United States
Environmental Protection
Agency
Office pf Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-81-005b
May 1981
Air
Application Of The
Empirical Kinetic
Modeling Approach
To Urban Areas
Volume II: Tulsa
-------
EPA-450/4-81-005b
Application Of The Empirical Kinetic
Modeling Approach To Urban Areas
Volume II: Tulsa
by
G.Z. Whitten and H. Hogo
Systems Applications, Incorporated
950 Northgate Drive
San Rafael, California 94903
Contract No. 68-02-3376
EPA Project Officer: Gerald L. Gipson
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
May 1981
-------
DISCLAIMER
This report has been reviewed by the Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the views and
policies of the U.S. Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement or recommendation for use.
i»9r ei 1
-------
ABSTRACT
The EKMA was evaluated using applications in the Tulsa area primarily
by comparing the trajectory model that forms the basis of the EKMA (OZIPP)
with other, more sophisticated, models. The study was carried out at
several levels, beginning with evaluation of OZIPP and ending with an
evaluation of the control-strategy predictions that result from employing
the EKMA isopleth methodology. The OZIPP trajectory model was compared
with the SAI Airshed Model and the SAI trajectory model, as well as with
some modified versions of the original OZIPP model.
The basic OZIPP model is a simple, moving-air parcel, or trajectory,
model that uses a detailed chemical mechanism for the surrogate propylene
and butane hydrocarbons. The model treats time-dependent, precursor-
emission factors along with expansion of the air parcel; entrainment is
treated by assuming that constant concentrations exist outside the parcel.
The primary emphasis of this comparison study was directed toward the
discovery of features in the basic OZIPP model that could explain differ-
ences in the results of the OZIPP model from those of some other model.
si
-------
CONTENTS
Disclaimer ii
Abstract i i i
Figures v
Tables vi
1. Introduction 1
Background and objectives 1
The empirical kinetic modeling approach 2
An overview of EKMA comparison evaluations 3
2. Summary, Conclusions, and Recommendations 5
Summary 5
Conclusions 5
Recommendations , 9
3. Model Applications 14
Definition of the models 15
Description of the Tulsa modeling region 23
Characterization of the Tulsa modeling days 23
Inputs used in the SAI Airshed Model simulations 29
Inputs used in the SAI trajectory model simulations 30
Inputs used in the Level II CBM/OZIPM and Level II
OZIPM model simulations 40
Inputs used in the Level III EKMA model simulations 49
4. Comparison of Model Results 60
SAI Airshed Model results 60
Comparison of SAI Airshed and SAI trajectory
model results 65
Comparison of SAI trajectory and Level II EKMA
model results 68
Level III EKMA model results 78
5. EKMA Isopleth Calculations 92
6. Sensitivity Studies 112
Sensitivity of the EKMA Model to emissions density
information 112
Sensitivity of the propylene/butane chemistry due
to variation in rate constants 114
Sensitivity of models to hydrocarbon and NOX
background concentrations 118
Sensitivity of Level II EKMA models to trajectory
selection involving wind shear 127
References 132
Appendix: Description •" Inputs Used in SAI Airshed Model
Simul at * ->ns 134
81 i 1v
-------
FIGURES
Number Page
3-1 The gridded Tulsa modeling region 24
3-2 Initial conditions in the ground-level grid cells on
3 August 1977 (isopleths in increments of 2 pphm) 31
3-3 Initial conditions in the ground-level grid cells
on 21 July 1977 33
3-4 The diurnal pattern of the mixing depth and top of the
modeling region 36
3-5 Trajectory path (CDT) 41
3-6 Level III straight-line trajectory to maximum 03 51
4-1 Comparison of NOX, PAR, ETH, BZA, PAN, and ozone for the
SAI trajectory and CBM/OZIPM models for 29 July 1977 71
4-2 Comparison between station-observed ozone and
CBM/OZIPM predictions 74
4-3 Level II EKMA trajectory for Tulsa, Oklahoma 80
4-4 Level III EKMA trajectory for Tulsa, Oklahoma 84
4-5 Level III EKMA trajectory for Tulsa, Oklahoma, using
the Carbon-Bond Mechanism 88
5-1 Level III trajectory for Tulsa, Oklahoma 97
5-2 Level III trajectory for Tulsa, Oklahoma, using the
Carbon-Bond Mechanism 103
6-1 Trajectory paths used in the EKMA sensitivity study of
grid resolution for 21 July 1977 113
si i
-------
6-2 CBM/EKMA Oklahoma simulation of 20 July 1987--trajectory
to the regional ozone maximum 119
6-3 Typical transport flight pattern 128
6-4 Comparison of trajectory paths based on two different
wind patterns for 29 July 1977 130
6-5 Total ground-level emissions of hydrocarbons for the
Tulsa region in kg/hr between the hours of 0800
and 0900 (CDT) 131
»*9r si i V1-
-------
TABLES
Number Page
3-1 Maximum Observed Ozone Concentration for the Four
Modeling Days in 1977 25
3-2 Boundary Concentrations Used in the Tulsa Simulations
(ppm) 35
3-3 Initial Conditions and Emissions Rates Used in the
Level II CBM/OZIPM Calculations for 21 July 1977 45
3-4 Initial Conditions and Emissions Rates Used in the
CBM/OZIPM Calculations for 29 July 1977 46
3-5 Initial Conditions and Emissions Rates Used in the
CBM/OZIPM Calculations for 3 August 1977 47
3-6 Initial Conditions and Emissions Rates Used in the
CBM/OZIPM Calculations for 2 September 1977 48
3-7 County-wide Emissions Estimates for 1977 50
3-8 Calculated Emissions Density for Five Oklahoma Cities 50
3-9 Initial Conditions Used in the Level III EKMA
Calculations for 21 July 1977 55
3-10 Initial Conditions Used in the Level III EKMA
Calculations for 29 July 1977 56
3-11 Initial Conditions Used in the Level III EKMA
Calculations for 3 August 1977 57
3-12 Initial Conditions Used in the Level III EKMA
Calculations for 2 September 1977 58
81
-------
4-1 Comparison of Ozone Predictions for the SAI Airshed,
SAI Trajectory, Level II, and Level III EKMA
Model S--1977 61
4-2 Summary of Peak Ozone Concentration Model Performance
Evaluation Measures 62
4-3 Summary of SAI Airshed Model Performance Evaluation
Measures 66
4-4 Aloft Concentrations of BZA and PAN Used in the SAI
Trajectory Model and the CBM/OZIPM Model 70
4-5 Comparison of SAI Trajectory and the CBM/OZIPM
Model Ozone Prediction 79
5-1 Emission Inventory Summary for the Tulsa Study Area 93
5-2 Control Strategies to Be Used in the Tulsa Modeling
Study 94
5-3 Ozone Predictions from EKMA Model Single-Day
Diagrams (ppm) 95
5-4 Ozone Predictions from Various Models for the Tulsa
Modeling Days (lowered ozone aloft in the EKMA
models) (ppm) 109
5-5 Required Percentage Reductions in Hydrocarbon
Predicted by the EKMA to Reach 0.12 (ppm) Ozone
Levels (3.7 percent NOX reduction assumed) 110
6-1 Emissions Rates for Trajectories Parallel to the One
to Vera for 21 July 1977 115
6-2 Initial Conditions Used in the EKMA Calculations
for 21 July 1977 116
6-3 Level II EKMA Ozone Predictions Using the Propylene/Butane
Chemistry for 21 July 1977 under Various Conditions
(observed ozone = 0.14 ppm) 117
6-4 SAI Trajectory Model Results Obtained by Varying
Hydrocarbon and NOX Concentrations in the Surface
and Aloft Layers 122
si i Vl11
-------
6-5 Effect of Precursor Conditions on Ozone Predictions
for 29 July 1977 for the CBM/OZIPM and Level III
EKMA Models 124
6-6 Calculation of Contribution Factors of Transported
Pollutants to Urban Levels--1977 125
6-7 Aircraft Measurements of NMOC 126
81
-------
1 INTRODUCTION
BACKGROUND AND OBJECTIVES
For several years the Office of Research and Development of the EPA
has sponsored a coordinated program consisting of fundamental chemical
research, smog chamber experimentation, and model development. The
objective of this program is to develop mathematical models capable of
simulating the dynamics of the chemical reactions and dispersion of the
gaseous pollutants found in the lower troposphere. These atmospheric
models are to be used both to gain an understanding of air pollution and
to predict the probable outcome of control scenarios. To this end, the
EPA Office of Air Quality Planning and Standards (OAQPS) has embarked on
an extensive program of model comparison and evaluation.
The specific purpose of the OAQPS program is to assess the suita-
bility of available photochemical oxidant modeling approaches for use by
states as planning tools in ozone air quality planning and management.
The program includes comparisons and evaluations of the various approaches
for typical U.S. cities. As part of the program, large, detailed data
bases are being assembled for several example cities.
The primary objective of this study is to evaluate one such approach-
—the EKMA—through the application of the photochemical trajectory model
(OZIPP) that forms its basis. The EPA has prepared highly detailed data
concerning precursor emissions, meteorology, and ambient air quality for
Tulsa and Philadelphia to test various air quality simulation models such
as the OZIPP and the SAI Urban Airshed Model. In addition, there are data
bases already available for St. Louis, Los Angeles, San Francisco, and
Sacramento. Because some models are grid-based and thus designed to
reproduce spatial and temporal variations within the metropolitan
airshed, detailed information files are required. The availability of
detailed data and of the results from parallel studies using sophisticated
models offers the following advantages for making an in-depth study of the
EKMA:
49r 81 19
-------
The simple application of the OZIPP trajectory model,
using either the actual available data inputs or the
corresponding data simulated in other models, should
generate results that can be compared both with ambient
air data and with the results of the other models. The
timing of results regarding data such as the ozone maxima
and precursor decay should also be comparable.
Tests of the sensitivity of the results to variations in
the input data can demonstrate important data requirements
for successful application of OZIPP. This information
should be useful in determining cost effectiveness for
future uses. The results of the sensitivity study are
also useful for improving the trajectory model in case the
present version does not adequately simulate observed
ambient data or the simulations of more complex models.
The application of the EKMA procedures to air quality
planning demonstrates the utility of the EKMA compared
with that of methods based either on simpler models or on
more complex models (such as the SAI Urban Airshed Model).
The potential effectiveness of OZIPP model improvements
can be assessed using the results of the complex models as
a standard.
THE EMPIRICAL KINETIC MODELING APPROACH
The Environmental Protection Agency (EPA) has developed a method for
estimating the emission controls that will be needed to meet the National
Ambient Air Quality Standard (NAAQS) for ozone (03) concentrations in
urban areas. This method, known as the Empirical Kinetic Modeling
Approach (EKMA), uses an isopleth diagram of ozone concentrations that is
related to hydrocarbon (HC) and nitrogen oxide (NOX) precursor levels.
Each point on the isopleth diagram represents the maximum one-hour ozone
level reached as a result of the specific HC and NOX combination described
by the abscissa and ordinate values of the isopleth diagram. The ozone
value generated from each HC and NOX combination is obtained by means of a
computer-based trajectory model--the Ozone Isopleth Plotting Package
(OZIPP) (Whitten and Hogo, 1978a). The OZIPP trajectory model contains
many simplifying assumptions and options that are designed to strike a
balance between factors such as the state of knowledge, data availability,
computer size, predictive accuracy, and overall cost.
<+9r 81 19
-------
AN OVERVIEW OF EKMA COMPARISON EVALUATIONS
This is the third of four studies performed at SAI to evaluate the
EKMA through comparison of the OZIPP trajectory model with other models.
The first study compared the SAI grid and trajectory models with various
versions of the trajectory model used in the EKMA for the Los Angeles area
(Whitten and Hogo, 1981); the second study involved two cities and two
grid models: The LIRAQ model formed the basis for the San Francisco area
comparison; for the Sacramento area the SAI model was used again (Whitten,
Hogo, and Johnson, 1981). The SAI model was also used for the present
study of the Tulsa area.
A comparison evaluation is done at several levels that provide a
spectrum of information. The independent application of the EKMA-OZIPP in
areas where other models have been applied merely validates the common
assumption that different models can sometimes produce different
results. This study attempts to go farther and to provide information
explaining why specific features of the different models lead to different
results. For instance, a specific feature of grid models, in contrast to
trajectory models in general, is their ability to simulate some wind shear
effects. The two models (grid and trajectory) have been found to produce
different results when wind shear effects are significant; the reverse is
also true (similar results can occur when wind shear is minimal). Thus,
specific differences in models can be associated with differences in their
results. Given the previous example, a prospective user would do well to
apply the trajectory model with a measure of discretion in a situation
known to have a high wind shear.
Different levels of evaluation are possible within the framework of
the EKMA. The control scenario estimations generated by the use of EKMA
isopleth diagrams are the end result of a series of steps involving model
assumptions and adjustments. At the beginning or base of the EKMA is a
trajectory model that is basically quite simple from the standpoint of
dispersion. This trajectory model can be applied in the absolute sense as
an atmospheric model on its own. The OZIPP computer code that contains
this model also has an array of options, each of which will be geared to
some specific standardized setting if the user does not exercise the
option.
The isopleth diagram is used in a relative manner so that, if all the
options and inputs fail to generate a simulated ozone value that agrees
with observations, the final discrepancy is eliminated by using the
observed value. Some discrepancy is almost always expected because,
except for the chemistry employed, the EKMA-OZIPP is a rather simplistic
81 19
-------
approach. The use of the observed ozone value rather than the simulated
value is intended to account for some of the simplifications inherent in
the EKMA-OZIPP.
The levels of evaluation in this study begin with the specific
features that define each basic model; other levels involve input data,
options, and, eventually, the methodology of constructing and using the
isopleth diagrams. At each level, the present study attempts to make
comparisons with other models, observations, and alternate methodol-
ogies. However, like many situations in nature, the evaluation of
atmospheric models does not necessarily fit a specific, well-organized
framework. Problems that are unique to some areas are not always apparent
ahead of time. Thus, this present study, which uses data from the Tulsa
area, is only part of an overall evaluation of the EKMA. Following
sections present the specific results and analysis of the Tulsa study as
they relate to the EKMA evaluation.
81 19
-------
SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS
SUMMARY
We have evaluated the EKMA primarily by comparing the trajectory
model that forms the basis of the EKMA (OZIPP) with other more
sophisticated models in applications to the Tulsa area. The study was
actually carried out at several levels, beginning with evaluation of OZIPP
and ending with an evaluation of the control-strategy predictions that
result from employing the EKMA isopleth methodology. The OZIPP trajectory
model was compared with the SAI Airshed Model and the SAI trajectory
model, as well as with some modified versions of the original OZIPP model.
The basic OZIPP model is a simple, moving-air-parcel, or trajectory,
model that uses a detailed chemical mechanism for the surrogate propylene
and butane hydrocarbons. The model treats time-dependent, precursor-
emission factors along with expansion of the air parcel; entrainment is
treated by assuming that constant concentrations exist outside the parcel.
CONCLUSIONS
The primary emphasis of this comparison study was directed toward the
discovery of features in the basic OZIPP model that could explain
differences in the results of the OZIPP model from those of some other
model. To that end, most of the OZIPP simulations utilized initial
conditions derived from the model used for comparison purposes. Thus, for
the most part, the study used the SAI Airshed Model data files to generate
appropriate OZIPP inputs. Although the utilization of these files tends
to circumvent the issue of normal-source data preparation, the comparison
of models is clarified by use of common data bases. Furthermore, costs
are minimized. For some of the OZIPP simulations, we used air quality
data rather than information common to the comparison model. This step
allowed the categorization of overall differences as either effects
resulting from the data utilized or those attributable to specific
features of the models.
This study, and similar studies by Whitten and Hogo (1981) and
Whitten, Hogo, and Johnson (1981) using Los Angeles, San Francisco, and
81 21
-------
Sacramento data, show that the OZIPP model can provide results that are
often quite similar to those of the more complex models. By carefully
separating the models according to specific formulation differences and
through careful analysis of intermediate results, the most important
feature in formulation that produced the difference in the results of two
models could often be isolated. Even though the absolute values predicted
by the OZIPP trajectory model can be significantly different from those of
the more complex models (one difference can result from the different
chemical mechanisms employed), the isopleth applications of the EKMA
(which uses the trajectory model results in a relative fashion) typically
show control predictions that correspond with those of the more complex
models.
The present study differs from the previous evaluations in that the
EKMA and Airshed Models were evaluated simultaneously instead of the EKMA
evaluation following that of the Airshed Model. This timing allowed a
certain interchange and feedback between the two studies:
> Isopleth diagrams were constructed to assist the
development of control scenarios for the SAI Airshed Model
study. The original diagrams showed quite clearly that
the background hydrocarbon levels used in the Airshed
Model study significantly affected the results of
potential control requirements. Therefore, the original
estimates of background hydrocarbons were reevaluated and
were found to be erroneously high. The subsequent
lowering of the estimates led to an additional base case
validation of the Airshed Model.
> One of the modeled days (29 July 1977) was characterized
by significant wind shear within the mixed layer. The
trajectory models could not produce results similar to
those of the Airshed Model, even when an average value for
the winds was used to define a trajectory path. The
sensitivity to trajectory path (due to the divergent
emissions patterns) was so great in this study, however,
that even the grid model may be inadequate. The version
of the grid model used in this study had only two model
layers within the mixed layer. Adequate simulation may
require more than two layers to account for the
combination of significant wind shear and highly divergent
emissions patterns.
The main conclusions reached in this study from comparison of the
grid and trajectory models in application to the Tulsa area are similar to
those of previous studies such as Whitten, Hogo, and Johnson (1981):
81 21
-------
> These two basic types of models tend to show close
agreement only in the absence of wind shear.
> Any trajectory model, when run in conjunction with a grid-
model study, can serve as a useful and inexpensive tool
for understanding, analyzing, and testing important and
sensitive elements of the situation being modeled.
> The SAI trajectory model serves as a source of data for
the EKMA models, as well as an important conceptual link
between these models and the SAI Airshed Model.
Comparisons of the SAI trajectory model and an EKMA model (CBM/OZIPM)
modified to most closely resemble the SAI trajectory model showed that
many features in the SAI trajectory model can often be eliminated or
simplified with a minimal effect on results. For instance, the two levels
within the mixed layer of the SAI trajectory model seem adequately
represented by a single layer in the CBM/OZIPM model. The same
photochemical kinetic mechanism, the Carbon-Bond Mechanism, is used in the
SAI Airshed Model, the SAI trajectory model, and the CBM/OZIPM model.
This mechanism is described as CBM-II by Whitten, Killus, and Hogo (1980).
Two important differences in model characterization were identifed in
this study, however:
> Surface deposition, neglected in the EKMA models, can
reduce ozone values by as much as 15 percent.
> Entrainment of PAN and reactive intermediate compounds
from aloft can increase ozone by as much as 10 percent.
The EKMA models normally entrain only 0^, N02, and hydrocarbons from aloft
and the concentrations aloft are assumed to be constant during the day,
whereas more sophisticated trajectory models, such as the SAI trajectory
model used in this study, simulate the same species aloft as in the mixed
layer, and the time-dependent chemistry aloft is simulated as well. There
were no measurements of PAN or other reactive intermediates aloft.
However, the SAI multilayer models (the Airshed Model and the trajectory
model) do simulate concentrations of these species aloft and subsequently
entrain them into the mixed layer when the inversion height rises.
Modification of an EKMA-type model to entrain the same amounts of these
species as are simulated aloft in the SAI model resulted in up to a
10 percent increase in ozone.
491" 81 21
-------
The single substitution of the CBM chemistry for the standard EKMA
propylene/butane chemistry produced a modified EKMA model that could be
used with the standard EKMA to compare the two chemistries. This
substitution showed that the propylene/butane mechanism frequently
generated less ozone than did the CBM. This reduced ozone generation was
found to be a result of some outdated rate constants; the key type of
reaction involved the conversion of NO to NC^ via peroxy radicals (R02 and
H02).
For the 1982 SIP studies the EPA has designated degrees of model
sophistication by Levels I through IV, which are described on
pages 65669-65670 in the November 14, 1979, Federal Register. In this
study Level I models are the SAI Airshed Model and the SAI trajectory
model. Versions of the OZIPP trajectory model used in this study, which
use trajectory paths derived from wind measurements, are considered to be
Level II EKMA. The important feature defining a Level III EKMA model is
the assumed trajectory relating early morning urban precursors to an
observed ozone peak (EPA, 1980, October draft).
Level III EKMA simulations that showed reasonable agreement with
observations were also performed, though the simulated values were low, in
general. However, the use of county-wide emissions in this model makes
detailed comparisons with the other models used in this study difficult
because the emissions data are different.
The CBM was also substituted for the regular propylene/butane
chemistry in the Level III EKMA model. The ozone levels generated by the
CBM in this test were, in general, higher than the simulated ozone levels
generated by the propylene/butane chemistry. Both the CBM and the
propylene/butane mechanism were used with their respective default
reactivities. Use of the CBM with reactivities suggested by the Tulsa
emissions data resulted in the generation of rather low ozone levels. A
revaluation of the Tulsa emissions data is in progress. Preliminary
results to date indicate reactivities approaching the CBM default level.
Those conclusions based on comparisons of control scenario
predictions for the SAI Airshed Model and the EKMA isopleths were,
unfortunately, limited by time constraints. At the time of this writing,
only one base case day (29 July 1977) had been used for control
simulations in the grid model study using a 1987 emissions inventory with
several options. For this comparison, Level III EKMA isopleth predictions
tended to be quite close to the SAI Airshed Model predictions.
Unexpectedly, the substitution of the CBM into Level III EKMA isopleths
resulted in significantly lower ozone predictions for 1987. The reasons
for this effect are complex and require further study.
81 21
8
-------
Those conclusions of the present study specific to the sensitivity of
various parts of the EKMA were as follows:
> Emissions density information—Although adjacent Level II
trajectory paths produced significantly different results
because of different emissions densities, the use of
broader trajectory paths with averaged emissions
eliminated this sensitivity.
> Rate constant adjustments in the propylene/butane
mechanism—Although many rate constants have now become
somewhat out of date, a critical rate involving peroxy
radical reactions with NO was identified. Moreover, the
ratio of this type of rate to the formation of nitric acid
from N02 was shown to be important.
> Surface and aloft precursors in the background air—These
were shown to play a critical role in the Airshed Model
study of the Tulsa area. Control strategy implications
were found to be highly sensitive to the overstated values
assumed for precursor levels in future background air.
> Trajectory path determination with wind shear—Attempts to
use average wind direction under wind shear conditions
demonstrated the importance of path determination over
widely varying emissions patterns for Level II models.
RECOMMENDATIONS
The following considerations are suggested for future studies:
> For cases in which differences in control strategy
predictions are found to be related to differences in the
chemistry employed in various models, it is important to
investigate the chemical reactions that produce these
differences. If appropriate smog chamber experiments that
would validate one or more types of mechanisms can be
identified, these experiments should be performed.
> The use of more complex models such as the SAI Airshed
Model should be accompanied by the use of the EKMA, both
for absolute trajectory simulations along the path to the
major ozone peaks and for control strategy guidance from
the isopleth diagrams. The studies performed to date show
-------
that unless the wind fields are complicated by such
factors as wind shear, the two types of models normally
give similar results. In the absence of complex
meteorological factors, important differences in results
would indicate the existence of problems in one or both of
the models that should be investigated.
> Identification of background concentrations and
reactivities must be carefully considered for both present
and future simulations. Severe control requirements can
be implied if the level of background pollutants is
overstated.
> The regular city-specific algorithm might be modified to
allow optional starting times in place of the present 0800
LOT starting time. Earlier starting times would emphasize
dependence on an emissions inventory, whereas later
starting times might emphasize dependence on initial
conditions.
> Modifications to the current version of the EKMA have also
been suggested in other studies. Although such
suggestions have not been addressed in this study, their
implementation might also be considered.
- The two-stage approach (in which stationary-source
emissions are held constant in the first stage and
mobile-source emissions are held constant in the second
stage) might be further developed so that local
districts can use EKMA isopleths to make the control
predictions for stationary emissions.
- Isopleths based on emissions inventories might be
further developed to perhaps replace the present
initial-condition-based isopleths. Urban HC and NOX
measurements are often not relevant to the initial
trajectory. Trajectories might originate prior to the
present time of 0800 hours. Path definition is not
important to moving-air-parcel models in general,
unless differences in emissions density and mixing
height exist. Preliminary calculations suggest that
precise definition of mixing height prior to 0800 hours
is unimportant. Wind velocities prior to 0800 hours
are typically low so that few grid squares would be
81 21 10
-------
81 21
crossed before the arrival of the wind at the more
easily definable trajectory paths now used, those
beginning at 0800 hours. Therefore, an important
factor in the success of an emissions-dominated model
would be minimal, or at least moderate, diversity of
the emissions density in the vicinity of the present
0800 starting area. To be sure, the present study
identifies the sensitivity to emissions diversity for
path definition in connection with attempts to simulate
a day exhibiting wind shear (see section 6). A
probable solution to that case would have been an
averaging of the emissions across several grid squares
upwind of the 0800 location of the main trajectory
path.
An emissions-oriented model is also potentially more
advantageous than the present initial-condition-and-
measured-HC/NOv model because of the poor mixing that
can occur prior to 0800. Measured urban HC and NOX
values often vary widely even when taken at the same
location as a result of local mixing and local
emissions. Using the HC/NOX ratio does not necessarily
mitigate the problem because the measured ratios can
still be dominated by local emissions and mixing
effects. Starting the model earlier and filling it
with well-mixed emissions from the region upwind of,
and within, the main urban area provides a more
realistic simulation of the air that arrives at the
downwind site where the ozone maximum occurs.
Furthermore, the ordinate and abscissa (which are
scaled relative to the emissions rather than initial
concentrations) would relate more closely to the
purpose of the EKMA (i.e., to estimate the changes in
peak ozone resulting from changes in emissions).
The mixing-depth algorithm should be improved so that
the changes with time correspond to observations
instead of producing a constant-dilution factor.
The transported pollutants, both in the surface layer
and aloft, might be varied by the user. The range of
variance would be between present levels and some
minimal natural background level. For some cities,
there is evidence to suggest that air transported into
11
-------
the area is actually returning from a previous day.
Other cities may be downwind of another urban area that
requires similar controls. The current procedure
requires two separate isopleth diagrams. (Again,
further development would be necessary to incorporate
this refinement.)
The use of atmospheric models under conditions of wind
shear within the mixed layer should be approached with
caution. Although Level II trajectory models are
basically inappropriate under such conditions, their use
to assess a specific area can still be helpful in Level I
grid model studies.
A generic concept can be associated with the EKMA that
goes beyond its present formulation and uses.
Developmental work on the EKMA should be continued to
update its formulation and improve its utility.
Specifically, the following steps are suggested:
- This study demonstrates that significant differences in
both absolute model results and in predicted control
requirements can sometimes occur when the Carbon-Bond
Mechanism is substituted for the propylene/butane
chemistry currently used in OZIPP. Further studies
should be undertaken to investigate and resolve
potential discrepancies between the two mechanisms.
- Point source emissions might be handled so that the
interaction with all possible urban HC/NOX
concentrations could be evaluated as part of a new-
source review procedure.
- Percent cutback diagrams (PCD) might be added to the
OZIPP code so that both the isopleth and the PCD are
generated.
- Temperature changes, deposition, variable reactivities,
nitrate chemistry, and sulfate chemistry can be added
to the OZIPP. Isopleths of species other than ozone
can also be generated by OZIPP.
- Statistical packages can be combined with the EKMA to
generate the probable number of exceedances resulting
from various control scenarios.
<+9r 81 21
12
-------
Credibility of the EKMA could be enhanced by providing a
means for considering urban hydrocarbon reactivity.
Either the latter should be related to the proplyene
reactivity used in OZIPP, or the chemical mechanism
utilized in the OZIPP should be changed to be compatible
with available reactivity information.
81 21
13
-------
SECTION 3
MODEL APPLICATIONS
This section describes the different photochemical air quality models
used for this study and the inputs associated with each model. Descrip-
tions of the Tulsa study region and the modeling days used in that study
are also presented in this section.
The purpose of this study is to compare ozone predictions between air
quality models of different complexity (in particular, to perform EKMA
calculations along with current SAI Airshed Model applications). These
models can be classified, on the basis of their complexity, according to
the levels of analysis described in the 1980 Federal Register notice on
ozone modeling:
> SAI Airshed Model—Level I analysis using a complex grid
model.
> SAI trajectory model—Level I analysis using a complex
trajectory model.
> OZIPM/CBM--Level II analysis using a simplified trajectory
model.
> City-Specific EKMA—Level III analysis.
> Standard default EKMA conditions—Level IV analysis.
The SAI Airshed Model is a three-dimensional grid model that mathe-
matically simulates the physical and chemical processes responsible for
photochemical smog. One of the advantages offered by a three-dimensional
grid model is the capability to consider spatial and temporal effects of
control strategies, whereas models of less complexity consider temporal
effects only. The structure of the model consists of an array of cells,
and the total volume represents an urban area. The horizontal dimensions
of each cell are constant, but the vertical thickness of the cells vary
according to the depth of the mixed layer throughout the simulation. The
»*9r si 2
14
-------
Airshed Model can simulate the advection of pollutant species through the
modeling region, the diffusion of material from cell to cell, the injec-
tion of primary source emissions into the modeled region, and the chemical
transformations of reactive species into intermediate and secondary
products.
The SAI Airshed Model was applied to four days in 1977--21 July, 29
July, 3 August, and 2 September—to validate model predictions with
station observations. Then, the model was used to evaluate control
strategies under consideration for the Tulsa area. The inputs to the SAI
Airshed Model serve as a data base for the series of less complex
models. Descriptions of each model are presented in the following
sections. Inputs to the SAI Airshed Model are summarized in the following
sections and described in detail in appendix A. Inputs to the less
complex models (based on the inputs from the Airshed Model) are presented
in the following sections.
DEFINITION OF THE MODELS
The following list briefly describes the series of models used in
this study. They are described in greater detail in the discussions that
compare them. Some models are distinguished only by the mode of applica-
tion.
> The SAI Urban Airshed Model.
> The SAI trajectory model. In terms of inputs, computa-
tions, and results at points along a selected trajectory
this model differs from the Airshed Model only through
the elimination of horizontal dispersion and convergent
and divergent winds.
> An EKMA-style model (OZIPM/CBM), which uses the same
chemistry (Carbon-Bond) and mixing heights as the SAI
trajectory model.
> The OZIPM model (as described above) but using the
regular chemistry.
> The city-specific OZIPP model as described by the EPA
(1980), but using Carbon-Bond chemistry.
> The city-specific OZIPP model incorporating the standard
EKMA chemistry as described by the EPA (1980).
i+9r 81 2
15
-------
In this section we wish to highlight the principal differences among
the models.
Comparison of the SAI Airshed Model and the SAI Trajectory Model*
These models were designed to complement each other and, therefore,
many algorithms are identical for both models (several subroutines of the
computer codes are actually shared); furthermore, they 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. Neverthe-
less, the trajectory model provides an inexpensive method for testing
parts of the grid model.
A special feature of the trajectory model is its ability to be
operated backward in time. This feature was used to generate the tra-
jectories used in this study. The time and location of either the
observed or the simulated ozone maximum were used to start the backward
trajectory to 0500 CDT. Thus, the forward trajectories always arrived at
the desired spot at the proper time for comparison with either observa-
tions or grid-model results. 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 location of the
origin of, and in the actual pathway taken by, the air parcel producing
the specific ozone maximum. Therefore, when compared with a grid model,
any trajectory model used as a tool for control strategies possesses the
inherent shortcoming of being incapable of accounting for these changes.
Comparison of the SAI Trajectory Model and the OZIPM/CBM Model
The following similarities and differences exist between the tra-
jectory model and the OZIPM/CBM developed for this comparison:
> Chemistry. Both models used in this study employ Carbon-
Bond chemistry identical to that of the Airshed Model.
However, the photolytic constants are computed differently
Detailed discussions of this comparison are available elsewhere (see
Reynolds et al., 1979).
81 2
16
-------
in the OZIPM/CBM and SAI models. For this study, we
modified the regular OZIPM/CBM values for N02 photolysis
using a constant factor so that all models agreed at 1300
COT. In the Airshed Model and trajectory model, the rates
of all other photolytic reactions vary with N02 photoly-
sis, with a fixed ratio for each; but in the OZIPM/CBM the
photolysis rates vary independently. Aldehyde photolysis
was adjusted in the OZIPM/CBM to give agreement between
the models at 1100 COT.
Actually, these differences in photolysis rates constitute
an advantage in sophistication for the OZIPM/CBM. The N02
photolysis rates normally used in the OZIPM/CBM are some-
what lower than those used in the SAI models, which are
based on actual measurements. Since the purpose of the
present study was to compare the OZIPM/CBM with computed
results of the Airshed Model, we were forced to alter the
OZIPM/CBM models to eliminate this difference.
The use of fixed ratios to N02 photolysis in the SAI
models in place of the variable ratios used in the
OZIPM/CBM 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, Killus, and
Hogo, 1980). The photolysis rates for aldehydes change
more with solar zenith angle than the N02 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 ultraviolet spectrum. The
short ultraviolet wavelengths are more affected by the
ozone in the stratosphere than are the longer wave-
lengths. At high zenith angles (morning and evening), the
solar rays must pass through a thicker layer of ozone.
Therefore, the ratio of aldehyde photolysis to N02
photolysis varies throughout the day, reaching a maximum
at solar noon. For this particular study, the OZIPM/CBM
aldehyde photolysis values were multiplied by a constant
value throughout each day so that the models agreed at
1100 CDT. Therefore, in the early morning the photolysis
rate of the aldehydes in the OZIPM/CBM would be somewhat
less than that used in the SAI models.
For this particular study, we chose to modify the
OZIPM/CBM 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
81 2
17
-------
better compare the fundamentals underlying these models.
The CBM version of OZIPM/CBM 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.
> Reactivity. The SAI trajectory model and Airshed Model
are operated with reactivity splits that vary with each
source category. However, the OZIPM/CBM does not have
this feature; instead, 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 reac-
tivity were in the temporal variations aloft and in the
variable emissions used in the SAI models compared with
the fixed reactivities employed in the emissions for the
OZIPM models. The reactivities used in the OZIPM/CBM were
determined from averages of emissions used in the tra-
jectory model; for averaging the hydrocarbons aloft, a
weighting factor was used each hour that was equal to the
positive change in mixing height.
In a sense, the N02/NOX ratio also acts as a reactivity
effect. The OZIPM/CBM uses a fixed ratio of N02/NOX (10
percent) for all NOX emissions, whereas the SAI models
allow this ratio to vary according to their source cate-
gory (the average ratio is typically about 5 percent). At
this time, we have not attempted either to change the
OZIPM/CBM or to fix the ratio in the SAI model, because we
feel the difference is trivial. For NOX aloft, the
OZIPM/CBM uses pure N02, which is c-lose to typical values
found in the trajectory model simulations because the
N02/NOX ratio is almost always greater than 0.85; often,
the values are greater than 0.99.
> Integration Scheme. The OZIPM/CBM uses a Gear-type method
without any steady-state approximations, whereas the
trajectory model uses the same Crank-Nicholson finite dif-
ferencing scheme and steady-state approximations that are
used in the Airshed Model. Again, this difference in
model sophistication is opposite from the direction
if9r 81 2
18
-------
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.
> Vertical Layers. The Airshed Model and trajectory model
share the same vertical layers and algorithms for numeri-
cal integration, but additional contributions are intro-
duced into the Airshed Model for convergent or divergent
winds. The OZIPM/CBM, 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 diffusion among all layers. The OZIPM/CBM
assumes instantaneous mixing within the entire mixed
layer, and no eddy diffusion is considered across the
boundary defined by the mixing height.
In our comparison studies, the effectiveness of multi-
layers and eddy diffusion was one of the central issues to
be evaluated 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 OZIPM/CBM requires fixed concentrations aloft, so
these must be chosen from some "proper" average of the
time-dependent concentrations computed in the SAI trajec-
tory model. With the average value used in this study we
attempted to weight most heavily the concentrations
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 Airshed Model
and trajectory model, but is neglected in the OZIPM/CBM.
> Mixing height. All models used in the OZIPM/CBM that are
based on the OZIPM computer code, as well as the Airshed
Model and trajectory model, can use identical mixing
heights with values that vary linearly in time between
specified values at each hour.
49r 81 2
19
-------
> Temperature. The temperature varies in the Airshed Model
and trajectory model, but the OZIPM/CBM currently employs
a fixed temperature. This difference can be signifi-
cant. Temperature affects both vertical diffusion and
chemistry in the SAI models, but it can affect only
chemistry in the OZIPM/CBM because vertical mixing is
assumed to occur instantaneously within the mixed layer.
> Emissions. The Airshed Model and trajectory model allow
for emissions into all layers; the OZIPM/CBM treats all
emissions in the mixing layer alike and does not directly
address emissions above the mixing layer. For the layer
aloft in the OZIPM/CBM, 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 trajectory model was used for the mixing layer in
the OZIPM/CBM. For this study, the OZIPM computer code
was modified to treat emissions in an absolute fashion
rather than as a multiple of the initial conditions. This
was implemented by using the transported surface layer
inputs as the initial concentrations. The normal initial
concentrations (the CALC mode in OZIPM) were specified as
1.0 for both HC and NOX, but these values were not added
as initial concentrations. They were used only for
emission fraction development because emissions are
expressed relative to these values in the OZIPM code.
> Emissions for the SAI models are given as total moles
emitted into each cell during each hour. To arrive at
values appropriate for the OZIPM/CBM we used the following
conversion procedure: the mixing volumes of the trajec-
tory model and the OZIPM/CBM were assumed to be equal
below the mixing height, such that
v ' A * Zmix •
where A is the area of the horizontal cross-section of the
air parcel defined in trajectory model simulations, and
ZTO-JX is the mixing height. If the volume is expressed in
cubic centimeters, the moles emitted in one hour (the
number available from the trajectory model computer
output) can be divided by this volume to give moles/cm"3
units. This, in turn, can readily be converted to the
OZIPM/CBM units of concentration, ppm, by using a conver-
sion factor of 2.445 x 10*° ppm moles"* cm3, which is
technically correct for a perfect gas at 298 K and 1
81 2
20
-------
atmosphere of pressure. The numbers calculated at this
point are almost directly applicable to the OZIPM/CBM
model. First, the use of 1.0 for the initial concentra-
tions of HC and NOX in the CALC mode implies that each
one-hour emissions number would be the concentration added
during the hour if the mixing height were the same as the
initial value ZQ. However, the computer codes are written
so that, at each instant in time, the emissions factor is
multiplied by Zg/Zmjx. Therefore, the Z,^ drops out and
the actual numbers (En) used are merely
En = A x 2.445 x 1010/Z0
where A and ZQ are expressed in centimeters.
In this particular study, it was occasionally necessary to
use special emissions factors. These factors were used
for those cases in which the trajectory model employed
substantial initial conditions for the first two layers
(up to 100 m) even though the mixing height was only 50
m. Since there was typically an order-of-magnitude
difference in concentration between the initial conditions
below 100 m and the layers above this level, there was no
way to compute a reasonable average value to be used in
the OZIPM/CBM for the layer aloft. Therefore, the initial
conditions between 50 and 100 m used in the SAI models
were treated as emissions in the OZIPM/CBM. These
"emissions" were added to the normal emissions during the
hour when the mixed layer in the SAI models jumped from 50
to 100 m. For these cases the appropriate OZIPM/CBM
emissions factors merely add to the normal emissions,
since the initial mixing height is 50 m and the height
added is also 50 m.
Comparison of the OZIPM/CBM Model and the (OZIPM) Model
The OZIPM/CBM and the OZIPM models differ only in the photochemical
mechanism each employs. The OZIPM model uses the propylene/butane
chemistry as described by Dodge (1978). The obvious intention of this
comparison was to evaluate only the differences between the Carbon-Bond
chemistry and the standard propylene/butane chemistry. Although the OZIPM
computer program was specifically written to handle such comparisons, some
complications were introduced when the OZIPP was operated with the pro-
pylene/butane chemistry. The Carbon-Bond chemistry used in this study
<+9r 81 2
21
-------
assumes that formaldehyde and ketones represent half of the total alde-
hydes found in the simulation, whereas the formaldehyde in the OZIPM is
only 40 percent of the total amount of aldehydes. The photolysis rates
used in the Airshed Model runs are based on a theoretical solar radiation
function that does not contain the updated NC^ absorption cross sec-
tions. The OZIPM simulations use the latest N02 cross sections, which are
approximately 10 percent lower than those used in the Airshed Model.
The reactivity of the two mechanisms was not typically adjusted in
this study. The total HC concentrations and emissions rates were identi-
cal for these two models, but the propylene/butane chemistry was typically
used with the standard 25 percent propylene, 75 percent butane, and 5
percent aldehyde reactivity.
Comparison of the OZIPM Models and the City-Specific EKMA (UZIPP) Model
The basic assumptions for the inputs to the OZIPM and OZIPP models
are essentially the same (i.e., taken from the Airshed Model data base),
but the method of defining inputs differs between the two models. Whereas
the OZIPM models consider initial conditions at 0500 CDT, the OZIPP (city-
specific EKMA) model simulations begin at 0800 CDT.
Another difference between the OZIPM and the city-specific EKMA is
the use of variable mixing-height inputs in the former, more sophisti-
cated, version. In the OZIPM, 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 OZIPP 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.
The OZIPM is based on trajectory paths that contain emissions of NMHC
and NOX specific to each path. The city-specific EKMA, on the other hand,
uses county-wide emissions estimates, which vary only between counties.
Inputs developed for the city-specific EKMA are based directly on
monitoring station observations, whereas the OZIPM model inputs are based
on the inputs of the Airshed Model. Although the Airshed Model inputs are
also based on monitoring data, the inputs are in some cases "tuned" within
the uncertainties of the measurement to give closer ozone agreements
between predictions and observations.
81 2
22
-------
DESCRIPTION OF THE TULSA MODELING REGION
The Tulsa area consists of an isolated metropolitan area (Tulsa,
Oklahoma) surrounded by smaller cities. Unlike the Los Angeles region and
numerous cities located in the northeastern United States, the Tulsa
region clearly defines the locations of the principal sources of emissions
and precursor transport. A map of the Tulsa area is shown in figure 3-1.
CHARACTERIZATION OF THE TULSA MODELING DAYS
From the extensive monitoring study reported by Eaton et al. (1979a,
1979b; Eaton and Dimmock, 1979) during the summer of 1977, four days were
chosen for the validation of the SAI Airshed Model. The maximum observed
ozone concentration for the four days is presented in table 3-1. We note
that throughout this section any reference to time represents time at the
beginning of the hour (e.g., 0500 CDT represents 0500-0600 CDT).
This section describes the meteorological and air quality conditions
associated with the 21 July, 29 July, 3 August, and 2 September 1977 study
days, considering, in particular, the general transport characteristics of
each day and the resulting pollutant patterns observed over the monitoring
network. Much of the following discussion was taken from Reid, Reynolds,
and Oliver (1980) and reflects different meteorological air quality
conditions of interest for SAI Airshed Model validation. As discussed in
the following sections, the same conditions for these validation days were
used throughout the series of models. We note that Eaton et al. (1979c)
have carried out an extensive examination of several interesting days for
the 1977 field study in Tulsa that includes the four days of interest
here. We have found their analyses to be quite helpful in preparing this
description.
Characterization of 21 July 1977
The 21 July 1977 study period was chosen because the meteorological
conditions recorded for a large portion of this day were representative of
those most frequently observed in Tulsa during the summer of 1977. A weak
high-pressure area situated over northern Florida produced southerly winds
in Tulsa throughout the morning and early afternoon. This flow pattern
was consistent with the 0700 CDT rawinsonde data collected at Oklahoma
City, which indicated that winds averaged over the lowest 1000 meters were
out of the southwest. Temperature data obtained from this sounding also
suggest the presence of stable air throughout the lowest 4000 meters of
the atmosphere.
<+9r 81 2
23
-------
748
NORTH
788
368*
328
O
O1.
c
288
OSAGE CO.
OCHE
' * '•'••' « ' * ' : i '"i'"M248
748 788
SOUTH
UTM Easting (kilometers)
Figure 3-1. The gridded Tulsa modeling region.
<+9r 81 1 8
24
-------
TABLE 3-1. MAXIMUM OBSERVED OZONE CONCENTRATION FOR THE FOUR MODELING
DAYS IN 1977
(ppm)
Time of Occurrence
Location of of Maximum Maximum
Maximum Concentration Observed
Date Concentration (CDT) Concentration
21 July Vera 1300 0.14
29 July Apache 1400 0.17
3 August Vera 1500 0.15
2 September Ochelata 1500 0.11
81 10 25
-------
The NWS surface weather map for 0700 CDT indicated a cold front
located in the panhandle regions of Oklahoma and Texas. By midafternoon,
reduced solar radiation caused by increased cloudiness marked the arrival
of the front in the Tulsa area. At this time the prevailing southerly
winds gave way to a more stagnation-like condition that comprised light
westerly and northerly winds at the surface. However, the average wind
direction observed over the lowest 1000 meters in Oklahoma City at 1900
CDT was from the southeast.
Precursors emitted during the morning were carried to the north,
resulting in a maximum hourly averaged 03 level of 0.135 ppm that was
observed at Vera. The ozone concentrations at Sperry and Ochelata were
higher than those reported at Skiatook and Wynona, which suggests that
pollutants from Tulsa were transported to the north or northeast and not
to the northwest. With the onset of cloudiness in the midafternoon,
observed ozone levels decreased at all stations in the network.
Characterization of 29 July 1977
This day is of particular interest because ozone concentrations
measured at the Apache station exceeded 0.16 ppm during four consecutive
hours in the middle of the day. In addition, the maximum value of 0.166
ppm at this station was the highest value reported for the entire 1977
summer study period. Another interesting facet of this day is the fact
that the ozone values observed at the three other city stations (post
office, health department, and Mohawk) generally did not exceed 0.12
ppm. Although the greatest distance between any two city stations was
less than 8 kilometers, substantial spatial gradients in the ozone field
were observed.
On 28 July, a large high-pressure area was centered over the north-
eastern portion of the country, and a stationary front was situated just
south of the Oklahoma and Texas borders. The Tulsa area experienced
overcast skies throughout the day, yielding lower-than-average tempera-
tures. In general, the winds were relatively light and out of the south.
By the morning of 29 July, the stationary front was positioned just
north of Tulsa, and during the predawn hours ozone concentrations at
several stations exhibited an increase. Eaton et al. (1979a) attributed
this unexpected occurrence to a disturbance caused by a frontal passage
that enabled ozone aloft to reach the ground. Ozone levels during this
time typically rose to about 0.04 ppm.
The 0700 CDT rawinsonde data taken at Oklahoma City indicated that
the winds in the lowest 1000 meters of the atmosphere were from the north,
i»9r 81 2
26
-------
whereas the winds in between 2000 and 3800 meters of the atmosphere were
out of the south. The early-morning surface winds in the Tulsa area were
from the southwest. At about 1100 CDT, the surface winds at Wynona and
Ochelata exhibited a shift to a north-northeasterly orientation. The
southwesterly and northeasterly winds in the southwest and northeast
portion of the modeling region, respectively, seemed to combine to produce
light, southwesterly winds in the Tulsa city area throughout the mid-
morning to early-afternoon period.
By 1400 CDT, the winds at Liberty Mounds exhibited a northerly
orientation that seemed to indicate a transition to northerly winds
throughout the network. In the afternoon, the urban plume was transported
to the south, which is illustrated by the fact that the ozone concentra-
tions at Liberty Mounds suddenly increased from a steady, early-afternoon
value of about 0.08 ppm to a value of 0.12 ppm at 1700 CDT. In the late
afternoon, the surface winds exhibited yet another shift, this time to a
southerly orientation. Ozone at Liberty Mounds then decreased to a level
more characteristic of ozone levels observed at that station earlier in
the day. At the same time, the ozone levels observed at the post office
site (to the north of Liberty Mounds) showed an increase. A peak hourly
concentration of 0.11 ppm was measured at 1900 CDT.
Characterization of 3 August 1977
This day is of interest because of the observation of relatively high
ozone concentrations at stations to the north of Tulsa. In fact, the
second highest hourly averaged ozone concentrations for the summer field
study conducted at Sperry, Vera, and Ochelata--0.135, 0.151, and 0.114
ppm, respectively — were reported on this day.
On 3 August, a stationary front was located near the Gulf Coast, and
a ridge of high pressure was situated to the east of Oklahoma. It is
likely that the front was responsible for the stalling of synoptic systems
that led to stagnation conditions in the Tulsa area. During the morning,
the skies were overcast and did not clear until about 1100 CDT. The
reduced solar radiation led not only to lower morning mixing heights, but
also to the delay of photochemical reaction phenomena until the late-
morning period.
The 0700 and 1900 CDT rawinsonde data recorded at Oklahoma City
indicated that the winds aloft were out of the south over the lowest 2500
meters of the atmosphere. Southerly winds were also observed at the
surface in Tulsa during the early-morning hours at all sites except
Liberty Mounds, which reported easterly winds. (We note that the winds
measured at Liberty Mounds throughout a significant portion of this day
81 2
27
-------
were inconsistent with those reported at the other stations.) Between
0800 and 1400 COT, the winds at each station (except Liberty Mounds) were
observed to undergo a systematic clockwise rotation. For example, the
hourly averaged directions reported at the health department site for the
hours 0800 to 1400 CDT were 189, 234, 288, 360, 45, 135, and 171 degrees,
respectively. By 1400 CDT, southerly winds were again reestablished in
the Tulsa area.
Because of the wind conditions cited above, early-morning precursors
were initially transported to the north and east. As the morning pro-
gressed, transport to the south and west probably brought an already
polluted air mass back into the vicinity of the city. With the onset of
southerly winds at midday, it is possible that additional emissions were
injected into the air parcel as it crossed the city and was transported to
the north. Ozone concentrations at Liberty Mounds exhibited a continuous
increase throughout the morning and peaked at 0.103 ppm between 1300 and
1400 CDT. It is likely that some of this ozone may have resulted from the
reaction of precursors that were emitted earlier in Tulsa and that were
subject to southern transport during a portion of the morning. Late-
morning ozone concentrations at other rural stations were somewhat lower.
Characterization of 2 September 1977
This day was selected for evaluating SAI Airshed Model performance to
include at least one period during which the supplemental upper-air
observation program was in effect. The primary influence on the meteorol-
ogy in Tulsa for this day was a high-pressure area centered over the
mountainous regions of North Carolina, Virginia, and West Virginia. The
winds were generally out of the south throughout the day, and the skies
were mostly clear. During the early-morning hours, the surface winds were
relatively light, indicating that little pollutant transport was occur-
ring. However, after 1200 CDT, surface winds in excess of 10 mph were
observed.
Three sets of integrated hydrocarbon grab samples were collected at
the Liberty Mounds, health department, post office, and Sperry sites
during the day. These were three-hour samples taken from 0600 to 0900,
0900 to 1200, and 1400 to 1700 CDT. During the 0600 to 0900 CDT sampling
period, the hydrocarbon composition at the two city sites was largely
representative of motor vehicle exhaust. At Sperry, however, a high
percentage of low-molecular-weight hydrocarbons was reported, indicating
that earlier emissions from the oil refining and storage facilities to the
south had been carried by the southerly winds to the Sperry area. The
0900 to 1200 CDT sample collected at the post office site also contained a
large fraction of low-molecular-weight hydrocarbons. Since the winds at
81 2
28
-------
this time were southerly to southwesterly, it is likely that these
measurements are also reflective of emissions from the petrochemical
complex. Hydrocarbon concentrations recorded in the 1400 to 1700 CDT bag
samples were relatively low, suggesting the effect of increased mixing and
removal by photochemical reaction.
Air quality measurements taken aloft using an instrumented aircraft
were obtained during both the early morning and the early afternoon. Each
flight consisted of a combination of vertical spirals and horizontal
traverses oriented upwind (to the south) and downwind (to the north) of
the city during the morning and afternoon periods, respectively. Although
the early morning NO and N02 concentrations aloft were found to be very
low, a background ozone level of about 0.06 ppm was observed. The
afternoon flight took place between 1302 and 1535 CDT, and a clear, urban
ozone plume was observed during the east-west horizontal traverses at 2500
feet MSL to the north of the city. A maximum ozone concentration of 0.156
ppm was measured on a horizontal traverse that passed just to the south of
Ochelata.
The measured ground-level ozone concentrations to the north of Tulsa
in the afternoon also indicate the presence of an urban plume. The plume
began to form at about 1100 CDT and was transported to the north. Ozone
concentrations at Ochelata were relatively constant during the time from
1000 to 1400 CDT, ranging between 0.053 to 0.060 ppm. During the next two
hours, the measured ozone levels increased, signifying the arrival of the
plume in the Ochelata area. The highest hourly averaged ground-level
ozone concentration of 0.108 ppm on this day was observed at Ochelata
between 1500 and 1600 CDT. At 1500 CDT, the winds shifted from a south-
erly to a southeasterly direction; the ozone values at Ochelata began to
decrease, and concentrations at stations to the west increased. The
maximum ozone concentrations at Skiatook and Wynona were reported at 1600
and 1800 CDT, respectively.
INPUTS USED IN THE SAI AIRSHED MODEL SIMULATIONS
Detailed descriptions of the methodologies used in describing the
meteorological, emissions, and boundary concentration inputs used in the
SAI Airshed Model are presented in appendix A of this document. In this
section, we summarize some of the inputs used for each of the four
modeling days.
The study region used in the SAI Airshed Model simulations was
defined during the development of the emission inventory by Engineering
Science (1979). The modeling region itself is 52 km by 124 km consisting
of 403 square grid cells. Each grid cell is 4000 meters in length.
49r 81 2
29
-------
Surrounding each side of the modeling region is a row of cells represent-
ing the boundary cells and containing the boundary conditions (concentra-
tions) used in the SAI airshed simulations. A map of the modeling region
is shown in figure 3-1, presented earlier.
Initial conditions for ozone, NOX, and hydrocarbons in the surface
layer were defined on the basis of monitoring station observations for
each of the modeling days. The surface initial conditions are shown
graphically for NOX and hydrocarbons in figures 3-2 and 3-3 for 21 July
and 3 August. A spatially uniform NMHC value of 0.2 ppmC was used for the
29 July modeling day.
Boundary conditions for the four modeling days are shown in table
3-2. These boundary conditions were used for the cells along the horizon-
tal edges and in the aloft layers of the modeling region.
Mixing height profiles were developed for each day using temperature
soundings taken at the Tulsa airport. Mixing heights were spatially
constant over the rural and urban regions during most of the day. Mixing
heights over the urban region were set higher in the morning and late
afternoon hours for 21 July, 29 July, and 2 September. Mixing height
profiles are shown graphically in figure 3-4 for the four modeling days.
Three-dimensional wind fields were generated from station measure-
ments for each of the modeling days. Wind fields for selected hours are
shown in appendix A (figure A-3). In general, winds are usually from the
south and southwest during the morning hours. Southerly winds may persist
throughout the day. On certain days the winds are from the north and east
during the afternoon hours, which leads to a turning of the air parcel
around the urban area. This effect occurred on 29 July and 3 August. Air
parcel trajectories derived from the SAI trajectory model (discussed in
the following sections) show the turning effect around the urban area.
Emissions inventories for the Tulsa area were developed by
Engineering-Science (1979). The emissions inventory reflects both ground-
level and elevated point sources of hydrocarbon and NOX. Emissions totals
are summarized in section 5 of this document (table 5-1). Spatial
distributions of each of the Carbon-Bond species in the emissions inven-
tory are shown in table A-6 of appendix A.
INPUTS USED IN THE SAI TRAJECTORY MODEL SIMULATIONS
The meteorological and emissions files used in the SAI trajectory
model are the same as those used by the SAI Airshed Model. By running the
SAI trajectory model in the backwards mode, we can determine a path to the
si 2
30
-------
NORTH
IB
_NYNO
20
OCHL
MNDS
1 I
10
SOUTH
j I
20
in
a
UJ
10
(a) For NOX (ppb)
Figure 3-2. Initial conditions in the ground-level
grid cells on 3 August 1977 (isopleths
in increments of 2 pphm).
81 1 8
31
-------
NORTH
30 h
H30
1 1 1 1 1 1 1
SOUTH
(b) For RHC (pphm)
Figure 3-2 (concluded).
U9T 81 18
-------
NORTH
i i i I J_ J \
81 1 8
10
S3UTH
(a) For NOX (ppb)
Figure 3-3. Initial conditions in the ground-level
grid cells on 21 July 1977.
33
-------
30
20
10
_H .'M3
N0RTH
10
30
I L L 1
_L
10
S0UTH
I I
(b) For RHC (pphm)
Figure 3-3 (Concluded).
»*9r si i e
34
-------
TABLE 3-2,
Species
NO
N02
°3
NMHC
Ethylene
Paraffins
Olefins
Aromatics
Carbonyls
Benzaldehyde
CO
BOUNDARY CONCENTRATIONS USED IN THE
TULSA SIMULATIONS
(ppm)
Boundary Concentrations
21 July
0.0001
0.002
0.06
0.24
0.0058
0.065
0.0014
0.012
0.012
0.00001
0.1
29 July
0.0005
0.002
0.08*
0.20
0.0049
0.053
0.0012
0.009
0.009
0.00001
0.1
3 Aug.
0.0005
0.002
0.06
0.21
0.0051
0.056
0.0013
0.010
0.010
0.00001
0.1
2 Sept.
0.0005
0.002
0.05*
0.17
0.0041
0.046
0.0010
0.008
0.008
0.00001
0.1
For 29 July, the ozone boundary concentrations on the sides of the
modeling region varied with location. The boundary values on the
north, south, east, and west boundaries were 0.07, 0.08, 0.06, and
0.08 ppm, respectively.
For 2 September, the O-j boundary concentration at the top of the
modeling region varied in time: From 400 to 1100 CST, it was set at
0.07 ppm and from 1100 to 2100 CST, it was set at 0.05 ppm.
81 10
35
-------
D- Q-
UJ UJ
0 D
C? O
5 S 0-
X X
CO
o
CM
CO
o
(O
o
QJ
2 I
3
•"3
i.
O
fO
CO
c
o
cr.
QJ
en
c
01
•o
QJ
a.
o
T3
c
(O
c.
O)
-o
X
E
QJ
OJ
re
o.
(D
C
S-
-o
CJ
ro
O)
3
O'
O
o
CO
o
o
00
o
o
VD
O
o
o
o
CM
(punoaG
81 1 8
36
-------
O
C\J
00
rn a:
i— i—
a. o.
LU LU
f""*i f"^
O O
i—i •—i Q-
X X O
I-H I—t f—
s: 2:
-------
O. CL.
LU O
O I--
CO
t/1
o
C\J O
1—I
CD
O I_
«—i 3
O
CO
CTi
rs
-------
00
U n: 3=
. UJ UJ
D O
(— C 1— C d-
x x o
.
CD
li—i t/o
ro
a
ICNJ 01
3
O
S-
01
X3
a>
CM
I CO
Ivo
o
o
o
o
o
CO
o
o
vo
o
o
o
o
CVJ
-a
QJ
-o
U
c:
O
i
CO
a>
en
BULXLH
i*9T 81 1 8
39
-------
maximum ozone observed for the day of interest. This path is based on the
wind fields in the first vertical layer only. As discussed earlier the
use of the wind fields in only the first layer can lead to differences in
the predicted maximum ozone between the SAI trajectory model and the SAI
Airshed Model on days with wind shear. From the locations of the maximum
ozone observed for each day presented earlier in table 3-1, we determine
the path to the maximum ozone. The trajectory paths for the four modeling
days are shown in figures 3-5(a) through 3-5(d).
INPUTS USED IN THE LEVEL II CBM/OZIPM AND LEVEL II OZIPM MODEL SIMULATIONS
On the basis of the results of the SAI trajectory model, emissions
and meteorological conditions are constructed for a Level II type of EKMA
model using Carbon-Bond chemistry and a Level II EKMA based on the
standard propylene/butane chemistry as used in the original version of the
OZIPP code.
The SAI trajectory model prints out the emissions and mixing heights
along the trajectory path at hourly intervals. These mixing heights can
be used as direct inputs to the OZIPM computer code. The emissions rates
from the SAI trajectory model are in units of moles/hr for each of the
Carbon-Bond species. These emissions are converted to ppmC/hr for
volatile organic compounds (VOC) by the following equation:
2
Emissions Rate (ppmC/hr) = [24450/(ZQ* 4000 )][Emissions Rate (moles/hr)]
where ZQ is the initial mixing height.
Initial NOX and VOC precursors (OZIPM) are taken from the initial
conditions from the average of the layers below the mixing height found in
the SAI trajectory model. NOX and VOC precursors aloft are determined
from the amount of precursors that entered the mixed layer from aloft in
the SAI trajectory model simulations. As a result of using this methodol-
ogy, the NOX and VOC aloft precursors are usually higher than measured
values. The chemistry in the SAI trajectory model will produce secondary
organic products (such as carbonyls) and ozone in the layer aloft. Thus,
if we were to compare assumed aloft conditions in the CBM/OZIPM (tables 3-
3 through 3-6) and actual measured data, we would find higher concentra-
tions in the OZIPM models. This method was used because the species aloft
continuously react in the SAI trajectory model, whereas OZIPM does not
treat precursors aloft until they are entrained. Emissions and meteorol-
ogical conditions for the four modeling days are shown in tables 3-3
through 3-6.
t+9r 81 2
40
-------
81 1 6
20
10
N0RTH
0CHE
1300]
VERR
SKIR
SPRY
noo>
P0;
1000)
RPCH
HLTH
0900
0800
0700
0600j
0500
HNDS
SOUTH
20
10
(a) For 21 July 1977
Figure 3-5. Trajectory path (CDT)
41
-------
20
10
NCRTH
li i 1 l t 1 I T T 1 I I T
PC HE
VEF.fi
SKIR
SPRY
HKk'K
1300
FCSTy
0800
0700
0600
;0500
MUDS
5CUTH
(b) For 29 July 1977
Figure 3-5 (Continued).
J
20
in
cr
UJ
10
<*9r 81 1 8
42
-------
20
10
NORTH
i—i—i—i—i—i—r~]—i
' ' J
1
OCHE
_J
VEF.fi
¥1500
SKIR
0600
0500
MUDS
I I I I I >..!.! < I > < '
SOUTH
(c) For 3 August 1977
Figure 3-5 (Continued).
81 18
43
-------
NCF.TH
20
10
11 i i i i i i i ; i i i j
rc*El500
VEF.P
nt;os
}
J
20
~1
t > • i i i i I i < i i
SOUTH
(d) For 2 September 1977
Figure 3-5 (Concluded).
to
cr
LU
10
81 1 8
44
-------
TABLE 3-3. INITIAL CONDITIONS AND EMISSIONS RATES USED IN THE
LEVEL II CBM/OZIPM CALCULATIONS FOR 21 JULY 1977
(a) Initial Conditions
Species
Surface
Aloft
NOX
NMOC
°3
0.003 ppm
0.163 ppmC
0.0356 ppm
0.003 ppm
0.13 ppmC
0.067 ppm
N02/NOX = 0.916.
Hydrocarbon reactivity:
OLE =
PAR =
ARO =
ETH =
CARB =
0.018
0.826
0.121
0.016
0.009
(b) Emissions Rates and Mixing Heights
Species
Time
(CDT)
0500-0600
0600-0700
0700-0800
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
1500
VOC
(ppmC/hr)
0.0002
0.0004
0.0024
0.0058
0.0144
1.2202
0.0723
0.0137
0.0041
0.0014
NOX
(ppm/hr)
0.0009
0.0037
0.1294
0.0573
0.0513
0.118
0.3333
0.0367
0.0113
0.0043
Mixing Height
at the Beginning
of Each Hour
(m)
50
50
50
50
100
200
450
600
600
600
700
81 10
45
-------
TABLE 3-4. INITIAL CONDITIONS AND EMISSIONS RATES USED IN THE
CBM/OZIPM CALCULATIONS FOR 29 JULY 1977
(a) Initial Conditions
Species Surface Aloft
NOX 0.0036 ppm 0.0016 ppm
NMOC 0.1356 ppmC 0.1066 ppmC
03 0.0326 ppm 0.0826 ppm
N02/NOX = 0.997.
Hydrocarbon reactivity:
OLE = 0.0409
PAR = 0.7886
ARO = 0.1296
ETH = 0.0320
CARB = 0.0089
(b) Emissions Rates and Mixing Heights
Mixing Height
Species at the Beginning
Time VOC NOX of Each Hour
(CDT) (ppmC/hr) (ppm/hr) (m)
0500-0600 0.0007 0.0002 50
0600-0700 0.0026 0.0007 50
0700-0800 0.0112 0.0031 50
0800-0900 0.0373 0.0076 50
0900-1000 0.0312 0.0076 100
1000-1100 0.1265 0.0340 200
1100-1200 1.4423 0.1751 300
1200-1300 0.8083 0.1949 600
1300-1400 0.3448 0.0766 700
1400 900
46
-------
TABLE 3-5. INITIAL CONDITIONS AND EMISSIONS RATES USED IN THE
CBM/OZIPM CALCULATIONS FOR 3 AUGUST 1977
(a) Initial Conditions
Species Surface Aloft
NOX 0.0323 ppm 0.0034 ppm
NMOC 0.2035 ppmC 0.1125 ppmC
03 0.0109 ppm 0.0702 ppm
N02/NOX = 0.999.
Hydrocarbon reactivity:
OLE = 0.0416
PAR = 0.7819
ARO = 0.1416
ETH = 0.0282
CARB = 0.0067
(b) Emissions Rates and Mixing Heights
Mixing Height
Species at the Beginning
Time VOC NOX of Each Hour
(CDT) (ppmC/hr) (ppm/hr) (m)
0500-0600 0.0098 0.0035 50
0600-0700 0.8317 0.0995 50
0700-0800 0.1633 0.0464 50
0800-0900 0.0610 0.0219 50
0900-1000 0.0335 0.0098 50
1000-1100 0.0240 0.0121 100
1100-1200 0.00 0.00 180
1200-1300 0.0632 0.0243 260
1300-1400 0.0866 0.0302 540
1400-1500 0.0220 0.0069 680
1500 800
81 10
47
-------
TABLE 3-6. INITIAL CONDITIONS AND EMISSIONS RATES USED IN THE
CBM/OZIPM CALCULATIONS FOR 2 SEPTEMBER 1977
(a) Initial Conditions
Species Surface Aloft
NOX 0.0065 ppm 0.0021 ppm
NMOC 0.2034 ppmC 0.1033 ppmC
03 0.0281 ppm 0.0712 ppm
N02/NOX = 0.9985.
Hydrocarbon reactivity:
OLE = 0.0364
PAR = 0.6913
ARO = 0.1610
ETH = 0.0845
CARB = 0.0268
(b) Emissions Rates and Mixing Heights
Mixing Height
Species at the Beginning
Time VOC NOX of Each Hour
(COT) (ppmC/hr) (ppm/hr) (m)
0500-0600 0.0054 0.0001 50
0600-0700 0.0113 0.0027 50
0700-0800 0.0273 0.0071 50
0800-0900 0.0264 0.0085 50
0900-1000 0.1616 0.0277 50
1000-1100 0.0288 0.0037 100
1100-1200 0.0017 0.00 100
1200-1300 0.0021 0.0003 600
1300-1400 0.0020 0.0004 1000
1400-1500 0.0031 0.0003 1000
1500 1000
it 9- 81 10
r 48
-------
For the Level II EKMA using propylene/butane chemistry, the standard
default reactivity of 25 percent propylene and 75 percent butane was used
along with the emissions and meteorological inputs discussed previously.
INPUTS USED IN THE LEVEL III EKMA MODEL SIMULATIONS
Emissions and meteorological inputs were developed as outlined in the
EPA (1980) guidelines document. All assumptions made for this study are
based on that document, which was in draft form during this study.
There are six input options available in the standard OZIPP Model
(light intensity, dilution, ozone transport, precursor transport, post-
0800 emissions, and reactivity). Emissions inventories for the Tulsa
study developed by Engineering Science (1980) were reported for each
county located in the study region. The county-wide emissions are
presented in table 3-7 along with the land area of each county. The NUX
and VOC emissions in units of kg/yr are converted to emissions densities
in units of kg/hr krrr (table 3-8).
Straight-line trajectories from the urban core to the location of the
maximum observed ozone were drawn and are shown in figure 3-6(a) through
3-6(d). As seen in these figures, the maximum observed ozone usually
occurred near Tulsa County for the four days.
Morning and afternoon mixing heights were determined from upper air
soundings taken at the airport (as discussed in appendix A). The initial
conditions used in the Level III EKMA calculations are presented in tables
3-9 through 3-12. Urban average 0600 to 0900 ambient concentrations of
NMOC and NOX are reported by Eaton et al . (1979a). Two monitoring
stations (Tulsa post office and health department) are located in the
urban core. The reported 0600 to 0900 NMOC and NOX concentrations from
these two stations were averaged and used to determine the post-0800
emissions fractions used in the Level III EKMA calculations for the four
modeling days. Ozone transport for each day was determined from observa-
tions made at the Liberty Mounds station located 37 km south of the urban
core. Eaton et al . (1979b) reported that the Liberty Mounds station was
influenced by the urban core 14 days out of the three-month study
period. Thus, Liberty Mounds represents the upwind station for the four
modeling days. Although NMOC and NOX precursors are not considered in
this section, on the basis of recommendations in the Level III guidelines,
the effects of using NMOC and NOX precursors (surface and aloft) in the
EKMA calculations are examined in section 6.
The assumed ozone concentrations aloft for each day were based on the
analysis by Eaton et al . (1979b), which compared aircraft measurements
si 2
-------
TABLE 3-7. COUNTY-WIDE EMISSIONS ESTIMATES FOR 1977
County
Tulsa
Washington
Creek
Osage
Okmulgee
Area
(km2)
1.48 x 103
1.10 x 103
2.42 x 103
5.88 x 103
1.81 x 103
Emissions
(kg/yr)
NMOC NOX
5.08 x 107 4.96 x 107
5.98 x 105 5.97 x 105
1.94 x 106 9.82 x 105
1.06 x 106 5.24 x 105
4.65 x 106 1.27 x 106
TABLE 3-8. CALCULATED EMISSIONS DENSITY FOR FIVE OKLAHOMA CITIES
County
Tulsa
Washington
Creek
Osage
Okmulgee
Emissions
(kg/hr km2)
NMQC
3.92
0.0621
0.0915
0.0206
0.293
NOV
3.83
0.062
0.0463
0.0102
0.0801
81 10
50
-------
NORTH
0
30
20
en
LU
TNO
SMfl
10
PC HE
1300 LCT
.VERR...... .
* 1200 LCT
/1100 LCT
; i 1000 LCT
L* 090° LCT
fflPCH
POsrOSOO LCT '
HLTH
MMDS
« I.I « 1 1 11 »
30
20
CO
cr
UJ
0
0
SOUTH
81 1 8
LCT = local civil time.
(a) For 21 July 1977
Figure 3-6. Level III straight-line trajectory
to maximum 0-.
51
-------
NORTH
0
20
IT)
LU
IS
10
< 1
7NO
-
- :
-
1 I
i i i t l l, 1 | i ( i t
C'CHE
VERfi ' . ' -
SKIR ; . ~
SPRY
• ~
i
HHHKT300 LCT
: flrCH - ^
1
- MMDS ""
1 1 1 J 1 ! J 1 1 1 ' ! 1 "V
20
en
ex
LJ
SOUTH
(a) For 29 July 1977
Figure 3-6 (continued).
81 1 8
52
-------
NORTH
ID
LU
0
^
P
0
rc
•• v-'T"11!' k i } ! ' 1, i | l i . i l
OCHE
7 NO
1500 LCT
^ • VEFfi . -
i A- 1400 LCT
yLisoo LCT
SKIfi ' /
SPRY Li 200 LCT _
L 1100 LCT _
A-1000 LCT _
hUi/-0900 LCT
7nFCH
P^T^ 0800 LCT '
; "" HLTH
_ —
t i i ... i i i i i . » J i ) • i ... ».V.N.
30
20
1C
171
CO
cr
LU
U9r 81 1 8
SOUTH
(c) For 3 August 1977
Figure 3-6 (continued).
53
-------
NORTH
»*9r 81 i e
0
30
20
in
loJ
10
-
""
~
Two
_
,_
_
~
_
-
'"•-
-
— »
-
:
1500 LCT ' -
OtflE
j
.-1400 LCT
-1300 LCT : _
; VERB ' . '•' -
T1200 LCT .-. , _
SKin \1100 LCT
SPRY
(ilOOO LCT •
j i
;io9oo LCT : ". _
!| RPCH ;
: DRDO L CT
HLTH
•«•
i
i J 1 J J 1 J 1 t J : i » ''"
30
20
en
cr
LJ
10
0
SOUTH
(c) For 2 September 1977
Figure 3-6 (concluded).
54
-------
NOX
NMOC
°3
0.0 ppm
0.0 ppmC
0.0 ppm
TABLE 3-9. INITIAL CONDITIONS USED IN THE LEVEL III EKMA
CALCULATIONS FOR 21 JULY 1977
(a) Initial Conditions
Species Surface Aloft
0.0 ppm
0.0 ppmC
0.07 ppm
N02/NOX = 0.25.
Hydrocarbon reactivity: standard OZIPP conditions,
Morning mixing height: 150 m at 0800 CDT.
Afternoon mixing height 600 m at 1200 CDT.
NMOC/NOX at 0600 to 0900 CDT = 6.15.
NMOC at 0600 to 0900 CDT = 0.283 ppmC.
NOX at 0600 to 0900 CDT = 0.046 ppm.
Design 03 = 0.14 ppm at Vera (1300 CDT).
(b) Emissions Fractions
Time
(CDT) NMOC NOX
0800-0900 0.155 0.294
0900-1000 0.155 0.294
1000-1100 0.155 0.294
1100-1200 0.155 0.294
1200-1300 0.079 0.150
81 10
-------
TABLE 3-10.
INITIAL CONDITIONS USED IN THE LEVEL III EKMA
CALCULATIONS FOR 29 JULY 1977
(a) Initial Conditions
Species
NOX
NMOC
Oo
Surface
0.0 ppm
0.0 ppmC
0.0 ppm
Aloft
0.0 ppm
0.0 ppmC
0.08 ppm
N02/NOX = 0.25.
Hydrocarbon reactivity: standard OZIPP conditions,
Morning mixing height: 150 m at 0800 CDT.
Afternoon mixing height: 900 m at 1400 CDT.
NMOC/NOX at 0600 to 0900 CDT = 9.09.
NMOC at 0600 to 0900 CDT = 0.482 ppmC.
NOX at 0600 to 0900 CDT = 0.053 ppm.
Design 03 = 0.147 ppm at Apache (1400 CDT).
(b) Emissions Fractions
Time
(CDT)
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
NMOC
0.091
0.091
0.091
0.091
0.091
0.091
0.255
0.255
0.255
0.255
0.255
0.255
81 10
56
-------
TABLE 3-11.
INITIAL CONDITIONS USED IN THE LEVEL III EKMA
CALCULATIONS FOR 3 AUGUST 1977
(a) Initial Conditions
Species
Surface
Aloft
NOX
NMOC
03
0.0 ppm
0.0 ppmC
0.0 ppm
0.0 ppm
0.0 ppmC
0.09 ppm
N02/NOX = 0.25.
Hydrocarbon reactivity: standard OZIPP conditions,
Morning mixing height: 150 m at 0800 CDT.
Afternoon mixing height 800 m at 1500 CDT.
NMOC/NOX at 0600 to 0900 CDT = 12.1.
NMOC at 0600 to 0900 CDT = 0.598 ppmC.
NOX at 0600 to 0900 CDT = 0.05 ppm.
Design 03 = 0.15 ppm at Vera (1500 CDT).
(b) Emissions Fractions
Time
(CDT)
NMOC
NO,
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
0.074
0.074
0.074
0.074
0.074
0.074
0.037
0.270
0.270
0.270
0.270
0.270
0.270
0.138
81 10
57
-------
TABLE 3-12.
INITIAL CONDITIONS USED IN THE LEVEL III EKMA
CALCULATIONS FOR 2 SEPTEMBER 1977
(a) Initial Conditions
Species
Surface
NOX
NMOC
°3
0.0 ppm
0.0 ppmC
0.0 ppm
Aloft
0.0 ppm
0.0 ppmC
0.07 ppm
N02/NOX = 0.25.
Hydrocarbon reactivity:
Morning mixing height:
Afternoon mixing height
standard OZIPP conditions,
150 m at 0800 CDT.
1000 m at 1300 CDT.
NMOC/NOX at 0600 to 0900 CDT = 8.75.
NMOC at 0600 to 0900 COT =0.52 ppmC.
NOX at 0600 to 0900 CDT = 0.059 ppm.
Design 03 = 0.11 ppm at Ochelata (1500 CDT).
(b) Emissions Fractions
Time
(CDT)
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
NMOC
0.085
0.085
0.085
0.085
0.043
0.0013
0.0013
NOX
0.229
0.229
0.229
0.229
0.117
0.004
0.004
81 10
58
-------
with the ground-level ozone measurements. Eaton reported that in most
cases, aircraft measurements of ozone are higher than the ozone measure-
ments at the surface. For two of the aircraft measurement days, the
difference between the airborne ozone measurements and the Liberty Mounds
measurements was less than 0.01 ppm. On both of these days, the measured
wind speeds at the health department during the time of the aircraft
measurements were greater than 13 mph. On the other aircraft measurement
days, the differences in ozone concentration between the aircraft and
surface measurements were in excess of 0.03 ppm. The wind speeds measured
at the health department were 6 mph or less.
For the Level III EKMA inputs, we based the ozone concentration
values aloft on the findings by Eaton. We expected that, at a minimum,
these concentrations would be greater than the 1100 to 1300 CDT Liberty
Mounds measurements by about 0.02 ppm and set the aloft ozone values
accordingly. For 3 August we kept the ozone aloft value equal to the
surface observations since the wind speeds at Liberty Mounds were fairly
high for this day compared with the other days.
To study the differences between the propylene/butane chemistry
currently employed in the EKMA with the Carbon-Bond chemistry, we replaced
the propylene/butane chemistry with the Carbon-Bond Mechanism without
making any other changes to emissions and meteorological inputs. A
default reactivity similar to the default 25 percent propylene reactivity
of the standard OZIPP chemistry was used in all the Level III EKMA/CBM
calculations. The following reactivities were taken from Killus and
Whitten (1981):
OLE 0.03
PAR 0.65
ARO 0.22
ETH 0.05
CARB 0.05.
81 2
59
-------
SECTION 4
COMPARISON OF MODEL RESULTS
This section presents a comparison of the absolute predictions for
each of the models discussed in the previous chapters. In almost all
cases, these models use meteorological and emissions inputs discussed in
section 3; specific exceptions are noted at the time of their discus-
sion. A summary of the ozone predictions for the four models is presented
in table 4-1.
SAI AIRSHED MODEL RESULTS
Ozone predictions and model performance for the SAI Airshed Model
have been discussed in detail by Reid, Reynolds, and Latimer (1980). We
wish to summarize here some of these performance evaluations for the four
modeling days. Table 4-2 presents the effects of some of the performance
measures on peak ozone concentrations. These measures were based on
comparisons of the maximum observed concentrations with the highest value
computed at a particular monitoring station even though the two values may
not have occurred at the same time.
As shown in table 4-2, both the median deviation and mean deviations
for 21 July 1973 are less than 0.01 ppm, indicating little bias in the
results. Since the median and mean absolute deviations are less than
0.015 ppm, the gross error in the peak predictions is also small. When
the absolute deviation is normalized by the measured value, we find that
the median and mean discrepancy between the computed and observed peak
concentrations is 14 and 19 percent, respectively. The ratio of peak
computed to peak measured concentrations can also be considered; the
median and mean values for this comparison are 0.99 and 0.94, respec-
tively, as indicated in table 4-2.
In table 4-2, we also see that the highest measured and predicted
ozone concentrations differ by less than 0.01 ppm for 29 July 1977.
Although little bias is indicated in table 4-2 for this day, the gross
error in predictions is slightly higher than that for 21 July 1977. For 3
August 1977, the Airshed Model tended to underpredict ozone, as shown by
H9>" 81 6
60
-------
TABLE 4-1.
Model
COMPARISON OF OZONE PREDICTIONS FOR THE SAI AIRSHED, SAI
TRAJECTORY, LEVEL II, AND LEVEL III EKMA MODELS—1977
SAI Airshed
SAI Trajectory
Level II EKMA
Level II w/CBM1"
Level III EKMA
Level III w/CBM§
(ppm)
Ozone Level
21 July
29 July
3 Aug
2 Sept
(Obs 03=0.14) (Obs 03=0.17) (Obs 03=0.15) (Obs 03=0.11)
0.14
0.15
0.10
0.15
0.10
0.12
0.17
0.14
0.11
0.14
0.12
0.15
0.12
0.15
0.10
0.15
0.13
0.18
0.12
0.09
0.08
0.09
0.10
0.13
OBS = observed.
' BZA and PAN entrained from aloft are used (see the text for further
discussion).
§ CBM (Carbon-Bond chemistry).
81 9
61
-------
TABLE 4-2. SUMMARY OF PEAK OZONE CONCENTRATION MODEL PERFORMANCE
EVALUATION MEASURES
Performance Measure 21 July* 29 July* 3 August* 2 September*
Highest observed value at
peak station*
Highest computed value at
peak station
Highest computed value at
any station
Highest computed ground-
level value on the grid
Deviation for all stations
Median
Mean
Standard deviation
N*
Deviation for all stations
with observed value
> 0.12 ppm
Median
Mean
Standard deviation
Nf
Normalized deviation for
all stations
Median
Mean
Standard deviation
N*
0.135
0.135
0.135
0.14
-0.003
-0.006
0.021
10
0.000
0.000
--
1
0.01
-0.06
0.26
10
0.166
0.174
0.174
0.21
-0.006
-0.002
0.024
10
0.008
0.010
0.041
3
-0.06
0.01
0.22
10
0.151
0.115
0.143
0.16
-0.007
-0.010
0.029
10
-0.014
-0.024
0.025
5
-0.04
-0.07
0.26
10
0.108
0.120
0.128
0.13
0.011
0.012
0.019
10
--
--
—
0
0.10
0.18
0.28
10
81 18
62
-------
0.012
0.015
0.015
10
0.014
0.018
0.015
10
0.021
0.022
0.020
10
0.016
0.018
0.011
10
TABLE 4-2 (Continued)
Performance Measure 21 July 29 July 3 August - 2 September
Normalized deviation for
all stations with observed
value > 0.12 ppm
Median 0.00 -0.24 -0.12
Mean 0.00 0.08 -0.19
Standard deviation -- 0.11 0.20
Nf 1350
Absolute deviation for all
stations
Median
Mean
Standard deviation
Absolute deviation for all
stations with observed
value > 0.12 ppm
Median 0.000 0.030 0.014
Average 0.000 0.030 0.025
Standard deviation -- 0.022 0.024
Nf 1350
Normalized absolute devia-
tion for all stations
Median
Average
Standard deviation
Normalized absolute devia-
tion for all stations with
observed value > 0.12 ppm
Median 0.00 0.24 0.12
Average 0.00 0.24 0.19
Standard deviation -- 0.19 0.20
Nf 1350
81 18
63
0.14
0.19
0.18
10
0.16
0.17
0.12
10
0.18
0.20
0.17
10
0.22
0.26
0.19
10
-------
TABLE 4-2 (Concluded)
Performance Measure
Ratio of highest computed
value to highest observed
value at each station
Median
Average
Standard deviation
Ratio of highest computed
value to highest observed
value at all stations with
observed value > 0.12 ppm
Median
Average
Standard deviation
N*
Correlation coefficient
for all stations
21 July* 29 July* 3 August* 2 September*
1.00
1.00
--
1
1.05
1.08
0.34
3
0.88
0.81
0.20
5
0.99
0.94
0.26
10
0.92
0.96
0.26
10
1.01
0.93
0.21
10
1.11
1.18
0.28
10
0.71
0.84
0.22
0
0.63
All concentration values are expressed as ppm.
N refers to the number of comparisons.
81 18
64
-------
the median and mean deviation for all stations with observed ozone values
greater than 0.12 ppm. The gross error (reflected in the normalized
median and average deviations) is higher for 3 August than for either 21
July or 29 July and is even higher for 2 September.
The bias in model predictions can be examined by computing the mean
deviation and mean normalized deviation of hourly pairs of predicted and
observed concentrations. The results of this exercise are presented in
table 4-3 for all four modeling days. For the 21 July 1977 day, the model
tended to underestimate ozone levels by about 17 percent. Considering
only those comparisons for which the measured value exceeded 0.12 ppm, we
found that the model showed a tendency to overpredict by about
4 percent. For 29 July 1977 and 3 August 1977, the model tended to
underpredict ozone by 14 and 23 percent, respectively, for all measured
ozone and by 20 and 22 percent, respectively, for measured ozone values
greater than 0.12 ppm. In contrast, the model tended to overestimate
measured ozone by 8 percent for 2 September 1977 (an attainment day).
COMPARISON OF SAI AIRSHED AND SAI TRAJECTORY MODEL RESULTS
As discussed in section 3, the SAI Airshed and SAI trajectory models
use the same meteorological and emissions inputs. However, the SAI
trajectory model is a two-dimensional model that uses only the first
vertical level of winds to describe the specific trajectory path. The
trajectory model follows the path of an air column the size of a single
grid cell using the same computer codes as the Airshed Model except for
those that involve horizontal dispersion, vertical winds, and wind
shear. We expect the SAI Airshed and SAI trajectory models to predict
similar ozone values for those days with low vertical winds and wind
shear. However, the SAI trajectory model could also predict results
similar to the SAI Airshed Model for those modeling days with high wind
shear. This could be coincidental if the emissions densities are evenly
distributed spatially or if the three-dimensional path of the air parcel
in the Airshed Model never passes over an area with high emission peaks.
Thus, SAI trajectory model results need to be analyzed in detail to
understand discrepancies with SAI Airshed predictions.
The SAI trajectory model results shown in table 4-1 for the four
modeling days were obtained without incorporating surface sinks. Such
results are more comparable with the CBM/OZIPM EKMA results discussed
later. The use of surface sinks can lead to a 10 to 15 percent reduction
in ozone predictions over those of the SAI models. Therefore, we expect
the trajectory model to predict higher ozone values than the Airshed
Model. For 21 July 1977, the SAI trajectory model predicted a maximum
ozone concentration of 0.138 ppm when surface sinks were used, and a value
81 6
65
-------
CTi
CM
CM
C/}
UJ
o;
Z3
oo
•a:
UJ
21
O
I—t
s
UJ
o
to
C£
oo
U.
o
en
"3
CM
1_ •
O O
*-
A
OJ
3 f
— O
§&
<0 CX
£CM
O
o d
,= 0^
•o
(o
QJ
^-
t—I f—I
o o
• •
o o
CM
en
•
o
o
VO CM
co co
• *
o o
CO
•
o
o
^ u") r^
• • CO
00 r-
o_ o
d o
CX CX
U"> tf>
o o
o o
• *
o o
o o
* •
o o
HH
CX CX
ID LT>
o o
o o
* •
o o
CO
r- co
r-t CO
• •
o o
nn
CX CX
O ^H
o o
• •
o o
I
CO VC
CO CM
• •
o o
EH
CX CX
LO ,-H
f—I I—I
o o
* •
o o
CO
en
•
o
o
4->
*-• r^
r^ CO
d o
o>
CO
d
o
CO CO
co co
• •
o o
ro
•a
(U
3
C/l
00
ro
LU
_l
co
OJ
U
c
«B
t
O
<4-
QJ
O-
OJ
s.
3
W1
CJ>
IB
t-
QJ
>
QJ
•o
•
•o
*J
c^>
1
-o
QJ
N
ra
o
c
o
o>
0)
O L.
ft)
ex,
> •«->
OJ I
•«-> -o
3 OJ QJ
t— IM CT
O C f *
I/I O r— i-
J3 f- 10 QJ
E >
•O fc. e£
I
•a
i^ Q> QJ
'I S"
. £ 1. •
I S QJ TD
I U
I CS> QJ
ro *J >
f— C «
OJ QJ C
i. f- O C
t. O "- O
O -^- *J -^-
(J U_
l«- 4^ *->
IB O t/>
U O JZ I
o u •—
tc 10 —
O UJ
O) I >
u in U
"O
C
(O
•o
OJ
1/1
O i-
4->
OJ
U)
o
0)
o
*£ O
U i.
«S t.
—1 QJ
o
(B
-------
r—
r^
CTl
CM
•o
I
o.
fC D.
, L. •
O O
•g
I
•O
o o
00 ID
O PO
• •
O C
o. o.
in oo
O r-l
o o
o o
Co c\j
• •
o o
O. Q.
*3- CO
r—< «—I
O O
• •
o o
en
•
o
o
t-i O
ID OO
O O
00
o
o
4->
*f fH
PO in
* *
O C
OJ
-o
"o
C
o
o
CO
I
03
T5
0>
S_
3 §1
•O CX
^CO
v^
J. •
0 0
*4_
A
r-v i!
r^ s rT
g 5°
4->
V>
01 "0
VI Mj
3 L
11 = E
<*> S8 !
QJ
£ co
o
L.
0 O
**_
A
QJ
a PO
— o
«o
2»
Ml. Pi.
ex ex ex ex
ex ex ex ex
en co en co
coin co.— i coin co.— i
CO «— I OO CO.— I OO
• • •• •• ••
OO OO OO OO
Hi H H
k^ U- U^- LiX
ex ex ex ex
r^ *t in in
por^ ^-ico vo«3- co^-i
COPO CO POCO OO
• • •• •• ••
OO OO OO OO
1 1
OJ
o
e a)
•0 fc.
E 3
L. i/>
O «
<*- OJ
t- £
O)
S.
1 1
1 1
•a -a
C 1 QJ 1 QJ
Ol • N • Q) 1 "N *
t- -o > •>- > 4-» T3 >T- >
4-> QJ QJ QJ ' — QJ QJ 3 QJ QJ QJ i — QJ QJ
•OMCD'CJ'O CT)-O »— »MCT>-D(CO1T3
f-^-«O E«O OC-'-'W E<"
> i— i. .S-l- • in O •— fc- -S-l-
QJ re QJ "0 o Q)^ ^'^-fooj^o QJ^
T3 E > -t-J C >-4-> E>4->C>4->
i_«OOC<;oO «Ol_«:OOEctOO
CO O C ••- O O
•o z :z •« > z z:
QJ QJ QJ
z: s: -D
•0 3
E Xi
E i-
o u
«t- 4->
L. 4->
QJ <
ex.
CO
1C
f- 10
X> i^
O
««-(_> i.
O -i- CD
4->
QJ «O »4-
L> E 0
CO) l-
QJ 4-» ^ O
i« en U fc-
J3 >> «» t-
et to -JO)
It II
i! ! ! r^
CO PO
en in
o d
o o
4-> -t->
PO in o ID
VD co in i— i v£>
do d o |Z
QJ
en
i re c
1 S_ O QJ
i 1/1 QJ •<- en
re -t-j > •*-> re
r- C «3 <0 S_
QJ QJ C •— 1 QJ
l_ f- O C QJ 1 >
L. O •'- O i- l/l «5
O •'- *J -i- J- 4-> S-
(_; M- re *J OC31.
l)_ *-> re (_) QJ O 3
t— QJ l/l 4-> -t- .C O
re o i/> •— o ^:
1_ (_> XT 1 OO U
c
o
•>- 4->
•— *-> c
IB W r— O)
I- •— »o E
O QJ t- C
CX I- 4-> O>
E J- « ••- ^
QJ O O. •— *
>— o oo «o ^~
.
on
QJ
3
re
>
-a
QJ
S-
3
l/l
re
QJ
E
•o
c.
ro
"D
QJ
+->
3
C-
E
o
u
It-
CD
I/)
s_
•r—
ro
CL
lf-
o
s_
QJ
JD
E
3
•z.
*
67
-------
of 0.154 ppm (table 4-1) without the use of surface sinks. This ozone
value of 0.138 ppm compares very well with the SAI Airshed Model predic-
tion of 0.14 ppm.
For 29 July 1977, the trajectory model ozone prediction of 0.14 ppm
(without surface sinks) was approximately 17.5 percent lower than the SAI
Airshed Model prediction of 0.17 ppm. Since wind data for 29 July show
high wind shear within the mixed layer during the morning hours, we cannot
conclude that the trajectory path to the maximum ozone concentration was
similar to that followed in the Airshed Model simulations for the air
arriving at the point of maximum ozone. Section 6 discusses methods of
determining alternative trajectory paths for this modeling day in place of
the use of surface winds only. We note that trajectory paths to cells
adjacent to the cell in which the monitoring station is located show wide
variations in ozone predictions. An ozone maximum of 0.17 ppm was found
in the cell just northeast of the cell that contained the monitoring
station.
For 3 August 1977, the ozone value predicted by the SAI trajectory
model was 25 percent higher than the 0.12 ppm ozone value predicted by the
Airshed Model, apparently because surface sinks were not incorporated in
the trajectory model simulation.
For 2 September 1977, the SAI trajectory model predicted lower ozone
concentrations than the Airshed Model, even though no surface sinks were
used in the trajectory model simulation. Investigation of the wind fields
used in the Airshed Model within the mixed layer showed some shear during
the early-morning hours. The first-level wind field showed low speeds,
but the second-level winds were, on the average, twice as fast (with a 10*
difference in direction).
COMPARISON OF SAI TRAJECTORY AND LEVEL II EKMA MODEL RESULTS
This section is divided into two parts: the first presents a
discussion of the predictions from the SAI trajectory model and the Level
II EKMA with Carbon-Bond Chemistry (CBM/OZIPM). The second presents a
comparison of ozone predictions for the Level II EKMA using propylene/-
butane chemistry with those of the CBM/OZIPM.
Although the results shown in table 4-1 for the SAI trajectory model
and the CBM/OZIPM are in close agreement, it was necessary to consider
entrainment of BZA and PAN (two species in the Carbon-Bond Mechanism) from
aloft. When we first began CBM/OZIPM calculations using the inputs
described in section 3, we found the CBM/OZIPM predictions consistently
low when compared with ozone predicted by the SAI trajectory model.
81 6
68
-------
Although the CBM/OZIPM predictions were within 10 percent of the SAI
trajectory model, we felt that the CBM/OZIPM could provide better agree-
ment. Detailed investigations of the predicted concentration profiles
from the SAI trajectory model and the CBM/OZIPM model indicated a major
discrepancy in BZA and PAN predictions between the two models. Upon
reexamination of the chemistry and the meteorological conditions for the
four modeling days, we discovered that the introduction of BZA and PAN
into the mixed layer from aloft was an important factor in the SAI
trajectory model simulations. The amounts of BZA and PAN simulated aloft
for these four modeling days are presented in table 4-4. The multilayer
SAI trajectory model continuously simulates the chemistry in each layer.
At any instant in time during the daylight hours, some of the overall
reactivity of an air parcel is contained in intermediate species, which
are equilibrated with the reacting system of sunlight and precursors, such
as PAN and BZA. Failure to entrain such simulated intermediates from
aloft necessitates the use of some reactivity within the mixed layer to
replace these intermediates to their equilibrated levels. Since the
standard version of CBM/OZIPM did not account for these two concentrations
from aloft, we modified the inputs to the CBM/OZIPM model and introduced
BZA and PAN as if they were part of the hydrocarbons found aloft for each
day. Table 4-5 presents the results obtained when BZA and PAN are
entrained in this manner. We see that the entrainment of BZA and PAN in
the CBM/OZIPM model produces results that are in much closer agreement
with those of the SAI trajectory model. For all CBM/OZIPM relative
prediction calculations (discussed in section 5), we included entrainment
of BZA and PAN. Our understanding of atmospheric chemistry indicates that
some such intermediates must exist during the day, even though they may
not be measured. Virtually any modern photochemical mechanism should
respond in a fashion similar to that of the CBM used in this study, so the
results are not considered to be specific to the CBM. Also, this addition
to the CBM/OZIPM model is not considered to be an exercise in curve-
fitting. One of the objectives of this project has been to determine the
specific differing features of the models that explain significant
differences in their results. This entrainment feature was added to the
CBM/OZIPM to confirm our postulation that such a difference between the
SAI trajectory model and the CBM/OZIPM would explain much of the differ-
ences in the result of the models.
A further comparison of concentration predictions between the SAI
trajectory model and the CBM/OZIPM is shown in figure 4-1, which presents
comparisons of NOX, PAR, ETH, BZA, PAN, and ozone for 29 July 1977. As
indicated in this table, species concentrations from both models are in
close agreement. Additional comparisons of ozone concentrations predicted
from the CBM/OZIPM calculations were made with data observed at monitoring
stations along each trajectory. The results of these comparisons are
shown in figure 4-2.
81 6
69
-------
TABLE 4-4. ALOFT CONCENTRATIONS OF BZA AND PAN USED IN THE
SAI TRAJECTORY MODEL AND THE CBM/OZIPM MODEL
(ppm)
Date—1977 BZZ PAN
21 July 0.001 0.001
29 July 0.0009 0.001
3 August 0.0011 0.0014
2 September >0.00005 >0.00005
i»9r 81 9
70
-------
Q.
D.
0.20-,
0.16-
0.12-
o 0.08-1
0.04-
0.0
* CBM/OZIPM
• SAI TRAJECTORY
0600
I
0800 1000 1200
Time (Hours, CDT)
1400
0.20-
0.16-
£0.012-J
C-
0.08-
0.04-
0.0-
* CBM/OZIPM
• SAI TRAJECTORY
0600
0800 1000
Time (Hours, CDT)
1200
1400
Figure 4-1. Comparison of NOX, PAR, ETH, BZA, PAN, and ozone
for the SAI trajectory and CBM/OZIPM models for
29 July 1977.
71
-------
0.005 -,
0.004 _
§. 0.003
Q.
0.002 -
0.001 -
0.0
CBM/OZIPM
SAI TRAJECTORY
0600
I I I
0800 1000 1200
Time (Hours, CDT)
1400
£
CL
-------
O.
Q.
0.0060-,
0.0048-
0.0036-
< 0.0024-^
Q-
0.0012-
0.0
* CBM/OZIPM
• SAI TRAJECTORY
0600
0800 1000 1200
Time (Hours, CDT)
1400
Q.
CL
UJ
O
O
U. C-V —
0.16_
0.12-
0.08-
0.04-
0.0
* CBM/OZIPM
• SAI TRAJECTORY
*
t *
t
•
0600 0800 1000 1200 1400
Time (Hours, CDT)
Figure 4-1 (concluded).
73
-------
^ »• » •+• ™" "• — "
c
o
o
5
1
e
o
•
•
.
•
CO UJ
z r>
0 -I
« «t
> o
Q£ (jj
LkJ 1—
»• CO O
CD •-.
0 0
LiJ
Z DC
O O.
P «t
<: ^
H- i<£
* O
_ _ 4 _ _ 4. _
*3 O
S N
•1 -
r o
_..._- -- -+ T - + - , +- ^ ~ .- i- - . T - +--•-
•
§
c *
.
o
c
0
©
o
0
c
§c
o
c
o
- *
0
c o
o
c
o
o
c. *
c
o
c
%
c
c
" c
o
-*•
5e
e
^
o
S
5
5
5
2
^^
5
o
^
o
o
o
0
2
s
^+— _ + —— + —— + — — + — — + — — >4— — *.— •-«...— + ——+.— _
^ ^ r^ C3 ^ '
•? — a w ci <
— - o o o <
9 o o e e
o
o
0
o
'
o
0
CO
^~
0
c
C\J
r—
o
o
r— •
, ^
f— ["*•»
/^ p*^.
0 0 OT
§ *
1 ^ rB_
3 3
O "-3
r—
CsJ
0 &
O = __
cr> ,_
o !~ jv.
o
o
CD
O
O
O
0
0
o
1-3
o
0
3 'C>
9
(Si
C
o
(J
•r—
-o
OJ
i.
D.
_.
a.
M
0
CO
o
-^
c
(O
a^
c
o
M
O
-a
a>
CD
to
O
C
o
•r-
4-3
03
to
C
OJ
O)
0)
-°
c
o
to
(O
a.
Q
0
Cs3
1
O)
CT>
(tudd) auozQ
61 18
74
-------
+ + - - _.-;_. _;-- + ---r : --- '- - - -r - -
*" C
O
* c
c
c
c
c _
c
r-
C
C
C
C
C
§*
5
^
"•
1
C
3
c
£ -^
/»•
^ ^
X
c
c
c
— .
c ^
c
c _
c c
00 C
0
^ <
K-* OO
> O
UJ 1—
oo o
CO i— i
4. O O
UJ
z cr:
C 0.
" el
»— ^
OO LU
•
* O
• O rt vS """ C K
5 0 « C = LI
- - — C C
s o o c e o
c _
5
c
c
c
c
c
c
c
w
c
c
c
c
c
c
ii
r- '
c :
= -
o
' O
LT>
o
o
. *3"
'
o
o
I- OO
o
o
>• c\j
o
o
h ^_
^— f*^.
O O^t
0 i—
o
O - >i
. O t/> i—
r- S- 3
o
jr CTI
OJ
0 E —
O -i- -Q
- o-, I— —
o
o
o
CO
•• o
o
^
o
- 0
o
§
o
o
'. 0
.
CD
^3
c:
o
o
OO
1
o>
S-
3
en
(ujdd) auozo
81 II
75
-------
*--*---*
6
§
c
c
o
Q£ i—
UJ t—
00 O
CO >—
O O
UJ
z cr
O O.
e>
ci
n —
2 £
e o
(Uldd) 9UOZQ
c
o
w
I
o
o
LTJ
o
o
o
o
ro
O
O
CM
O
o
Q i—
O
4->
g« 3
°3 ?
O ct
OJ
E
gP
cri
o
o
o
CO
o
o
o
o
§
o
o
o
•o
O)
c
o
u
CM
i
81 18
76
-------
^"
•
(/>
sy
0
^•r
^"i.
ce
UJ
+ oo
CO
O
z
o
^•M
1^ J). MT «•+»•»«
-5 N fl «C — « <
e> i. * * n - <
c o c c e o <
c o e o ~ c t
o
0
1 /^
L* J
o
0
•=d-
o
o
-------
A second set of calculations was made with the same Level II EKMA,
using the propylene/butane chemistry instead of the Carbon-Bond chemis-
try. All other inputs remained the same as those discussed in section
3. No attempt was made to resolve differences in ozone predictions
between the SAI trajectory model and this version of Level II EKMA because
of the different chemistries. The absolute ozone predictions (shown in
table 4-1) obtained with the propylene/butane chemistry are, in general,
lower than those obtained from the CBM/OZIPM. These Level II EKMA
predictions are also lower than the CBM/OZIPM predictions when BZA and PAN
are not entrained (see tables 4-1 and 4-5). Investigative analysis
indicates that one explanation of the lower predictions from this Level II
EKMA model could be found in the outdated reaction rate constants in the
propylene/butane mechanism (see section 6). Figure 4-3 shows the temporal
ozone profile for the four modeling days along with observations made
along the trajectory path.
LEVEL III EKMA MODEL RESULTS
Since the Level III EKMA, SAI trajectory, and Level II EKMA model
inputs are developed through different methodologies, we will not compare
their absolute ozone predictions with those of previous models. Instead,
we compare Level III EKMA ozone predictions with observations and with
results obtained with the use of a different chemical mechanism in the
Level III EKMA models. As seen in table 4-1, the Level III EKMA ozone
predictions tend to be lower than the observed maximum ozone for all of
the four modeling days. Note that the city-specific EKMA predictions are
also usually lower than those of preceding model predictions. This result
may be due in part to the conservative assumptions made in estimating NMOC
levels aloft.
Figure 4-4 shows the ozone temporal profiles for each of the modeling
days along with station observations obtained along the straight-line
trajectories. A second set of Level III EKMA model calculations obtained
using the Carbon-Bond chemistry shows results similar to those obtained
with the Level II EKMA calculations using Carbon-Bond chemistry (when
compared with observations). In most cases, the Level III EKMA/CBM
predictions tend to be higher than those of the standard Level III EKMA.
Figure 4-5 shows the temporal ozone profiles for each of the modeling days
along with station observations obtained along the straight-line trajec-
tories. Analysis of the detailed computer calculations shows that the
results can be explained by the aldehyde concentrations and photolysis
rates used in the propylene/butane mechanism, which are much higher than
those used in the CBM. The CBM reactivities were taken from the Level II
EKMA study, whereas default reactivity was used for the propylene/butane
mechanism.
81 6
78
-------
TABLE 4-5. COMPARISON OF SAI TRAJECTORY AND THE CBM/OZIPM MODEL
OZONE PREDICTIONS
(ppm)
Predicted 03 Predicted 03
Predicted 03 (CBM/OZIPM) (CBM/OZIPM)
Date--1977 (SAI Trajectory) w/o BZA and PAN) with BZA and PAN)
21 July 0.154 0.138 0.154
29 July 0.138 0.125 0.140
3 August 0.150 0.138 0.150
2 September 0.091 0.091 0.091
81 9
79
-------
o
o
o
o
IS)
s
N
O
o
C5
o
o ®
(uidd) auozQ
1 •—
1
1
1
1 O
1 0
+ «*•
1 "~
1
1
1
1
1 0
I"
1 , —
1
1
1
I o
0
+ ^
1 ^
i
1
1
1
1
1
0
! o
+ •—
f ^^.
i tr
i i
: s{
t 2 =
! S
\ *-*
1 0>
I E
1 0 T-
1 O I—
+ en
1 O
i
i
i
i
i
" 0
i °
+ CO
o
1
1
1
1
1
| 0
0
i°
1
1
1
1
1
! o
0
+ s
!-.
r^
en
r—
>>
3
•-3
i—
CM
•" ^*
(O
*••-''
re
o
o
o
O)
(O
UJ
QJ
>
-------
CO
2
1—
£
tt
IO
CD
O
g
§
V)
«
e
•
O
UJ
o
o
Lu)
a:
a.
UJ
O
+ — —
o
® _
Q
§§
,
0
o
o
0
o
§
* c
0
§l
o
o
1
o
s
0
§
o
§
o
o
§
0
§
0
o
0
o
0
§
1
5
o
0
5
£
1
1
o
1
Xh
^ N C CO ^ ^ <
*•* ^ J* Ift W *^ I
— e c c e e <
• • • • • •
A d«s A A 4fe
? **"**
»
fc
^_
•D
ID
s
c
+J
c
o
«-^
oo
'
•53-
QJ
3
en
(iudd) euozo
i+9r 81 18
81
-------
8
8
8
O
o
8
i
0
i
6
O
8
0
o
^^ *
8
o
o
o
o
o
o
o
1
o
o
o
o
o
o
0
° Q
o
0
8
8
8
o
* o
GO O
CO UJ 4. O
| g g
*"- > §
2» ^» o
^? ^1 ®
to o i
00 •-" 0
00 X
UJ o
SC OC £i
^3 ^^ Q
+ *~ S p°
1 ^^ 3fc* O
^" «b O
«^ uj e
* o 1
i CO O PI * * CO <
N O e N « « - <
2 2 • e e • ® <
e » « e • e ® <
(uidd) auozo
tt9r 81 18
82
o
o
LO
0
o
^™
o
§
0
o
CM
O
O
r~
^~
r**^
^™»* f***1
l_— CT>
o •—
•4-*
o * ^
0 10 3
C3 L- ^*
"~ 0 <
^^^ ro
^
£ ^— *.
0 -^ Ji
O ^— %1— ^
0
o
0
CO
o
o
o
f-x.
o
o
•S
o
CD
o
tn
9°
9
9
s>
*~~*
TD
CU
3
C
^3
c
o
o
09
i
CU
J_
3
CD
-------
§
o
o
o
o
g
c
+ »
0
1
o
c _
o
c
o
o
o
0
»
o
s
o
o
f
IS) UJ
O _J
«-> et
H- >
^^
£ S
LU H-
CO «-i
0 0
l~lJ
Z CtL
O D-
•— i
^ I
\- s
C/) UJ
* O
h_—4— — 4— _4— _4— — 4— — 4— — 4— — 4_— .
^ o> « n e
^ N » » *
t e © o ®
• • • • •
s 0 • © •
+ 9r 81 18 (^J 9UOZ0
-
•
•
•
•
5
°O
O H
O
o
§
o
o
o
o
5
g
o
o
O
2
§
5
c
S
2
£
o
Sfc,
o
g
c>
^— — 4>-»4»~i4— i_4.>._.
n <
C4 — <
e e e
• •
e e «
o
. o
in
o
o
• ^
o
o
CO
o
o
(NJ
r*-
o
, f-^
"~ &;
O i.
<_> 0)
JD
0 - E
Oi" QJ
2^ ^
0 QJ
_C v)
"* CVJ
O
0 E •—•
cri I— —
o
O
o
CO
o
o
o
f —
o
o
§
o
o
LO
r C~^
>
•
»
. — .
x;
CL1
•o
u
c
o
' —
CO
1
^f
OJ
s-
3
cn
u.
83
-------
•
to
o
H~l
5
LU
to
CD
O
z
o
I-*
1—
£
to
*
.
r,
•H
O
O
0
I
o
e
o
0
o
o
o
o
e
o
o
2
2
°p
o
o
o
Q
o
o
o
o
o
to
0 °O
K"" ^?
0 O
s
~ o
_-
O • i—
o in
CO fc. >l
i— 3 •—
0 =
.C "~3
^^ r_
0
J
(O
o
.c
(O
^
0
•1
fO
3
H*
S-
,0
14-
i-
o
4-*
O
OJ
»T3
»
^~*
f
1 1 1
i_i
»— »
'oJ
O)
_J
w-
1
"*J
Ol
i.
3
cn
U-
(iudd) auozo
49
r 81 18
84
-------
0
g
o
o
o
o
o
H
o
§
o
1
•
§
1
o
8
8
o
o
o
0
o
0
o
o
z: O
O O
•— i O
»— c/o O
> 0 0
0:1—1 O
UJ (—
co ~ o
00 0
lit O
z £ o
2 °- e
»- « 5e
^ s §
c^ UJ we
88
* O
I-——4— — 4— — 4—— 4— —4— — 4— —4— — 4— —4——4— — 4—— 4— —*•— — •
>^ 0. a e- o- e c <
^ — e>Ni»<|»'C><
•t ^M ^b ^^ ^^ ^^ ^P (
»e*«9«»a
(uidd) auozo
o
0
. CO
o
o
£;
o
o
^o
r-
O
0
LO
§
r—
*— [^
s g
0 . "~
S 52 ^
•~ = 5
o ^
^L en
CO
ol _
O i_ J3
CXJ ^ —
0
^_
1 —
o
o
o
o
o
g
o
o
> CO
> 0
>
»
9
•c
QJ
(Continu
r)-
i
«cj-
QJ
3
Cn
U.
81 18
85
-------
o
o
•
*
>
•
•
•
GO
O
K- 00
0
Q£ »— <
UJ (—
CO O
03 I-H
O Q
UJ
Z 01
O 0.
I-H
i— S
GO UJ
* 0
+ M M + M .
O
6
o
o
0
o
o
0
o
o
§
o
o
o
o
o
0
o
o
o
o
o
o
o
o
0
o
o
o
o
0
o
o
g
o
o
Q
o
o
o
* D
o
o
o
g
g
o
"s"*1"*""**** *" »« ** 8 *" "" « r> i
2 - S § S 8 \
6 b 6 6 6 6 <
> CO
r—
o
0
f^
r—
o
0
r—
o
0
LT>
r—
o
o
• «3-
o
o
ro
o
o
^
o
o
^
o
§
1
0
o
en
o
0
o
CO
) O
~>
1—
0
o
„
(/)
3
o
c5
1
•"~
1 —
4-J
to
3
en
ro
1
O
*J-
'd-
O)
3
O>
ONOZU OLO.C
si 18
86
-------
•
1
1- "
Q
PM
•
K
O '
0
0
o
o
* o
o
o
O '
o
§
^
o
o
o
o
§0
i
o
°0
* c
0
o
t/) O
0 §
1— i^> C
•a: z H
> 0 O
OL i—
LU 1— O
0
>
»
»
CTl
S-
a>
JO
QJ
a.
QJ
CvJ
•D
.
* *
-o
OJ
-o
3
0
c
o
1
QJ
4-
3
Cn
(Uldd) 8UOZQ
87
-------
•
• *
•
00
o
h-
(X.
LU
oo
OQ
O
2
O
»— t
ft
OO
*
1
r "" "*
M
3
O
0
0
0
o
8
0
o
o 0
H
o
o
0
H
o
o
o
4
o
0
o
o
° 0
o
o
4
0
o
£ o
2 °0
1 —
0 0
i— i O
0 0
LU 0
o: o
Q_
i *°
LU O
0
0 0
o
« + ««^«*««^M»-«»««*«-»-- + «- + ««+t.«4«§
" "" n ** "* "" "* n ** "* ""*"«""** ** "* rx "" « ** »«""*" ** i
W 0 0 < » N i
i+ «-" O O O O <
6 6 6 ci 6 6 <
CO
o
g
o
o
10
0
o
Ln
•~
o
o
•«*•
r— —
o
o
^2
o
•§
o
o
• 1—
o
o
• o
o
o
§
o
) °
i
^~.
Q
O
«\
CO
J-
3
0
-C
"~
CO
(O
o
CD
C
O
CO
I
C
o
o
O)
en
c
-------
+ MM + MM + MM + MIH + »HH<+»-ll-l + l-• rt tt
I
o
o
o
o
* *•* +
•4
O
N
6 6
o M o z ui o.o.r
r>
O
o
o
* o o
o o
« + ..««• CO
8°
o
O O*t
O i—
o « >,
O 1/1 i—
+ OO S- 3
r— 3 ""O
O
-— CO
O •!-
•»• O (-
Ovi
0
o
O
o
Q)
E
C
O
LT>
O>
3
CD
81 18
89
-------
1
1
•
•
•
'
00
o
1— I
1—
oc
LU
00
CO
o
0
1— 1
«t
oo
*
4. M M +
p«
N
d
o
8
o
•
o
§
o
o
Q
O
O
O
8
°8*
o
o
0
o
§
0
o
8
a
o
o
a
o
o
0
a
D
o
0
z: °
0 °^
i— i O
1- 0
o o
t—t
S o
£ °o
«:
i * 8
LU O
§
o 8
§
~~§~~ + ~~t,~~*~~&~~*~~j.~~ + ~~t~~ + ~~l~~*~~i
>« M M O O O (
6 6 6 6 6 6 (
o
o
• CO
o
o
1^
§
to
o
g
0
1
o
o
CO
o
o
CM
r- ™
o
0
,_
o
8
o
0
en
o
c
o
i- 00
5 O
3
_
Q
o
«t
CO
5-
3
0
•r™
1—
p;
2
-M
to
3
cr.
-------
•
•
•
•
.
1/7
Z
0
1—4
t—
0
1—4
Q
LU
a:
o.
z
0
1— 1
1—
^£
i—
' ^}
*
h 1-1 «- +
P4
r>
•4
D
••1+MW+ + + + + + + +
0
O
o
o
0
o
8
0
o
'
o
o
o
o
o
•
o
o
1
o
o
8
o
g
o
o
o
o
o
o
Q
"1*
o
o
o
°°
3
r~-
f^.
i —
h- $_
O CD
O fi
E
S- Q.
O
-------
SECTION 5
EKMA ISOPLETH CALCULATIONS
In this section, we discuss the use of the EKMA models to estimate
the results of control scenarios designed to produce ozone reductions
sufficient to meet federal or state standards. Predictions obtained using
the EKMA isopleth diagrams are compared with the predictions obtained from
the SAI Airshed Model. Unfortunately, at the time of this writing, only
one modeling day (29 July 1977) in the Airshed modeling effort had been
investigated using a 1987 emissions inventory. Thus, we can compare only
results of the SAI Airshed and EKMA models for the 29 July 1977 modeling
day. We also present Level III EKMA predictions for 21 July 1977 and 3
August 1977 so that they can be compared with Airshed Model results at a
future date. No effort is made to compare EKMA predictions for 2
September 1977 since the maximum observed ozone on that day was less than
the federal standard of 0.12 ppm.
Emission inventory summaries are presented in table 5-1. These
inventories were developed by Engineering Science (1979) for the modeling
year 1977 and for two future-year scenarios (1982 and 1987). Current SAI
Airshed modeling efforts have concentrated on the 1987 predictions. Table
5-1 indicates that Reactive VOC is reduced 31.5 percent and NOX is reduced
3.7 percent in the period from 1977 to 1987. Additional control measures,
which should lead to greater ozone reductions, are presented in table
5-2. These additional measures would produce a 47.5 percent VOC reduction
in the period from 1977 to 1987. SAI Airshed modeling efforts to date
have performed simulations using options I, II, and V (table 5-2) with 29
July 1977 meteorological conditions.
In all of the original 1987 Airshed simulations, the initial,
boundary, and aloft conditions were kept at the same values as those used
in the 1977 validation runs. The reason for this is that the 1977 initial
conditions were considered to be at background concentrations (0.20 ppmC
for NMHC). Any concentrations greater than the background values would
have been changed to reflect the emissions changes. Since the Level II
EKMA models use the same inputs as the SAI Airshed and SAI trajectory
models, the Level II EKMA models were also run with constant initial
conditions. Table 5-3 shows ozone predictions obtained from the SAI
si 11
92
-------
TABLE 5-1. EMISSION INVENTORY SUMMARY FOR THE TULSA STUDY AREA
Year
1977
1982
Source
Point
Areaf
Line
Total
Point
Area^
Line
Total
Emissions
tons/day
Reactive VOC
(ppmC)
NO/
(ppm)
104.4
29.3
65.4
199.1
62.7
39.1
44.3
146.1
126.0
24.9
58.3
209.2
1987
Point
Area^
Line
Total
66.3
41.9
28.2
136.4
127.3
33.7
40.3
201.3
NOX as N02.
* Not including mobile sources.
81 is
93
-------
TABLE 5-2. CONTROL STRATEGIES TO BE USED IN THE TULSA MODELING STUDY
1977 1987 Percentage
(tons/day) (tons/day) Reduction
Option I
Present strategy as presented
by the state of Oklahoma 199.1 136.4 31.5%
Option II
Option I
Vehicle inspection and
maintenance (VIM) N/A 130.9 34.3
Option III
Option II
Oil/water separators
Fiberglass products N/A 117.7 40.9
Option IV
Option III
Stage I vapor recovery N/A 112.1 43.7
Option V
Option IV
Rideshire, park and ride
Fleet conversion N/A 110.3 44.6
Option VI
Option V
Stage II vapor recovery
Transit improvements
Bicycle paths
Traffic flow improvements N/A 104.4 47.5
N/A = Not applicable.
81 13
94
-------
TABLE 5-3. OZONE PREDICTIONS FROM EKMA MODEL SINGLE-DAY DIAGRAMS
(ppm)
Date
SAI Level III
Airshed Percentage EKMA
Model Change in Model
Results CL Results
Percentage
Change in
Level III
EKMA with
CBM Model
Results
Percentage
Change in
0,
29 July 1977 0.17
29 July 1987—option I 0.14
29 July 1987—option II 0.138
29 July 1987—option III 0.131
(observed 0, = 0.19
21 July 1977 0.14
21 July 1987—option If
21 July 1987—option IIt
21 July 1987—option Vf
(observed 0, = 0.14 ppm)
3 August 1977 0.12
3 August 1987—option I*
3 August 1987—option II
3 August 1987—option V1^
(observed 0, = 0.15 ppm)
17.651
18.8
22.9
0.145
0.14
0.125
0.11
0.105
0.09
0.132
0.13
0.12
14.7S
17.6
26.4
21.4
25.0
35.7
12.0
13.3
20.0
0.12
0.11
0.095
0.08
0.075
0.05
0.12
0.115
0.10
29.455
35.3
44.1
42.9
46.4
64.2
20.0
23.3
33.3
Percentage change in ozone =
(Observed 0 ) - (Predicted 0 )
3 3_
(Observed 0 )
x 100.
Airshed simulations for these days had not been performed at the time of the writing of this document.
81 13
95
-------
Airshed Model using the control options I, II, and V (table 5-2). For the
Level III EKMA simulations, we generated an isopleth for the base year
using initial conditions described in section 3. Although the initial
precursor conditions remained the same for 1977 and the future year (1987)
in the Airshed Model simulations, the ozone value aloft was lowered from
0.08 ppm to 0.06 ppm in the future-year Level III EKMA simulations. A
future-year isopleth with a lowered value of ozone aloft was generated by
following the recommended procedures outlined in the Level III guidelines
(EPA, 1980). The ozone value aloft was lowered by 0.01 ppm in the future
year diagrams for all of the modeling days. Figures 5-l(a), (b), and (c)
show the base case ozone isopleth diagrams for 21 July 1977, 29 July 1977,
and 3 August 1977, respectively. Figures 5-l(d), (e), and (f) show the
future-year diagrams for the same three days. A second set of base case
and future year isopleths was generated for the three modeling days using
the Carbon-Bond chemistry. These diagrams are shown in figure 5-2.
We compared current Airshed Model predictions with Level III EKMA
predictions using emission reduction options I, II, and V on 29 July, 21
July, and 3 August. For these comparisons we used the base case diagrams
[figures 5-1 (a) through (c) and figures 5-2 (a) through (c)] developed
for the Level III EKMA and the Level III EKMA using the CBM. By using the
base case diagram we can directly compare EKMA predictions with the
airshed predictions since the ozone concentrations aloft were kept
constant in the base case EKMA diagrams. Results of these comparisons are
shown in table 5-3. From table 5-3 we see that the percentage change in
ozone between the Level III EKMA using propylene/butane chemistry is
similar to that in the airshed results, whereas the percentage change in
ozone from the Level III EKMA using the CBM shows a greater sensitivity to
the same emission reductions.
We also compared currrent Airshed Model predictions with Level III
EKMA predictions using emission reduction options I, II, and V (table 5-2)
in the future-year diagrams of 29 July, 21 July, and 3 August [figure 5-1
(d) (e), and (f) for the EKMA with propylene/butane chemistry and figure
5-2 (d),(e), and (f) for the EKMA using Carbon-Bond chemistry]. Results
of this exercise are shown in table 5-4, which indicates that with similar
reductions in hydrocarbon and NOX the Level III EKMA models (both
propylene/butane and CBM chemistry) produce ozone predictions lower than
those of the SAI Airshed Model. Results obtained with the Level III EKMA
model using the Carbon-Bond chemistry demonstrate the largest change in
ozone level from these scenarios.
Following the Level III EKMA guidelines (October 1980, draft), we
estimated the hydrocarbon reductions needed to comply with the federal
ozone standard of 0.12 ppm. The results of this exercise are presented in
table 5-5. As seen from this table, the standard Level III EKMA required
«+9r 81 11
96
-------
6CTO LO'O
90-Q *D'D
tO'D
oro
£0'0
90*0 *0'0
Kdd'XBN
CO'O
tD'O
oo -cf
OJ
in
ra
o
OJ
to
CTi
>,
3
fO
o
.c
ro
i/i
=5
S-
o
i.
o
OJ
>
OJ
I
U1
o>
J-
en
81 3
97
-------
o>
fO
• o
O>
ro
r-
r~
CTl
•-3
o-
CM
-o
QJ
C
O
ZIJ0 60' 0
HJJ'XBM
90"0
£0*0
00 '(P
O
s_
C7>
"+9r Bl 3
98
-------
Zl'0 0l"0 S0"0 90"0
-------
6CTO
swa
so'o .sfco'o ecro
J L_ --L-
501'0 6CTO SLQ''Q BQ'D SW'O EO'O SIO'O
Wdd'XEN
OO'Cf
s_
m
0!
-------
ero
O)
OJ
s_
cc
ero si'o zro BCTO 90*0
EO'D
OO'I
c
o
i
LO
3
cn
13
101
-------
zro oro so'o 90*0
zro
01 -
80*0 90'0
Wdd'XBN
-c
Ol
-o
Z3
c
c
I
LD
OJ
i.
3
CD
i+9r 81 3
102
-------
90-0
£0-0 Z0'0 10'0
Wdd'XON
«t9r 81 3
103
-------
Zl'0
OJ
10
IB
U
O>
CTi
>,
3
"-3
CTl
C\J
Ol
3
C
O
o
CM
I
LO
O)
(.
01
Zl'0 0t"0 80-0 90'0
Wdd'XON
00
81 3
104
-------
90-0 £0*0
QJ
i/)
TO
O
QJ
in
«C
m
•c
O-1
C
O
i
in
A0"0 90'0 S0'0
'r B! 3
105
-------
90-0 S0'0
to
o>
>>
OJ
i-
3
+->
3
<4-
r~^
oo
CTi
>>
3
'O
C\J
3
C
CNJ
in
O)
Wdd'XON
U9r 81 3
106
-------
zr0 0r0
J-
«
OJ
OJ
t-
cc
3
•-3
— CTi
-o
OJ
C
o
CNJ
I
in
CJ
u
zr0 0r0
80"0 90'0
Wdd'XON
0*0 no'D"
fl 3
107
-------
90 "0 S0'0
Z.0'0 90'0 S0'0
0 £0'0
Wdd'XON
re
O)
fc.
f-.
00
en
in
3
T3
OJ
•o
3
CVJ
I
LO
QJ
S_
t9r 81 3
108
-------
TABLE 5-4. OZONE PREDICTIONS FROM VARIOUS MODELS FOR THE TULSA MODELING DAYS
(lowered ozone aloft in the EKMA models)
(ppm)
Date
29 July 1977
29 July 1987— option I
29 July 1987— option II
29 July 1987— option V
(observed 0, = 0.17 ppm)
SAI
Airshed
Model
Results
0.17
0.14
0.138
0.131
#
Percentage
Change in
Ozone
0.0%
17.6
18.8
22.9
Level III
EKMA
Model
Results
__
0.138
0.135
0.12
*
Percentage
Change in
Ozone
_
18.8%
20.5
29.4
Level III
EKMA with
CBM Model
Results
..
0.11
0.108
0.09
*
Percentage
Change in
Ozone
_
35.3°;
36.5
47.1
21 July 1977
21 July 1987—option If
21 July 1987—option IIf
21 July 1987—option Vf
(observed O = 0.14 ppm)
0.14
0.0
0.102
0.10
0.08
27.1
28.6
42.9
0.08
0.07
0.05
42.9
50.0
64.3
3 August 1977 0.12
3 August 1987—option If — — 0.128 8.0 0.112 25.3
3 August 1987—option IIt — — 0.125 10.0 0.108 28.0
3 August 1987—option Vf — — 0.11 26.7 0.085 43.3
(observed 0, = 0.15 ppm)
(Observed 0 ) - (Predicted 0 )
* 33
Percentage change in ozone = x 100.
(Observed 0 )
3
Airshed simulations for these days had not been performed at the time of the writing of this document.
t9r 81 13
109
-------
TABLE 5-5. REQUIRED PERCENTAGE REDUCTIONS IN HYDROCARBON PREDICTED BY
THE EKMA TO REACH 0.12 (ppm) OZONE LEVELS (3.7 percent NOX
reduction assumed)
Date—1977
21 July
29 July
3 August
2 September
Design
03
0.14
0.17
0.15
Level III EKMA
17.0
44.6
37.6
Level III EKMA
with CBM
10.7
28.1
28.2
Design ozone for this day was less than 0.12 ppm.
81 13
110
-------
percentage reductions 1.3 to 1.5 times those of the Level III EKMA using
the Carbon-Bond chemistry to reach an ozone level of 0.12 ppm. Some
reasons for this sensitivity are discussed in the next chapter; however,
the main reason for the sensitivity appears to result from the 10-hour
simulation time combined with the polarity of reactivity found in the 25-
percent high reactivity propylene--the 75-percent low reactivity butane
combination used in standard Level III EKMA. At similar overall ozone
generation capacity, the two chemistries tend to produce different diurnal
03 profiles as discussed in section 4. The high reactivity of propylene
tends to cause the standard chemistry to make ozone faster than occurs
with the CBM until the propylene runs out; then, ozone forms more slowly
than the ozone generated by the CBM. The CBM has more levels of reac-
tivity, which tend to even out the rate of ozone generations. However,
with this more even rate of ozone generation, the CBM often produced
simulations that did not reach an ozone peak within the 10 hours simulated
by the OZIPM model. This lack of ozone peaking is exaggerated by condi-
tions of low HC/NOX ratio. Hence, the control scenarios that emphasize HC
control tend to lower the HC/NOX ratio and, in turn, the emphasis on ozone
peaking results in high sensitivity. Thus, the standard mechanism peaks
before the 10-hour cutoff point more often than it does in the CBM.
Therefore, HC control using the CBM produces many ozone values taken
during the last (of 10) hour by the OZIPM algorithm, even though the
simulated ozone may still be rising rapidly in the model. Under similar
conditions, the propylene/butane chemistry using 25-percent propylene
could have produced an early peak that would perhaps be higher than the
truncated peak produced by the CBM. Hence, longer simulation times may be
required if the CBM were to be used in the EKMA.
81 11
111
-------
SECTION 6
SENSITIVITY STUDIES
This section discusses four areas of sensitivity of the EKMA models
to various input parameters and model characteristics investigated in this
study:
> Emission density information.
> Ozone predictions obtained with the propylene/butane
mechanism.
> Surface and aloft precursor conditions.
> Sensitivity of Level II EKMA models to the choice of the
trajectory path.
SENSITIVITY OF THE EKMA MODEL TO EMISSIONS DENSITY INFORMATION
Emission densities with a 2-kilometer spatial resolution were
originally developed by Engineering Science (1979); in the Airshed
modeling effort, the grid cell size was increased to 4 kilometers. A
recent study of the San Francisco region (Whitten, Hogo, and Johnson,
1981), concluded that the size of the grid cells can become important in
the EKMA predictions if sharp differences are seen in adjacent grid
cells. Ozone predictions can change significantly depending upon the
emission fractions used in the calculations. In an effort to test the
sensitivity of the EKMA model to grid resolution, we defined two trajec-
tories parallel to the original trajectory, which had been defined to
reach the maximum observed ozone for the 21 July 1977 modeling day.
Figure 6-1 shows the two new trajectories (one east and one west of the
original trajectory).
The SAI trajectory model predicted similar results for ozone concen-
trations from the eastern trajectory (0.145 ppm ozone) and the original
trajectory (0.154 ppm ozone), whereas predictions from the western
81 15
112
-------
20
OT
UJ
10
N0RTH
:x:x^x:x*x:x:*x::?l:xx::l.x:x:*':x:x!':+f^*t.~:<:-:f. :-x-l. ::.x:*.-:::-::!.-:x::*x-:::
BCHE
VERfl
SKJR
06001
0500 .'
DS
Vx'f
S0UTH
20
to
-------
trajectory were lower (0.127 ppm). The emissions patterns for the new
trajectories are shown in table 6-1 and the initial conditions for all
three trajectories are shown in table 6-2. By averaging emissions and
initial conditions from the original trajectory with the one next to it we
also created two additional trajectories twice the original grid size.
Performing EKMA calculations with these last two trajectories, we found
that both gave similar results. The western trajectory resulted in an
ozone value of 0.151 ppm, while the eastern trajectory resulted in an
ozone value of 0.158 ppm.
SENSITIVITY OF THE PROPYLENE/BUTANE CHEMISTRY DUE TO VARIATION
IN RATE CONSTANTS
Since its formulation, many of the rate constants used in the
original propylene/butane mechanism (Dodge, 1978) have been updated with
more recent measurements and evaluations (Atkinson and Lloyd, 1980). To
test the sensitivity of the propylene/butane mechanism to rate constant
changes, we investigated the effects on ozone predictions of perturbations
in the following rate constants:
> N02 photolysis constant
> Aldehyde photolysis constants
> Radical and NOX sink reactions
> Reactions of peroxy radicals (R02 and H0£) with NO.
The N02 photolysis constant is calculated using a theoretical
absorption curve. In the Tulsa monitoring program, total solar radiation
was measured in units of langleys/min that can be related to the N02
photolysis constant. From these measurements and from the theoretical N02
photolysis constant, we found that the theoretical constant used in the
EKMA was approximately 10 percent lower than the measured values. Table
6-3 shows results obtained by raising the EKMA N02 photolysis constant 10
percent using the 21 July meteorology. The difference in the ozone
predictions is approximately a 3 percent increase from the base case ozone
prediction of 0.096 ppm.
The photolysis of aldehydes can be important in photochemical smog
mechanisms. The propylene/butane mechanism has a higher rate of aldehyde
photolysis than the CBM. To estimate the differences in ozone predictions
caused by aldehyde photolysis, we lowered the aldehyde photolysis rates in
the propylene/butane mechanism so that the two mechanisms would correspond
H9I" 81 15
114
-------
TABLE 6-1. EMISSIONS RATES FOR TRAJECTORIES PARALLEL
TO THE ONE TO VERA FOR 21 JULY 1977
20
Trajectory
Trajectory
Time
(CDT)
0500-0600
0600-0700
0700-0800
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
Time
(CDT)
0500-0600
0600-0700
0700-0800
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
West of Vera
HC (ppmC/hr)
0.0006
0.0008
0.0222
0.0130
0.0285
0.5983
0.3207
0.0218
Trajectory
West of Vera
NOx (ppm/hr)
0.0
0.0
0.0003
0.0016
0.0067
0.1054
0.0633
0.0078
East of Vera
HC (ppmC/hr)
0.0014
0.004
0.017 .
0.0391
0.0789
0.9088
0.2386
0.027
Trajectory
East of Vera
NOX (ppm/hr)
0.0007
0.0015
0.0052
0.0046
0.0128
0.0924
0.0621
0.009
115
-------
TABLE 6-2. INITIAL CONDITIONS USED IN THE EKMA CALCULATIONS
FOR 21 JULY 1977
Surface
Trajectory
NO
N02
N0x
N02/NOX
OLE
PAR
ARO
ETH
CARB
HC
°3
CO
West of Vera
0.00001 ppm
0.0031 ppm
0.0031 ppm
0.9968 ppm
0.0029 ppmC
0.0643 ppmC
0.0720 ppmC
0.0115 ppmC
0.0120 ppmC
0.1627 ppmC
0.0354 ppm
0.1000 ppm
Trajectory West of Vera
NO
N02
NO
OLE
PAR
ARO
ETH
CARB
HC
°3
CO
0.0001
0.0011
0.0012
0.0015
0.0611
0.0559
0.0098
0.0136
0.1419
0.0613
0.8158
Trajectory to Vera
0.00001
0.0031
0.0031
0.9968
0.0029
0.0643
0.072
0.0115
0.012
0.1627
0.0356
0.1000
Aloft
Trajectory to Vera
0.0003
0.0016
0.0019
0.0013
0.0587
0.0488
0.0090
0.0149
0.1329
0.067
1.4778
Trajectory East of Vera
0.00001
0.0032
0.0032
0.9969
0.0029
0.0643
0.0720
0.0115
0.0120
0.1627
0.0349
0.1000
Trajectory East of Vera
0.0004
0.0022
0.0027
0.0012
0.0582
0.0482
0.0089
0.0158
0.1323
0.07
1.9453
20
116
-------
TABLE 6-3. LEVEL II EKMA OZONE PREDICTIONS USING THE PROPYLENE/BUTANE
CHEMISTRY FOR 21 JULY 1977 UNDER VARIOUS CONDITIONS (OBSERVED
OZONE = 0.14 PPM)
Predicted Ozone
Sensitivity Run (ppm)
Base case 0.096
Raised N02 photolysis = 10 percent 0.099
Raised N02 photolysis and lowered aldehyde photolysis = 0.093
60 percent
Raised R02 + NO reactions to a ratio of 0.86 with 0.131
OH + N02 reaction
81 16
117
-------
at 1100 CDT. To make the chemistry of the propylene/butane and Carbon-
Bond mechanisms as similar as possible, we kept the N02 photolysis rate
increased by 10 percent for both mechanisms. From the results of this
exercise shown in table 6-3, we see that the predicted ozone is
approximately 3 percent lower than the base case value when the aldehyde
photolysis rates are lowered by 60 percent.
A third area of difference between the two photochemical smog
mechanisms concerns the peroxy radical reactions that convert NO to N02
(i.e., R02 + NO reactions). We performed a sensitivity test by raising
the R02 + NO reactions to 6900 ppm^min'1 from the 1200 to 1800 ppm'^in"1
value normally used. The significant factor in these kinetics is found in
the ratio of ozone production, which takes place through the conversion of
NO to N02 by the peroxy radicals, to the loss of both radicals and NOX via
the nitrate production caused by the hyroxyl radical reaction with N02. In
the CBM, this ratio is 0.86; therefore, using a value of 6900 ppm^min"1
in the propylene/butane mechanism for all reactions of the type R02 + NO
would bring the ratio up from about 0.2 to 0.86.
Results of this exercise are also shown in table 6-3. As we can see,
the predicted ozone values increased by approximately 36 percent over the
base case value. The predicted ozone value of 0.13 ppm is much closer to
the observed value of 0.14 ppm for both the 21 July day and the CBM/OZIPM
predicted value of 0.15 ppm. Thus, we see that updating the R02 + NO
reaction rates can provide one explanation for the low simulated ozone
concentrations predicted by the propylene/butane mechanism. Although
changing this one type of reaction provided an explanation for much of the
difference routinely seen between the results obtained from the CBM and
the propylene/butane mechanism, this change should not be construed as the
means to easily update the propylene/butane mechanism for use in EKMA
applications. Before an updated version of the mechanism can be used, a
revalidation against smog chamber data must be performed.
SENSITIVITY OF MODELS TO HYDROCARBON AND NOX BACKGROUND CONCENTRATIONS
In the original Airshed Model study of the Tulsa area, it was
believed that the base year hydrocarbon concentration of 0.2 ppmC repre-
sented typical and insignificant background values. Thus, it was felt
that the future year simulations would not require reduced initial and
aloft hydrocarbon concentrations. However, it was later determined that
the hydrocarbon concentrations found in the Tulsa region are as low as
0.04 ppmC. CBM/OZIPM isopleths generated to assist the original Airshed
modeling effort predicted that a hydrocarbon reduction of 94 percent might
be required to achieve the federal ozone standard of 0.12 ppm from the
highest simulated ozone level [figure 6-2(a)] if the 0.2 ppmC level
81 15 118
-------
0S'I OS'I EC'I DE'B 09'0
c
re
o
o
c/1
c
o
fO
U
O
T3
C
3
O
o
re
CQ
re
x
ro
ill
C
O
fsi
O
ro
c
o
•i—
cr,
OJ
O)
u
Ol
co
o
CNJ
o
o
4->
ro
00
ro
o
"ro
BfZ P.8*t OS'I 03 "I 06'0 09'0
Hdd'XON
CNJ
I
CD
3
CT>
20
119
-------
03'I
(J
re
CQ
-o
-------
existed. The isopleth diagram was generated using the emissions and
meteorology of the specific trajectory to the maximum predicted ozone of
0.21 ppm from the base case Airshed Model simulation. A monitoring site
was not located at the site of this maximum occurrence. The OZIPM
computer code was set up to keep the initial conditions for each calcula-
tion constant and to vary only the emissions.
A second isopleth was generated using these same inputs, but initial
and aloft conditions were allowed to vary proportionally with the emis-
sions changes [figure 6-2(b)]. In the guidelines for the use of the
standard city-specific EKMA, the initial and aloft precursor concentra-
tions are referred to as "transported" pollutants in the surface and aloft
layers. Varying the concentration proportionally with the emissions
reductions on the isopleth diagram certainly represents an extreme case,
since some natural and transported pollutants would not vary between base
and future years. However, the use of fixed concentrations for these
background pollutants represents the other extreme. The two approaches
tend to bracket the problem. For many urban areas, there is undoubtedly
some carry-over—or return of previously emitted, but unreacted--
precursors. Surely, the level of such pollutants would be reduced
according to the level of local controls. Additionally, numerous controls
(e.g., those for automobile-related emissions) are common to all urban
areas because of federal regulations, and some of the pollutants trans-
ported from upwind sources would vary in proportion to local controls.
However, all of these background pollutants can vary only between present
levels and some natural, and currently uncertain, background level.
For the diagram using modified background concentrations, an esti-
mated hydrocarbon reduction of only 53.8 percent is needed for attainment
of the ozone standard. The primary reason for the considerable difference
between the two isopleths in this case is that the initial conditions
dominate emissions in the total precursor concentration for the simula-
tions. This situation is unlike that of the Los Angeles simulations in
which the emissions tend to dominate the modeling region (Whitten and
Hogo, 1981).
To further exemplify the importance of the initial and aloft condi-
tions in the Tulsa simulations, we used the SAI trajectory model to
simulate several sensitivity runs in which the hydrocarbon and NOX
concentrations in the surface and aloft layers were varied. For a base
case, we used the trajectory to the maximum ozone predicted by the Airshed
Model. In each of the runs, we lowered the hydrocarbon and/or NOX by 50
percent. The results, presented in table 6-4, show that the hydrocarbon
concentrations are the key pollutant. It should also be noted that the
surface concentrations were changed only at the 0500 starting time of the
model; by the normal EKMA starting time of 0800, the 50 percent change in
the mixing layer concentration still amounted to a 42 percent reduction.
81 is
121
-------
TABLE 6-4. SAI TRAJECTORY MODEL RESULTS OBTAINED BY VARYING
HYDROCARBON AND NOX CONCENTRATIONS IN THE SURFACE
AND ALOFT LAYERS
Predicted Ozone
Concentration Level (ppm)
Base case 0.21
Hydrocarbon reduced by one-half 0.174
NOX reduced by one-half 0.21
Both hydrocarbon and NOY reduced by one-half 0.174
t*9r 81 16
122
-------
We also investigated the sensitivity of ozone prediction to the
addition of precursor transport for the CBM/OZIPM and Level-111-type EKMA
models. For the CBM/OZIPM model, we cut the aloft hydrocarbons in half
while keeping the surface hydrocarbons at the original values (see
section 3). This change reflects the observed measurements more
closely. The average aloft hydrocarbon concentrations were approximately
half those found in the surface bag samples (see section 3). Results
obtained using the lower hydrocarbon concentrations aloft for the 29 July
1977 day are shown in table 6-5. We see that the ozone maximum was
reduced by 13 percent as the result of a 50 percent HC reduction aloft.
For the Level III EKMA models, no precursor transport (except ozone) was
considered in the previous sections. Current Level III guidelines (EPA,
1980) do not recommend using surface and aloft precursor transport for
normal applications unless the user decides otherwise. To test the
sensitivity of the Level III EKMA models to precursor transport, we
developed averaged surface and aloft hydrocarbon transport as inputs for
the 29 July 1977 day.
On the basis of procedures outlined in appendix B of the Level III
guidelines (EPA, 1980), we determined the mean contribution factor of the
Liberty Mounds station on the urban core by comparing 0600 to 0900
observed NMOC between the Liberty Mounds station with the two monitoring
stations located in the urban core. Results of this comparison are
presented in table 6-6, which shows only complete NMOC samples. When NMOC
samples are reported for both urban stations, an average is taken between
the two stations. From these comparisons, we found the median contribu-
tion factor to be 0.36. This factor times the 0600 to 0900 average NMOC
used for each modeling day gives the NMOC precursor in the surface
layer. NOX contribution factors were also determined for the same days
for which there were calculated NMOC contribution factors. The results of
this exercise are shown in table 6-6(a). The median contribution factor
is 0.044. From the Level III guidelines, we readjusted the post-0800
emissions by 1.0 minus the median contribution factors. Aircraft samples
were made between 26 August and 17 September 1977 for hydrocarbon, NOX,
and ozone. Table 6-7 shows aircraft measurements of NMOC at different
altitudes taken in the area around Liberty Mounds (figure 6-3) during the
morning hours. From the aircraft data, we averaged the measurements taken
at the 1200-ft and 2500-ft level to determine the average NMOC and NOX
aloft. Of the six days of aircraft measurements, all observations were
fairly consistent except for those of 2 September. The NOX measurements
were on the average twice as high as the measurements observed on the
other days. The hydrocarbon measurements were approximately 50 to 75
times higher on 2 September than on any of the other days. The average
NMOC aloft is 0.062 ppmC (not including the 2 September measurements) and
0.002 ppm for NOX aloft (not including the 2 September measurements). If
all the measurements are averaged, the NMOC could be 0.452 ppmC and NOY
t*9>" 81 15
123
-------
TABLE 6-5. EFFECT OF PRECURSOR CONDITIONS ON OZONE PREDICTIONS FOR
29 JULY 1977 FOR THE CBM/OZIPM AND LEVEL III EKMA MODELS
Predicted Ozone
Model (ppm)
CBM/OZIPM (base case) 0.15
CBM/OZIPM (hydrocarbon aloft reduced by one-half) 0.13
Level III EKMA (base case) 0.12
Level III EKMA (added surface and aloft percursors) 0.13
<*9r 81 16
124
-------
TABLE 6-6. CALCULATION OF THE CONTRIBUTION FACTORS OF
TRANSPORTED POLLUTANTS TO URBAN LEVELS—1977*
Date
21 August
22 August
12 September
16 September
25 September
26 September
27 September
28 September
Liberty
Mounds
0.008
0.012
0.002
0.003
0.004
0.004
0.002
0.001
(a) NOX
(ppm)
Post
Office
0.024
0.102
0.082
0.124
0.050
0.090
0.044
0.025
' Average contribution factor = 0.104;
median contribution factor
Date
22 July
23 July
21 August
22 August
12 September
16 September
18 September
25 September
26 September
27 September
28 September
29 September
Liberty
Mounds
0.536
0.327
0.277
1.042
0.106
0.171
0.068
0.121
0.279
0.149
0.083
0.076
= 0.044.
(b) NMOC
(ppmC)§
Post
Office
--
—
0.767
3.684
0.348
0.871
—
0.390
0.653
0.300
0.278
"
Health
Department
0.022
0.097
0.054
0.129
0.028
0.090
0.032
0.027
standard
Health
Department
--
--
--
1.127
--
0.566
—
—
0.516
—
—
"
Contribution
Factor'
0.360
0.124
0.037
0.023
0.143
0.044
0.063
0.037
deviation = 0.112;
Contribution
Factor
—
--
0.36
•*•
0.43
0.30
0.24*
--
0.31
0.48*
0.50
0.30
"
* Only complete hydrocarbon samples are presented.
Average contribution factor = 0.365; standard deviation = 0.095;
median contribution factor = 0.36.
Contribution factors are calculated using the average of the
values obtained for the post office and Health Department sites,
49r 81 16
125
-------
TABLE 6-7. AIRCRAFT MEASUREMENTS OF NMQC
Date
26 August 1977
2 September 1977
3 September 1977
11 September 1977
16 September 1977
17 September 1977
Time
ICDJI
0722
0725
0731
0737
0729
0736
0743
0752
0726
0731
0738
0746
0724
0727
0734
0742
0726
0729
0736
0743
0739
0741
0745
0751
Altitude
(ft)
1200
2500
5000
7550
1100
2600
5000
7600
1250
2500
5000
7700
1500
2500
5000
7500
1500
2700
5000
7500
1500
2500
5000
7500
NMOC
(PPmC)
0.0416
0.0407
0.0995
0.0259
3.114
1.699
2.134
1.18
0.097
0.039
0.0164
0.0636
0.0445
0.0475
0.0201
0.0215
0.0879
0.0725
0.0213
0.0345
0.0885
0.0563
0.095
0.236
81 18 126
-------
would be 0.003 ppm. We decided to use an NMOC value of 0.062 ppmC and a
NOX value of 0.002 ppm since the measurements of 2 September are question-
able.
From the results of the exercise shown in table 6-5, we see that the
predicted ozone maximum value was raised by 8 percent when the HC precur-
sors were added.
The high sensitivity to the background HC levels assumed in the
original SAI Airshed Model validation effort has been demonstrated in this
study. This sensitivity applies to both the base case and control
predictions. Largely as a result of this demonstration using the EKMA
models and the SAI trajectory model, the study using the SAI Airshed Model
is currently being reevaluated using a combination of more reactive
emissions, fewer background hydrocarbons, and observations that have been
adjusted downwind (to account for incorrect KI calibration) and shifted in
time (to correct misinterpretations of daylight, rather than standard,
time). These pending changes in the Airshed Model results do not neces-
sarily invalidate the comparisons made to date between the Airshed Model
and the various trajectory models studied here. The verifying factor in
the comparisons conducted to date is the use of common data in the various
models and, thus, the comparisons show differences in results accruing
from differences in model formulation and implementation.
SENSITIVITY OF LEVEL II EKMA MODELS TO TRAJECTORY SELECTION
INVOLVING WIND SHEAR
In this section, we discuss problems that can occur from the use of
trajectory-specific emissions for days with significant wind shear. As
noted in the previous discussion of the ozone predictions for the 29 July
1977 day, the SAI trajectory model results (and consequently, the
CBM/OZIPM EKMA models results) do not agree with those of the SAI Airshed
Model. This lack of agreement is apparently related to wind shear during
the early morning hours. Figure 6-3 shows the two trajectory paths used
for the SAI trajectory model calculations. One trajectory path is based
on the surface-level wind fields, whereas, the other is based on an
average of the first- (surface) and second-level winds. We see from
figure 6-3 that the trajectories are significantly different during the
hours from 0500 to about 1000 COT; yet after 1000 CDT, the trajectory
paths are quite similar.
The SAI trajectory model predicted ozone concentrations of 0.14 ppm
from both of the trajectory paths, and the Airshed Model predicted an
ozone value of 0.17 ppm. We note that the two trajectory paths represent
two extremes in terms of the possible trajectory taken by air flowing in
»t9l" 81 15
127
-------
13363C
AFTERNOCT
PATTERN
Oxaulfte
Figure 6-3. Typical transport flight pattern.
81 20
128
-------
the Airshed Model. The appropriate average path, one comparable to the
Airshed Model simulations, would be somewhere between the two paths shown
in figure 6-3 for the early-morning hours. The critical factor in path
selection is the gridded emission inventory. If the grid squares with
high emissions are bypassed, the model will not have sufficient precursor
concentrations to generate the proper ozone level. In the present case,
all emissions between these two path extremes contribute to the air parcel
that finally forms at 1000 CDT and moves to the monitoring site. This
area is illustrated by the shaded portion of figure 6-4.
Figure 6-5 shows the spatial distribution of emissions of reactive
hydrocarbons during the hour from 0800 to 0900 CDT. From this figure we
see that the two specific trajectories used in the SAI trajectory model
calculations did not pick up the high emissions located in cell (6, 17).
i+9r 81 15
129
-------
N0RTH
20
(O
LU
10
0CHE
VERfl
SKIR
SPRY
1400
1300JPHU400
•1200
MNDS
20
IT)
-------
NORTH
::.:.;.;:l:.:.:.; jl. ;...;! v. ..< '. '.of '. ;.;:! :.|.|.y _"_;...; 4 .;. 1 7| : "
CO
UJ
I -• "vi ! '. .-. I .-.-<'
OCHL
VERfl
SKIfl
SPRY
HNDS
30
CO
cr
10
SOUTH
Figure 6-5.
Total ground-level emissions of hydrocarbons for the
Tulsa region in kg/hr between the hours of 0800 and
0900 (CDT)
20
131
-------
REFERENCES
Atkinson, R., and A. C. Lloyd (1980), "Evaluation of Kinetic and
Mechanistic Data for Modeling of Photochemical Smog," ERT No. P-A040,
Environmental Research and Technology, Inc., Westlake Village,
California.
Dodge, M. C. (1978), "Effect of Selected Parameters on Predictions of a
Photochemical Model," EPA-600/3-77-048, U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina.
EPA (1980), "Guideline for Use of City-Specific EKMA in Preparing Ozone
SIPs," EPA-450/80-027, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina.
Eaton, W. C., and F. E. Dimmock (1979), "Study of the Nature of Ozone,
Oxides of Nitrogen, and Nonmethane Hydrocarbons in Tulsa, Oklahoma:
Vol. II. Data Tabulation," EPA-450/4-79-008b, U.S. Environmental
Protection Agency, Research Triangle Institute, Research Triangle
Park, North Carolina.
Eaton, W. C., et al. (1979a), "Study of the Nature of Ozone, Oxides of
Nitrogen, and Nonmethane Hydrocarbons in Tulsa, Oklahoma," EPA-450/4-
79-008a, U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina.
Eaton, W. C., et al. (1979b), "Study of the Nature of Ozone, Oxides of
Nitrogen, and Nonmethane Hydrocarbons in Tulsa, Oklahoma—Vol. Ill,
Data Analysis and Interpretation," EPA-450/4-79-008c, Research
Triangle Institute, Research Triangle Park, North Carolina.
Engineering Science (1979), "Emission Inventories for Tulsa, Oklahoma for
SAI Model Application," Engineering Science, McLean, Virginia.
Jeffries, H., D. Fox, and R. Kamens (1975), "Outdoor Smog Chamber
Studies: Effect of Hydrocarbon Reduction on Nitrogen Dioxide,"
University of North Carolina, Chapel Hill, North Carolina.
si m 132
-------
Killus, J. P., and G. Z. Whitten (1981), "User's Guide to the Carbon-Bond
Mechanism," SAI No. 75-81-EF81-90, Systems Applications, Inc., San
Rafael, California.
Reid, L. E., S. D. Reynolds, and D. A. Latimer (1980), "Evaluation of
Airshed Model Performance in Tulsa," Technical Memorandum No. 3, SAI
No. 223-ES80-164, Systems Applications, Inc., San Rafael, California.
Reynolds, S. D., et al. (1979), "Photochemical Modeling of Transportation
Control Strategies--Vol. I. Model Development, Performance Evaluation,
and Strategy Assessment," prepared for the Federal Highway
Administration, Office of Research, Washington, D.C.
Sellers (1965), Physical Climatology (University of Chicago Press,
Chicago, Illinois).
Turner, D. B. (1969), "Workbook of Atmospheric Dispersion Estimates," AP-
26, U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina.
Whitten, G. Z., H. Hogo, and R. G. Johnson (1981), "Application of the
Empirical Kinetic Modeling Approach (EKMA) to Urban Areas,"
SAI No. 32R-EF80-139R, Systems Applications, Inc., San Rafael,
California.
Whitten, G. Z., and H. Hogo (1981), "Comparative Applications of the EKMA
in the Los Angeles Area," SAI No. 10R-EF80-73, Systems Applications,
Inc., San Rafael, California.
Whitten, G. Z., and H. Hogo, (1978), "User's Manual for Kinetics Model
and Ozone Isopleth Plotting Package," EPA-600/8-78-014a, Systems
Applications, Inc., San Rafael, California.
Whitten, G. Z., J. P. Kill us, and H. Hogo (1980), "Modeling of Simulated
Photochemical Smog with Kinetic Mechanisms," EPA-600/3-80-028a,
Systems Applications, Inc., San Rafael, California.
81 i* 133
-------
APPENDIX
DESCRIPTION OF INPUTS USED IN SAI AIRSHED MODEL SIMULATIONS
The following description of the inputs used in the SAI Airshed
simulations are taken from Reid, Reynolds, and Oliver (1980). Many of the
inputs derived for the less complicated models as discussed in the main
part of the document are based on the inputs developed for the Airshed
Model.
PREPARATION OF INITIAL AND BOUNDARY CONCENTRATION INPUTS
Three input files—AIRQUALITY, BOUNDARY, and TOPCONC—are required
for the specification of initial and boundary conditions of all pollu-
tants. The AIRQUALITY file specifies the concentrations of all species
for all grid squares at the start time of the run. The data available for
constructing this file are the monitoring data from the eight RTI sites
and the two Tulsa City/County Health Department (TCCHD) stations, the
hydrocarbon bag samples, and the aircraft flights. We used two methodolo-
gies for constructing gridded fields of initial conditions from the data--
one for constructing the gridded ground-level concentrations, and one for
extrapolating the concentrations aloft from the ground-level values.
To construct the initial and boundary concentration files, we
examined the Tulsa air quality data base in order to characterize the
levels and distribution of pollutants in the study area. Using this
information and drawing on our previous experience in modeling other urban
areas, we tried to identify pollutant concentrations and to establish
reasonable methodologies for preparing these inputs. In cases in which
there were little or no data specific to a modeling day, we chose to use
average values rather than to develop a correlation between the parameter
of interest and another parameter that was actually measured. This proce-
dure enabled us to avoid estimating some parameter of interest on a day
during which the conditions may have been quite different from those used
in a correlation analysis.
The available NOX and ozone data collected at the eight RTI monitor-
ing stations were generally adequate to characterize the ground-level
134
-------
initial and boundary values for these pollutants. However, we had
considerably fewer data with which to characterize the ground-level
hydrocarbon initial and boundary conditions. To further complicate
matters, pollutant concentrations aloft were measured on only one of the
four days selected for model simulations. Of course, the range within
which we could vary these hydrocarbon concentrations was limited by the
uncertainty in the available data. The difficulty arises in separating
the range of uncertainty from the range of real fluctuations. Although an
average value for the latter can be considered to best represent the HC
levels, a broad range of true uncertainty allows a more arbitrary choice
of HC levels. The estimated HC levels are currently (as of this writing)
being reevaluated; these original estimates involved HC/NOX ratios (which
varied over a wide range), actual HC measurements (which were extremely
limited), and NOX measurements (which varied over a wide range). As a
result, considerable uncertainty exists in our specification of the
initial and boundary concentration inputs, especially those for hydrocar-
bons.
Preparation of Gridded Ground-Level Initial Concentration Fields
We began the calculations at 0500 CDT as described earlier in section
3 of the main report. The hourly average concentrations observed at the
monitoring sites for the hour ending at the starting time of the run were
employed as the inputs to the interpolation routines that were used to
calculate the gridded concentration fields. We used these measured values
to estimate conditions at 0500 CDT rather than, say, the average of the
0400 to 0500 and 0500 to 0600 hourly average observations to avoid the
possible influence that emissions subsequent to the simulation starting
time would have on the measured values, and thus, the resulting initial
concentration fields. In this study, we used an inverse distance weight-
ing interpolation routine to construct the gridded surface fields. This
relationship for the estimated concentration c^ in grid cell k can be
written as follows:
? Cm'1 ^k"1
ck -^- —3 ,
L, ik
1=1
where cm ^ is the measured concentration at monitor i, n is the total
number of measured values, and r^ is the distance from the monitor to the
center of grid cell k.
49r 81 5 135
-------
In addition to the two TCCHD ozone monitors, eight RTI monitoring
stations were used to estimate NO, NOp, and 63 ground-level concentra-
tions. Since there is only one site (Liberty Mounds) in the southern
third of the region, it was necessary to add "fictitious" monitoring
stations in the southern end of the region to prevent the calculated
concentrations in this area from being unduly influenced by the values
observed at other stations to the north. In past modeling applications we
have sometimes found it necessary to add stations in the corners of the
modeling region and to give these stations background values that reflect
the rural nature of the outlying portions of the grid. Otherwise, the
mathematical routine for calculating gridded ground-level initial concen-
tration values will extrapolate using data that are often collected in the
more source-intensive portions of the grid. We added four "fictitious"
stations in the southern third of the grid—one in each corner, one on the
eastern edge, and one on the western edge, each about eight grid cells
from the southern boundary. The pollutant concentrations at each of these
stations were set to the background levels cited in table A-l, except for
ozone, which was given a value of 0.04 ppm.
The amount of spatial variation in the early morning NOX and 03
monitoring data differed from day to day. For 21 July, the NOX data
ranged from 5 ppb at the rural stations to 70 ppb at the post office
site. The post office observations exhibited the highest early morning
NOX concentrations. The NOX measurements for 29 July were low with a
range from 5 to 25 ppb. The third of August demonstrated the greatest
variation in NOX observations and the initial condition field reflects
this, as can be seen in figure A-l(a). The data for 2 September again
reflected the low values of NOX observed on most days, with a high of 23
ppb. Most of the early morning ozone measurements for all the modeling
days were around 30 ppb, with extreme values of 10 and 50 ppb. For this
ozone to exist before dawn, the NOX must have been predominantly in the
form of N02 because of the rapid reaction of NO with ozone.
On any particular day, continuous hydrocarbon measurements were made
at only two sites. Since these measurements did not provide sufficient
data to specify the spatial variability of the ground-level hydrocarbon
initial conditions, we found it necessary to tap other sources of hydro-
carbon data. Additional available hydrocarbon data are the 0600 to 0900
CDT bag samples routinely taken at Liberty Mounds, health department, and
the post office. Although these samples were taken during the morning
rush hour, the data from Liberty Mounds should represent upwind concentra-
tions, since there is little traffic in the area, and since few emissions
sources exist near the site. The Okmulgee Refinery, which is about 20 to
25 km south of the Liberty Mounds monitoring station, is the only poten-
tially significant nearby emissions source. To assess the impact of the
refinery, we used the formulas given by Turner (1969) for calculating
81 5 136
-------
TABLE A-l.
Species
NO
N02
°3
NMOC
Ethylene
Paraffins
Olefins
Aromatics
Carbonyls
Benzaldehyde
CO
BACKGROUND CONCENTRATIONS USED IN THE
TULSA SIMULATIONS
(ppm)
Background Concentrations
21 July
0.0005
0.002
0.06
0.24
0.0058
0.065
0.0014
0.012
0.012
0.00001
0.1
29 July
0.0005
0.002
0.08*
0.20
0.0049
0.053
0.0012
0.009
0.009
0.00001
0.1
3 Aug.
0.0005
0.002
0.06
0.21
0.0051
0.056
0.0013
0.010
0.010
0.00001
0.1
2 Sept.
0.0005
0.002
0.05f
0.17
0.0041
0.046
0.0010
0.008
0.008
0.00001
0.1
For 29 July, the ozone boundary concentrations on the sides of the
modeling region varied with location. The boundary values on the
north, south, east, and west boundaries were 0.07, 0.08, 0.06, and
0.08 ppm, respectively.
For 2 September, the Og boundary concentration at the top of the
modeling region varied in time: From 400 to 1100 CST, it was set at
0.07 ppm and from 1100 to 2100 CST, it was set at 0.05 ppm.
137
81 10
-------
20
10
.WYNO
NORTH
10
1 I T
OCHL
MNDS
I I I I I I -I 1 1 1 1 L
SOUTH
20
10
Figure A-l.
(a) For NO (pphm)
rt
Initial conditions in the ground-level
grid cells on 3 August 1977 (isopleths
in increments of 2 pphm).
81 1 8
138
-------
NORTH
I I I I I I I I I
i I I i i I I
SOUTH
(b) For RHC (pphm)
Figure A-l.(concluded).
81 18
139
-------
pollutant concentrations downwind of point sources, and assuming D
(neutral) stability, we calculated the expected hydrocarbon concentrations
at Liberty Mounds that could be attributed to the refinery. For the case
in which the centerline of the plume was directly over the monitoring
site, the expected concentration is 0.08 ppm, and for the case in which
the centerline was 2 km to the east or west of the site, the expected
concentration is 0.015 ppm. It is unlikely that the refinery plume was
directly in line with the monitoring site for three hours. Furthermore,
since the nonmethane hydrocarbon concentration measured at Liberty Mounds
on the modeling days was between 0.17 and 0.24 ppm, it appears that the
influence of the refinery is not significant. Therefore, we used the
nonmethane hydrocarbon measured at Liberty Mounds between 0600 and 0900
CDT as the basis for estimating the background concentrations of the
carbon-bond species.
In preparing the gridded hydrocarbon ground-level initial concentra-
tion fields, the 0600 to 0900 CDT nonmethane hydrocarbon value measured at
Liberty Mounds was assumed to represent the hydrocarbon level that would
exist at the other monitoring sites unless a continuous hydrocarbon
instrument happened to be located at a particular site, in which case we
used the larger of the 0600 to 0900 Liberty Mounds value or the actual
nonmethane hydrocarbon concentration for the hour ending at 0500 CDT.
This methodology produced a fairly uniform initial concentration field,
except for higher values in the downtown area. Available measurements
suggest that there is a relatively high background concentration of
hydrocarbons in the Tulsa area, possibly attributable to natural sources,
oil production activities, or other upwind anthropogenic sources.
The NMOC values assumed to exist at monitoring locations and used to
construct the ground-level initial concentration fields when data were not
otherwise available at these locations are summarized in table A-l, which
lists the background NMOC values. The only instance in which the continu-
ous NMOC differed significantly from the Liberty Mounds values occurred on
3 August, when at the post office site, a NMOC concentration of 0.60 ppm
was observed. The effect of this measurement on the calculated ground-
level initial concentrations of reactive hydrocarbons is shown in figure
A-l(b).
Since there were limited data for CO, S02, and total suspended par-
ticulate (TSP), we chose to use only low background concentrations for
these species. Of these contaminants, only CO is included because it is
one of the reactants in the chemical kinetic mechanism. Because the chem-
istry is not too sensitive to CO concentrations, use of the natural back-
ground value of 0.1 ppm for initial conditions seemed a reasonable basis
for establishing this model input. For the 2 September simulation, the
model was employed to calculate aerosol concentrations, and thus it was
-------
necessary to have initial conditions of SC^ and aerosols. For SC^ we used
a constant 0.01 ppm and a 0.0 yg/m3 aerosol concentration.
The chemistry package in the SAI Airshed Model contains five differ-
ent hydrocarbon species that are referred to as the Carbon-Bond species:
> ETH
> PAR
> OLE
> CARB
> ARO.
The available nonmethane hydrocarbon data must be split among the Carbon-
Bond species. The bag sample data collected in Tulsa were used to
estimate the splits for each species according to the factors presented in
table A-2. At the start of the simulation period when the mixing layer is
relatively shallow, a large portion of the modeling region is located
above the inversion base. As a result, we decided to use the bag samples
collected by the aircraft to estimate the hydrocarbon splits. In this
study, we averaged all the valid data and used the same splits for all the
runs. The resulting splits employed in estimating initial and boundary
conditions are summarized in table A-3. Averaged data rather than day-
specific data were used to avoid placing too much emphasis on a single-
composition determination.
The gas chromatographic (GC) measurements used to analyze the hydro-
carbons from the bag samples did not report carbonyls, so it was necessary
to estimate the splits for these species from other sources. In past
model applications we assumed about 5 percent of the initial NMOC to be
carbonyls, and thus we have employed this value in the Tulsa simulations.
The carbonyl assumption is based on our work in simulating smog cham-
ber experiments (J. P. Killus, private communication, 1979), the hydrocar-
bon fraction of the predictions at the end of simulations for other
cities, and the carbonyl fraction in the emissions inventory for Tulsa.
To test the reasonableness of the assumption, we calculated the carbonyl
fraction from the model results for the last hour (2000 to 2100 COT) of
the 21 July simulation. Using the hydrocarbon predictions at the 10
stations, we found that the average carbonyl fraction was 0.06 with a
range between 0.02 and 0.14. In addition, there was a positive correla-
tion with distance from downtown Tulsa. Therefore, a carbonyl fraction of
0.05 seems quite reasonable. The justification for using carbonyl splits
from the end of a simulation day lies in the assumption that the per-
centage split does not vary significantly from one day to the next. Also
involved is the assumption that the chemical mechanism that generates
reactive intermediates such as the carbonyls is representative of the
actual atmospheric chemistry.
81 5 141
-------
TABLE A-2. SPLITTING FACTORS FOR THE HYDROCARBON SPECIES
MEASURED IN THE TULSA STUDY
(ppm/ppmC of HC species)
Carbon-Bond Species Equivalents
Hydrocarbon Species PAR OLE* ARO1" ETH*
Ethane
Ethylene -- -- -- 0.5
Propane 0.5*
Acetylene
Isobutane 1.0
n-Butane 1.0
Propylene 0.33 0.33
Isopentane 1.0
n-Pentane and cyclopentane 1.0
Trans-2-pentene + isoprene 0.6 0.2
2,2-Dimethylbutane 1.0
Methylpentane 1.0
1-Hexene 0.67 0.17
Hexane 1.0
Dimethylpentane 1.0
Methylhexane 1.0
2,2,4-Trimethylpentane 1.0
n-Heptane 1.0
Methylcyclohexane 1.0
2,3,4-Tn'methylpentane 1.0
Methylheptane 1.0
n-Octane 1.0
n-Nonane 1.0
n-Decane 1.0
n-Undecane 1.0
n-Dodecane 1.0
Benzene
Toluene 0.14 — 0.14
Ethylbenzene 0.25 — 0.13
Xylene 0.25 — 0.13
n-Propylbenzene 0.33 -- 0.11
Ethyltoluene 0.33 — 0.11
Trimethylbenzene 0.33 ~ 0.11
Butylbenzene 0.40 -- 0.10
Each olefin and ethylene contain two carbon atoms.
": Each aromatic contains six carbon atoms.
§ A factor of 0.5, rather than 1, is used for propane to account for the
lower reactivity of this compound relative to that of the other
species that contain single-bonded carbon atoms.
si 10 142
-------
TABLE A-3. HYDROCARBON SPLITS USED IN THE TULSA SIMULATIONS
Hydrocarbon Splits*
Carbon-Bond Type
Ethylene
Paraffins
Olefins
Aromatics
Carbonyls
Benzaldehyde
(fraction of NMOC
by Carbon-Bond type)
0.024
0.268
0.006
0.05
0.05
0.00001
Standard
Deviation
0.026
0.138
0.009
0.036
—
_ _
Ethane, acetylene, and benzene, as well as 50 percent of
the propane, were considered unreactive in our analysis of
the available ambient hydrocarbon composition data.
81 10 143
-------
Preparation of Aloft Initial Concentration Fields
Although the monitoring data provide the primary means for estimating
ground-level initial conditions, they give no direct information about the
vertical profile of pollutant concentrations. The aircraft flights pro-
vide the only source of air quality data aloft. However, only one of the
modeling days corresponds to a flight day; thus, the aircraft data can be
used only as a guide for the other three days. When the mixing depth is
rising rapidly, observations at upwind monitors can also give an indica-
tion of the concentrations aloft. The ozone data aloft, measured during
morning aircraft flights, indicate a consistent pattern. Of the five
morning flights conducted during southerly transport conditions, all but
one (17 September) measured a fairly constant ozone concentration of
between 0.06 and 0.08 ppm. As shown in table A-4, the average ozone aloft
measured during the morning spiral over Liberty Mounds agrees quite well
with the late morning (1000 to 1200 CST) ground-level value reported at
this site. In specifying the ozone levels aloft, we considered the late
morning ozone measurements reported for those stations that were not
influenced by emissions from the Tulsa city area. The aloft measurements
of NOX were generally all below 0.01 ppm, corresponding to the observa-
tions at the upwind stations.
Because of the paucity of aloft data on three of the modeling days,
we used a fairly simple vertical profile for initial conditions. The
cells within the mixed layer were given the calculated surface concentra-
tion, and the cells above the mixing height were set at the aloft values
estimated from the aircraft flights (when available) and the observations
at upwind stations. For the 21 July, 29 July, and 3 August simulations,
we set the initial NO and N02 concentrations above the mixed layer to 0.5
and 2 ppb, respectively. Ozone concentrations above the mixed layer were
set to 0.06 ppm for the 21 July and 3 August simulations and to 0.08 ppm
for the 29 July run.
For the 2 September 1977 simulation, day-specific data were available
from the aircraft flights for specifying pollutant concentrations aloft.
Figure A-2 shows the data collected during the morning spiral over Liberty
Mounds on 2 September. The vertical profiles of NO and NOX are essen-
tially constant and at near-zero levels. The ozone profile shows more
variation, with the concentration near the ground at about 0.05 ppm,
rising to 0.08 ppm at approximately 600 m above the ground and then
decreasing to between 0.05 and 0.06 ppm.
We examined various methods for estimating the background concentra-
tions of hydrocarbons which were employed to characterize hydrocarbon
»*9r 81 5
144
-------
TABLE A-4.
Date
August
26
COMPARISON OF AIRBORNE AND GROUND-LEVEL OZONE
CONCENTRATIONS AT LIBERTY MOUNDS IN 1977
(ppm)
Average Ozone
Aloft
0.06
Late Morning
Ground-Level Ozone
Between 1000 and 1200 CST
0.053
September
2
3
16
17
0.06
0.07
0.07
0.05
0.055
0.065
0.063"
0.051
On 16 September 1977, the skies were overcast until 1100 CST,
after which the ambient temperature increased significantly.
Therefore, the ozone concentration cited here was measured
between 1100 and 1300 CST on that day.
"+9r 81 10
145
-------
(T*» • oooi) aaaiuiv
+ m
O
i/l
I
i
I
^C'
£
•t
-4
1
ol
M
«!
o!
SS
di
en
r~-
cr>
o
o
Q
T3
c
re
c
o
•»->
re
o>
u
3
O
0)
$-
O)
•r—
Q.
re
o
O
CD
^~ ^J
••- JE:
c CD
S- •!-
O i—
re re
CDi—
C 3
•r- t—
rs
t3 r-.
OJ r—
u s.
OJ O)
O -4->
Q.
re O)
4-> I/)
re
Q CM
CM
I
g
=3
CD
>; oooT)
81 18
146
-------
concentrations aloft. The data from Liberty Mounds and Wynona were more
representative of background concentrations than were the data taken in
the city of Tulsa. Furthermore, since the winds were primarily from the
south on the modeling days, and since more data were obtained at Liberty
Mounds than at Wynona, we selected the Liberty Mounds data. Since a large
portion of the modeling region at the start of a simulation is situated
above the mixed layer, we chose to analyze the aircraft bag sample data in
order to determine the characteristic composition of the background
hydrocarbons. Using these average composition figures, the 0600 to 0900
Liberty Mounds NMOC measurements, and assuming a flat initial hydrocarbon
vertical concentration profile, we were able to establish the complete
initial hydrocarbon concentration fields input to the model.* A summary
of these hydrocarbon levels is provided in table A-l.
Preparation of Boundary Concentration Fields
The boundary concentrations on the sides and top of the modeling
region for the 21 July and 3 August simulations were set equal to the
initial concentrations above the mixed layer. Thus, the boundary values
were constant both spatially and temporally. For the 29 July run, differ-
ent values of the ozone concentration were specified for each boundary.
Wind and air quality data reported at the Wynona and Ochelata stations
suggest that different air masses influenced these two stations. Based on
the observed ozone concentrations at times of inflow to the modeling
region, we set the ozone values on the western, northern, and eastern
boundaries to 0.08, 0.07, and 0.06 ppm, respectively. Ozone concentra-
tions on the southern boundary were set to 0.08 ppm based on the Liberty
.Mounds observations and the ozone concentrations specified aloft.
Hydrocarbon boundary concentrations were set to the background values
cited in table A-l.
On 2 September, the ozone measurements made by the aircraft were used
to specify the boundary concentrations at the top of the region. As the
height of the region for this simulation increased, the top of the grid
passed through the ozone gradient noted in the previous section. To
account for the variation in the ozone aloft we varied the ozone concen-
tration in the TOPCONC file. Ozone was set to 0.07 ppm for the hours
before 1200 CDT and to 0.05 ppm for later hours. This variation produced
the added effect of setting the ozone concentration above the inversion
base to 0.07 ppm in the initial conditions. The initial conditions above
Subsequent analysis has indicated that this procedure for estimating
hydrocarbons aloft may lead to an overestimation of paraffin and
aromatic concentration levels.
81 5
147
-------
the mixing height were similarly affected. The morning horizontal
aircraft traverses on 2 September and the Liberty Mounds surface observa-
tions suggest that the boundary concentrations for NO and NC^ at the edges
of the region can be set to 0.5 and 2 ppb, respectively.
PREPARATION OF METEOROLOGICAL INPUTS
The meteorological inputs to the SAI Airshed Model determine the
transport characteristics and rate of mixing of the pollutant cloud.
Meteorological factors often account for the fact that the NAAQS is
violated on some days and not on others, even when the emissions patterns
are similar. Therefore, the windfield and the hourly mixing depths are
important inputs to the airshed model. At the same time, upper air
meteorological data were collected in Tulsa only on a limited number of
special study days. As a result, the three-dimensional structure of the
wind field was one of the most difficult model inputs to specify. We note
that the initial and boundary concentrations aloft were equally difficult
to estimate, again because of a dearth of measurements aloft on three of
the days of particular interest in this study.
The Wind Field
In addition to the airport site, which collected instantaneous values
every hour, four RTI sites measured hourly average wind speed and direc-
tion. As with the construction of the initial conditions, the three-
dimensional windfield was calculated in two steps: first, the surface
field was prepared, then the upper-level windfields. The surface wind-
field was developed using the ground-level measurements plus synthesized
data and an interpolation routine that weights the observations by the
inverse of the distance from the grid square. Mathematically, this
relationship can be written
n
E .
i=l uirik
u, =
k n -1
£ ik
1=1
n
E -i
1.1 Vik
k n „ -1
ik
81 5 =
148
-------
where u^ and vk are the calculated horizontal velocity components for grid
cell k, u-j and v-j are the measured velocity components at station i, n is
the number of wind measurements, and r^ is the distance between monitor-
ing site i and grid cell k. A smoothing algorithm, which replaces each
velocity component estimated above by the average velocity component
calculated over the block of nine surrounding cells, was used to smooth
the calculated wind field to eliminate any discontinuities introduced by
the interpolation procedure. Finally, we calculated the average measured
wind speed and the average computed wind speed over the modeling region
resulting from the preceding calculations. The wind speed in each grid
cell was then scaled by the ratio of the average measured speed to the
average computed speed. This last operation was implemented since the
vector averaging processes embedded in the interpolation and smoothing
steps yield a computed flow field with wind speeds that are lower than
those exhibited by the original observations.
The wind field calculated from the reported data only was compared
with the pollutant observations. The comparison was carried out both
visually and by exercising the airshed model. It was clear from the
comparison that some modifications were necessary for most of the modeling
days. For instance, there were hours for which the wind observations
suggest a more complex flow pattern than that which could be calculated by
simply interpolating the observations.
Two modifications were made to the wind data—adjustments to the
actual measurements and the addition of synthesized data for locations
relatively far from any observation sites. These modifications were subj-
ectively based on wind observations, air quality observations, and synop-
tic weather maps obtained from the National Weather Service. Synthesized
data were developed using available information to draw streamlines and
adding synthesized wind data so that the calculated field follows the
streamlines.
On 21 July 1977 the winds were from the south for almost the entire
day with little variation from site to site. Since no problems were
encountered in calculating the surface windfield for this day, no modifi-
cations were made to the reported wind data. The surface wind field for
selected hours on this day are shown in figure A-3(a).
The air quality data for 29 July 1977 suggest a very sharp spatial
gradient in ozone concentrations across the city. The monitor at Apache
81 5
149
-------
UJ
ocn
CEO
Cd 21
UJ—•
CM
ii i . i i i i i i i i r i - i i i i i
tn
o
o
o
o
o
en
O>
a:
10
to
0)
_C
i-
o
in
LD
-o
01
O
Ol
o
00
S-
o
Lu
O CO
CEQ
ct: z
LU —•
LU
LU
LU
•>•
LU
LU
>
LU
-a
"ai
•o
c
tn
«-"(-
<=}
tn
«-
CD
on
__( ) [ | 1 ^^ 1
H-
_
-
— UJ
f w v T~T * ^ "t
u
—
-» . « ^ — a < ^ — «—
— a
*"!* 1* 1" r1^"^"
0
oo
0
CNJ
• T, u i- ' • '•*-!»
J^-A-^--»- ^r-«^«=
>•
• fl • • "t "• "J
— — — — _ .^
a
* - - rf£ * * *
—-*-»-»--—
*T* r* i*1!* ^r*
o
CsJ
~-k$ 4s. 1^— ifr*l*J-»-.
-1 -• "M "V. "T f
"T ^ "W "• 1_ V
JB *^B W "• • •
LJ ^
f- •« f ^~-« B M
•*- ^ ^t~ "f- T™ —
O
i^4_S-^,^-LjKL?
a
c
_
_
*~"*~'*^*~'~"'*~''*~~'*~"_
(n _
£ —
-
-
—
^j^j^-^j^^,^ jt _j^^ ~
—
-
C
3n
0
in
O
cn
o
o
(T)
o
o
CO
OJ
4->
-------
o to
cc o
en 2
LU •—'
> 3
CL
LU
>
LU
LU
to
o
o
o
in
o
o
o
LH
LJ
LU
CD
•a
LU
CD
CJ
O
O
CO
O
O
C\J
T3
Ol
O
O
CO
ef.
i+9r 81 18
cn
151
-------
LU
OLD
cr a
0:2
UJ •—i
LU
CM
en
u
o
o
o
o
o
en
-r-0
tn
OJ
o ^
LU
O CO
CE O
LO
C
r
>'»'»
» v »
» x x
X » X
» > X
XXX
* * *
XXX
•-XX
»; X X
C. x x '
•. V x,i
k>V
C
r
D
")
„! vlx Ix Ix
> > x x x
» X X X X
x » x x x
X » X X X
UJ
> * » X X
X ^ X X X
X X X X X
X X X X V
X X X ^. V
•< x. X.X.. x.
K. ms.mv'v^«.
^ l" IMX 1
D
r)
0
C\J
IX f*v 1**. ^*. ^. ^ ^ ^- ^ ^v ^ ^
XX,VX.VV*s^,*\^ ^\ ^.
x,^%»*%.*tVVVK^
» x, * % %%%%%V%.%
KjV VV*vVV^VV ^ ^
V V V *- V X V J^ *^^- *^ ^
x v v t v v v*^ ^£~ v x. x
Q- i ID
x x v x v v x^. v v \ \
1s" Is" l" I 1* p 1* 1 1 1 II
o
CM
0
w—t
iV'TX IV|KIVIMVIVIU
tn
|k|* 1*1*1* 1*1*1*1*!
O
C
< 1 « 1 « 1 *
_
_
-
_
—
-
c
In
*-•
o
in
en
o
o
oo
o
o
r-
•a
(U
3
cr
c.
o
(-3
CO
0)
i.
3
cn
81 18
152
-------
LT>
o _>
UJ
O CO
a a
or z
UJ •— i
> x
CE
O
cn
-\X\N
-\ \ \ ^
_X X X X
_X X X X
-X X X X
* x x v
_» » x x
» * * *
~t t t /
-f f f f
Vlf f
o
cn
•«r cn
_i _i
-» LU •* UJ
> >
UJ LU
_l _J
O
rsj
H»*»^^x^»-*- t J J
f t t f t t f t f f f f f f f
t t it 1 t t t t t f f f f t
,tkl*,ttt.Wt'tWttt
"* o
OvJ
CXI «-•
•* LU ^ LU
> >
LU LU
_l —1
0
<***?????/ SSSs
it***/1///1// SS,
f / / / SJ&S//S///7
t t / / /£///////?
t f f f /'////'/// /7
///////// / / / /7
///////// ////7
f t t t t f ///////7
/ / / //////////7
/ / / //////////7
0 C
^
' t—
_ cn
0
^ 2
— < o
r 1
Q
'Ln o
r CO
r
D
tn
v^
z:
•o
ai
3
C
C
o
O LO
a a
or z
LU ^H
LO
O
cn
o
UJ
o
r-j
LU
1 * '* '
t> A
<• r
TT<
J l
t 4 »
1« * * 4
* t t *
/ft*
/, X. C.t
I I II
\ \
\\
,\.N
111!
A i \ ,\
Ln
o
cn
O
CNJ
cn
o
o
CNJ
O
O
•a
OJ
CO
I
•a:
O)
i.
=3
cn
81 18
153
-------
o en
era
(SI
LU
>
\
LJ
o
o
CD
o
o
r-
o
LU
in
o
c
o
ocn
CE D
on -z.
LU—i
> 2
cr
CO
o
o
LD
O
O
ID
c
o
CJ
<£
0)
3
CT>
81 18
154
-------
UJ
o en
CL O
CC 2
LLJ
LO
yt ' • • ' '
7/777? '
-777777 l
-7/77/7 '
7/7777;;
~L f* L l&i L I.
L'/ 'f'f'f'f'ltl t
-f!f!ta'ta\\\
\
' \ \ \\
1 A A A
c f f f r lk
'7
77'
f,f
o
en
o
CNJ
\\\\XN\\V
J I L
iumN
mu\-
.^::>i^ ."w.-.-
o
o
en
o
o
CD
C_) *"
CO 3
C'OT
cr D
a; z
ITJ
o
o
rn
in,
^-,J-.^^LvLJ-^UI J ,1 ,i J ±-±-l~±~2~±.
1—I—I—T
-«—*-—^—^—-*—«-
•-of-'
^-.^^.•••^^--.^^ -
^^ --^^-»*,-«^.»—«*-^«tfc^
W
'~-
-------
CD en
CCD
CM
LU
v / /'/ y
7 777 7
/
un
en
o
o
o
o
o
o
in
en
LU
ocn
cr CD
cc z:
C\J
_l
LU
LU
in
en
o
o
o
O
o
en
•o
01
c
o
CO
Ol
S-
3
CD
81 18
156
-------
LU
ocn
CEO
LTZ:
UJ •—<
>2
LU
(\J
LU
>•
UJ
in
cn
o
o
rn
o
o
c\j
SS SS// o
__o
•o
01
LU
UJ
CD CO
CEO
cez
LU—'
C\J
en
o
o
o
CM
o
o
01
-M
c
o
CO
I
OJ
s-
13
cn
81 18
157
-------
LL)
ocn
CL
m
t\l
UJ
I 5-1 II III 1 III I I I I I I Illlllll
- in
in
o
o
r-
o
o
CO
-o
^ o
UJ
CD en
O
en
o
o
o
o
CD
3
C
O
O
QJ
S-
3
cn
81 18
158
-------
Sen
£9 *\ -1 *-^ -1
C£ ^- ^ LU — LU ••
LU •— ' > >•
>• 2 LU LU
cr _i _i
O 0
LO ^ rxj
;. _ „_ ^ ..^^ „
«-»-»- »- -^•~'»^»^»-J^*--»^»^^~.'»-.* >
LU LU
— 1 _J
O
«-H £
I —
1- 1* 1* 1* HHHKHHK
0 C
3/>
^_
CD
O
O _
0
Q
«— t
1
- %
CO
o
D
cr>
LU
LU
>
LU
LD
LU
0 0
CO C\J
kUlfcU^L U UJ» 4 J-J O
n_
bJ >- 5^ ,
_, ^ -, ^ ^. ^ ^ -, -, fe, . . -c. ,
.^. ______ 1- 1 • • « •
-
- * "*" *"*-^ *-*-*-*-«-• •-«-«-*-*-»
- C3
•- r t-1-1 v?i i1 1-1 t-i *i i "i *i "i
0 0
tn CM
o
•— * C
«l «l J -J 1 J^l_UL-L.UUlt»l>l>l&
-
jn g
= z -
-
-
^iiTrrrii"! 1 1 1 1 t 1
0 C
^Ln
i-«i
Jrl
LO
D
s_
-------
UJ
oco
£ §
5- O
oto
era
CNJ
tn
o
o
o
c\j
o
o
-a
at
•o
3
O
c
o
o
ro
s_
3
Ol
81 18
160
-------
recorded 0.16 ppm ozone levels from 1100 to 1500 CDT, but the three other
city stations, Mohawk, health department, and Post office, did not measure
ozone levels above 0.10 ppm for these hours. We attempted to adjust the
windfield to reflect the observed ozone concentrations by adding data
along the sides of the region and modifying some of the wind observa-
tions. We received a copy of the strip chart recorded at the Tulsa health
department and used these raw data to adjust the downtown wind data. The
strip chart pointed out that the wind direction, if automatically
averaged, appears to vary rapidly between 20° and 340", but the reported
hourly average was about 180°. As a result, we modified the wind direc-
tion in the city area to be reflective of the strip chart recording.
Examples of the wind field throughout this day are shown in figure A-3(b).
For the 3 August 1977 simulation two changes were made to wind
data. The data collected at Liberty Mounds were not used to prepare the
wind file for this day because the wind measurements at Liberty Mounds
were inconsistent with the other measurements and because the wind direc-
tion monitor was inoperative from 1300 to 1800 CDT. The other change made
to the data was the subjective specification of wind data at six addi-
tional locations for the hours between 0800 and 1400 CDT. During these
hours the windfield was quite changeable; for example, in this time
period, the monitor in downtown Tulsa observed a 360° turning of the
wind. We used the surface wind data plus the three-hour surface weather
maps to construct possible flow patterns for each hour. The synthetic
data were added so that the final interpolated windfield would follow the
observed patterns. Figure A-3(c) gives examples of the wind field for
this day.
The winds were primarily from the south on 2 September 1977. During
the middle of the day similar wind directions were observed at all of the
sites, but the wind speeds at Liberty Mounds and Ochelata were much lower
than those measured at health department and Wynona. We increased the
wind speeds at Liberty Mounds and Ochelata to match the speeds at the
other two monitoring sites, though this had little effect on the model
predictions. Some modifications were also made to the wind directions
reported at Ochelata in midafternoon.
At about 1500 the winds shifted from southerly to southeasterly at
the health department and Wynona sites. Southwesterly directions reported
at Ochelata at this time considerably influenced the calculated flow pat-
terns north of the city. Although the air quality data suggest that pol-
lutants were carried to the Skiatook and Wynona areas in midafternoon, the
wind directions calculated using the measured directions at Ochelata
yielded wind inputs with a southerly orientation (the southeasterly winds
at Wynona countered by southwesterly winds at Ochelata). Therefore, we
modified the directions at Ochelata after 1500 CDT to be more consistent
<*9r 81 5
161
-------
with those measured at the health department and Wynona. The wind field
for selected hours on this day are shown in figure A-3(d).
The Oklahoma City rawinsondes and the Tulsa pibals on selected days
were the basis for estimating wind shear. Since the upper-air data were
sparse, a relatively simple algorithm was used for calculating the winds
aloft. Winds aloft usually show less spatial variation than do surface
winds and are more reflective of the larger-scale synoptic conditions.
Therefore, the spatial variations exhibited in the surface winds were
decreased with height, which resulted in relatively uniform wind veloci-
ties aloft, reflective of the data taken from the rawinsonde measurements.
For the hours between rawinsondes, the winds aloft were averaged from
the rawinsonde measurements closest in time. An examination of the
limited pibal data gathered in Tulsa and the corresponding rawinsonde data
collected in Oklahoma City indicates that the wind direction in Tulsa was
about 20° different (in a clockwise direction) from that in Oklahoma
City. Thus, for days when no upper-air data were reported for Tulsa (21
July, 29 July, and 3 August), we employed the Oklahoma City data and
rotated the wind by 20* in a clockwise direction. Since pibals were
released in Tulsa at 0640 and 1035 CDT on 2 September, we employed these
data without modification in the specification of winds aloft.
Figures A-3(a) through A-3(d) include the average wind for each
level. This gives some idea of the wind shear resulting from the applica-
tion of this methodology.
The Mixing Height
The height of the inversion base for the modeling days was estimated
using the rawinsondes taken twice daily at Oklahoma City and the measured
surface temperatures at the airport in Tulsa. For the 2 September
simulation, we employed the morning and afternoon aircraft spirals to
establish the vertical temperature profile in the Tulsa area. The
rawinsondes were released at 0700 and 1900 CDT. For all the modeling days
the morning sounding indicated a surface or near-surface inversion, and
the afternoon sounding indicated that the inversion had broken up within
the lowest 1000 m of the atmosphere. The one exception to this was the 21
July sounding. On this day cloud cover in the afternoon caused the mixing
height to rise to only 800 m above ground.
To estimate the mixing depth between the two rawinsondes, we extrapo-
lated the surface temperature at the Tulsa airport upward at the adiabatic
lapse rate until it intersected the measured vertical temperature pro-
file. In cases for which the temperature at a particular height differed
<»9r 81 5
162
-------
significantly between the morning and afternoon soundings, we employed a
straightforward linear interpolation (in time) to establish suitable
temperature profiles throughout the day. These adjustments were made to
reflect atmospheric heating and cooling effects. An example of the mixing
depth estimation procedure is shown in figure A-4 for 21 July. The
diurnal variations of the mixing depth for each modeling day are shown in
figures A-5(a) through A-5(d).
The mixing depth was kept spatially constant over the entire modeling
region. However, the minimum mixing height was set at-100 m for the city
and at 50 m for the rest of the grid to account for the heat island effect
and the increased mixing expected in the more populated sections of the
modeling region. No other spatial variation was included because we
lacked data and physical features that would suggest spatial differences
in the height of the mixed layer.
Other Meteorological Inputs
Six meteorological scalars that vary in time only are required as
inputs to the SAI Airshed Model. Information needed to calculate each
scalar is contained in the routinely collected meteorological data. In
this study, the temperature gradient above and below the inversion was
calculated from the rawinsonde data. The concentration of water vapor was
calculated from humidity data collected at the Tulsa International Airport
by the National Weather Service. The atmospheric pressure was set to the
nominal value for this area.
The N0£ photolysis rate constant was based on the solar radiation
data collected at Skiatook using the equation k} = 0.4 x solar radiation
(in langleys) (Jeffries, Fox, and Kamens, 1975). This formula does not
consider the attenuation of solar radiation with height due to light
scattering from aerosols. The airshed model does calculate this attenua-
tion when aerosols are simulated, as in the 2 September simulation. Since
the aerosol concentrations calculated for the 2 September run were not
high, the variation of the photolysis rate with height was not considered
significant. We do not expect higher calculated aerosol concentrations
from the model simulations for the other days. Thus, ignoring the
attenuation for the other days is expected to have had little effect on
the model predictions.
One adjustment was made to the above methodology for specifying the
photolysis rate constants. At 1600 CDT on 21 July the solar radiation
measured at Skiatook fell to near-zero levels due to cloud cover. Exami-
nation of the NOX and 03 data collected during the afternoon of 21 July
suggests that photochemical reactions did not stop at 1600 CDT or even an
i+9r 81 5
163
-------
4500
4000
3500
3000
gj 2500
c
* 2000
1500
1000
500
I *8
1800 CST
SOUNDING
Surface Temperatures
St2 Measured 1n Tulsa
300° 304° 308" 312e
Potential Temperature (*K)
316"
320e
81 1 8
Figure A-4. Temperature soundings in Oklahoma City and Tulsa
surface temperatures for 21 July 1977.
164
-------
o
loo
CO
I CM
GO
Q)
O
s-
CTv
i—i
>>
3
•~D
«—I
00
i_
o
o
QJ
I-
O)
TD
O
QJ
Q-
O
0.
o>
•o
00
O)
M-
O
c
S-
O)
ro
D-
ro
C
S-
LO
I
cu
I-
3
O
O
CVJ
O
o
o
o
o
00
o
o
VD
O
O
O
O
00
(puncuO
sj3}aiii)
81 18
165
-------
en
i—i
>>
3
00
o
(sua^aui)
IT)
«£
S-
3
cr>
81 18
166
-------
o
CM
CO
I/O
O
ro
O
OJ
S-
3
O
CD
ro
o
Q- D.
UJ O
.
z: o
CO
O
O
CM
O
o
o
o
o
00
o
o
o
o
o
o
CM
T3
O)
o
o
LT>
I
-------
1 1 1
1 1
D- D-
UJ UJ
O O
CD CD
zi zr
i—i i—i D.
X X O
_J Z O
«=C < "-1
o: co tu
Z3 o: uj
ex. => a:
g
CO
I CO
(O
o
CM CD
I -C
+J
M-
o
s_
3
O
O)
-Q
O)
•M
a.
O
o
o
o
-o
OJ
T3
3
O
c
o
IT)
a>
S-
cn
81 18
163
-------
hour or two later. Upon review of the surface wind data, we found that
the wind directions reported at Wynona, Ochelata, and health department
shifted from a southerly to westerly or northwesterly orientation at 1500,
1700, and 1800 CDT, respectively. Since the drop in solar radiation mea-
sured in the northwestern portion of the modeling region seemed to be
correlated with the shift in the surface wind directions, we suspected
that the solar radiation might not have been spatially uniform over the
modeling region. To account for the observed ozone, we averaged the 21
July solar radiation data with the data taken on the next sunny day.
The following tablulation was used to estimate the exposure class,
which was also based on the solar radiation data:
Solar Radiation
(langleys) Exposure Class
0. - 0.25 0
0.25 - 0.75 1
0.75 - 1.25 2
> 1.25 3
The last meteorological input is the surface temperature, which was
set spatially constant for the entire region based on the hourly values
measured at the Tulsa airport. These values and the hourly values for the
other meteorological variables for each day are summarized in table A-5.
EMISSIONS INPUTS
The EPA contracted with Engineering-Science to develop a complete
emissions inventory for the Tulsa area. The only modification we made to
this inventory was to double the grid spacing.
Two issues were raised during this project concerning changes to the
emissions inventory. One was the differences in mobile sources on
Fridays. The other involved possible changes in the reactivity of per-
chloroethylene. Two of the modeling days, 2 September and 29 July 1977,
were Fridays. We discussed the difference between average weekday and
Friday emissions with Mr. Jerry Howell of the Indian Nations Council of
Governments (INCOG). We concluded that mobile source emissions on Friday
might be from 10 to 15 percent greater than those on average weekdays, but
there were few data available to quantify any spatial or temporal changes
to the emissions inventory. Since the uncertainty in the complete emis-
sions inventory is likely to be at least as great as any proposed change,
we decided to make no modifications to the inventory.
81 5
169
-------
0>
J- O
O) o
Q-- — '
0)
ovJCMCMOjoooocooooooocorocvjojcMOocM
oo
V. * •—
OJ S- E
-t-> +J O.
ro c a.
3 OJ >
>
3
Ol C S-
1_ -i- oj
3 >>
4-> +J (O ^~
<0 C _l E
S- O) —
OJ '<- T3 O
Q. T3 O) o
& n) X •^~
CU i- -i-
H- C3 S
000000000000^00000
OOOOOOOOOOC3OOOOOO
I I I I I I I I I I I I I 1 I I I
uu.
o
oo
LU
=D
3
•
^XJ
I-
o
LJL.
^-
TO
"*'™*x
OJ
3
S_
O)
D.
E
O)
QJ
O
ro
4->
C
»— -
0
T3 «
Qj ^»-^
X
••—
^D VD VO VXD VO v£D VO ^ ^O VO ^ ^O VD »jD ^D VO VO
co ro ro ro ro ro ro ro ro ro ro ro ro co ro ro ro
OOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOO
I I I I I I I I I I I I I I I I I
CD
O)
$_
3 I/I
«/l
o — icoo^Ofr>O
ooo<— ico^-Lfi
'— •cooco'— l
0000
(DodooddoooocJooooo
E \or~~coo^O'-icofo^j-Lnvor^oocriO
OOO OOOi—1«—ii—(«—*<—ii—it—it—II-HI—ICM
81 10
170
-------
OJ
s_
re O
S- »
O) ~_-
D.
OJ
s_ re
0)
s-
-l-> O-
c: o.
O) -
U
c:
O
O
ooooooooooooooooo
ooooooooooooooooo
ooooooooooooooooo
T3
O)
-1-1
c
O
C_3
CT1
i—I
>>
3
01 c s_
S_ -r- d)
= >1
+j +j (d .—•
—-
"C
Q. -I— QJ
E 13 X
d; fO -i—
I— s- 5:
O
ooooooooooooooooo
ooooooooooooooooo
ooooooooooooooooo
i i i i i i i r i i i i i i i i i
(D
S_ in
3 CO
co «3
O i—
CL CJ
X
OOi— I
OOO
CO
co
O coi — i
-l-> C I
o o c
C\J4->
o re
sd" V£> ^- CO
OLn^-cot^^-OLOcriOr-criCTivoo
oooi—(CMoo^-«^-«^-«3-ooroooi—1000
oooo ooooooooooooo
CU (—
E oo
•c- C_3
ooooooooooooooooo
ooooooooooooooooo
COCTiO
OJ
OOOO O O t— I i — I i — li — I r— I i— I i — li — ii — li — I
49P 81 10
171
-------
O)
s-
3
4_> ^~.
03 c_>
d) ^—'
Q.
O>
c
o
U E
4-> Q-
C Q.
Ol -
U
C
o
OOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOO
ooooooooooooooooo
C\JC\JCVJCMC\JC\JCMC\JCVJOJOJCVJCMC\JCVJC\JCSJ
T3
01
c
o
IT)
CO
CTl
tn
3
CD
ro
S-
o
O) C S_
S_ ••- <1)
3 >>
4-) +J ro ^--.
"3 C _J E
S- OJ •«->.
O) ••- T3 O
D. T3 d; o
E
3 to >,
ooooooooooooooooo
ooooooooooooooooo
I I I I I I I I I I I I I I
t. C
O) OJ
a>
O
OOOOOOOOOOOOOOe-Hi-i^-i
ooooooooooooooooo
ooooooooooooooooo
a>
S- 00
3 00
oo re
O i—
Q. O
X
LlJ
OOO»—t
oo
•i-
oo
O ooi— i
4-* C I
O O C
CVJ+J
o
-------
S_
3
S- o
O)
Q.
a;
csj CM ro vo oo 01 i—it—i c\j CM ro m CM i—icot^r-v
CMCMCMCMCMCMrooororoooooooroojoooo
c
o
O) i- E
4-> -l-> Q.
CD
<
01
i.
O)
JD
E
a>
4->
Q.
OJ
oo
O) C S_
S- •!- O)
•4-J 4-3 (T3 -—•
c x —•
Ol S- •!-
O)
s_
O)
coooooooroporororooooorooororoooro
OOOOOOOOOOOOOOOOO
ooooooooooooooooo
ooooooooooooooooo
I I I I I I I I I I I I I I I I I
O)
O i-
O)
rO
re
rO
S-
O) O) "O
Q. -i- O)
O
n3 -i-
5- 5!
CD
ooooooooooooooooroco
ooooooooooooooooo
ooooooooooooooooo
ooooooooooooooooo
01
S- in
3 1/1
t/1 (O
o •—
Q. O
X
OOO-— <•— i
r-i«— 1000
(/I
• i- 4J
01 C
>> (O
O
-(-> C I
O O C
-C O •!-
Q- E
O) --- '
a:
O
OOO>—lfMrO«5fLOLn<3-«d-CMi—li—IOOO
ooooooooooooooooo
tn
CM
O)
E
ooooooooooooooooo
ooooooooooooooooo
^j-Lr)ior^.aDcriO<—tojoo^tijniot-^cocyiO
OOOOOOr-li—I r—I i—t,—I i—I ^H i—I,—(!—IOJ
81 10
173
-------
Perchloroethylene is a solvent widely used in dry cleaning. Engi-
neering-Science considered this compound unreactive when they compiled the
Tulsa emissions inventory. After further investigations, we concluded
that perchloroethylene is less reactive than propane, which is considered
to be half as reactive as the higher-carbon-number paraffins. Emissions
from dry cleaning establishments account for less than 5 percent of total
hydrocarbon emissions, and since perchloroethylene has a relatively low
reactivity, we chose not to alter the reactivity of perchloroethylene. If
we had changed the reactivity to the low value, we would anticipate a
change in the ozone predictions of no more than a few percent.
Tables A-6(a) through A-6(g) present the total gridded emissions for
NO, N0£ paraffins, olefins, aromatics, ethylenes, and aldehydes for the
period from 0100 to 2300 COT. This is the time period over which the
model simulations were conducted.
MISCELLANEOUS INPUTS
Additional inputs required by the SAI Airshed Model include gridded
fields of two ground-surface characteristics—surface roughness and a
surface deposition factor. The surface roughness is used to calculate the
vertical diffusivity coefficient. The surface deposition factor is
employed in the calculation of the surface deposition velocity, which is
used to estimate the rate of pollutant removal from the atmosphere by the
surface uptake processes described earlier in section 5. Since these two
variables depend on land use, vegetation, and the types of development
found in each grid square, their values can be estimated employing maps
that describe land use. U.S. Geological Survey maps having a scale of
1:250,000 for the Tulsa area were useful for this task. In addition,
land-use maps prepared by the local planning agency were consulted in
developing gridded fields of surface roughness and surface deposition
factors. Table A-7 gives the estimates of surface roughness and deposi-
tion velocity for each land-use category we considered.
»*9r 81 5
174
-------
8 8
o o
888
o o o
888888888
888888
O O O O O O
8 g 8 8 8
O O O O O
8 8
o o
8 § 8
o o o
888088000
ooooooooo
Sa n N < n
n - * n o
o o o o o o -
o o o o o
§o c - c
o c c c
o o o c c
8 8
o o
888
000
8888o-fitt-
ooooooooo
8888
O O O O
8 g g S i
O O O C C
8 8
O O
888
000
g888SS58£
OOOOOOOOO
-N--OOOO
8888
O O O O
S 5 S S g
o o c o o
00
o
H—1
oo
00
8 5
O O
8 2
O O
8 8 §
O O O
S £ S
O O O
oooooooooo
O W •* N W N
*- m n - o o
o o 5 o
o o o o
o o o o o c
-onaio-^o
OooonnpjN
oooooooo
: s 5
o-nnnooooooooo
o o o
o o o
oooon — rt*
oooooooo
* 5
— OOOOOOOOOOOOCC
c:
o
8 o
o o
g g
o o
888
O O O
S 8 8
o o o
8_ n M ,0 n — nt
o o o n o o o
oooooooo
ooooooooo
f* o- n CD ni •«
r» K * » n *
o * *• « o o
n B n » » 9-
o ri n •« - o
o - o o o ©
» n n n
o o o o
o o o o c
o o o o o o o
8 8
o o
888
O O O
85188888
O - ^
CD
8S885S8888S888
oooooooooooooo
85885888888888
OOOOOOOOOOOOOO
88000888880880
oooooooooooooo
g 8
oooooooooooooo
8 § 8 S §
O O C C C
8w£~»wOQnc
onooooo-
ooooooooo
o o o
o o o
8881
o o o o o o
ggggSggg
oooooooo
o o o - c
ooooooooo
ooooooccc
8 8
B R
888888888888
88888888
o o o c
o c o c o
S R R S
S S S
1 l
175
-------
8 S
O O
8 8
o o
8 8
o o
8 8
o o
8 8
O O
8 8
o e
8 8
o o
8 8
o o
8 8
o o
8 8
o o
88
o o
8 8
o e
888
o o o
8 8
o o
888
o o o
8 8
8 8
o o
8 8
o o
8 8
o o
8 8
o o
5 8
o e
8 3
e o
8 3
o o
8 Z
a o
8 8
o o
8 8
O O
888
o o o
8 8
O O
888
o o o
8 8
O O
8 8
o o
8 8
o o
8 8
o o
8 5
O O
8 3
o o
3 3
e e
S 2
o o
3 :
o o
3 8
O O
85
O O
8 8
o o
888
O O O
8 8
o o
888
O O O
8 8
o o
888888
8 8
o o
8 I
o o
S S
e e
— n
o o
- 2
o o
8 3
o, e
e o
o o
8 8
o o
888
o o o
8 8
o o
858
000
8 8
o o
8 8
O O
8 5
o o
o o
o o
8 o
o o
o 8
o o
5 S
e e
o o o o
s s
o o
8 S
o o
8 8
o o
888
O O O
8 8
o o
888
o o o
o o
o o
o o
o o
5 8
o o
8 8
o o
5 8
o o
8 e
o e
e o e o
n n
o o
8 3
o o
88
o o
8 8
o o
888
o o o
8 o
o o
n » o
o o 5
o o o
OJ
=3
C
8 S
o o
8 8
o o
8 8
o o
8 8
o o
8 o
o o
5 3
o o
S o
o o
2 3
o o
n -
o o
S 3
o o
88
o o
8 8
o o
885
o o o
3 3
o o
O O O
O O O
c
o
-------
88888
§ 8
o o
888
O O O
88
O O
88
O O
888
O O O
888
O O O
8 8 8
o o o
8888
0000
O O O O O O
888
000
8 8
O O
888
000
o o
o o
o o
o o
ooooooooo
5888
O O O O
000
000
o o o
o o c
8 S S
o o o
o o
o o
o o o
000
8 £
o o
8 8
o o
8 S Z
o o o
o o o o
o o o o
Sv «
o o
o o o
3n o
o o
o o o
0)
3
C
C
O
to
<:
I/I
re
000
o o o
8 5
o o
006
000
S 8
o o
8 5
o o
— o o
Srt «
o o
o o o
5 S 5 8
o o o o
8 8 S
o o o
o o o
000
N 4 4
o o o
000
888
o o o
o o
o o
o o o
o o o
o o
o o
n *•
o o
O O O
o o o o
o o o o
o o o
o o o
o o o
o o o
o o o o o
s =
o o
s s
o o
ooooo
o o
o o
o o o
o o o
r *
o o
no-*
n n «•
oo-
n n *
n tv o
o- » o
« «• n
w - o
o o o
o o o
8 8 S
o o o
*
o o
o o
« r>
o o
o o
o o - o o o
— o o
o o o
rv a> «
CD r> o
o o o
— m r>
n n -
O - o
S 2 §
O O O
o o o o
o o o o
o o o
o o o
o o o o o o o
o o o o
IK O O O
n o o o
o o o
o o o
o o- o
o o o
o o o
o o o
— o o
m m o
O O O
S 2 §
CD O O
8m c
o o
o o o
S 8 g
O O O
CD
088
o o o
o S S
000
o o §
o o o
*• o
o o
o o
o o
o o
o o S
o o -
n o
o o
o o o
o o o
o o
o o
o o
o o
o o o
5* o-
- o
o o o
258
O O O
n *. o
o o o
o o o o
o o o o
o o o
o o o
o o c
o o c
o o o
o o o
o o o o
o o o o
o o o
o o o
o o o
o o o
5 o o
000
8 5
o o
o o o
8 S
o o
5 8
o o
o o n
o o o
o o o o o o
o o o o
o o o o
o o o
o o o
o c o
o
o
§
ooo
ooo
— — r»
OOO
ooo
oo
oo
rt —
OO
oo
oo
ooo
OOO
ooo
oo
oo
O —
OO
oo
oon
ooo
nr)»
OOO
ooo
ooo
ooo
nom
--O
ooo
o o o
o o o
ooo
— r% w r>
0000
0000
o o o
o o o
o o o
•c — n —
o o o o
o o o o
o o o
o o o
O O C
o o c
888
O O O
8 8
o o
888
000
S8888888888S88888
O O O C
177
81 18
-------
8888888888888
OOOOOOOOOOODO
§888888888888
ooooeoooooooo
88888888S8S8S
ooooooooooooo
888888885S38S
ooooooooooooo
8 8
o o
888888
o o
o o
3 8
o o
— — ft
o o o
o o
§
o o
5 8
o o
8 8
o o
o o
o o
5 8
o o
88888888
OOOOOOOQ
88888888
oooooooo
S 8 8 8 § 8 8 8
oooooooo
88888888
8 8 S
o o c
•o
O)
3
C
C
O
I
UJ
_1
CO
8888888880000
ooooooooooooo
S588S888833S5
OOOOOOOOOOOOO
88888888888008
O)
O
8888888885888
o o o o o o o oooooo
8888888888888
ooooooooooooo
o -
o o
I 8
o o
S 2
o o
8 8
o o
8 8
n S
O O
S S
o o
= 8
o o
- o
o o
58888888
OOOOOOOO
o o c
8 S 8
O O O
o o o
o 8
O O
8590500-
o 5 o o o S o
oooooooo
o o c
o e o o
•> — » N
•• •• o o
— o o o
8 8
O O
oooooooon
S 8
o o
88888888888888888
ooooooooooooooooo
88888888888888855
oooeooooooooooooo
o o
o o
oo
oo
80
o
oo
o o o o
0000
8858
o o o o
oooooooo
oooooooo
-0000000
oooooooo
oooooooo
88888888
oooooooo
88888888
o o c
o o c-
8 8
8 S s
8888888888888
ooooooSooSooo
ooooooooooooo
8888888888888
OOOOOOOOOOOOO
88888888
oooooooo
5 8 8 8 8 S g 8 § S
888888888888888888888588888888
8
8
o
8
R
O O C
oooooBoo
OOOOOOOO
88888888
ooeooooo
oooo
88888888
oooooooo
S 8
o o
81 18
173
-------
8 8
o o
8 S
o o
8 8
o o
8 § 8
o o o
§ 8
O O
8 i
o o
8 8
o o
8 8
o o
8 8
o o
8 8 8
000
8 8 § § 8
o o o o o
8 g
C C
8 8
o o
8 8
o o
8 8
o o
8 8
o o
888
O O O
8 §
O O
8 S
O O
8 8
o o
S 8
o o
§ 8
O O
888
O O O
O O O O O O C
8 8
O O
8 8
o o
8 8
o o
8 8
O O
858
O O O
o o
o o
o o
o o
8 S
o o
2 *
O O
8 S
O O
088
O O O
So o o o
o o o o
o o o o o
8 8
O O
8 8
O O
8 8
o o
8 8
o o
8 8
o o
8 8
o o
o o
o o
5 o o
o o o
885
o o o
8 S
O O
S S
o o
8 o
o o
2 8
O O
o o o o o
O O
o o
8 8
o o
888
o o o
n o o
ooo
000
o o o o o o
o o
o o
oooooooc
8 S
o o
o o
o o
o o
o o
8 8
o o
O O O O O
n D-
o o
o o
S R
O O
o o o o o
888
OOO
oooooooc
O)
3
o
C -l->
O "3
O E
^ O
VO <
I
8 8
o o
8 8
o o
o o o o
8 8
o o
8 8
o o
8 8
O O
8 8
o o
8 8
o o
5 5 o
OOO
888
OOO
888
OOO
8 8
O O
8 8
O O
o -
o o
8 8
O O
8 8
O O
S S
o o
8*
-
OO
S o
o o
- o
o o
o 8
O O
888
OOO
OOO
OOO
- p o
o 5 o
OOO
o o o o o n, o
o o o o o
o o o o o
o o o o o o o
QJ
80
o
CD
o o
o o
o o
§o
o
o o
8 8
O O
888
OOO
S 8
o o
8 8
o o
8 K
o o
S.M « K M
O O O O
o o
o o
o o
o o
OOO
OOO
o o c c
O O O C
o o o o o o c
8 8
o o
8 8
o o
8 5
o o
8 8
O O
888
OOO
8 8
O O
8 8
o o
8 8
O O
8 8
o o
S 8
o o
888
OOO
8 8 8 § g
O O O O O
g i
8 8
o o
8 8
o o
8 8
O O
O O
o o
o o
888
000
8 8
o o
8 8
o o
8 S
o o
8 8
o o
8 S
o o
888
OOO
§0000
0000
o o o o o
8 8
O O
go
o
o o
8 8
O O
8 8
O O
888
OOO
8 8
O O
8 8
O O
8 o
O O
88
OO
8 8
O O
- o o
000
OOO
o o o o o o c
g §
o o
§ §
o o
§ §
o o
8 §
o o
§ § §
OOO
g g
o o
8 S
o o
O g
o o
8 g
O O
§ 5
o o
OOO
ooo
o o
o o o o o
8 8
o o
R 8
8 8
n 8
8 8
S S
88888
8 8
o o
8 8
o o
8 8
o o
8 8
O O
8 8
o o
888
OOO
r* n rt
n w w
1 i
179
-------
8 8 8
a e e
8888
o o o o
888
ooo
8 8
o o
8 8
O O
888
000
8 8
o o
88888
888
o o o
838
O O O
888
000
8888
o o o o
888
o o o
8 8
3 S
g 8
888
000
8 8
o o
888
000
g § §
o o o
888
O O O
8888
o o o o
858
o o o
0000000$
588
ooo
8 8
o o
888
000
o o o
888
O O O
8888
O O O O
808
o o o
8 o
o o
s s
o o
o o
e o
888
o o o
8 8
o o
888
000
8 5 8
o o o
CD
3
•i- 0)
4-> C
c o
5 •*
CD
-------
T3
O>
C I/)
•r- O)
+J -0
o _c
On j
U*1
"O
1
ej;
LJ O'.
S
| —
8
- O
g
: o
g
n
8
- 0
g
- O
g
O
8
O
g
O
§
g
•Q
0
8
n
O
8
O
„ °
O
§
rv
O
8
O
g
O
g
0
g
8
0
g
0
g
o
g
o
g
o
g
8
O
8
O
O
o
0
8
0
g
o
8
O
8
0
8
O
8
8
O
g
o
8
o
8
O
8
O
8
g
O
8
0
8
0
8
O
g
O
g
o
n
g
o
8
O
8
S
O
8
0
8
o
8
O
8
O
§
8
O
8
O
8
0
8
O
§
0
8
O
O
ft
8
o
8
O
8
8
0
g
O
8
o
g
o
g
o
8
8
O
8
O
8
o
8
O
g
O
8
0
8
O
8
O
8
g
O
g
O
8
o
g
o
8
O
8
8
O
8
0
8
0
8
O
8
0
8
O
8
0
8
O
8
8
0
8
0
S
o
8
O
8
O
8
g
O
8
O
8
O
8
O
§
O
8
O
IV
g
o
g
o
g
8
O
8
O
8
o
g
o
g
o
g
g
o
g
0
g
0
g
o
o
o
o
g
o
8
o
8
O
8
8
O
8
O
8
o
8
O
8
O
8
8
O
8
0
g
0
g
o
o
O
o
g
0
g
0
g
o
g
g
o
g
0
g
o
8
O
g
o
g
8
O
8
O
O
o
0
8
o
8
0
8
0
n
8
O
8
O
8
O
O
8
O
O
o
0
o
o
0
g
8
O
8
O
8
O
8
O
8
O
g
o
g
0
8
O
O
o
o
g
o
o
o
g
o
8
O
g
8
O
8
O
8
0
8
O
§
O
8
O
8
O
8
0
8
O
O
o
o
0
o
0
o
8
0
8
8
O
8
0
8
O
8
O
8
O
8
O
ra
8
O
8
O
8
8
O
O
o
o
o
0
o
8
O
8
8
O
S
0
8
O
8
O
8
0
8
O
8
o
g
o
g
g
o
s
o
0
0
8
0
8
0
8
8
0
8
0
8
O
8
O
8
O
8
0
9-
8
0
O
O
0
o
o
-
o
o
o
s
o
o
o
8
o
o
0
8
o
8
O
8
O
8
O
8
O
BD
8
O
O
o
n
o
§
o
s
o
o
n
n
o
0
o
n
o
n
o
o
o
o
o
o
o
8
O
8
O
8
O
8
0
8
Q
s
o
f)
o
0
0
0
0
s
0
0
8
o
8
0
8
0
8
o
8
0
8
O
8
0
O
o
o
o
n
o
o
n
O
O
0
o
0
o
n
o
o
n
o
8
O
8
O
8
O
8
O
g
O
8
O
8
o
o
0
8
0
0
0
0
o
8
0
8
0
0
0
8
8
o
8
0
8
0
8
O
8
0
8
D
8
O
8
o
0
8
O
O
O
8
o
8
O
o
o
g
8
o
g
o
8
O
S
o
o
o
o
8
O
g
o
8
O
8
Q
8
O
O
o
8
o
8
O
o
o
8
8
O
8
O
8
O
8
0
§
o
g
o
g
o
8
O
g
Q
8
O
O
o
8
0
g
o
8
o
8
8
o
8
O
8
O
8
O
§
O
8
O
8
O
8
O
8
8
O
8
O
8
o
8
O
8
O
8
8
O
8
O
8
O
8
O
§
O
8
O
8
O
8
O
8
g
O
g
o
g
0
g
0
8
0
8
8
O
8
0
8
O
8
O
o
o
0
8
O
8
0
8
O
8
8
O
g
o
g
o
g
o
g
o
g
g
o
g
0
8
O
8
O
o
o
0
8
O
8
O
8
O
8
8
O
8
0
8
o
8
O
8
O
8
8
O
8
O
8
O
8
O
8
0
8
O-
8
O
8
O
8
8
0
8
O
8
o
8
O
8
O
8
8
O
8
O
8
O
g
o
§
o
8
O
8
O
8
0
g
g
o
g
O
g
o
g
o
g
o
g
g
o
g
o
8
O
8
O
8
O
8
O
8
O
8
O
8
8
O
8
O
0
o
8
0
8
O
8
8
O
8
O
8
O
8
O
§
O
g
o
8
O
8
O
g
8
O
8
O
O
o
o
r
o
g
o
g
o
o
o
8
O
8
O
O
o
o
g
o
8
O
8
O
8
O
8
8
O
8
O
O
O
8
O
8
O
§
8
O
8
O
O
o
o
g
0
§
o
o
o
o
8
O
8
O
O
o
o
o
0
o
o
c
£
c
i
o
8
o
c
c
g
0
o
o
c
8
o
c
~
c
~
s
0
81 i
131
-------
TABLE A-7. SURFACE AND VEGETATION FACTORS BY LAND USE CATEGORY
Land Use Category
Woods-brushwood
Grass or crops
General urban
Residential
Commercial
Central business district
Industrial
Parks
Water
Surface Roughness
(m)
2.0
0.1
0.5
3.
1.
.0
.0
1.2
0.002
Vegetation Factors
2.0
1.0
0.5
1.0
0.2
0.2
0.2
1.0
4.0
Source: Sellers (1965).
The vegetation factors are dimensionless and represent the pollutant
uptake effectiveness of a particular type of land use (or vegetation
cover) relative to that of alfalfa.
81 10
182
-------
TECHNICAL HEPOHT DATA
(I'lrasc reuJ lmjr..ciiijns un tl ,• n-iu;c in. lure u"","'< Inn-)
1. REPORT NO.
2.
3. RtCIPltNT'S ACCESSION-NO.
[4.TITCE AND SUBTITLE
5. REPORT DATE
Application of the Empirical Kinetic Approach (EKMA)
to the Tulsa Area
|6. PERFORMING ORGANIZATION CODE
7.AUTHOPUSI
H. Hogo, G. Z. '.-.'bitten, and S. D. Reynolds
8. PERFORMING ORGANIZATION REPORT NO
49r-81-EF81-116
[9. PERFORV.'NG ORGANIZATION NAME AND ADDRESS
Systems Applications, Inc.
101 Lucas Valley Road
San Rafael, California 94903
10. PROGRAM Ei.fcf/ENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NA.VE AND ACDRESS
Air Management Technology Branch
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF RE PORT AND PERIOD COVERED
14. SPONSORING AC5NCY CODE
15. SUPPLEMENTARY NOTES
Gerald L. Sipson
The EKMA was evaluated using applications in the Tulsa area primarily by
comparing the trajectory model that forms the basis of the EKMA (OZIPP) with
other, more sophisticated, models. The study was carried out at several levels,
beginning with evaluation of OZIPP and ending with an evaluation of the control-
strategy predictions that result from employing the EKMA isopleth methodology.
The OZIPP trajectory model was compared with the SAI Airshed Model and the SAI
trajectory model, as well as with some modified versions of the original OZIPP
model.
The basic OZIPP model is a simple, moving-air parcel, or trajectory, model
that uses a detailed chemical mechanism for the surrogate propylene and butane
hydrocarbons. The model treats time-dependent, precursor-emission factors along
with expansion of the air parcel; entrainment is treated by assuming that
constant concentrations exist outside the parcel.
The primary emphasis of this comparison study was directed toward the
discovery of features in the basic OZIPP model that could explain differences in
the results of the OZIPP model from those of some other model.
7.
KEY WORDS AN'3 DOCUMENT ANALYSIS
DESCRIPTORS
b iDENTiFiERC/OPEN' ENOCO TERV.G
COSAT i 1 u'lj Group
Control Strategies
Photochemical Pollutants
•iodels
!EKi1A
OZIPP
•. DISTRIBUTION
10. SECURITY CLASS (Jlus
21. NO. O£ PAotS
190
20. SECURITY CLASS (Tins pj
22. PRICE
Form J220-1 U-73)
-------
|