United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-84-005
February 1984
Air
GUIDELINE FOR
USING THE
CARBON-BOND
MECHANISM IN
CITY-SPECIFIC
EKMA
-------
EPA-450/4-84-005
February 1984
GUIDELINE FOR USING THE
CARBON-BOND MECHANISM IN
CITY-SPECIFIC EKMA
Office Of Air Quality Planning And Standards
Office Of Air And Radiation
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
This report has been reviewed by the Office of Air
Quality Planning and Standards, U.S. Environmental
Protection Agency, and approved for publication.
Mention of trade names or commercial products is not
intended to constitute endorsement or recommendation
for use.
n
-------
TABLE OF CONTENTS
Page
List of Tables v
Li st of Fi gures vi
Acknowledgements vi i
1.0 Introduction 1
2.0 The CB-3 Mechani sm 5
2.1 Organic Reactivity 6
2.2 'Use of CB-3 In OZIPM 11
3.0 Procedures for Applying EKMA/CB-3 14
3.1 Selection of Modeling Cases 17
3.2 Development of Model Inputs 18
3.2.1 Model Inputs Without Significant Changes in
Recommended Procedures 20
3.2.2 CB-3 Mechanism 23
3.2.3 Organic Reactivity 24
3.2.4 Precursor Transport 32
3.3 Predicting Peak Ozone 44
3.3.1 Procedures for Making Ozone Predictions 44
3.3.2 Comparisons of Predictions With Observations .... 47
3.3.3 Review and Adjustment to Model Inputs 49
3.4 Computing VOC Emission Reductions 52
3.4.1 Derivation of Empirical Data 53
3.4.2 Generating Base Case Diagrams 54
3.4.3 Generating Future Case Diagrams 55
3.5 Selection of the VOC Emission Reduction Target 57
4.0 Using the CB-3 Mechanism With OZIPM 61
4.1 OZIPM/CB-3 Benchmark 64
4.2 Prediction of Peak Ozone 65
4.3 Generating a Base Case Diagram 69
4.4 Generation of Future Case Diagram 69
-------
TABLE OF CONTENTS (continued)
References R-l
Appendix A A-l
Appendix B B-l
Appendix C C-l
Appendix D D-l
Appendix E E-l
Appendix F F-l
Appendix G G-l
IV
-------
LIST OF TABLES
Number Page
2-1 Carbon Numbers for CB-3 Organic Species 10
3-1 OZIPM/CB-3 Model Inputs 19
3-2 Ranges of Urban NMOC Composition 30
3-3 Recommended Continental Background 37
3-4 EKMA/CB-3 Sensitivity to Precursor Concentrations Aloft . 39
3-5 Example Illustrating Effect of Model Predictions on
Selection of Control Target 59
4-1 Summary of OZIPM Codes for Model Input Data 62
4-2 Input Data For Benchmark Run 66
4-3 Example Inputs For Predicting Peak Ozone 67
4-4 Example Inputs For Generating A Base Case Isopleth
Diagram 70
4-5 Example Inputs For Generating A Future Case Isopleth
Di agram 72
A-l CB-3 Mechani sm A-2
B-l CB-3 Mechani sm Inputs For OZIPM 8-2
B-2 Explanation of OZIPM/CB-3 Inputs B-4
C-l Bond Groups Per Molecule (in alphabetical order) C-2
C-2 Exampl e Probl em - Part 1 C-7
C-3 Exampl e Probl em - Part 2 C-10
C-4 Exampl e Probl em - Part 3 C-12
-------
LIST OF FIGURES
Number Page
2-1 Example of Carbon Bond Lumping Procedure 8
3-1 Example Ozone Isopleth Diagram 15
-------
ACKNOWLEDGEMENTS
The principal author of this document is Gerald L. Gipson. The
contributions of a number of colleagues who contributed significantly to
the concepts expressed in this document are gratefully acknowledged. At
EPA, thanks are extended to Dr. Marcia Dodge, Dr. Basil Dimitriades and
Dr. Edwin Meyer for their technical dialogue and review. Special thanks
are extended to Dr. Harvey Jeffries and Mr. Ken Sexton of the University
of North Carolina for their willingness to participate in this project,
the significant time spent in discussions, and their valuable contributions.
This project would not have been possible without the continuing assistance
of Dr. Gary Whitten and Mr. Jim Kill us at Systems Applications, Incorporated.
The amount of time spent, their patience and their attention to detail,
as well as their valuable contributions, are all very much appreciated.
Finally, special thanks are extended to Mrs. Carole J. Mask for typing,
editing and helping assemble the document.
vi i
-------
1.0 INTRODUCTION
In March of 1981, the U.S. Environmental Protection Agency (EPA)
issued guidelines for applying the city specific Empirical Kinetics
Modeling Approach (EKMA).l EKMA is a procedure that can be used to
estimate reductions in emissions of Volatile Organic Compounds (VOCs)
that are needed to achieve the National Ambient Air Quality Standard
(NAAQS) for ozone. Application of city-specific EKMA according to the
March 1981 guidelines entails using the Ozone Isopleth Plotting Package
(OZIPP) to relate peak ozone concentrations to its precursors - Nonmethane
Organic Compounds (NMOC) and Oxides of Nitrogen (NOY).2>3 OZIPP is a
A
computer program that incorporates a simplified trajectory model and a
chemical kinetics mechanism (known as the DODGE mechanism) that mathe-
matically simulate ozone formation. Since the issuance of the March 1981
guidelines, the use of other chemical mechanisms with EKMA has been
suggested.4,5 in response, supplemental guidance on using other mechanisms
was circulated to EPA Regional Offices in December of 1981.6 The supple-
mental guidance contained generalized recommendations regarding the
application of other mechanisms, but did not provide specific details on
how any one particular mechanism might be incorporated in an EKMA modeling
analysis. The purpose of this document is to provide more specific
information regarding the use of one alternative mechanism - the Carbon-
Bond III mechanism (CB-3) developed by Systems Applications, Incorporated.7
The March 1981 guidelines dealt exclusively with using the DODGE
chemical mechanism and the EKMA technique. Those guidelines contained
-------
recommended approaches for formulating OZIPP input variables and applying
EKMA to estimate the VOC emission reduction needed to achieve the ozone
NAAQS. Many of these recommendations are appropriate for the CB-3
mechanism as well. In a few instances, however, the recommendations were
based on the results of sensitivity tests conducted with the DODGE
mechanism. Because the CB-3 and DODGE mechanisms do not always exhibit
the same sensitivity to a particular model input variable, some of the
March 1981 recommendations need to be modified when CB-3 is used with
EKMA. This document focuses primarily on those modifications that are
needed to use the CB-3 mechanism with EKMA in an appropriate manner, but
also summarizes the recommendations that remain unchanged.
The recommended approach for applying EKMA with the CB-3 mechanism
(hereafter referred to as EKMA/CB-3) parallels the one outlined in the
March 1981 guidelines for using EKMA/DODGE. As a consequence, the format
of this document is similar to that of the 1981 guidelines. Procedures
are recommended for transforming available emissions and aerometric data
into model input values. Alternative procedures are suggested for those
cases in which sufficient information is available to warrant their use.
Also, every effort has been made to provide so called "default" values
that can be used in the event of missing data. Because of the similarities
between an EKMA/CB-3 and EKMA/DODGE analysis, this document is intended
to serve as a companion to the March 1981 guidelines. The reader will
more than likely find it useful to refer to the 1981 guidelines for
discussions on the concepts underlying the modeling analysis or for
details on any particular portion of the modeling approach.
-------
As discussed in the March 1981 guidelines, the EKMA technique is
applied by using, the OZIPP computer program that internally incorporates
the DODGE mechanism. While OZIPP could be modified to replace the DODGE
mechanism with CB-3, extensive revisions to the computer code would be
required. However, an alternative program that accepts as input any
chemical mechanism is available. This program, Ozone Isopleth Plotting
With Optional Mechanisms (OZIPM) is very similar in structure to OZIPP,
but provides the flexibility needed for dealing with optional mechanisms.8
Since it is considerably easier to deal with OZIPM rather than modify
OZIPP, the discussions that follow will focus exclusively on using the
CB-3 mechanism with the OZIPM program.
The remainder of this document is divided into three chapters.
Chapter 2.0 contains a discussion of the CB-3 mechanism and its relationship
to the OZIPM program. Chapter 3.0 describes the modifications to the
March 1981 guidelines that are needed for an EKMA/CB-3 application.
Finally, Chapter 4.0 illustrates how the model inputs described in
Chapters 2 and 3 are actually used with OZIPM in order to apply the EKMA
technique.
One final point should be made concerning the evolution of the
recommendations contained herein. As noted previously, the CB-3 mechanism
was developed by SAI. Because of their unique expertise with CB-3 and
their familiarity with EKMA, EPA contracted with SAI to develop a
methodology for using CB-3 within the city-specific EKMA framework.9 The
recommended methodology was used by the EPA to formulate an initial set
-------
of guidelines which were discussed with representatives of EPA, SAI and
the University of North Carolina (UNC). Subsequent to these discussions,
the guidelines were revised, and then circulated to the same representa-
tives, and to other interested parties for further comments. The final
guidelines thus represent a synthesis of ideas expressed by a number of
interested groups.
-------
2.0 THE CB-3 MECHANISM
As the name implies, CB-3 is the third in a series of evolving
chemical kinetics mechanisms. Each of the successive carbon bond
mechanisms contains revisions that reflect increased knowledge of the
photochemistry leading to ozone formation. The CB-3 mechanism is currently
the most recent version of that generic series. It has been designed to
simulate laboratory smog chamber experiments using detailed data bases,
as well as atmospheric situations in which much less information is
typically available. While a comprehensive discussion on the scientific
basis of the CB-3 mechanism is beyond the scope of this document, some
introductory material on basic concepts is included below for those
unfamiliar with CB-3.
A distinguishing feature of any chemical mechanism is the manner in
which organic reactivity is treated. Because the construction and use of
a mechanism that includes all atmospheric species is virtually impossible,
individual organic species must be combined, or lumped, into some sort of
functional group, or groups. Thus, the discussion of any chemical mechanism
must necessarily address the manner in which organic chemistry is repre-
sented in the mechanism. The concepts underlying the treatment of organic
reactivity in CB-3 are discussed in Section 2.1 below.
As noted in Section 1.0, use of the CB-3 mechanism in a city-specific
EKMA analysis is most easily accomplished with the OZIPM computer program.
While the CB-3 mechanism has been designed in general form for use with
any photochemical model, some adjustments are normally required to "fit"
-------
the mechanism to any particular computer code, and OZIPM is no exception.
In addition, OZIPM requires a special input format for chemical mechanisms
that warrants some explanation. These items are addressed in Section 2.2
below.
2.1 ORGANIC REACTIVITY
As described in Section 2.0, a characteristic that typically
distinguishes chemical mechanisms is the manner in which organic compounds
are represented in the mechanism. A number of approaches have been
taken, but most have focused on lumping similar species into a single,
identifiable molecular species that represents the chemistry of that
particular class of compounds. For example, propylene might be used to
represent the chemistry of all alkenes. The CB-3 mechanism is somewhat
different in that the primary functional organic groups are based on
various types of structural units (e.g., single-bonded carbon atoms) as
opposed to molecular type (e.g., alkanes). As will be seen below, this
kind of structuring results in some organic species being represented by
more than one functional group.
In CB-3, seven functional groups are used to represent the role of
organic species, each based on various types of carbon bonds:
(1) single-bonded carbon atoms, termed paraffins and represented
by PAR;
(2) slowly reacting double bonds, almost exclusively ethylene
and represented by ETH:
-------
(3) relatively reactive double bonds, termed olefins and
represented by OLE;
(4) reactive aromatic compounds, termed aromatics and represented
by ARO;
(5) carbonyl compounds such as aldehydes and ketones, termed
carbonyls, and represented by CARB;*
(6) highly photolytic, a -dicarbonyl compounds such as methyl
glyoxal and biacetyl, termed dicarbonyls and represented by DCRB; and
(7) nonreactive compounds, represented by NR.
Of the seven groups listed above, users will be mostly concerned with
Groups 1-5 and 7. DCRB compounds are primarily formed as products of
chemical reaction. Hence, they are present, by and large, as a result of
reactions occurring amongst the first five functional groups.
Just as important as the definition of the functional groups themselves
is the manner in which individual organic species are apportioned to
those groups. As noted above, a particular organic compound is assigned
to a CB-3 group, or groups, on the basis of molecular structure. To
illustrate the procedure, consider the propylene molecule which contains
one single carbon-carbon bond and one double carbon-carbon bond (see
Figure 2-1). In the CB-3 mechanism, the propylene molecule is represented
by 1 paraffin and by 1 olefin. In essence, the molecule has been
*In addition to aldehydes and ketones, olefins with internal double
bonds are included in this group in order to eliminate intermediate
species that have short lifetimes.
-------
H
H H
C = C - C
I I
H H
Propyl ene
- H
H H
I I
C = C
I
H
H
I
C - H
I
H
1 OLE
1 PAR
Figure 2-1. Example of Carbon Bond Lumping Procedure
-------
apportioned on the basis of the carbon-carbon bonds: the double-bond
represented by OLE, and the one single bond by PAR. Similar classifica-
tions have been determined for scores of other compounds, and they provide
the basis for establishing the overall reactivity of an urban mix.
In the propyl ene example discussed above, note that the number of
carbon-atoms associated with PAR is one, while the number for OLE is two.
A general principle underlying use of the carbon-bond mechanism is that
the number of carbon atoms associated with any individual carbon-bond
group is fixed. (Table 2-1 shows these characteristic carbon numbers for
all six reactive carbon-bond functional groups.) By making use of the
carbon numbers, concentrations of each CB-3 group can be determined from
concentrations of individual organic species. To illustrate, consider
the propylene example discussed above, and further assume that the con-
centration of propylene is 3 ppmC. Since propylene is represented in
CB-3 by one PAR and by one OLE, the 3 ppmC total propylene concentration
must be apportioned to these two carbon-bond groups. Of the three carbon
atoms in a propylene molecule, one is PAR and two are OLE (see Figure 2-1).
Thus, one-third of the carbon atoms can be thought of as PAR, and two-thirds
as OLE. Since concentration is proportional to the number of carbon atoms,
the concentrations of PAR and OLE in the CB-3 mechanism would be 1 ppmC and
2 ppmC, respectively.* This same concept can be extended to multicomponent
mixtures as well. In such cases, concentrations of the individual organic
i .e., Cp/\R = 1/3x3 ppmC and COLE = 2/3 x 3 ppmC
-------
Table 2-1. Carbon Numbers for CB-3 Organic Species.
Carbon-Bond Group Number of Carbon Atoms
Paraffins (PAR) 1
Ethylene (ETH) 2
Olefins (OLE) 2
Aromatics (ARO) 6
Carbonyls (CARB) ' 1
Dicarbonyls (DCRB) 3
10
-------
species are first apportioned to their respective CB-3 group. The total
concentration of any particular CB-3 group is then obtained by summing
the contributions due to the individual organic species. This procedure
will be more fully discussed in Chapter 3.
In using the CB-3 mechanism with the OZIPM program, absolute
concentrations of the individual CB-3 groups are not directly input to
the model. Rather, the total NMOC concentration is specified, and the
fraction of carbon attributable to each CB-3 group is input. For example,
assume that the total NMOC concentration is 2.0 ppmC, of which 1.4 ppmC
is PAR (as determined by the procedure described in the preceding
paragraph). Then the apportioning factor, or carbon fraction, for PAR
would be 0.70, indicating that 70% of the total carbon is categorized as
PAR. The CB-3 carbon fractions correspond conceptually to the DODGE
mechanism apportioning factors of 25% propylene, 75% butane, and 5% added
as aldehydes. The CB-3 fractions can be determined on a city-specific
basis, or a special set of default values can be used. Both options will
be discussed in Section 3.5.
2.2 USE OF CB-3 IN OZIPM
The general form of the CB-3 mechanism that is recommended by SAI
for use with OZIPM is contained in Appendix A. The rationale and
background information leading to this recommendation are contained in
Reference 9. More extensive information on the evolution of the carbon
bond mechanism in general can be found in References 7, 11, 12, 13, and
14.
11
-------
As noted in Section 1.0, the OZIPM program is designed to accept any
alternative mechanism, provided it is coded in a specified format. In
addition to coding the chemical reactions and corresponding rate constants,
all photolytic reactions and primary organic functional groups must be
identified. Furthermore, those chemical species that undergo photolysis
must be given special, pre-defined names. All of this information is
input to OZIPM by means of the MECH option (see Reference 8 for details).
The specific inputs necessary to use CB-3 in OZIPM are contained in
Appendix B (details of the format are contained in Reference 8 and sum-
marized in Appendix B). It should be noted that to conform with OZIPM
input requirements, the names of two species in the CB-3 mechanism listed
in Appendix A must be changed: CARB to HCHO and DCRB to ALD2.* The
species NR represents the nonreactive portion of organic compounds. Note
»
that it is included as part of a "do-nothing" cycle, and does not affect
the other reactions nor the amount of ozone formed. Finally, two addi-
tional reactions have been added to reflect the effect of tropospheric
background on ozone formation (reactions 90 and 91 in Appendix B). The
basis for these reactions will be discussed in Section 3.2.5,
The discussions in Section 2.1 and above have provided a general
overview of the CB-3 mechanism and its relationship to the OZIPM program.
In most instances, consideration of the details of the mechanism will not
Even though the species labels HCHO and ALD2 are used in OZIPM, the CB-3
nomenclature of CARB and DCRB will be used throughout the text.
12
-------
be required in any particular model application. The major concern in
most applications is the determination of the total NMOC concentration,
and then the specification of the carbon-bond fractions required to
apportion the total carbon concentration to the individual carbon-bond
groups (i.e., PAR, ETH, OLE, ARO, CARB, DCRB, and NR). The procedures
for developing these and other model inputs for an EKMA application are
the subject of the next chapter.
13
-------
3.0 PROCEDURES FOR APPLYING EKMA/CB-3
Although the March 1981 guidelines deal explicitly with OZIPP and
the DODGE mechanism, many of the concepts described in that document are
relevant to the use of EKMA with other mechanisms as well. For example,
selecting the cases to model and the manner in which isopleth diagrams
are used to compute VOC emission reductions are unaffected by choice of
chemical mechanism. Nevertheless, use of CB-3 with OZIPM does require
some special considerations. This chapter will focus primarily upon
these circumstances, but will also describe, in general terms, all other
facets of conducting an EKMA modeling analysis. Again, the reader may
refer to the March 1981 guidelines document for details regarding some
aspects of the modeling methodology.
The ensuing discussion of using CB-3 with EKMA can perhaps be
facilitated by a brief overview of the general modeling procedure. The
OZIPM program is used to generate ozone isopleth diagrams that explicitly
relate peak hourly ozone concentrations to initial (i.e., 8 a.m.) ambient
levels of the ozone precursors NMOC and NOX (see Figure 3-1). The diagrams
are used with a measured peak ozone concentration and a city's NMOC/NOX
ratio to compute, on a percentage basis, the VOC emission reduction
needed to lower the observed peak to the level of the standard. While
isopleth diagrams are explicit functions of initial NMOC and NOX, the
positioning of the ozone isopleths on the diagram are also affected by
model input variables that are related to meteorology, emissions occurring
throughout the day, and pollutants transported from areas upwind of the
14
-------
0.2 0.4 0.8 0.8 1,0 1.2 1.4 1.8 1.8
0.2 0.4 0.8 0.8 1.0 1.2 1*4 1-6 1.8 2-0
KttHC.PPMC
Figure 3-1. Example Ozone Isopleth Diagram.
15
-------
city under review. Because these factors vary from day to day, the
highest VOC emission reduction estimate will not necessarily correspond
to the highest, observed ozone peak-l^ To account for this phenomenon,
the modeling approach recommended in the March 1981 guidelines consisted
of:
(1) modeling a number of high, observed ozone peak concentrations;
(2) computing the amount of VOC emission reduction needed to
lower each peak to the level of the standard;
(3) selecting a final VOC emission reduction target that is
consistent with the statistical form of the ozone standard.
Subsequent to the distribution of these recommendations, EPA issued
supplemental guidance further recommending that predictions of peak ozone
be compared to observed levels.6 If the agreement between predictions
and observations is found to be poor, review and possible adjustment to
key model inputs are suggested prior to computing VOC emission reductions.
While good agreement between predictions and observations does not com-
pletely insure accurate control estimates, successful prediction of
observed ozone peaks does provide some confidence that the chemical and
physical processes leading to ozone formation are being adequately simulated.
The modeling procedure described in the preceding paragraph can be
divided into five basic steps:
(1) selecting the observed ozone peaks to model;
(2) formulating the model inputs;
16
-------
(3) predicting peak ozone;
(4) computing VOC emission reductions; and
(5) selecting the overall VOC emission reduction target.
The remaining discussion is divided along these lines. For reference,
only items 2, 3 and 4 above contain information that is new or significantly
different from that found in the March 1981 guidelines. The other two
topics are included for completeness, even though no major modifications
have been-made to the recommended approaches.
3.1 SELECTION OF MODELING CASES
As noted in Section 3.0, the highest VOC control estimate may not
correspond to the highest observed ozone concentration. Further, the
statistical form of the ozone NAAQS permits on average, one daily maximum,
1-hour average ozone concentration above 0.12 ppm per calendar year
at each site.1^ Consideration of these two factors led to the recommendation
that a number of observed peaks above 0.12 ppm be modeled. The VOC
emission reduction target is then selected from these results in a manner
that is consistent with the statistical form of the ozone NAAQS. For an
EKMA/CB-3 analysis, the same procedure is recommended.
One additional issue could possibly arise when EKMA/CB-3 is used to
replicate a modeling analysis conducted with EKMA/DODGE. In this situation,
one may desire to use the original set of modeling results to reduce the
modeling candidates for EKMA/CB-3 to some smaller subset. Given the
complex nature of the nonlinear interactions embedded in the OZIPM model,
17
-------
the case with the highest YOC emission reduction may be different with
EKMA/CB-3 than with EKMA/DODGE. As a result, the arbitrary elimination
of any modeling case runs the risk of affecting the final VQC emission
reduction target. Therefore, using EKMA/DODGE results to screen modeling
cases from consideration with EKMA/CB-3 is not recommended.
Recommended Procedure: The recommended procedure for selecting
the cases to be modeled is identical to that delineated on page 10 of the
March 1981 guidelines document. Summarizing, the five (5) highest daily,
maximum ozone concentrations at each site should be selected as candidates
for modeling. Only ozone peaks that occur within or downwind of the
urban area under review should be included. The five highest values
should generally be chosen from the most recent three (3) years during
which measurements were made at a site.
3.2 DEVELOPMENT OF MODEL INPUTS
As just described, the five highest, daily maximum ozone peaks at
each site are selected for modeling. Two basic objectives of the modeling
analysis are to predict the observed ozone peaks, and to compute the VOC
emission reductions needed to lower each observed peak to the level of
the ozone NAAQS. To best accomplish these objectives, the model inputs
should be based on the atmospheric conditions associated with each observed
peak. Thus, their derivation ought to be done on a case-by-case basis.
In some instances, however, insufficient or inadequate data preclude such
a determination, and appropriate approximations or defaults are needed.
The major purpose of this section is to describe the methodologies
recommended for deriving the model input values under both circumstances.
Table 3-1 summarizes the model input variables that require
consideration, regardless of the intended purpose of the model simulation.
18
-------
Table 3-1. OZIPM/CB-3 Model Inputs.
Model Input Variables Section* New Recommendations
> CB-3 mechanism -- X
> Sunlight intensity 3.1.1
> Dilution 3.1.2
> Post-0800 emissions 3.1.5
> 03 transport 3.1.3
> Precursor transport 3.1.4 X
> Initial N02/NOX ratio 3.1.6
> Organic reactivity 3.1.6 X
* Refers to section numbers in the March 1981 guidelines (Reference 1).
19
-------
Procedures for estimating many of these inputs have been discussed in
the March 1981 guidelines, and the corresponding section numbers are
shown for reference. With EKMA/CB-3, however, some supplemental guide-
lines for developing model inputs need to be provided, and are so noted
in Table 3-1. While the discussion below will focus primarily on the
new procedures, those that remain unchanged are briefly reviewed for
completeness. Subsequent to that review, the recommendations for deriving
the other model inputs are discussed in more depth.
Before discussing each of the model input variables, one final point
should be added. The recommendations discussed below deal with model
inputs that correspond to conditions associated with the observed ozone
peak (i.e., so called base-case conditions). Some of these conditions
might be expected to change in future years subsequent to the implementa-
tion of VOC control programs. Factoring these potential changes into the
modeling analysis will be discussed in Section 3.4. Thus, the recom-
mendations discussed below concerning the derivation of model input
values will necessarily focus on data corresponding to emissions and
atmospheric conditions associated with a particular ozone peak observed in
the base case.
3.2 1 Model Inputs Without Significant Changes In Recommended
Procedure?
A number of model input variables are unaffected by the choice
of chemical mechanism and, as a consequence, many of the recommendations
contained in the March 1981 guidelines are also appropriate for CB-3. As
20
-------
might be expected, these variables primarily relate to the physical
processes affecting ozone formation (as opposed to chemistry related
variables). Examples include sunlight intensity, 03 transport and
dilution. While a detailed discussion of the procedures used to derive
these inputs will not be repeated here, a brief review is included for
background.
Light Intensity. The OZIPM program uses a city's latitude,
longitude and time zone, and the day of the year being modeled to generate
the appropriate diurnal pattern of photolytic reaction rates. While
updates have been made to some of the photolytic rates, these have been
incorporated in the CB-3 mechanism related inputs. Thus, no changes need
be made for this set of model inputs.
Dilution. In the OZIPM model, dilution occurs as a result
of the rise in atmospheric mixing height that typically occurs between
early morning and mid-afternoon. The mixing height can be viewed as the
top of a surface-based layer of air which is well-mixed due to mechanical
and thermal turbulence. Specific inputs to OZIPM include the early
morning mixing height, the maximum afternoon mixing height, the time that
the mixing height rise begins, and the time at which the maximum mixing
height is finally attained. In the March 1981 guidelines, procedures were
provided for estimating the early morning mixing height and maximum after-
noon mixing height from available radiosonde measurements. In the absence
of such measurements, appropriate defaults were listed. Further, the
OZIPM program will internally calculate the rate of rise in mixing height
21
-------
based upon a characteristic curve developed by Schere and Demerjian.1^,!?
This is the procedure that is generally recommended for EKMA/CB-3 modeling
analyses. However, OZIPM also contains an option whereby mixing heights
can be specified for hourly intervals of the simulation period. This
option can be used in place of the general recommendation whenever suf-
ficient information is available to make such estimates.
Post 0800-Eim'ssions. Post 0800-emissions refer to emissions
occurring along the trajectory subsequent to the start of the model
simulation. The actual model inputs are expressed as the fractions of
initial NMOC and NOX concentrations that should be added each hour to
represent the effect of fresh precursor emissions. The March 1981
guidelines delineated the computational procedures that can be used to
calculate emission fractions from the emissions data, the initial mixing
height, and initial precursor concentrations. These same procedures are
recommended for EKMA/CB-3 as well. (Note that the reactivity of the
organic emissions will be addressed in Section 3.2.3.)
Initial N02/NQX. The March 1981 guidelines recommended
a default value of 0.25. Alternatively, the initial N02/NOX ratio could
be derived from early-morning, urban core measurements of NO and N02-
No changes to these procedures are needed for application of EKMA/CB-3.
Ozone Transport. Ozone may be transported into a particular
city either (1) within the surface-based mixed layer, or (2) above the
early-morning mixed layer with downward mixing into the surface layer
22
-------
taking place as that layer increases in depth during the day. The March
1981 guidelines delineated procedures for deriving estimates of these
quantities based on either direct measurement, indirect measurement or
default estimate, and these are the procedures recommended for EKMA/CB-3
applications as well. However, some question has arisen as to the
appropriateness of the assumption imbedded within OZIPM (and OZIPP) that
the concentration of ozone aloft is constant throughout the simulation
period.4 Since little information is likely to be available to address
this issue on a city-by-city basis, and since the effects of ozone aloft
are likely to be most pronounced in mid-morning (i.e., the time at which
the mixing height rises most rapidly), the assumption of a constant level
aloft is still deemed most appropriate. However, if specific information
is available to support a different approach, then appropriate adjustments
could be made to the modeling procedure. It should be added that any
such adjustments would require modifying the OZIPM computer code.
Summary of Recommended Procedures. The procedures recommended
for formulating model input values for light intensity, dilution, post-0800
emissions, initial N02/NOX, and ozone transport are identical to those
delineated in the March 1981 guidelines. However, if sufficient city-
specific information is available, then alternative approaches may be
used to estimate a diurnal mixing height profile, or a diurnal profile in
the concentration of ozone aloft.
3.2.2 CB-3 Mechanism
An introduction to the CB-3 mechanism was presented in
Chapter 2, and the specific inputs that are needed to enter CB-3 into
OZIPM have been included in Appendix B. Consequently, no more detailed
discussion of the mechanism will be presented here. However, it should
23
-------
be added that the CB-3 mechanism recommended by SAI has undergone a wide
range of tests, some of which have led to the selection of specific
mechanism parameters. Further, the mechanism labeled as CB-3 in this
document differs in several details from the mechanism identified as "CB-3"
in References 7 and 24. In applying procedures described in this document,
the mechanism described in Appendix B is the recommended version.
Recommended Procedure. The form of the CB-3 mechanism
recommended for EKMA/CB-3 applications is contained in Appendix B. Pages
B-2 through B-3 show the mechanism in the format that is required for
input to OZIPM, with an accompanying explanation on page B-4.
3.2.3 Organic Reactivity
The fundamental concepts underlying the treatment of organic
reactivity in the CB-3 mechanism were described in Section 2.1. As noted
in that section, the organic reactivity input that is required by OZIPM
consists of specifying a set of apportioning factors, or as they are more
commonly termed, carbon-fractions. Specification of these fractions
permits the OZIPM progam to apportion total NMOC concentration into the
individual carbon groups - PAR, ETH, OLE, ARO, CARB, DCRB and NR.* (The
apportioning procedure is carried out within the model for the NMOC
concentrations that occur both initially and as a result of subsequent
post-0800 emissions.) For comparison, the apportioning factors for the
DODGE mechanism are 25%, 75% and 5% for propylene, n-butane and aldehydes,
respectively.
Although DCRB is an organic carbon-bond group, initial concentrations
and emissions of this group are very small compared to the others.
Thus, the carbon fraction input for this group is normally zero.
24
-------
Three basic approaches are possible for estimating the
carbon-fractions. The recommended approach consists of using a set of
default fractions that have been derived by SAI through analyses of
available ambient organic species data, emissions inventory data, and
review of pertinent, scientific experimental results.?>9 The second, an
alternative approach, requires the analysis by gas chromatography (GC) of
individual organic species' concentrations in ambient air within the city
under review. Typically, this approach requires a special field study.
Another alternative approach involves the analysis of VOC emissions
inventory data. In this case, the carbon-fractions are derived directly
from the emissions data, provided that information is available on the
quantity of species emitted by individual sources or source categories.
Each of the three approaches is discussed in more detail below.
The recommended approach of using a default value rather than
making a city-specific determination arises primarily as a consequence of
two factors. First, the default values are estimated by SAI to be
representative of typical urban reactivity based on ambient sampling
results conducted in a number of locales.7 While some city-to-city
variations in organic composition are to be expected, the default recom-
mendations should adequately represent most U.S. cities.7*9 The second
factor relates to the resource requirements associated with the two
alternative approaches. Since they require either a special ambient
sampling program or the compilation of emissions inventory data that is
more detailed than normal, the resources needed to carry out these programs
25
-------
can be significant. Thus, use of the default values is normally recom-
mended because of the expense associated with the more resource-intensive
approaches.
Instead of using default values, carbon-fractions can be computed
from GC analysis of ambient samples. References 18 and 19 discuss the
monitoring aspects of GC analysis, and will not be repeated here. However,
it should be noted that GC analysis is not an automated technique, and is
most often done on a special study basis. Thus, a monitoring program of
limited duration "is the most pragmatic approach for developing the informa-
tion needed to compute carbon fractions. While it is difficult to prescribe
exactly the number of samples needed, enough should be analyzed to ensure
that representative, average carbon-fractions can be computed. Since the
carbon-fractions will be used to apportion initial concentrations and
concentrations due to fresh emissions, the most appropriate sampling
period is one prior to the onset of significant photochemical reaction.
As with continuous total NMOC monitoring, sampling during the 6-9 a.m.
time period within the area of maximum emission density (i.e., usually
the center city) is generally recommended. Ambient samples for GC analysis
can be collected either by integration over a period of one hour or more,
or by grab sample in a few seconds. In general, the integrated method is
preferable since the possibility of measuring short term fluctuations in
species concentrations will be minimized.
The third basic approach involves the derivation of carbon-
fractions from emissions inventory data. The technique makes use of a
26
-------
set of "split factors" that distributes total VOC emissions from a
particular type of source, or source category, into individual organic
species, which can then be aggregated according to the appropriate carbon
bond groups.20 The split factors can be determined on a source-by-source
basis by source testing, or taken from literature sources such as
Reference 21. Obviously, if a substantial number of source tests are
undertaken, then this approach can be extremely resource intensive. In
practice, the apportionment of total VOC emissions to carbon bond groups
can be a rather intricate process, and should be conducted with the aid
of a photochemical modeling specialist familiar with the carbon-bond
mechanism. Publications that discuss the actual computational procedures
include References 7, 9, 20, 21, 22, and 23. Because of the complexities
involved, the details of the procedure will not be discussed in this
document.
The three recommended approaches described in the preceding
paragraphs lead to the derivation of a single set of carbon-fractions
which apply to both the initial concentrations and the concentrations
resulting from fresh emissions. Conceptually, the possibility exists for
a separate set of carbon-fractions to be developed for the initial
concentrations and the concentrations of the fresh emissions. Taking
this concept one step further, carbon-fractions could even be developed
for each hour of fresh emissions in order to account for potential spatial
and temporal variations in the emissions of different species. (Obviously,
the latter would require that an emissions inventory be of sufficient
27
-------
spatial and temporal resolution to identify such differences.) While
these concepts are intuitively appealing from a modeling point of view,
their inclusion in an EKMA/CB-3 modeling analysis is not recommended
because the added sophistication does not justify the extra expense. If
any of these last concepts are adopted, modifications to the OZIPM computer
code will be required, since OZIPM is presently structured to handle only
one set of carbon-fractions.
Regardless of the technique employed in their derivation, the
carbon-fractions are used to apportion t'otal concentrations of organic
compounds which are based upon ambient measurements. Of the two organic
compound monitoring techniques (i.e., continuous and GC), both rely on a
flame ionization detector that is relatively inefficient in responding to
many oxygenated compounds such as aldehydes and ketones (i.e., these
techniques measure hydrocarbons only). SAI has estimated that, initially,
total carbonyl compounds (i.e., those including aldehyde and ketones, as
well as some surrogate carbonyls) are about 5% of total nonmethane hydro-
carbon concentrations.^ Only about 1% of the total carbon that is measured
can be classified in the carbonyl group (i.e., surrogate carbonyls). The
remainder of the carbonyls (i.e., 4% of the nonmethane hydrocarbons that
are measured) is attributable to oxygenates that are not detected. Thus,
the carbon-fractions should normally sum to 1.04 (or 104%).* If ambient
*Note that this concept corresponds to the procedure used with the
DODGE mechanism where total nonmethane hydrocarbon (NMHC) is split
into 25% propylene and 75% butane, but 5% of the NMHC is added as
aldehydes.
28
-------
measurements of aldehydes are available, a city-specific determination of
the carbonyl fraction can be made. However, these measurements tend to
be complex, using techniques that have not been standardized since they
are mostly conducted by research groups. As a consequence, carrying out
a special aldehyde monitoring program cannot be routinely recommended.
Whenever city-specific estimates are made by the techniques
discussed above, special care must be taken to insure that the computed
carbon-fractions represent a realistic distribution of NMOC species. SAI
has developed ranges of probable carbon-fractions in order to ascertain
whether the possibility of computational or measurement error has been
introduced (Table 3-2). The use of any carbon distribution with an
outlying carbon fraction(s) is not recommended with the version of CB-3
discussed in this document. The CB-3 mechanism has been tested over
these ranges, and specific mechanism parameters have been chosen on this
basis. Use of values outside this range will require a reformulation of
the CB-3 mechanism. Thus, if a carbon-fraction falls outside of a prob-
able range, the derivation of that value should be reviewed to insure
that no errors have been made. In the event that no errors can be found,
and the modeler is sure that use of a distribution with outlying carbon-
fractions is warranted, the CB-3 mechanism should be modified. Obviously,
consultation with a photochemical modeling specialist thoroughly familiar
with the details of the CB-3 mechanism will be required to make any
changes to the mechanism.
29
-------
Table 3-2. Ranges of Urban NMOC Composition.*
Carbon-Bond Group** Carbon-Fraction
PAR 0.50-0.70
ETH 0.02-0.11
OLE 0.02-0.07
ARO 0.10-0.40
CARB 0.03-0.10
NR 0.05-0.22
* From Reference 9
** DCRB assumed to be negligible
30
-------
Recommended Procedure. The carbon-fractions recommended
for use in an EKMA/CB-3 analysis are listed below:
PAR = .58
ETH = .04
OLE = .03
ARO = .19
CARB = .05
DCRB = .00
NR = .15
They can normally be used unless sufficient information is available to
derive city-specific information by one of the methods discussed below.
Alternate Approach. If analyses of ambient air samples by
gas chromatography are available for a particular city, the results can be
used to derive carbon-fractions. The ambient samples should be taken in
the high emission density area (normally the urban core) within the 6-
9 a.m. Local Daylight Time (LOT) period during the ozone season. Inte-
grated samples are generally preferable to instantaneous grab samples.
It is desirable that enough samples be analyzed to provide a representative
average. For supplemental information regarding monitoring aspects, the
reader is referred to References 18 and 19, and for details on how carbon-
fractions are computed from the sampling results, the reader is referred
to Appendix C of this document.
Alternate Approach. If a detailed, speciated VOC
emissions inventory is available, then those data may be used to compute
carbon-fractions. For details of the procedures for compiling a speciated
inventory and apportioning the emissions into carbon bond classes, the
reader is referred to References 7, 9, 20, 21, 22, and 23. The aid of a
photochemical modeling specialist familiar with the CB-3 mechanism will
normally be required to conduct such an analysis.
Caveat: If either of the alternative approaches is used,
the resultant carbon-fractions should fall within the ranges shown in
Table 3-2. If they do not, it is strongly recommended that the data and
computations be thoroughly checked to insure that no errors have been
introduced. If the problem cannot be resolved, use of the default carbon
fractions listed in the recommended procedure above is preferable.
Because some mechanism rate constants were derived on the basis of the
ranges given in Table 3-2, use of carbon fractions that fall outside of
this range with the recommended form of the CB-3 mechanism may lead to
unrealistic results, and is not recommended.
31
-------
3.2.4 Precursor Transport
The guidelines for applying EKMA/DODGE contained a recommendation
that precursor transport (both in the surface layer and aloft) not normally
be considered. Sensitivity tests conducted with EKMA/DODGE revealed that
VOC control estimates are not substantially altered by typical levels of
precursor transport.1 Also, the measurements that are typically required
to characterize the levels of precursor transport can be expensive,
difficult to make and generally require special field studies. Given
these problems, the routine consideration of precursor transport in
EKMA/DODGE is difficult to justify. (Nevertheless, the 1981 March guide-
lines do provide for its consideration for the benefit of those who
desire to account explicitly for its effect.) Sensitivity tests conducted
with EKMA/CB-3, however, suggest that these same recommendations are
not always appropriate for EKMA/CB-3 analyses. Thus, some supplemental
guidance in this area is needed.
First, consider the role of NOX transport. Sensitivity tests
conducted with EKMA/DODGE have revealed that control estimates are not
critically affected by typical levels of NOX transport. The same general
finding holds true for EKMA/CB-3 as well. Thus, routine consideration of
NOX transport is not generally recommended, but the procedures outlined
in the March 1981 guidelines can be used to incorporate it in an EKMA/CB-3
analysis if so desired. As noted in the 1981 guidelines, measurements
that may not be routinely performed are required to estimate NOX transport
levels. The reader is referred to Appendix B of the 1981 guidelines
for additional details.
32
-------
As noted above, EKMA/DODGE exhibits relatively little sensitivity
to typical organic precursor levels that may be transported into a city
from upwind areas. Such is not always the case with EKMA/CB-3, however.
As a consequence, dealing with precursor transport in an EKMA/CB-3 analysis
warrants special consideration. For analysis purposes, it is convenient
to divide the precursor transport into two components: (1) naturally
occurring background, and (2) "manmade transport" that is generated by
significant, upwind source areas.9 Natural background represents a
ubiquitous component of ambient organic compounds that is irreducible
(i.e., the background organics will likely be unaffected by the implementa-
tion of VOC control programs). On the other hand, "manmade transport"
levels will depend on a city's location relative to other source areas
and the meteorological patterns that affect transport between source
areas. Consequently, one would expect the manmade transport levels to
vary from city-to-city, and that future levels might be reduced as a
result of the implementation of VOC control programs in upwind source
areas. Thus, the approaches for dealing with each are necessarily
different, and the discussion below will be divided along those lines.
Natural Background. SAI has investigated the role of
background organics vis-a-vis photochemical modeling with the CB-3
mechanism.9»24 jn their studies, SAI has subdivided naturally occurring
background into two components: (1) a tropospheric component that occurs
on a global scale, and (2) a continental component that is associated
with the surface-based mixed layer over the continental portion of the
33
-------
United States. For purposes of the ensuing discussion of these two types
of background, reactivity is defined as the rate at which hydroxyl radicals
react with atmospheric organic compounds to form peroxyl radicals, e.g.,
OH + HC > R02' (3-1)
The peroxyl radicals are responsible for converting NO to N02 in the
sequence of reactions that eventually lead to significant ozone formation.
High concentrations of slowly reacting species can have the same reactivity
as low concentrations of highly reactive compounds. True background
mixtures will usually react slowly, and be relatively stable in terms of
concentration. Species that react more rapidly are normally associated
with fresh emissions.
As noted above, tropospheric background is intended to represent
background reactivity on a global scale. It is based on measurements of
light hydrocarbons and CO taken over the eastern Pacific Ocean. The
reactivity of this mixture has been calculated, and photochemical mechanisms
used to estimate the equilibrium concentration of other compounds, such
as carbonyls.9 It should be added that some of the compounds that make
up the tropospheric background are normally considered unreactive because
of their relatively slow rate of reaction with hydroxyl radicals (examples
include methane and ethane). Nevertheless, the sum effect of their
presence should be accounted for in an EKMA/CB-3 analysis. However,
because the concentrations of the tropospheric background compounds do
not vary substantially, their effect can be accounted for in the CB-3
34
-------
mechanism directly. SAI has recommended that the tropospheric component
be included in the chemical mechanism by means of the following two
reactions:*
OH > ME02 k = 28 min-1 (3-2)
OH > H02 k = 88 mlrr* (3-3)
As can be seen in Appendix B of this document, these two reactions have
been included as reactions 90 and 91 in the recommended form of the CB-3
mechani sm.
In addition to the natural tropospheric background just discussed,
boundary layers over land masses receive a variety of organic emissions
that result in a continental background contribution over and above that
of the tropospheric component. These emissions occur as a result of
various biogenic, geogenic and anthropogenic processes. In general,
distinctions among these sources cannot be made from ambient measurements
alone because the composition of organics emitted from these sources
exhibit considerable overlap, and because the composition of some emissions,
especially those from biogenic sources, has not been well characterized.
It should be added that the measurements referred to here have been taken
in relatively "clean air" within the continental United States. Thus,
this continental background represents an irreducible component of ambient
organic compounds, although a portion may actually be the result of
anthropogenic activity.
* Source: Reference 9
35
-------
The continental background concentrations recommended by EPA
are, in turn, based on recommendations by SA! which consider measurements
taken in rural areas, and represent continental "clean air" conditions.
In their review of the data, SAI found that the concentrations of organic
species attributable to continental background gradually decreased with
height until the tropospheric background was approached at heights typically
above the early morning mixed layer. SAI recommends that the total mass
loading of the continental background be represented by a single concen-
tration that represents the average level within the surface-based mixed
layer. To account for variations in the height of this layer, SAI further
recommends that the average concentration should be inversely proportional
to mixing height. For EKMA/CB-3 applications, this recommendation can be
implemented by deriving the continental background concentration relative
to the initial mixing height. The estimated continental background
concentration should then be assumed to prevail throughout the surface
layer in the model. For an initial mixing height of 250 meters, SAI
recommends that the concentrations of carbon-bond species shown in Table 3-3
be added to the urban initial NMOC concentration.9 If the mixing height
is greater than 250 meters, then the total background NMOC concentration
should be reduced by a factor that is proportional to the ratio of 250
meters to the higher initial mixing height. For example, if the initial
mixing height were 500 meters, the total NMOC concentration of .038 ppmC
(Table 3-3) would be reduced by a factor of 1/2 (i.e., 250 meters = 1/2)-
500 meters
Thus, continental background can be viewed as the result of continuous
36
-------
Table 3-3. Recommended Continental Background.
Recommended Computed
Species Concentration, ppmC Carbon Fraction
PAR 0.020 .53
ETH 0.002 .05
OLE 0.001 .03
ARO 0.005 .13
CARB 0.010 .26
DCRB 0.000 .00
NR 0.000 .00
TOTAL 0.038* 1.00
* The 0.038 ppmC total is relative to a 250 meter initial mixing height.
For other initial mixing heights, the total concentration should be
adjusted by the following equation (with the same carbon fraction
used):
(250 meters)
CBKG = (o.oss ppmc) R^
where
= adjusted background concentration, ppmC
H0 = initial mixing height, meters
37
-------
areawide emissions producing ambient concentrations that are proportional
to the depth of the mixed layer.
Manmade Transport. The tropospheric and continental
backgrounds discussed above represent a ubiquitous and irreducible
component of ambient organic concentrations.* As noted above, the
possibility exists that organic precursors generated by upwind source
areas may be transported into a city. In the discussions that follow,
the manmade component is to be viewed as the contribution of organic
precursors over and above those occurring naturally.
In the context of the OZIPM model, manmade transport could
occur in the surface layer, aloft (i.e., above the early morning mixed
layer), or both. Of particular importance with EKMA/CB-3 is the concen-
tration of organics aloft. To illustrate its importance, a series of
sensitivity tests were conducted with EKMA/CB-3 in an effort to quantify
the potential effect, and the results are summarized in Table 3-4.
Predictions of peak ozone in the base case are not substantially affected
by the assumed levels aloft, but VOC control estimates are relatively
sensitive to the assumed levels.** Further, measurements taken aloft in
a number of cities encompass the levels used in these sensitivity tests.26*27
* Irreducible in this context means that the background organics will
likely be unaffected by the implementation of control programs.
** It should be added that this same sensitivity may not be exhibited
under all conditions. Nevertheless, the model inputs chosen for these
tests are similar to those used for many urban areas.
38
-------
Table 3-4. EKMA/CB-3 Sensitivity to Precursor Concentrations Aloft.
Precursor concentration aloft,* ppmC 0.000 0.010 0.020 0.030 0.040 0.050
Prediction of peak ozone, ppm 0.223 0.223 0.227 0.229 0.231 .232
VOC control estimate,t % 39 43 43 46 48 51
* See page 43 for the assumed composition of precursors aloft
t All estimates made assuming a 40% reduction in levels aloft in future years
39
-------
These measurements indicate that levels aloft vary from city to city.
A comprehensive review of NMOC measurements aloft is underway, and this
review may enable recommendations concerning default estimates to be made
in the future. The only alternative approach that can currently be
recommended is to measure organic precursors aloft on a city by city
basis. Unfortunately, such measurements are difficult to make, and very
resource intensive. Nevertheless, they appear to be necessary to
characterize typical levels aloft.
Appendix B of the March 1981 guidelines describes the
measurements that are needed to estimate concentrations of precursors
aloft being transported into an urban area. However, it is more than
likely that measurements will not be available for some of the days that
need to be modeled. In these cases, the March 1981 guidelines recommended
that median concentrations from all available measurements be used. This
procedure can be enhanced by associating particular measurements with
prevailing wind direction and possibly atmospheric conditions associated
with ozone episode conditions. For example, a particular city could be
located such that a heavily populated and/or industrialized area lay to
the south, but predominantly rural areas extend to the west. Early
morning organic measurements taken upwind or aloft would provide some
indication of the manmade transport. However, the manmade transport
contribution would be expected to be much greater for a windflow out of
the south rather than one from the west. Thus, the available measurements
could be grouped according to wind direction, in order to estimate manmade
40
-------
transport levels on other days with similar flow patterns.* The measure-
ments could be further stratified according to ozone episode conditions
(i.e., days with high ozone levels versus those with low levels). Again,
median values for each category could be used as estimates for those
cases in which day specific measurements are unavailable.
A second factor to consider with regard to these types of
measurements is the heights at which they are taken. The results of
studies conducted previously to characterize organics aloft suggest that
concentrations vary with altitude.27 Any measurement program that is
undertaken should attempt to characterize the concentrations of organics
within the layer that will be entrained into the afternoon mixed layer as
a result of the increase in mixing height later in the day.** For example,
if the mixed layer normally grows from 250 meters to a height of 1500
meters by the afternoon, then the measurements should be taken at varying
heights between 250 meters and 1500 meters. The OZIPM model inputs
should then be based on the average concentrations.found in the 250 meter
to 1500 meter layer. If measurements are available at only one height,
then these measurements alone will have to suffice.
"*"Ideally, wind measurements should be taken aloft at altitudes below
the maximum afternoon mixing height. If only surface wind measurements
are available, relationships between surface wind and wind aloft on
days with NMOC measurements aloft should be used in categorizing days
on which surface wind measurements alone are available.
** Note that the measurements should be taken in the early morning, before
the growth in the mixed layer begins, in order to determine the
concentrations that will be entrained later in the day.
41
-------
Given that measurements of organic compounds aloft are available,
some additional adjustments should be made in order to estimate the final
OZIPM model inputs. First, the tropospheric background embedded in the
CB-3 mechanism accounts (at least partially) for unreactive compounds
that may exist aloft, and for about 0.010 ppmC of all reactive compounds.
Thus, if concentrations of individual species are available (e.g., by GC
analysis), then only the reactive constituents need be considered when
estimating total concentrations of organics aloft. (Table C-l of
Appendix C identifies those compounds classified as unreactive.) Further,
the total concentration aloft that is derived from the measurements
should be reduced by 0.010 ppmC to avoid "double-counting" the reactive,
tropospheric background component. Finally, if the measurements aloft
are made by Flame lonization Detector (FID), oxygenated compounds (i.e.,
carbonyls) wiil not be detected. While precise estimates are difficult,
SAI's study of photochemical modeling results suggest that roughly 15% of
nonmethane hydrocarbons transported into an urban area can be classified
as carbonyls.9 in the absence of specific measurements of oxygenated
compounds, the 15% value is recommended as a default.
To illustrate the recommendations described above, assume that
nonmethane hydrocarbon measurements, including only reactive compounds,
totaled 0.040 ppmC. First, the 0.010 ppmC reactive tropospheric component
would be subtracted from this level, and then the resultant increased by
15% to account for unmeasured carbonyls (i.e., [0.040-0.010] x 1.15 =
0.035 ppmC). Thus, 0.035 ppmC would be the OZIPM estimate for organic
42
-------
compounds aloft. However, the composition of the levels aloft must also
be specified. If measurements of individual species are available, then
the average composition derived from the measurement concentrations can
be used. If such measurements are not available, then the following
default carbon-fractions are recommended:
PAR = 0.61
ETH = 0.06
OLE = 0.03
ARO = 0.15
CARB = 0.15
DCRB = 0.00
NR = 0.00
The default was calculated from the fractions used for the continental
background (Table 3-2), but adjusted to reflect only a 15% carbonyl
composition. This default composition is consistent with SAI's estimate
for manmade transport, although slightly lower in carbonyls and ethylene,
and slightly higher in aromatics and paraffins.9
Recommended Procedure. (1) To account for a global,
tropospheric background of organic compounds, two reactions have been
included in the recommended form of the chemical mechanism. No additional
input is required of the user. (2) A specified set of concentrations
have been recommended to account for a continental background (Table 3-3).
These backgrounds represent irreducible components of atmospheric organic
compounds, and are recommended for all EKMA/CB-3 analyses. The continental
background concentration should be considered in the surface layer of the
model only. (3) While measurements for manmade transport are complex and
resource-intensive, the sensitivity of EKMA/CB-3 control estimates to this
factor suggests its explicit consideration. Such measurements should be
taken above the mixed layer in the early morning at varying heights in
order to establish the average concentration in the layer that will be
entrained as the mixing height grows to its maximum. Since tropospheric
background accounts for some of the species concentrations measured aloft,
only the reactive components should be considered and the measurements
should be reduced by 0.010 ppmC to account for the reactive tropospheric
background component. If measurements of oxygenated compounds are not
available, then the resultant concentration should be increased by 15%
43
-------
(all CARB) to account for these compounds. The composition of the organics
aloft can be determined from the measurements by the procedures described
in Appendix C. Alternatively, the composition listed above can be used.
In the event NMOC data aloft do not exist for certain of the days to be
modeled, wind data should be examined so as to categorize each day with
measurements. If sufficient information exists, the days with measurements
should be further categorized by ozone episode conditions. Then, for
those days without measurements, the median NMOC level of a particular
category can be used as an estimate for a day that corresponds to that
category.
3.3 PREDICTING PEAK OZONE
In one study, it was found that EKMA could yield a lower control
estimate for a case when peak ozone is underpredicted as compared to one
in which peak ozone is more accurately predicted.4 In extreme cases of
underprediction, a solution may not be possible with EKMA. In a similar
fashion, a large overprediction could lead to a control estimate that is
higher than that obtained when good agreement is found. As part of the
supplemental EKMA guidance issued in-December of 1981, EPA addressed this
problem by recommending that predictions of peak ozone be made on a
routine basis, and appropriate adjustments or compensations be made if
poor agreement is found.6 In this section, the procedures for making the
predictions, comparing them with observations and making appropriate
adjustments are described. It should be added that the discussion below
is not peculiar to the CB-3 mechanism, but applies to use of any mechanism
with the EKMA technique.
3.3.1 Procedures for Making Ozone Predictions
In Section 3.2, most of the OZIPM model inputs that are needed
either to predict peak ozone or to estimate VOC emission reductions were
44
-------
discussed. In order to make predictions of peak ozone, one additional
set of model input variables is needed: the concentrations of NMOC and
NOX that are representative of the initial (i.e., 8 a.m.), urban core
levels. These model inputs are the most critical for making predictions,
and should be estimated on a case-by-case basis. Because of the model
sensitivity to these inputs, use of mean or median values compiled from
measurements taken across a number of days may lead to erroneous results.
Thus, day-specific measurements should normally be used to make these
estimates.
*
As for estimating the initial concentrations, the recommended
procedure is to make use of ambient NMOC and NOX measurements routinely
taken in the urban core. The initial concentrations are intended to
represent the NMOC and NOX that is initially present within the mixed
layer at the start of the model simulation (i.e., 8 a.m.). While several
approaches could be taken, the recommended method is to use the 6-9 a.m.
average concentration measured by collocated NMOC and NOX monitors within
the urban core. If more than one pair of measurements are available from
a set of such monitors, then the 6-9 a.m. average concentration at each
monitor should be averaged to obtain an overall, urban average NMOC and
NOX concentration. Algebraically, the above procedure can be expressed
as follows:
N
(CNMOC)O = 1=1
and
45
-------
where
N
£ C(CNOX)6-9]f
(cNOx)o = initial concentrations of NMOC
and NOX (in units of ppmC
and ppm, respectively) input
to OZIPM simulation
g^i. nCNox)6-93i = tne 6"9 a-m- average concen-
trations of NMOC and NOX (in
units of ppmC and ppm,
respectively) taken in the
urban core (or high emission
density area) at site i
N = total number of collocated
monitors for which day-specific
NMOC and NOX measurements are
available.
As noted above, the initial NMOC and NOX concentrations are
derived from day-specific measurements of NMOC and NOX. In-some instances,
an NMOC measurement may not be available for the day being modeled. In
such a case, the initial NMOC concentration can be approximated by making
use of the median NMOC/NOX ratio and a day-specific measurement of NOX
alone, provided it is available. The initial NMOC concentration for use
with the OZIPM simulation can be computed as the product of the median
NMOC/NOX ratio and initial NOX concentration, or
(CNMOC)O = (cNOx)0 (NMOC/NOX) (3-5)
where
= the 1'nitial NMOC concentration for the OZIPM
simulation, ppmC
= tne initial NOX concentration calculated by
equation 3-4b, ppmC
46
-------
(NMOC/NOX) = the median NMOC/NOX ratio as derived according
to the procedures outlined in Section 3.4
It should be emphasized that this approach is an approximation, and the
one described in the preceding paragraph is preferable.
With the estimates of initial NMOC and NOX, and the corresponding
day-specific inputs listed in Table 3-1, the CALCULATE option of OZIPM
may be used to perform a single model simulation. An example simulation
will be described in Section 4.2, and additional information is contained
in Reference 2. Thus, no additional discussion will be included here.
Recommended Procedure. The CALCULATE option of the OZIPM
program should be used to predict peak ozone for comparison with the
observed peak. The model inputs discussed in Section 3.2 should be used,
with initial concentrations that have been derived according to
equations 3-4a and 3-4b, using data that are specific to the day being
modeled. In the event that day-specific NMOC measurements are unavailable,
the initial NMOC concentration can be approximated by means of equation 3-5,
with the recognition that some uncertainty may be introduced in the
analysis. If day-specific measurements of NMOC and NOX are not available,
predictions of peak ozone cannot be made. In this case, computation of
VOC control estimates are recommended, but without the requirement of
reasonable agreement between prediction and observation.
3.3.2 Comparisons of Predictions With Observations
The principal output of concern obtained with a CALCULATE
simulation is the predicted ozone. By numerically integrating the
differential equations describing ozone formation processes (i.e., chemical
reaction, emissions, dilution, etc.), instantaneous concentrations of
ozone are computed throughout the simulation period. (For example output
see Section 4.2). From this computed profile of instantaneous ozone
concentrations, the OZIPM program calculates the maximum 1-hour average
47
-------
concentration occurring during the model simulation. It is this maximum
1-hour average concentration that is generally recommended for comparison
with the observed ozone peak. The model performance measure that is
generally recommended is the relative deviation of the prediction from
the observation, or
DEV = x 100 (3_6)
where
DEV = deviation of the model prediction from the observation,
percent
Cp = maximum 1-hour average predicted peak ozone, ppm
C0 = observed peak ozone, ppm
If the relative deviation is found to be within +_ 30%, then agreement
between the prediction and the observed peak is judged to be sufficient
to proceed with control estimate calculations. If the model underpredicts
by more than 30% (i.e., DEV < - 30%) or overpredicts by more than 30%
(i.e., DEV > + 30%), then review of, and possible adjustment to, key
model inputs according to the discussion of Section 3.3.3 below is
warranted. It should be noted that the observed ozone peak (not the
predicted) is recommended for subsequent control calculations.
Recommended Procedure. The relative deviation of the
model prediction from the observed peak should be computed according
to equation 3-6 above. The model predicted peak to be used in this
computation is the maximum 1-hour average ozone concentration calculated
by the OZIPM program. If the computed deviation is within + 30%, then
the model results are sufficiently accurate for control estimate
48
-------
calculations. If the deviation is outside the +_ 30% envelope, then the
procedures discussed in Section 3.3.3 should be applied in an attempt to
improve the simulation results.
3.3.3 Review and Adjustment to Model Inputs
If inadequate agreement between a model prediction and an
observed peak is found, review of the model inputs is recommended. The
objective of this review is to investigate whether some modifications
to key model inputs can be made in order to improve the model predictions.
This review should focus on those model inputs that most critically affect
predictions of peak ozone. Of most importance are the initial NMOC and NOX
concentrations, dilution and post-0800 emissions. Adjustment of these
inputs, within the uncertainties associated with their development, is
warranted if improvements in model predictions can be made. Obviously,
any errors that may have been made in their derivation should be corrected
as we!1.
While specific recommendations in trouble-shooting poor model
performance are difficult to make, some general guidelines can be made
depending on the nature of the problem, be it an underprediction or an
overprediction. Consider the case of underprediction first. Causes of
underprediction could result from initial concentrations being too low
or dilution being too great (i.e., the initial mixing height being too
low, the maximum afternoon mixing height being too great, or both).
These inputs should be reviewed to see if any errors have been made in
their estimation. If some uncertainty exists with regard to the data
from which they were derived, then the inputs can be adjusted within that
49
-------
range. For example, assume that an initial NMOC concentration had been
derived from 6-9 a.m. measurements taken at three monitoring sites.
However, the 6-9 a.m. concentration at one site (say site A) deviated
substantially from the concentrations measured at the other two sites
(say sites B and C). Then, improved agreement between model predicted
and observed ozone might be found if the initial NMOC concentration for
the day in question was derived solely from the measurements taken at the
two sites in agreement (i.e., sites A and B). Note that, if the initial
concentrations and/or the initial mixing heights are adjusted, corres-
ponding modifications should be made to the post-0800 emissions as well.
In addition, alterations to the initial mixing height would require
changes to the assumed continental background concentrations as discussed
on pages 36-37. Another possible cause for underprediction could lie in
the organic reactivity inputs that are used with the CB-3 mechanism. The
values that are being used should be checked to insure that they lie
within the recommended ranges listed in Table 3-2. However, day-specific
adjustments are not recommended.
Guidelines for correcting a problem of overprediction are
similar in concept to those for underprediction. For example, initial
concentrations could be too high, and/or dilution too low. Again, these
inputs might be adjusted within the range of reasonable uncertainty. As
an example, assume that mixing height data were computed using two sets
of radiosonde measurements and the averages used as model inputs. In a
case of overprediction, then, the largest of the two afternoon maximum
50
-------
mixing heights might be input rather than the average. In addition to
these types of adjustments, it is recommended that the ozone predicted at
the time of the observed peak also be reviewed.* A situation could exist
whereby the observed peak occurs relatively early in the simulation
period, and the model predicted maximum 1-hour average ozone occurs late
in the simulation. If the ozone level predicted at the time of the
observed peak agrees to within +_ 30% of the observed level, then that
result indicates adequate agreement for control calculation purposes, and
no further adjustments need be made.
As noted above, two key model inputs that substantially affect
model predictions of peak ozone include the initial NMOC and NOX
concentrations, and initial mixing height. The possibility exists that
mass balance techniques could be used to evaluate the appropriateness
of a particular set of initial concentrations and an initial mixing
height. For example, one could test by means of a simplified box model
whether or not a city's emissions are sufficient to generate the measured
initial concentrations within a mixed layer corresponding to the postulated
initial mixing height. While such an approach is intuitively appealing,
such calculations may not be able to account properly for ventilation,
The predicted, hourly average ozone concentration at the time of the
observed ozone peak can be approximated from the instantaneous predic-
tions by averaging the instantaneous predictions corresponding to the
hours bracketing the time of the observed peak. For example, if the
observed peak occurred between 1 and 2 p.m., then the instantaneous
predictions at 1 p.m. and 2 p.m. could be averaged. Alternatively,
more rigorous integration techniques could be used to compute the
integrated, hourly average.
51
-------
and for advection of pollutants from source areas nearby the precursor
monitors. Nevertheless, it does provide one means of assessing the
reasonableness of the postulated model inputs. Again, it should be added
that any adjustment to the initial concentrations and/or initial mixing
height will affect the post-0800 emission fractions, necessitating their
recomputation.
Recommended Procedure. Model inputs can be manipulated
within reasonable ranges to determine if improvement can be made in the
agreement between the model predicted peak ozone and the observed peak.
Specifically, the key model inputs are initial NMOC and NOX concentrations,
mixing heights and post-0800 emissions. In addition, CB-3 organic reactivity
inputs should be checked to insure that they conform to the recommended
ranges listed in Table 3-2, but day-specific adjustments are not recommended.
While the other model inputs can be reviewed and adjustments made where
appropriate, their relative importance vis-a-vis predicting peak ozone is
not deemed as critical as the aforementioned variables. Finally, model
inputs should only be adjusted within the range of reasonable uncertainty,
and not just selected such that good agreement between the model prediction
and observed peak is obtained. Finally, if acceptable agreement cannot be
found, control estimates should still be made and the procedures discussed
in Section 3.5 applied.
3.4 COMPUTING VOC EMISSION REDUCTIONS
The procedures for computing VOC emission reductions from ozone
isopleth diagrams have been described in Reference 1, among others. In
the first step, the OZIPM program is used to generate a base-case isopleth
diagram. This is normally accomplished by using the same model inputs
that are used to make predictions of peak ozone, except that the CALCULATE
option is replaced with the ISOPLETH option. A base, or starting, point
is then located on that diagram using two pieces of empirical data - a
city's prevailing NMOC/NOX ratio and the observed ozone peak for the case
being modeled. If changes in VOC and/or NOX emissions are the only
52
-------
changes projected, then the base-case diagram can be used by itself to
make the necessary computations. However, if changes in factors such as
ozone transport or precursor transport are projected to take place concur-
rently with changes in precursor emissions, then a second, future case
diagram must be generated. This section will focus on developing the
empirical data that are needed, generating the base case diagram, and
factors to consider in the generation of a future case diagram. For
details of the computational procedures that are involved in making VOC
emission reduction estimates, the reader is referred to Reference 1, the
March 1981 guidelines.
3.4.1 Derivation of Empirical Data
As noted above, two pieces of empirical data are used to
establish a starting point on the base case isopleth diagram. The first
is simply the ozone peak that was measured on the day being modeled. The
second is the NMOC/NOX ratio prevailing in the city under review. The
procedures that are recommended for estimating these values for an EKMA/
CB-3 application are identical to those delineated in the March 1981
guidelines.^ They are briefly summarized below.
Recommended Procedure. Details of the recommended procedures
for estimating the ozone peak and a city's prevailing NMOC/NOX ratio that
should be used to establish the starting point on a base-case isopleth
diagram are contained on pages 39 through 43 of the March 1981 guidelines.1
In summary, the ozone peak is the maximum 1-hour average level measured
at the site/day under review. The NMOC/NOX ratio is determined from 6-
9 a.m. measurements taken at collocated monitoring sites within the urban
or high emission density area. If measurements are taken at only one
monitoring site, then the ratio to be used in EKMA calculations should be
the median of the ratios found on all days that are being modeled for
which accompanying NMOC and NOX data are available. Use of a day-specific
53
-------
ratio is recommended only when data are available at more than one monitoring
site, and the individual ratios at each site do not differ by more than
+_ 30% from the average ratio.
3.4.2 Generating Base Case Diagrams
Generation of a base case diagram with OZIPM is carried out by
using the model inputs described in Section 3.3 and the ISOPLETH option.
The only additional inputs associated with the ISOPELTH option are those
controlling the NMOC and NOX scales of the diagram. As with EKMA/DODGE,
these values should be chosen such that the starting point is located
towards the right-hand portion of the isopleth diagram in order to
facilitate accurate VOC emission reduction calculations. This topic is
addressed in greater detail in the OZIPP User's Manual, Reference 2.
As was noted in Section 3.3.3, some situations may arise in
which predicted peak ozone values agree more closely with observed levels
if the predicted value corresponds to the time of the observed peak,
rather than to the maximum value occurring during the simulation. Such a
factor might suggest that, for control calculation purposes, the length
of the simulation should be shortened to correspond to the time needed to
reach the observed peak. In general, full 10-hour simulations are recom-
mended for these cases, but the option exists to shorten the simulation
using the OZIPM TIME option (see Reference 8). However, modeling analyses
suggest that ozone peaks are likely to occur later in the day as VOC
controls are implemented.25 AS a consequence, when a simulation length is
shortened for the generation of the base case diagram, a future case
diagram generated with a full 10-hour simulation length is recommended.
54
-------
This procedure will insure that the VOC control calculations will properly
account for the lengthening of the time to peak ozone associated with the
implementation of a VOC control program.
Recommended Procedure. Standard techniques for generating
base case diagrams are recommended. An example problem is discussed in
Section 4.3. One possible option that can be invoked is one of shortening
the simulation length to correspond to the time of the observed peak.
The approach would normally be taken only if the ozone predicted at the
time of the observed peak agreed much more closely with the observation
than the simulation maximum. In any event, a future case diagram using a
full 10-hour simulation period is recommended.
3.4.3 Generating Future Case Diagrams
As explained above, the generation of a future case diagram is
only necessary if changes in factors other than precursor emissions are
projected to take place. Such factors could include projected changes in
ozone transport, precursor transport, and/or a possible shift in organic
reactivity. (Note that only one future base diagram is needed to
incorporate all of these changes.) This section will focus exclusively
on how these changes are estimated. The procedures for generating the
diagrams are identical to those used for the base case diagram, with only
the relevant model inputs changed. For additional background material,
the reader is referred to pages 49-61 of the March 1981 guidelines document.
The March 1981 guidelines provided procedures for projecting
changes in ozone and precursor transport due to the implementation of YOC
control programs upwind of a city. Basically, these recommendations remain
intact. With regard to ozone transport, the diagram on page 54 of the
March 1981 guidelines document can be used to project the changes in ozone
55
-------
transport. Likewise, the guidelines document recommended that NMOC
precursor transport levels could be reduced by 40% if a city was impacted
by nonattainment areas upwind, 20% if otherwise. Again, these same
recommendations hold with the added caveat that only "manmade" NMOC
transported levels should be reduced (see Section 3.2.5). In addition,
no adequate procedure currently exists to project how the composition of
the manmade, transported organic compounds might change in the future.
As a consequence, the same composition of manmade and background organics
is recommended for both base and future cases.
The possibility exists that, as VOC emission controls are
implemented, a shift in a city's organic reactivity could take place.
This could occur if particular NMOC species are controlled to a greater
or lesser extent than others, and/or if the control program itself caused
a shift in the emissions of particular species. Conceptually, this
possiblity can be accommodated in the modeling analysis by generating a
future case diagram using the projected change in NMOC composition (i.e.,
a set of carbon-fractions representing the future case). Because of
the great uncertainties associated with making such projections, attempting
to account for reactivity shifts is not generally recommended. The only
means available for making these projections is through an analysis of a
detailed, speciated, VOC emissions inventory reflecting the imposition of
the VOC control program. Accounting for projected changes in organic
reactivity requires that (1) a speciated emissions inventory of acceptable
detail is available, and (2) some procedure for estimating how a particular
56
-------
control program may change the reactivity of any source or group of
sources. Because of the difficulties associated with each, such projec-
tions are likely to contain significant uncertainties, and thus are not
generally recommended.
As discussed in the previous section, one additional factor
could precipitate the generation of a future case diagram. If the
simulation length is shortened in the generation of the base case diagram,
then a future diagram using a full, 10-hour simulation period is recommended
in order to account for the possibility that the time to peak ozone may
increase when VOC controls are implemented. Again, this factor can be
incorporated simultaneously with the other potential changes that were
discussed above.
Recommended Procedure. Procedures for generating future
case diagrams are similar in concept to the recommendations contained
in the March 1981 guidelines document. For ozone transport, the recom-
mended procedures are identical. A similar situation exists for NMOC
transport, except that only the manmade levels can be reduced, with all
background levels held constant. Further, a constant composition of
background and manmade transport, from base to future case, is recommended.
A similar recommendation exists for the composition of the city's organic
composition. However, if an adequate data base exists (i.e., a detailed,
speciated emissions inventory), then a shift in organic reactivity can be
accommodated, provided sufficient information is available to do so.
Finally, if any base-case diagram was generated with a shortened simulation
period, then a future case diagram reflecting a simulation length of at
least 10 hours is recommended for VOC reduction calculations.
3.5 SELECTION OF THE VOC EMISSION REDUCTION TARGET
After all site/day combinations have been modeled, the final step of
the modeling analysis involves the selection of the overall VOC emission
reduction target. In essence, this procedure is dictated by the form of
57
-------
the ozone NAAQS, and is identical to the method recommended in the March
1981 guidelines document.1 In summary, a control target is selected for
each site that permits, on average, one hourly-average concentration
above 0.12 ppm per year. This corresponds to selecting the fourth highest
control level if three years of data are available, the third highest
control for two years of data, and the second highest control estimate
for only one year. The overall control target is then chosen as the
highest of the site specific control estimates to insure that the ozone
standard is attained at all sites.
The only additional factor that could affect the procedure just
described is the consideration of model predictions versus observations.
Recall from Section 3.3.3 that a VOC emission reduction estimate should
not be used when the model predicted peak ozone disagrees with the observed
peak by more than +30%. However, it has been observed that substantial
underpredictions of base case, peak ozone may lead to control estimates
which are too low.1* Conversely, significant overpredictions of base
case, peak ozone may yield control estimates which are too high. Under
some circumstances, this finding enables one to use control estimates for
those days in which base case, peak ozone is poorly predicted. To
illustrate, consider an example in which model predictions and control
estimates have been made for a site with three years of ozone data (see
Table 3-5). Note that for Day 1, peak ozone is substantially underpredicted,
and the control estimate is the highest of all days. If any improvements
were made to predicted peak ozone, the control estimate for this day
58
-------
Table 3-5. Example Illustrating Effect of Model Predictions on Selection
of Control Target.
Observed Predicted Relativet Control Rank of
Day
1
2
3
4
5
Ozone, ppm
0.27
0.22
0.20
0.18
0.15
Ozone, ppm
0.18
0.20
0.22
0.18
0.21
Deviation, %
-33
- 9
+10
0
+40
Estimate, %
55
47
51
45**
42
Control 1
1
3
2
4
5
predicted - observed
t Deviation = observedx
** Control Target = fourth highest control estimate (for 3 years of data)
59
-------
would likely be increased even more. Since the control estimate for Day 1
is already higher than the control target (i.e., 45%), any improvements
in model predictions would not affect the selection of the final control
target. Thus, the results from Day 1 can be used, even though the model
significantly underpredicted peak ozone. The converse situation occurs
for Day 5. In this case, any improvements in model predictions would
likely reduce the control estimate for that day, again having no bearing
on the choice of the final control target. If the model prediction is
poor, but neither of the situations described above occur (i.e., over-
prediction and high control estimate, or underprediction and low control
estimate), then it is recommended that the site/day be discarded, and
replaced by the day with the next lowest peak ozone concentration.
Recommended Procedure. To obtain the final VOC emission
reduction target, see the procedures described.on pages 11 through 16 of
the March 1981 guidelines. In general, a candidate control estimate is
chosen for each site based on the number of years of data and the statistical
form of the ozone standard (i.e., fourth highest control for three years,
third highest for two years, and the second highest for one year). Of
the candidate site-specific control estimates, the highest one is selected
as the overall VOC emission reduction target. However, all cases in
which predictions and observations disagree by more than 30%, should be
discarded, unless:
(1) peak ozone is underpredicted and the VOC reduction
estimate is greater than the candidate site-specific estimate;
(2) peak ozone is overpredicted and the VOC reduction
estimate is lower than the candidate site-specific estimate.
In the event that a day is eliminated, the next lowest peak at the site
in question should be added for modeling.
60
-------
4.0 USING THE CB-3 MECHANISM WITH OZIPM
The previous chapter focused on the derivation of model inputs and
the modeling procedure for an EKMA/CB-3 application. In this chapter,
attention is focused on how the model variables are input to OZIPM. As
was mentioned earlier, OZIPM is very similar in structure and operation
to the OZIPP model. While a detailed description of these computer
programs is beyond the scope of this document, a brief review should
facilitate further discussion. For more background information and
additional detail, the reader is referred to both the OZIPP and OZIPM
User's Manuals, References 2 and 8.
Functionally, the OZIPM program can operate in one of two ways: (1)
perform a single simulation in which peak ozone is predicted, and (2)
generate an ozone isopleth diagram. The two functions are activated by
means of an input record with the code letters CALC or ISOP, respectively,
placed in the first four character locations. Additional inputs can
follow in one of six 10-column fields: Field 1 includes columns 11-20,
Field 2 corresponds to columns 21-30, etc. Other model inputs are handled
in a similar manner. Four letter codes are used to identify particular
types of model input variables. (Table 4-1 lists several of the more
commonly used ones.) Some of these options also require that input data be
coded on input records immediately following the option record. In these
cases, the data also follow the 10-column field format, except that the
fields begin in column 1.
61
-------
Table 4-1. Summary of OZIPM Codes For Model Input Data.
CODE Type of Input Data
MECH Chemical mechanism
PLAC Light intensity
DILU Mixing heights
IRAN 03, NMOC and NOX transport
EMIS Post-0800 emissions
REAC Reactivity inputs
62
-------
All model input variables have been Initialized to default values.
As a consequence, if any inputs are to be changed, over-riding values
must be entered prior to either the CALC option or the I SOP option. As
for the numerical data, virtually all inputs are in floating-point format
(i.e., decimal rather than integer). In order to minimize the possibility
of coding errors, numerical data can always be entered with accompanying
decimal points. This convention will be followed throughout
this document.
The remaining portion of this chapter deals with the appropriate
structure of the model input data that should be used with OZIPM and the
CB-3 mechanism. Recall that the CB-3 mechanism is itself an input to OZIPM,
As a consequence, the first block of input records in any OZIPM/CB-3 run
should always be those input records listed in Appendix B on pages B-2
through B-3. In the discussions below, this block of data will always
be indicated by the single code word MECH. Note that this really implies
a total of 87 records.
The remainder of this chapter is primarily devoted to the coding and
structure of the other model inputs. First, a benchmark run will be
discussed so that a prospective user may insure that OZIPM is functioning
properly, and that CB-3 is correctly coded. The subsequent sections deal
with problems of predicting peak ozone, generating a base case isopleth
diagram, and generating a future case diagram.
63
-------
4.1 OZIPM/CB-3 BENCHMARK
When dealing with a model and chemical mechanism as complex as OZIPM
and CB-3, special emphasis should be placed on insuring that the program
is operating correctly, and the mechanism has been properly coded and
entered to the program. The easiest way to check is by comparison with a
benchmark. The input data for the benchmark has been kept as simple as
possible in order to avoid the possibility of introducing errors that
could be due to some of the other model inputs. Thus, before proceeding
to city-specific simulations, replication of the benchmark simulation
discussed here is recommended.
The input data required to generate the OZIPM output for comparison
with the benchmark are shown in Table 4-2. Note that the MECH option
actually indicates all of the CB-3 mechanism inputs (Appendix B) are to
be positioned in front of the remaining input records. The DILU option
is set so as to eliminate dilution from the simulation. The REAC option
is used for organic reactivity input data. The 7.0 (located anywhere
within columms 21-30) indicates that there are seven organic compounds
whose respective carbon-fractions will follow in the next input record.
The order must be the same as that used in the CB-3 mechanism inputs,
which in this case is PAR, ETH, OLE, ARO, CARB, DCRB, and NR (see
Appendix B). The carbon-fractions used in the benchmark correspond to
the default values discussed in Section 3.2.3. They are entered in
consecutive 10-column fields (i.e., 1-10, 11-20, 21-30 etc.). Following
the reactivity data is the input record instructing OZIPM to perform a
64
-------
single simulation with an initial NMOC concentration of 1.0 (within
columns 11-20), and an initial NOX concentration of 0.100 (within columns
21-30). The value of 1.0 (within columns 31-40) causes a detailed printout
of species concentrations, chemical reaction rates, photolytic rate
constants, etc. Finally, "blank record" following the CALC option indicates
that a blank record is always the last record of an OZIPM input stream.
The output that is generated by OZIPM using the inputs in Table 4-2
is shown in Appendix D. Prospective users should find relatively close
agreement between their output and that shown. However, some discrepancies
will likely occur as a result of differing computer systems. As a rule
of thumb, predicted ozone concentrations should agree fairly closely to
the third decimal place. It is also worthwhile to check to insure that
all rate constants, especially those for photolytic reactions, are correct.
Rate constants for all reactions are printed for each hour of the model
simulation (see Appendix D).
4.2 PREDICTION OF PEAK OZONE
Table 4-3 shows the input structure for a sample case in which a
single simulation is to be conducted for the purpose of predicting peak
ozone. Again, the MECH record is included to indicate the CB-3 input
records. Following these records is the record labeled PLAC, which
contains information on the city's latitude, longitude, time zone, and
the date of the day being modeled (i.e., June 24, 1980 in this example).
A free format input record follows, indicating the city's name. The next
record contains the initial and maximum afternoon mixing heights. The
65
-------
Table 4-2. Input Data For Benchmark Run.
1
2
3
<+
5
MECH
OILU
EEtC
.53
CALC
100.
7.0
.04
1.0
100.
.03
.100
.19
1.0
.05 .00 .15
I .100 1.0
BLANK RECORD
NOTE: (1) All code words begin in column 1.
(2) The code word MECH represents the 87 records contained in
Appendix B.
(3) Numerical entries are contained in 10-column fields (i.e.,
columns 1-10, 11-20, 21-30, etc.).
(4) "BLANK RECORD" indicates that the last record in an OZIPM
input data set should not contain any entries.
66
-------
Table 4-3. Example Inputs For Predicting Peak Ozone.
1
2
3
4
5
6
7
3
9
10
11
12
13
14
15
16
17
MECH
PLACE
CITY
DILU
EMIS
.02
. 35
.07
REAC
.53
TPAN
.033
.00
.040
.00
CALC
39.9
NAME
250.
-8.0
.02
.35
7.0
.04
.01
.53
.61
1.10
75.1
1235.
.17
.02
.35
.03
.07
.05
.06
.120
5.0
.17
.19
.19
-7.0
.03
.03
1930.
.17
.03
.05
-7.0
.13
.15
6.0
.10
.03
.00
0.0
.26
.15
24.
.02
.03
.15
0.0
.00
.00
BLANK RECCSD
NOTE: (1) All code words begin in column 1.
(2) The code word MECH represents the 87 records contained in
Appendix B.
(3) Numerical entries are contained in 10-column fields {i.e.,
columns 1-10, 11-20, 21-30, etc.).
(4) "BLANK RECORD" indicates that the last record in an OZIPM
input data set should not contain any entries.
67
-------
post-0800 emissions data are encoded on the next four records, beginning
with the code work EMIS. The reactivity inputs (i.e., REAC) are identical
to those discussed in the previous section. The TRAN option is used to
input information on transported (and background) pollutants. The first
two numerical values are the surface and aloft concentrations of ozone
(i.e., 0.01 and 0.07 ppm, respectively). The next two fields are for
surface and aloft concentrations of NMOC. The OZIPM convention for
entering these data is to code a "-7.0" within columns 31-40 and 41-50.
(The minus sign indicates concentration data are to follow, and the seven
refers to the number of organic species.) Then the actual input data are
coded in the next records, with the first entry giving the total NMOC
concentration, and subsequent entries containing the carbon fractions.
In this example, 0.038 ppmC is input on the 12th record to account for
background reactivity in the surface layer (Table 3-3). The following
record indicates that 0.040 ppmC is to be included as manmade transport
aloft (see Section 3.2.5). The two sets of carbon-fractions for the seven
CB-3 organic species that are shown in Table 3-3 and on page 43 are coded
following the total NMOC concentrations within columns 11-20, 21-30, etc.
(Note that one additional record is needed to input the carbon fraction
for the seventh species in each case, i.e., record numbers 13 and 15.)
The final two fields on the TRAN record refer to surface and aloft trans-
port of NOX, both of which are assumed zero in this example. Finally,
the CALC option directs OZIPM to perform a simulation with the initial
NMOC and NOX concentrations shown in the fields following the CALC code.
Again, "blank record" following the CALC option refers to the need for a
blank input record at the end of the input stream.
68
-------
The output generated using this input stream is contained in Appendix E.
Worthy of note is the fact that the predicted ozone concentrations are
printed for each hour of the simulation period. Also note that the maximum
1-hour average concentration occurring during the simulation period is
printed near the end of the output. It is this value that is normally
compared to the observed peak, as discussed in Section 3.3.
4.3 GENERATING A BASE CASE DIAGRAM
The example discussed in this section is a follow-on to the one just
described. Suppose that, using the same model data, the goal is to
generate an ozone isopleth diagram rather than make a single prediction
of peak ozone. Here, the diagram is presumed to represent existing, or
base case, conditions. This can be accomplished by using the same input
stream that was used in the previous example, but simply replacing the
CALC option with the ISOP option (see Table 4-4). Since no other informa-
tion is included on the ISOP input record, default values are used to
determine the NMOC and NOX scales on the resultant diagram. Appendix F
contains the OZIPM generated output.
4.4 GENERATION OF FUTURE CASE DIAGRAM
The final OZIPM application example is another follow-on to the
preceding two examples. The isopleth diagram described in the preceding
section represented base case conditions. Assume now that a future case
diagram is desired in order to reflect projected changes in ozone aloft
and manmade transport. In this example, ozone aloft is reduced from
69
-------
Table 4-4. Example Inputs For Generating A Base Case Isopleth Diagram,
1
2
3
4
5
6
7
8
9
10
11
12
13
!<+
15
IS
17
MECH
PLACE
CITY
DILU
EMIS
.02
.35
.07
REAC
.53
TRiM
.033
.00
.CiO
.00
ISCP
BLANK
39.9
NAME
250.
-3.0
.02
.35
7.0
.04
.01
.53
.61
RECORD
75.1
1235.
.17
.02
.35
.03
.07
.05
.06
5.0
.17
.19
.19
-7.0
.03
.03
1930.
.17
.03
.05
-7.0
.13
.15
6.0
.10
.03
.00
0.0
.26
.15
24.i
.02
.03
.15
0.0
.00
.00
NOTE: (1) All code words begin in column 1.
(2) The code word MECH represents the 87 records contained in
Appendix B.
(3) Numerical entries are contained in 10-column fields (i.e.,
columns 1-10, 11-20, 21-30, etc.).
(4) "BLANK RECORD" indicates that the last record in an OZIPM
input data set should not contain any entries.
70
-------
0.07 ppm to 0.06 ppm,* and manmade transport of organic compounds from
0.040 ppmC to 0.024 ppmC. As shown in Table 4-5, these are the only
inputs that are changed from those needed to generate the base case
diagrams. Appendix 6 contains the output, which differs only slightly
from that in Appendix F. The base and future case diagrams shown in
these appendices could then be used to compute the VOC emission reduction
that is needed to lower the ozone peak observed on this day to a level of
0.12 ppm.
Such an estimate is arrived at by using the graph on page 54 of the
March 1981 guidelines.
71
-------
Table 4-5. Example Inputs For Generating A Future Case Isopleth Diagram.
i.
2
3
4
5
6
7
3
Q
10
11
1 ?
13
14
15
16
17
MECH
PLACE
CITY NAME
DILL)
EMIS
.02
.35
.07
PEAC
.53
7R,\N
.033
.00
.024
.00
ISOP
39.9
250.
-3.0
.02
.35
7.0
.0^
.01
.53
.61
75.1
1235.
.17
.02
. 35
.03
.06
.05
.06
5.0
.17
.19
.19
-7.0
.03
.03
1930.
.17
.03
.05
-7.0
.13
.15
6.0
.10
.03
.CO
0.0
.26
.15
24.0
.02
.03
.15
0.0
.00
.00
BLANK RECORD
NOTE: (1) All code words begin in column 1.
(2) The code word MECH represents the 87 records contained in
Appendix B.
(3) Numerical entries are contained in 10-column fields (i.e.,
columns 1-10, 11-20, 21-30, etc.).
(4) "BLANK RECORD" indicates that the last record in an OZIPM
input data set should not contain any entries.
72
-------
REFERENCES
1. G. L. Gipson, W. P. Freas, R. K. Kelly, and E. I. Meyer, Guideline
for Use of City-specific EKMA in Preparing Ozone SIPs,
EPA-450/4-80-027, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina, March 1981.
2. G. Z. Whitten and H. Hogo, User's Manual for Kinetics Model and
Ozone Isopleth Plotting Package, EPA-bUU/8-/8-Ql4a, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, July 1978.
3. Ozone Isopleth Plotting Package (OZIPP), EPA-600/8-78-014b, U.S.
Environmental Protection Agency, Research Triangle Park, North
Carolina, July 1978.
4. H. E. Jeffries, K. G. Sexton and C. N. Salmi, Effects of Chemistry
and Meteorology on Ozone Control Calculations Using Simple Trajectory
Models and the EKMA Procedure, EPA-450/4-81-034, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, November
1981.
5. W. P. L. Carter, A. M. Winer and J. N. Pitts, Jr., "Effects of
Kinetic Mechanisms and Hydrocarbon Composition on Oxidant-Precursor
Relationships Predicted by the EKMA Isopleth Technique," Atmospheric
Environment, Volume 16, No. 1, January 1982.
6. Richard G. Rhoads, memorandum to Director, Air and Hazardous Materials
Division, Regions I-X, "Effects of Chemistry and Meteorology on
Ozone Control Calculations Using Simple Trajectory Models and the
EKMA Procedure," December 3, 1981.
7. J. P. Killus and G. Z. Whitten, A New Carbon-Bond Mechanism for Air
Quality Simulation Modeling, Final Report for Contract 68-02-3281,
Environmental Sciences Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, February 1982,
8. G. Z. Whitten and H. Hogo, User's Manual for Ozone Isopleth Plotting
With Optional Mechanisms (OZIPM). Draft Report for Contract 68-02-2428,
U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina, March 1978 (final report in preparation).
9. J. P. Killus and G. Z. Whitten, Technical Discussions Relating to
the Use of the Carbon-Bond Mechanism in OZIPM/EKMA, EPA-45Q/4-84-009,
U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina, November 1983 (in press).
R-l
-------
10. G. J. McRae, J. A. Leone and J. H. Seinfeld, Evaluation of Chemical
Reaction Mechanisms for Photochemical Smog, Part I:Mechanism
Descriptions and Documentation, Interim Report for Cooperative
Agreement 810184, Environmental Sciences Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, North
Carolina, September 1983.
11. G. Z. Whitten and H. Hogo, Mathematical Modeling of Simulated
Photochemical Smog. EPA-600/3-//-011, U.S. Environmental Prot
Agency, Research Triangle Park, North Carolina, January 1977.
12. G. Z. Whitten, H. Hogo and J. P. Killus, "The Carbon-Bond Mechanism:
A Condensed Kinetic Mechanism for Photochemical Smog," Environmental
Science and Technology, Volume 14, No. 6, June 1980.
13. G. Z. Whitten, J. P. Killus and H. Hogo, Modeling of Simulated
Photochemical Smog With Kinetic Mechanisms, EPA-bUU/j-su/UZSa, U.S.
Environmental Protection Agency, Research Triangle Park, North
Carolina, February 1980.
14. G. Z. Whitten, H. Hogo, M. J. Meldgin, J. P. Killus, and P. J. Bekowies,
Modeling ofSimulated Photochemical ^m99 ^th Kinetic Mechanisms,
EPA-600/3-79-Q01a, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina, January 1979.
15. Code of Federal Regulations, "National Primary and Secondary Ambient
Air Quality Standards," Title 40, Part 50.9.
16. Addendum 1 to the Use's Manual for the Kinetics Model and Ozone
Tsbpieth Plotting Package (OZIPP), U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina, Apirl 1981.
17. K. L. Schere and K. L. Demerjian, "A Photochemical Box Model for
Urban Air Quality Simulation." Proceedings, 4th Joint Conference on
Sensing of Environmental Pollutants, American Chemical Society,
November 1977.
18. H. Singh, Guidance for the Collection and Use of Ambient Hydrocarbon
Species Data in Development of Ozone Control Strategies,
EPA-450/4-80-008, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina, April 1980.
19. Guidance for Collection of Ambient Nonmethane Organic Compound (NMOC)
Data for Use in 1982 Ozone SIP Development, and Network Design and'
Siting Criteria for the NMOC and NOX Monitors. EPA-45Q/4-80-011,
U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina, June 1980.
R-2
-------
20. Procedures for the Preparation of Emission Inventories for Volatile
TTrganlc Compounds. Volume II: Emission Inventory Requirements for
Photochemical Air Quality Models. EPA-450/4-79-018, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, September
1979.
21. Volatile Organic Compound (VOC) Species Data Manual, EPA-450/4-80-015,
U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina, July 1980.
22. Guidance for Applying .the Airshed Model to Urban Areas,
EPA-450/4-80-OZO, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina, October 1980.
23. Emissions Inventories for Urban Airshed Model Application in the
PTnTadelphia AgcR, EPA-4bU/4-8?-UUb, U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina, April 1982.
24. G. Z. Whitten, J. P. Killus and R. G. Johnson, Modeling of Simulated
Photochemical Smog with Kinetic Mechanism, Final Report submitted in
fulfillment of EPA Contract 68-02-3281, Dr. Marcia C. Dodge, Project
Officer, Environmental Sciences Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, 1982.
25. G. L. Gipson, Comparison of Three Ozone Models: Urban Airshed, City-
specific EKMA and Proportional Rollback, EPA-45Q/4-82-OUZ, U.S.
Environmental Protection Agency, Research Triangle Park, North
Carolina, March 1982.
26. M. W. Chan, D. W. Allard and I. Tombach, Ozone and Precursor Transport
Into an Urban Area - Evaluation of Approaches, EPA-450/4-79-Q39,
U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina, December 1979.
27. W. C. Eaton, M. L. Saeger, W. D. Bach, J. E. Sickles, II, and
C. E. Decker, Study of the Nature of Ozone, Oxides of Nitrogen and
Nonmethane Hydrocarbons in Tulsa, Oklahoma -Volume III: Data
Analysis and Interpretation, EPA-450/4-79-008c, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, September
1979.
R-3
-------
APPENDIX A
CB-3 Mechanism Recommended for OZIPM
-------
Table A-l. CB-3 Mechanismt
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Reaction
N02 > NO + 0
0 + (02) + (M) - > 03
NO + 03 > N0£ + 02
N02 + 03 ---> N03 + 02
N02 + 0 > NO + 02
OH +'03 > H02 + 02
HQ2 + 03 -> OH + 202
OH + NO? -> HNOo
c. o
02
OH + CO > H02 + C02
NO + NO + (02) ---> N02 + N02
NO + N03 > N02 + N02
N02 + N03 + (H20) --- > 2HN03
NO + H02 > N02 + OH
H02 + H02 ---> H202 + 02
X + PAR >
02
OH + PAR > ME02 + H20
02
0 + OLE > ME02 + AC03 + X
0 + OLE > CARB + PAR
02
OH + OLE > RA02
03 + OLE > CARB *- CRIG
03 + OLE > CARB + MCRG + X
02
0 + ETH > ME02 + H02 + CO
Rate Constant
at 298 K
(ppm~l nrin-1)
1.0*
4.40 x 106**
26.6
0.048
1.3 x 104
100
2.40
1.60 x 104
440
1.50 x 10'4
2.80 x 104
26.0
1.20 x 104
1.50 x 104
105
1200
2700
2700
3.70 x 104
0.008
0.008
600
Activation
Energy
(°K)
0
0
1450
2450
0
1000
1525
0
0
0
0
-1.06 x 104
0
0
0
560
325
325
-540
1900
1900
800
A-2
-------
Table A-l. CB-3 Mechanismt (continued - 2)
Rate Constant Activation
at 298 K Energy
Reaction (ppnr1 min'1) (°K)
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38a.
38b.
38c.
38d.
39.
40.
41.
0 + ETH > CARB + PAR
02
OH + ETH > RB02
03 + ETH > CARB + CRI6
02
NO + AC03 > N02 + ME02
02
NO + RB02 -> N02 + CARB + H02 + CARB
02
NO + RA02 -> N02 + CARB + H02 + CARB
02
NO + ME02 > N02 + CARB + ME02 + X
02
NO + ME02 > N02 + CARB + H02
NO + ME02 > NRAT
03 + RB02 > CARB + CARB + H02 + 02
03 + RA02 > CARB + CARB + H02 + 02
OH + CARB > CR02 + X
02
OH + CARB > H02 + CO
02
OH + CARB > ACOs + X
CARB > CO + H2
CARB > QQ
202
QQ ___> H02 + H02 + CO
QQ > ME02 + ME02 + X + XCO
xco > x + co
N02 + AC03 > PAN
PAN > AC03 + N02
H02 + AC03 > stable products
600
A
1.20 x 104
0.0024
A
1.04 x 104
A
1.20 x 104
A
1.20 x 104
3700
7400
900
5.0
20
100
9000
8200
(0.24)*
(.564)*
8867
1133
104
7000
0.022
1.50 x 104
800
-382
2560
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1.35 x 104
0
A-3
-------
Table A-l. CB-3 Mechanismt (continued - 3)
Rate Constant Activation
at 298 K Energy
Reaction (ppnr1 min-1) (°K)
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
H02 + ME02 > stable products
NO + CRIG > N02 + CARB
N02 + CRIG -> NOs + CARB
CARB + CRIG > Ozonide
NO + MCRG > N02 + CARB + PAR
N02 + MCRG > NOs + CARB + PAR
CARB + MCRG - > Ozonide
CRIG > CO + H20
CRIG > stable products
02
CRIG > H02 + H02 + CO
MCRG > stable products
02
MCRG > ME02 + OH + CO
02
MCRG > ME02 + H02
02
MCRG > CARB + H02 + CO + H02
02
OH + ARO > RARO + H20
02
OH + ARO > HOo + OPEN
02
NO + RARO > N02 + PHEN + H02
02
OPEN + NO > N02 + DCRB + X + APRC
APRC > DCRB + CARB + CO + X
02
APRC > CARB + CARB + CO + CO
PHEN + N03 > PHO + HNOs
PHO + N09 > NPHN
9000
1.20 x 104
8000
2000
1.20 x 104
8000
2000
670**
240**
90**
150**
340**
425**
85**
6000
A
1.45 x 104
4000
6000
104**
104**
5000
4000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
600
400
0
0
0
0
0
0
A-4
-------
Table A-l. CB-3 Mechanismt (continued - 4)
Rate Constant Activation
at 298 K Energy
Reaction (ppm-1 min-1) (°K)
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
PHO
OPEN
OH +
DCRB
PHEN
CR02
OCRB
MONO
OH +
°3 -
OlD+
0*0
03 -
+ H02 ---> PHEN
+ 03 > DCRB + X + APRC
02
PHEN > HOo + APRC + PAR + CARS
02
---> H02 + AC03 + CO
+ OH ---> PHO
02
+ NO ---> N02 + CARB + AC03 + X
+ OH - > AC03 + CO
- > OH + NO
NO ---> HONO
-> 0*0
---> 0
+ (H2Q) ---> OH + OH
--> 0
5.00 x 104
40
3.00 x 104
(0.02 x KI)***
104
1.20 x 104
2.5 x 104
(3.1)*
9770
(0.53)*
4.44 x IQlO**
6.8 x 109**
1.0*
0
0
0
0
0
0
0
0
0
0
0
0
* Sunlight-dependent; rate constant is scaling factor for OZIPM input.
** Units of min-1.
*** Sunlight-dependent; rate constant is scaling factor to be multiplied
by rate constant for Reaction 1.
t Source: Reference 9.
A-5
-------
APPENDIX B
CB-3 INPUTS FOR OZIPM
This appendix contains a listing, and corresponding explanation, of
the inputs that are required to incorporate the CB-3 mechanism in the
OZIPM program. The inputs themselves are contained in Table B-l. For
reference, each input record has been numbered sequentially. Table B-2
contains a brief explanation of the model input values. For a more
comprehensive discussion of these inputs, the reader is referred to
Reference 8, the User's Manual for the OZIPM program.
-------
Table B-l. CB-3 Mechanism Inputs For OZIPM,
1
2
3
4
5
6
7
3
9
10
11
IZ
13
14
15
16
17
13
19
20
Zl
'yy
23
24
25
26
27
23
29
30
31
32
33
34
35
36
37
33
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
MECH
1.
67.
PAR
1.
NO 2
0
03
03
0
03
03
NO 2
-CO
NO
N03
NO 3
H02
K02
PAR
PAR
OLE
OLE
OLE
OLE
OLE
ETH
ETH
ETH
ETH
AC03
P.B02
RA02
ME02
KE02
ME02
RE02
RA02
OH
HCHO
HCHO
HCHO
HCHO
QQ
Q3
xco
AC03
PAN
AC03
ME02
CRIG
CRIG
CRIG
MCRG
MCRG
MCRG
99.
71.
ETH
2.
NO
N02
NO 2
CH
H02
CH
OH
NO
NO
N02
NO
H02
X
OH
0
0
CH
03
03
0
0
GH
03
NO
NO
NO
NO
NO
NO
03
03
HCHO
OH
OH
N02
H02
H02
NO
N02
HCHO
NO
N02
HCHO
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
33
81
82
83
39
40
41
42
43
44
45
46
47
48
7.0
99.
OLE
2.
NO
03
NO 2
N03
NO
H02
OH
H02
N02
NO 2
NO 2
ME02
ME02
HCHO
RA02
HCHO
HCHO
KE02
HCHO
RB02
HCHO
N02
N02
N02
N02
NO 2
HCHO
HCHO
CR02
H02
X
CO
QQ
H02
ME02
X
PAN
AC03
N02
NO 3
N02
N03
7.0
73. 38.
ARO HCHO
6. 1.
0
N02
NO 2
OH
AC03 X
PAR
CRIG
MCRG X
H02 CO
PAR
CRIG
ME02
H02 HCHO
H02 HCHO
ME02 HCHO
HCHO H02
HCHO H02
HCHO H02
X
CO
AC03
H02 CO
ME02 X
CO
NOE
HCHO
HCHO
HCHO PAR
HCHO PAR
1.0
37. 67.
ALD2 NR
3. 1.
l.OOE+0
4.40E+6
2.66E+1
4.8E-2
1.30E+4
l.OE+2
2.4E+0
1.6E+4
4.4E+2
1.50E-4
2.8E+4
2.6E+1
1.2E+4
1.5E+4
l.OE+5
1.2E+3
2.7E+3
2.7E+3
3.7E+4
8.0E-3
8.0E-3
6.CE+2
6.0E+2
1.2E+4
2.4E-3
1.04E+4
HCHO 1.2E+4
HCHO 1.2E+4
X 3.70E+3
7.40E+3
9.0E-f2
5.0E+0
2.0E+1
l.OE+2
9.0E+3
8.2E+3
2.4E-1
5.64E-1
8.867E+3
XCO 1.133E+3
l.OE+4
7.0E+3
z.zt-z
1.5E+4
9.0E+3
1.2E+4
8.QE+3
2.0E+3
1.2E+4
3.0E+3
2.0E+3
1450.0
2450.0
1000.
1525.0
-10600.
560.0
325.0
325.0
-540.0
1900.0
1900.0
300.0
800.0
-332.0
2560.0
13500.0
8-2
-------
Table B-l. CB-3 Mechanism Inputs For OZIPM. (continued)
57
53
59
60
61
62
63
64
65
66
67
63
69
70
71
72
73
74
75
76
77
73
79
80
SI
82
33
34
25
8f>
37
CRIG
CRIG
CRIG
MCRG
MC=G
MCRG
MCRG
ARO
ARO
RARO
OPEN
AFRC
AFRC
PHEN
PHO
PHO
OPEN
OH
ALD2
PHEN
CR02
AL02
HCNO
CH
03
01D
010
OH
OH
NR
03
OH
OH
NO
NO
H03
N02
H02
03
PHEN
OH
NO
OH
NO
49
50
51
52
53
54
55
56
57
58
59
bO
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
90
91
92
99
CO
H02
ME02
ME02
HCHO
RARO
H02
N02
N02
ALD2
HCHO
PHO
PHEN
AL02
H02
H02
PHO
N02
AC03
OH
HCNO
DID
0
OH
H02
ME02
NR
0
H02
CH
H02
H02
OPEN
FHEN
ALD2
HCHO
HCHO
HN03
X
APRC
AC03
HCHO
CO
NO
OH
CO
CO
H02 CO
H02
X APRC
CO X
CO CO
APRC
PAR HCHO
CO
AC03 X
6.7E+2
2.4E+2
9.0E+1
1.5E+2
3.4E+2
4.25E+2
8.5E+1
6.0E+3 600.0
1.45E+4 400.0
4.0E+3
6.CE+3
l.OE+4
l.OE+4
5.0E+3
4.0E+3
5.0E+4
4.0E+1
3.CE+4
2.0E-2
l.OE+4
1.2E+4
Z.5E+4
3.1EO
9.77E+3
5.4E-1
4.44E+10
6.8E+9
8.8E+1
2.6E+1
l.OE+0
l.OE+0
B-3
-------
Table B-2. Explanation of OZIPM/CB-3 Inputs.
Rec # Columns Value Description
1
1
1
1
1
2
2
2
2
3
4
4
4
4
5
5
5
5
6-87*
1- 4
11-13
21-23
31-33
61-63
1- 2
11-13
61-63
1- 3
1- 3
11-13
51-52
1- 2
11-12
51-52
1- 4
7-10
13-16
17-18
25-28
33-36
41-44
49-52
55-64
66-72
MECH
86.
7.0
7.0
1.0
1.
71.
67.
67.
PAR
ETH
NR
1.
2.
'
1.
_
_
_
_
_
-
-
-
-
-
Signifies mechanism inputs to follow
Identification number of last chemical reacti
Number of photolytic reactions
Number of organic species
Number of photolytic reactions with rates
proportional to the N02 photolysis rate (kj)
Identification number of photolytic reaction
Identification number of photolytic reaction
Identification number of photolytic reaction
Identification number of photolytic reaction
rate proportional to N02 photolysis rate (k-j)
Name of first organic species
Name of second organic species
Name of seventh organic species
Carbon number of first organic species
Carbon number of second organic species
Carbon number of seventh organic species
Reactant 1
Reactant 2
Reactant 3
Identification number of reaction
Product 1
Product 2
Product 3
Product 4
Reaction rate
Activation energy
on
1
2
7
with
* Record numbers 5 through 87 contain the CB-3 chemical reactions.
Thus, a general description of the input variables is provided.
B-4
-------
APPENDIX C
Computation of Carbon-Bond Fractions From GC Data
In this appendix, the computation of carbon-fractions from results
of gas chromotagraphic (GC) analysis is discussed. As noted earlier, GC
analysis actually measures the concentrations of individual organic
species, which must then be grouped according to the CB-3 organic
reactivity classes. In order to keep the computations relatively simple
for illustrative purposes, hypothetical examples are discussed. For more
detailed discussion, including consideration of "real" data, the reader
is referred to References 7 and 9.
At the heart of the computational procedure is the definition of how
individual species should be categorized according to carbon bond type.
SAI has provided such definitions for approximately 200 individual species,
and these are reproduced in Table C-l. These definitions, or species
profiles, give the number of bond types found in each CB-3 category.
Using this information, along with the carbon numbers shown in Table 3-1,
it is possible to compute concentrations of individual carbon bond classes,
and then determine percentage of carbon in each class. These computations
will be illustrated below by the examples in Tables C-2 and C-3, respectively.
Table C-2 presents example calculations for a hypothetical example.
The individual species that might be detected by GC analysis are shown in
column 1, and their associated concentrations, in units of ppbC, are shown
in column 2. The remaining columns are associated with the carbon bond
computations.
C-l
-------
Table C-l. Bond Groups Per Molecule (in alphabetical order).
Species Profiles by Bond Group
Saroad
No.
43814
43820
43813
45225
45208
45207
43218
46201
43245
43224
43312
43296
43276
43299
43291
43280
43279
43234
50001
43274
43277
43271
43278
43308
43311
43452
50002
43310
43229
43225
43228
50004
43275
43230
43223
43211
43270
43298
43295
43293
43297
45221
50025
43503
43404
Chemical Name
1,1, 1-TR ICHLOROETHANE
1,1, 2-TRICHLQROETHANE
1,1-DICHLOROETHANE
1,2, 3-TR IMETHYLBENZENE
1 , 2 , 4-TR IMETHYLBENZENE
1 , 3 , 5-TRIMETHYLBENZENE
1,3-BUTADIENE
1,4-DIOXANE
1-HEXENE
1-PENTENE
l-T-2-C-4-TM-CYCLOPENTftNE
2 , 2 , 3-TR IMETHYLPENTANE
2 , 2 , 4-TR IMETHYLPENTANE
2,2, 5-TR IMETHYLPENTANE
2 , 2-DIMETHYLBCTANE
2,3, 3-TR IMETHYLPENTANE
2,3, 4-TR IMETHYLPENTANE
2 , 3-DIMETHYL-l-BUTENE
2,3-DIMETHYLBUTANE
2 , 3-DIMETHYLPENTANE
2 , 4-DIMETHYLHEXANE
2 , 4-DIMETHYLPENTANE
2 , 5-DIMETHYLHEXANE
2-BUTYLETHANOL
2-ETHCKYETflANOL
2-ETHQXYETHYL ACETATE
2-ETHYL-l-BDTENE
2-METHOXYETHANOL
2-METHYL PENTANE
2-METHYL-l-BDTENE
2-METHYL-2-BDTENE
2-METHYL-2-PENTENE
2-METHYLHEXANE
3-METHYL PENTANE
3-METHYL-l-BUTENE
3-METHYL-l-PENTENE
3-METHYL-T-2-PENTENE
3-METHYLHEPTANE
3-METHYLHEXANE
4-METHYL-T-2-PENTENE
4-METHYLHEPTANE
A-METHYLSTYRENE
A-PINENE
ACETALDEHYDE
ACETIC ACID
OLE
.
-
-
-
-
-
2
1
1
1
-
-
-
-
-
-
-
1
-
-
-
-
-
-
-
1
-
-
1
-
-
-
-
1
1
-
-
-
-
-
-
1
-
-
PAR
_
-
-
3
3
3
-
2
4
3
8
8
8
8
6
8
8
4
6
7
8
7
8
5
3
4
4
2
6
3
3
4
7
6
3
4
4
8
7
4
8
2
8
1
1
ARC
.
-
-
1
1
1
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
1
-
-
CARB
_
-
-
-
-
-
-
2
-
-
-
-
-
-
-
-
-
-
-
-
1
1
2
-
1
-
-
2
2
-
-
-
2
-
-
2
-
1
-
1
-
ETH ONREACT
2
2
2
-
- -
_ _
- -
_ _
- -
-
-
-
- -
- -
- -
_ _
- -
- -
- -
- -
- _
- -
-
- -
- -
- -
- -
- -
- -
_ _
- -
- -
- -
_ _
- -
_ _
- -
- -
1
C-2
-------
Table C-l. Bond Groups Per Molecule (in alphabetical order), (continued - 2)
Species Profiles by Bond Group
Saroad
No.
43551
43702
43206
43704
50015
50020
50026
45201
50024
43213
43510
50003
43115
43116
43117
43511
43512
43289
43294
43513
43290
43807
43804
43443
43803
43217
43227
50019
43248
43264
43273
43242
43292
43207
50027
43320
43823
43802
50018
43450
50017
45103
50012
43287
43285
Chemical Name
ACETONE
ACETONITRILE
ACETYLENE
ACRYLONITRILE
ANTHRACENE
B-METHYLSTYRENE
B-PINENE
BENZENE
BENZYLCHLQRIDE
BUTENE
BOTYRALDEHYDE
C-3-HEXENE
C-7 CYCLOPARAFFINS
C-8 CYCLOPARAFFINS
C-9 CYCLOPARAFFINS
C3 ALDEHYDE
C5 ALDEHYDE
C6 OLEFINS
C7-OLEFINS
C8 ALDEHYDE
C8 OLEFINS
CARBON TETRABROMIDE
CARBON TETRACHLORIDE
CELIOSOLVE ACETATE
CHLOROFORM
CIS-2-BUTENE
CIS-2-PENTENE
CRYOFLOURANE (FREON 114)
CYCLOHEXANE
CYCLOHEXANONE
CYCLOHEXENE
CYCLOPENTANE
CYCLOPENTENE
CYCLOPROPANE
D-LIMONENE
DIACETONE ALCOHOL
DICHLORODIFLDOROMET9ANE
DICHIOROMETflANE
DIMETHYL ET3ER
DIMETHYL FORMAMIDE
DIMETHYL-2 , 3 , DIHYDRO-1H-INDENE
DIMETHYLETHYLBENZENE
DIMETHYLNAPHTHALENE
DOCOSANE
EICOSANE
OLE
«.
-
-
1
-
1
1
-
-
1
-
1
-
-
-
-
1
1
-
1
-
-
-
-
-
-
-
-
-
1
-
1
-
1
-
-
-
-
-
-
-
-
PAR
3
1
1
1
8
-
8
1
1
2
3
4
7
8
9
2
4
4
5
7
6
-
-
4
-
2
3
-
6
5
4
5
3
3
6
5
-
-
2
5
4
6
22
20
ARO
_
-
1
1
-
-
1
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
1
1
1
-
-
CARE
_
1
-
-
-
-
1
-
-
-
1
1
1
-
-
-
2
-
2
2
-
-
1
-
-
-
-
2
1
-
-
-
-
-
-
-
-
-
ETH
|
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
DNREACT
^
1
1
-
-
-
-
5
-
-
-
-
-
-
-
-
-
-
-
-
1
1
-
1
-
-
2
-
-
-
-
-
-
-
-
1
1
-
3
-
-
C-3
-------
Table C-l. Bond Groups Per Molecule (in alphabetical order), (continued - 3)
Species Profiles by Bond Group
Saroad
No.
43202
43433
43438
43302
43812
43351
43219
43721
45203
43288
43203
43815
43370
43601
50011
43502
43368
43367
43286
43282
43232
50005
43281
43231
43214
43306
43304
43446
43451
43215
43120
45105
43109
45106
43112
45104
43106
43105
45234
43108
43107
43114
43122
43121
45108
Chemical Name
ETHANE
ETHYL ACETATE
ETHYL ACRYLATE
ETHYL ALCOHOL
ETHYL CHLORIDE
ETHYL ETHER
ETHYLACETYLENE
ETHYLAMINE
ETflYLBENZENE
ETHYLCYCLOHEXANE
ETHYLENE
ETHYLENE DICHLORIDE
ETHYLENE GLYCOL
ETHYLENE OXIDE
ETHYLNAPHTHALENE
FORMALDEHYDE
GLYCOL
GLYCOL ETHER
HENEICOSANE
HEPTADECANE
HEPTANE
HEPTENE
HEXADECANE
HEXANE
ISO-BUTANE
ISO-BUTYL ALCOHOL
ISO-PROPYL ALCOHOL
ISOBUTYL ACETATE
ISOBUTtfL ISOBUTYRATE
ISOBUTYLENE
ISOMERS OF BUTENE
ISOMERS OF BUTYLBENZENE
ISOMERS OF DECANE
ISOMERS OF DIETflYLBENZENE
ISOMERS OF DODECANE
ISOMERS OF ETHYLTOLUENE
ISOMERS OF HEPTANE
ISOMERS OF HEXANE
ISOMERS OF METHYLPROP. BENZENE
ISOMERS OF NONANE
ISOMERS OF OCTANE
ISOMERS OF PENTADECANE
ISOMERS OF PENTANE
ISOMERS OF PENTENE
ISOMERS OF PROPYLBENZENE
OLE PAR
_ 0.4
3
3
2
- -
3
4
1
2
8
-
_ _
2
1
6
_ _
1
. 1
21
17
7
1 5
16
6
4
4
3
"~ 6
7
3
2
4
10
4
12
3
7
6
4
9
8
15
5
3
3
ARO
_
-
-
-
-
-
-
1
-
-
-
-
-
1
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
1
-
1
-
1
-
-
1
-
-
-
-
-
1
GARB
^
-
2
-
-
1
-
-
-
-
-
-
-
-
-
1
1
1
-
-
-
-
-
-
-
-
-
-
1
1
2
-
-
-
-
-
2
-
ETH
mm
-
-
-
-
-
-
-
-
-
1
1
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
UNREACT
1.6
1
-
-
2
-
-
1
-
-
-
-
1
-
-
-
-
-
-
-
-
-
-
-
C-4
-------
Table C-l. Bond Groups Per Molecule (in alphabetical order), (continued - 4)
Species Profiles by Bond Group
Saroad
No.
43113
43111
45107
43110
45102
43243
43444
43119
50022
45212
45205
43201
43432
43301
43445
43801
43552
43560
43559
43209
50016
43261
43262
43272
43819
50010.
43118
45801
43212
43435
43305
43238
43255
43260
43220
43303
45209
43259
43258
45101
43284
43235
50021
45211
45204
Chemical Name
ISOMERS OF TET3ADECANE
ISQMERS OF BRIDECAKE .
ISCMERS OF TRIMETHYLBENZENE
ISOMERS OF DNDECANE
ISCMERS OF XYLENE
ISOPRENE
ISOPROPYL ACETATE
LACTOL SPIRITS
M-CRESOL (3-METHYLBENZENOL)
M-E1HYLTOLUENE
M-XYLENE
METHANE
METHYL ACETATE
METHYL ALCOHOL
METHYL AMYL ACETATE
METHYL CHLORIDE
METHYL ETHYL KETDNE
METHYL ISGBUTYL KETDNE
METHYL N-BUTYL KETONE
METHYLACETYLENE
METHYLANTHRACENE
METHYLCYCLOHEXANE
METHYLCYCLOPENTANE
METHYLCYCLOPENTENE
METHYLENE BROMIDE
METHYLNAPHTHALENE
MINERAL SPIRITS
MONOCHLQROBENZENE
N-BUTANE
N-BUTYL ACETATE
N-BDTYL ALCOHOL
N-DECANE
N-DODECANE
N-PENTADECANE
N-PENTANE
N-PROPYL ALCOHOL
N-PROPYLBENZENE
N-TETSADECANE
N-TRIDECANE
NAPHTHA
NONADECANE
NONANE
0-CRESOL (2-METHYLBENZENOL)
0-ETHYLTOLUENE
0-XYLENE
OLE PAR
14
13
3
11
2
1 1
5
8
- -
3
2
- -
- -
1
8
- -
3
5
5
1.5
9
7
6
1 4
5
7
5
4
5
4
10
12
15
5
3
3
14
13
8
19
9
- -
3
2
ARO
mm
-
1
-
1
-
-
-
1
1
1
-
-
-
-
-
-
-
-
-
1
-
-
-
-
1
-
-
-
-
-
-
-
-
-
1
-
-
-
-
-
1
1
1
GARB
B
-
-
-
2
-
-
1
-
-
-
-
-
-
1
1
1
-
-
-
-
-
-
-
1
-
-
-
-
-
-
-
-
-
-
-
-
1
-
-
ETH
^
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
DNREACT
^m
-
-
-
-
-
-
-
-
-
1
3
-
1
-
-
1.5
-
-
-
-
1
-
1
1
-
-
-
-
-
-
-
-
-
C-5
-------
Table C-l. Bond Groups Per Molecule (in alphabetical order), (continued - 5]
Species Profiles by Bond Group
Saroad
NO.
43283
43233
43265
50023
45206
43817
45300
50006
43204
43504
43434
43205
43369
43602
50013
43208
45216
45220
43123
43309
45215
43390
45232
45202
43216
43226
45233
43824
43811
43821
43740
43822
50014
43241
43000
43860
45401
Chemical Name
OCTADECANE
OCTANE
OCTENE
P-CRESOL (4-METHYLBENZENOL)
P-XYLENE
PERCHLOROETHYLENE
PHENOLS
PROPADIENE
PROPANE
PROPRIONALDEHYDE
PROPYL ACETATE
PROPYLENE
PFOPYLENE GLYCOL
PROPYLENE OXIDE
PROPYLNAPH1HALENE
PROPYNE
SEC-BUTYLBENZ ENE
STYRENE
TER PENES
TERT-BUTYL ALCOHOL
TERT-BUTYLBENZENE
TETRAHYDPOFURAN
TETSAMETHYLBENZENE
TOLUENE
'mANS-2-BDTENE
TRANS-2-PENTENE
TRI/TETSAALKYL BENZENE
TSICHLOROEIHYLENE
TR ICHLOROFLOUROMETHANE
TR ICHLOROTR IFLOUROETHANE
TRIMETHYL AMINE
TR IMEOHYLFLUGROSILANE
TRIMETHYLNAPHTHALENE
UNDECANE
DNKNOWN SPECIES
VINYL CHLORIDE
XYLENE BASE ACIDS
OLE
_
-
1
-
-
-
-
-
-
-
-
1
-
_
-
-
-
-
1
-
-
-
-
-
-
-
-
-
-
-
.-
-
-
-
.1
-
-
PAR
18
8
6
-
2
-
-
1
1.5
2
4
1
2
2
7
2
4
1
8
-
4
3
4
1
2
3
5
-
-
-
3
-
7
11
4
-
2
ARC
_
-
-
1
1
-
-
-
-
-
-
-
-
-
1
-
1
1
-
-
1
-
1
1
-
-
1
-
-
-
-
-
1
-
.25
-
1
GARB
«
-
-
1
-
-
-
2
-
1
-
-
1
-
-
-
-
1
-
-
-
1
-
-
2
2
-
-
-
-
-
-
.32
-
-
ETfl
«
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
1
-
-
-
.16
1
DNREACT
^
-
-
-
-
2
6
-
1.5
-
1
-
-
1
-
1
-
-
-
4
-
-
-
-
-
-
-
1
2
3
-
C-6
-------
Table C-2. Example Problem1" - Part 1
Computation of Carbon Bond Concentrations
Measured Compound
Species
1
Ethyl ene
Propylene
n-butane
Trans-2-butene
2,3 dimethyl -
butane
Toluene
M-xyl ene
Benzene
Totals
ppbC
2
20
30
170
10
100
70
40
60
500
Species Profile*
PAR ETH OLE ARO CARB NR
3456 78
1
1 1
4
2 2
6
1 1
2 1
6
Carbon-Bond Concentration
PAR
9
10
170
5
100
10
10
305
ETH OLE ARO CARB NR
10 11 12 13 14
20
20
5
60
30
60
20 20 90 5 60
* Blank entry corresponds to zero
'''This is a hypothetical problem, and is not necessarily intended to be
indicative of the NMOC composition of ambient air.
C-7
-------
Columns 3-8 show the species profiles, as taken from Table C-l. Columns
9-14 contain concentrations of the individual carbon bond groups that are
attributable to the organic compound shown in column 1. These concentrations
are calculated according to the following formula:*
cij = T (C-l)
i (SP)I-J(CN)I-
1=1
where
C-jj = concentration of carbon-bond group i due to species j,
ppbC
Cj = measured concentration of species j , ppbC
(Sp)-jj = species profile number for carbon-bond group i and
species j (Table C-l)
(Cfj)-j = carbon number for carbon-bond group i (Table 2-1)
To illustate, consider the apportionment of 30 ppbC of propylene into its
individual components. From Table C-l, we see that one bond is defined
as olefin and one as paraffin. Thus,
(1.0H30 ppbcHl)
CPAR = U.OHD + i.o(Z) = 1Q ppt>c
and
(1.0H30 ppbc)(2)
COLE = (i.ujcu + i.uuj = 20 PP&C
In all subsequent formulae, the number 7 refers to the total number of
carbon-bond groups requiring inputs: PAR, ETH, OLE, ARO, CARB, DCRB,
and MR. However, since DCRB is always zero in this example, it has
been omitted from the computations in Tables C-2, C-3 and C-4.
C-8
-------
Note that the governing equation, C-l, results in the measured concentration
being multiplied by the fraction of carbon defined' for each carbon-bond
group.
After all of the species concentrations have been apportioned to the
carbon-bond groups, then the fraction of carbon corresponding to each
group can be calculated. To use the same notation as before:
N
E cij
j=l (C-2)
fi = 7 R
I E Cij
1-1 J-l J
where
fj = carbon fraction for carbon-bond group i
C-jj = concentration of carbon-bond group i due to species j
N = total number of species measured
Stated more simply, the concentrations of each individual carbon bond
group are divided by the total concentration measured. Table C-3 shows
the results of these computations for the example problem presented in
Table C-2. Note that, initially, the sum of carbon-fractions is 1.0.
Since oxygenates are not normally measured by GC analysis, 0.04 would be
added to carbonyls to account for their presence. Thus, the final set of
carbon fractions that is consistent with OZIPM/CB-3 sum to 1.04, and they
are shown in the right-hand column of Table C-3.
C-9
-------
Table C-3. Example Problem - Part 2
Computation of Carbon Fractions
CB-3 Class Concentration, ppbl ]
PAR
ETH
OLE
ARO
CARS
NR
305
20
20
90
5
60
Initial Carbon Fraction^
0.61
0.04
0.04
0.18
0.01
0.12
Final Carbon Fraction-
0.61
0.04
0.04
0.18
0.05
0.12
TOTAL 500 1.00 1.04
1 From Table C-2
2 Computed by dividing the carbon bond concentration by the total
concentration (i.e., 500 ppbc)
3 CARB adjusted to account for unmeasured oxygenates
C-10
-------
The preceding discussion focused on computations associated with the
analysis of one GC sample. For OZIPM/CB-3 applications, however, a number
of samples are recommended. Since it is extremely unlikely that all
samples will yield identical results, some method is required to reconcile
differences. The approach recommended here is to average fractions across
the samples, and then normalize those results to the value of 1.04. In
general notation,
K
£ fik
fi = k=1 (C-3)
~
and
JL
fi = 1.04 7 (C-4)
z Ti
1=1
where
T-j = mean carbon-fraction found from K samples
f-ik = carbon-fraction for carbon bond group i and sample k
K = total number of samples
/\
fi = normalized carbon-fraction for carbon bond group i
Table C-4 summarizes an example set of computations. Of course, the
normalized fractions (right-hand column) would typically be used in an
EKMA/CB-3 computation.
C-ll
-------
Table C-4. Example Problem - Part 3
Computation of Average Carbon Fraction
Carbon Fraction
Sample Number
Species
PAR
ETH
OLE
ARO
CARB
NR
1
0.61
0.04
0.04
0.18
0.05
0.12
2
0.59
0.02
0.06
0.22
0.05
0.10
3
0.60
0.06
0.02
0.20
0.05
0.11
Average^
0.60
0.04
0.03
0.20
0.05
0.11
Normal ized^
0.61
0.04
0.03
0.20
0.05
0.11
Total 1.04 1.04 1.04 1.03 1.04
1 Example: PAR = (0.61 + 0.59 + 0.60J/3 = 0.60
2 Example: PAR = (1.04/1.03)(0.60)^ 0.61
C-12
-------
APPENDIX D
Output for Benchmark Run
-------
o
o
o
o
o
ef-
ts
to
O
-3-
o
O
O
CJ r~t r-t
O
:r
u
o
u
o
X
o
u
ci: CM or
u o <
21 I D-
M O
o: uj
u i:
o
x
2 ^ 8 9 S
o
X
CJ
o
o z
X O
X
o
CO
o
-
o
X
z
o
Kl
o
to
o
o
z. o
to
o
to
o
to o
o z
o
u
§ §
o o
X X
o
O
o
CO
QL
rH Cvl Kl
D-2
-------
o
o
o
o
o
-------
o
o
o
o
o
f
o
o
LI
O
O
o
o
CJ ^ r-l
oj
co
O
o
u
CE
Q.
O
U
u
X
o
X
PJ
o
z
o
CJ
o
u
o
Q:
Q.
<
CK
<
Q-
O
U
o
o
Q. X
O O.
O O M
X = O
U O
X S X
u
ct
O
u
O
5 o
X U
OJ O
o x
LJ o
O
a;
cj o
cj eu o X O
O O -J CJ x
z X < 3: u.
O PJ
£ §
10
O
o
z o
O 1-4
X O
o
X
X
o
CJ
z
X
EL
X
o
§ §
o
o:
u
CD
o;
u
o
o:
o
a
a.
o
u
o:
O
a
z
uj g o
a. SI 51
a.
o
x
o
PJ
o
c:
o
X
o
fO
o
X
o
to
o
in
in
-o
in
r-
in
to
in
o
o
-------
o
o
o
t-i
t-
u
u
<
a:
o r«
O rH
O
L.-1
< n.
z o
O
U.
I-
<
>-
<
u
o
3
O
o
X O
o
CD
O
O
o
M
h-
u
O
CO
*
u
<
a
o
x
u
X
X
u
u
1/5
O
U
<
a:
o
x
a.
in o
r- o
o co
i-t O
<
o
O >H
a:
X
o
2 o
M O
O O
UJ fH
a
UJ
X O
> o
(>
0
Nl
tD
O
O
o
a.
LU <
r ->
M O
I- V)
a
S
z
o
o
z:
o
n
LU
X
o
<
o:
o
<
ct
cc o
-------
o
CO
to
CO
to
o
vn
o
CM
»H CM r-t
O O
O O
O O
o
o*
O
o
CM
O
o
o
Ul
o
M
O
CO
o
o
CJ
o
o
CM
O
o
o
(-1 fH fl CO r-*
o
o
-3-
O
CJ
o
O
-o
O f- O O
in r-t o o
|H |H rH CO
CO
CO
o
o
O UJ
ce
r-< UJ
o
in LU
M to
o
o
CM
O
CM
O
ro
o
>T
3-
UJ
o;
<<
M
(-
U
z
o
U
<
a:
a:
a:
O
O
A
UJ
o
CO
CO
CO
CO
o
o
CM
o
i
f-
o
CO
Ci
M
O
O
£
UJ
K
U
o
I
o
X
I-
D-6
-------
o o
c; ~
X u
vfOOOOOOOO
,3-00000000
cooooooooo
D: o O
< C3
a. Q. u.
000
o o o
CO O O
000
o o o
000
oooooooooo
_ u
O P Cl.
U UJ <
03 o
OJ O
o o o o
o o o o
OOOCJOOOOOO
o
H O
M K O
CO O
2 < Ul
a: (M
X <
o >
z o
Q O O
> >-* o
It- o
UJ (J O
D < O
UJ O
0. O
o <
O < < X
O O rH O
o o ro o
o o o o
o o o o
<7* O O O
M3 O O O
n c;
o o o o
o o o o
in o o o
(J O O O
o o
o o
o o
to in
o o
I I
in o rH o
o o o o
<7* O rH O
in
UJ
-------
<7" O xO
000
r*. "via vj- in r-* m
O O O O O O
cO cj f>
CJ rH O
0:00
< CJ X
Q. CC &_
rH co o
O O O
1 ) t
oo ro . Lft in u-i r-- in r-
o o o o o o o
rH fO O O
O t
U
U
c:
o
rH
CJ r-
O U
CJ CJ O
000
O rH O
I I I
*Q O Is*
0s O rH
CO O CJ
-£> in
0s -D
in r- m un co in I
o o o o o o
co r- in in
0000
in
0
in
r^.
X <
O h-
2 O
s a;
o u u.
r- o
O i-H
ITl CJ
rH O
r-f 0^ CO
O CO
c- fi
co a~
r^ in in trt r^ m o
o o o o o o o
111(111
o
o
CO
UJ ^
o o
UJ O
D - t
a. o
o <
c: D:
U H-
x. a
LJ
r >-
M a
o
o
cu
po o
O -i
z a
O
fl U
ro
O
o
UJ
~0
fu
CJ C
O UJ
UJ
Q.
O
0
<
o
o:
<
cr
K
O
U
0
o
CO
o
1
o
CO
CJ
CJ
CJ
o
rH
O
1
O
*
0
t
CO
O
1
CO
CO
rH
**
O
1
ro
-------
CT- CT- in
O O O
I
I
O O
o: ^
X u X
cj in
*r KI o
in co »
in -j- o
o o o
to -o -3- *
o o o o
in m in
o o o
CM
D: O O
t r- co
o o
d- i> cj
-JD "J CO
* f- CO
000
o in in y
a co
o o
CO CJ CO
Kl CO CO
o
CO
cd
X <
O >
2O
x cc r:
QUO.
a co -o
ooo
i i i
rO O rH
r- co <»
o r*~ co
ro
o
r^ ^f in in r- in o
0000000
LJ U
a <
CJ 2
r^ o HJ
O «£ a
Z o: o
J- fl SO
CJ CO CJ
0s o in
cj in o
o> IH in
oooooooo
t f t I I I I I
Q. O
o <
CL a:
a. u.
o o
" o;
to u < a.
o x a; 2
cj cj r- .H
O 0 0 O
1 1 1 1
cj so ro o
CJ CJ O O
r- in co in
ooo
f 1 1
(7* O CJ O
CJ r-t ITl O
Kl
o
r-t CJ CJ O
oooooooo
(J i
X O
o o o
u o: u
O < co
CO <
CD
in IH
i
|H CO I-H
in
uj r: o
O -J < rH
2 o a. o
<\i ro ?o in
o o o I-H
i i i i
CJ O f"- -3-
CO CO vT CJ
0s o e* cj
r-» CO CJ i-t
in in irt r^
o o o o
CJ O LT| O
in CO tO r-H
CJ O
CJ O Z
o uj cr o
-z. E o x
UJ >
Ju o:
o o o o
I I I I
o- in co ro
r-< CJ O in
-------
in LTI r-* u^ vo
o o o o o o o
o u
z <
a:
OOO
I I L
-O KI r-H
CO CJ CQ
co KI in
O (H O
t 1 t
co cj r^-
CJ *f <0
i
o o o o o o o
-
X < CO
Oh- - o
N i> K)
OOO
in o
CJ O
CO ^T
>f -f ~ M O
I H O
UJ CJ O
O *C O
-j c:
< L-
m o uj
o < c.
z a o
*o r- -o
OOO
I I I
sr fH tn
^D CJ O
r-. o co
OOO
I I t
r^ KI in
CM a- m
r-H - J-
o_ o
o <.
c: o:
a. u.
o o o o
o o
Z K
M o < nc
o =: a: ^
KI
-------
O O
c; n
X o
co
O O CO
CO CU If)
vJ- UO CO
O O O
I t I
o r^ ro
CO CO CJ
,0 v± r-- in o
o o o o o o o
I
I
i O O O
rHsfCMrHinCJrOOOO
w a a
o u -J
x i: <
Ifl O K»
O O O
I 1 I
rH CO rO
cj cO cO
CJ Kl x£
l
l/l co irt - M O
xl o
UJ CJ O
o < o
_i o;
< U.
M LiJ
x c: x
O o a.
CJ Z
ro o LU
O < a.
z a o
h- O
o o
C\j OJ LI
r-t O O
r*» CT- in
l l l
O CO CJ
to o in
CJ vT rH
I
I
CO CO cQ
o o o
I 1 1
r- ct ro
-T O rH
r- ro rH
O O O O O O O
I I I 1 I I 1
^OrHrHCJrHCMrHOOO
.
U I-
X O
O O O
u cr o
O < < X
UT CO rH
ro ro CM
UJ O
-J M
>- t-
a. u
o <
a D:
a. u_
o o
X CK
_ ,
O X o: Z
o o o o
Kl CJ O
* C4 fO LO
m in co
o o o
1 I I
-o co r-. o
o r-- o o
in ro Kl O
in ^ r^- o
in ^ o ^
> i i
CO
o
(J
CO
o
UJ >
z: ce
Zoooooooooo
O i i i i f ' t t i i
L)
< rH Kl
UJ
X
o
H-
o
D-11
-------
O O rH
lf\ ± h- cO
o o o
I } I
O 03 C-
O rH CJ
IT) CJ id
m cj KI
I
I
LI LI v"f Kl MT in
O O O O O O
CJ IT) O <* O CJ U"l O O rH
OOrHUlrHOMOOrH
rHrHrHrHLnrHOJOOO
2: o
< in
I PS
ro m r»
O t- U.
CJ LU <
o o o
< i t
^ O CO
J
rH
O CJ
CJ o: Q
o cj -J
x z: <
CD Z
M UJ
Ln
o
1
0s
c^
o
t
r**.
^
o
t
x*
CO
*>
o
f^.
f1^.
o
CJ
o
o
1
ff*
o
Kl
O
i
O
Kl
C7>
O O O O
o o o o o o o
UJ ^T
a a
>~ M
X (
LU a
a <
-j Q:
< u.
O
O
o
o
o
Kl O LU
O < a.
Z o: O
vO Is- -O
O O O
CJ O
tn o V
O cO CO
o o o
co ^T IH cr- r-» in -j* o o o
f^rHrHrHrHiHKlOOO
oooooooo
CO
CO
UJ _^
UJ O
-J w
> t
a. o
o <
D: o:
a- u.
CO
o
O O
KI u < a
o x a: i:
C71 rH CJ O
oooooooo
^O
z: a
O O O rH
< I I I
c- r^ cj co
N co ^o (7^
CO O O rH
cu in co in
o o o o
rH O
>r -o co r^
o o o o
I I I I
V)
UJ
r-
<
(sjocO'OWr^LfioLrii-*
oooooooooo
1 1 t I 1 I I I 1 1
U"I*OO''OCOWC'JOOO
UJvOCjr-tcJO-rHO-coaj.o
-------
CM Kl
o o
o r» K»
o o o
I I I
sO rH CO
rH O rH
U*l CJ CJ
CO O IT!
O O O
I 1 1
rH CO f
O O O
co
O O O O O
O O
* ^£
c:
O CO CJ
CO O -O
<7* O rH
in o *o
o o o o o
o <
-J c:
< u.
O < Q.
Z C£ O
CO O cO
O O O
in
co
LJ O
Kl
>- r- rH
a. u o
O -i
a: o:
a. u_
o o
3: a:
Kl u <
O I Q: -
O O O O
lit!
rH C-H (> O
cj r*« in o
rH O *£> in
O rH O
co co
CO CT1 0s CT*
o o CM in
O CO
o -3-
I 1 I i 1
tt
UJ
o
u
<
u
o
r-
O
UJ
X
D-13
-------
CJ Kl
O O
o: i:
X U
o o o
t t I
in o o
f^ o o
CJ CO vO
CO O O
000
rH O O O O O O O
111(111 1
O LA O LTt J'l CJ CJ O OJ
K>r-.rHrHrH-0COOrH
fO
O
CJ
O
X
0
OJ
>-
LJ
Q
UJ
UJ
-J
>-
Q.
O
C£
G.
U
X
t:
K
t t
2
o
t-H
r
(J
<
a:
u.
b
h-
O
rH
U
-<
o:
u.
O
)-t
i
u
c;
u.
_l
r-
O
L_
-O
in
cj
KI
K»
o
cr
co
0
0
o
o
0
o
o
r^
-o
o
r-l
O
^
CD
rH
*O
CJ
ce 9 g
a a: ul
u
o P £
U UJ <
CD CJ
cj o: a
O LJ -J
X il <
rH UJ
x Q: 3
c u a.
CJ Z
ro o UJ
g < c.
g o
K) U -t K
o x D; zi
Kl
ooo
U u: U
r^ ft .-! s**
O
1
CM
O
>3-
rH
O
!
O
CO
i-H
in
o
I
rH
rH
U1
O
Is-
O
i-H
ro
rH
Lft
O
1
rH
i t
in
rH
iH
O
\
O
r-*
CJ
m
^
0
o
rH
CO
O
1
LT)
UO
ro
Kl
o
10
o
o
rl
0
O
O
1
K»
r-
in
o
1
rH
CJ
O
1
^
K)
CJ
1^.
m
0
i
0^
<5
o
1
Lf|
O
10
0
. !
1
CJ
-£i
O
1
rH
CJ
in
O
1
O
-T
CO
o
Is-
<0
o
o
0
vT
m
0
i
Kl
CO
CJ
rH
I
0
1
0
^
1
C3
1
-
M Q
*£> <> in o
o o o o
to o
- O rH
ooooooo
rHOOinCJrHrHOOO
in CT- ^> in co LI in
ooooooo
iriCJincOrHrHCJOOO
ooooooo
oooooooo
LI in in r^ r* r-^ vo -t
oooooooo
ooooooooo
I I I I I I I I I
vTrH*OrHinf^r^r^iHO
KlrHrHO-OrH-T-O-OO
rHrHi-HrHOCJinincOO
D;
a.
o
eg
p-l
o
CJ CJ CJ ,
I I I
o
I
CJ
(7-
O
O
-O
CJ O
CJ O "
O LU cr o
z ^ cr 3:
UJ >
r: D:
fO <3- !> O
O O O O
Kl rH
O O
H- H-
O Kl
O "1
in o-
CO CO CO
o o o
CT* cj \o ^ r*- cj r^> ro
CJ O O O Kl CO -3" CJ
r^OLIrH COOrHrH
cn
LU
I-
<
r-
VI
UJ
r-
o:
Zoooooooooo
O I I I i I I I I I I
(J
UJ
o;
O
u
<
u
M
I-
>-
^J
O
s
a
r~
-T
D-14
-------
cJ ro
O O
o; i:
X O =
tn co -
rH O O O
CO CO l/l Kl Kl
O iH O
i i t
%o o co
to ui cr-
0s CO Kl
1/1 o o in o -o t/i
O i-( o O O O O
-O
CO
-
X <
OK
r O
o
O
co co in
r- KI ^H
o Is- - M
CJ Z
Kl O LU
O < C.
r a o
in co r-
000
co o a-
O rH O
t I i
x£ -O Kl
in r^ -r
o o a-
oooooooo
>- h-
D. (J
O <
ct o:
a. u_
O
O
o o
z a-
o x a s:
rH O CO O
CJ sT *O O
_
UJ
iu IH o co
a; *
.oooooooooo
o
u
(-
<
o:
o
r~
o
a
m
CJ
ro
D-15
-------
CJ Kl
o o
r-l o -o m
o «-* o o o o o
rHOCJOO-
o
in
CM
X < m
Or- 0
in
0
i
to
m
-3"
in
0
i
CM
^
o
o
0s
CO
*
to
o
1
o
in
^.
CM
0
1
G"
cr-
rH
O
1
O
o
in
1-1
o
rH
1
CO
to
in
*
oooooooooo
<
en
o
(J
o
rH
in
o
i
c-
co
o
2OOOOOOOOOO
O I I I i ' I i i I I
< P-H
DC O
U O
rH PI
UJ >
i: o:
O O
O O
r- o
<
a:
a:
UJ
z
o
X
a.
D-16
-------
<
I
ro o")
or:
o
M 1T|
w t- a\
O U O
2 <£ <0
a: o
x <
O r-
Z O
r-
O O
>- rH
X H
LU O
Q <
-I C£
< U-
uj o r*.
-i M tn
>- H- O
a u o
o <
ct cc
a. u.
CJ
03
cJ ro
O 0
QL II
X U X
CJ
C£ O O
< D-i X
Q. [X D.
U
o £ tl
U UJ <
O CJ
cj D; a
O CJ -J
X II <
o z
HH LU
X C X
O U CL
CJ Z
Kl O UJ
O *C CL
z a: o
000
i > t
*3- ca r--
H » 1 C"1
>3~ ro rO
. i
r-t cj in
rH f^ CO
o o o
! 1 1
f*» f"- CJ
ro ro r-
r- 0 rH
ro co r-
rH rO O
O O rl
1 1 1
o ro 0s
o co r*-
cj ^ ro
CJ -tf rH
in o ro
O rH O
1 1 1
o in m
CJ -T »
rH f-J rH
rO rH rH
CO 0 vO
o o o
1 1 1
T O 03
* « (
>j ro co
in ^ r^
o o o
1 1 1
in cj ro
Is* to in
ro co co
ro co r-t
o o o
i i i
o *c> ro
rH CJ *>
r-* r-* N?
>
rH rH rH
1
in co o
O O CD
1 1 1
iH O O
*J 0s CJ
ro o o
N FH ro
I i i
in o cj
O O r-H
i i i
rH rH CO
0s r-- u")
in Is- cj
CO >O CO
i i
f^» CJ >O
O rH CD
1 1 1
rH rH O
1 t I
cj o r^-
^r in -J-
J1 rH CJ
t
r- cj cj
i i
r-. o o
O rH O
1 1 i
^ CT* CJ
r*- ro in
«3" vT >O
rO rH CJ
O o
x o:
ro u < K
o x a: =:
ro
O O O
u a: u
O < < x
o o o o
i i i i
r* ifl 3-
o in cj o
ro o OD o
& o >x o
r- co vo cj
O O O rH
rH O O O O
v* ,0 O vf CO fO O CJ
sT cj ro
rHrH-TrHOrHrHOrO
co in ro
o o o
s£) O CO LA CO O O
O rH O O O O O
rOCJrHCJrHrHrOOOO
in o r- in o o
O rH O O O O
xo LCI in m
o o o o
rHCjrOCJrHrHrHOOO
oooooooo
I I I I
ooooooooo
-------
o
0
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
o
o
o
o
o
in
LU
t-
+ o
-------
APPENDIX E
OZIPM Output Corresponding to Inputs Shown in Table 4-3
-------
) O
) O
) O
Kl
O
O
1/J
sT
f>
O
O
U1
O
O
O
fo
O
O
O
O
ro
O
ui
-------
O
u
X
o o
X 2: CM
u u o
3: x x
o
X
O
u
O
u
o
u
CM O
CvJ O X
O LU O
i z: x
u
X
u
X
o
u
o
o
o
o
o o
u u
X X
o o
X X
u u
X X
X
o
o o
r. 2
O O
-------
o
o
PJ
o
ef-
ts
in
T
O
u
(J
CK
O.
o
o
o
I
u
X
CM
o
a:
o
o
o
o: cc
o. <-J 2
Z X I
(J
o:
a.
o
o
u o o
=:<_>;£
wo
o
o
X
CJ O CM
O X O
X 0. 2
O
U X
< o
o
z a
O r-l
x o
X
o
o
X
c
z o
X
O
o
o
X
X
D.
X O
o -z.
X
o
o o
a: o:
u a
o
o:
o
c:
O
a:
a.
o
U
o:
a.
o
X
a.
a.
o
X
o
CM
o
c:
fj
O
X
o
o z
1.1
If!
-o
in
CO
in
o
-o
K1
O
-9-
-O
Ift
o
-o
o
co
o
o-
-o
CM
O-
E-4
-------
o
o
o
a
t-
u
o
o
o
u.
a
u
in
a
u
O
I-
g
a!
0
0
H
IT!
r*
UJ
o
t
M
t?
5
_J
O
It
UJ
6
NJ
UJ
>H
H-
80
00 LOCAL DAYLIGHT
o- to
iH iH
*3" O
CJ \-
o
o
O 03
UJ UJ
h- E
< l-l
O t-
r-1
o
Kl
rH
O
O
a.
X.
_J
o
01
(S
w
a
_j
_j
o
LU
UJ
(-
r
o
u.
o
UJ
z
M
y
o:
UJ
i
UJ
a
^
o
w
i-
^
o
< a.
z o
M 1-
u.
»
2;
O
M
UJ
X
0
in o
o: z
UJ >H
> r:
M H
INHIHG OF EACH HOUR)
CD
UJ
CQ
UJ
^~
1-
1-
0
o
co
UJ
M
H
-T
0
rH
rH
-
UJ
o
-j
<
2
UJ
o;
<
UJ
a:
LU
X
K
tO
o
rH
f-
<
Q:
t
HZ
UJ
o
o
tj
Q
UJ
r
O
a.
CO
2:
a:
»-
E-5
-------
z: E
a. c.
a. a.
x
o
X
o
I CC
< 3
H O
o
a
a:
<
u
o
a:
a
u
o
a:
Q
Z o:
W LU
a.
a.
O Q
UJ
O ~
< LU
o
CO
O
O o:
2 §
o =:
M
H- CK
U UJ
< Q.
££
o
rH
O
o
1^-
o
to
<
<
o:
o
u
O
o
= t-
LU
2
O
in
in
O O
M M
o o
U
<
o
a
u
o
a:
a
in
M Q
n z
LU ^)
O
if> o;
p o
Z Z
o o
o z
o:
3
O
o
<
o:
C.
x
LU
10
in
tH
z:
LU
in
o
C
U
X
o
2:
o
a.
13
o:
O
o
O
Lfl
o
<
ce
z z
o o
U Z
E-6
-------
K-
IT O iH
< O O
\- o 03
fO CO i-H r-1
or o o
H CO
CO O
z
o
hH
CM r-
O U
Z <
Q:
u_
0
0
o
in
0
CJ
a
co
in
co
CO
ro
^
in
co
o
CO
o
CJ
CT-
in
CO
-0
ro
r-
IT-
CO
in
CO
-------
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
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
Kl ~
sn
-------
APPENDIX F
OZIPM Output Corresponding To Inputs Shown in Table 4-4
-------
o
o
o
CJ
o
-r
o
o
-o
o
X
u
a:
o
u
o
X
o
2
X
o
O C2
o <
< o.
ct
u
u
r:
o
x
o:
£
c;
u
C-J
o
c\j ro
g ro g o O
o
O
3:
O O
2 2
OX
o
z
X
o
o
X
X
o
s o
o
z
X
o
to
o
o
z.
fO
o
ro
o
o
z:
O
u
§ i
ro
O
CJ
o
o
X
x
i-
UJ
o
u
CJ
O
UJ
c:
a.
tn
o
UJ
X
F-2
-------
o
o
o
o
o
o
o
o
o
o
o
CO
CO
Kl
O
fO
Ki
PJ
o
O
u
o
o
X
u u g
o
u
o:
<
a.
c:
<
a.
O
u
o
u
CJ O
CJ O X
o u a
o o
X X
u o
X X
o
o
Kl
O
u
o
X
o
u
o
z
o o
u O
X X
g
x
O O CJ
x x o fa
u u c: o o cr
x x u x x u er
Z o
< u
x a. <
C\J Kl
i i
o
u
o
X
O
z
to
o
o
o
X
X
o
X
o
i
o
X
o
X
u
X
o
z
o
a
x
o
u
X
g
u
x
o
u
X
o
X
u
X
a-
o-
cr
er
O M »-* M
ui rv r>* rr
H <_> U U
a
u
r:
a
u
a:
u
IB to
1-1 o:
D: u
u Z
13
cr
in
Kl
CO
Kl
Kl
CO
O
f
CO
-------
o
o
o
o
.H xT
O O
i-l O
O iH
o
u
u
c:
o.
u
X
o
I X
o
u
o
o: D:
a.
-------
o
o
r-
o
in
o
PI
o
z z
o o
a.
a.
K.
O
U-
a
LU
i-
_t
13
y
^
10
z
<^
CT
z
O UJ
u r:
UJ Z
6
1- Nl
rH
CD UJ
O rH
-J 1-
AL DAYLIGHT TIME
U
o
_J
0 0
CO 0
rH rH
-J- O
CJ r-
0
0
>O CO
UJ UJ
1- £
Q 1-
tD
M
2
O
_J
j
O
LU
UJ
X
O
a:
u.
Q
UJ
k_j
y~
tt
UJ
UJ
a
-^
o
M
t-
-J
IH
a
Kl
CVJ
2:
M
U.
t
O
in
CU
_j
^
M
t-
HH
( (
to
J
X
o
HH
UJ
6
to
ex
UJ
5
rH
O
rH
a.
o
h-
10
1
o
o
CO
a:
K
V^
UJ
M
rH
t-
o:
O
X
u
UJ
u.
o
(3
S
z
M
O
UJ
03
UJ
X
r-
H~
^0
CO
l_
z
ts
M
UJ
X
o
HH
X
r
o
o
Ul
rH
O
O
rH
O
O
Kl
rH
O
O
CJ
l-t
0
0
(H
rH
0
0
0
r4
0
O
O*
o
o
CO
UJ
M
H-
iH
Kl
*-H
rH
tri
rH
rH
rH
Ifl
^r
0
rH
rH
LI
^y
,
%J-
O
03
K»
t
rH
O
if\
rH
rH
sT
O
O
e\j
h-
O
rH
UJ
X
ETH FRACTION . 040 OLE Fl
0
CO
in
2
O
M
J~
U
a:
u.
a:
D.
^
H-
M
"Z.
rH
HCHO FRACTION .050 ALD2 Fl
0
a*
rH
2;
o
M
t-
u
OS
LU
O
c;
^
rH
1-
rH
Z
M
0
in
rH
Z
o
^_
U
a:
u.
0
D; in
~ CM
-J X
< O
HH H
H \
rH CM
Z 0
vi
M
-------
a.
a.
x
o
o
o
DC
^
U
o
QC
a
>-
X
0
0
UJ
o
KJ
O
h-
U.
O
.J
<
M
h-
M
M
U.
O
o
1
u
<
c:
u.
UJ
X
(-
(O
<
a
UJ
to
CO
LU
K
a.
x
UJ
to
o
M
CO
CO
hH
z:
UJ
to
g
D
M
I-
2
O
U
o
~
a:
UJ
a.
a
UJ
t
»-
t-H
T~
UJ
O
t-t
1
<
cz
to
UJ
LJ
^
O
u
6
CO
a:
u
o
DC
O
>-
X
a
^
D
0
C:
o
^
u
CO
~^
o
z.
0
n- cj
o
0
W3
-------
CM
3"
CJ
in
o
X
<
E
0
CM rH
r» ro
-3- in
rH CT-
O OJ
.04847
X
E
g
.07416
X
E
0
rH
rH
o-
in
rH
X
E
g
o
OJ
in
CM
X
E
O
.19053
X
E
Z
.14587
X
E
O
o
0
CO
CO
o
X
g
.24715
X
E
g
ro
in
ro
CM
CM
X
0
7
.19709
X
E
O
CM
in
in
in
rH
X
E
g
.07946
X
E
O
.18037
X
E
2
.17310
X
E
O
.16909
X
O
.15795
X
E
O
o
CO
O
X
E
O
rH
O*
CJ
ro
rH
X
O
CJ
CD
CJ
ro
rH
X
r
o
2:
r-4
ro
rH
X
o
-r
.14591
X
E
g
.12810
0
rH
\-
<
00000
00000
14286
r-
o
0
o
0
o
0
ff
o
in
ro
91906
in
in
14286
r-~
o
CO
OJ
-3-
rH
r-~
40469
o
40469
o
40469
-o
40469
-o
40469
-0
42047
in
42047
in
42047
m
42047
in
42047
in
o
CJ
SO
in
-o
-------
X
E
H
O
.09333 t
X
z:
i
o
.10017 t
X
i
o
£)
rH
ro
H
O
|H
CO
0
in
CJ
X
E
r-
0
CJ
in
rH
X
E
1-
O
.13158 h
X
z:
t
o
.00394 t
X
t-
o
-I 06063"
X
o
in
in
ro
CJ
X
a
E
1-
O
.16599 t
X
E
V-
O
.12294 h
X
E
O
.08243 t
O
fH
rH
in
CJ
X
H
O
.20708 ^
X
o
rH
CJ
rH
X
r-
0
.11309 h
r-
co
o
in
o
CO
to
-o
CO
-i-
~O
T
rH O
ro co
o
-3-
O
O
in
ro
o
o
O
in
ro
CJ
o
CO
CO
CO
o
o
ro
in
-y
CO
-3-
rH CO
iH -3"
3- r~
r^ r-i o
CO CJ rH
CO -O CO
.H rH
^ ro
o in
o
-o
in
CT-
co
ro
co
O
o
O CO
rH IT!
in r-
o- £>
ro
cj
r-
o
in
o
o
in
cj
ro
3-
ro
c-
o
CO
CO
ro
rO
F-8
-------
X
«t
X
-
CO
cu
v£)
co
CO CO CO CO CO I-H
CO CO CO CO CO CO
o o o o o
-------
o
o
-» CU O *
o
-o
a.
o c.
O w
CO
CJ
o
o
to
o
z o x Q. Q. r:
F-10
-------
APPENDIX G
OZIPM Output Corresponding To Inputs Shown in Table 4-5
-------
o
o
o
o
o
o
CM
O
o
-o
o
o
u
x
o
u
C\J
O
O
u
CE
<
0.
CD
a:
u
Csj Kl CJ
fO O O O O X
O = 2 = = O
CM
O
-
O
UJ
z:
O CM O O
= O Z X
U < CJ (_J
z o: x =
O CM O
O X
-------
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
ri
ro
u
I
o
(J
X
o o
x x c\i
u u o
o o
X X
o
u
o
u
o
u
CM O
-------
o
o
o
CO
o
I
u
o
x
o
X
o
u
o
o
a
<
o:
<
a.
o
u
D- X
o a.
o ro
x o
U "Z.
X X
U
a:
a.
o
2
O
CE
<
O
X
o
2
O
X
o
I
Z CJ
OJ O W
S -I O
f J O
-------
0
0
C-
o_
o
o
0
0
RAC
D2
X
o
Lfl
Kl
C\J
0
0
a
o
o:
o
<
Q
_J
<
CJ
o
O
Z
I
u
0
0
3
O
o
W
i s
rH I-H
r- H
CJ O
< <
a a
LU Lu
r- U
UJ X
o
in
o
a
o:
<
u
o
a:
a
>-
x
3
U
CO
z
o
u
CO
o
0
Kl
t-
M
u
o
r-
O
O
Lfl
§ S
r- fvJ
rH
CD LU
O M
O
U_
LU
r z
O rH
o;
E X
Q; C3
LU *-H
t- UJ
UJ X
0 z
z o
O M
r- a:
M 2 rH
Q W t-
UJ
OJ
z
o
u
o:
o
CD
X
M UJ
H X
.CJ
a:
O o
ce o: in
< z c\j
g
Z Z Z O
HI M W 2
UJ
a:
Z O
O M
M O
o
u
a:
o uj
Q. CJ
CO <
Z u.
< n:
D: 3
I- 01
G-5
-------
a
C-
X
o
o
rj
o
^
CJ
f£
u
o
c;
a
>-
I
o
-o
o
UJ
o
o
1-
u.
o
_J
^
<
HI
(-
Z
M
u
O
2
o
K
o
<
a:
u.
UJ
z
H-
<
a
UJ
to
UJ
a:
a.
x
LU
01
0
HI
U1
U1
M
E
UJ
in
z>
2
HI
H;
O
(J
5
o
X
cc
UJ
a.
a
UJ
(
H-
HI
T~
LU
2"
O
HI
H-
<
1
^
O
o
u
o
a
o:
<:
u
o
o:
a
>-
"T*
Q
13
o
D:
ta
o
ca
i
~r
O
Z
r~
>o
in
j-
M
f-i
CK
3
0
3Z
O
fvj
0
o
CJ
0
o
CJ
0
>
0
0
r-1
O
r-1
O
r-
rH
O
r-
rH
*
^
o
HI
H-
Q:
u.
M
h-
M
HI
U_
O
O
H-
(J
<
a
u.
UJ
X
H-
<
a
UJ
U}
UJ
a.
x
UJ
U*l
o
HI
in
M
T*
UJ
tn
o
HI
H;
6
u
^^
K
3
O
X
C£
UJ
a.
a
UJ
H-
HI
r
UJ
5
H.
<
a
z
UJ
u
o
X
g
Q
^3
O
ID
U
a
t
Q
Z
1^
O
in
*
10
w
rH
fy
0
X
0
ro
0
o
ro
0
0
to
0
t
0
o
rH
0
U1
m
o
U1
K)
O
in
Kl
~7"
O
HI
(-
O
a:
"
6-6
-------
X
<
X X
«£ <£
XX
< <
IT- in
K> fO
IO If) CM
O -
X
ft
E
X
<
2:
X
-------
1
g
to
o*
r-
o-
o
X
<
g
r~
03
r-
c-
o
X
<
z:
g
03
CJ
CO
O-
O
X
<
g
r^
o
r-
o
.H
X
<
g
rH
to
o
CM
rH
X
<
£.
g
o
O
O
r^
0
X
<
Z:
g
r-»
~o
o
(O
o
X
<
z:
g
0
o
a*
0
rH
X
<
O
CM
to
03
o
r-H
X
<
£
g
to
r~
rH
CO
O
X
<
z:
g
in
vO
r^.
rH
to
X
<
o
-z.
t
to
vj"
LTI
CM
X
<
o
z.
-o
o
o
J
-0
CO
0
X
<
§
X
-------
t-
o
o
X X
< <
n s:
9 9
CO
o
CO
o
in
o
ro
in
CO
in
in
CO
in
CO
in
CJ
in
.3-
cu
-o
m
o
CO
o
PO
o-
ro
CO
-o
-o
CO
in
o
-o
CO
o
o
o
G-9
-------
o
o
O
CO
a.
o a.
o ~-
o
CO
+ +
O
-------
o
UJ
6-11
-------
TECHNICAL REPORT DATA
(Please read Insmtctions on the reverse before completing)
4.
9.
REPORT NO. ' 2. 3. RECIP
EPA-450/4-84-005
TITLE AND SUBTITLE 5. REPO
Guideline for Using the Carbon Bond Mechanism in
City-.specific EKMA 6-PERF
AUTHOR(S) 8. PERF
Gerald L. Gipson
PERFORMING ORGANIZATION NAME AND ADDRESS 10. PRO
U.S. Environmental Protection Agency
Office uf Air Quality Planning and Standards I'i'.'coN
MDAD, AMTB, Mail Drop 14
Research Triangle Park, North Carolina 27711
12. SPONSORING AGENCY NAME AND ADDRESS 13. TYP
14. SPO
Same
lENT'S ACCESSION NO.
RT DATE
February 1 984
ORMING ORGANIZATION CODE
DRMING ORGANIZATION REPORT NO.
GRAM ELEMENT NO.
TRACT/GRANT NO.
E OF REPORT AND PERIOD COVERED
NSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The document describes how to use the Carbon Bond 3 (CB-3) chemical mechanism in
the city-specific EKT1A model as a means for estimating emission control requirements
needed to demonstrate attainment of the ozone NAAQS. Topics addressed include (a) an
overview of the CB-3 mechanism, (b) procedures for applying the CB-3/EKMA model, and
(c) special computer considerations to be taken into account when using CB-3 with
city-specific EKMA.
17. KEY WORDS AND DOCUMENT ANALYSIS
a
DESCRIPTORS b. IOENT] HERS/OPEN END
ozone
control strategies
photochemical pollutants
models
SIPs
EKMA
OZIPP
carbon bond mechanism
18. DISTRIBUTION STATEMENT 19. SECURITY CLASS (This
, , ,. 4. j 20. SECURITY CLASS (This
Unlimited
ED TERMS c. COSATI Field/Group
ReportJ 21. NO. OF PAGES
157
page} 22. PRICE
EPA Farm 2220-1 (Rtv. 4-77) PREVIOUS EDITION is OBSOLETE
------- |