailed city-specific EKMA
analysis (using Dodge chemistry) of ozone data collected at 10 air quality
monitoring stations over the three-year period of 1979-1981. The design
ozone was 0.171 ppm at Trenton, New Jersey, on 24 June 1980. Therefore
the 13 July and 19 July 1979 days evaluated in this study might not corre-
spond to the final design used in regulatory planning.
85068r37
-------
LL.
O
CO
c
1— 1
t—
•
UJ _J
=5
O
co cr>
0 C
o z
o
>- CO
or »-H
•
f-H
CM
UJ
_J
CO
t—
f—
CD
>
CD
^
3 4->
••3 C
£
CT^ E
f-H C
fO
4_)
+j
**
r—
CD
>
CD
__l
^>
*3 C
E
CO C
CD
i
+_
^— '
»— i
•—
*CD
>
CD
"*
^
^H
»— c
,_
CD
>
^J
$
^-i
»— <
•— '
•j;
>
CD
— 1
+-
t—t
t—>
•— '
'CD
CD
4-
^"H
»— 1
r_
CD
CD
^J
UJ
cr
f1.
o
c
CO
i
cc
o
UJ
cr
c
c
c
CO
i
cc
o
SE *
d. CO
t— 1 1
NJ CD
C 0
^^^,
y $
o. CM
»-« i
rvj cc
C 0
f
**""
to
c
CO
f-H 1 1 |
CO
CO 1 1 1
cr i i i
f-H
O 1 1 1
CM 1 1 1
*"
*f 1 1 1
CM
IT) 1 1 1
CV
C 1 1 1
ID
& 1 I 1
O 1 1 1
in
O 1 II
CM •ij' F^. CT>
CO CM CM f-H
CM *3- CO 00
CO CM CM f-H
IT) CO CO C^
*J- CO CM f-H
O
x> -c ac jc
C C +J
3 C 3 C T3 f-
o o o o ex
t_ «f_ g ^^^ __
C- J^ l^ J^ W CO C
«o o-r- Of- OCD 3 f-
o jr°-c o JE o o 4->
jo o 4-> ac ^_> jo c o
f- f- O 3
c ex ex eo-o
f- f- f- f- a;
C/i 4*^ (/) ^^ (/) TC3 ^
CDC CCC CCC CC
CT> O Ot/>
OT3 303 303 3j*t/>
C T3 CXI XJCXJ XJO«f—
OO CDOCD
C
"*" 2!
V) V
C ' ' '
4J C
O *C
XJ • •
CD C
(. cy o
•f^» *P~
X> t- *J
3 XJ *J
O C (0
C- 3
cn o ^y*
H JO C
O i—
JE in O a<
C
CD CD i- O
jr jr CM
4-> 4-> O O
CD O CD
(/) C 4-> to
3 O
0 C CD
CD jo
3 «C
i- CD X5 X»
CD CD
o To c. o
C 4J CD i-
C V> X>
CD o jo a*
C- M • O C-
a» T- m o.
X t- CD c
o t- o n
«C JZ L. ct
z: «o xi rs
i^ Cy XJ CD
UJ JZ C (0 C
4-» 3 <0 O
C- 0 JO
O C JO XI
«*- O CD CD
r— 3 IO
o c 4-» ID jo
f- O C >
t- X> O CD
*O M CD 3
C CD f- C i—
CD t- C. O «C
o a> o M >
v> x jr o
CD
C CD CD C
O O JC to O
f- f- 4-> tJ M
4-> L. O O
O «U r—
3 C CD CD CD
XJ CD XJ to «/»
CD O O «C «*J
ct to E co cc
* * +-
*
J-
00
63
-------
SECTION 4
SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS
SUMMARY
EKMA was evaluated primarily by comparing the box model (OZIPM) that
forms its basis with more sophisticated models in applications to the
Philadelphia area. The study was carried out at several levels, beginning
with evaluation of OZIPM and ending with an evaluation of the control-
strategy predictions resulting from application of the city-specific EKMA
(OZIPM) based on EPA guidelines (EPA 1984). The OZIPM model was compared
with the UAM and Systems Applications Trajectory model, as well as with
modified versions of the original OZIPM model. The basic OZIPM model is a
simple moving-air-parcel, or box, model that can be used with any chemical
mechanism. The OZIPM model treats time-dependent, precursor-emission
factors and expansion of the air parcel; entrapment is treated by
assuming constant concentrations outside the parcel.
CONCLUSIONS
The primary emphasis of this comparison study was the identification
of features in the basic OZIPM model that could explain differences
between the results of the OZIPM model and those of other models. For the
most part, the study used UAM data files to generate appropriate OZIPM
inputs. The use of these files allows the comparison of models having a
common data base and comparison of EKMA simulations for regulatory
analysis: OZIPM can be run using either UAM inputs or by followinq EPA
guidelines (EPA, 1981, 1984).
This study and similar studies by Whitten and Hogo (1981), Whitten
et al. (1981) and Hogo et al. (1981) using data for Los Angeles, San Fran-
cisco, Sacramento, and Tulsa show that the OZIPM model can provide results
that are often quite similar to those of the more complex models. By
identifying specific differences in model formulation and analyzing inter-
mediate results, the formulation feature that produced the difference in
the results of two models could often be isolated. Even though the abso-
lute values predicted by the OZIPM model can differ significantly from
those of the more complex models (such differences can result from the
85068r3 5
64
-------
different chemical mechanisms employed), EKMA isopleths (which use OZIPM
results) typically yield control predictions that correspond with those of
the more complex models.
The main conclusions reached in the comparison of UAM and OZIPM
results in application to the Philadelphia area are similar to those of
previous studies (e.g., Whitten et al., 1981):
(1) These two models tend to show close agreement.
(2) Any box model (e.g., OZIPM) when run in conjunction with a grid
model (e.g., UAM), can serve as a useful and inexpensive tool
for understanding, analyzing, and evaluating important and
sensitive elements of the situation being modeled.
(3) The Systems Applications Trajectory Model serves as a source
of data for the EKMA type models, as well as an important con-
ceptual link between these models and the UAM.
Comparisons of the Systems Applications Trajectory Model and an EKMA-
type model (OZIPM/CBM) modified to most closely resemble the Systems
Applications Trajectory Model showed that many features in the Systems
Applications Trajectory Model can often be eliminated or simplified with a
minimal effect on results. For instance, the two levels within the mixed
layer of the Systems Applications trajectory model seem adequately repre-
sented by a single layer in the OZIPM/CBM model. The same photochemical
kinetic mechanism, the Carbon-Bond II Mechanism (Whitten et al., 1980),
is used in the UAM, the Systems Applications Trajectory Model, and the
OZIPM/CBM model. For comparison, the OZIPM/CBM was also run with CBM-III
chemistry (Killus and Whitten, 1984).
For the 1982 SIP studies, the EPA categorized degrees of model
sophistication by Levels I through IV (pages 65669-65670 of the 14 Novem-
ber 1979 Federal Register). The Level I models used in this study are the
UAM and the Systems Applications Trajectory Model. Versions of the OZIPM
box model used in this study that use trajectory paths based on UAM inputs
are considered to be Level II EKMA models. Level I and Level II models
are distinguished by the degree of complexity. Briefly, Level I models
use gridded data, whereas Level II models use data from only those grids
that the trajectory path passes through. A trajectory model can be
classified as either a Level I or Level II model based on its level of
complexity. The primary feature that defines a Level III EKMA model is
the assumed trajectory relating early-morning urban precursors to an
observed ozone peak (EPA, 1981).
ssoeer2 5
65
-------
Level III EKMA simulations that showed reasonable agreement with
observations were also performed. However, the use of county-wide emis-
sions 1n this model makes detailed comparisons with the other models used
1n this study difficult because their emissions data differ.
RECOMMENDATIONS
The following considerations are suggested for future studies:
For cases 1n which differences in the control strategy predictions
of various models are found to be related to differences in their
chemistry, 1t is important to investigate the chemical reactions that
produce these differences. Where appropriate smog chamber experi-
ments that would validate one or more types of chemical mechanisms
can be identified, these experiments should be performed.
The use of more complex models such as the UAM should be accompanied
by the use of EKMA, both for absolute trajectory simulations along
the path to the major ozone peaks, and to obtain control strategy
guidance from the EKMA isopleth diagrams. The studies performed to
date show that unless the wind fields are complicated by such factors
as wind shear, the two types of models normally give similar
results. In the absence of complex meteorological factors, important
differences in results would indicate the existence of problems 1n
one or both of the models that should be investigated.
Identification of background concentrations and reactivities must be
carefully considered for both present and future simulations. Severe
control requirements can be implied if the level of background pollu-
tants is overstated.
The regular city-specific algorithm might be modified to allow op-
tional starting times in place of the present 0800 LOT starting
time. Earlier starting times would emphasize dependence on an emis-
sions inventory, whereas later starting times might emphasize depend-
ence on initial conditions.
Modifications to the current version of the EKMA have also been sug-
gested in other studies. Although such suggestions have not been
addressed in this study, their Implementation might also be con-
sidered:
(1) Isopleths based on emissions Inventories might be further
developed to perhaps replace the present initial-condition-
based isopleths (Hogo, et al., 1981). Urban HC and NO
^
85068T3 5
66
-------
measurements are often not relevant to the initial trajec-
tory. Trajectories might originate prior to the present
0800 LOT time. Mathematically, path definition is not
important to moving-air-parcel models unless differences
in emission density and mixing height exist. Preliminary
calculations suggest that precise definition of mixing
height prior to 0800 is unimportant. Wind velocities prior
to 0800 are typically low, so that few grid squares would
be crossed before the arrival of the air parcel at the more
easily definable trajectory paths now used, i.e., those
beginning at 0800 hours. Therefore, an important factor in
the success of an emissions-dominated model would be mini-
mal (or at least moderate) diversity of the emissions
density in the vicinity of the present 0800 starting
area. In the present study, the path defined by the aver-
age of the winds in the first two layers passed over
significant emissions sources, but the paths defined by
winds in the surface and second layers were on either side
of the main emissions sources. Since models are often for
episodic conditions, the path must be forced over major
emissions if the early morning winds are uncertain and
significant emissions diversity exists.
An emissions-oriented model may also be potentially more
advantageous than the present initial-condition-and-mea-
sured-HC/NOy model because of the poor mixing that can
occur prior to 0800. Measured urban HC and NOX values
often vary widely, even when taken at the same location, as
a result of local mixing and local emissions. Use of the
HC/NOX ratio does not necessarily mitigate the problem
because the measured ratios can still be dominated by local
emissions and mixing effects. Starting the model earlier
and filling it with well-mixed emissions from the region
upwind of, and within, the main urban area may provide a
more realistic simulation of the air that arrives at the
downwind site where the ozone maximum occurs. Furthermore,
the ordinate and abscissa (which are scaled relative to the
emissions rather than to the initial concentrations) would
relate more closely to the purpose of the EKMA (i.e., to
estimate the changes in peak ozone concentrations resulting
from changes in emissions).
The use of any atmospheric model under conditions of wind shear with-
in the mixed layer should be approached with caution and extensive
confirmation. Although Level II box models are basically Inappro-
priate under such conditions, their use to assess a specific area can
ssoeer2 5
67
-------
still be helpful 1n Level I grid model studies. For example, by
using averaged wind data, a box model can be used to determine a
preliminary trajectory. The results from this trajectory can be
used to assess input data to the airshed.
A generalized concept can be associated with EKMA that goes beyond
its present specific formulation and uses; developmental work should
be continued to update its formulation and improve and broaden its
utility. Specifically, the following steps are suggested:
(1) Point source emissions might be handled so that their
interaction with all possible urban HC/NOX concentrations
could be evaluated as part of a new-source review proce-
dure.
(2) Percent cutback diagrams (PCD) might be added to the OZIPM
code so that both the isopleth and the PCD are generated.
A PCD is a diagram of ozone isolines, but the ordinate and
abscissa represent the percent control of NO or VOC
required to meet the 0.12 ppm 03 standard using the EKMA
procedure. A VOC/NOX ratio is required to generate a
PCD. Such a diagram graphically illustrates all possible
combinations of VOC and NOX control implied by the EKMA
procedure to reach the 0.12 ppm 03 line on a regular EKMA
ozone isopleth diagram.
(3) Statistical packages could be combined with EKMA to gener-
ate the probable number of exceedances resulting from vari-
ous control scenarios.
(4) Isopleths of many species other than ozone could be use-
ful. Some candidates are nitric acid, hydrogen peroxide,
and sulfate species relevant to acid deposition. Also,
toxic species (e.g., formaldehyde) can be studied using the
generic-model-and-isopleth approach.
(5) A "two-stage" version of EKMA should be further
developed. Local authorities can effectively control
only stationary sources, whereas mobile sources tend to
be controlled at the federal level. Hence, a first-stage
EKMA would vary only mobile emissions and the second stage
would vary only stationary emissions. This would give
local authorities a much clearer picture (isopleth) of
the effectiveness of stationary source controls. Also,
the difference in reactivity between mobile and stationary
sources could be more accurately modeled using the staged
EKMA.
68
85068r2 5
-------
REFERENCES
Allard, D., M. Chan, D. Marlin, and E. Stephens (1981), "Philadelphia
Oxidant Data Enhancement Study - Analysis and Interpretation of
Measured Data," EPA-450/4-81-011, U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina.
Braverman, T. W., and J. L. Haney (1985), "Evaluation and Application of
the Urban Airshed Model in the Philadelphia Air Quality Control
Region," EPA-450/4-85-003, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina.
DER (1983), Commonwealth of Pennsylvania "Supplement to 1982 Revision of
the State Implementation Plan for Ozone and Carbon Monoxide for the
Pennsylvania Portion of the Philadelphia Air Quality Control Region,"
Bureau of Air Quality Control, Department of Environmental Resources.
EPA (1981), "Guideline for Use of City-Specific EKMA in Preparing Ozone
SIPs," EPA-450/80-027, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina.
EPA (1982), "Emissions Inventories for Urban Airshed Model Application in
the Philadelphia AQCR," EPA-450/4-82-005, Research Triangle Park,
North Carolina.
EPA (1984), "Guideline for Using the Carbon-Bond Mechanism in City-
Specific EKMA," EPA-450/4-84-005, U.S. Environmental Protection
Agency, Research Triangle Park, Noth Carolina.
Hogo, H., G. Z. Whitten, and S. D. Reynolds (1981), "Application of the
Empirical Kinetic Modeling Approach to Urban Areas," EPA-450/4-81-
005b, Systems Applications, Inc., San Rafael, California.
Jones, K. H., et al. (1983), "The Rational and Need to Consider an
Alternative to EKMA, "Journal of the Air Pollution Control
Association, Vol. 33, No. 4, pp. 330-332.
85068^3 6 69
-------
Killus J. P., G. Z. Whitten (1981), "A New Carbon-Bond Mechanism For Air
Quality Simulation Modeling," SAI No. 81245, Systems Applications,
Inc., San Rafael, California.
Killus, J. P. (1982), "Background Reactivity Estimates for Atmospheric
Mode ing Studies," presented at the XV Informal Conference on Photo-
chemistry, June 27-July 1, 1982, Stanford, California.
Killus, J P., and G. Z. Whitten (1983), "Effects of Photochemical Kinetic
Mechanisms on Oxidant Model Predictions," Systems Applications, Inc
San Rafael, California (83039r).
Killus, J. P., and G. Z. Whitten (1984), "Technical Discussions Relating
to the Use of the Carbon-Bond Mechanism in OZIPM/EKMA," EPA-450/4-84-
009, U.S. Environmental Protection Agency, Research Triangle Park
North Carolina. *
Reynolds, S. 0., et al. (1979), "Photochemical Modeling of Transportation
Control Strategies-Vol. I. Model Development, Performance Evaluation,
and Strategy Assessment," prepared for the Federal Highway
Administration, Office of Research, Washington, D.C.
Whitten, G. Z., and H. Hogo, (1978), "User's Manual for Kinetics Model
and Ozone Isopleth Plotting Package," EPA-600/8-78-014a, Systems
Applications, Inc., San Rafael, California.
ph;J; Zr i'cP' ™}]^' 3nd "• Hog° (1980>* "^^9 of Simulated
Photochemical Smog with Kinetic Mechanisms," EPA-600/3-80-028a
Systems Applications, Inc., San Rafael, California.
Whitten, G. Z. H. Hogo, and R. G. Johnson (1981), "Application of the
o?Pnnl , 6tlC Mode11n9 Approach (EKMA) to Urban Areas," EPA-450/4-
81-005a, Systems Applications, Inc., San Rafael, California.
Whitten . G. Z.. and H. Hogo (1981), "Comparative Applications of the EKMA
in the Los Ange es Area," SAI No. 10R-EF80-73, Systems Applications,
Inc., San Rafael, California.
Whitten, G. Z., and H. Hogo (1985), "Control Strategy Sensitivity to
Specific Chemical Reactions," report in preparation.
85068p3 6
70
-------
APPENDIX
DESCRIPTION OF INPUTS USED IN URBAN
AIRSHED MODEL SIMULATIONS
The following description of the inputs used in the .Urban Airshed
Model (UAM) simulations are taken from Braverman and Haney (1985). As
discussed in the main volume of this report, many of the inputs derived
for less complicated models are based on the inputs developed for the UAM.
MODELING REGION SPECIFICATIONS
The modeling region of the Philadelphia airshed specified in this
study is 180 x 170 km, or a total area of 30,600 km2. This region covers
parts of Pennsylvania, Delaware, Maryland, and New Jersey and includes the
metropolitan areas of Philadelphia, Pennsylvania; Wilmington, Delaware;
and Trenton, New Jersey (see Figures 1 and 2).
The modeling region specifications (e.g., grid origin, grid size
number of horizontal cells, etc.) are contained in what is known as the
REGION packet. For the Philadelphia application, the modeling grid con-
tains 36 by 34 horizontal cells that are fixed at 5000 m by 5000 m, and 4
vertical cells (layers) that vary in thickness depending on the hourly
mixing height and the height of the top of the modeling region. There are
two vertical cells below and above the mixing height. During the simula-
tions, the cells, or layers, below the mixing height may attain the mini-
mum cell thickness specified (for this application, 50 m) during the night
whei the mixing height is at a minimum. As the mixed layer thickens dur-
inn the day, the vertical thickness of layers 1 and 2 grows, while the
th.ckness of the upper layers (3 and 4) decreases. The specified height
for the top of the modeling region is fixed for each hour and is used in
preparing the REGIONTOP file. The UAM REGION packet in all input files
was specified as follows:
UTM ZONE: 18
X ORIGIN: 387000 m EASTING
Y ORIGIN: 4340000 m NORTHING
GRID SIZE: 5000.0 m
esoeeb r 8
-------
TOP OF REGION: 1630.0 m
NX: 36 cells
NY: 34 cells
NZ: 4 cells
MINIMUM CELL THICKNESS-LOWER LAYERS: 50 m
MINIMUM CELL THICKNESS-UPPER LAYERS: 50 m
PREPARATION OF INITIAL AND BOUNDARY CONCENTRATION INPUTS
Three input files—AIRQUALITY, BOUNDARY, and TOPCONC—are required
for the specification of initial and boundary conditions of all pollu-
tants. The AIRQUALITY file specifies the concentrations of all species
for all grid squares at the start time of the run. The data available for
this file is constructed using monitoring data from 16 surface air quality
sites, 18 surface meteorological sites, and 3 upper-air meteorological
sites. In addition, for this application a helicopter was employed to
collect air quality and meteorological data above the surface in an
attempt to quantify pollutant transport aloft. Within the Philadelphia
urban area, special stations were also established for the collection of
hydrocarbon data. Two methodologies were used for constructing gridded
fields of initial conditions from these data—one for constructing the
gridded ground-level concentrations and one for extrapolating the concen-
trations aloft from the ground-level values.
To construct the initial and boundary concentration files, the Phila-
delphia air quality data base were examined to characterize the levels and
distribution of pollutants in the study area. This information was used
to identify pollutant concentrations and to establish methodologies for
preparing these inputs. The available NOX and ozone data collected at th*
16 surface air quality stations were generally adequate to characterize
the ground-level initial and boundary values for these pollutants. Inputs
to the simulation (e.g., wind fields, mixing heights) were prepared after
examining day-specific air quality and meteorological data collected from
the network of stations in the region. To simulate the carryover of
emission precursors on 13 July from the previous day, ozone readings at
the Vineland, New Jersey monitor along the southeast boundary of the
simulation grid were used to specify hourly ozone boundary conditions.
An early-morning (0000-0700 EST) nonmethane hydrocarbon (NMHC) con-
centration of 0.15 ppmC was specified for cells below the mixing height,
based on an analysis of data from other Northeast Urban Corridor cities
under similar meteorological conditions (stagnation and carryover). The
hydrocarbon species split of NMHC was chosen to represent an aged urban
air mass as recommended by Killus (1982). This value, along with those
85068b p 8
72
-------
for NO, $$2* an(* PAN (specified after examining late-afternoon downwind
observed data at Downington, Pennsylvania and Van Hi Seville, New Jersey
for days on which these sites were influenced by the Philadelphia urban
plume), were used as input to the EKMA/OZIPM trajectory model to obtain
the daytime variation of these species for use as hourly boundary condi-
tions along the southeast Inflow boundary. The OZIPM simulation predicted
ozone values similar to those recorded at the Vine!and monitor, thus giv-
ing credance to these boundary condition values.
On 19 July, observed NO and N02 data for the hour of 0500-0600 EST
were used as Input values for the northeast and east boundaries below the
mixing height for the hours of 0000-0600 (Braverman and Haney, 1985).
Since hydrocarbons were not measured at the Van Hiseville monitor, esti-
mates of the influx of total reactive hydrocarbons across the northeast
and east boundaries below the mixing height were specified by multiplying
the hourly NOX concentrations at the Van Hiseville monitor by the Phila-
delphia surface layer emission inventory hydrocarbon/NOx ratio of 6. The
total reactive hydrocarbon value was then speciated into Carbon-Bond
Mechanism components, using the carbon-bond fractions of the emission
inventory.
Preparation of Gridded Ground-Level and
Aloft Initial Concentration Fields
Initial conditions for all species were specified using all available
monitoring data in the Philadelphia region. The simulation of 13 July
commenced at midnight, requiring concentration values corresponding to
this time. Rather than using the hourly averaged observed value for each
species for the midnight hour, initial condition concentrations were cal-
culated for each monitor by averaging the observed hourly average pollu-
tant concentrations from 2300-2400 EST on 12 July with the 0000-0100 EST
concentration on 13 July to obtain a two-hour averge value for midnight.
A Poisson interpolation routine was used to construct the gridded surface
fields.
Sixteen surface air quality stations were used to estimate NO, N02,
and 03 ground-level concentrations. In past modeling applications it was
sometimes necessary to add stations 1n the corners of the modeling region
and to give these stations background values that reflect the rural nature
of the outlying portions of the grid. However, in this study, initial
conditions for surface grid cells without monitors were obtained by
employing a Poisson Interpolation, which extrapolates these values using
data collected 1n the more source-Intensive portions of the grid. After
the surface field was computed, a vertical Interpolation method was chosen
1n which the background concentration at the top of the modeling region
85068b r 8
-------
(TOPCONC) was used 1n levels 3 and 4, and the level 2 value was obtained
by a linear interpolation between the surface value and the level 3
value. Using this method, all grid cells in all levels were initialized
with appropriate concentrations for all species.
The amount of spatial variation in the early-morning NO and 0? moni-
toring data differed from day to day. For 13 July, the NO data ranged
from 0.003 ppm at Summit Bridge to 0.127 ppm at the downtown Camden
site. The Camden observations exhibited the highest early-morning NO
concentrations. The NOX measurements for 19 July were lower, with a range
of 0.006 to 0.080 ppm.
PREPARATION OF METEOROLOGICAL INPUTS
The meteorological inputs to the UAM determine the transport charac-
teristics and rate of mixing of the pollutant cloud. Meteorological fac-
tors often account for violation of the NAAQS on some days and not on
others, even when the emissions patterns are similar. Therefore, the wind
field and the hourly mixing depths are important UAM inputs. Since upper
air meteorological data were collected in Philadelphia only on a limited
number of special study days, the three-dimensional structure of the wind
field was one of the most difficult inputs to specify.
Calculation of The Wind Field
Similar to the construction of the initial conditions, the three-
dimensional wind field was calculated in two steps: The surface wind
field was prepared first, followed by the upper-level wind fields. The
surface wind field was developed using the ground-level measurements and
synthesized data, and an interpolation routine that weights the observa-
tions by the inverse of the distance from the grid square. Mathemati-
cally, this relationship can be written:
Vn-
z
i=l
85068b r 8 74
-------
v -
1=1 V1rik
k n _ -1
1=1
where uk and v^ are the calculated horizontal velocity components for grid
cell k; u^ and v^ are the measured velocity components at station 1; n is
the number of wind measurements; and r^ is the distance between monitor-
Ing site i and grid cell k. A smoothing algorithm, which replaces each of
these velocity components by the average velocity component calculated
over the block of nine surrounding cells, was used to smooth the calcula-
ted wind field to eliminate any discontinuities introduced by the interpo-
lation procedure. Finally, the average measured wind speed and the aver-
age computed wind speed was calculated over the modeling region resulting
from the preceding calculations. The wind speed in each grid cell was
then scaled by the ratio of the average measured speed to the average
computed speed because the vector-averaging processes embedded in the
interpolation and smoothing steps yield wind speeds lower than the obser-
vations.
Observed surface wind data were available at the following locations:
Station
Northeast Philadelphia Airport, Pennsylvania
Philadelphia International Airport, Pennsylvania
McQuire Air Force Base, New Jersey
Willow Grove Naval Air Station, Pennsylvania
Lakehurst Naval Air Station, New Jersey
Trenton-Mercer County Airport, New Jersey
Killville Airport, New Jersey
Greater Wilmington Airport, Delaware
Chester, Pennsylvania
Bristol, Pennsylvania
Allegheny-Philadelphia, Pennsylvania
Summit Bridge, Delaware
Downington, Pennsylvania
Lumberton, New Jersey
Van Hi Seville, New Jersey
Norristown, Pennsylvania
NW
NE
SE
SW
UTM Easting
(km)
499.0
478.6
534.1
488.7
556.9
515.6
494,3
448.6
467.8
510.0
491.7
441.0
436.0
518.0
559.0
473.5
430.0
550.0
535.0
410.0
UTM Northing
(km)
4436.0
4414.6
4431.3
4449.8
4431.5
4459.0
4357.0
4390.7
4409.3
4439.5
4425.2
4376.0
4426.0
4423.0
4439.5
4440.0
4480.0
4490.0
4380.0
4360.0
ssoesb r 8
75
-------
..
To obtain wind vectors for levels 2 and 1 fnr
r^s s;r,'~ • •
level 4 vector for each arid r.ii ?H 9 SUrfaCe vector and tne
V2 s (V4 ' V1)*(NODE2/1500)
V3 s (V4 * V1)*(NODE3/1500)
85068b r 8
-------
where
NODE2 = the height of the node of level 2,
NODE3 = the height of the node of level 3,
V^ = the speed of the vector for level 4,
Vj = the speed of the vector of the surface wind, and
a = 0.2
This method is only one means of interpolating for the winds in levels
2 and 3. In this equation, a is the surface friction coefficient that
varies as underlying features on the surface vary. If the a coefficient
had been dropped, a straightforward linear interpolation would have
resulted from the equation; however, the magnitude of the winds in levels
2 and 3 would have been underestimated with this procedure. Only one
value of a was used in this application to reflect an average value of
surface friction throughout the region. A nonlinear interpolation could
also have been used by solving for a in each grid cell and using the
resulting a in the interpolation; however, this method was not used in
this application.
A full three-dimensional flow field was thus created for each hour.
Either because of noise in the wind sampling network, or because of the
nature of the vectors sampled at two distinct points, the resulting flow
field contains a great deal of divergence. If the field is employed by
the Urban Airshed Model with the divergence left in, spurious artificial
vertical motions will result in unrealistic vertical transport. The last
phase of the wind field preparation involves the elimination of nearly all
divergence.
The program DIVFREE was written to read the three-dimensional file
and eliminate most of the divergence in the flow field. This program was
adopted from an EPA algorithm developed by Clark and Eskridge (1977) and
is the final step in the creation of the wind field. As a quality control
check, the wind vectors are then plotted on the airshed grid and examined
for reasonableness before use in the Urban Airshed Model simulation.
Trajectory analysis can be performed using this final file to determine
the origin both in space and time of the parcels affecting the peak ozone
concentrations.
Because of the availability of upper-air data within the airshed
region on 19 July, the procedure used for specifying winds in levels 2, 3,
and 4 differed from that used for the upper-level winds for 13 July.
85068b r 8
77
-------
Radiosonde or pibal wind measurements were available for three sites:
Wilmington, Delaware, Downtown Philadelphia, and Trenton, New Jersey.
These data were plotted graphically for review purposes. Figures A-l,
A-2, and A-3 present measured winds aloft up to 3250 m above Wilmington,
Downtown Philadelphia, and Trenton, respectively for 1200 EST on 19
July. The data plotted in these figures show a light surface flow (2 to 4
m/s) in a general southeasterly direction, while the wind speeds aloft
(1250 to 3000 m) were moderate with a general westerly direction. All
available upper-level wind data were plotted in this manner and utilized
to estimate specific hourly inputs for each level. After examining the
data, a constant average wind vector was obtained for each hour for levels
2, 3, and 4.
Mixing heights for 13 July 1979 were estimated every half hour for
urban and rural areas for daytime, transitional, and nighttime regimes.
Urban and rural mixing height cell designations are included in the dis-
cussion of the terrain file, page A15. Since no soundings were available
for Philadelphia on this day, the daytime mixing heights between 0600 and
1400 EST were estimated using the available morning (0700 EST) and evening
(1900 EST) radiosonde soundings at JFK and Dulles airports. The procedure
used to estimate daytime mixing heights is similar to the methodology used
by Holzworth (1972). First, spatially averaged surface temperatures for
the Philadelphia region were computed for each hour using all available
meteorological measurement sites. The resultant average hourly surface
temperatures were plotted on temperature-versus-height graphs of each
sounding. The height of the intersection between the dry adiabat of the
surface temperature and the temperature sounding was then taken as an
estimate of the mixing height.
Between 0600 and 1200 EST, the mixing height was computed as the
average of the values estimated from the JFK and Dulles airport morning
soundings. An average of the JFK and Dull as morning and afternoon sound-
ings was used to compute mixing heights between 1200 and 1400 EST. An
exception to this procedure was made for estimating rural mixing heights
between 0600 and 0900 EST, before the dissipation of the rural surface
stable layer. During this period rural mixing heights were subjectively
estimated to increase from the overnight value of 100 m to the 0900 EST
value. Because of the general nature of the "adiabatic" procedure used
for estimating daytime mixing heights, no other distinction was made
between urban and rural mixing heights during the 0600-1400 EST time
period.
The overnight (0000 through 0600 EST) mechanically dominated mixing
heights were set to a constant value of 250 m in urban areas and 100 m
in rural areas. These "default" values were obtained from analyses of
esoesb r 8
78
-------
£>£L TffH = 250 METERS
PL'L TffS = 2 M/S
FIGURE A-l. Pibal sounding at Wilmington, Delaware on 19 July
1979, 1200 EST.
79
-------
DEL TRH = 250 METERS
DEL TRS = 2 M/S
2 pibal sounding at Philadelphia, Pennsylvania on 19 July
1 -7 / .7 j 1 i. D U t J I .
80
-------
QELTffH = 250 METERS
VELTRS = 2 M/S
FIGURE A-3. Pibal sounding at Trenton, New Jersey on 19 July
1979, 1300 EST.
81
-------
overnight mixing heights in Philadelphia, St. Louis, and other cities
(Godowitch,-1984; Godowitch, et al., 1984b; Bornstein, 1968; Clarke,
1969).
Mixing heights during the evening transition between convectively
dominated and mechanically dominated mixing are difficult to estimate,
particularly in the absence of local measurements. In order to approx-
imate this transition so as to avoid sharp discontinuities in mixing
height in a manner consistent with the capabilities of the Airshed Model,
mixing heights were decreased at a rate of ~2-3 m/min after the time of
maximum average surface temperature at 1400 EST (Noonkester, 1976; Kaimal,
et al., 1982). This rate was used for urban mixing heights until 2300 EST
when the overnight value of 250 m was assumed to be applicable. In rural
areas, this rate was applied until 1800 EST, at which time the mixing
height was decreased more rapidly to reach the overnight value of 100 m by
2000 EST (Godowitch, 1984b).
For 19 July the hourly averaged mixing height values were estimated
using available radiosonde observations and sodar data for the Philadel-
phia area. The radiosondes were released in downtown Philadelphia; the
sodar was located at Summit Bridge, Delaware. Both urban and rural mixing
heights were specified on the half hour beginning at midnight. The fol-
lowing procedures were used in specifying the mixing heights used in the
simulation:
Urban mixing heights from 0500 through 1500 EST were estimated using
vertical potential temperature profiles obtained from Philadelphia
upper-air soundings. For each hour, the mixing height was considered
to be at the base of the layer in which the potential temperature
increased rapidly with height. Upper-air sounding data were avail-
able from the urban site for 0500, 0714, 0937, 1150, and 1500 EST.
For hours between soundings, potential temperature profiles were
obtained by interpolation;
Urban mixing heights between 0000 and 0500 EST were set at the valje
estimated from the 0500 EST sounding;
The data indicate that the maximum mixing height was reached at 1200
EST. Urban mixing heights following the time of the 1500 EST sound-
ing were determined by decreasing the mixing height at a rate of ~2-3
m/min (Noonkester, 1976; Kaimal, et al., 1982) to reach the 250-m
overnight default value by 2100 EST.
85068b r 8
82
-------
Other Meteorological Inputs
Six meteorological scalars that vary 1n time only are required as
inputs to the Urban Airshed Model. Information needed to calculate each
scalar is contained in the routinely collected meteorological data. In
this study, the temperature gradient above and below the Inversion was
calculated from the rawinsonde data. A value of 24000 ppm was specified
for the concentration of water for both days. The atmospheric pressure
was set to 1.0 atm for both days.
The exposure class categories chosen for each hour were specified
after reviewing the available meteorological data including surface tem-
peratures, solar radiation data, and synoptic weather summaries. As
detailed in Braverman and Haney (1985), the exposure class is a measure of
near ground-level stability due to surface heating or cooling and can be
estimated from insolation as follows:
3 , strong )
2 , moderates daytime insolation
1 , slight )
exposure =
class
4
0 , heavy overcast day or night
, > -g- cloud cover )
3 / nighttime cloudiness.
-2 , < -g- cloud cover)
The exposure class categories chosen for each hour were specified
after reviewing the available meteorological data Including surface tem-
peratures, solar radiation data, and synoptic weather summaries for both
days.
A computer program developed by Schere and Demerjian (1977) was used
to calculate layer-averaged N02 photolysis rate constants based on month,
day, year, latitude, longitude, time zone, time of day, mixing height, and
measured solar radiation data. The first seven parameters were used to
calculate zenith angles and corresponding clear sky theoretical surface
N02 photolysis rate constants. The mixing heights, along with the
parameters used to calculate clear sky theoretical surface N02 photolysis
rate constants, were used 1n the calculation of clear sky theoretical
layer-averaged N02 photolysis rate constants. The methodology used to
determine the theoretical surface and layer-averaged N02 photolysis rate
constants 1s described by Demerjian et al. (1980). The total measured
esoesb r 8
83
-------
solar radiation data 1n langley/min was multiplied by the constant 0.40
mln"1 min/langley derived by Jeffries et al. (1982) to convert the solar
radiation data to empirical surface N02 photolysis rate constants (min'1)
The empirical layer-averaged N02 photolysis rate constant was calculated
by the following equation:
Empirical layer-
averaged N02 photolysis
rate constant
aLe constant
Clear sky
theoroetical surface
NO photolysis
rate constant
Clear sky
theoretical layer
averaged NO
Photo1ysis
rate C0nstant
A complete set of inputs used in the METSCALARS file for 13 July and
19 July is contained in Tables A-l and A-?.
EMISSIONS INPUTS
The gridded minor point, area, mobile, and elevated point source
emission inventory was prepared for the EPA in 1981 by Engineering
Science, Inc., (EPA, 1982). (The boundary of the gridded inventory Is
shown in Figure A-4.) The bold line defines the area for which gridded
emission estimates were available for the 1979 inventory. Hourly (local
daylight time) emission values for total NOX and total hydrocarbon are
presented in Table A-3. Total daily emissions by source type are pre-
sented in Table A-4 for NO, N02, and hydrocarbons. This table also give?,
the average hydrocarbon splits for the entire emissions inventory. The
gridded spatial distribution of ground-level hydrocarbon emissions is
presented in Figure A-5. The gridded spatial distribution of ground-levpl
NOX emissions is presented in Figure A-6.
TERRAIN
The terrain of the modeling region was classified according to land-
use values estimated while compiling the emission inventory (EPA, 1982).
For grid squares not contained in the emission inventory region, land-use
classifications were obtained from a United States Geological Survey
(USGS) map with a 1:250,000 scale. Each land-use category was assigned
a surface roughness value and an estimate of surface uptake velocity
(vegetation factor) according to Wesely (1983) (Table A-5). Land-use
classification by grid cells for the Philadelphia modeling region is pre-
sented in Figure A-7.
85068b r 8
84
-------
z
o
£T
^c
i
^
_
co
o
UJ
1C
CO
CXL
»— i
**
Ti
•
>_
_J
""^
co
«— i
UJ
31
1—
o:
O
u_
UJ
_J
1—4
U-
o:
DL
O Q.
O
0 C
o
L. .t-
O> 4->
^T3 L»
3 4->
to o:
•i- O
«/> h-
>> o
'o u_
^ o
0
Z (O
O£
QJ
t- (^
3 IO
in
O
o. CO ^^ ^t" CV CO CO t** CTv PO CO UD ^^ tf>
CO VO tO VO CD •" * ^" ID ^^ ^* C7^ VD Ot CM t-H
1 1 IOOOOOOOOOOOOOOO 1 1 1 1
O O O O '•"' «"•' •"•* CM CVJ CM CO CO CO CO CO CM CM »-H i— 1 O O O
oooooooooooooooooooooo
oooooooooooooooooooooo
oooooooooooooooooooooo
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
VOVDVOIOVOIOIOVOVOIOIDVOVOVOIOVOVOVOVOVOIOVO
oooooooooooooooooooooo
oooooooooooooooooooooo
oooooooooooooooooooooo
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
oooooooooooooooooooooo
oooooooooooooooooooooo
C\J CO ^^ if) ^O r^ 00 O1^ C3 t~< C\l CO *^ IO ^O f*^ CO O^ CO »™H CO ^f
OOOOOO^O»— it— it— it— it— it-Ht— it— if— «f— iCMCMCMCM
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
oooooooooooooooooooooo
oooooooooooooooooooooo
^^ C^ CO ^^ ^O ^O ^^ CO C^ <^^ ^H ^^ CO ^^ 1^^ 4O ^^ GO ^^ ^^ ^^J CO
OOOOOOOOOf-H«-Hi-Ht— l«-Hi-^^-li-Hf— If— ICVJCVJCVJ
JC
c
85
-------
01
O *•—*
CM
O
oooooooooooo
oooooooooooo
Ol
£_
3
to
O
Q.
X
O.
Q.
00
or
O
oo
I—
Z3
Q.
oe:
«c
_i
<0
or
Q.
to to
o to
o o o o o o o o' o* o* o" o o* o o" o* o o" o" o* o'
ooc
o o o o o
O O O O O ,-t rH
o o
QQ
0»
I >
c o
O) ^
u <:
00
CD C= O O CD O
O O
O <= O O O O O C=
O O
???????
c=
o o
c o
o o
88888888888880000
•—I CM CO ^ Lf)
OOOOOOOOO^i^^;^;
•'••'Illllll^
88888880000
O t—I CM CO ^f IT) VO
o o o o o o o
0008888888
vor^oocr>o«-ieMco^-
'-«'-«•—««—«CMCMCMCMCM
OOOOOOOQOOOOO
o o o o o
cr> o i—i CM co
«—I CM CM CM CM
°
86
-------
0
NORTH
10
T
20
i i i i i i i i
30
+
H
10 20 30 40 50
KILOMETERS
30
20
-------
TABLE A-3. HOURLY EMISSIONS OF NOX AND HYDROCARBON (TONS/HR)
USED IN THE 1979 PHILADELPHIA EMISSION INVENTORY
Hour
0000
0100
0200
0300
0400
0500
0600
0700
0800
0900
1000
1100
1200
1300
1400
1500
1600
1700
1800
1900
2000
2100
2200
2300
Total
Total
(LOT)
- 0100
- 0200
- 0300
- 0400
- 0500
- 0600
- 0700
- 0800
- 0900
- 1000
- 1100
- 1200
- 1300
- 1400
- 1500
- 1600
- 1700
- 1800
- 1900
- 2000
- 2100
- 2200
- 2300
- 2400
(tons/day)
(tons/year)
Total
NOX
23.413
21.070
19.886
19.712
20.461
24.625
33.658
53.234
53.623
49.794
50.629
51.720
46.660
50.451
51.662
54.883
58.815
52.584
42.650
35.866
32.448
30.728
29.480
26.664
934.718
341,172.000
%
2.50
2.25
2.13
2.11
2.19
2.63
3.60
5.70
5.74
5.33
5.42
5.53
4.99
5.40
5.53
5.87
6.29
5.63
4.56
3.84
3.47
3.29
3.15
2.85
100.0
Total
HC
17.255
15.090
12.413
11.777
11.444
14.422
25.677
81.214
94.468
89.978
89.809
84.016
77.34
75.132
75.917
81.869
74.434
67.097
56.362
39.228
35.537
33.102
31.363
22.956
1217.897
444,532.000
%
1.42
1.24
1.02
0.97
0.94
1.18
2.11
6.67
7.76
7.39
7.37
6.90
6.35
6.17
6.23
6.72
6.11
5.51
4.63
3.22
2.92
2.72
2.58
1.88
100.0
85068b 8
88
-------
cc
£
UJ
z
*"*
^
•-H
OL
ii i
O
_J
**"*
o_
en
en
t-H
UJ
3Z
r—
Z
•— 1
^~*
•5
cV>
^"•^
UJ
0-
^~
H~
UJ
o
ct:
~^
o
CO
CO
CO
z
o
CO
CO
I-H
'JE
LU
__l
•"•*
cC
o
_J
o
•a-
i
«c
UJ
_J
CO
o
3
o
CO
o en co vo
vo en co CM
en ^**> co en
en o «-H in
co t-H en en
*-4 in *t ^
in O in r-.
vo co m en
r-. co in vo
O CO CT> f»»
r>- «— « vo co
in co vo co
•« •»
^H f"H
co en co CM
CM CM en in
co r^ vo en
««•««•
t-H vo t-H in
o ^- «a- <*>
en co co vo
in vo t-H
t-H CM t-H
CM CM t-H CO
en co P"*- in
r^ vo 'sf in
P">» CO «5fr ^^
co en vo
t-H ^H VO
*s- ^- co in
^j- co in in
t-H CM CM CO
CM in in r>.
t-H co en CM
CO CM CO
VO f"- CO VO
co co o en
O CO CO ^f
^»» «^ r*» co
in t-H o co
CM •-! CM r«*
in en r>. co
in r»> en co
en vo o vo
•IMAM
co f-H in vo
CO ^" O ^"
r^ o r^ vo
in CM CM vo
C flJ
*— tl
*^^ \r
O 4-* QJ L.
a. c o 3
i- l_ O
•O O 3 CO
O C
O 4-> O
O ^ 4-> ^H
CM d O C
in co nj o
« t_ JO
t-H C U- C.
O TO
j3t3OOCOt—tCMCn
CM O «OOV)^l-CM^-COin
CM I-H o co ) J3
• • 3C t.
t-H CO t- CO
r^ o or
*-H »— O.
^_
-------
en
LU en
e_> en
Qi —
ID 31
O LU
en
o
ex o
LU oc
ct
cr _i
ex
G/DRY
CB
oa
IS
eg
1SU3
(S
00
IS
C\J
0000
ca
C3 ^3
^3 ^3
^3 ^a
(M LT)
I I
IS
O IS S>
S IS S
S S IS
<\i in ~
(S
Q
CS
(S
T3
OJ
-C
to
i.
•r~
rtJ
(T3
•r~
.C
Q.
"oi
no
O
Q.
OJ
SI
CO
SI
C\J
to
t/1
•r—
OJ
c
o
jQ
S.
o
CO
-------
85 5
fSJ ^
JC
m
(\J
I I I I I I I I I I
SI
00
cs
ct
o
I I I I I I I I I I I I I I
(S IS
m CM
(ill
cr
a;
a>
CO
t.
to
(O
«o
a.
OJ
to
c
o
•r-
co
to
O
CD
J-
OJ
10
O)
u
to
M-
$-
=3
to
vo
LJJ en
O J-
1-1 O
1S3M
91
-------
TABLE A-5. SURFACE ROUGHNESS AND VEGETATION FACTOR VALUES
Surface Surface Uptake
Roughness Resistance Velocity Vegetation
Land-Use Category (m) (s/cm) (cm/s) Factor
Mixed Grassland and Cropland
Deciduous Forest
Coniferous Forest
Urban Area
Ocean Water
0.10
1.0
1.0
1.0
0.001
1.0
0.6
1.5
3.0
20
1.0
1.7
0.7
0.3
0.05
1.0
1.7
0.7
0.3
0.05
85068D 8
92
-------
Legend:
G = Mixed Grassland
ana cropland
D = Deciduous Forest
C = Coniferous forest
U = Urban Area
N = Ocean Water
10
I I I I | I I
20
30
4500
4460
446B
4440
4420
4400
43E<0
436B
4341
407 427 447 467 487
Hosting (km)
507
527
547
567
FIGURE A-7. Land-use classifications for the Philadelphia airshed
modeling region.
93
-------
REFERENCES
Ames, J., T. C. Meyers, L. E. Reid, D. C. Whitney, S. H. Golding, S. R.
Hayes, and S. D. Reynolds (1978), "SAI Airshed Model Operations
Manuals Volume I—User's Manual and Volume II—Systems Manual," EPA-
600/8-85/007 a,b, the U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina.
Braverman, T. W., and J. L. Haney (1985). "Evaluation and Application of
the Urban Airshed Model in the Philadelphia Air Quality Control
Region," EPA-450/4-85-003, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina.
Bornstein, R. D. (1968), "Observations of the Urban Heat Island Effect
in New York City," J. Appl. Meteor.. Vol. 7, pp. 575-582.
Clark, T. L., and R. E. Eskridge (1977), "Non-Divergent Wind Analysis
Algorithm for the St. Louis RAPS Network," Environmental Sciences
Research Laboratory Report EPA-600/4-77-049, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina.
Clarke, J. F. (1969), "Nocturnal Urban Boundary Layer over Cincinnati,
Ohio," Mon. Wea. Rev.. Vol. 97, No. 8, pp. 582-589.
Demerjian, K. L., K. L. Schere, and J. T. Peterson, (1980), "Theoretical
Estimates of Actinic (Spherically Integrated) Flux and Photolytic
Rate Constants of Atmospheric Species in the Lower Troposphere," in
Advances in Environmental Science and Technology. Vol. 10, pp. 369-
459, J. Pitts and R. Metcalf, eds., John Wiley & Sons, New York. New
York.
[EPA] U.S. Environmental Protection Agency, (1982), "Emissions Inventories
for Urban Airshed Model Application in the Philadelphia AQCR," 450/4-
82-005, Research Triangle Park, North Carolina.
Godowitch, J. M. (1984), personal communication.
85068 9
-------
Godowitch, J. M., J.K.S. Ching, and J. F. Clarke (1984a), "Spatial
Variation of the Evolution and Structure of the Urban Boundary
Layer," submitted for publication to Bound. Layer Meteor.
Godowitch, J. M., J.K.S. Ching, and J. F. Clarke (1984b), "Evolution of
the Nocturnal Inversion Layer at an Urban and a Nonurban Location,"
submitted for publication to J. Climate and Appl. Meteor.
Holzworth, G. C. (1972), "Mixing Heights, Wind Speeds, and Potential for
Urban Air Pollution Throughout the Contiguous United States," Office
of Air Programs Publication No. AP-101, U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina.
Jeffries, H. E., R. M. Kamens, K. G. Sexton, and A. A. Gerhardt (1982),
"Outdoor Smog Chamber Experiments to Test Photochemical Models," EPA
Cooperative Agreement No. 805843, University of North Carolina,
Chapel Hill, North Carolina.
Kaimal, J. C., N. L. Abshire, R. B. Chadwick, M. T. Decker, W. H. Hooke,
R. A. Kropfli, W. D. Neff, F. Pasqualucci, and P. H. Hildebrand
(1982), "Estimating the Depth of the Daytime Convective Boundary
Layer," J. Appl. Meteorol.. Vol. 21, pp. 1123-1129.
Killus, J. P. (1982), "Background Reactivity Estimates for Atmospheric
Modeling Studies," presented at the XV Informal Conference on
Photochemistry, June 27-July 1, 1982, Stanford, California.
Noonkester, V. R. (1976), The Evolution of the Clear Air Convective Layer
Revealed by Surface Based Remote Sensors," J. Appl. Meteorol., Vol.
15, pp. 594-606.
Schere, K. L. and K. L. Demerjian (1977), "Calculation of Selected
Photolytic Rate Constants Over a Diurnal Range," EPA-600/4-77-015,
U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina.
Reid, L. E., S. D. Reynolds, and D. A. Latimer (1980), "Evaluationn of
Airshed Model Performance in Tulsa," Technical Memorandum No. 3, SAI
No. 223-ES80-164, Systems Applications, Inc., San Rafael, California.
Wesely, M. L. (1983), "Turbulent Transport of Ozone to Surfaces Common
1n the Eastern Half of the United States," In Trace Atmospheric
Constituents: Properties, Transformations, and Rates, p. 345-370,
Advances in Environmental Science and Technology, Vol. 12, S. E.
Schwartz, ed. (H. Wiley-Interscience Publication).
85068 9
-------
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing/
1. REPORT NO.
EPA-450/4-81-005C
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Application Of The Empirical Kinetic Modeling
Approach To Urban Areas
Volume III: Philadelphia
5. REPORT DATE
October 1985
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
G. Z. Whitten, H. Hogo, N. M. Yonkow, R. G. Johnson
and T. C. Meyers
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Systems Applications, Inc.
101 Lucas Valley Road
San Rafael, California 94903
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
Contract No. 68-02-3376
12 SPONSORING AGENCY NAME AND ADDRESS
Air Management Technology Branch
Office Of Air Quality Planning & Standards
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
EPA Project Officer: Gerald L. Gipson
The Empirical Kinetic Modeling Approach (EKMA) was evaluated using the results of
model applications in the Philadelphia area, primarily by comparing results of the box
model (OZIPM) that forms the basis of the EKMA with those of more sophisticated models.
The study was carried out at several levels, beginning with evaluation of OZIPM and
ending with an evaluation of the control strategy predictions resulting from use of the
EKMA isopleth methodology. OZIPM results were compared with those of the Urban Airshed
Model (UAM), which is listed as one of the EPA's preferred models, and the Systems
Applications Trajectory models, as well as with those obtained with modified versions
of the original OZIPM model. The basic OZIPM model is a simple moving air parcel, or
Sox, model that uses a detailed chemical mechanism for surrogate propylene and butane
ydrocarbons. The model treats time-dependent precursor-emission factors and expansion
f the air parcel. Entrainment is treated by assuming that constant concentrations
'.ist outside the parcel. The primary emphasis of this comparison study was identifi-
ition of features inthe basic OZIPM model that could explain differences between the
of the OZIPM model and those of other models.
V
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Control Strategies
Photochemical Pollutants
Models
EKMA
OZIPP
Ozone
Unlimited
19. SECURITY CLASS (This Report)
Unlimited
21. NO. OF PAGES
102
20. SECURITY CLASS (This page)
Unlimited
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREV.OUS ED.T.ON is OBSOLETE
-------
------- |