United States Office of Air Quality EPA-450/4-81-031d
Environmental Protection Planning and Standards September 1981
Agency Research Triangle Park NC 27711
_
-------
This report was furnished to the U.S. Environmental Protection
Agency by Systems Applications, Incorporated in fulfillment of
Contract 68-02-2870. The contents of this report are reproduced
as received from Systems Applications, Incorporated. The opinions,
findings and conclusions expressed are those of the author and not
necessarily those of the Environnental Protection Agency. Mention
of company or product names is not to be considered as an endorsement
by the Environmental Protection Agency.
-------
EPA-450/4-81-031d
Comparative Application
Of The EKMA
In The Los Angeles Area
EPA Project Officer: Edwin L Meyer, Jr.
Prepared by
U.S. Environmental Protection Agency
Office of Air, Noise and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
September 1981
-------
CONTENTS
DISCLAIMER i i
LIST OF ILLUSTRATIONS v
LIST OF TABLES vii
EXECUTIVE SUMMARY ix
I THE MODEL SERIES 1-1
II MODEL COMPARISONS II-l
A. Comparison of SAI Grid and Trajectory Models II-l
B. Comparison of the SAI Trajectory Model and the
CBM-EKMA Model II-2
C. Comparison of the CBM EKMA with a Similar
Model Using the Regular EKMA Propylene/Butane
Chemi stry II-9
D. Comparison of the Propylene/Butane EKMA Models
Using Adjusted or Standard Photolysis Rates 11-10
E. Comparisons of the Regular City-Specific EKMA
Model and the Same Model Using Time-Dependent
Mixing-Height Input 11-10
F. Comparisons of the City-Specific EKMA and the
Standard EKMA 11-10
III COMPARISON OF MODEL RESULTS III-l
A. A Comparison of Grid and Trajectory Modeling
Results II1-21
B. A Comparison of CBM-EKMA and SAI Trajectory
Modeling Results 111-23
C. Chemical Mechanism Comparison Results 111-24
-------
D. Comparison of Linear and Exponential
Mixing-Height Results 111-27
E. Comparison of the City-Specific EKMA with the
Standard EKMA Results 111-27
IV THE USE OF ISOPLETH DIAGRAMS TO PREDICT OZONE LEVELS IV-1
A. Comparisons of Isopleth Predictions IV-3
B. A New Type of Isopleth Diagram IV-20
REFERENCES R-l
-------
ILLUSTRATIONS
III-l Schematic Comparison of Model Results 111-3
III-2 Trajectory Used in the EKMA Study of 26 June 1974 III-5
III-3 Trajectory Used in the EKMA Study of 27 June 1974 III-6
111-4 Trajectory Used in the EKMA Study of 4 August 1975 111-7
III-5 Trajectory Used in the EKMA Study of 26 June 1982 III-8
III-6 Trajectory Used in the EKMA Study of 27 June 1982 III-9
III-7 Trajectory Used in the EKMA Study of 4 August 1982 111-10
III-8 Comparison between Constant Dilution and Variable
Dilution Rate for 27 June 1982 111-28
IV-1 Standard OZIPP Run (City Specific EKMA) for
26 June 1974 IV-6
IV-2 Standard OZIPP Run (City Specific EKMA) for
27 June 1974 IV-7
IV-3 Standard OZIPP Run (City Specific EKMA) for
4 August 1975 IV-8
IV-4 Standard OZIPP Run (City Specific EKMA) for
26 June 1982 IV-9
IV-5 Standard OZIPP Run (City Specific EKMA) for
27 June 1982 IV-10
IV-6 Standard OZIPP Run (City Specific EKMA) for
4 August 1982 IV-11
IV-7 Standard Ozone Isopleth Conditions IV-12
-------
IV-8 Standard OZIPP Run (City-Specific EKMA) for
26 June 1974: Method (3) IV-22
IV-9 Standard OZIPP Run (City-Specific EKMA) for
27 June 1974: Method (3) IV-23
IV-10 Standard OZIPP Run (City-Specific EKMA) for
4 August 1975: Method (3) IV-24
IV-11 Standard Ozone Isopleth Conditions: Method (3)
Emission Changes IV-25
IV-12 CBM-EKMA Trajectory: 26 June 1974—Base Case IV-28
IV-13 CBM-EKMA Trajectory: 27 June 1974—Base Case IV-29
IV-14 CBM-EKMA Trajectory: 4 August 1975—Base Case IV-30
IV-15 CBM-EKMA Trajectory: 26 June 1982—
Different Trajectory IV-31
IV-16 CBM-EKMA Trajectory: 27 June 1982—
Different Trajectory IV-32
IV-17 CBM-EKMA Trajectory: 4 August 1982—
Different Trajectory „ IV-33
IV-18 EKMA Ha Trajectory: 26 June 1974—Base Case IV-35
IV-19 EKMA Ila Trajectory: 27 June 1974—Base Case IV-36
IV-20 EKMA Ila Trajectory: 4 August 1975—Base Case IV-37
IV-21 EKMA Ila Trajectory: 26 June 1982—
Different Trajectory IV-38
IV-22 EKMA Ila Trajectory: 27 June 1982—
Different Trajectory IV-39
IV-23 EKMA Ila Trajectory: 4 August 1982—
Different Trajectory IV-49
VI
-------
TABLES
III-l Ozone Concentration for NDI Modeling Days III-2
III-2 Initial Conditions Used in the City-Specific EKMA III-ll
III-3 Initial Conditions Used in the EKMA Study
(CBM-EKMA Model) 111-12
III-4 Hydrocarbon Reactivities Used in the CBM-EKMA Study 111-13
III-5 Hourly Emission Fractions and Inversion Bases Used
in the EKMA Study: 26 June 1974 111-14
III-6 Hourly Emission Fractions and Inversion Bases Used
in the EKMA Study: 27 June 1974 111-15
111-7 Hourly Emission Fractions and Inversion Bases Used
in the EKMA Study: 4 August 1975 111-16
III-8 Hourly Emission Fractions and Inversion Bases Used
in the EKMA Study: 27 June 1982—Different
Trajectory 111-17
III-9 Hourly Emission Fractions and Inversion Bases Used
in the EKMA Study: 27 June 1982--Different
Trajectory 111-18
111-10 Hourly Emission Fractions and Inversion Bases Used
in the EKMA Study: 4 August 1982—Different
Trajectory 111-19
III-ll A Comparison of the SAI Trajectory Model and the
EKMA: 27 June 1974 111-25
IV-1 Percentage of Emission Changes IV-4
IV-2 Design 03 and HC/NOX Ratios Used in the EKMA
Predictions IV-16
vn
-------
IV-3 Comparison of Ozone Predictions for the City-Specific
EKMA, the Standard-Conditions (Level IV) EKMA, and the
SAI Airshed Model IV-17
IV-4 A Comparison of 1982 Ozone Predictions for the Isopleth
Models (Ozone Concentrations in ppm) IV-21
IV-5 Ozone Predictions from CBM-EKMA and EKMA Ha Models IV-41
VI 1 1
-------
EXECUTIVE SUMMARY
The Office of Air Quality Planning and Standards (OAQPS) of the U.S.
Environmental Protection Agency (EPA) has for some time been examining the
attractiveness of using grid-based, complex atmospheric simulation models
for predicting and estimating reactive-pollutant concentration patterns
that might result if a specific emissions control strategy were adopted in
a particular urban area. Because of the complexity and the cost of such
models, less complex models like EKMA are also being considered by several
states for widespread applications in regulatory programs. As part of a
major program, OAQPS is sponsoring a study of the use of complex grid
models as a means for evaluating the simpler, less resource-intensive
models. The study investigates three major approaches: (1) simplification
of input information to existing complex models; (2) direct comparison of
simple and complex model performance; and (3) use of complex models in a
limited number of prototype cities representing a variety of
meteorological/air quality/emissions combinations.
The primary objective of the study presented in this report is to
compare the performance of the simple models with that of the more complex
models.
The input files for the SAI Urban Airshed Model for three meteoro-
logical days were used to provide data for a comparison study of a series
of trajactory models. Emission files projected for the year 1982 were
also available so that comparisons between models could be made in order
to project emission changes. The model series was developed to provide a
basis of analysis in which the simplest model (the standard EKMA) could be
compared to the most complex (the SAI Urban Airshed Model). The major
findings of this trajectory model comparison study are organized according
to the model series as follows:
Although the SAI grid and SAI trajectory models share many
algorithms in common, the two models may produce quite
different results from the same input data. Along any
given trajectory path the only major differences between
the models are found in the grid model's ability to
simulate wind shear, convergence, and divergence.
However, the observations often agree more closely with
IX
-------
the simulations of the trajectory model than with those of
the grid model. Hence, the wind fields used in the grid
model may require some reevaluation. Comparisons for
regions in which simple wind patterns are expected are
being carried out in St. Louis and Tulsa. It is also
recommended that trajectory model simulations be used to
supplement grid-model simulations in general, especially
for trajectories to the maximum observed or simulated
ozone sites.
> The EKMA-type model with Carbon-Bond chemistry (CBM-EKMA)
produced simulations of ozone levels fairly close to the
levels simulated by the SAI trajectory model. Differences
in the time dependence of ozone development for the two
models suggest that the variable temperature used in the
SAI trajectory model, when compared with the fixed
temperature used in the CBM--EKMA, may cause the most
significant difference in the results between the two.
> Comparisons of the CBM chemistry and the propylene/butane
chemistry used in the regular EKMA model show that major
differences can occur under certain circumstances. These
differences apparently stem from reaction rate constant
difference and from the balanced spectrum of species
treated in the CBM compared with the polarized reactivity
of propylene and butane. The differences appear to be a
function of HC/NOX concentration, and the two mechanisms
lead to different control-strategy implications. There-
fore, further study is required.
> Changes in propylene percentage or photolysis did not lead
to significant differences in maximum ozone levels when
the propylene/butane chemistry was used for this set of
trajectories.
> Results using linear changes in mixing heights were
compared with results from the city-specific EKMA, which
uses only two mixing heights interpolated with an exponen-
tial function. The mixing-height changes found in the
input files used for this particular study were, in
general, adequately interpolated by the exponential
function used in the EKMA.
> The city-specific EKMA trajectory results often agree
with, or come close to, points on the standard EKMA
isopleth by virtue of large compensating differences in
-------
the models. Treatment of continuous emissions in the
city-specific model tended to generate more ozone;
however, large changes in mixing height tended to generate
less ozone in the city-specific model than in the standard
EKMA isopleth, which neglects continuous emissions and
assumes a small change in mixing height.
Isopleth predictions by all the models tended to reflect the trajec-
tory model results except when assumptions necessary to generate the
diagrams differed from assumptions made in the SAI grid and trajectory
model simulations.
A new type of isopleth diagram has been introduced during this
study. The new diagram is based on emissions that use early trajectory-
model starting times in order to minimize dependence on initial
conditions. Further comparison of the two types of isopleths is
necessary, but the findings of this study suggest that the new method
predicts changes in ozone levels that are more consistent with changes in
emissions than did the original isopleth approach.
XI
-------
I THE MODEL SERIES
The various .applications of the SAI Urban Airshed Model to the Los
Angeles area provide a special opportunity for a comparison of the results
of those applications with the results of other models, particularly with
those of the SAI trajectory model and with versions of the trajectory
model used in the EKMA. If we arrange these models in a series on the
basis of model complexity, with a high degree of commonality among the
models, comparisons of results between adjacent models can be made in a
way that minimizes the number of model differences involved.
A similarity in the results of adjacent models would indicate that
for the particular case studied the differences in complexity were not
important. Correspondingly, a large difference in results would appar-
ently indicate that the features uniquely addressed by the more complex
model were significant for that case.
This section describes such a model series and comparison study using
six Airshed Model runs. Only three meteorological days were used for the
study: 26 June 1974, 27 June 1974, and 4 August 1975. The SAI Airshed
Model was run for each meteorological day, using emissions files for the
1975 base year and projected emissions for 1982.
The following list briefly describes the series of seven models used
in this study. More detailed descriptions of the models are given as each
model comparison is discussed. Some models are distinguished only by the
mode of application.
1) The SAI Urban Airshed Model.
2) The SAI trajectory model (the only difference between
this model and l)--in terms of inputs, computations, and
results at points along a selected trajectory—is the
elimination of horizontal dispersion and convergent and
divergent winds).
By model complexity we mean the spatial and temporal detail with which
the modeling is executed and also the detail with which the physical and
chemical processes are represented.
1-1
-------
3) An EKMA-style model (CBM-EKMA), with the same chemistry
(carbon bond) and mixing heights as the SAI trajectory
model (in this case the main differences are in the use
of fixed concentrations aloft, elimination of eddy dif-
fusion, and the use of a Gear-type integration scheme
instead of a steady-state scheme with finite differenc-
ing).
4) The same EKMA-style model (EKMA Ha), with the standard
propylene/butane chemistry but with modified photolysis
rates.
5) The same EKMA-style model (EKMA lib) with standard
photolysis rates.
6) The same EKMA-style model (EKMA He), with constant
dilution determined from two mixing heights (i.e., the
regular city-specific OZIPP model).
7) The "standard" Los Angeles EKMA.
1-2
-------
II MODEL COMPARISONS
A. COMPARISON OF SAI GRID AND TRAJECTORY MODELS*
The model series begins with the SAI grid model and ends with the
"standard" Los Angeles EKMA trajectory model. The first "step" in the
series is from the SAI grid to the trajectory model. This step involves a
major transition for the entire series from the grid concept to the moving
air parcel trajectory concept. Therefore, we must carefully compare the
differences and similarities between these two SAI models.
The two models were designed to complement each other and many
algorithms are identical for both models (several subroutines of the
computer codes are actually shared); furthermore, both 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 of the grid-model
files. Because the other levels in the grid model often advect with
velocities that are different from this lowest level, the two models often
generate different results as the simulations progress in time. Never-
theless, the trajectory model provides an inexpensive method for testing
parts of the grid model.
A special feature of the SAI trajectory model is that it can be
operated backward in time. This feature was used to generate the trajec-
tories used in this study. The time and location of either the observed
(for the base-case days) or the simulated (for the 1982 days) ozone maxi-
mum were used to start the backward trajectory to 0400 hours. Thus the
forward trajectories always arrived at the desired spot at the proper time
for comparison with either observations or grid-model results. A major
shortcoming inherent in the use of any trajectory model as a tool for con-
trol strategies is that changes in emissions often lead to changes in the
time and location 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 origin of the trajectory
Detailed discussions of this comparison are available elsewhere (see
Reynolds et al., 1979).
II-l
-------
paths. Because the trajectory model is fixed to a specific path, it would
overestimate the effectiveness of emission controls at the monitoring site
if used for control-strategy estimates.
B. COMPARISON OF THE SAI TRAJECTORY MODEL AND THE CBM-EKMA MODEL
The step in the model series from the SAI trajectory model to the
model we call the CBM-EKMA represents the first step into the generic set
of EKMA-type models; all these models can be run with the computer program
known as OZIPM. Therefore, comparisons of the SAI trajectory model and
the EKMA-type model, which is as similar in design as is possible without
major changes in computer coding, are particularly significant.
The following similarities and differences exist between the SAI
trajectory model and the EKMA developed for this comparison:
> Chemistry—Both models use a carbon-bond chemistry that is
also identical to the chemistry currently used in the SAI
grid model. However, the photolytic constants are
computed differently in the EKMA and SAI models. For this
study, we modified the regular EKMA values for N0£
photolysis using a constant factor so that all models
agreed at 1300 PDT. In the Airshed Model and trajectory
model, all other photolysis rates vary with N0£ photoly-
sis, with a fixed ratio for each; but in the EKMA, the
photolysis rates vary independently. Aldehyde photolysis
was adjusted in the EKMA to give agreement between the
models at 1100 PDT.
Actually, these differences in photolysis rates constitute
an advantage in sophistication for the EKMA. The N02
photolysis rates normally used in the EKMA are somewhat
lower than those used in the SAI models because the latter
formerly used an algorithm based on older data. Inciden-
tally, the SAI models are now being updated to the EKMA
N0£ photolysis rates, which are those recommended by
Demerjian, Schere, and Peterson (1980). Since the purpose
of the present study was to compare the EKMA with previ-
ously computed results of the Airshed Model, we were
forced to alter the EKMA models to eliminate this differ-
ence, rather than wait for the older SAI studies to be
rerun with the newer photolysis rates for N02-
The improved treatment of aldehyde photolysis rates in the
EKMA over that of the Airshed Model is similar to use of
II-2
-------
the numerical scheme of Gear rather than the steady-
state/Crank-Nicholson method. Such improvements would be
too costly for the SAI model. The use of fixed ratios of
N0£ photolysis in the SAI models in place of the variable
ratios used in the EKMA model may lead to some minor
differences in chemistry. The ratios are at a maximum at
solar noon and a minimum at sunrise and sunset (Whitten,
Kill us, and Hogo, 1980). The photolysis rates for
aldehydes change more with solar zenith angle than the NC^
photolysis rate does because aldehydes photolyze at the
short wavelength end of the solar spectrum, whereas N02
photolyzes nearer the long wavelength end of the ultravio-
let 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
N0£ photolysis varies throughout the day, reaching a
maximum at solar noon.
For this particular study, the EKMA aldehyde photolysis
values were multiplied by a constant value throughout each
day so that the models agreed at 1100 PDT. Therefore, in
the early morning the photolysis rate of the aldehydes in
the EKMA would be somewhat less than that used in the SAI
models. This effect would add to the expected increase in
reactivity in the mornings for the SAI models as a result
of the steady-state approximations. Since the steady-
state approximation assumes an instantaneous buildup of
the steady-state species, the chemistry would be expected
to proceed faster in the early morning for models using
this approximation than for models of the EKMA type.
These models do not use steady-state conditions and, in
addition, the species build up gradually because of the
slow photolysis rates that occur in the morning. However,
the comparisons made in this study usually showed the
opposite, namely, that the EKMA models were slightly more
reactive in the mornings. The reasons for the occurrence
of this anomaly are unknown at this time.
For this particular study, we chose to modify the CBM-EKMA
to make it as much like the Airshed Model and trajectory
model as possible. By removing all possible potential
variables between models, we hoped to be able to better
compare their underlying principles. For comparisons
involving the regular EKMA, we chose to use that model in
II-3
-------
the manner in which it might normally be applied. The CBM
version of EKMA has normally been applied to assist users
of the Airshed Model to estimate potential control
scenarios. For such applications the photolysis
adjustments discussed here are normal. Perhaps for this
study an additional series of tests with CBM-EKMA using
the regular EKMA photolysis rates would have been helpful,
but they were not performed during that phase of the
contract.
> Reactivity—The SAI Trajectory and Grid models are
operated with reactivity splits that vary with each source
category. However, the EKMA-type models do not have this
feature; reactivity is assumed to be constant for all
hydrocarbon emissions throughout the day. For this study
the OZIPM computer code was modified so that three
different hydrocarbon reactivities could be used: 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 were in the
temporal variations aloft and in the variable emissions
used in the SAI models compared with the fixed reactivi-
ties employed in the emissions for the EKMA models. The
reactivities used in the CBM-EKMA were determined from
averages of emissions used in the SAI trajectory model;
for hydrocarbons aloft a weighting factor was used each
hour that was equal to the positive change in mixing
height.
In a sense, the N02/NOX ratio also acts as a reactivity
effect. The EKMA-type models all use a fixed ratio of
N02/NOX (10 percent) for all NOX emissions, whereas the
SAI models allow this ratio to vary according to their
source category (the average ratio is typically about 5
percent). At this time, we have not attempted either to
change the EKMA or to fix the ratio in the SAI model,
because we feel the difference is trivial. For NOX aloft,
the EKMA uses pure NC^, which is close to typical values
found in the SAI trajectory model simulations; the
calculated N02/NOX ratio is almost always greater than
0.85, often with values greater than 0.99.
> Integration Scheme—The EKMA uses a Gear-type method with-
out any steady-state approximations, whereas the SAI
trajectory model uses the same Crank-Nicholson finite dif-
ferencing scheme and steady-state approximations that are
II-4
-------
used in the SAI grid model. Again, this difference in
model sophistication is opposite to the direction intended
by our series of models. However, our previous tests
using smog chamber experiments suggest that the dif-
ferences introduced by the numerical integration schemes
should be minimal. Although the Gear integration scheme
is well established for use in chemical kinetics calcula-
tions, the computer time and storage requirements are
large. Use of more efficient schemes, such as the Crank-
Nicholson technique employed in the SAI Airshed and
trajectory models, requires the consideration of many
species in steady state. In addition to.reducing the
requisite size of the system, the steady-state approxima-
tion helps eliminate a problem arising from stiffness.
This mathematical property, common to most chemical
kinetic systems, is a result of some parts of the system
changing much more quickly than others.
Although the steady-state approximations for the CBM-have
been carefully chosen and tested, there is no guarantee
that the differences in results between the stiff-stable
Gear method and the steady-state Crank-Nicholson method
will be small compared with the other model differences
studied here (such as eddy-diffusion and multiple vertical
layers). Physically, the steady-state approximation
implies that the mass of the steady-state species is
insignificantly small and that these species are created
and destroyed instantaneously,, The first implication
leads to mass balance problems when a species considered
in steady state is large in concentration compared with
its precursors or products. The second implication is
most severely tested in atmospheric models during sunrise
and sunset conditions when the photochemistry may be
proceeding extremely slowly.
Vertical Layers—The SAI grid .and trajectory models share
the same vertical layers and algorithms for numerical
integration, but additional contributions are introduced
into the grid model for convergent or divergent winds.
The EKMA, of course, 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 SAI models account for mixing within the mixed
layer, mixing within the inversion layer, and eddy diffu-
sion among all layers. The EKMA assumes instantaneous
mixing within the entire mixed layer, and no eddy diffu-
II-5
-------
si on is considered across the boundary defined by the mix-
ing height.
In our comparison studies, the effectiveness of multiple
layers and eddy diffusion was one of the central issues to
be evaluated by means of the models in this region of our
model series. A problem that we tried to avoid, which
does confound the comparison, is the entrainment of
pollutants from aloft. The EKMA requires fixed concentra-
tions aloft, so these must be chosen from some "proper"
average of the time-dependent concentrations computed in
the SAI trajectory model. With the average used in this
study we attempted to weight most heavily the concentra-
tions entrained when the mixing height rises most signifi-
cantly. 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 considered 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 affect-
ing the surface layer from concentrations aloft is eddy
diffusion. Eddy diffusion is treated in the SAI grid and
trajectory models but is neglected in the EKMA.
Mixing height—All models used in the EKMA that are based
on the OZIPM computer code, as well as the SAI grid and
trajectory models, can use identical mixing heights with
values that vary linearly in time between specified values
at each hour. However, the SAI models have a special
arbitrary feature that can produce mixing-height profiles
different from those of the EKMA. Minimum (50 m) and
maximum (300 m) values for the heights are specified in
the SAI models used in this study. When the mixing height
is between 100 meters and 600 meters, the SAI models have
two equal levels below the mixing height, and two equal
levels between the mixing height and the top of the
modeling region (1000 m). For all such cases the EKMA and
the SAI models use mixing heights that vary with time in
the same fashion. However, these two models may not have
identical mixing heights for heights less than 100 m or
more than 600 m. The SAI models linearly interpolate a
value of mixing layer thickness called DIFBREAK at each
time step from the mixing-depth input at each hour.
However, below 100 meters and above 600 meters the size
and the number of the cells with rapid mixing does not
correspond to two equal cells below DIFBREAK. If DIFBREAK
II-6
-------
falls below 52.5 meters the SAI models use one cell of 50
meters with rapid mixing. When the computed value of
DIFBREAK goes above 52.5 meters, two cells of 50 meters
each are suddenly treated with rapid mixing parameters.
Above 600 meters, two cells of 300 meters each are used
until the computed value of DIFBREAK reaches 630 meters,
at which time a third cell of 200 meters is suddenly
added, bringing the effective mixing height to 800 meters.
Temperature—The temperature varies in the SAI grid and
trajectory models, but the EKMA presently employs a fixed
temperature. This difference can be significant.
Temperature directly affects both vertical diffusion and
chemistry in the SAI grid and trajectory models, but it
can affect only chemistry in the EKMA models because
vertical mixing is assumed to occur instantaneously within
the mixed layer.
Emissions—The SAI Airshed Model and trajectory model
allow for emissions into all layers; the EKMA treats all
emissions in the mixing layer alike and does not directly
address emissions above the mixing layer. For the layer
aloft in the EKMA, the emissions must be accounted for in
the determination of the fixed concentrations aloft. The
sum of the emissions used in the lower layers of the SAI
trajectory model was used for the mixing layer in the
EKMA.
The EKMA was modified for this study to treat emissions in
an absolute fashion rather than as a multiple of the
initial conditions. This was implemented by treating the
initial conditions as pollutants in the transported
surface layer for the EKMA, but the normal initial condi-
tion (for the CALC mode in OZIPP or OZIPM) was specified
as 1.0 for both HC and NOX. The OZIPM computer code was
modified slightly to ignore this specification of 1.0 as
an initial condition. However, the multiplication of this
number times the emissions was retained to allow the
generations of isopleths on a scale relative to the base
case emissions.
Emissions for the SAI grid and trajectory models are given
as total moles emitted into each cell during each hour.
To arrive at values appropriate for the EKMA we used the
following conversion procedure: The mixing volumes of the
SAI trajectory model and the EKMA were assumed to be equal
below the mixing height, such that
II-7
-------
where A is the area and Z^^ is the mixing height. If the
volume is expressed in cubic centimeters, the moles
emitted in one hour (the number available from the trajec-
tory model computer output) can be divided by this volume
to give moles-cm"^ units. This, in turn, can be readily
converted to the EKMA units of concentration, ppm, by
using a conversion factor of 2.445 x IQr® ppm moles'* 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 models used
in the EKMA that are coded in both OZIPP and OZIPM.
First, the use of 1.0 for the initial concentrations of HC
and NOX in the CALC mode implies that each one-hour emis-
sions number would be the concentration added during the
hour if the mixing height were the same as the initial
value Zg. However, the computer codes are written so
that, at each instant in time, the emissions factor is
multiplied by Zg/Z^. Therefore, the Z^ drops; out and
the actual numbers (En) used are merely
En = A x 2.445 x 1010/Z0
where A and Zg are in centimeter units.
In this particular study it was necessary to use special
emission factors occasionally. These factors were used
for those cases in which the SAI trajectory model employed
substantial initial conditions for the first two layers
(up to 100 m) even though the mixing height was only 50
meters. Since there was typically an order-of-magnitude
difference in concentration between the initial conditions
below 100 meters and the layers above this level, there
was no way to compute a reasonable average to be used in
the EKMA for the layer aloft. Hence the initial condi-
tions used in the SAI models between 50 meters and 100
meters were treated as emissions in the EKMA models.
These "emissions" were added to the normal emissions
during the hour when the mixed layer in the SAI models
jumped from 50 meters to 100 meters. For these cases the
appropriate EKMA emissions factors merely add to the
11-8
-------
normal emissions, since the initial mixing height is 50
meters and the height added is also 50 meters.
C. COMPARISON OF THE CBM-EKMA WITH A SIMILAR MODEL USING THE REGULAR
EKMA PROPYLENE/BUTANE CHEMISTRY
The obvious intention of this comparison was to evaluate only the
differences in the carbon-bond chemistry used by the SAI models and that
used in the standard versions of the EKMA. Although the OZIPM computer
program was specifically written to handle such comparisons, some compli-
cations were introduced when the EKMA was operated with the
propylene/butane chemistry. First, the N02 and aldehyde photolysis rates
were changed to make the carbon-bond and the propylene/butane mechanisms
compatible. The N02 photolysis rates presented no problem since they were
modified to be identical with the model using the CBM; this adjustment was
discussed in the comparison of the SAI and CBM-EKMA models. For the
aldeydes, it was assumed that the molecular concentrations of formaldehyde
and acetaldehyde were always equal. The CBM used in this study assumes
that the amounts of formaldehyde plus ketones is equal to the sum of the
higher aldehydes. The aldehyde photolysis for the EKMA propylene/butane
mechanism needed to be adjusted downward by a factor of three to ensure
that the radical inputs for both mechanisms were equal at equal levels of
total carbonyl concentration. Many studies, such as Whitten, Killus, and
Hogo (1980), have shown that the input rate of radicals from aldehyde
photolysis is about the most sensitive single factor in smog chemistry.
We felt that a comparison of the CBM-EKMA and a model using the standard
propylene/butane chemistry of EKMA would show the differences in the
chemical mechanisms more clearly if the radical input rates were
matched. However, we were concerned with the effects of these adjustments
to the standard chemistry, so the next step in the model series involved
only the reestablishment of the value of standard photolysis factors.
The reactivity of the two mechanisms was occasionally adjusted in
this study. The total HC concentrations and emissions rates were identi-
cal for these two models, but the standard EKMA propylene/butane chemistry
was typically used with the standard 25 percent propylene, 75 percent
butane, and 5 percent aldehyde reactivity. When the two mechanisms did
not produce similar results, attempts to match results were sometime made
by varying the reactivity of the propylene/butane mix.
II-9
-------
D. COMPARISON OF THE PROPYLENE/BUTANE EKMA MODELS USING ADJUSTED OR
STANDARD PHOTOLYSIS RATES
As just discussed, the model closest in the model series to the CBM
used specially adjusted NOp and aldehyde photolysis rates. For the
individual trajectory-model runs, the comparison, at this point in the
model series, illustrates the sensitivity to the photolysis constant
adjustments. For the control-strategy assessment study using isopleth
diagrams, we chose this point in the model series to convert from iso-
pleths based on emissions inventories (the models were started at 0400) to
isopleths based on 0800 initial precursor concentrations. The details of
the isopleth construction and use are discussed later.
E. COMPARISONS OF THE REGULAR CITY-SPECIFIC EKMA MODEL AND THE SAME
MODEL USING TIME-DEPENDENT MIXING-HEIGHT INPUT
The only difference between the EKMA just discussed, which uses the
standard photolysis rates, and the regular city-specific EKMA regarding
the OZIPP computer code is the use of variable mixing-height inputs in the
former, more sophisticated version. In the propylene/butane version of
the EKMA just discussed, instantaneous mixing heights are computed by
linear interpolations between hourly input values rather than by the OZIPP
method, which uses an exponential interpolation between the morning and
afternoon mixing heights. The exponential increase in mixing height with
time provides a constant dilution factor. In addition to the potential
problems associated with the assumed shape of the mixing-height time
profile, the standard EKMA cannot consider a decrease in mixing height
that may follow an increase, nor can it consider a resumption in mixing-
height growth following a decrease or a pause. However, the Los Angeles
area has relatively less variance in mixing heights than other cities, so
this aspect of the model series is not severely tested in this location.
F. COMPARISONS OF THE CITY-SPECIFIC EKMA AND THE STANDARD EKMA
This particular study involved the city of Los Angeles, which pro-
vided the basis for the standard EKMA. Hence the longitude and latitude
used, as well as the time of year, are quite similar for these two
modeling approaches. The main differences occur in mixing heights,
materials aloft, and post-0800 emissions.
11-10
-------
Ill COMPARISON OF MODEL RESULTS
The results of applying the seven models over the six trajectories
are given in table III-l and comparisons of certain results are illus-
trated in figure III-l. The paths of the six trajectories are illustrated
in figures 111-2 through 111-7. In each case the trajectory was deter-
mined by running the SAI trajectory model backward with the chemistry
turned off. For the three base-case days, the trajectories were deter-
mined so that the maximum observed ozone occurred in the trajectory
model. For the 1982 days, the location of the predicted maximum ozone
grid square in the SAI Urban Airshed Model was used to start the backward
trajectory. Details of the input data are given in tables III-2 through
111-10.
The early-morning start (0400) of the trajectory used in this study
raises questions about uncertainties in wind patterns and mixing depths
that can occur in the predawn hours. Therefore, the following information
is given to help clarify the use of trajectory models such as the EKMA-
type model in simulating the hours up to the normal starting time of 0800:
> Wind speeds in the early hours are not typically rapid so
that the distance traveled up to 0800 would not be
great. Hence the critical input data involved here, the
emissions densities, should not be significantly affected
by an ill-defined pathway. A general rule to consider
would be the resolution and spatial variance of the
emission inventory compared to the length and uncertainty
of the early-morning pathway.
> Chemistry in the predawn hours is not sensitive to
concentration, so the actual mixing heights used in the
model up to 0800 are not critical (within some reasonable
range). When the photochemistry becomes active around
0800, the mixing depth becomes more important, but this
quantity is better defined. Since emissions are treated
in an absolute fashion, the mixing heights used prior to
0800 would not affect the concentrations at 0800. Of
course the initial conditions used at the start of the
simulations (in the present case, 0400) would be affected
III-l
-------
oo
>-
ct
O
CJ
a:
o
o
o
o
tvl
o
CQ
° " 88Sg 8SS8
i«r ' ~ ' '
M
O
•o
l_
TS 3C ~+ r*> t\i ~* O^*
«uj OOOO* O* O
w
t*o
•o
k.
03 21 o
»^ CsJ (NJ •*• O P^J
dodo do
**
UJ
V £
* 00ifl0 — ^SooO
z oooo ooooo
<_>
WTi
o
L. p O
O 4* —— *£>
*> *j O *"> f> f^J «~*
. u
o_
*2
*—£tj O*VOO« csjmmr^--*
4C »^ »• frj r*> PSJ ^j ro ^u *-* f*o ""^
^U^ oodd dodod
s!
K S
I §
I 1
:§ a'-. *•
g?g|-5l=SlS
^*^f<^«U4-'
C-C^3— Sc-C
s 5 §
Lb.SU.
O O. O
LU O U.
00 CO Ch
|||| jfllf
2|1
u I —
SI1
X u
-------
0.4
0.3
0.2
0.1
0.0 0.1 0.2 0.3 0.4
SAI Trajectory
(a)
0.4
0.3
0.2
0.1
+25%
0.0 0.1
0.2 0.3
SAI Trajectory
(b)
0.4
FIGURE III-l. SCHEMATIC COMPARISON OF MODEL RESULTS
III-3
-------
0.4
0.3
0.2
0.1
I
I
0.0 0.1 0.2 0.3 0.4
CBM-EKMA
(c)
CO
I
0.4
0.3
0.2
0.1
0.0 0.1
0.2 0.3
CBM-EKMA
(d)
FIGURE III-l (Concluded)
0.4
III-4
-------
1SU3
o
..i;r;;.T! 4 ::{.:;.._. i'!.'-'Tr
O
o
I
o
o
o
•a-
el
{-.^iiiuiifi^i £ H-.
r^^ -.• •_-.-..-.•.- ....-.- n T
CM
I
UJ
o:
C5
' x: /•'• •'•'••Xy' - * : .... •
' i '• T " V ' T t t IT 't " i '• ' < L jt i
1S3H
III-5
-------
1S«3
o
.+ .:): + .... 4.:. . 4- . 1
csj
O
oo
LU
UJ
Q
LU
oo
o:
o
"3
«S
ct
C3
t t r
t T f T < T
1S3H
III-6
-------
. 4. v . * . .-4..
. .. *
CM
LT>
to
o
U-
o
S
V)
ce
O
cc
i s
T < T t
T ' t 't I
• T t t
1S3H
III-7
-------
. *• • •?..••*.-..
I. .. •!;. .1 . 1. ;.;.; { ; I . 1 > ...I T
. .; j. . ...* ..X.£.'. ...* ]
CSJ
00
UD
CM
OL
O
»—
«_J
LU
rr>
o:
LO
I
' t '•'•' + •' T' < •" T V "• T " •' i T J .'' t- •' + •' • t :.' t r ' 4 t •'»'•'• 'ii'
o
1S3N
I1I-8
-------
1SU3
o
cc
-s
r^
CSJ
u_
O
>-
a
•=>
I—
CO
<
o
-------
1S03
o
rsi
CSJ
CO
cr.
CJ
3
^~
U-
o
o
i i O
. I «•>
o;
O
oo
ex
o
LlJ
•-3
Ct
a:
' i!' '•
tt T'<
111-10
-------
X
o
ro *± VD VD O ^J"
«—I O r-H O VO CO
CO VO
VD •—« CT>
CO
O
ro o Q.
O •— a.
en o r~-
in CFi VD CO CO OO
o o o o o o
o o o o o o
o
UJ
o.
O)
o »—
(O E
t»- ro 5.
i_ O O.
us -1—
oo
in O-i VD O ro VD
«!• in oo in r-^ in
O O O O O O
o o o o o o
o
UJ
oo
oo
o
t-
1—4
o
z
o
CM
I
c
o
^O ^~+-
i. 4-> O
«3 It- E
u o o.
O r- Q.
j_ «C ^-«
ID
in
in
rH O
* * * *
o o o o
o m
•
>4- «O E
L. O O.
3 o n.
oo j- »—•
co co !-. ro
80 >* o
o o o
o o o
o o
o o o
«Ji O CTi VD «5fr in
co in cri in «a- o
nj- vT> r~- co co ro
o o o o o o
CD
CD
to x
<4- o
PO f-~ r^ co
in in oo **•
O rH O O
>— < ro
O O
O O O O O O
o
a.
X
o
ZT
*V I
V Nl
O
Z
n t
UJ
4->
(O
O
«— 1
r-~
O
«•
r~>
CT>
i— i
O)
C
3
O
VD
CO
0
r»-
O
«•
r»~
c^
<— i
VD
C
O
r^
CO
CO
1 —
o
in
r^
CM
i— 1
4-*
ISI
3
CT>
3
«— (
O)
c
rj
VD
CO
ro
00
o
CO
CO
Oi
f— 1
o>
c
3
<-D
r^
co
ro
r-
O
CO
00
o>
1— <
4-*
l/l
3
cn
3
et
«3-
§
C?
II
1
•r-
4->
cn
c
4->
J_
ft)
4->
tn
*
III-ll
-------
ro o Q.
O r- Q.
CTi r^
LO VO CO OO CO
o o o o o o
O O O O O O
LU
O
0)
U
1- O Q.
3 ^-«
OO
O O O O •—i •—i
<—i t—4 »—i o O O
X X X • • •
»—i 1-4 i—t o O O
LU
I
00
o
o
00
c
o
$_•!-> ^^
LO f-- O Lft
i™-* o^ m »-H *jf~ ro
t—i o LO •—i t—i ro
O O O O O O
:*:
LU
x<*- e
o o 5.
Z r- O.
«c —
c\j CM r^^. ro *^" o^>
o o -a- o o t-H
o o co o o o
O CD O CO O O
o
LU
oo
o
O L. -—•
fl3 tJ O
14- U E
J_ O Q.
3 l_ 0.
00 T3 —*
CTi CM LO CTi <—' «—I
oo r*. tn »-i CO t-i
LO ^* r*^ ro ro ro
O —i O O O C
o
o
CO
o
3
oo
x e
O CL
z; Q.
ro Lft OO t—i O O
rr O VD ro CM ro
O CM O O O O
o o o d o d
ro
i
CM
O
CM
VD
o ur>
LT
O O CV
O O O •—i <—i O
LU
CO
IT5 CM
f^ CM CM CO
CTi CO CO CT1
CD CU t/> Ct) QJ *S*
C C 3 C C 3
3 3 CT) 3 3
O ""3 3 <-} <-} 3
CM CM
CM CM
111-12
-------
I/I
c
o
>-
O
00
u
£
rj ro o ro o\ cr
m m n csi CM tsj
O O O O O o
O O d O O O
«-« t-i •-( ro a\
W v T f) ro n
000000
......
o o o o o o
QC
cp ui
OOOWld—i
CvJ
evj
OO
i-l— 1CVJ
dodddcj
1-1 w r^ PO in
ff> oo CO i«~ ro o>
«cudvcr^r^^
......
o o o o o o
V « f~» r~- in — i
ro ro n CM ro r^i
O «D O O CD O
d d d o d o
CD
O
o
•R
QC
<
OJ ^ KCi ^ *— (
r«- r** tp CO co r^
O C\J O O i-i O
d d d d d d
o o o o o o
^H r*» f^ ID
&* *~* KO *~< •—< oo
«\jmc\j«jrocM
dododo
OO i —
ID O O
O O O O O O
m »-c oj ip CM ro
O o O O O O
d d d o d o
O^ \£> CD CO C^ CO
JggggSS
d d d d d o
po r^. ID tn ir> trt
n csj CM cvj fvi OJ
O O O O O O
......
OOOOOO
tO IT) ID ID ID ID
(VCVJCMCM(SJCM
o o o o o
o fo C\J m ro ro
tevovototovo
......
OOOOOO
CVJ CM CM CO CNJ C\J
OOOOOO
d d d o d o
ID (M
«• r^ «M ro eo
fv. gr\ CD CD O**
Q; O) in OJ O>
P § §• § §
•^ --S 5 T "i
VD Iv. ic r^
PO C\J * ^1 C*J
111-13
-------
TABLE III-5. HOURLY EMISSION FRACTIONS AND INVERSION BASES
USED IN THE EKMA STUDY: 26 JUNE 1974
Hydrocarbon NO^
Time Inversion Base Emissions Emissions
(PST) (meters) (ppmC) (ppm)
400 100.0 0.000 0.000
500 100.0 0.105 0.029
600 127.3 0.090 0.020
700 251.0 0.169 0.042
800 297.5 0.324 0.080
900 338.9 0.327 0.089
1000 402.7 0.289 0.055
1100 499.4 0.185 0.032
1200 594.5 0.174 0.027
1300 600.0 0.129 0.034
1400 584.2 0.182 0.037
1500 518.2 0.131 0.201
1600 485.0 0.082 0.070
III .1.4
-------
TABLE III-6. HOURLY EMISSIONS FRACTIONS AND INVERSION BASES
USED IN THE EKMA STUDY: 27 JUNE 1974
Hydrocarbon NOX
Time Inversion Base Emissions Emissions
(PST) (meters) (ppmC) (ppm)
400 50.0 0.000 0.000
500 50.0 0.058 0.006
600 100.0 0.415 0.121
700 183.2 0.839 0.209
800 273.9 0.742 0.160
900 387.8 0.524 0.095
1000 471.9 0.331 0.050
1100 600.0 0.353 0.054
1200 800.0 0.261 0.057
1300 800.0 0.326 0.071
1400 800.0 0.221 0.185
1500 1000.0 0.134 0.443
1600 800.0 0.368 0.063
111-15
-------
TABLE III-7. HOURLY EMISSIONS FRACTIONS AND INVERSION BASES
__ USED IN THE EKMA STUDY: 4 AUGUST 1975
Hydrocarbon NO^
Time Inversion Base Emissions Emissions
(PST) (meters) (ppmC) (ppm)
400 290.2 0.000 0.000
500 285.6 0.007 0.003
600 270.9 0.021 0.006
700 270.3 0.055 0.014
800 255.6 0.086 0.020
900 283.3 0.054 0.011
1000 302.5 0.025 0.005
1100 310.6 0.017 0.005
1200 366.8 0.046 0.009
1300 426.3 0.035 0.074
1400 429.0 0.028 0.016
1500 429.7 0.072 0.010
1600 368.3 0.008 0.002
111-16
-------
TABLE II1-8. HOURLY EMISSION FRACTIONS AND INVERSION BASES
USED IN THE EKMA STUDY: 26 JUNE 1982--DIFFERENT
TRAJECTORY
Time
(PST)
400
500
600
700
800
900
1000
1100
1200
1300
1400
1500
1600
Inversion Base
(meters)
100.
100.
117.8
226.1
265.6
321.9
414.3
483.0
578.7
588.0
589.2
512.9
440.8
Hydrocarbon
Emissions
(ppmC)
0.000
0.001
0.034
0.111
0.286
0.323
0.221
0.204
0.188
0.153
0.101
0.058
0.015
NOX
Emissions
(ppm)
0.000
0.001
0.003
0.061
0.064
0.055
0.076
0.084
0.054
0.097
0.029
0.020
0.007
111-17
-------
TABLE III-9.
HOURLY EMISSION FRACTIONS AND INVERSION BASES
USED IN THE EKMA STUDY: 27 JUNE 1982—DIFFERENT
TRAJECTORY
Time
(PST)
400
500
600
700
800
900
1000
1100
1200
1300
1400
1500
1600
Inversion Base
(meters)
50.0
50.0
50.0
101.8
167.5
228.5
294.8
372.8
427.4
481.0
559.1
800.0
800.0
Hydrocarbon
Emissions
(ppmC)
0.000
0.000
0.000
0.00015
0.0004
0.013
0.071
0.200
0.529
0.407
0.339
0.194
0.159
NOX
Emissions
(ppm)
0.000
0.000
0.000
0.001
0.004
0.007
0.020
0.340
0.371
0.176
0.122
0.060
0.055
111-18
-------
TABLE 111-10,
HOURLY EMISSION FRACTIONS AND INVERSION BASES USED
IN THE EKMA STUDY: 4 AUGUST 1982—DIFFERENT
TRAJECTORY
Time (PST)
400
500
600
700
800
900
1000
1100
1200
1300
1400
1500
Inversion
Base (meters)
204.2
224.6
189
191,
219.1
257
282
302
346
334.1
341.2
401.2
Hydrocarbon
Emissions
(ppmC)
0.000
0.003
0.004
0.002
0.019
0.045
0.054
0.073
0.055
0.028
0.037
0.037
NOX
Emissions
(ppm)
0.000
0.005
0.002
0.00044
0.010
0.032
0.024
0.024
0.013
0.009
0.015
0.013
111-19
-------
by the ratio of mixing heights used between 0400 and
0800. To some extent, the pollutant concentrations would
also be modified by the concentrations assumed aloft.
Nevertheless, the main reason for starting the model
earlier than 0800 is to enhance the dependence on emis-
sions and suppress the dependence on initial conditions.
Thus the general rule here would be to ensure that the
contribution of the concentrations from 0400, or whatever
early starting time is used, is small compared to the
contribution of the emissions between 0400 and 0800, or
whatever time the photochemistry begins. Critical factors
to check would be the 0400 or initial mixing height as
compared with the 0800 mixing height and the concentra-
tions used aloft.
In the present study the model comparisons are all based
on common, time-dependent, gridded wind field and mixing-
height sets that were generated for the SAI grid model.
Any uncertainties in the preparation of these input files
apply to the grid model as well as to the rest of the
models used in the present study. However, the two
mitigating factors associated with the slow wind speed and
insensitivity to predawn mixing heights just discussed
also apply to the grid model.
Although grid models have an inherent advantage over
trajectory models in dealing with complex wind shear
effects in the predawn hours, in actual grid model
applications, the typical grid sizes used are much larger
than the predawn wind shear patterns. Furthermore,
monitoring data on the real wind patterns is rarely, if
ever, available on a scale small enough to model the
patterns precisely even if the grid size of the airshed
models were small enough. Hence, in practical applica-
tions, the grid models have no real advantages over the
trajectory models in their ability to handle predawn
uncertainties in wind patterns and mixing heights.
The emission density of the Los Angeles region is rather
uniform over large areas; therefore, the uncertainties in
wind patterns during the predawn hours do not affect the
emission densities used in the models. This particular
point is really an application of the general rule just
discussed, namely that, especially in the Los Angeles
area, the slow (in time) wind speeds combine with a slowly
111-20
-------
rising (in space) emission density to minimize the effects
of uncertainties in wind patterns.
It should be noted also that there are large differences in initial
conditions among the six specific trajectories. In all cases, the values
chosen came from the next highest model in the model series. For example,
the initial N02/NOX ratio shown in table III-3 for the CBM-EKMA on 26 June
1974 was 0.629. This number was obtained from the computer output
generated by the SAI trajectory model for the time of 0400 PST. In table
III-2, the city-specific EKMA N02/NOX ratio for the same trajectory was
0.73, which came from the CBM-EKMA computer output at 0800 PDT.
A. A COMPARISON OF GRID AND TRAJECTORY MODELING RESULTS
We first compare the results of the SAI grid model and the SAI
trajectory model. Because the only possible difference in results along
the trajectory paths must be due to the result of horizontal dispersion
and dilution effects (the two models are identical except in concentration
changes caused by horizontal effects such as convergence), the large dif-
ference in results for some days (notably 26 June 1974, 4 August 1975, and
27 June 1982) is striking. The maximum observed ozone concentration
occurs in Fontana for the first two of these days and the grid model
results show poor agreement with observations, whereas the trajectory
models all show at least fair agreement with the observations.
Comparison of the results produced by the two models for these two
days illustrates a very important point about the use of sophisticated
grid models--that is, trajectory models provide a diagnostic tool useful
in understanding discrepancies between observations and grid-model simu-
lations. In the present case, a review of grid model intermediate results
revealed that the low simulated ozone values for Fontana are significantly
affected by wind shear effects computed from the topographic influences of
the steep mountains nearby. Evidently, the grid-model wind field algo-
rithm leads to unrealistically high dilution. Since such effects are
ignored in the trajectory models, it appears that for grid cells so
affected the grid-model results are less dependable.
The 27 June 1982 trajectory illustrates a somewhat different prob-
lem. The grid-model simulation predicts a rather high ozone level because
of a high (0.3 ppm) ozone concentration aloft near Upland. The ozone is
entrained when the mixing layer thickens in the afternoon. The high ozone
aloft originates several grid squares away, where a simulated wind-flow
convergence zone pushes polluted air aloft. This air is, in turn, carried
to the Upland region by the winds aloft. Since the winds aloft are not
parallel to the surface winds, the trajectory model entrains much cleaner
111-21
-------
air, using the same inversion rise as the grid model. In fact, the ozone
in the trajectory model is diluted during the period in which air along
the same path in the grid model is enhanced by ozone from the air aloft.
However, the impact of the ozone aloft may be exaggerated in this case
because of a questionable feature of the grid model mixing-height algo-
rithm. This feature is now being investigated, partly as a result of the
present comparison study.
Trajectory models cannot normally simulate important effects such as
wind shear. In one case we found a method for coping with this effect.
Two values are shown in table III-l for the 4 August 1982 SAI trajectory
mModel simulation at Upland. The value in parentheses was obtained by
considering elevated point sources. The other value ignores elevated
point sources for this trajectory as do the values from all the EKMA-type
models for this particular day. There was a substantial wind shear
through the mixing layer up to 1000 hours in the morning for this particu-
lar trajectory. At the same time, the atmosphere was only moderately
unstable in the cool mixing layer. These two effects of wind shear and
minimal mixing combine to force the elevated NOX emissions down to the
side of the trajectory path modeled. Of course this effect is exaggerated
to some extent in the present algorithms because the path of the trajec-
tory model is defined by only the surface-layer wind field of the grid
model.
Some sophisticated averaging process might have been used for the
grid model layers below the mixing height, based on such factors as the
vertial mixing rate and wind speed, to determine an alternate trajectory
path. In this case, a partial treatment of the elevated sources might
have been more appropriate. However, the success of the simple treatment
of the trajectory path and the simple solution of ignoring the elevated
sources seemed to preclude the need for any further improvements during
the present study. Nevertheless, a concurrent study at SAI has encoun-
tered a severe wind-shear day in the Sacramento region that does not
respond to the simple ignorance of elevated sources, so that averaging of
the vertical wind fields has been necessary to generate reasonable
simulations.
Incidentally, the trajectory model simulations imply that the site of
maximum simulated ozone at Upland for this day could result from the
presence of wind shear below the mixing height. The absence of wind shear
would apparently allow the elevated NOX emissions to suppress ozone along
the trajectory path.
111-22
-------
B. A COMPARISON OF CBM-EKMA AND SAI TRAJECTORY MODELING RESULTS
The nine pairs of results shown in table III-l, which compare the SAI
trajectory model with the CBM-EKMA, show agreement to a standard deviation
of less than 8 percent. Considering the many differences between these
two models this agreement appears to indicate that the most important
features of the more complex trajectory model are reasonably represented
in this CBM-EKMA. Future studies might create more models between these
two variations for purposes of investigating additional model features.
As an example, model results have been found to be sensitive to tempera-
ture. In a recent study, which used a different Los Angeles trajectory
than any reported in this work, it was found that the same EKMA, used with
CBM chemistry, produced 0.41 ppm ozone when run at a constant 305 K and
0.38 ppm ozone when run at a constant 300 K (Whitten and Hogo, 1980).
Temperatures in the SAI trajectory model typically vary between about
290 K and 314 K in the course of a simulation. An overall impression,
gained from comparing the detailed outputs of the SAI trajectory model and
the CBM-EKMA, is that the formation rate of ozone in the morning hours is
usually somewhat faster in the CBM-EKMA than in the SAI trajectory model;
however, the situation reverses in the afternoon. This effect would be
consistent with the constant temperature used in the EKMA, which is higher
than the morning temperatures used in the SAI trajectory model and which
is lower, of course, during the afternoon. This overall impression of the
relative rates of ozone formation is opposite to the behavior that is
normally expected as a result of differences such as steady-state approxi-
mations versus full Gear-type simulations and differences in the aldehyde
photolysis rate.
A very brief investigation of diurnal emission patterns was made for
the 26 June 1982 simulation based on the 1974 path. Two values are shown
in table III-l: the value shown in parentheses was computed using the
1974 pattern of emissions; however, the emissions levels were uniformly
reduced to reflect the reductions along the total path. The regular value
was computed directly from the 1982 emissions files of the SAI grid model,
as were all other 1982 values shown in table III-l. Since the results
differ by only 5 percent, the pattern change was only moderately impor-
tant.
The CBM-EKMA model results originally showed the largest difference
from results of the SAI trajectory model for the trajectory to Upland of
27 June 1974. Although observed maximum ozone of 0.49 ppm occurred at
Upland that day, the SAI grid and trajectory models both simulated only
0.32 ppm ozone, whereas the CBM-EKMA originally simulated only 0.28 ppm
ozone. After a considerable effort was made to understand the main
reasons for the discrepancy, it was discovered that carbon monoxide levels
111-23
-------
in the SAI trajectory model have a significant influence on ozone produc-
tion. The ozone result shown (0.30) for the CBM-EKMA in table III-l
includes CO emissions in order to bring the CO levels close to the values
seen (approximately a 3 ppm average) in the SAI models.
To estimate the importance of physical dispersion differences between
the SAI trajectory model and the modified CBM-EKMA we operated both models
without chemistry. The comparison data for 27 June 1974, are presented in
table III-ll. Data for ethylene (ETH) and the single-bonded carbons (PAR)
are not completely comparable because the reactivity (splits between the
various CBM species) varies with time in the SAI trajectory model but is
constant in the EKMA. The overall similarity indicates that such factors
as multiple levels, eddy-diffusion, and variable reactivity are not
critical for this trajectory.
Since the results of the runs with chemistry are similar for all of
the comparisons shown in table III-l, it is tempting to speculate that the
multiple levels, eddy-diffusion, and variable reactivity treatments are
never very important. However it is probably better to determine for each
specific case to be considered in the future whether or not any of these
treatments appear to be especially important. For instance, variable
reactivity would be an especially important treatment if a large new
source was to be considered along the trajectory path and if the hydro-
carbon reactivity of this source was considerably different from the
average of the existing sources.
C. CHEMICAL MECHANISM COMPARISON RESULTS
A comparison of chemical mechanisms appears to show that significant
differences can occur as a result of the different model chemistries.
Three columns of results in table III-l compare the EKMA trajectory models
that differ only in chemistry. The first two of these columns compare the
CBM-EKMA with the propylene/butane-EKMA, which had equal photolysis
rates. The next two columns compare the same propylene/butane chemistry
with and without the adjusted photolysis rates. Figure III-l shows some
of the same comparisons.
Some trajectory runs were also made adjusting only N0£ photolysis (up
by 13 percent), but with standard EKMA aldehyde rates (3 times higher than
used in the CBM). Virtually no difference was found for the peak ozone
values when the results were compared to the fully adjusted photolysis
results, yet the ozone level did begin to rise much earlier in the day.
Some trajectory runs were also tried using higher reactivities for the
propylene/butane emissions mix (propylene percentages up to 35 were
used). Again, the ozone level rose earlier in the day but the maximum
111-24
-------
TABLE I11-11. A COMPARISON OF THE SAI TRAJECTORY MODEL AND THE EKMA:
27 JUNE 1974
(a) NOX Concentrations
Trajectory EKMA
Time Model Model
(PST) (ppm) (ppm)
400 0.043 0.043
500 0.071 0.072
600 0.072 0.073
700 0.055 0.055
800 0.074 0.073
900 0.091 0.089
1000 0.091 0.088
1100 0.087 0.078
1200 0.073 0.071
1300 0.078 0.077
1400 0.085 0.084
1500 0.124 0.122
(b) ETH Concentrations
400 0.0096 0.0096
500 0.0124 0.0118
600 0.0132 0.0124
700 0.0119 0.0116
800 0.0130 0.0132
900 0.0143 0.0144
1000 0.0147 0.0148
1100 0.0142 0.0138
1200 0.0139 0.0125
1300 0.0143 0.0129
1400 0.0148 0.0135
1500 0.0157 0.0139
(c) PAR Concentration
400 0.358 0.358
500 0.442 0.430
600 0.406 0.395
700 0.268 0.265
800 0.304 0.303
900 0.338 0.333
1000 0.340 0.336
1100 0.307 0.300
1200 0.284 0.280
1300 0.298 0.291
1400 0.320 0.314
1500 0.337 0.329
111-25
-------
ozone level was not significantly affected. Evidently the CBM chemistry
produces more ozone for amounts of emitted HC late in the day than does
the propylene/ butane chemistry.
Comparison of either the CBM-EKMA or models using propylene/butane
chemistry with observed data would not provide an appropriate assessment
of the quality of the chemical mechanisms. Mechanism comparisons must be
done at many levels and should include these considerations:
> How well do the basic reactions of each mechanism compare
with known and evaluated reactions? For example, the
propylene/butane mechanism uses a value for the rate
constant of the H02 reaction with NO that is low compared
with recent evaluations.
> How important are the reactions that are not updated? For
instance, it was found that updating the reaction rate
constant for the HC^ reaction with NO in the
propylene/butane mechanism tended to remove most of the
difference between its results and those obtained with the
CBM-EKMA for the late afternoon period discussed earlier.
> How well do the mechanisms simulate a set of smog chamber
experiments covering a wide range of concentrations and
HC/NOX ratios?
> How well do the mechanisms treat the conversion of
atmospheric species into the surrogate species utilized in
the mechanisms?
> How well do the mechanisms respond to differences in
reactivity?
Once a mechanism has been put into an atmospheric model and used to
simulate a specific situation, a large number of elements are introduced
that can either compensate or add to deficiencies in the chemistry.
Therefore, a comparison of mechanisms with observed data in such cases is
inappropriate except as an additional element to be considered along with
the comparisons just discussed.
More investigative work in the area of mechanism comparison is
needed; however, other consistent explanations exist, in addition to the
reaction updating just mentioned, that might eliminate the ozone peak
discrepancy between the propylene/butane mechanism and the CBM. For
instance, the CBM chemistry has a more balanced distribution of reactivity
that the propylene/butane mix cannot simulate. When the propylene used in
111-26
-------
the standard EKMA has virtually all reacted, the remaining butane is too
unreactive to maintain a significant rate of ozone buildup. The detailed
computer outputs showed that the propylene fraction remaining often fell
below 5 percent between 1200 and 1500 hours, meaning that 95 percent of
the remaining HC consists of butane. The CBM, however, uses additional
species, with intermediate reactivity such as ethylene and aromatics that
provide a more constant rate of ozone buildup.
The effect observed in this study does not appear to be as simple as
either of the explanations given because dilution, continuous emissions,
HC/NOX ratio, and overall concentration all seem to interact differently
with the two mechanisms. For instance, the 1982 results are much closer
in agreement for the three pertinent columns of table III-l (columns 4, 5,
and 6) than for the base-case years (1974 and 1975). In addition to the
implication of involvement of HC/NOX ratios and concentrations, the
control-strategy predictions for the two chemistries can obviously be
different. This point will be discussed later.
D. COMPARISON OF LINEAR AND EXPONENTIAL MIXING-HEIGHT RESULTS
Study results obtained from a comparison of dilution determined from
linear interpolation of hourly mixing-height input data, and from exponen-
tial interpolation of only two mixing heights are not significantly
different. For this particular series of trajectories the temporal
behavior of the mixing heights closely resembled the exponential curves
implicit in the regular EKMA. The most significant difference in the
mixing-height comparison columns of table III-l occurred for the trajec-
tory of 27 June 1982 in which the standard model, using the exponential
curve, predicted 0.07 ppm ozone compared to a prediction of 0.11 ppm ozone
for the more sophisticated model, which used hourly mixing-height
inputs. Figure III-8 shows the temporal behavior of mixing heights for
this case.
E. COMPARISON OF THE CITY-SPECIFIC EKMA WITH THE STANDARD EKMA RESULTS
The results of the final comparison shown in table III-2 contrast the
city-specific model with the model used to generate the standard EKMA
isopleth diagram. This comparison involves several differences in model
input data. In general, initial aldehyde and ozone concentrations,
pollutants aloft, post-0800 emissions, and initial N02/NOX ratios all
favor ozone production for the city-specific model in contrast to the
standard EKMA isopleth version. However, similarities and differences in
the results of these two versions are most easily explained by differences
in the mixing heights used by the city-specific model. The standard EKMA
111-27
-------
IOC
800
900 100U
1100 1200
Time (PST)
1300
1400 7500
FIGURE III-8. COMPARISON BETWEEN CONSTANT DILUTION AND VARIABLE
DILUTION RATE FOR 27 JUNE 1982
111-28
-------
isopleth assumes a mixing-height change of 24 percent, whereas the city-
specific simulations use changes of 58 to 365 percent. The city-specific
model allows mixing-height inputs that correspond to morning and afternoon
levels. Hence the city-specific model predicts higher ozone levels
because of the many effects noted above than does the standard model if
the mixing-height changes are relatively small through the day.
For 4 August 1982, the mixing-height change was the smallest for this
series, giving a predicted ozone level of 0.20 ppm, compared to 0.13 ppm
for the standard EKMA. For this case, when the change in mixing height
was large, the city-specific model actually predicted lower ozone concen-
trations than did the standard model. For the 27 June 1974 trajectory,
the change in mixing height was 265 percent, giving a predicted ozone
concentration of 0.21 ppm, compared to the standard model ozone prediction
of 0.24 ppm. This mixing-height explanation is not complete, however.
The comparison for 27 June 1982 is more appropriately explained by the
ozone concentration aloft because the mixing-height change was very large
(365 percent); yet the two models predicted similar ozone levels near 0.06
ppm--the ozone concentration aloft in the city-specific model was 0.081
ppm for this case.
111-29
-------
111-30
-------
IV THE USE OF ISOPLETH DIAGRAMS TO PREDICT OZONE LEVELS
The EKMA uses ozone isopleth diagrams according to a specific set of
guidelines to estimate control scenarios that should produce ozone
reductions sufficient to meet federal or state standards. The isopleth
diagrams themselves are constructed according to a set of assumptions by
computer simulation of many individual precursor-emissions combinations,
using the same trajectory model and trajectory path.
In the previous sections we compared an EKMA trajectory model to the
more complex SAI trajectory model and to the SAI Urban Airshed Model.
Using computer programs OZIPP and OZIPM, it is possible to construct ozone
isopleth diagrams from many repeated individual applications of any of the
models previously discussed. In this section we explore some of the
assumptions used in the construction and application of isopleth diagrams
in the EKMA. In addition, we offer some suggestions that may improve the
reliability of this modeling approach.
Isopleths based on EKMA trajectory models create some uncertainties
when used to predict the maximum observed ozone levels that result from
changes in precursor emissions. These uncertainties are inherent either
in the nature of the models themselves or in the assumptions necessary to
construct the isopleth diagrams. Some of the factors involved are:
> Fixed monitoring sites—As discussed earlier, the time and
location of maximum observed ozone concentrations might
change as a result of a change in emissions. Unless the
wind direction is constant, air arriving at a monitoring
site at different times of the day can originate from
quite different locations. Thus, the trajectory path to a
monitoring site may be much different at different
times. Often, some areas of an urban airshed have
different HC/NOX ratios and concentrations from those of
the trajectory path that lead to the site of maximum
observed ozone. Application of even a uniform emissions
control policy could conceivably lead to different degrees
of improvement (if not deterioration) in other areas of
the airshed. Therefore, the location of the monitoring
site at which maximum observed ozone occurs may vary in
future years.
IV-1
-------
Uneven changes in emissions--Mobi1e-source emissions are,
for the most part, regulated at the federal level, yet
stationary sources are usually subject to local control.
Therefore, the uniform control of either HC or NOX is
rarely even in time and space. Although individual
trajectory-model simulations might provide acceptable
levels of accuracy for the base-case year and the future
year, the isopleth approach is not readily adaptable to
nonuniform emission changes.
Under the present algorithms, each isopleth diagram must
use the same emission pattern for the entire diagram.
However, the impact of alternate emission patterns can be
estimated by changing the pattern and generating a new set
of isopleths. Thus, non-uniform emission changes could be
handled through an iterative process in which the amount
of necessary average control is first estimated on the
original diagram. Then, a new diagram is constructed on
the basis of the pattern projected for realistic implemen-
tation of the first estimate. Next, any corrections
resulting from the changes in the emission pattern are
reintroduced on the original diagram. Finally, a third
diagram could be constructed if the expected emission
pattern was significantly different from that used in the
second diagram. The process would then be repeated until
the combined precursor reductions and emission pattern
changes resulted in the desired ozone level.
Another method for dealing with the problem of interaction
between federal mobile-source control and local station-
ary-source control has been suggested. This approach
first involves the construction of an isopleth diagram in
which every point has an equal contribution of stationary
sources in the absolute sense. Such a diagram would
supposedly characterize the response to mobile-source
controls only; then, an estimate of changes in vehicle-
mile reductions and mobile emission reductions for some
future year should produce an estimated future ozone level
on the diagram using the normal EKMA method. Next, a
second diagram would be constructed in which every point
corresponds to the estimated future mobile emissions; this
second diagram would supposedly characterize the response
to stationary-source controls. The local air pollution
management district could then base their local control
estimates on this second isopleth diagram.
IV-2
-------
> Transported pol1utants--As seen in this study, assumptions
that are used for pollutant concentrations aloft, or that
are found in the initial conditions, can significantly
affect simulated maximum ozone results based on future
emissions scenarios. The grid-model simulations used in
this study were purposely intended to be conservative
estimates of air quality improvement. Hence the pollutant
levels in the air aloft were not rolled back with emis-
sions reductions. The initial (i.e., 400 PST) conditions
were also not rolled back in the 27 June 1982 simulation.
A key feature of the EKMA when isopleth diagams are involved is the
relative nature of this approach. Whether or not the trajectory model
involved simulates the absolute ozone concentration at the point of the
isopleth where one begins is not important to the use of the EKMA. To be
sure, a gross lack of agreement in the absolute sense would be of primary
concern regarding the ability of the combination of model and input data
to appropriately simulate relevant atmospheric conditions. Nevertheless,
the inherent "calibration" or adjustment involved when the model is fixed
at the starting point on an isopleth diagram eliminates many minor
problems associated with atmospheric monitoring and modeling. This simple
and straightforward adjustment feature of the EKMA provides a distinct
advantage over a complex airshed model in predicting control
implications. When complex airshed models don't quite agree with the base
case air quality data, it is difficult to predict future air quality.
This fact is by no means an indication of the quality or appropriateness
of these two types of models. The complex models have many adjustments in
their input files that can be justified within some range, and straight-
forward calibration procedures have yet to be established.
A. COMPARISONS OF ISOPLETH PREDICTIONS
In this section, predictions obtained using the EKMA isopleth
diagrams are compared with the predictions based on the SAI grid and
trajectory models. Since the isopleth method normally involves predic-
tions that utilize base case observations, the observed data is also
used. However, appropriate comparisons, starting with observed data, can
be made only between the various EKMA-type models because the SAI grid and
trajectory models do not have the calibration feature of the isopleth
approach.
The emission files used in the Airshed Model study form the basis for
the percentage emission changes shown in table IV-1. Those trajectories
discussed in section III were used. However, as shown in table IV-1, a
IV-3
-------
TABLE IV-1. PERCENTAGE OF EMISSION CHANGES
Total Area Specific Trajectory
Date HC NOX HC NOX
June 26
74 + 82 -47 -21 -48 -17
June 27
74 * 82 -44 -19 -49 -15
August 4
75 - 82 -47 -20 -50 -14
June 26
82 - 74 +87 +27 +84 +20
June 27
82 - 74 +79 +23 +98 +14
August 4
82 + 75 +89 +25 +99 +24
IV-4
-------
new comparison, based on a total emission inventory for the entire gridded
area, is also included because percentage changes in emissions for the
total area are commonly used with the EKMA. Although the specific
trajectory paths are the same as those discussed in the previous sections,
the changes in emissions shown in table IV-1 are percentage changes for
the total trajectory. In work described in the previous sections, the
1982 initial conditions and hourly emission inputs were taken from the
model used for specific comparison, as just discussed. For the isopleth
predictions, the initial conditions, reactivity, and diurnal changes in
emissions must always be derived from the base case values unless separate
isopleth diagrams are used for the predicted values. Hence, those
percentage changes shown in table IV-1 that are based on 1982 data imply
that the base case values for trajectory reactivity and diurnal emissions
also came from the 1982 specific trajectory model.
The order of discussion of the EKMA models does not follow the order
of the model series presented earlier. This change was made because the
modified EKMA-type models were used with a different style of isopleth
diagram that is more appropriately introduced later in this section. The
isopleths generated for the regular city-specific EKMA model, using the
unmodified OZIPP computer code, are presented in figures IV-1 through IV-6
for the six specific trajectories used in this study. Figure IV-7 is the
so-called "standard" EKMA or Level IV diagram. In addition to the
isopleth lines generated by the OZIPP computer code, figures IV-1 through
IV-7 contain several lines and special points of interest that are
explained below.
The isopleth diagrams were used in several different ways to allow
comparisons among the various models used in this study. In the previous
sections the model comparisons involved steps in a series. Thus, a given
model could be properly compared only with adjacent models in the
series. However, the calibration feature of the isopleth diagrams as used
in EKMA enables us to calibrate any one model to observed data or to any
other model. On the other hand, prediction comparisons of specific
calculations between series-adjacent models in section III involved
parameter changes between base and predicted simulations that cannot
readily be incorporated into single-isopleth diagrams. As discussed
previously, examples of these parameter changes are reactivity and diurnal
emissions patterns. When a grid model is compared with a trajectory
model, the change could be in the trajectory path itself if maximum ozone
predictions form the bases of comparison.
Figures IV-1 through IV-6 typically show two HC/NOX ratio lines; 3
points labeled 1,2, and PREDICTED 03; an AIRSHED-PREDICTED 03 point along
each HC/NOX ratio line; and an OBSERVED 03 point on the solid HC/NOX ratio
line. Figure IV-7 is divided into three parts; each part has points
labeled 1 to 6 and three ratio lines on which the design 03 is indicated.
IV-5
-------
*Z'0 OZ'O 9PO
a:
o
o
LJ
O.
I
a.
o.
(XJ
o
-------
09'0
en
c\j
Di
O
LJ_
UJ
UJ
O.
oo
0.
o.
I—I
r-j
O
O
ce.
«t
o
cn
ooJcP
IV-7
-------
ID
I
o:
O
UJ
o
UJ
Q.
O.
O.
O
a:
CD
zro oro
ao*o
Wdd'XBN
IV-8
-------
VZ'O
OZ'Q 91*0 Zl'O
80'0 WO
CM
CO
CT>
<£>
CM
o:
O
D.
oo
I
o.
o.
o
Qi
83*0
OZ'O
arc zro
Wdd'X0N
80^0
IV-9
-------
09'0
OS'O
OKO
0£'Q
OZ'O
oro
OS'O
Ofr'O 0£'0
Wdd'XBN
OZ'O
oro
00-0°
CM
CO
(NJ
o:
o
u_
c_>
LU
Cu
I
D.
a.
o
o
to
ID
I
o;
o
IV-10
-------
90*0
icro
00
rs
QC
O
UJ
o
o
LU
CL
to
I
D.
O.
O
o
1/1
vo
I
cc:
=>
C3
rO'Ci £0*0
I-MJ •:,;";
IV-11
-------
OE'O SZ'O
OZ-0 Sl'O
j I
•o
O
4->
O)
en
c
»^-
t/i
3
C
O
u
•r*
TJ
CD
O.
O
O
x
Q_
O
CO
<:
CO
r*».
i
~. LU LU CO UJ UJ CO
1 I Z Z => Z 2 =>
L: \
vO
CM CM
tfl
=5
*£.
=5 CS
.
vD r-*
- CM CM
• » I I ' '
r- CM CO «S- Lf> VO
n- '
01 "0
sro
or '0
00 "0"
IV-12
-------
sz'o oz'o sro oro so'o
IT) CM
CM CM co
t>. cyi co oo cy>
UJ LU z z =)
O =>
=5
-------
oe
[•0 nt'O GO'O C.3'0
•o
o
o
a.
0)
o
•o
0)
in
(O
o
O
o
•a
a>
u
o
o
LU
o:
ST'O
CI'O
IV-14
-------
Method 1 indicates how EKMA performance might compare with that of
the Airshed Model if generalized (i.e., area-wide) controls were called
for with EKMA, but in practice the controls were not distributed evenly.
The "generalized" control was derived by comparing the total gridded
emissions inventories for the region used by the Airshed Model in the base
case and 1982 simulations. To calibrate EKMA to the Airshed Model, the
design 03 was taken from the highest 03 simulated in the Airshed Model (at
a monitoring site), and the HC/NOX ratio was derived from averaging the
simulated airshed precursor ratios from 0600 to 0900 PDT at the downtown
Los Angeles monitoring site. Figures IV-1 through IV-6 illustrate the
design 03 values as AIRSHED-PREDICTED 03; the HC/NOX ratios are
represented by a solid line in each figure; and the results of applying
this method are shown as the points labeled 1. Each of the six figures
(IV-I through IV-6) represents a trajectory-specific isopleth based on the
city-specific calculations described in section III. Figure IV-7a (Level
IV EKMA) shows the same set of design 03 values and HC/NOX ratio lines as
figures IV-1 through IV-6, but the results for the six specific
trajectories are labeled 1 through 6 in figure IV-7. All the design 03
and relevant HC/NOX ratios are also presented in table IV-2. The EKMA
isopleth results are compared with the Airshed Model predictions in table
IV-3.
Method 2 shows how the EKMA might perform compared with Airshed Model
predictions when trajectory-specific emissions changes are applied. To
calibrate the EKMA to the Airshed Model, the design 03 values used in
method 1 are used, but the HC/NOX ratios are derived from the 0600 to 0900
PDT average HC and NOX concentrations found in the SAI Trajectory Model
simulations. The HC/NOX ratios represent a trajectory-specific ratio that
may be very different from the HC/NOX ratios defined in method 1. The
HC/NOX ratios are shown in figures IV-1 to IV-7 as a dashed line. The
results from applying method 2 are shown as points labeled 2 in figures
IV-1 to IV-6. Figure IV-7(b) shows the same 03 design values and HC/NOX
ratios used in method 2 on a Level IV EKMA and the results from all six
emissions changes.
Method 3 (prediction method) shows the EKMA predictions using design
03 values from monitoring station observations. To apply this method, we
used the design HC/NOX ratios from method 1 together with the total area
emission changes. Design 03 values are shown in figures IV-1 to IV-3 as
OBSERVED 03 on the HC/NOX ratio (represented by a solid line). Results
from applying the total area emission changes are shown as points labeled
PREDICTED 03 in figures IV-1 to IV-3. Figure IV-7(c) shows the prediction
method results for the three days.
The results obtained with this set of isopleth diagrams can be
presented in several ways. Previous model comparisons involved steps in a
IV-15
-------
TABLE IV-2. DESIGN 03 AND HC/NOX RATIOS USED
IN THE EKMA PREDICTIONS
(a) Method (1):
> Total area emission changes
> Downtown Los Angeles 6:00 to 9:00 a.m.
HC/NOX from Airshed Model
> Design 03 from Airshed Model
Date Design 03 HC/NO
x
26 June 1974 0.20 6.69
27 June 1974 0.32 8.92
4 August 1975 0.20 9.39
26 June 1982 0.29 6.69
27 June 1982 0.32 8.92
4 August 1982 0.23 9.39
(b) Method (2):
> Trajectory-specific emission changes
> Trajectory-specific 6:00 to 9:00 a.m. HC/NOX
> Design 03 from Airshed Model
Date Design 03 HC/NOX
9.52
4.72
9.72
9.52
4.72
9.72
26 June 1974
27 June 1974
4 August 1975
26 June 1982
27 June 1982
4 August 1982
(c) Prediction
0.20
0.32
0.20
0.29
0.32
0.23
Method:
> Total area emission i
> Downtown
Los Angeles
HC/NOX from Airshed Model
> Observed design 03
Date Design 03 HC/NO
x
26 June 1974 0.34 6.69
27 June 1974 0.49 8.92
4 August 1975 0.32 9.39
IV- 16
-------
TABLE IV-3. COMPARISON OF OZONE PREDICTIONS FOR THE CITY-SPECIFIC EKMA,
THE STANDARD-CONDITIONS (LEVEL IV) EKMA, AND THE SAI
AIRSHED MODEL
(a) Method (1):
> Total area emission changes
> Downtown Los Angeles 6:00 to 9:00 a.m.
HC/NOX from Airshed Model
> Design 0-j from Airshed Model
Date
26 June 1974
27 June 1974
4 August 1975
26 June 1982
27 June 1982
4 August 1982
City-Specific
EKMA
0.12
0.24
0.14
0.39
0.42
0.31
#3
-40
-25
-30
+35
+31
+35
Level IV
EKMA
0.13
0.25
0.15
0.37
0.39
0.29
#3
-35
-22
-25
+28
+22
+26
Whole
Region
0.29
0.32
0.23
0.20
0.32
0.20
Specific
Trajectory
0.13
0.27
0.11
0.31
0.25
0.24
(b) Method (2):
> Trajectory-specific emission changes
> Trajectory-specific 6:00 to 9:00 a.m. HC/NOx
> Design 0-j from Airshed Model
Date
26 June 1974
27 June 1974
4 August 1975
26 June 1982
27 June 1982
4 August 1982
City- Specific
EKMA
0.14
0.14
0.14
0.35
—
0.31
£>3
-30
-56
-30
+21
__
+35
Level IV
EKMA
0.14
0.14
0.14
0.35
0.43
0.29
/o3
-30
-56
-30
+21
+34
+26
Whole
Region
0.29
0.32
0.23
0.20
0.32
0.20
Specific
Trajectory
0.13
0.27
0.11
0.31
0.25
0.24
IV-17
-------
TABLE IV-3 (Concluded)
(c) Prediction Method:
> Total area emission changes
> Downtown Los Angeles 6:00 to 9:00 a.m. HC/NOX
from Airshed Model
> Observed design 0,
Date
26 June 1974
27 June 1974
4 August 1975
City-Specific
EKMA
0.19
0.36
0.22
A03*
-44
-27
-31
Level IV EKMA
0.24
0.39
0.25
-29
-20
-22
Percentage change in ozone =
(Predicted 0 ) - (Design 0 )
(Design 0 )
x 100
IV-18
-------
series; the isopleth approach allows a "recalibration" back to the point
on which one wishes to base the relative comparison. On the other hand,
the predictions of the more sophisticated models involve changes in
reactivity, diurnal emission patterns, and even the trajectory path
itself, which predictions based on single isopleth diagrams cannot
simulate.
The three methods of comparison shown in table IV-2 are based on the
values of maximum ozone concentrations found in the SAI airshed
simulations. For the base case days (1974 and 1975) we used the maximum
ozone concentrations found in the SAI airshed simulations at the
monitoring sites (station maximum ozone). For the future-year days (1982)
we used the maximum ozone concentration found anywhere in the modeling
region (regional maximum ozone) in the SAI airshed simulations. Often the
regional maximum ozone value is greater than the maximum ozone value at
the station due to the continuation of the chemistry and transport. This
is especially true when station maximum ozone observations occur early in
the afternoon. Thus, the design 0-j values for 1982 shown in table IV-2
are higher than those for 1974 and 1975.
The Airshed Model "predictions" included in table IV-3 are repeated
from table1 III-l; for the changes across the entire modeling region, the
Airshed Model numbers are actually the transposed design 03 values found
in table IV-2. The results presented in table IV-3 show that on a
relative basis the EKMA tends to be more sensitive to emission changes
than the Airshed Model over the entire basin, even though the changes
predicted along specific trajectories are similar. A perplexing exception
to this similarity is seen in the method (2) comparison based on the
27 June 1974 trajectory. For this case the design HC/NOX and emissions
changes were made for the specific trajectory; however, agreement with the
Airshed Model was much better when basin-wide emissions and the downtown
HC/NOX ratio were used. The 27 June 1982 trajectory-specific comparison
[method (2)] disclosed another problem; the specific HC/NOX ratio of 4.72
was so low that the design ozone level could never be reached. A similar
situation was reported by Feldstein et al. (1979) for the San Francisco
area.
The second type of comparison involves the use of EKMA diagrams to
predict future observations based on past observations. In this case,
comparison of the SAI grid and trajectory models is not as appropriate
since these models are not "calibrated" to past observations. For this
comparison, the model series is used and the predictions of the various
EKMA type models are of interest. The Airshed Model base case HC/NOX
ratios were used for the city-specific and Level IV predictions shown on
the isopleth diagrams by PREDICTED 03. These HC/NOX ratios were used
because the observed ratios have been a subject of controversy and have
IV-19
-------
changed because of recalibration. The ratios generated by the emission
inventory in the grid model for the downtown area provide a self-
consistent reference across the model series for this study. These
relative predictions, based on observations, are shown in table IV-3,
which is similar to table III-l showing the absolute predictions.
In the third method of comparison, the base case ozone values and the
HC/NOX ratios for the city-specific and Level IV predictions were taken
from the absolute prediction study. Thus, the predictions were based on
the 0800 PDT HC and NOX values generated by the more sophisticated model
in the series, namely, the model with a modified mixing-height algorithm
started at 0500 PDT hours. Since both the city-specific and Level IV
models start at 0800, the initial HC/NOX values are the same. These
predictions are also shown in table IV-4; the predictions are shown
graphically on the isopleth diagrams by (3) in figures IV-8 through IV-10
for the city-specific case and figure IV-11 for the standard Level IV
diagram. This relative prediction comparison is discussed next with the
presentation of the modified isopleths. Table IV-4 shows that the city-
specific EKMA and Level IV EKMA models predict identical concentrations
for 1982 using the 1974 trajectories; however the city-specific EKMA tends
to predict lower ozone concentrations for 1982 than the Level IV EKMA when
the 4 August 1975 trajectory is used.
B. A NEW TYPE OF ISOPLETH DIAGRAM
In the past, the EKMA has used an isopleth diagram based on initial
concentrations of HC and NOX during the period from 6:00 to 9:00 a.m.,
more specifically, those at 0800 hours. When the regular EKMA isopleth
diagrams are used to estimate changes in maximum ozone concentrations
resulting from changes in HC and NOX emissions, the observed 6:00 to 9:00
a.m. HC/NOX ratio is used with the maximum observed ozone concentration to
locate a starting point on the isopleth diagram. Relative changes in the
HC and NOX values making up the ordinate and abscissa of the isopleth
diagram are equated to relative changes in emissions to find the predicted
ozone value. Alternatively, the relative changes in the HC and NOX values
necessary to reach some desired ozone value are equated to the necessary
relative changes in emissions.
This section describes a new method of locating the starting point on
a new isopleth diagram that is based on the relative emission inventories
of HC and NOX. There are several reasons for introducing a new type of
isopleth diagram:
IV-20
-------
TABLE IV-4. A COMPARISON OF 1982 OZONE PREDICTIONS FOR THE
ISOPLETH MODELS (OZONE CONCENTRATIONS IN PPM)
SAI City-Specific Level IV
Date Trajectory CBM-EKMA EKMA I la EKMA EKMA
26 June 1974
design 03 = 0.30 0.21 0.18 0.13 0.20 0.22
design 03 = 0.19 — -- 0.11 0.14 0.14
design 03 = 0.20 — — — 0.14 0.14
27 June 1974
design 03 = 0.32 0.29 0.17 0.19 0.20 0.22
design 03 = 0.30 — 0.18 0.19 0.20 0.20
design 03 = 0.21 — -- 0.13 0.14 0.14
4 August 1975
design 03 = 0.34 0.22 0.22 0.23 0.23 0.26
design 03 = 0.35 — 0.21 0.24 0.24 0.26
design 03 = 0.26 — -- 0.18 0.17 0.20
design 03 = 0.28 -- -- — 0.21 0.21
IV-21
-------
90 0 SO'O fcO'O £0*0
LQ'Q 30'0 SLVQ HI 0 ?0 I.
Q
O
LU
Z
3
T3
o:
O
LU
Q.
Q.
Q.
O
Q
C£
-------
sro zro BQ-Q gcro ECTO
IZ'O
Ch
CM
ex:
O
O
LU
0.
CL.
CL
O
o;
-------
zro oro
80-
zro oro
80'0 9Q-Q
Wdd'XCN
frO'O
00 '0°
o
o
!-» O
eC
LU
CL
oo
Q.
Q.
O
O
cr
Z3
C3
IV-24
-------
OE'O SZ'O
oo
UJ
oo
oo
o
o
LU
oo
o
g
o
ac
UJ
O.
O
oo
IP—4
UJ
O
O
o
Q
5
oo
IM-25
-------
> The ambient HC/NOX ratio is often difficult to measure
accurately. (Specifically, the HC values are the most
difficult to measure.)
> Even accurate measurements often show a large degree of
scatter from hour to hour, day to day, or from location to
location, within the same urban area. Expression of an
appropriate average thus becomes difficult and uncertain.
> The observed ratios of 6:00 to 9:00 a.m. HC/NOX concentra-
tions rarely agree with the ratios of HC/NOX emissions,
even though the morning precursors should be essentially
unreacted.
> The observed ratios of 6:00 to 9:00 a.m. HC/NOX concentra-
tions may not agree with simulated ratios found in very
sophisticated atmospheric models. This lack of agreement
is rarely, if ever, used to evaluate the quality or
appropriateness of such models.
> It has never been demonstrated that relative changes in
6:00 to 9:00 a.m. concentrations are a good measure of
relative changes in the emission inventory. On the
contrary, the opposite was shown to be true by Whitten,
Meldgin, and Roth (1977). They reported that a simulated
24 percent reduction in total emissions of HC resulted in
a 41 percent reduction in 6:00 to 9:00 a.m. HC concentra-
tions using the SAI grid model in Denver.
> The present EKMA diagrams do not always relate to a base
case trajectory model simulation that can be used to
validate, or gain some indication of, model performance.
If the EKMA trajectory is dominated by post-0800 emissions
of an HC/NOX ratio that is different from the initial
condition values, then validation is very difficult.
Thus the main feature of the new type of diagram is the elimination
of the 6:00 to 9:00 a.m. HC/NOX dependence. The ordinate and abscissa
become the relative emissions of NOX and HC. To suppress dependence on
the initial conditions, the model starting time is moved sufficiently
forward so that the local emissions significantly dominate the production
of ozone. The justification for an early starting time was given earlier
in the discussion of the CBM-EKMA trajectory model. Initial conditions
are treated only as transported pollutants in the surface layer. The
following list details specific features of the new method:
IV-26
-------
> Ambient HC and NOX measurements are not required.
However, any such available data can be used to validate
or provide some measure of performance for the trajectory
model using the base case emissions inventory, which
defines the (1,1) point of the isopleth diagram.
> Emission inventories, which are required for the new
method, are already required under Section 172.(b) for
nonattainment areas.
> Since the scales of the ordinate and abscissa of the new
diagrams are relative emissions, the predictions of the
new EKMA method relate directly to changes in emissions.
Dependence on initial concentrations is minimized, whereas
dependence on those emissions that are to be controlled is
maximized.
> The line from the origin through the (1,1) point using the
new method becomes the equivalent line to the HC/NOX ratio
line in the original EKMA. That is, the design or staring
point ozone is located along this line. However, this
line has a new feature not readily apparent in a compar-
ison with the original EKMA. The distance between the
(1,1) point and the observed ozone is an instant and clear
measure of performance. Except for an unequal bias of
errors between the HC and NOX inventories, the combined
dispersion, chemistry, and inventory errors should lie
along this line.
Examples of the new type of isopleth diagram are shown in figures IV-
12 through IV-17 for the CBM-EKMA and figures IV-18 through IV-23 for the
EKMA Ha model. The center of the diagram, the (1,1) point, corresponds
to the CBM-EKMA trajectory model runs of table III-l discussed earlier.
The (2,2) point corresponds to a doubling of both HC and NOX emissions.
Initial concentrations and precursor concentrations aloft were rolled back
(or up) linearly according to the location on the isopleth diagram. Ozone
values aloft were adjusted by the following formula:
= °-04
where HC and NOX refer to the location on the isopleth diagram. The use
of the isopleth diagram is similar to the original EKMA version except
IV-27
-------
09 'I
OS11
OZ'l
06'.0 09*0 OE'O
oo
to
o
LjJ
tO
vo
CJ
a:
o
i
I
CO
a:
o
orz oo*i os'i oz*i 06
09 J0 DC "0 00 '0°
IV-28
-------
08"! OS'l OZ't 06^0 09'0 0£ 0
on
ec
o
UJ
CO
rs
CM
QL
O
I—
O
UJ
rs
a:
LU
I
CQ
UJ
a:
orz
09'
OS'I
OZ'I 06'0
Wdd'XBN
IV-29
-------
OS1I OZ'I 06*0 09*0 Qt'Q
on
OO
•a:
CO
i
Lf)
to
O
o:
o
I
CD
o:
3
CD
orz
00J0°
IV-30
-------
08'I
o
o
UJ
-3
I
CNJ
CO
Cl
-3
CM
O
UJ
C£
I
CO
IT)
>-H
I
UJ
o:
orz DO •
os't
OZ'I 06 -
Wdd'XBN
09*0 OE'O 00-0°
IV-31
-------
tn
ID
IN
Ol'Z 09'I OS* I OZ't 06-0 09*0 0£'0 00-C'
oe
LU
CM
00
o
CM
CO
o
UJ
QC
IV-32
-------
08'I OS'I OZ'l 06*0 03'0 0*0
Ol'Z O8't OS't OZ't 06'0 09'0
DC
o
UJ
<~3
<.
cc
O
i
CM
00
O
OC
O
UJ
'O
-------
that the line through the (1,1) point in the new diagram replaces the 6:00
to 9:00 a.m. HC/NOX ratio line of the original diagram. The difference
between the observed maximum ozone and the simulated ozone value at the
(1,1) point represents the cumulative discrepancy between the model and
the observations resulting from inaccurate measurements, an inaccurate
emission inventory, or inadequacies of the model itself.
The various prediction methods discussed earlier were applied to
these new isopleths (figures IV-18 through IV-23) to produce the results
given in tables IV-4 and IV-5. The results in table IV-5 can be compared
with those given in table IV-3. However, the changes in the new isopleths
for transported pollutants tend to generate higher sensitivity to control
than the normal isopleths (and the SAI Airshed Model and trajectory
model), which keep these pollutants constant. Nevertheless, results
obtained with trajectory-specific emission changes [method (2)] show the
greatest sensitivity to the specific trajectory path using the new
isopleths when compared with the normal isopleths, which all predict 0.14
ppm using the base case (1974 and 1975) trajectories. The CBM-EKMA
predicts slightly lower ozone concentrations than any of the other
versions of EKMA. The predictions for 27 June 1982 show the greatest
variance among the models; the new isopleth versions show significantly
less predicted ozone.
It should be noted that the 27 June 1982 base case isopleths in both
the regular and new versions have a rather unusual general shape that was
reported earlier by Whitten and Hogo (1978). In this report, it was shown
that such shapes are caused by two maxima that occur during the simula-
tions. The OZIPP and OZIPM codes pick the highest of the two peaks so
that the resultant diagram is really a composite of two more normal
diagrams that would plot only one peak. A more serious problem with some
isopleths based on this 27 June 1982 trajectory involved the use of design
ozone values that could not be located on the diagram. This problem was
caused by the gross difference, originally seen in table III-l, between
the SAI Airshed Model and trajectory model results (0.32 ppm and 0.12 ppm,
respectively) for this trajectory. Hence, the "calibration" of the
relative isopleth method was too severe when it was necessary to extend it
to a model that, on an absolute scale, predicted much lower levels. The
EKMA-style models all originally simulated rather low ozone levels, as did
the SAI trajectory model.
The rollback of transported pollutants cannot totally explain the
gross difference between the 1982 CBM-EKMA prediction and the SAI trajec-
tory model prediction based on the 27 June 1974 data shown in a comparison
of the results in table IV-4. As originally shown in table III-l, the
CBM-EKMA used the specific trajectory information along the same 1974 path
from the SAI trajectory model outputs using the 1982 emissions. In table
IV-34
-------
-3
VO
Qi
o
"3
CO
a:
OI'Z O8't OS*l OZ'l 06*0 09-Q 0£*0 00*cP
IV-35
-------
08'T OS'l OZM
OE'O
06'0 09'0 OE'O 00'.
o
LU
o->
•"3
r^.
CSJ
DC
O
o
LU
•~3
<:
cr
DC.
O
DTZ OB'l
-------
t/1
LU
t/")
CO
I
I
U.O
o
(_3
LU
o
CO
I
Jl'Z 08 'I OS'I OZ'l 06-0 09'0 0£J0 00JcP
IV-37
-------
OO'l OS'l
CD
OTZ OB'l
cc
o
LU
i
i
CM
CO
13
"3
VO
CM
D£.
e
o
UJ
"2
CM
I
-------
O9'l
OS'T
OZ'I
O&'O 09*0
cr
o
UJ
O
I
C\J
co
UJ
•as.
<-3
r**.
CM
Q£
O
CSJ
CM
I
oo * r
IV-39
-------
(N
CO
QC
o
o
i
00
en
rs
«•
a:
o
c\j
i
UJ
C£
C3
orzos Jt os1*! oz*i 06 r'n og^o oe*o oo'cP
IV-40
-------
TABLE IV-5. OZONE PREDICTIONS FROM CBM-EKMA AND EKMA Ha MODELS
(a) Method (1):
> Total area emission changes
> Downtown Los Angeles 6:00 to 9:00 a.m.
HC/NOX from Airshed Model
> Design 0-. from Airshed Model
Date
CBM-EKMA
EKMA Ha
26 June 1974
27 June 1974
4 August 1975
26 June 1982
27 June 1982
4 August 1982
0.12
0.18
0.14
0.55
0.60
0.33
-40
-44
-30
+90
+80
+44
0.12
0.19
0.14
0.53
—
0.31
-40
-41
-30
+83
—
+35
(b) Method (2):
> Trajectory-specific emission changes
> Trajectory-specific 6:00 to 9:00 a.m. HC/NOx
> Design 0, from Airshed Model
Date
CBM-EKMA
to,
EKMA Ila
to,
26 June 1974
27 June 1974
4 August 1975
26 June 1982
17 June 1982
4 August 1982
0.12
0.15
0.13
0.54
0.71
0.33
-40
-53
-35
+86
+122
+43
0.11
0.17
0.13
0.53
—
0.31
-45
-47
-35
+83
—
+35
IV-41
-------
TABLE IV-5 (Concluded)
(c) Prediction Method:
> Total area emission changes
> Downtown Los Angeles 6:00 to 9:00 a.m.
HC/NOX from Airshed Model
> Observed design 0-j
Date
CBM-EKMA
AO/
EKMA Ha
26 June 1974
27 June 1974
4 August 1975
0.18
0.25
0.21
-47
-49
-34
0.19
0.26
0.22
-44
-47
-31
(Predicted 0 ) - (Design 0 )
Percentage change in ozone = • x 100
(Design 0 )
IV-42
-------
III-l, the value shown is 0.27 ppm for the CBM-EKMA compared with the 0.29
ppm value simulated by the SAI trajectory model, yet the isopleth result
using the same trajectory predicts a value of 0.17 ppm. Investigation of
the detailed computer outputs revealed that the primary explanation of
this difference is found in the initial conditions used in the SAI
trajectory model for 1982. These initial conditions were quite high in
concentration for the first two 50-meter levels but the mixing height was
initially only 50 meters. Hence, the rise in mixing height at 0600 (PST)
brought in a significant amount of pollutants. In the original CBM-EKMA
calculation this effect was simulated by a special addition to the
emission vector at that hour (see table III-6).
IV-43
-------
REFERENCES
Anderson, G. E., M. J. Hillyer, and G. Z. Whitten (1978), "Photochemical
Plume Dynamics of NOX," EF78-13, Systems Applications, Incorporated,
San Rafael, California.
Anderson, G. E., et al. (1977), "Air Quality in the Denver Metropolitan
Region: 1974-2000," EPA-908/11-77-002, Systems Applications,
Incorporated, San Rafael, California.
Demerjian, K. L., K. L. Shere, and J. T. Peterson (1979), "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. 9 (Wiley
Interscience, New York, New York, in press).
Feldstein, M., et al. (1979), "Anatomy of an Air Quality Maintenance
Plan." J. Air Pollut. Control Assoc.. Vol. 29, pp. 339-363.
Hayes, S. R. (1978), "Performance Measures for Air Quality Disperson
Models," EF78-93, Systems Applications, Incorporated, San Rafael,
California.
Kill us, J. P., and G. Z. Whitten (1980), "Pollutant Tradeoff Ratios Appro-
priate to the Potrero Power Plant Emissions," EF79-100R, Systems
Applications, Incorporated, San Rafael, California.
Reynolds, S. D., et al. (1979), "Photochemical Modeling of Transportation
Control Strategies—Volume I. Model Development, Performance Evalua-
tion, and Strategy Assessment," DOT-FH-11-8529, Systems Applications,
Incorporated, San Rafael, California.
Reynolds, S. D., et al. (1977), "Continued Research in Mesoscale Air Pol-
lution Simulation Model ing--Volume V. Refinements in Numerical
Analysis, Transport, Chemistry and Pollutants," EF77-142, Systems
Applications, Incorporated, San Rafael, California.
Tesche, T. W. et al. (1979), "Evaluating Simple Oxidant Prediction Methods
Using Complex Photochemical Models," EPA-68-02-2870, Systems
Applications, Incorporated, San Rafael, California.
R-l
-------
Trijonis, J. (1977), "Empirical Studies of Ambient Nitrogen Dioxide Air
Quality and N02 Precursor Relationships," TSC-PD-A160-2, Preliminary
Draft of the Final Report, Technology Service Corporation, Santa
Monica, California.
Whitten, G. Z., M. J. Meldgin, and P. M. Roth (1977), "A Preliminary
Evaluation of the Potential Influence of Varying HC/NOX Ratios on the
Design of Oxidant Control Strategies," EF77-33R, Systems
Applications, Incorporated, San Rafael, California.
Whitten, G. Z., J. P. Kill us, and H. Hogo (1980), "Modeling of Simulated
Photochemical Smog with Kinetic Mechanisms. Vol. I—Final Report,"
EPA-600/3-80-028a, Systems Applications, Incorporated, San Rafael,
California.
Whitten, G. Z., and H. Hogo (1980), unpublished results.
R-2
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-450/4-81-031 d
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Comparative Applications of the EKMA in the
Los Angeles Area
5. REPORT DATE
&
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
G. Z. Whitten and H. Hogo
8. PERFORMING ORGANIZATION REPORT NO.
SAI 10R2-81-EF80-73
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Systems Applications, Incorporated
950 Northgate Drive
San Rafael, California 94903
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-2870
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The ability of seven models, ranging in sophistication from the SAI Urban Airshed
Photochemical Grid Model to a published set of ozone isopleth curves (i.e.,
"standard" or Level IV EKMA) to predict peak ozone is examined for three days
observing high ozone in the Los Angeles Basin. The impact of control strategies
simulated with each of the models is also compared. Impact of specific differences
among models such as different means of considering horizontal and vertical diffusion
and use of different chemical mechanisms is also examined.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Ozone
Photochemical pollutants
Photochemical models
EKMA
Urban Airshed Model
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
Unlimited
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
21. NO. OF PAGES
99
20. SECURITY CLASS (Thispage)
22. PRICE
------- |