&ER&
            United States
            Environmental Protection
            Agewv
            Office of Air Quality
            Planning and Standards
            Research Triangle Park NC 27711
EPA-450/4-80-027

March 1981
Guideline For Use
Of City-specific EKMA
In Preparing Ozone SIPs

-------
                                             EPA-450/4-80-027
Guideline For Use Of City-specific EKMA
            In  Preparing Ozone SIPs
                  Air Management Technology Branch
                  Monitoring and Data Analysis Division
               U.S. ENVIRONMENTAL PROTECTION AGENCY
                   Office Of Air, Noise, and Radiation
                Office of Air Quality Planning and Standards
               Research Triangle Park, North Carolina 27711
                        March 1981

-------
This report 1s Issued by the Environmental Protection Agency to report
technical data of Interest to a limited number of readers.  Copies are
available free of charge to Federal employees, current contractors and
grantees, and nonprofit organizations - in limited quantities - from the
Library Services Office (MD-35), Research Triangle Park, North Carolina
27711; or, for a fee, from the National Technical Information Service,
5285 Port Royal Road, Springfield, Virginia  22161.
                   Publication No. EPA-450/4-80-027
                                  11

-------
TABLE OF CONTENTS

                                                                       Page

Summary	     v

List of Figures	     x

List of Tables	    xi

1.0  Introduction	      1

2.0  Estimating Control Requirements for 1982 Ozone SIP s	      7

     2.1  Selection of Days to Model 	     10
     2.2  Day-Specific Control Estimates 	     10
     2.3  SIP Control Requirement	     11

3.0  Generation of Day-Specific Isopleth Diagrams	     16

     3.1  OZIPP Modeling Data	     17
          3.1.1  Light Intensity	     19
          3.1.2  Dilution	     19
          3.1.3  Transport of Ozone	     20
          3.1.4  Precursor Transport 	     24
          3.1.5  Post-8:00 a.m. Emissions	     26
          3.1.6  Reactivity	     38

     3.2  Empirical Data	     39
          3.2.1  Dally Ozone Design Value	     40
          3.2.2  NMOC/NO¥ Ratios	     40
                        rt
     3.3  Use of OZIPP to Generate Isopleth Diagrams 	     43

4.0  Calculation of Control Estimates	     47

     4.1   Use of Single Day-Specific Diagrams	     48
     4.2  Concurrent Changes in Emissions and Other Factors	     49

          4.2.1  Concurrent Reductions 1n Local  Precursors and
                 Transported Ozone 	     52
          4.2.2  Concurrent Reductions 1n Local  Precursors and
                 Transported Precursors	     58
          4.2.3  Concurrent Changes In Local Precursors and
                 Diurnal  Emission Patterns 	     59
          4.2.4  Consideration of Multiple Changes 	     60

5.0  Acknowledgment	     62

Appendix A.   Estimation of Mixing Heights for Use 1n OZIPP 	    A-l

Appendix B.   Consideration of Transported Precursors 	    B-l

References	    R-l

-------
SUMMARY
     This summary outlines the procedure used 1n applying city-specific EKMA 1n
State Implementation Plans (.SIPs) for ozone.  City-specific EKMA consists of two
components:  the OZIPP model and the EKMA procedure (I.e., OZIPP/EKMA).

     OZIPP (Ozone Isopleth Plotting Package) is a computer program which allows
the user to plot maximum hourly ozone concentrations as an explicit function of
initial (e.g., 8 a.m.) ambient concentrations of the precursors non-methane
organic compounds (NMOC) and oxides of nitrogen (NO ).  The resulting output 1s
                                                   ^
an ozone isopleth diagram.  Because the OZIPP model allows consideration of
meteorological inputs, pollutants transported from upwind and precursor emissions
which are appropriate for each city under review, the Isopleths are also implicit
functions of these variables.

     EKMA (Empirical Kinetic Modeling Approach) is a procedure for applying the
ozone isopleth diagram obtained with OZIPP to estimate the Impact of controlling
urban volatile organic compound (VOC) and/or NO  emissions on peak hourly ozone
                                               A
concentrations.  Ambient 6-9 a.m. NMOC/NO  ratios measured 1n the urban core and
                                         J\
peak hourly ozone measured within or downwind of the city are needed to apply an
Isopleth diagram in the EKMA, procedure.

     Because urban VOC and/or NO  emission reductions are likely to be more
                                ^
effective under some conditions than others, there is not necessarily a simple
1:1 correspondence between high observed ozone concentrations and emission controls
estimated as necessary to attain the National Ambient A1r Quality Standard (NAAQS)
for ozone.   Thus, in order to ensure that the standard 1s attained, 1t 1s neces-
sary to perform city-specific EKMA modeling for several days.  This provides more

-------
complete assurance  that  the  standard will be met 1f a control target 1s achieved.
Consideration of  the  five  days with the highest observed dally maximum ozone
concentrations at each site  affected by the urban area 1s believed to be sufficient.
     The procedure  for applying city-specific EKMA is outlined in the following
seven steps.  Each  step  is discussed 1n greater depth 1n the indicated section(s)
of this report.

          1.   Select the  five days with highest maximum hourly ozone concentra-
tions observed at each site  affected by the urban area over the past three years
(Section 2.1).

          2.   For  each  of these days, estimate the 6-9 a.m. NMOC/NO  ratio
                                                                    ^
prevailing in the urban  core (Section 3.2.2).

          3.   Using  the input data summarized 1n Table SI, apply the OZIPP
computer program  to generate ozone isopleth diagrams for each day.  The exact
procedure 1s described in  User's Manual for Kinetics Model and Ozone Isopleth
Plotting Package. EPA-600/8-78-014a.  The resulting diagrams denote sensitivity
of peak hourly ozone  concentrations to locally generated precursors, given a set
of "current" accompanying  conditions (e.g., a specified concentration of trans-
ported ozone from upwind sources).

          4.   Using  the ozone concentrations obtained in Step 1  and the NMOC/NO
                                                                                J\
ratios derived in Step 2,  establish a "starting point" on each of the ozone
isopleth diagrams obtained 1n Step 3 (Section 4.1).

          5.   If there  1s a change(s) in the base case conditions assumed to
prevail in generating Isopleth diagrams obtained in Step 3, it is necessary to
                                        VI

-------
Table SI.  Inputs Needed to Generate an Ozone Isopleth Diagram
Input

(1) Light Intensity
    Atmospheric
    Dilution
(3) Present Transport
    of Ozone
(4) Present Transport
    of Precursors
 (5) Post 8 a.m. Emissions
Mow Used and Obtained

Enter the datn and a city's latitude, longitude and time
zone to the OZIPP program.

Use National Weather Servlre (NWS) or Urban Radiosonde data
and urban surface temperatures to estimate R a.m. mixing
height, maximum (tally mixing height and the Mmr> which the
maximum daily mlxim) height occurs.  In the event NWS or
urban data arc unavailable, use values compiled In Table A-l.

The model assumes that ozone Is transported In two layers:
at the surface and aloft.  Ozone transported at the earth's
surface Is generally assumed to be zero.  Ozone transported
aloft can be estimated using mid-morning ozone concentrations
observed at a continuously operated surface monitor located
upwind of the city on each day considered.

Measurements and past model applications have indicated the
Impact of transported precursors is generally negligible.
Hence, these are ordinarily assumed to be zero.  In cases
where thorn Is reason to believe precursor transport is high,
special procedures nro needed.  Those procedures roqiiire fi-9 a.m.
measurements of precursor concentrations upwind of the city
obtained by simulation of organic species.

Derived using 6-9 a.m. ambient NMOC and NO  measurements In the
urban core on each day modeled, seasonal countywlde VOC and NO
emir.-.Ion inventories anil  I nforin.it I on concerning the location of
the maximum observed ozone concentration on each day modeled.
Sectlon(s) of Report
Describing Process
In Detail	

Section 3.1.1
Section 3.1.2
Appendix A
Section 3.1.3
Section 3.1.4
Appendix 0
Section 3.1.5
(6) Reactivity
Recommended default values should he used.
 Section  3.1.6

-------
generate a second set of  Isopleth diagrams.  This second set of diagrams reflects
the Impact of changing the set of "current" accompanying conditions described 1n
Step 3.  For example, a second set of diagrams is most frequently used to allow
the user to consider changes  in ozone transported Into a city.  In such an appli-
cation, the second  set of diagrams is obtained using an assumed future concentra-
tion of transported ozone.  Suggested procedures for estimating future transported
ozone are described in Section 4.2.1.

          6.   VOC and/or NO  controls needed to reach 0.12 ppm 0« on each day at
                            X                                    «3
each site are estimated using the starting point on each diagram representing the
current conditions  (Step  4} and calculating controls needed to reach the 0.12 ppm
ozone Isopleth on the corresponding diagram obtained in Step 5.  This procedure
enables the user to consider concurrent changes 1n local precursor emissions and
in such variables as transported ozone from upwind sources.

          7.   As a result of Step 6, there is one estimate of VOC controls
needed to reach a peak hourly ozone value of 0.12 ppm at each site for each day.
Each of these estimates 1s valid only for the best estimate of changes in NO
                                                                            /\
between the base year* and 1987.  The site specific VOC emission reduction is the
fourth highest control estimate, assuming a three-year data base.   If there were
only one year of data at  a site, the control requirement would be the second
highest control estimate; two years of data, the third highest control estimate,
etc.  Because of possible significant changes in emissions or emission patterns
over longer periods, a three-year sampling period 1s the maximum which should
* Normally the base year is 1980.

-------
normally be considered.  The SIP VOC emission reduction for the urban area  is the
highest site-specific control estimate.   Selection of the target control  goal  for
a SIP is illustrated 1n Table 2.2.
                                      ix

-------
LIST OF FIGURES                                                           Page
3-1  Example Isopleth Diagram	       18
3-2  Examples of Acceptable Monitoring Locations for Estimating
     Transported Ozone 	       25
3-3  Example Formulation of Post-8:00 a.m. Emission Density Scheme  .       30
3-4  Illustration of Emission Fractions Calculated Using
     Initial Conditions	       35
3-5  Example Calculation of Emission Fractions 	       37
3-6  Example Calculation of the Design NMOC/NO  Ratio	       44
                                              A
4-1  Example Emission Reduction Calculation Using a Single
     Ozone Isopleth Diagram	       51
4-2  Future Ozone Transport as a Function of Present Transport ...       54
4-3  Example Emission Reduction Calculation Considering Changes
     in Transported Ozone	       57
A-l  Flow Chart for Table A-3	       A-9
B-l  Vertical Pollutant Profile Illustrating a Shallow Layer of
     Elevated NOV Concentrations Aloft 	       B-5
          %      /\
B-2  Example Emission Reduction Calculation with the Explicit
     Consideration of Precursor Transport in the Surface Layer ...       B-10
B-3  Illustration of the Treatment of Contiguous Urban Areas ....       B-l6

-------
LIST OF TABLES                                                             Page
S-1  Inputs Needed to Generate an Ozone Isopleth Diagram 	     vii
1-1  Data Requirements for Use of City-Specific EKMA	       3
2-1  Tabular Estimation of Site-Specific Control Requirements	      12
2-2  Selecting the SIP Control Requirement: An Example	      14
3-1  Data Requirements for Calculating Mixing Heights According
     to the Recommended Procedure	      21
3-2  Recommended Procedure for Calculating Emission Fractions	      36
3-3  OZIPP Options for Model Input Data	      45
4-1  Control Calculations Using a Single Isopleth Diagram	      50
4-2  Control Calculations Using Two Isopleth Diagrams	      53
A-l  NWS Radiosonde Stations and Climatologlcal Mixing Heights for
     Selected Cities 	      A-3
A-2  Preferential Order of Data Selection	      A-4
A-3  Procedures for Estimating Mixing Heights	      A-7
A-4  Worksheet for Computing Mixing Heights	      A-8
A-5  Surface and Sounding Data	      A-10
A-6  Morning Mixing Height Determination 	      A-13
A-6A Example	      A-14
A-7  Maximum Mixing Height Determination	      A-l5
A-7A Example	      A-l7
B-l  Typical Effects of NMOC Transport Aloft on Control Estimates. . .      B-2
B-2  Typical Effects of NMOC Surface Layer Transport on Control
     Estimates	      B-7
B-3  Explicit Treatment of Surface Layer Precursor Transport 	      B-9
                                       XI

-------
1.0  INTRODUCTION



     This document is intended to fulfill several purposes:





          1)  to describe application of a relatively simple modeling procedure,



city-specific EKMA, which is useful for estimating the effectiveness of reducing



emissions of volatile organic compounds (VOC) and/or oxides of nitrogen (NOX) in



reducing peak ozone values downwind or within a city;





          2)  to indicate how available emissions, air quality and meteorological



data are used in the modeling procedure;





          3)  to provide one means for Air Pollution Control Agencies to check



the completeness and suitability of analyses performed with the model to support



the need for urban VOC and/or NOV emission controls.
                                y\




     The city-specific EKKA modeling procedure has also been referred to as



"Level III analysis."  These terms are synonymous.  City-specific EKKA consists



of two components:  the CZIPP model and the EKMA procedure (i.e., OZIPP/EKMA).1 >2'3



"OZIPP" (Ozone Isopleth Plotting Package) is a computer program which allows the



user to plot maximum hourly ozone concentrations as an explicit function of



initial (8 a.m.) ambient concentrations of non-methane organic compounds (NMOC)



and NO  within the urban area.    The resulting output is an ozone isopleth diagram.
      A


Although the ozone isopleths are explicitly plotted as a function of initial NMOC



and NO , they are also implicit functions of numerous other variables.   For
      /\


example, the role of subsequent VOC and NO  emissions, meteorology and  pollutants
                                          A


transported from areas upwind of the modeled city are also considered in OZIPP.

-------
      "EKMA"  (Empirical  Kinetic Modeling Approach)  is a procedure for applying the
ozone isopleth diagram  generated with  the OZIPP to estimate the impact of changing
urban VOC  and/or  NOV  emissions on peak hourly ozone concentrations.  Two pieces
                    /\
of  information must be  available to apply EKMA:  the 6-9 a.m. ambient NMOC/NO
                                                                             /\
ratio prevailing  in the urban area, and the peak hourly ozone concentration
measured within or downwind  of the city.  These two pieces of information provide
a starting point  on the isopleth diagram from which to estimate control requirements.

      The conceptual basis  for the Isopleth diagram is a Lagrangian model (OZIPP)
which follows  the evolution  of ozone from precursors within a migrating column of
air.   The  column  of air extends from the earth's surface throughout the mixed
layer.  All  pollutants  are uniformly mixed within  the column.  From early morning
to mid-afternoon,  the height of the column Increases due to the diurnal rise in
mixing height.  This  results in simultaneous dilution of pollutants within the
column and entrainment  of  additional pollutants from above the mixed layer.
Concurrently with the foregoing dilution and entrainment, fresh precursor emis-
sions encountered as  the column is transported by  the wind are injected into it.
All resulting  pollutants within the column of air  are assumed to react chemically
in accordance  with the  chemical kinetic mechanism  in OZIPP.

      The data  base needed  to develop isopleth diagrams in accordance with the
model  formulation described  in the previous paragraph is summarized in Table 1-1.
Such  a data  base  is consistent with what has been  described on pages 65669-65670
in the November 14, 1979,  Federal Register as a "Level III" analysis.   The data
are adequate for  insuring  that the observed peak ozone value has, indeed, occurred
within or  downwind  of the  city being reviewed.  However, the data are not adequate
for tracing  the exact trajectory leading to the peak ozone value, nor for precisely

-------
Table 1-1.  Data Requirements for Use of City-Specific EKMA
     Data Item
                                                  Requirements
1.  Air Quality Data

    A.  Ozone Monitors
                                        - Minimum of 3 sites; one upwind
                                        the second on downwind edge of
                                        city and the third 15-40 km
                                        downwind.  (See EPA-450/2-77-C21a
                                        and Section 3.1.3)
    B.  Precursor Data
                                        - Minimum of 1 site with collocated
                                        NMOC and NO  monitors, with two
                                        sites deslrSble.  (See
                                        EPA-450/4-80-011 and Section 3.2.2)
2.   Meteorological Data

    A.   Upper Air Data
                                        - National Weather Service Rawlnsonde
                                        data 1f available.  (See Section 3.1.2
                                        and Appendix A)
        Surface Temperature and
        Pressure Data
                                        - Hourly surface temperature and
                                        pressure data at one site in the
                                        urban area.  (See Section 3.1.2)
        Surface Wind Data
                                        -Minimum of two sites; one in high
                                        precursor area and one additional
                                        site.   (See Section 3.2.1)
3.  Emission Data

    A.   VOC Emissions




    B.   NO.  Emissions
                                        -  A seasonally adjusted,  countywide,
                                        annual  emission rate  for  each
                                        broad source category.   (See
                                        Section 3.1.5)

                                        -  A seasonally adjusted,  countywide,
                                        annual  emission rate  for  each
                                        broad source category.   (See
                                        Section 3.1.5)

-------
estimating emissions  injected  into a particular column of air.  The procedures
described 1n this  document  therefore make assumptions about the trajectory of the
modeled column of  air.   The column is assumed to be within the urban area at
8 a.m. and to migrate downwind in a straight Hne to the site observing the peak
ozone concentration by  the  time the peak is observed.*  Although use of city-
specific input data should  improve model accuracy., the limited data base and the
resulting assumptions made  in  applying the city-specific EKMA (i.e.,, Level III
analysis) still make  it uncertain that the model will always perform well in an
absolute sense.  That is, one  would not necessarily expect to be able to measure
NMOC and NO  downtown and then use the isopleth diagram to predict peak ozone
           /\
concentrations.  However, sensitivity studies have shown that differences
resulting from incomplete input data or gross assumptions employed by the model
tend to exhibit proportional impacts on the positions of various ozone isopleths
on an isopleth diagram.     Thus, the model should perform satisfactorily when
applied in a relative sense (e.g., to estimate a change in peak ozone accom-
panying changes in local precursor levels).

     City-specific EKMA is  expected to be widely used in preparing State
Implementation Plans  (SIPs) submitted to attain the National Ambient Air Quality
Standard (NAAQS) for ozone.  The NAAQS for ozone is now in a statistical form
(i.e., daily maximum ozone  concentrations are not expected to exceed 0.12 ppm
more than once per year).   Because urban VOC and/or NO  emission reductions are
*  Because a review of St. Louis RAPS data has suggested that ordinarily the
   diurnal variation  in  urban NMOC/NO  ratio  is small, it is not regarded as
   critical that the  column of  air bexwith1n  the urban area at precisely 8 a.m.
   It should be sufficient if the air is within the urban area by mid-morning.
   If extensive wind  data suggest that the column of air is still upwind of the
   urban area at noon or later, problems of this nature need to be addressed
   on a case by case  basis 1n consultation with the EPA.

-------
likely to be more effective under some conditions than others, there 1s not
necessarily a simple 1:1 correspondence between high observed ozone concentrations
and estimated emission controls needed to attain the ozone NAAQS.  Thus, 1n order
to ensure that daily maximum ozone concentrations are not expected to exceed
0.12 ppm more than once per year, 1t 1s necessary to apply city-specific EKMA to
several different days observing high ozone concentrations.  This differs from
common past practices in which a single ozone design value was selected and
control requirements were calculated based solely on this value.  Chapter 2.0
describes the procedure, underlying rationale, and key issues which arise 1n
applying city-specific EKMA on several days.

     Data needs for applying city-specific EKMA on each day selected pursuant to
guidance in Chapter 2.0 have been summarized in Table 1-1.  Chapter 3.0 discusses
how these data are transformed Into input to city-specific EKMA.  The discussion
concerning each input variable follows a consistent format.  First, 1t 1s assumed
that valid data have been collected as specified in the Federal Register.  The
most appropriate procedure for transforming these data into each input variable
is identified.  However, it is also recognized that the available data will not
always conform exactly to the Federal Register specifications.  For example, in a
few cases, the data base may be more detailed than what is specified in the
Federal Register.   Procedures are described for utilizing such a data base if so
desired.   More often, some data may be missing or subsequently determined to be
invalid.   Procedures for deriving model inputs under such circumstances are also
identified.   In the event that quality assured NMOC and NO  data cannot be obtained
                                                          /\
for a sufficient number of days to provide a reliable estimate of the prevailing
NMOC/NOX  ratio, city-specific EKMA cannot be used.

-------
     Once all of the input has been prepared consistently with guidance in



Chapter 3.0, the OZIPP model is used to generate ozone Isopleths for each of the



days identified as necessary.  The EKMA procedure is then applied to each set of



isopleths to estimate the amount of VOC and/or NO  controls needed to reduce the
                                                 /\


peak hourly ozone concentration to 0.12 ppm on each day.   From these estimates,



the amount of VOC and/or N0v control 1s selected such that a peak hourly 0,
                           X                                              0


concentration of 0.12 ppm will not be expected to be exceeded on more than one



day per year.  Chapter 4,0 illustrates the application of city-specific EKMA to



estimate controls for each day.  The determination of the SIP control requirement



is described in Section 2.3.

-------
2.0  ESTIMATING CONTROL REQUIREMENTS FOR 1982 OZONE SIPs
     Prior to discussing procedures for estimating control requirements with
city-specific EKMA, 1t is appropriate to address the National Ambient Air Quality
Standard (NAAQS) for ozone.  The NAAQS affects the choice of ozone values Input
to EKMA, as well as the stringency of a city's calculated control requirements.

     The National Ambient Air Quality Standard for ozone 1s attained when the
expected number of days per calendar year, with maximum hourly average concen-
trations above 0.12 ppm, 1s less than or equal to one.   This differs from the
previous photochemical oxidant standard which specified a concentration "not to
be exceeded more than once per year."   As with the previous standard, air quality
data are examined on a site by site basis and each site must meet the standard.
The Guideline for the Interpretation of Ozone A1r Quality Standards recommends
that a period of three successive years of air quality data be used as the basis
                                           o
for determining attainment of the standard.   As more years of data are included,
a greater chance exists for minimizing the effects of an extreme year caused by
unusual meteorology.  However, extending the number of successive years too far
increases the risk of averaging data during a period 1n which a real shift in
emissions and air quality has occurred.

     Two important differences with past practices occur as a result of the new
standard.  First, only one hourly value is considered for each site on each day.
This can affect the set of candidate "ozone design values" used to estimate the
amount of controls needed to reach 0.12 ppm ozone.  Formerly, extremely rare
episode days with a number of very high ozone values at any given site were
weighted disproportionately.   The new standard inherently recognizes this problem
by allowing only one concentration (and therefore one control estimate) to be

-------
considered per site per day.
     Second, the phrase, "... expected number of days per calendar year..."
alleviates a concern with the former standard that those States which have main-
tained conscientious monitoring programs over the years are penalized.  Since the
expected or "average" number of days per year with daily maximum ozone concentra-
tions greater than 0.12 ppm is of concern, this means that there can be more than
one day observing ozone concentrations in excess of 0.12 ppm if the data base is
two or more ozone seasons in length.
     Another implication of the NAAQS is that the frequency distribution of ozone
concentrations at each monitoring site which occurs after the implementation of
controls is the key consideration in demonstrating attainment.   This has always
been the case.  In the past, however, very simplistic models (e.g., rollback,
envelope curves) were used to demonstrate attainment.  Minimal  use was made of
meteorological or air quality data.  Under such circumstances,  choosing the ozone
value to input into a model in order to calculate control requirements was very
straightforward.  The design value was simply the second highest value observed.
The degree of control needed to attain the NAAQS, however, is a function of many
things in addition to observed ozone concentrations during the  base period.  For
example, controls needed to attain the NAAQS are a function of pollution trans-
ported from upwind sources, prevailing NMOC/NOV ratios and atmospheric dilution.
                                              A
Therefore, it is conceivable that the second highest ozone design value would not
require the second highest control requirement to attain the NAAQS if one uses
city-specific EKMA or more sophisticated models.  Of paramount  interest is the
frequency distribution of control estimates calculated with such models.  Depend-
ing on the length of the period of record at any given monitoring station, one

-------
would choose the control estimates which would Insure that, on average, the dally
maximum hourly ozone concentration would not exceed 0.12 ppm more than once per
year at any monitoring site.  For example, if the period of record at one site
were three years, the fourth highest calculated control estimate would be chosen
as demonstrating attainment at that site.  The control requirement needed to
demonstrate attainment for the city as a whole 1s whatever 1s necessary to demon-
strate attainment at all ozone monitoring sites.  Therefore, 1t is necessary to
model a number of high ozone days to determine the degree of control required to
attain the ozone standard.   This need, while always present, can now be adequa-
tely addressed due to increased data availability.  Given this situation, demon-
stration of attainment could entail generating isopleths for (1) each site and
each day above 0.12 ppm ozone, or (2) for each set of meteorological conditions,
with associated probabilities of occurrence, that may result 1n violations of the
NAAQS.  Depending on the nature of the ozone problem for a given city, either
approach would require generating several isopleth diagrams.  Because 1t is
relatively easy to generate these diagrams and apply EKMA, procedures described
in the next Section recommend examining several days and monitoring sites without
very elaborate procedures to eliminate a number of candidate days prior to the
modeling.   Persons using more sophisticated modeling procedures would most likely
find 1t preferable to develop screening procedures to eliminate a number of
candidate days prior to the modeling exercise.
     Even  for city-specific EKMA, given the size of monitoring networks and the
expected number of observed ozone concentrations greater than 0.12 ppm in many
urban areas,  the computer costs and manpower requirements for the two approaches
mentioned  above might exceed the resources available to many State and local

-------
agencies.  Therefore, 1n developing a procedure for estimating urban area control

requirements, consideration has been given to resource requirements, model  com-

plexity and model input requirements, as well as to the form of the ozone standard.


2.1  SELECTION OF DAYS TO MODEL

     Control estimates obtained with city-specific EKMA can be sensitive to such
                                                             2
factors as the level of ozone transported into an urban area.   Thus, control

estimates based on the "classical design day" could understate the degree of

control required for the city to attain the NAAQS.  As a result, several days

should be selected for modeling.


     Recommended Procedure:  For each ozone monitoring site, the five (5) days
with the highest daily maximum ozone concentrations should be selected as candi-
dates for modeling.  In selecting the days for each site, only those days on
which ozone values are observed downwind or within the urban area should be
considered.  Determination of what is "downwind" on a given day requires the
review of prevailing wind data for that day.  Such a review may, in fact, reveal
that a nominally "downwind" site is in actuality upwind of the city on a parti-
cular day (see Figure 3-2 for an example of "upwind" versus downwind areas).  If
this is the case, the ozone value observed at that site should not be modeled.
The days should be chosen from the most recent three (3) years for which meas-
urements were made at the site.  By limiting the number of days to model, this
approach represents a compromise between a strict interpretation of the ozone
standard and consideration of model complexity and local agency resources.
However, the control estimates obtained with this procedure should not differ
significantly from those obtained by modeling each site and day above the ozone
standard.  EKMA, when used in this manner, serves as a "screening procedure" for
identifying the control strategy design day.  With more sophisticated models,
other (e.g., statistical) techniques may prove to be more viable for selecting
days to model for control strategy evaluations.

     Alternate Procedure:  Alternative attainment demonstrations might include
modeling each day in which the NAAQS was exceeded during the past three years, or
modeling a statistically determined design day or set of meteorological classes.
Probabilistic approaches, however, may be more difficult to develop and implement
than simply using EKMA itself as the screening procedure.


2.2  DAY SPECIFIC CONTROL ESTIMATES

     Once the days to be modeled ere selected, control estimates must be calculated
                                        10

-------
for each of these days.  Considerable duplication 1n high ozone days 1s likely

for many of the sites 1n the monitoring network.  When this occurs, the same

isopleth diagram can usually be used to make control estimates for a number of

sites.


     Recommended Procedure:  The OZIPP should be used to generate an Isopleth
diagram for each day Identified as described in Section 2.1.  Preparation of the
model input data 1s described in Section 3.1.  Development of a second Isopleth
diagram may be necessary if different control measures are implemented concur-
rently, e.g., reductions In future levels of ozone transported aloft.  Procedures
for handling this situation are presented in Chapter 4.

          Two additional pieces of information are needed to calculate control
estimates for each day and site:  the dally maximum ozone value at each site and
the urban 6-9 a.m. NMOC/NO  ratio.  Section 3.2.2 provides procedures for deter-
mining the appropriate ratto for use in city-specific EKMA.

          Having assembled the necessary data and generated an ozone Isopleth
diagram for each selected day and (if necessary) each site,* the VOC reduction
necessary to lower the peak ozone selected value to 0.12 ppm 1s calculated for
each day at each appropriate site.  The procedures for calculating control esti-
mates using the isopleth diagrams are described 1n Chapter 4.


2.3  SIP CONTROL REQUIREMENT

     The control requirement for use in the 1982 ozone SIPs 1s determined from

the control requirements estimated for each site.


     Recommended Procedure:  Given the form of the ozone standard, the SIP control
requirement Is that control estimate with frequency of occurrence of 1/365 for
the controlling site.   Thus, having modeled the five highest days at each site,
it is assumed that the upper portion cf the control requirement distribution has
been defined.   Table 2-1 indicates which control estimate would be selected as
the control requirement for each site.   The SIP control requirement is the highest
of the site-specific control requirements.
   Often, it will not be necessary to generate as many diagrams as implied by
   this procedure.  For example, if high ozone values occur at two or more
   sites on the same day, with rare exceptions, the same diagram can be used.
   One exception to this generalization is discussed in Section 3.1.5.
                                        11

-------
Table 2-1.  Tabular Estimation of Site-Specific Control  Requirements
                  Number of Complete    Data Point Used
                  Ozone Seasons         for Site-Specific
                  Monitored*	Control Requirements	.	

                         1              Second highest control  estimate

                         2              Third highest

                         3              Fourth highest
*  Monitoring data at a site is considered to be complete if valid daily
   maximum hourly concentrations exist for at least 60 days during the
   ozone season.  If a site does not meet this criterion during any of the
   last three ozone seasons, the second highest control estimate calculated at  a
   site is the site-specific control requirement.  In order for a season's daily
   maximum ozone values to be valid, there can be no systematic pattern of
   missing potential peak hours in the data base.
                                        12

-------
     Table 2-2 contains an example of the procedure for selecting  SIP control
requirements for urban areas.  In constructing this example,  it  is assumed that
the urban area has three ozone monitoring sites:  Site A within  the city,  Site B
in the predominantly downwind direction from the city, and Site  C  in the
predominantly upwind direction of the city.   Site A has two (2)  years of ozone
data, Site B has three (3) years of data, and Site C has only one  (1) year of
ozone data.
     The first step in calculating the SIP control requirement is  to list the
ozone maxima and dates of the five highest ozone days at each site.  A site 1s
only Included for a specific day if the wind data indicate that  the site is
actually downwind of the urban area.
     An examination of the high days at each site reveals some duplication in the
dates of high ozone days among sites.  For example, in Table  2.2,  June 8 appears
for both Site A and Site B.  For this case,  the same isopleth diagram can usually
be used at each site.  Control estimates are then made for each  site/day combi-
nation using the appropriate ozone isopleth diagram.
     Experience with the model, and an examination of the differences in such
factors as transported ozone levels between  the day to be modeled  and other days
with higher ozone maxima, may provide a means for limiting the number of sites or
the number of isopleth diagrams to be generated for each site.   In the example in
Table 2-2, because of the duplication of days among the sites, only two isopleth
diagrams were generated for the Site A, and  only one for Site C.   Then,  using  the
selection criteria presented in Table 2-1, the site-specific  control  requirements
were obtained (the control estimates enclosed in boxes in Table  2-2).   In  Table 2-2
                                        13

-------
Table 2-2.  Selecting the SIP Control Requirement:  An Example
Site A
P/
.24
.18
.16
.15
.14

Date
10/1/80
6/8/81
10/2/80
8/13/81
8/21/80

% R
60
57
(45)
39
*
                              Site B

                               %+
                              .22

                              .19

                              .19

                              .18

                              .17
                        Date     % R

                        6/8/81     58

                        8/25/80   54

                        10/2/80   55

                        8/26/80  (51)

                        7/26/79   49
Site C**

 ik+    Date      % R

.16    77/31/81     45

.15     9/17/81    (38)

.15     6/1/81      *
     SIP Control Requirement = 51% obtained from Site B.
  *  Isopleth diagrams were not generated for these days since transport and
     mixing heights were similar to days with significantly higher ozone
     values; thus, the control requirements would be less.

 **  On these days, the Site C was actually downwind of the city as indicated
     by local wind data.

  +  The daily maximum Q3 at each site.
Assumptions:
(1)   Urban Area has three monitoring sites,  Site A,  Site B
     and Site C.

 (2)  Site A has two years of ozone data,  Site  B has  three
     years of data, and Site C  has only one  year of  ozone
     data.

 (3)   Control requirements (%R) may not decrease monotonically
      due to daily variations in transported ozone levels,
      mixing heights,  NMOC/NOV  ratios,  etc.
                             X
                                        14

-------
the SIP control requirement 1s the maximum of the site specific control require-
ments, I.e., the 51 percent estimate Site B for this example.
     In summary, the method for estimating urban area control requirements for
1982 SIPs involves (1) the repeated application of the OZIPP model to generate
Isopleth diagrams for each of the five highest ozone days from the past three
years at each site, (2) the use of the EKKA procedure to provide a control esti-
mate for each site-day, and (3) the selection of the SIP control requirement as
the control  estimate needed to demonstrate attainment at all monitoring sites for
the urban area as a whole.
                                        15

-------
3.0  GENERATION OF DAY-SPECIFIC ISQPLETH DIAGRAMS
     As described in Section 2.1, a number of days with elevated ozone concen-
trations are selected for modeling.  An ozone isopleth diagram 1s generated for
each day using the OZIPP computer model.  Each model application requires the
use of day-specific modeling Inputs.  Day-specific empirical data (or suitable
surrogates} are then used in conjunction with each diagram to calculate the VOC
and/or NO  controls necessary to reduce the peak hourly ozone concentrations
         ^
observed on each day to 0.12 ppm.  The empirical data are used to locate a
starting point on each diagram, from which control calculations are made.
Because the scales of the diagram axes can be varied 1n OZIPP, the empirical
data must be developed prior to generation of a diagram to Insure that the
starting point 1s located in an appropriate position on the d1agra».  This
Chapter describes the methodologies appropriate for deriving the day-specific
•edeling da£t. the day-spec If ft empirical data, and the procedures for generatta?
the diagrams.  The control calculations are described in Chapter 4.

     In describing the techniques for developing modeling Inputs, more than one
procedure is usually presented.  The first method listed is generally preferable,
given the level of detail of information normally available.  However, 1t 1s
recognized that 1n some Instances more detailed information may be available and
could be Included 1n the analysis, 1f desired.  On the other hand, Information
may not always be available, or reliable.  Consequently, optional methods of
developing Input data are also presented.  If no information at all 1s available,
default values, or methods, are sometimes included.  In practice, the methods
used to derive Input to the model Involve a trade-off between the resources
required to collect and process data, and the enhanced credibility of the analysis
                                        16

-------
 by more completely  representing physical processes.  The preferred approaches
 are designed  to  provide a fairly complete representation, given data assumed to
 be commonly available.  To insure specificity to a particular city under review,
 city-specific determinations of Input data are strongly recommended, with default
 measures  recommended only as a last resort.
 3.1  OZIPP MODELING DATA
     The  conceptual basis for the OZIPP/EKMA model has been described References 1,
 2 and 3.  The following discussion focuses on the procedures for deriving the
 input data for using OZIPP to generate isopleth diagrams for each day investigated.

     As shown in Figure 3.1, ozone isopleths are expressed as explicit functions
 of initial NMOC and NO  concentrations.  However, the positioning of the ozone
                      j\
 isopleths on  the diagram is also an implicit function of meteorology, pollutant
 transport, emissions, and chemistry.  In the OZIPP model, several variables can
 be input  to the model to.tailor the diagram to a specific situation, I.e., the
 atmospheric conditions occurring on that day.  The OZIPP input variables for
 developing day specific diagrams can be grouped according to the following
 categories:
          1)  light intensity
          2)  dilution
          3)  0., transport
          4)  precursor transport
          5)  emissions
          6)  reactivity
 In the discussion that follows, procedures for deriving the input variables in
each group are described.
                                        17

-------
                                             1.2   1.4   1.5    1.9
                                                                      2.0
Figure 3-1.  Example Ozone Isopleth Diagram
                                       18

-------
     3.1.1  Light  Intensity
          The OZIPP program  utilizes latitude, longitude, day of the year, and
time zone to calculate  the appropriate variation of photolytic rate constants.
The model does not provide for adjusting the attenuation of light Intensity due
to cloud cover.  Hence,  conditions without significant cloud cover are assumed,
corresponding to the  types of conditions normally found on days with high ozone
concentrations.  Studies have shown that control requirements are relatively
insensitive to variations in light intensity, so this assumption is not perceived
as critical.

          Recommended Procedure:  The recommended approach for specifying light
intensity is straightforward.The local latitude and longitude for the center
of the city should be used.  The day of the year and the appropriate time zone
should be specified, as  described in EPA 600/8-78-014a.3
     3.1.2  Dilution
          In OZIPP, dilution 1s assumed to occur as a result of the rise in
atmospheric mixing height which normally occurs between early morning and mid-
afternoon.  The mixing height is the top of a surface-based layer of air which
is well-mixed due  to mechanical and thermal turbulence.  The mixing height rise
is governed by the specification of four variables Input to OZIPP:  the 0800
local  civil  time (LCT)* mixing height, the maximum mixing height, the time at
which the rise begins in the model (e.g., 0800 LCT), and the time at which the
mixing height reaches its maximum.
* In most areas 1n the United States, local civil time is Identical to local
  daylight time during the ozone season.  In areas which do not switch to
  daylight savings time, local civil time is the same as local standard time.
  The OZIPP program was designed to operate on the basis of Local Daylight
  Savings Time.  For those areas which do not switch to daylight savings time,
  the model may be made to operate on standard time by adding one (1) hour to
  the time zone described in Section 3.1.1.  See also page 39 of Reference 3.
                                       19

-------
          Recommended Procedure:  The recommended approach for estimating the
0800 LCT mixing height and the maximum mixing height uses temperature sounding
data routinely compiled by the National Weather Service.  Supplemental surface
temperature and pressure data are also needed.  Table 3-1 contains guidelines
for selecting the required data, while the detailed procedures for calculating
the morning and maximum mixing heights are contained 1n Appendix A.  A minimum
0800 LCT mixing height of 250 meters above ground level (AGL) is recommended.
This value is based on studies of the St. Louis and Philadelphia early morning
urban mixed-layer,11 and is recommended even 1f the morning mixing height com-
puted using the NWS sounding 1s lower.  The 250 m minimum is due to mechanical
turbulence caused by Increased surface roughness in the urban area.  Also, any
maximum mixing height greater than twice the cllmatological value contained in
Table A-l of Appendix A or Reference 12 should be checked as described in
Appendix A.  If no alternate data are available in this case, twice the clima-
tological value is recommended 1n place of the computed value.

          Alternate Procedure:  Direct measurements of the urban vertical tem-
perature profile may be used in place of the NWS data if they are available.
Mixing heights can be found from local urban radiosonde measurements, helicopter
soundings, or by sodar (I.e., acoustic radar).  Identical computational pro-
cedures as those described in Section A-2 of Appendix A should be employed.  As
for the recommended procedure, a minimum value of 250 meters should be used if
the computed morning mixing height is less than 250 meters.

          Default Procedure:  If none of the information needed for the first
two procedures is available, then the climatological values listed in Table A-l,
or obtained from Reference 12 may be used.  If the latter is used, the average
summertime maximum mixing height on non-precipitation days should be used.  For
the default procedure, the mixing height rise may be assumed to continue from
250 meters at 0800 LCT until the time of the cllmatological maximum daily surface
temperature.   If the latter is unknown, then a climatological value of 1500 LOT
(i.e., 1400 LST) can be assumed.


     3.1.3  Transport of Ozone

          The transport of ozone and its precursors into and from urban areas

has been reported by numerous investigators. '       The discussion in this

section is limited to the transport of ozone; precursor transport is discussed

in the following section.


          The two primary mechanisms by which ozone 1s transported into an urban

area are:
                                        20

-------
Table 3-1.  Data Requirements For Calculating Mixing Heights According
            to the Recommended Procedure
     Surface Temperature Measurements  -  For each day modeled, the urban
     surface temperature at 0800 LCT and the maximum temperature occurring
     between 0800 and 1800 LCT are required.  The most suitable data are those
     taken from a well ventilated shelter at a site near the center of the urban
     area.
     Surface Atmospheric Pressure Data  -  For each day modeled,  the urban
     surface barometric pressures for the same time and location  of the
     temperature measurements are needed.  If these are not available,  local
     National Weather Service (NWS) or Federal Aviation Administration  (FAA)
     pressure data may be used.
     NWS Radiosonde Data  -  For each day modeled, mixing heights are
     calculated from vertical  temperature profiles obtained from radiosondes
     taken by the NWS at several sites throughout the United States.   Appendix A
     contains information on selecting an appropriate site and obtaining the
     necessary data.
                                        21

-------
               1)  advection of ozone along the earth's surface, and

               2)  advection of ozone aloft, typically at night and during early
morning hours above the ground-based mixed layer, with downward mixing when the
mixing layer increases later in the day.

Ozone transported at the surface is subject to surface reactions and scavenging
by other species (e.g., NO) emitted during the night.  As a result of nighttime
atmospheric stability, ozone transported aloft does not come into contact with
scavengers emitted during the night.  Thus, overnight advection of ozone aloft
appears to be the more significant mechanism of transport from one urban area to
another.1'13

          Control strategies designed to attain and/or maintain the ozone standard
in individual urban areas must take into consideration the impact of transported
ozone on peak afternoon concentrations.  The procedure of using OZIPP with the
EKMA technique for the case where future transported ozone concentrations might
be significantly different from current levels is explained fully in References T
and 3, and described in Chapter 4 of this report.

          A.   Present Transport of Ozone at the Surface
               The chief Impact of ozone transport near the surface is expected
to be the more rapid conversion of NO to N02«  When Incoming ozone near, the
ground was varied using OZIPP, the impact on maximum ozone concentrations was
generally found to be negligible.  This finding held true for incoming concen-
trations as high as 0.12 ppm ozone.   In addition, several field studies have
shown that ozone transported along the surface tends to be minimal.  '*
                                        22

-------
                Recommended Procedure:   Based on the previous discussion,  1t 1s
 recommended that, for most situations,  the value for present ozone transported at
 the surface be set equal  to zero for each day modeled.
                Alternate  Procedure:   If the situation arises,  as a result of the
 geographical  location of  the city under study, that consideration of surface
 ozone transport is desirable, then dally estimates of ozone surface transport may
 be obtained from urban monitoring sites.   It is suggested that the 6-9  a.m.  LCT
 average ozone concentration at an urban site(s) be used as the estimate of the
 concentration of ozone transported Into the urban area along the surface  for the
 given day.*

           B.    Present Transport of  Ozone Aloft
                As noted above,  1t appears that unscavenged ozone transported
 aloft is likely to be the component  of  transport having the greatest Impact on
 maximum afternoon ozone levels  observed within or downwind from cities.   Thus,
 daily estimates of ozone  aloft  are needed for control  strategy development with
 OZIPP/EKMA.   Techniques for estimating  the level  of ozone transported aloft have
 been  the subject of two studies.13'18   Five different  techniques,  which were
 considered  to be feasible were  field tested in Philadelphia  during the  summer of
 1978.     The  five methods are:   1) use  of fixed ground  based stations; 2)  use of
 airborne measurements  in  a dedicated aircraft;  3)  use of  airborne  measurements
 with  a  portable instrument package;  4)  use  of free  11ft balloon  soundings; and 5)
 use of  soundings  by tethered  balloon.   Reference 13 contains a detailed description
 of each  of these  techniques as well  as  a  discussion of the findings of the study.
 Of the  five measurement techniques evaluated,  surface measurements at fixed
 sites, airborne measurements by dedicated instrumented aircraft and soundings by
ozonesonde beneath  a free  balloon were judged  to be practical means of providing
 information on ozone transported aloft.
* See related footnote on page 39
                                        23

-------
                Recoimiended Procedure:   In  selecting the technique to use, consid-
eration was  given  to such factors  as  technical capability and available funding,
and  the intended use of the data.   Ozone measurements taken on the day being
modeled are  recoil-mended as the best estimate  of ozone aloft.  These measurements
should be obtained at surface monitoring sites upwind of the city during the
first hour after breakup of the nocturnal  inversion.  An acoustic radar (sodar)
can  be used  to  determine the time  of  inversion breakup for the day.  If the time
of the breakup  of  the nocturnal radiative  Inversion is not known, the 1000-1200
Local Standard  Time average ozone  concentration recorded at the upwind monitor
should be used  as  the transport estimate.  A major advantage of surface measure-
ments is that 1t is the only method which  allows continuous measurements, and
thus assurance  that measurements exist  for days or for times of day which are
later determined to be of Interest.   The s1te(s) should be located in as rural a
location as  possible so as not to  be  appreciably affected by local sources of
precursors.  The distance such upwind sites should be located from a city depends
on the extent of urban development.   Because  it is desirable not to measure
pollutants which are redrculated  from  the city under review, a distance of
40 km or more upwind from the urban core should be sufficient.  This distance
perhaps could be reduced for smaller cities.  Figure 3-2 depicts orientations
for  acceptable  upwind sites.

                Alternate Procedure;   Information on the vertical distribution of
ozone transported  above the surface layer  in  the early morning may be used
directly if  it  is  available.   Such information might include aircraft or free-
lift ozonesonde measurements.   The reader  is  referred to Reference 13 for a
detailed discussion of these techniques.

                Missing Data:   In the event that an estimate of transport is not
available for a given day being modeled, the median transport value from the
remaining days  being modeled should be  used as a default value.  This procedure
applies to all  the  data gathering  techniques  described above; fixed site, air-
craft and ozonesonde observations.


     3.1.4   Precursor Transport

          Just  as  for ozone,  precursor  pollutants could be transported in both

the surface  layer  and aloft.   However,  consideration of precursor transport is

usually not  considered essential for several reasons.   First, transported pre-

cursor concentrations  tend  to  be substantially less than concentrations within
             2
urban areas.     Secondly,  in most areas  transported organic species are likely to

be less reactive than  urban  species because of the rapid reactions of the more

reactive species.
                                        24

-------
                                                   (B)
                                                       STAGNATION
              Denotes upwind area
Figure 3-2.  Examples of acceptable monitoring locations for estimating
             transported ozone.
                                        25

-------
          Reconmended Procedure;  Several studies have suggested that signifi-
cant transport of precursors  Into an urban area 1s not an Important consideration
in most cases.2,13,18  Consequently, transported concentrations of NMOC and NOX,
in both the surface  layer  and aloft, are recommended to be set to zero.
                                                                   2
          Sensitivity studies presented 1n Appendix B and elsewhere  Indicate
that, for a wide range of  assumed levels of transported NMOC, there 1s not an
appreciable impact on estimated VOC control requirements using city-specific
EKMA.  Further, consideration of transported precursors requires collection of
additional data.  However,  if 1t is still desired to include consideration of
transported precursors in  the analysis, the following alternative procedure may
be used.

          Alternate  Procedure:  Utilize the methodology presented in Appendix B
to estimate local emission  control requirements for a city subject to significant
and pervasive precursor transport from upwind sources.

     3.1.5  Post 8 a.m. Emissions
          As was previously described, the OZIPP/EKMA model is similar in concept
to a Lagrangian photochemical  dispersion model in that ozone and precursor
concentrations within a well  mixed column of air are modeled as the column
traverses the city.  The column of well-mixed air is assumed to originate 1n the
urban core, and begin moving  at 0800 LCT toward the site of the peak ozone
concentration.  Thus, the  emissions occurring subsequent to 0800 LCT are deter-
mined by the space-time track of the column.  Within OZIPP, emissions data are
input to the model for each hour after 0800 LCT, and are expressed relative to
the initial concentrations.   Thus, there are two basic, and separate, problems
in deriving the necessary  emissions Information:

               1)  determining emissions along the space-time track of the theo-
retical column^
                                        26

-------
               2)  expressing the emissions derived in 1) relative to the initial
concentrations.  These initial concentrations are assumed to be the result of
local urban emissions occurring prior to 8 a.m.
The first of these involves the use of an emissions inventory to estimate the
emissions occurring along the assumed path.  The second involves the calculation
of relative emission fractions that are directly input to the OZIPP model.  Each
of these is discussed separately below.
          A.   Preparation of Emissions Data
               The first step in preparing the available emissions data for use
in OZIPP/EKMA is to specify the space-time track of the column trajectory.  The
column is assumed to begin moving from the urban core at 0800 LCT directly
towards the site of the peak ozone concentration.  The rate of movement is
estimated using the time and location of the peak ozone concentration.   The
column is assumed to move at uniform speed (throughout the day) to the site of
the peak ozone value such that it arrives at that site at the time of the
observed peak.*  For example,  assume the observed peak one-hour-average ozone
concentration on a particular day was measured between 1400-1500 LCT at a site
35 kilometers downwind of the urban core.   The column speed would be calculated
*  Recall from Chapter 2 that some duplication in high ozone days is likely for
   many of the sites in the monitoring network.  A rigorous modeling approach
   would involve specifying trajectories for each site/day combination,  and
   generating the corresponding diagrams.   However, in most instances,  the
   trajectory leading to the peak ozone observed on that day should be  a suf-
   ficient approximation of the other trajectories.  Hence, only one diagram can
   be developed for the day and be used for all sites.  The exception would be
   cases in which significantly different  post 8 a.m.  emissions occur along the
   different trajectories between the urban core and the different monitoring
   sites.  For this situation,  separate site/day diagrams need to be developed.
                                        27

-------
by dividing 35  kilometers by  seven hours (i.e., the elapsed time between Q800 LCT
and 1500 LCT).  The  location  of the column at any particular hour would be
determined by assuming  the column -moved in a straight line between the urban
core and the site of the peak at the uniform speed.  (For the example, the speed
would be 5 km/hour.)  The assumed trajectory may not be identical to the actual
trajectory corresponding to highest observed ozone concentrations.  However, the
wind data should be  checked to assure the modeler that the peak ozone concentra-
tion observed at a monitor station is, in fact, downwind from the city.  Observa-
tion of urban NMOC/NO   ratios made during the St. Louis RAPS do not suggest the
                     .A
prevailing urban ratio  varies greatly during the morning.  Further, several
hours are required for  ozone  to reach its peak.  Hence, the trajectory described
above should be an acceptable approximation, provided the column of air passes
over the city by mid-morning  (e.g., 1000-1100 LCT).
               With  the trajectory path identified, the next step in preparing
the emissions data is to use  the emission inventory to calculate the hourly
emissions.  The procedure depends to some degree on the spatial and temporal
resolution of the emission inventory.  In the recommended methodology that
follows, a seasonally adjusted, annual county-wide emissions inventory of
reactive VOC* and NOV is assumed to be available.  This level of detail will
                    J\
result using procedures presented in Reference 19.  An alternate method addresses
procedures that can  be  used with inventories of greater temporal and/or spatial
*  In the discussion of post 8 a.m. emissions  in Section 3.1.5, the term "VOC,"
   volatile organic compounds, refers to the sum of reactive organic emissions
   included in emission inventories.  The term "NMOC," non-methane organic  .
   compounds, refers to ambient measurements of all organic compounds other than
   methane.
                                       28

-------
resolution.  Regardless of the resolution of the inventoryt the goal of this
step  is to develop emission densities representative of the area over which the
column 1s assumed to pass each hour.  Thus, the emission density represents the
average emission density "as seen by the column" for each hour between the
simulation start and the time of the measured peak.
               Recommended Procedure:  For this approach, a county-wide emissions
inventory 1s assumed; I.e., total reactive VOC and NO  emissions on a seasonally
adjusted annual basis are available at the county level only.  The simplest
method of deriving the necessary Information 1s to compute the county-wide
emission density by summing total emissions (point plus area source) and dividing
by the area of the county.  Hourly resolution of the Inventory can be estimated
by assuming emissions occur uniformly throughout the year.  For example, one
would divide the seasonally adjusted annual emissions by 8760 hrs/yr.  The
emission density at any hour 1s thus determined by the county in which the
trajectory path is located for that hour.  For those hours 1n which the trajec-
tory  path crosses a county boundary(s), the county emission densities may simply
be averaged.

               The procedure just described can best be illustrated by example
(see  Figure 3-3).  Part A of the figure shows the information typically available.
Here  it assumed that the peak ozone concentration measured on the example day
occurred at a site 35 kilometers downwind of the urban area between 1400 and
1500 LCT.  The first step is to depict the trajectory path on an appropriate map
by marking off hourly segments (see Part B).  Step 2 Involves deriving the
county-wide emission densities on an hourly basis (Part C).  Finally, in Step 3,
the sequence of emissions is determined by using the emissions for the county in
which each hourly segment lies.   For those cases 1n which a segment lies in two
counties, the emission densities of the two counties were averaged.  Part D
shows the results of Step 3.

               The approach just described 1s a relatively simple one which 1s
consistent with the amount of information typically available.  However, as
                                        29

-------
A)  Information Available
            Peak ozone measured
            between 1400-1500 LCT.
            Site is 35km downwind
            of center-city (cc).
                                                                        County Emissions Data
                                                                                   Adjusted         Adjusted
                                                                                     VOC              NOX
                                                                                   Emissions        Emissions
                                                                     Area,km2        kg/yr            kg/yr
                                                                                   5.45x1O7         2.96x1O7
                                                                                   3.93x1O7         3.38x1O7
                                                                                   1.68xl06         1.17xl06
B)  STEP 1:  Determine Trajectory Segments
                                             A straight line is drawn between the center-city and the downwind
                                             site.  The line is divided into seven equal  segments, representing
                                             the number of hours between 800 and 1500 LCT.
C)  STEP 2:  Calculate, hourly, county-wide emission densities
                     Area, km2
                         150
                         500
                         300
                Adjusted
              VOC Emissions
                  kg/yr
                 5.45x1O7
                 3.93x1O7
                 1.68xl06
                  Adjusted
                VOC Emission*
                density, kg/hr km2
                       41.5
                        9.0
                         .6
                 Adjusted
              NO  Emissions
                xkg/yr
                  2.96x1O7
                  3.38x1O7
                  l.UxlO6
          Adjusted
       NOV Emission*
       defisity. kg/hr km2
             22.5
              7.7
               .4
                    *Emission density = emissions/(8760 x area)
D)  STEP 3:  Specify Sequence of Emission Densities
          Hour
           1
           2
           3
           4
           5
           6
           7
Time. LCT
    8-9
    9-10
    10-11
    11-12
    12-13
    13-14
    14-15
Trajectory Segment
Location (County)
        A
        A/B
        B
        B
        B/C
        C
  Adjusted
VOC Emission
density, kg/km2 hr
        41.5
        25.3
         9.0
         9.0
         9.0
         4.8
          .6
   Adjusted
NO  Emission
density, kg/km2 hr
        22.5
        15.1
         7.7
         7.7
         7.7
         4.1
          .4
 Figure 3-3.   Example Formulation of Post 8 a.m.  Emission Densities Scheme
                                                      30

-------
described below, more sophisticated techniques may sometimes be desired.  It

should be remembered, however, that city-specific EKMA (Level III analysis) does

not require a precise definition of an air parcel trajectory.  Use of spatially

and temporally detailed inventories implies a need for greater efforts to define

trajectories.  Such efforts are more consistent with the use of more sophisticated

models.  Although the use of very detailed emission Inventories 1s inconsistent

with the degree of sophistication inherent in city-specific EKMA, methods for

improving spatial and temporal resolution are possible.  These are described 1n

the following paragraphs.


               Alternate Procedure;  In the recommended approach, emission
densities were determined on a county-wide basis.  In some Instances, a finer
spatial resolution may be desired.  For example, suppose total emissions are
available for the county, but the county encompasses a very large land area, of
which the urban area is a relatively small part.  In this case, one may wish to
suballocate the emissions to areas smaller than the county"levels.  For example,
the very large county might be divided into a rural portion and an urban portion,
with emissions apportioned to each area.  The methodology for apportioning will
usually depend on available data and the particular situation encountered.
Consequently, no specific guidance can be given here, but the concepts contained
in References 19 and 20 may be adopted.  If an allocation 1s performed to a sub-
county level, the sub-county areas should not be resolved to less than 100 square
kilometers.  Resolution to smaller areas is not recommended because of the
inability of the model to consider horizontal gradients in pollutant concentra-
tions.

               Alternate Procedure;  A second way in which the recommended
approach could be made more precise is more detailed consideration of the temporal
distribution of emissions.  In the recommended approach, hour by hour differences
in emissions are not considered.  If more detailed information is available,
these data could be employed to derive more precise temporal resolution of
emissions as described in Reference 20.  In the absence of geographically specific
data, a city-wide average may generally be used for distributing temporal emission
patterns.  For example, the distribution could be calculated as the average of
the temporal distribution of mobile and stationary sources.

               Alternate Procedure:  In some cases, emissions inventory data
will be available that Is of a finer resolution than a county-wlde/annual basis.
For example, an hourly gridded emissions inventory may be available.  This
procedure addresses the problem of using an hourly emissions inventory with
grids finer than 10x10 kilometers on a side.  However, the development of a
gridded inventory specifically for use with the OZIPP/EKMA technique is not
recommended.


                                        31

-------
               Recall that the model underlying the OZIPP/EKMA procedure does
not consider the existence of horizontal concentration gradients.  The emission
density "seen by the column" during any hour should represent an average.  Con-
sequently, 1t is recommended that emissions inventory data resolved to grids
smaller than 10x10 kilometers Ci.e., 100 km2) be aggregated to a minimum size of
100 square kilometer grids.  The larger grid squares conform more readily to the
assumptions underlying the model.  The larger grid cells should be located such
that the average emission density is representative of the area it covers.
Thus, for example, a 10x10 km grid square might be centered on the center city,
and the new grid network for OZIPP/EKMA developed from there.
               The procedure for determining the emissions encountered by the
column of air simulated with OZ1PP is similar to the one described for the
recommended methodology.  However, in this case the network with the larger grid
squares is used 1n place of the county-wide data.  The resulting emissions
schedule 1s established by the location of the trajectory segments with respect
to the grid cells.  If a trajectory segment crosses a grid boundary, the emission
densities of the adjacent grids may be averaged.

          B.   Derivation of Emission Fractions
               In Section A, the procedure for estimating the sequence of emis-
sions was described.  The emission sequence represents emissions injected into
the modeled column of air each hour subsequent to the 0800 LCT simulation starting
time.  As indicated in the example problem (see Figure 3-3), the emission densities
comprising the sequence of emissions are expressed 1n terms of mass per unit
area (e.g., kg/km2).  In OZIPP, however, the emissions must be expressed relative
to initial concentrations.  Thus, it is necessary to translate the sequence of
emissions (in mass units) to fractions of initial concentrations.  This is
accomplished by first deriving an initial emission density which is tied to the
initial concentration.  The emission fractions are then computed by dividing the
hourly post 8 a.m. emission densities (i.e., from the sequence of emissions) by
the initial emission density.  In essence, this procedure is the same as calcu-
lating the concentrations that would be generated by the emissions within the
imaginary column of air after one hour, and then dividing those computed
                                        32

-------
concentrations by the initial concentration to obtain the appropriate fractions.

               The approach recommended for computing emission fractions 1s one
in which an initial emission density 1s computed from Initial conditions.  The
initial emission density is calculated as the emission density necessary to
generate in one hour the initial concentrations observed within the column.  So,
if an initially empty column were "exposed" to the calculated initial emission
density for one hour, the concentrations of precursors in the column at the end
of the hour would equal the observed Initial level.  The emission density for
any hour after 0800 LCT can then be related to the Initial concentrations by
means of the initial emission density.   For example, 1f the emission density for
hour 1 (i.e., 0800-0900 LCT) is one-third that of the initial emission density,
then one-third of the initial concentration would be generated by the emissions
for that hour.  The emission fraction can therefore be calculated using the
following equations:

               Q0=aC0Ho                                      (3-1)
             and
                    Qi
               E1 = ui                                           (3-2)

             where
               QQ = calculated initial  eirisslon density (kg/km2)
               CQ = initial  precursor concentration (ppm)
               Hp =  initial  mixing  height (kilometers)
               Q.J  =  emission density for  hour  i  from the  sequence of emissions
                    (kg/km2)
               E.J  =  emission fraction for hour i
                                        33

-------
                a  =  conversion  factor for converting from volumetric to

                     mass  units


The methodology 1s further  Illustrated 1n Figure 3-4, and the procedures for

deriving the necessary data are described below.


               Recommended  Procedure:  The first step in calculating the emission
fractions is to compute an  initial emission density for MO  and VOC.  This
requires the specification  of an  initial concentration for each precursor and the
initial mixing height.  The initial concentrations can be obtained from ambient
6-9 a.m. measurements of  NMOC and NO  on the specific day being modeled.  If NMOC
or NO  ambient data  for the specific day being modeled are unavailable, then the
median 6-9 a.m. NMOC and  NO concentration for all days being modeled should be
used for each precursor.

               The computational  steps necessary to calculate the emission
fractions are outlined in Table 3-2 and illustrated in Figure 3-5.  A brief
explanation of each  step  follows:

                     Step  1:   The initial concentration to be used in the emission
                              fraction calculations is the 6-9 a.m. average pre-
                              cursor concentration 1n the urban area.  (Average
                              concentrations are calculated for both NMOC and
                              NO  .)  The procedures for calculating the 6-9 a.m.
                              average concentrations at individual monitors are
                              described in Section 3.2.2, and should be followed
                              here.  If more than one monitor 1s located in the
                              urban area, the 6-9 a.m. concentrations at each
                              monitor should be averaged to obtain an the overall
                              urban average 6-9 a.m. concentration.

                     Step  2:   An  Initial emission density 1s calculated for VOC
                              and NO , individually.  The recommended conversion
                              factors (a) for VOC and NO  are 595 kg HC/ppmC km3
                              and 1890 kg NO /ppm km3, respectively.  The deri-
                              vation of the conversion factors is described in
                              Table 3-2.  For converting NMOC to VOC, it is
                              assumed that one ppmC represents a molecular
                              weight equivalent to CHa-s9.  For NO , a molecular
                              weight of 46 is assumed because inventories list
                              NOY as equivalent N02.
                                A

                     Step  3:   The hourly emission fractions are calculated by
                              dividing the emission densities for each hour by
                              the initial emission density.  This 1s done sepa-
                              rately for each pollutant.
                                        34

-------
Urban Area
                              Time:   0800
                                                   C  = Column concentration at
                                                    0   0800 LCT
                                                   H  = Mixing Height at 0800  LCT
                    Assume the precursor concentration in the
                    column is CQ.   Therefore, QQ " C  H
                              Time:   0800 -  0900
Urban Area
             Assume the emission density for the hour 0800-0900 equals 1/4
             of the emission density calculated from the initial  concen-
             tration-
             The concentration generated by the emissions occurring between
             0800 and 0900 would be 1/4 of the initial concentration-
             Therefore, the emission fraction for Hour 1  would be
             1/4, or .25:
                             '
                                        -  -
              Figure 3-4.  Illusiration of Emission Fractions Calculated
                          Using Initial Conditions.
                                        35

-------
Table 3-2.  Recommended  Procedure For Calculating Emission Fractions


Step 1:   Calculate  the  6-9  a.m. urban average NMOC and NO  concentrations for
          the day  being  modeled.  If data from more than one urban monitoring
          site are available,  average the Individual 6-9 a.m. levels.  If data
          for the  specific day being modeled are not available, then use the
          median of  the  6-9  a.m. averages for all days being modeled.


Step 2:   Calculate  the  initial emission density for VOC and for NO  :
                           Co Ho
               where
                    Q  =  Initial emission density, kg/km2

                    HQ =  Day-Specific mixing height at 0800 LCT (kilometers)

                     « =  Conversion factor*
Step 3:   Calculate  individual  hourly emission fractions:
               where

                    E.J = Hourly emission fraction for hour i

                    Q.. = Emission density for hour 1, kg/km2

                    QQ = Initial emission density calculated in Step 2 above,
                         kg/ km2
*  Conversion factors:  « For NMOC to VOC = 595 kg/km3 ppmC


                        « For NOV         = 1890 kg/km3 ppm
                                A
These conversion factors were calculated assuming a molar density of dry air at
25°C (i.e., 1 mole/24.4 liters).  For NMOC to VOC conversion, the molecular
equivalent to one ppmC is assumed to be CH2>5 (i.e., a molecular weight of 14.5
is assumed).9  For NO, a molecular weight of 46 1s assumed since NO  emission
inventories are complied as equivalent N02)-
                                        36

-------
A)  Available Information
    • Air Quality Data
         NMOC [=] ppmC-
         NOX  L=J  Ppm
Urban
Monitor
1

2

Pollutant
NMOC
NOX
NMOC
NOV
6-7
2.3
.250
1.7
.190
7-8
2.0
.200
1.7
.210
8-9
1.5
.180
1.6
.170
6-9 AVR.
1.9
.210
1.7
.190
    •  Emissions Schedule Shown
      In Figure 3-3
    •  0800 LCT Mixing Height = .25km

B)  Step 1:  Calculate Urban Average 6-9 LCT Concentration
         NOV
                     1.9 + 1.7
                                 = 1.8
                      .210+.190
.200
C)  Step 2:   Calculate Initial  Emission Densities

       For VOC
            Qo - a Co Ho • (595 k9/km3 ppraC)  (1.8  ppmC)  (.25 km) = 268 k9/km2
       For NOX
            Qo = «Co Ho = (1890 kg/km3 ppm)  (.200 ppm)  (.25 km) =   95kg/km2
D)  Step 3:   Calculate Hourly Emission Fractions

                                 VOC Hourly
                                       NQX Hourly

Hour
1
2
3
4
5
6
7

Time, LCT
8-9
9-10
10-11
11-12
12-13
13-14
14-15
Emission Density,*
kg/ km2
. 41.5
25.3
9.0
9.0
9.0
4.8
.6
NMOC Emission
Fraction**
.15
.09
.03
.03
.03
.02
.00
Emission Density* NOX Emission
kg/km2
22.5
15.1
7.7
7.7
7.7
4.1
.4
Fraction**
.24
.16
.08
.08
.08
.04
.00
      * Hourly Emission Densities From Figure  3-3
     ** Emission Fraction = (Emission Density)/(Initial Density)
  Figure 3-5.   Example Calculation of Emission  factions
                                                    37

-------
     3.1.6  Reactivity
          The OZIPP model incorporates a chemical kinetic mechanism to describe
the reactions taking place among pollutants.  The kinetics model  used 1n OZIPP
represents a detailed sequence of chemical reactions which has been proposed for
a mixture of propylene, n-butane and NO .  The chemical mechanism used 1n the
                                       J\
kinetics model is based on information obtained in smog chamber experiments with
                                  22  23
propylene and n-butane separately.        The kinetics model predictions were
matched against Bureau of Mines (BOM) smog chamber data obtained by irradiating
automobile exhaust.    Initial proportions of propylene and n-butane were then
adjusted so that consistently close agreement was obtained with observations in
the BOM chamber.  Of the available smog chamber studies, the ones using auto-
motive exhaust are thought to use a mix of reactants most representative of the
mixes found in urban atmospheres.

          In OZIPP, total NMOC is represented by the sum of propylene and
n-butane, with aldehydes related to the total "NMOC" concentration.  The reactivity
is determined by specification of three variables:  1) the fraction of NMOC that
is propylene; 2) the fraction of NMOC that is to be added as aldehydes; and 3)
the fraction of initial total NO  that is N0?.  The proportional  mix of propylene
                                A           £
and n-butane for which OZIPP yields peak ozone predictions comparable to that
observed with automobile exhaust in the BOM chamber is 25 percent and 75 percent,
respectively, with an additional 5 percent of the initial NMOC added as aldehydes.
Propylene-butane mixes have not been established for any atmospheric mix other
than the automotive mix used in the BOM chamber.  Unless satisfactory corre-
spondence between other atmospheric mixes and other propylene-butane mixes can be
established, the 25 percent propylene, 75 percent butane, plus 5 percent aldehydes,
should be used.
                                        38

-------
          The third variable to consider Is the fraction of the Initial N02 that

is NOX.  Sensitivity studies have suggested that the effect of the Initial N02/N0>

mix on peak ozone concentrations is small.    Consequently, specification of this

variable is not deemed critical.


          Recommended Procedure:  The 25 percent propylene-75 percent butane mix,
with 5 percent of the initial NMOC added as aldehydes, should be assumed for the
model simulations.  This mixture corresponds to the reactivity of automotive
exhaust irradiated in a smog chamber.  Unless satisfactory correspondence can be
established between other propylene-butane mixes and other atmospheric mixtures
which differ appreciably from automotive exhaust, the 75 percent/25 percent mix
should be used.

               A default value of 25 percent for the initial N02 to NO  ratio is
recommended.  Sensitivity studies have shown that peak ozone concentrations (and
hence control estimates) are relatively insensitive to any particular assumption.
However, day-specific information may be used if desired.  For this case, the 6-
9 a.m. average N02 should be divided by 6-9 a.m. average NO  to obtain an N02/N0
ratio at each monitor located in the central core of the uroan area.   If more
than one monitor is located in the urban area, then an average ratio may be
obtained by averaging the individual ratios.*


3.2  EMPIRICAL DATA

     Two pieces of empirical data are needed to establish a starting point on the

ozone isopleth diagram for calculating control requirements.  The first is the

maximum one-hour average ozone concentration observed at the site of interest.

The degree of emission control  necessary to reduce this "peak" to 0.12 ppm is to

be calculated; hence, the peak level will  be termed the daily design value.
   Strictly speaking, this latter procedure is more consistent with the use of
   measured 6-9 a.m. urban ozone concentrations to estimate ozone transported
   into the city within the morning mixed layer Csee page 23).   Howevei; use
   of the default value makes little difference 1n the resulting control  estimates.
                                        39

-------
     The second piece of information needed is the NMOC/NO^ ratio.  This ratio is
derived from the 6-9 a.m. concentrations of NMOC and NOV within the urban area.
                                                       y\
These measurements correspond conceptually to the NMOC and NO  concentrations on
                                                             j\
the axes of the isopleth diagrams.  The ratio will be termed the design ratio.
The procedures for deriving both the design ozone values and the design ratios
are described below.

     3.2.1  Dally Ozone Design Value
          The dally ozone design value is used 1n conjunction with the NMOC/NOX
ratio to establish a starting point on the Isopleth diagram for calculating
control estimates.  The "dally design value" 1s used to refer to the measured
maximum hourly ozone value for the day being modeled.*  For use with the isopleth
diagram, the design value should be expressed 1n ppm units rounded to two decimal
places.

          Recommended Procedure:  A daily maximum 1 hour value is obtained for
each site which is downwind of the city, and/or within the city in the case of
light and variable winds on the day for which the isopleth diagram is to be
developed.  Surface wind data should be examined to assure that the site is not
"upwind" of the city.  Based on the results of field studies and reviews in which
ozone gradients downwind from urban areas were examined, peak ozone concentrations
should generally be observed within 15-45 km downwind of the central business
district.2,25~2*  Specific siting criteria for the monitoring of photochemical
pollutants are discussed elsewhere.2,26"27

     3.2.2 NMOC/NOX Ratios
          The prevailing 6-9 a.m. LCT NMOC/NOV ratio measured in the urban core
                                             /\
of the city 1s the second piece of empirical data required to define the starting
point on the Isopleth diagram.  The design ratio is viewed as a characteristic of
*  Note that an ozone design value exists for each site for which the day being
   modeled 1s among the  five highest ozone days.
                                        40

-------
the city which would prevail during the remainder of the morning and early after-
noon in the absence of chemical reactions.  The ozone isopleth diagram expresses
peak ozone concentrations as a function of the initial concentration of NMOC and
NOV.  Thus, the 6-9 a.m. LCT NMCC/NOV ratio 1s considered to be the appropriate
  A                                 ./»
ratio for use with the Isopleth diagram since this ratio 1s consistent with the
                              28
conceptual basis of the model.    To ensure that representative ratios are obtained,
the NMOC and NOV instruments should be collocated in the central core of the
               A
urban area.  The slteCs) should be located in an area of relatively uniform
emission density and not significantly influenced by any individual source.  More
detailed guidance on siting NMOC instruments is contained 1n Reference 29.
Guidance on the operation of NMOC Instruments 1s available in Reference 30.

          Significant discrepancies have been found between NMOC/NO  ratios
                                                                   A
calculated on the basis of ambient measurements and those obtained from emission
               31
inventory data.    Reasons for the lack of correlation between the two ratio
calculation procedures have not been resolved.  As a result, only ambient NMOC/NO
                                                                                 A
ratios should be used with EKMA since these ratios are consistent with the con-
ceptual basis of the model and the emission ratios have been shown to be poor
surrogates for these ambient ratios.  To ensure maximum usefulness of the contin-
uous NMOC data, careful attention to calibration, operational procedures, and
quality assurance 1s necessary.  Presently available instruments are capable of
yielding acceptable data at concentrations above about O.S.ppmC 1f they are care-
fully maintained and calibrated.''

          The NMOC data are to be collected during the season of peak ozone
concentrations Csummerl.  Because NMOC concentrations are apt to be relatively
high in central urban locations at those times of the day (early morning) when
                                        41

-------
these measurements are required for use in EKMA more confidence can be placed in.

the estimate.  Data from a recent study of continuous NMOC Instruments suggest

that relative standard deviations of 30 percent are likely for well operated and

maintained instruments when concentrations are above 0.5 - 1.0 ppmC.  However,

because of the uncertainties in individual NMOC readings, the NMOC/NO¥ ratio
                                                                     J\

calculated at a single site for a single day is not recommended for use 1n city-

specific EKMA.  Considering FID Instrument reliability and model sensitivity, the

following procedure is recommended for calculating NMOC/NO  ratios.
                                                          /\


          Recommended Procedure:

          1.  Individual NMOC/NO  ratios at a site are calculated as the ratio of
the 6-9 a.m. LCT average NMOC ana NO  concentrations, i.e., the average of the
hourly concent-rations for hours 6-7, 7-8, and 8-9 LCT, respectively.  Ratios
should not be calculated for any day with less than two valid hours for either
NMOC or NOV.
          y\

          2.  If precursor measurements from more than one urban site are avail-
able for the same day as the ozone design value, and the individual ratios at
each site do not differ by more than 30 percent from the average ratio, then the
design ratio (DR) is the average of the individual 6-9 a.m. NMOC/NO  ratios.  If
data from only one site are available for a specific day, the design ratio for
that day should be calculated using the procedure described in (3) below.


                       n
               DR   =  z   (R.)     = E
                        =1
                          n
          where
               R, =  6-9 a.m. (NMOC/NOJ ratio at Site 1
                 I                     f\
               ff =  the average ratio
               n  =  the number of sites


          3.  If only one site measuring both NMOC and NO  is available, or if
the values from  different sites are missing or not comparable (I.e., some ratios
differ by more than 30% from the mean ratio), the following procedure is recom-
mended.  The ratio for use with the isopleth diagram should be calculated as the
median of the ratios observed on all those days being modeled vdth accompanying
NMOC and NO  data.
                                        42

-------
               DR  = median IR--}, for a single site
                              j
                 or
               DR  = median" IK.}, for multiple sites
          where
                j  = a high ozone day, and j = 1,2»...,N
When the monitor fs located In a relatively high and uniform emission density
area, good agreement has been found between ratios calculated on Individual high
ozone days and more robust measures..-1,2,^8  Thus, use of the tnedlan ratio 1s
viewed as providing a robust estimate of the dally design ratio which 1s consist-
ent with the conceptual basis of the model.
          4.  If no collocated pairs of NMOC and NO  monitors were operated 1n
the urban area during one or two of the years 1n the suggested three year period,
1t 1s recommended that the procedure tn (3) above be followed using the available
NMOC/NO  data to choose a median NMOC/NO  ratio.  In this case, the median ratio
would be used with ozone data from all years without the appropriate precursor
concentration data.

     Two examples of calculating NMOC/NO¥ ratios for use with EKMA/OZIPP are
                                        A
contained 1n Figure 3-6.

3.3  USE OF OZIPP TO GENERATE ISOPLETH DIAGRAMS
     An ozone Isopleth diagram 1s to be generated for each day Investigated.
Section 3.1 described the procedures for deriving the day-specific modeling data,
and Section 3.2 described the formulation of the empirical data.  This section
briefly addresses the procedures for producing a day-specific diagram using these
data.  The details are described in Reference 3, the OZIPP User's Manual.
     The modeling data are input to OZIPP by means of option cards.  Table 3-3
shows the appropriate option card to be used with each category of model  vari-
ables described in Section 3.1.   There are only two other important considerations.
First, the diagram must contain the starting point defined by the intersection of
the design NMOC/NOV ratio and the daily design ozone level.   This is accomplished
                  A
by selecting appropriate NMOC and NQ.. scales on the abscissa and ordinate,
                                    n
                                        43

-------
Example 1.


   Given:  Ratios at five urban core sites on the day being
          modeled are respectively 9.1, 6.2, 6.4, 6.5 and 9.8
   Find:   The NMOC/NO  ratio for use in EKMA
   Solution:  First calculate the average ratio

              F    9.1  + 6.2 + 6.4 + 6.5 + 9.8
              R =  -.         g
              R =  7.6

    Note that all the ratios are within +_ 305- of R",  i.e., all
    the ratios are.between 5.3 and  9.9.  Then,  the  design ratio
    is

              DR = R = 7.6
Example 2.
    Given:  Assume that only one site is available for the study.
            Assume also that the NMOC/NOX ratios are available for
            five of the design days.   These ratios are 8.8, 8.6,
            15.5, 9.7, and 14.3, respectively.
    Find:   The design ratio for use with all  the design days.


    Solution:  Since only one site is available, the design ratio
               is

                     DR = median {8.8, 8.6, 15.5, 9.7, 14.3}

                     DR = 9. 7
FIGURE 3-6.  Example Calculations of the Design NMOC/NC  Ratio
                                                       A


                             44

-------
Table 3-3.  OZIPP Options For Model Input Data
Input Data                    Section                  OZIPP Option
Light Intensity                3.1.1                   PLACE
Dilution                       3.1.2                   DILUTION
03 Transport                   3.1.3                   TRANSPORT
Precursor Transport            3.1.4                   TRANSPORT
Post-8:00 a.m. Emissions       3.1.5                   EMISSIONS
Reactivity                     3.1.6                   REACTIVITY*
*  Since OZIPP default values are recommended, this option may be omitted.
                                        45

-------
respectively,  Reference. 3 provides guidance for choosing the proper scales using
the ISOPLETH option.  Secondly, tsopleths corresponding to the ozone dally design
value and to 0.12 ppm should &e Incorporated In the diagram to facilitate control
calculations.  This 1s also accomplished by using the ISOPLETH option.   Conse-
quently, a day-specific Isopleth diagram can be generated by proper specification
of the ISOPLETH option 1n addition to those listed 1n Table 3-3.
                                        46

-------
4.0  CALCULATION OF CONTROL ESTIMATES
     The previous Chapter described procedures for deriving day-specific ozone
isopleth diagrams.  These diagrams are used to calculate VOC emission controls
necessary to reduce the peak ozone level observed at a particular site on an
individual day to 0.12 pom.  The SIP requirement 1s determined from those results
according to the method described 1n Section 2.3.  This Chapter describes the
techniques that are used to calculate the VOC emission reduction estimates for
each site-day 1n order to determine the SIP control requirement.

     In estimating the degree of VOC control necessary to reduce peak ozone
levels to 0.12 ppm, the role of NOX should be accounted for.  This is accomr
pi 1 shed by estimating the change 1n NO  between the base period and the future
                                      A
period which is of Interest.  For the 1982 SIPs, the most straight-forward
procedure for estimating this change 1s to project the expected percent change
in total region-wide NOV emissions between the base period and 1987.  -Factors to
                       A
consider include growth, anticipated effects of various control programs, and
technological advancements.

     The first step in calculating controls using a diagram is to establish a
starting point on the diagram.  This point is defined by the intersection of the
design NMOC/NO¥ ratio line with the isopleth corresponding to the ozone dally
              ^
design value, and represents the base case conditions (i.e., base emissions,
base transport, etc.).  All other points on the diagram represent the effects of
changing precursor emissions relative to the base case assuming that everything
else remains constant.  Thus,  control requirements may be calculated for a given
day using a single diagram if:
                                        47

-------
          11  Initial precursor concentrations and post 8;OQ a.m.  emissions are-
reduced proportionally, and
          2}  All other non-emission factors (e.g., transported ozone}  remain
constant.
     A second situation arises tf there 1s a concurrent change 1n  emissions and
some other factor, or If post 8:00 a.m. emissions are not reduced  proportionally
with Initial concentrations.  However, meteorological conditions must be assumed
to be the same 1n the future case as 1n the base case, e.g., mixing heights will
be the same for both diagrams.  In this case, two Isopleth diagrams are required:
(1) a diagram applicable for base case conditions, and (2) a diagram applicable
for the future (i.e., the post-control period 1n 1987), which reflects the con-
current changes described above.  This situation 1s discussed in Section 4.2.
4.1  USE OF SINGLE DAY-SPECIFIC DIAGRAMS
     An isopleth diagram derived for a specific day corresponds to the base
period.  If non-emission conditions corresponding to the base period are expected
to remain unchanged, then the single day-specific diagram can be used to estimate
the VOC control requirements.  The single diagram cannot be used if one of the
following situations is expected to occur between the base period  and the future
period of interest (e.g., 1987):
          1)  ozone transported into the area changes
          2)  precursors transported into the area change
          3)  the diurnal emission pattern changes grossly (.I.e.,  hourly emissions
change independently of one another}.
                                        48

-------
If one or more of these situations 1s anticipated, the procedures described 1n



Section 4.2 must be followed.





     The procedure for arriving at the necessary VOC emission reduction 1s out-



lined 1n Table 4-1 and an example 1s shown 1n Figure 4-1.  The starting point on



the diagram 1s located by the Intersection of the design NMOC/NCL ratio line
                                                                A


with the Isopleth corresponding to the dally ozone design value.  The NMOC and



N0¥ coordinates of this point correspond to the base case conditions to which
  A


all changes 1n NOY and/or NHOC must be referenced.  The next step 1s to estimate
                 ^


the change 1n NO  between the base state and the future period.  The base NOV
                A                                                           X


level 1s adjusted by the percent change 1n total NOX emissions anticipated



between the two periods.  A post-control point 1s then located on the 0.12 ppm



isopleth at the point reflecting the projected change In NOV.  The final step 1n
                                                           J\


the calculation Is to compute the percent change 1n NMOC required to reduce NMOC



from the base point to the post-control point.  This percent reduction Is the



degree of VOC emission control necessary to reduce the peak ozone level to



0.12 ppm.





4.2  CONCURRENT CHANGES IN EMISSIONS AND OTHER FACTORS



     Consideration of different control measures that are implemented concur-



rently necessitates using a different procedure than that described 1n Section 4.1.



The major examples of concurrent implementation of control measures include:





          1)   reduction in local precursors and reduction of ozone transported



into a city;




          2)   reduction in local precursors and reduction of precursors trans-



ported into a city;
                                        49

-------
Table 4-1.  Control Calculations Using A Single Isopleth Diagram


Step 1 :   Locate the base-case point on the diagram by finding the Intersection
          of the design NMOC/NO  ratio line with the Isopleth for the ozone
          dally design value.  Cet the NMOC and NO  coordinates of this point
          be signified by (NMOC)i and (NOYh, respectively.
                                         A


Step 2:   Calculate the post-control NO  coordinate by adjusting the base
          NO  level by the expected percent change 1n NO  emissions between the
          base and post-control periods.  Let the post-control NO  coordinate be
          signified by (NOJ2.  Thus,                            *
                          A
                                      ANO
               (NOX)2 = (N0xh X (1 -H

          where
               (NOV)2 = post-control NOV coordinate
                  X                    A

               (N0v)i = base case NOV coordinate
                  X                 X

               ANOV   = expected change in NOV, percent
                  A                          A


Step 3:   Locate the post-control point on the diagram by finding the post-
          control HOY coordinate, (N0¥)2» on the 0.12 ppm ozone Isopleth.  Let
          the NMOC coordinate of thisxpo1nt be (NMOC)2.


Step 4:   Compute the percent reduction necessary to lower (NMOC)i to (NMOC)2.
          Thus,


               % Reduction in VOC = [1 - (NMOC)? ] x 100
                                         "
                                        50

-------
GIVEN:  03 Dally Design Value     = .24

        Design NMOC/NOX           =8:1

        Anticipated Change in N0¥ = -20%
                                J\
        Base case diagram shown below

FIND:   Percent reduction in VOC emissions needed to reduce ozone from .24
        to 0.12 ppm

SOLUTION:
                            BflSt ^PSf DlflGRP.f.  03 «!'.CrT L .12


STEP 1: The base case point is found by the intersection of the 8:1 NMOC/NO
        ratio line with the .24 ozone isopleth (Point 1) on the diagram.  At
        Point 1, (NMOCh = 1.64 and (N0vh = .205
                                       /\

STEP 2: The post-control NO  coordinate is calcualted as follows:
                           y\

               (NO)2 = (.205) x (1 -      = .164
STEP 3: The post-control point is located at the intersection of the .164 NOX
        coordinate and the 0.12 ppm ozone isopleth (Point 2).  At Point 2,
        (NMOC)2 = 0.46

STEP 4: The VOC emission reduction is calculated as

               % reduction = (1 -    6) x 100 = 72%
Figure 4-1.  Example emission reduction calculation using a single ozone
             isopleth diagram
                                       51

-------
          31  reduction  1n  local precursors and gross changes 1n diurnal emission
patterns;
          4}  combinations  of the above.

In each case, the procedure for calculating YOC emission reductions requires the
development of an additional Isopleth diagram.  The VOC emission reduction is
then calculated using both  the base case and the additional diagram.  Each of
the above four situations are considered below.

     4.2.1  Concurrent Reductions in Local Precursors and Transported Ozone
          When calculating  emission reductions, consideration should ordinarily
be given to the situation when future transported ozone concentrations may be
different from the base  period levels as a result of the imposition of control
measures upwind of the city of interest.  Before calculating VOC emission reduc-
tions, an isopleth diagram  must be generated for the post-control case, i.e.,
the case in which the ozone transported into the area is reduced.  It must be
recognized that it is difficult to generalize about what should be assumed
concerning future levels of transported ozone.  One source of this difficulty
arises as a result of the differing characteristics of "upwind areas."  For
example, 1n the Northeastern United States, one could expect greater reductions
in "future transported ozone" than in many other areas.  The greater reductions
would occur because of large designated nonattainment areas for ozone which are
upwind of northeastern cities.  Nonattainment areas are subject, at a minimum,
to emission controls arising from the Federal Motor Vehicle Control Program
(FMVCP), application of  Best Available Control Technology (BACT) on new sources
of VOC, phasing out of old, poorly controlled stationary sources and application
of Reasonable Available  Control Technology (RACT) on existing sources.  In many
                                        52

-------
other areas of the country however, upwind areas are predominantly unclassified
or attainment areas.  These areas are subject only to the FMVCR and emission
reduction which might occur as the result of existing, poorly controlled sources
being retired and replaced with new sources having BACT.  The procedures described
below require case by case judgments as to whether areas upwind of a city being
modeled are primarily nonattalnment areas or otherwise.  Once this judgment 1s
made, these procedures may be used to estimate the degree of reduction expected
due to control programs Implemented upwind of the city.  Following this discus-
sion 1s a description of the quantitative techniques for estimating VOC emission
reductions.

          A.   Estimating Future Transport of Ozone
               Recall from Section 3.1.3 that ozone may be transported 1n both
the surface layer and aloft.  As was described 1n that section, several studies
have shown that advectlon of ozone aloft appears to be the more significant
mechanism of transport from one area to another.  If control programs are imple-
mented in upwind areas, ozone transported into the city is likely to be reduced.
However, in most cases, the source area and the level of future controls are not
likely to be known to any degree of certainty.
               Recommended Procedure:   Because of the considerable uncertainty
in the location and future control levels of the source area(s) for ozone trans-
ported into the urban area, the relationship depicted in Figure 4-2 is recommended
for estimating the future ozone transport level given the level of present
transport.   The solid curve in Figure 4-2 was derived on the basis of changes in
VOC emissions which are projected assuming a national mix of source categories;
national estimates of projected growth in stationary source emissions and vehicle
miles traveled (VKT); anticipated impact of applying reasonably available control
technology CRACT) to stationary sources and the impact of the Federal Motor
Vehicle Control Program on mobile sources; and consideration of natural back-
ground levels.2,32'35  It was assumed that future ozone levels would not exceed
the NAAQS.   The solid curve is most appropriate for use by cities subject to
impacts from large upwind nonattainment areas.  The dashed curve in Figure 4-2
                                        53

-------
Q.
Q.
S-
o
Q.
c/J
£1
o
N
C

d)
S-
                  Present Ozone Transport,  ppm
  Figure 4-2  Future Ozone Transport as a Function of Present Transport

-------
differs from the  solid curye only 1n that the  Impact of FACT 1s  ignored.  The

dashed curve is most appropriate for use when  a city 1s Isolated and not Impacted

by  large designated nonattainment areas.



                    Without Information to the contrary,  future transport along

the surface should be assumed equal to zero.   If  significant non-zero concen-

trations were  found for present ozone transport along the surface, then future

ozone transport levels should be obtained using the relationships shown in

Figure 4-2.





          B.   Calculation of 1/OC Emission Reductions



               The computational procedure for determining the required VOC



emission reductions is outlined in Table 4-2,  and an example problem is shown in



Figure 4-3.  The  procedure is similar to that  described in Section 4.1.  The



starting point for the calculation is found in exactly the same manner as



described in Section 4.1, I.e.» by the intersection of the design NMOC/NO  ratio
                                                                         A


line and the ozone isopleth corresponding to daily ozone  design value on the



base case diagram.  Again, the NMOC and NO  coordinates of this point correspond
                                          J\


to  the base case  conditions to which any changes  in NMOC  and/or NO  must be
                                                                  A


referenced.  The  post-control  point is found exactly as was previously described,



except that a diagram corresponding to future  transported  ozone level is used.



Changes expected  1n NOY are first calculated by adjusting  the base NOV level
                      A                                              A


(i.e., the NO  coordinate of the starting point) by the assumed percent change
             A


in  NO  expected between the base period and the post-control  period.  The post-
     A


control  point 1s  found on the 0.12 ppm isopleth of the future case diagram at



the post-control  NO  level.   The NMOC reduction is then determined by calculating
                   A


the percent reduction necessary to reduce NMOC from the base point to the post-



control  point.   This percent reduction is the  VOC emission reduction necessary



to  reduce the peak ozone concentration to 0.12 ppm, assuming that ozone trans-



ported aloft changes in the future.
                                        55

-------
Table 4-2.  Control Calculations Using Two Isopleth Diagrams


Step 1 :   After developing a Base case diagram, generate a future-case diagram
          for the post-control period.  For example, 1f transported ozone Is
          expected to be reduced, a diagram reflecting that reduction should be
          generated.


Step 2:   Locate the base-case point ph_the base-case diagram by finding the
          intersection of the design NMOC/NO  ratio line with the isopleth for
          the ozone dally design value.  Let the NMOC and NO  coordinates of
          this point be signified by (NMOCh and (NO h, respectively.
                                                    A


Step 3 :   Calculate the post-control NO  coordinate by adjusting the base NO
          level by the expected percentrchange in NO  emissions between the Base
          and post-control periods.  Let the post-control NOV coordinate be
          signified by the CNOj2.  Thus,                   x
                                           ANO
                    CNOX)2 = (N0xh x (1 +


               where
                    (NO )2 = post-control NOV coordinate
                       A                    X

                    (NO h = base case NO  coordinate
                       A                 X

                     ANOV  = expected change in NOV, percent
                        X                         X


Step 4:   Locate the post-control point on the future-case diagram by finding
          the post-control NOV coordinate, (NOV)2, on the 0. 12 ppm ozone Isopleth.
          Let the NMOC coordinate of this point be (NMOC)2.


Step 5:   Compute the percent reduction necessary to lower (NMOC)i to (NMOC)2-
          Thus,
               % Reduction in VOC = [1 -       z ] x 100
                                        56

-------
                                      = .24

                                      = 8:1
GIVEN:  03 Design Value

        Design NMOC/NOX

        Present Transport of 03 Aloft = .12


        Future Transport of 03 Aloft  = .09 (from Figure 4-2)

        Anticipated change in NO      = -202

        Base case and future diagrams shown below.


FIND:   Percent reduction in VOC needed to reduce ozone from .24 to .12 ppm
SOLUTION:
                        NIIMC.P'TC
    STEP 1:  The future case diagram shown above was generated by using the OZIPP input data for the base
             case diagram, except that ozone transported aloft was changed from 0.12 to 0.09 ppm.

    STEP 2:  The base case point is found by the intersection of the 8:1 NMOC/NO  ratio line with the .24
             ozone isopleth (Point 1) on the base case diagram.  At Point 1, (NMOCh = 1.64 and (NO h = .205

    STEP 3:  The post-control  NOX coordinate is calculated as follows:


                   (NOX)2 = (.205) x (1 - ^) = .164
         4:  The post-control point is located on the future case diagram at the intersection of the .164
             NOX coordinate and the 0.12 ppm ozone isopleth (Point 2).  At Point 2 (NMOC)2 = .61
    STEP


    STEP 5:  The VOC emission reduction is calculated as:
                   % reduction = (1 -     ) = 63%
  Figure 4-3.   Example emission  reduction  calculation  considering changes  in  transported  ozone.
                                                        57

-------
     4.2.2  Concurrent Reductions In local Precursors and Transported Precursors
          As was described  In Section 3,1.4, the transport of precursors 1s not
considered essential Inmost Instances, especially If upwind reductions parallel
local reductions.  However, If precursor transport Is deemed critical, procedures
that may be employed are addressed 1n detail 1n Appendix B.  The conceptual
framework for considering precursor transport 1s discussed below, along with
procedures for estimating changes 1n future levels of precursors transported
into an area.

          Conceptually, the procedures for considering precursor transport are
exactly the same as those described for ozone transport.  The major problem 1s
estimating the precursor concentrations transported Into the area.   This in part
depends on the geographical location of the city, and is addressed 1n detail in
Appendix B.  Because control programs are likely to be implemented upwind, the
concentrations of precursors transported into an area are likely to change.
Described below are procedures that may be used to estimate the changes in pre-
cur^sor levels transported into an area.  These may be applied to both surface
layer transport and transport aloft where applicable.
          Recommended Procedure:  Because of the considerable uncertainty in
future control levels for the source area(s) for precursors transported into an
urban area, the following reductions are recommended:  NMOC transported concen-
trations may be reduced up to 40% in cities being impacted by upwind nonattain-
ment areas and up to 20% in other cities.  NO  levels are not projected to
change significantly, so no change should be assumed.  The basis for these
recommendations is the projection of VOC and NO  emissions that was described in
Section 4.2.1 on estimating future transported ozone levels.
          Alternate Procedure;  If specific information concerning the primary
source of precursors transported into the urban area, and the scheduled controls
for that area are available, then estimates of future transport may be developed
using a proportional model.  However, the source area should be clearly identified
through the use of trajectory analysis and the assumptions concerning the imple-
mentation of future controls should be fully documented.
                                        58

-------
          The procedures for considering precursor transport aloft are exactly
the same as those described for ozone transport.  A base case diagram is used in
conjunction with a diagram representing the post-control state.  The first repre-
sents the base case with existing transported levels, while the second diagram
corresponds to the situation in which changes 1n precursor transport are expected.
The procedures outlined 1n Table 4-2 are also applicable to this case.  The
procedure for considering precursor transport 1n the surface layer 1s more complex,
and is discussed in Appendix B.
     4.2.3  Concurrent Changes 1n Local Precursors and Diurnal Emission Patterns
          Gross changes in the diurnal emission pattern expected between the base
period and the post-control period can be Incorporated into an analysis.  It must
be emphasized that these changes must be significant and on a scale as large as,
or larger than, the spatial and temporal resolution of the emissions Inventory
used 1n the analysis (see Section 3.1.5).  Because of the assumptions inherent 1n
the model, evaluation of small scale changes cannot be made using EKMA.  Thus,
for example, the OZIPP/EKMA approach could not be used to evaluate the effects of
constructing a new highway.  However, it could be used to assess the effects of
large scale changes, such as a suburban county substantially increasing emissions
relative to an urban county as a result of rapid growth.

          The procedures for incorporating changes in diurnal emission patterns
are similar to those for treating changes 1n transported pollutants.   A base case
diagram is developed using the existing emission information, and a diagram
representative of future emissions is developed incorporating the expected changes
in the diurnal emission pattern.   The expected changes must be derived from
locally applicable projections.  For example,  an increase in emissions in one or
                                        59

-------
more counties might  be  anticipated due to population growth or industrial expan-
sion.  This expected increase would be incorporated in the analysis by developing
a future case diagram using the emission fractions corresponding to the projected
emissions increase.   Thus, if a two-fold increase in county-wide emissions (in
the absence of controls) is assumed for a county, the emission fractions for the
affected county would be doubled for generating the future case diagram.  Control
calculations are performed exactly 1n the manner described in Table 4-2.
          Normally,  consideration of changes in diurnal emission patterns will
not be necessary.  The  modeler will have to decide whether or not any anticipated
changes are significant enough to warrant special consideration.
     4.2.4  Consideration of Multiple Changes
          The previous  three sections summarized the procedures for calculating
control estimates for cases in which local precursors are reduced and one other
significant change is expected to occur.  If more than one additional change is
anticipated, the same procedures can be used.  In addition to the base case
diagram, one diagram reflecting all future changes is developed.  For example,
assume transported ozone, transported precursors, and gross changes in the
diurnal emission pattern are all expected.  The base case diagram would be devel-
oped using the base  case transported ozone, transported precursors, and diurnal
emission fractions.   The second diagram would be developed using the future
transported ozone levels, future transported precursor levels, and the emission
fractions corresponding to the projected diurnal emission pattern.  This new
diagram would incorporate all changes simultaneously, and any calculated VOC
emission reductions  would reflect these assumptions.  However, the future case
diagram would still  use the same meteorological conditions, e.g., parcel path and
                                        60

-------
mixing heights, as the base case.  The calculations themselves would be carried
out just as outlined in Table 4-2.
                                         61

-------
5.0  ACKNOWLEDGEMENT
     The principal authors of this document are Gerald L. Gipson, Warren P.  Freas,
Robert F. Kelly, and Edwin L. Meyer, Jr.  The authors are indebted to members of
the EPA Working Group on Ozone Modeling for their many helpful written and verbal
comments.  The excellent typing and general clerical support from Mrs. Carole Mask
is also gratefully acknowledged.
                                        62

-------
                 APPENDIX A
ESTIMATION OF MIXING HEIGHTS FOR USE IN OZIPP

-------
     In OZIPP, the rate of dilution of atmospheric pollutants 1s governed by the
diurnal change 1n mixing height.  The mixing height Is the top of a surface-based
layer of air which is well-mixed due to mechanical and thermal turbulence.  As
described in Section 3.1.2, the input variables required for OZIPP include:  the
mixing height at 0800 LCT, the maximum mixing height, the time at which the
mixing height begins to rise If it starts to rise after 0800 LCT, and the time at
which the mixing height reaches its maximum.  The rate of rise 1s computed
internally by OZIPP.

     Three different procedures for determining daily morning and afternoon
mixing heights were outlined 1n Section 3.1.2.  The recommended procedure entails
the use of temperature soundings taken routinely by the National Weather Service
at various locations throughout the United States.  If more direct measurements
are available (e.g., radiosondes taken in the urban area or sodar data), they may
be used instead of NWS data.  If neither of the above two sets of measurements
can be used, then the use of 250 m for the 0800 LCT mixing height and the climato-
logical mean value for the maximum mixing height is recommended.  The procedures
to be followed for each approach are described below.

A.I  RECOMMENDED PROCEDURE USING NWS RADIOSONDES
     Temperature soundings are taken by the NWS at sites throughout the United
States.  Soundings are usually taken every 12 hours at 1200 and 0000 Greenwich
Mean Time (GMT), corresponding to 0800 and 2000 Eastern Daylight Time, respec-
tively.  Therefore, to estimate daily mixing heights, (1) a NWS site must be
selected which is representative of the city of Interest, (2) appropriate
sounding data and urban surface data must be obtained, and (3) these data must be
used to compute the morning and maximum mixing heights.  Each of these steps is
discussed below.
                                        A-l

-------
     A.T.I- Site Selection
          In selecting a NWS site as the basis for mixing height estimation,
care should be taken to Insure that the site 1s meteorologically representative
of the city of Interest.  Table A-l contains recommended sites for a number of
cities.  Backup sites are listed for those cases In which radiosonde data may
not be available for a given day, or 1f the site has significantly different
meteorological conditions.  Examples of the latter are the case 1n which a
surface front lies between the sounding site and the city or the city 1s clear
but cloudiness or precipitation occurs at the sounding site.
     A.1.2  Selection of Day Specific Data
          The dally morning mixing height for the model 1s normally estimated
using the 1200 GMT (0800 EOT) sounding, while the maximum mixing height is esti-
mated using the 0000 GMT (2000 EOT) sounding.   In some cases, these soundings
may not be available or appropriate and alternate approaches will  be necessary.
Table A-2 summarizes the order of preference in selecting the radiosondes for
estimating the daily mixing heights.  The actual data may be obtained from the
National Climatic Center (NCC).*
          In addition to the sounding data, surface temperature and pressure
data are also needed for each day modeled.  The urban surface temperature at
0800 LCT (or the average temperature between 0800-0900 LCT)  and the maximum
*National Climatic Center, Federal Building, Asheville, NC  28801
 Telephone:  (704) 258-2850, x203
 Please allow about four weeks for NCC to fill an order.
                                        A-2

-------
Table A-!.
     City
Allentown, PA
Baltimore. MD
Boston, MA
Bridgeport, CT
Chicago, IL/IN
Cincinnati. OH/KY
Cleveland, OH
Dayton, OH
Denver, CO
Detroit, MI
Fresno, CA
Hartford. CT
Houston, TX
Indianapolis, IN
Los Angeles. CA
Louisville, KY/IN
Milwaukee, WI
Nashville, TN
New Haven, CT
New York. NY/NJ
Philadelphia. PA/NO
Phoenix, AZ
Pittsburgh, PA
Portland, OR
Providence, RI

Richmond, VA
Sacramento, CA
St. Louis, MO/IL
Salt Lake City. UT
San Bernardino, CA
San Diego, CA
San Fransdso, CA
Scranton, PA

Seattle. WA
Springfield, MA
Trenton, NO
Ventura-Oxnard, CA
Washington DC/MD/VA
Wilmington, DE

Worcester, MA
Youngstown, OH
         NWS Radiosonde Stations

     PRIMARY                 BACKUP(S)
                         Cllnatologlcal
                         Mixing Heights
                           (m AGO

                                   MAX
 NYC. NY; At! City, NO
 Dulles AP, VA
 Portland, ME
 NYC, NY; At! City, NJ
 PeoHa, IL
 Dayton, OH
 Dayton, OH
•Dayton, OH
•Denver, CO
 Flint, MI
 Oakland, CA
 Albany, NY
 Victoria. TX
 Dayton, OH
 Vandenberg AFB, CA
 Dayton, OH
 Green Bay. MI
*Nashv1lle, TN
 NYC. NY; Atl City, NJ
•NYC, NY; Atl City, NJ
 NYC. NY; Atl City. NJ
 Tucson, AZ
•Pittsburgh, PA
 Salem. OR
 New York. NY

 Dulles AP. VA
 Oakland. CA
 Salem. IL
•Salt Lake City, UT
 San Diego, CA
•San Diego, CA
 Oakland. CA
 NYC, NY; Atl City, NJ

 Oullayute. WA
 Albany, NY
 NYC, NY; Atl City, NJ
 Vandenberg AFB.'CA
•Dulles AP, VA
 Dulles AP, VA;
  Atl City, NJ
 Albany. NY
 Pittsburgh, PA
Albany, NY; Dulles AP, VA          1825
Wallops Is., VA; Atl City, NJ      1825
Albany, NY; Chatham, MA            1375
Albany. NY                         1500
Green Bay. WI                      1575
Huntlngton, WV                     1650
Buffalo. NY                        1650
Huntlngton. WV                     1661
Grand Junction, CO                 3358
Dayton, OH                         1700
Vandenberg AFB. CA                 2000
NYC, NY; Atl City, NJ              1500
Lake Charles. LA                   1525
PeoHa, IL; Salem; IL              1600
San Diego, CA                       603
Nashville. TN                      1700
PeoHa, IL                         1575
Jackson. AL                        1845
Albany, NY                         1450
Albany, NY                         1512
Dulles AP. VA                      1700
Wlnslow. AZ;                       3250
Dayton, OH; Dulles AP, VA          1794
Kedford, OR; Qullavute, WA         1575
Albany, NY; Chatham, MA;           1350
 Atl City, NJ
Greensboro. NC; Wallops Is., VA    1725
Vandenberg AFB. CA                 1600
PeoHa. IL; Monett, MO             1625
Grand Junction, CO                 3673
Vandenberg AFB, CA                 1200
Vandenberg AFB, CA                  564
Vandenberg AFB, CA                  625
Albany, NY; Atl City, NJ           1850
 Dulles AP, VA
Salem, OR                          1398
NYC, NY; Atl City, NJ              1600
Dulles AP, VA                      1700
San Diego, CA                       610
Wallops Is., VA                    1884
Wallops Is., VA; New York, NY      1700

Portland, ME; Chatham, MA          1500
Buffalo, NY; Dayton, OH            1700
• This station should be used unless the data 1s missing  for all  the  times
  listed 1n Table A-2.   However,  1f a frontal passage  occurs between  the  time of
  maximum ozone and the time of the launch of the 0000 GMT  sounding (normally
  about 2300 GMT), the  1200 GMT sounding from th?t site should  be used.

  NOTE:  The NYC. NY radiosonde station was replaced by Atlantic  City. NJ on
         September 2, 1980.
                                         A-3

-------
Table A-2.  Preferential Order of Data Selection
          Morning Mixing Height Estimate

          1.   1200 GMT Sounding at Primary Site
          2.   0600 GMT Sounding at Primary Site*
          3.   1200 GMT Sounding at Backup Site
          4.   0600 GMT Sounding at Backup Site*


          Maximum Mixing Height Estimate

          1.   0000 GMT Sounding at Primary Site
          2.   1800 GMT Sounding at Primary Site*
          3.   1200 GMT Sounding at Primary Site
          4.   0000 GMT Sounding at Backup Site
          5.   1800 GMT Sounding at Backup Site*
          6.   1200 GMT Sounding at Backup Site
*  Soundings are not normally taken at these times,  but may be  available  1n some
   instances.
                                        A-4

-------
temperature occurring prior to 1800 LCT are needed to estimate the morning and
maximum mixing height, respectively.  The surface temperature data should be
measured to the nearest 0.1 °C at a well ventilated site.  The site should be
located near the center of the urban area.  Surface atmospheric pressure measure-
ments are needed at the same time and location of the urban surface temperature
measurements, if at all possible.  If these measurements are not available, a
local NWS or Federal Aviation Administration weather reporting station's barometer
reading may be used.

          If the elevation of the pressure reading and the urban temperature
site are different, an adjustment should be made to the pressure measurement
using equation (3).
          where
               Z .   = the elevation, in meters above sea level (mASL), of the
                      pressure measurement
               Zsfc = t'ie e^evatl'on (mASL) of the urban temperature measurement
               Pobs = ^e Pressure» in millibars, at Z -
               Psfc = the pressure, in millibars, at the urban temperature site

NOTE:  ZQks will be equal to zero meters ASL when a pressure reduced to sea
level is used.

The value of P -  from equation (3) is an approximate value and can be rounded
to the nearest whole millibar.
                                        A-5

-------
     A. 1.3- Nixing Height  Esttetfon
          The procedures for estimating the 0600 LCT mixing height and the
maximum mixing height are  outlined In Table A-3.  These procedures have been
automated.  Appropriate computer listings and a User's Guide can be obtained
from the authors.  The procedures 1n Table A-3 are designed for use with the
worksheet displayed  In Table Ar4.  Figure A-T contains a flow diagram of the
process.  The procedures use the mandatory and significant pressure levels
reported for each sounding (Table A-5).  The steps lead to determination of the
height at which the  adlabatic lapse rate (extended from the surface temperature
and pressure) intersects the vertical temperature profile.  (For background
information, the reader Is referred to References 21, 36-37.)  An example
problem is presented 1n Section A.4.

          In some instances, the mixing heights estimated by this procedure way
not be representative.  If the 0800 LCT morning mixing height is estimated to be
less than 250 meters, then a value of 250 meters should be used.  This assumed
minimum value for the 0800 LCT mixing height accounts for the effects of mixing
due to mechanical turbulence caused by Increased surface roughness in the urban
     11 38
area.  '    Similarly, if  the city's maximum mixing height 1s greater than twice
the cHmatological maximum value (e.g., see Table A-l), the surface temperature
and pressure used and the  choice of sounding site should be checked for repre-
sentativeness using  the guidelines 1n A.1.1 and A.1.2 above.  If no backup data
are available, twice the cHmatological value should be used as the maximum.
Also, a maximum mixing height less than or equal to the morning mixing height,
or less than one-third the cUmatological maximum mixing height value is suspect.
Using data from a backup site may provide a more realistic value.  However, if
                                        A-6

-------
 Table A-3.   Procedures for Estimating K1x1ng Heights


      Step 1  — For reference,  the  Information at  the  top of Table A-4 should be listed
 (e.g., date, city, etc.).   If  the  naming mixing  height 1s  to  be calculated, the 0800 LCT
 surface data are used.  If the maximum mixing height  Is to  be  calculated, the data cor-
 responding  to the time of  maximum  temperature (I.e.,  between 800-1800 LCT) are used.  In
 the row labeled URBAN SURFACE  DATA, enter the following Information:  1) the elevation of
 the urban temperature site In  meters above  sea level;  2) the surface pressure In millibars
 (this value  1s P,fc)> and  3) th& surface temperature  In degrees Celsus (°C).

          Convert the surface  temperature In column four to degrees Kelvin (°K) by adding
 273.2, and enter the result In column five.  This value Is  T f (CK).

          Use Equation 1 below and the values Just entered  to  calculate the potential
 temperature  at the surface (0C.  In eK to the nearest  0.1 °K) and enter this value under
 column six  ue(cK)".          sfc
          G     fin  °IO  - T
          e     un   K;    '
                                            (In mb)°-286
            sfc  vm   *'  •  'sfc V1"  *'\—1000 mb

     Step 2 —  Using the  temperature sounding data, find the highest pressure level other
 than the sounding's  surface value that Is less than the pressure at the urban surface.*
 From this pressure level  on the sounding, enter the height  (If listed), pressure and
 temperature (1n °C)  Into  the row marked "(2)" on Table A-4.

     Step 3 —  Convert  the temperature at this le*vel to the Kelvin scale and enter 1n
 column 5.   Compute the  potential temperature (6 ) to the nearest 0.1"K using the pressure
 (P, In mb)  and  temperature (T  1n °K) at this level 1n Equation 2 below:

                                  /  .     . \.Q 286
          0  /<„  OK\  _  T  /,  o|/\  f (in mb) I                   /.x
          «p nn  K;    ip un  KJ ^ jA.^ mD' /                   IO


 Enter the value of 9. found from Equation (2) Into the same row under the column labeled


     Step 4 --  If the potential temperature, "9," of the last row that was entered 1s
 greater than the  potential temperature 9 - , and this Is the first level above the surface,
 then 250 meters should  be used as the mixing height.  Otherwise, go to Step 5.  If It 1s
 less than or equal to 8 . , then enter the height (1f given), pressure and temperature of
 the next lowest pressure  level found on the sounding Into the next row of Table A-4 and
 return to Step  3.

     Step 5 —  The mixing height 1s between the last two levels entered Into Table A-4.  If
 height values are given for both of these levels, the elevation of the mixing height can be
 found using Step  6.   If one of the levels does not have a height value, use linear Inter-
 polation to find  the  pressure value for the potential temperature value of 8 -  + 0.1 "K.
 Enter this  pressure  value Into the row marked "MIXING HEIGHT" at the bottom 8fcTable A-4
 under the column  "PRESSURE In mb."  Proceed to Step 7.

     Step 6 —  From the two levels where height 1s given on the sounding surrounding the
 mixing height level,  use  linear Interpolation to find the height (1n meters ASL) at the
 value 6 -   +  O.TK (I.e.. the potential  temperature 0 at the mixing height).  Enter the
 value found  by  linear Interpolation Into the row labeled "MIXING HEIGHT" under the column
 "HEIGHT (mASL)" and proceed to Step 8.

     Step 7  --  Use linear Interpolation  to find the height above sea level of the mixing
 height using  the  pressure at the mixing  height (found 1n Step 5)  and the pressure levels on
 the sounding  above and  below the mixing  height pressure that have both pressure and height
 values.   Enter  the height value found Into the row "MIXING HEIGHT" under the column marked
 "HEIGHT, (mASL)" and  proceed to Step 8.

     Step 8  —  Subtract the elevation of the urban site (mASL) from the height (mASL) of
 the mixing  height.  The result 1s the height of the mixing height 1n meters above the
 surface of  the  dty (mA.GL).   Enter this  value Into Table A-4.

NOTE;  Despite the fact  that pressure and height,  and potential temperature and
      height, are not linearly related,  linear Interpolation does not produce
      significant errors  over the limited ranges used above.

*  For example, 1f the  urban surface pressure 1s 985 mb,  and the sounding pressures are:
   1005, 1000,  963, 850 mb,  etc.,  963 mb 1s  the "highest pressure level  that 1s  less than
   the pressure at the  urban surface."   850  mb 1s the "next lowest pressure level" needed
   In  Step  4.


                                            A-7

-------
Table A-4.  Worksheet for Computing Mixing Heights
    Date:                     Time  Of Mixing Height For Input Into Model:
                                           Sounding Method:

                                       LCT.  Surface elevation:
City:

Time of Sounding:

Location of Sounding:

LOCATION OF URBAN SURFACE DATA (IF DIFFERENT THAN ABOVE)  -
tiiASL
LEVEL
Urban
Surface
Data (1)
(2)











MIXING
HEIGHT
HEIGHT
(mASL)












PRESSURE
(mb)












TEMP.
(°C)












0 + 0.1 °K PRESSURE HEIGHT
(°K) (mb) (mASL)
TEMP.
(°K)












9
(°K)












REMARKS
Kfc







•



HEIGHT HEIGHT USED IN
(mAGL) MODEL (mAGL)
/
                                       A-8

-------
                              - ENTER URBAN SURFACE DATA
                              INTO TABLE A-If,
                              - CONVERT SURFACE TEMPERATURE
                              TO "X
                              -mro egf  USING EQUATION (i).
                         3}- ON TABLE A-tf LIST DATA FROM
                         *^ SOUNDING FROM THE FIRST PRESSURE
                          LEVEL ABOVE THE URBAN SURFACE LEVEL.
                             i - CONVERT TEMPERATURE TO °K AND
                              USE EQUATION (2) TO COMPUTE PO-
                              TENTIAL TEMPERATURE (9_) FOR
                              THE LEVEL JUST ENTERED.
                                                          - ENTER NEXT PRESSURE
                                                            LEVEL INTO TABLE A-f
                                             TES
                                   - MIXING HEIGHT IS
                                   AT esfc + 0.11.
                                      - ARE THE
                                 'HEIGHT VALUES GIVEN
                                   THE LAST TWO ROWS
                                     ENTERED INTO
                                      JABLE A-4,

                                             NO
                            TES
                             - USE LINEAR INTERPOLATION
                             TO FIND THE PRESSURE AT
                                                        OF
- FIND THE HEIGHT (mASL)
OF THE MIXING HEIGHT BY
LINEAR INTERPOLATION FROM
THE PRESSURE AT THE MIXING
HEIGHT.
                                                                        USE LINEAR INTER-
                                                                      POLATION TO FIND THE
                                                                      HEIGHT  (mASL) AT

                                                                       mijcing height.
                           fl/- FIND THE HEIGHT ABOVE THE
                            URBAN AREA  (mAGL) OF THE MIX-
                            ING HEIGHT  (THIS GIVES THE
                            ANSWER).           	
Figure A-1.   How Chart for  Table  A-3.   Numbers in circles  are step  numbers  in Table A-3.
                                           A-9

-------
Table A-5.  Surface and Sounding Data
Hour Starting at. LCT

          8
          9
         10
         11
         12
         13
         H
         15
         16
         17
         18
        Surface Data

           Temperature °C

                23.2
                23.9
                25.8
                27.3
                28.7
                29.3
                30.1
                30.4
                30.8
                31.4
                31.2
                  Pressure,  mb

                     1010.3
                     1010.7
                     1010.8
                     1010.
                     1010.
                     1010.0
                     1009.6
                     1009.2
                     1008.8
                     1008.6
                     1008.5
                           Sounding Data
          1200 GMT Sounding
                                    0000 GMT Sounding
Pressure (mb)  Height (m ASL)  Temp.  (°C)     Pressure (mb)  Height  (m ASL)  Temp. (°C)
   S 1015
   M 1000
   S  967
   M  850
      827
      817
   M  700
      680
      661
      608
   M  500
      491
      453
      438
   M  400
      388
      349
      324
   M  300
   S  267
   M  250
   M  200
   M  150
   S  148
    8*
  139

 1550
 3168
 5860
 7560
 9640

10890
12370
14190
 23.0
 23.0
 24.4
 16.2
 14.2
 13.6
  4.6
  5.6
  5.6
  0.4
 -8.3
 -9.3
-12.7
-13.9
-18.7
-20.1
-26.3
-29.7
-33.7
-39.5
-47.7
-51.7
-60.9
-61.5
S 1012
M 1000
M  850
S  831
S  791
S  778
S  760
M  700
S  628
S  560
M  500
M  400
S  371
M  300
S  265
M  250
S  205
M  200
M  150
S  127
S  120
M  100
M   70
                                 50
                                 30
                                 20
                                 15
    8*
  114
 1537
 3164


 5860
 7560

 9650

10900

12370
14190
16690
18900
21040
24350
27030
 31.0
 30.6
 16.4
 15.4
                13,
                11,
                11,
  7.0
  1.6
 -1.5
 -7.3
-18.9
-21.7
-33.1
-39.9
-42.9
-52.9
-53.3
-61.1
-64.9
-61.7
-63.3
-58.5
-54.5
-49.9
-44.7
-42.1
Note:  M = Mandatory Levels and S  = Significant  Levels
       If NWS Data are used, both  the  mandatory  and  significant levels
       are needed.

       The 0000 GMT Sounding 1s the following  day  in GMT.
       * The lowest level  of the sounding  should not be used 1n the mixing height calculations.
                                       A-10

-------
the low afternoon mixing height 1s due to the existence of a surface-based
stable layer, an adjustment to the procedures outlined In Table A-3 can be
employed.  Replace the "Urban Surface Data" with the following data from the
sounding site:  1) the maximum temperature, 2) the estimated or observed surface
pressure at the time of maximum temperature, and 3) the height of the sounding
surface level.  Then compute the mixing height according to the procedure 1n
Table A-3.  If this problem occurs on a majority of modeling days, then an
alternative, more representative site should be used for all the modeling days.

A.2  USE OF ALTERNATE DATA
     Other, more direct measurements of mixing height may be used to Increase
the representativeness of the estimated values.  These methods Include direct
urban temperature sounding and sodar data.  The measurements should be taken
over the urban area near the center of the city at 0800 LCT, and close to the
time of the climatological maximum surface temperature.  It 1s not recommended
that these measurements be taken specifically for the OZIPP/EKMA techniques;
however, they may be employed 1f available.  Examples are discussed below.
          1)  Local Urban Radiosonde — the methods described in Section A. 1.3
can be used to find the mixing height from radiosondes taken within the urban
area as opposed to NWS sites.  The radiosonde surface temperature and pressure
should be used in place of the URBAN SURFACE DATA.
          2)  Urban Helicopter Soundings — Similarly, vertical temperature
profiles obtained from helicopter soundings can be used in place of the NWS
soundings.  The urban site surface temperature and pressure should be used as
the URBAN SURFACE DATA.
                                        A-ll

-------
          3)  Sodar  ~  (also  known as Acoustic Radar) the mixing height found by
sodar (1n mAGL) can  be  used directly 1n the model.
NOTE:  Regardless of the  procedure applied, the limitations concerning the
morning and maximum  mixing heights that were described in Section A.1.3 should
be observed.
A.3  USE OF CLIMATOLOGICAL MEANS
     If radiosonde data are not available, 250 m should be used for the 0800 LCT
mixing height and the city-specific cllmatological mean value may be used for
the maximum mixing height.  Table A-l lists representative values for several
cities, and Reference 12  contains information for the contiguous United States.
If Reference 12 is used,  values for summer, non-precipitation days should be
used.  The appropriate  starting and ending times of the mixing height rise in
the model are 0800 LCT  and the time of the maximum temperature.  If the latter
is unknown,.1400 1ST  (1500 LOT) may be assumed.
A.4  EXAMPLE PROBLEM
     To illustrate the  procedure described 1n Section A.1.3, an example problem
is included for reference.  Table A-5 shows relevant data typically available.
Note that both the 1200 GMT and the 0000 GMT soundings are used 1n the calcula-
tions, the former for the morning mixing height and the latter for the maximum
mixing height.  Table A-6 shows the individual computational steps for the
morning mixing height calculation, while Table A-7 shows the same for the maximum
mixing height.
                                        A-l 2

-------
Table A-6.  Morning Mixing Height Determination —

Example from Table A-5:
     08 LCT Temperature = 23.2°C
     Maximum Temperature after 08 LCT = 31.4°C at 17 LCT
     08 LCT Pressure = 1010.3 mb
     Pressure at time of maximum temperature (1700 LCT) = 1008.6 mb
     Time of morning mixing height = 0800 LCT
     Time of maximum mixing height = 1700 LCT
Problem:
     Find the 0800 LCT mixing height using data from the sounding shown in
     Table A-5 (i.e., the 1200 GMT sounding).  A worksheet is shown as Table A-6A.
     The elevation of the urban surface site 1s 62 mASL.
Solution:
     STEP 1
          Enter 62., 1010.3 and 23.2 into row (1) of Table A-6A (URBAN SURFACE DATA)
          TEMP (°C) = 23.2
          Converted to °K = 23.2 + 273.2 = 296. 4°K
          Enter 296.4 into row (1) of Table A-6A under "TEMP(°K)"
          Using Equation (1) on the Urban Surface Data:
               0    = ?Qfi 4°K /'3 mb\
               ysfc   ^6-4 K MOOO.  mbj
               Gsfc = 295>5°K
     STEP 2 - Enter 139., 1000.  and 23.0 into row (2) of Table A-6A.
     STEP 3 -  23.0 + 273.2 = 296. 2°K
               Using Equation (2):

               Q  - ?Qfi 7°K /J99.P. m^\
               6p ' Z96'Z K C1000 mb}
               0  = 296. 2°K (Enter this value into Table A-6A)

     STEP 4 - Op (296. 2°K) is greater than Qsfc (295. 5°K).
Since 9  is from the first level  above the surface,  the 250 m default value
should Be used for the 0800 LCT  mixing height.
                                        A-13

-------
Table        Example  (Hypothetical  Data)
Date: PATE OF MoOE-tiNfr   Time of Mixing Height For Input Into Model: 080O EPT
City: CITY TO BE M»oeLCP              Sounding Method:
Time of Sounding:  0$OO          LCT. Surface Elevatio
Location of Sounding: fl/AME OF SotirtO/Nfr  SITE
LOCATION OF URBAN SURFACE DATA (IF DIFFERENT THAN ABOVE)  - Sr«rrr
, URBAN KAPIOSO/V0E
         mASL
1
LEVEL
Urban
Surface
Data (1)
(2)









MIXING
HEIGHT
__ ~ 2
HEIGHT
(mASL)
U
13?








3
PRESSURE
(mb)
/ 0/0.3
1000.








Osfc+ 0.1 °K PRESSURE
(°K) (mb)
2 95. D ^~"
4 1
TEMP.
33.^
W.o








5
TEMP.
a*.*
W6.^








HEIGHT
(mASL)
O
f 6
oO
*i ^^ £"
(A IO.O
3iu.a








7
REMARKS
^
9 AT TMIS LtVtl
IS Hlt-HtfTHAtl
^SFC+aiV.








HEIGHT HEIGHT USED IN
(mAGL) MODEL (mAGL)
O 2£"O
                                   A-14

-------
Table A-7.  Maximum Mixing Height Determination


Example from Table A-5:
     08 LCT Temperature  = 23.2°C
     Maximum Temperature = 31.4°C at 17  LCT
     08 LCT Pressure = 1010.3  mb
     17 LCT Pressure = 1008.6  mb
     Time of morning mixing height = 0800 LCT
     Time of maximum mixing height = 1700 LCT
Problem:
     Find the maximum afternoon  mixing height  using  data  from the sounding shown In
     Table A-5 (I.e., the 0000 GMT sounding).  A worksheet  1s shown as Table A-7A.  The
     elevation of the urban site 1s 62 mASL.
Solution:
     STEP 1 -                         »
               Enter 62, 1008.6  and 31.4 Into  row (1) of Table  A-7A  (Urban Surface Data)
                    TEMP (°C)  =  31.4CC
                    Converted  to °K = 31.4 + 273.2 = 304.6°K
                    Enter 304.6°K Into row (1) of Table A-7A under "TEMP  (°K)
                    Using Equation (1) on the  Urban  Surface Data
                                         lftnfi  , mh  -0.286
                         0    -  304 6CK  f]zz°'   Pi
                          sfc    •""•" *  Mooo   mo
                    Osfc = 303.9

     STEP 2 -
               Enter 114., 1000., and 30.6 Into Table A-7A.
     STEP 3 -
               30.6°C +  273.2  =  303.8°K
               Using Equation  (2):

                    0 = 303.8CK

                    6 = 303.88K

     STEP 4 - 303.8°K 1s less  than 303.9°K
              Therefore, enter 1537., 850.  and 16.4  Into  Table  A-7A.
              and return to STEP 3.
     STEP 3 -       16.4°C + 273.2 = 289.6°K
               Using Equation  (2)
                                  850 mb  •°'286

                    0 = 303.4°K

     STEP 4 - 303.4°K 1s less  than 303.9°K
              Therefore, enter 831.  and  15.4 into  Table A-7A (note that there 1s no
              height value for this pressure level)  and return  to STEP 3.
                                      A-15

-------
Table A-7 (Continued)

     STEP 3 -       15.4°C  -f 273.2 = 288.6°K

               Using Equation (2):
                                 a,,  mk  - 0.286

                    °p ' 288-6°K (TOT*
                    0   = 304.3CK

     STEP 4 - 304.3°K  Is greater than 303.98K

     STEP 5 -

               °sfc +  °*10|< " 303-9°K * 0<1°K = 3M.08K

          Using linear Interpolation from potential temperature (0) to pressure since  a
          highest  value 1s  not given for the 831 mb pressure level

                    0  (°K)         Pressure (mb)

                    303.4              850
                    304.0              P mixing height
                    304.3              831
          P mixing  height =  831  mb -  (850 mb jff1Bmb){3M.O;K - 304.3°K)


                          .  831  . 09 mbH-0 3°K)
                          =  837.3 mb
                    The  pressure at the mixing height (rounded to the nearest whole
                    millibar)  1s 837 mb.
     STEP 7  -
               Use linear Interpolation to find the height above sea level  of the
               mixing height.   Enter 3164. and 700. Into Table A-7A.

                    PRESSURE  (mb)       HEIGHT (mASL)

                         850             1537.
                         837            Z mixing height
                         700             3164.


          Z mixing height - 1537 . * ("64 . O537 m^mb - 850 mb)


                          - 1537 n, + 062? m)M3 mb)
                          - 1&3/ m +    -ISO mb

               Z  mixing  height  = 1678 m

     STEP 8 -
               1678 mASL - Height of mixing height
              -62 mASL - Elevation of urban surface site
               1616 mAGL = Mixing height 1n meters above the urban area.

          1616 m  1s the  height  of the maximum mixing height to be used In the model
          with the time  of 1700 LCT.
                                          A-16

-------
Table        Example (Hypothetical Data)
Date:                    Time of Mixing Height For Input Into  Model: I~]OO£PT
City:                                   Sounding Method: NWS.
                      o f} r\f*\                                   o
Time of Sounding:      A\J\J\J    LCT. Surface Elevation:         o.  mASL
Location of Sounding:
LOCATION OF URBAN  SURFACE DATA (IF DIFFERENT THAN ABOVE)  -
1
LEVEL
Urban
Surface
Data (1)
(2)









MIXING
HEIGHT
" 2
HEIGHT
(mASL)
U
//f
ISY7
—
3W





9sfC(°°)10K
3o*o
3
PRESSURE
(mb)
/oo8.£>
IOOQ.
ISO.
%3/.
100.





PRESSURE
(mb)
03.tA1





HEIGHT USED IN
MODEL (mAGL)
/6/6.
                                 A-17

-------
             APPENDIX B
CONSIDERATION OF PRECURSOR TRANSPORT

-------
      Transport  of pervasive,  high  precursor  concentrations into a city as a
 result  of emissions  from  upwind  sources  presents difficulties 1n applying city-
 specific  EKMA.   One  difficulty is  estimating the concentrations of precursors
 transported  into the urban  area.   Such estimates generally require special
 monitoring programs  to  detect transported pollutant levels.  Often, data may not
 be  available for a specific day  being investigated.  This is further complicated
 by  the  fact  that transported  pollutant concentrations are strongly affected by
 the meteorological conditions on a given day, making any generalizations regarding
 treatment of transport  difficult.  In spite  of these difficulties, general quanti-
 tative  techniques  have  been developed to account explicitly for the role of
 precursor transport.  As  will be described,  some sensitivity analyses have been
 performed using  these techniques.  The results indicate that consideration of
 precursor transport  1s  not  normally necessary, and is not recommended.  However,
 the techniques are presented  for those cases in which explicit treatment of pre-
 cursors is perceived by a user as  an Important consideration in the overall
 modeling  analysis.

     Three topics are discussed below:  1) the transport of precursors aloft; 2)
 the transport of precursors in the surface layer; and 3) the case in which the
 nonattainment area consists of more than one high-emission density area.
 B.I  PRECURSOR TRANSPORT  ALOFT
     Most evidence suggests that concentrations of precursors transported aloft
 are usually very low.13,11*,17,18  Generally, concentrations of NMOC measured
above the morning mixing  height have been found to be less than 0.1  ppmC.  NO
 levels transported from upwind areas above the morning mixed layer are usually
much less than concentrations attributable to the urban area.   Table B-l  summarizes
                                        B-l

-------
Table B-1. -Typical Effects of NMOC Transport Aloft on Control Estimates*



                                    Base Case NMOC Transported Aloft, ppm
03 Design Value
.18
.18
.18
.24
.24
.24
.30
.30
.30
NMOC/NOY
t\
8:1
12:1
16:1
8:1
12:1
16:1
8:1
12:1
16:1
.05
+4
+3
+1
+3
+2
+3
+2
+2
+2
.10
+8
+5
+3
+5
+4
+5
+3
+4
+3
.15
+12
+8
+4
+8
+7
+7
+5
+5
+4
   Entries in the table represent the typical difference between considering
   precursor transport aloft versus neglecting Its effect.   For example, if
   the control estimate obtained by considering transported NMOC were 50%,
   and that computed using the normal OZIPP/EKMA technique (I.e., neglecting
   transport) were 45%, a difference of 5% in control level would be recorded
   in the table.
                                        B-2

-------
the results of a sensitivity analysis that was performed to assess the effects of
considering NMOC transport.  These tests were conducted using conditions likely
to be associated with what 1s believed to be unusually high concentrations of
precursor transport.  Since areas experiencing significant precursor transport
would likely be subject to relatively high ozone concentrations transported
aloft, the concentration of ozone transported aloft was assumed equal to 0.12 ppm.
However, control estimates were made assuming that both ozone and precursors
transported aloft would be reduced according to the recommended approach described
in Sections 4.2.2 and 4.2.3.   The table shows the difference in control estimates
obtained by explicit consideration of precursor transport aloft versus using the
OZIPP/EKMA technique in the normal manner.  For example, if the control estimate
obtained by considering precursor transport explicitly were 50%, and that obtained
using the normal method were 45%, the difference of +5% would be reported in the
table.  The results reveal that differences in control estimates are small  for
concentrations of NMOC transported aloft which are less than about 0.1 ppmC.  As
a result of these findings, and the uncertainties and difficulties associated
with measuring precursors aloft, explicit consideration of precursor transport
aloft is np_t normally recommended.  If, however, conditions are such that explicit
treatment 1s perceived by the user as essential, the required additional measur-
ements and the manner in which these measurements can be employed 1n an OZIPP/
EKMA application are described below.

     The only appropriate method for estimating precursor concentrations aloft 1s
by direct measurement,  such as airborne measurements or those taken on tall
towers.   Surface measurements of precursors may not be good indications of
precursor transport aloft because of possible influences of surface layer
                                        B-3

-------
transport and local generation.  Thus, precursor transport aloft can only be

considered 1f direct measurements are available.  These measurements should be

made 1n the early morning, upwind of the city, at a height above the morning

mixed layer but below the anticipated maximum daily mixing height.  Enough

samples should be taken to Insure that five measurements are available on days

for which precursor transport is likely to be important (i.e., on which the

prevailing wind flow may be from highly industrialized or urbanized areas upwind

and conditions are conducive to ozone formation).  Normally, this requires a five

to six week sampling program.  Reference 13 contains guidance on obtaining direct

measurements.


     Explicit Consideration of Precursors Transported Aloft:  If direct measure-
ments of precursors aloft are available on the day being investigated, they may
be used directly in the modeling analysis.  If no measurements were taken on the
day being modeled, but are available for days with somewhat similar meteoro-
logical conditions (e.g., wind is from the same general direction), then the
median value of these measurements may be used.  At least five measurements taken
on days conducive to high ozone formation are necessary for estimating the median
value.  In either of the above cases, care should be taken to insure that the
measurements used are truly representative of the vertical concentration profile
above the mixed layer.  For example, a measurement of high concentration above
the mixed layer may reflect the presence of a plume within that layer, but may
not be representative of concentrations extending throughout the layer found
above the mixed layer.  Figure B-l illustrates the problem.

          The quantitative treatment of precursors transported aloft is analogous
to that of ozone transported aloft (see Section 4.2.1).  A base case diagram is
first developed with the precursor concentrations transported aloft estimated
according to the procedure described 1n the preceding paragraph.  A post-control
diagram Is then generated using precursor concentrations aloft which reflect the
effects of upwind control programs.  Section 4.2.2 describes the methodologies
for estimating future concentrations of precursors transported aloft.   The compu-
tational procedures necessary to estimate the necessary VOC reductions are also
described in Section 4.2.2, and summarized 1n Table 4-2.
                                        B-4

-------
                     I—I—I—I	i—I—i—i——i—i—i—i
                 0.00     0.25     0.50     0.75     1.00     1.35
1


1 1 1
1

1 I 1
2

i •
1

i
•4

I i i
1

6
|
i i i i
IS

i i i
c
                      OZOME,HO,riO.-.,ritlD TEHFEPHTIJPE 'S nLTITUDE
Figure B-l.   Vertical Pollutant  Profile Illustrating a Shallow Layer of
              Elevated NO^ Concentrations Aloft.
                                         B-5

-------
B.2  PRECURSOR TRANSPORT  IN THE SURFACE LAYER
     Similar to transport of precursors aloft, most evidence suggests that urban
concentrations resulting  from precursors being transported within the surface
layer are usually negligible.13,11*,17,18  Furthermore, the contribution of pre-
cursors transported in the surface layer to urban levels 1s accounted for to some
degree by using the normal OZIPP/EKMA technique.  For example, the design NMOC/NOX
ratio determined from urban measurements reflects contributions from both the
urban area and the upwind area.  To the extent that precursor reductions in both
areas parallel one another, consideration of precursor transport will have little,
if any, effect on local emission reductions.  Table B-2 summarizes the results of
some sensitivity tests designed to assess the effect of incorporating the trans-
port of NMOC within the surface.  These tests were similar to those conducted for
the transport of NMOC aloft.  In addition to surface layer transport of NMOC, the
concentration of ozone aloft was assumed to be relatively high, I.e., 0.12 ppm.
However, control estimates were made assuming that both ozone aloft and precursors
transported in the surface layer would be reduced by 40%, as described in
Sections 4.2.2 and 4.2.3.  The table shows the difference in control estimates
obtained by explicitly considering surface layer transport versus using the
OZIPP/EKMA in the normal manner (I.e., surface layer transport is not explicitly
considered).  The differences are generally small, especially for contributions
from upwind areas less than 30%.  Thus, because of the 1nsens1tivity, and because
of the approximations and assumptions that must be made in order to evaluate
surface layer transport explicitly, consideration of precursor transport within
the surface layer 1s not generally recommended.  However, if consideration of
precursor transport is perceived as a critical Issue 1n the modeling analysis,
the additional measurements which are necessary and the manner in which they are
                                        B-6

-------
Table B-2.  Typical  Effects  of NMOC  Surface Layer Transport on Control Estimates*
                                       Contribution of Surface Transport
                                       to Urban Levels, percent**

03 Design Value         NMOC/NO¥            5-10   10-20   20-30   30-40
                              /\      _____^_^_^_^__^_^_^^__^__^__	

     .18                 8:1                +1      +2    .   +3       +3

     .18                 12:1                +2      +1       +2       +3

     .18                 16:1                +1+1       +1       +2
     .24                 8:1

     .24                12:1

     .24                16:1


     .30                 8:1

     .30                12:1

     .30                16:1
*  Entries 1n the table represent the  typical  difference between considering NMOC
   transport in the surface layer versus  neglecting Its effect.   For example,
   if the control estimate obtained by considering  transported NMOC were 50%,
   and the estimate derived using the  normal OZIPP/EKMA technique (I.e.,
   neglecting transport) were 45%, a difference  of  +5% would be entered in the
   table.

** A 10% contribution implies that 90% of the  urban,  6-9 a.m.  concentration
   is due to local generation and 10%  is  due to  transport from upwind.
+3
+2
+2
+1
+3
+4
+1
+4
+5
-3
+4
+8
+1
+2
+2
+1
+2
+2
0
+3
+4
+1
+6
+7
                                        B-7

-------
used in the analysis are described below.
     The determination of  surface layer transport 1s accomplished by means of
surface based measurements taken at the upwind edge of an urban area.  It 1s
recognized, however, that  upwind measurements will not always be available for
the days of interest.  The technique described below makes use of the available
upwind measurements.  References 13 and 29 contain guidance on monitor siting and
sampling.  However, deployment of a continuous NMOC analyzer for the sole purpose
of estimating precursor transport is not recommended.  Rather, collection of a
limited number of discrete samples, analyzed chromatographically with the species
summed to yield total NMOC, is preferable for estimating upwind NMOC.  These
upwind samples should be collected between 6-9 a.m. LCT.  In order to adequately
characterize precursor transport, samples for a minimum of five days on which
meteorological conditions favor significant transport (I.e., days on which the
prevailing wind flow may be from highly industrialized on urbanized areas upwind
and conditions are conducive to ozone formation) should be available.  Normally,
a five to six week sampling program will be adequate.

     Explicit Treatment of Surface Layer Transport:  Table B-3 summarizes the
computational  procedures for considering surface layer transport of precursors.
The concepts employed to account for surface layer transport are 1) to estimate
the contribution of transported precursors to urban levels using available upwind
measurements and corresponding urban area measurements, and 2) to incorporate
these relative contributions in the OZIPP/EKMA technique.  The first two steps
involve estimating typical relative contribution of precursor transport to urban
levels.  Thus, for example, a median contribution factor of .5 implies that 50%
of the precursor concentrations measured in the urban area are due to precursor
transport.  Contribution factors should be computed separately for NMOC and/or
NO    Steps 3 through 5 entail quantifying precursor transport relative to the
normal  OZlPP/EKMA base case point.   The precursor concentration to be entered in
OZIPP as surface layer transport is calculated in Step 4, and the design NMOC/NO
ratio is adjusted to remove the relative contribution of precursor transport
(Step 5).  In Step 6, a new base case diagram is generated to reflect the effects
of precursor transport in the surface layer.  As Indicated In Step 7, the esti-
mation of future transport, development of a post-control diagram, and subsequent
                                        B-8

-------
Table B-3.  Explicit Treatment of Surface Layer Precursor Transport
1.  Compute the ratio of upwind measurements to urban average levels  for each
    day on which upwind measurements are available.


         (CF).  =  j^upwlndJi

                   lSirbanM
    where

         (CF).      = fractional contribution of transported levels to urban
                      levels for day i

                      upwind 6-9 a.m. precursor measurement for day i


         (C  .   )i  = 6-9 a.m. urban average concentration (if data from more
           uroan      than Qne „„,„.,• tor are available, the 6-9 levels  for each
                      monitor are averaged).

2.  Determine the typical transport level by taking  the median of the contn's,
    bution factors calculated in Step 1.  Let this factor be represented by CF.

3.  Generate a base case isopleth diagram neglecting the contribution of surface
    layer transport.  Determine the base case point  by finding the intersection
    of the design NMOC/NO  ratio with the isopleth corresponding to the daily
    design value.  Let the coordinates of this point be represented by (NMOC)j
    and (NO )j.  (Note that this is identical to what is done when precursor
    transport is not considered.)

4.  Estimate the relative contributions of transported precursors by  multiplying
    the contribution factor(s) by the corresponding  base coordinate(s) found in
    Step 3.  Thus,

                                          x (NMOCh

                                          x (NOJ,
               (CTRAN}NO
                        X
            where                       x

                               =  Relative Transported Concentration


               (CF)            =  Contribution Factor calculated in Step 2

               (NMOC)lt (N0x)j =  Base case coordinates found in Step 3

 5.  Compute an adjusted NMOC/NO  ratio to remove the contribution of transported
     precursors.

               Radj =    C
          where

               R    = normal  NMOC/NOX ratio calculated according  to Section 3.2.2

               R j. = an NMOC/NO  ratio adjusted to remove the  contribution of
                aaj   transported

     Develop a new base case  diagram.  The relative precursor concentrations
     calculated in Step 4 are entered in the OZIPP model  as the concentrations of
     precursors transported in the surface layer.   [If post 8 a.m.  emissions are
     considered, the emission,fractions calculated according to Section  3.1.5 must
     also be divided by l.-(CF).]


     Follow the procedures described in Section 4.2.2  for calculating VOC
     emission reductions.  The diagram developed in Step  6 serves as the base
     case diagram, and the adjusted NMOC/NO  ratio calculated in  Step 5  should be
     used as the design ratio.


                                     B-9

-------
GIVEN:  03 Design Value                     =  -24

        Design NMOC/NOX                     =  8:1

        Present Transport of 03 Aloft       =  .12

        Future Transport of 03 Aloft        =  .09

        Anticipated change in NOX           =  -30%

        Median contribution factor for NMOC =  .25

        Median contribution factor for NOX  =  0.0


FIND:   Percent reduction in VOC needed to reduce  ozone from .24  to  0-12 ppm, assuming that future surface
        transport of NMOC will be reduced by 40%.
SOLUTION:
        STEPS 1 and 2:  The median contribution factors are 0.25 for  NMOC and  0.0 for NOX-  Therefore, only
                        the transport of NMOC will be considered.

        STEP 3:  The base case diagram (neglecting transport)  is shown  below.  The base case point is found
                 in the normal fashion, i.e., the intersection of the .24 isopleth with the 8:1 NMOC/NOX<
                 The coordinates of this point are (NMOC)! = 1.85 and (NO h = .231
STEP 4:  The relative contribution of NMOC transported in the surface layer is calculated as follows:


                ^RAN^MOC =  (-25) x 1<85 =  <46

                       n   =  0.0
STEP  5:  The adjusted NMOC/NOX ratio  is computed as follows:
                ADJ
  Figure B-2.   Example emission reduction calculation with the explicit consideration of precursor transport
               in  the surface  layer.

                                                 B-10

-------
STEP 6:  The new base case diagram is shown below.  The following modifications were made to the input data
         that was used to generate the original base diagram (Step 3):  1) surface layer NMOC transport was
         set to .46 as opposed to zero for the original diagram; and 2) the emission fractions for NMOC were
         each divided by 1.0 - .25 =  .75.


                                                   " ~~   V*     ~"s"~ ' * "*':  ^
                                       •    Jr    V" '  "" " 'l* ' 11 • ' '    *-*

                                                                                      a

         The starting point (Point 1) on this diagram is found by finding the intersection of the adjusted
         NMOC/NOV ratio line (6:1 in this case) with the .24 ozone isopleth.   The coordinates of this point
         are (NM6c)i = 1.52 and (N0xh = .253.


STEP 7:  A future case diagram is developed to reflect the expected changes in transported pollutants.  Thus,
         the input data used to generate the base diagram (Step 6) is modified as follows:  1) ozone transported
         aloft is set to .09 rather than .12 (see Section 4.2.2); and 2) NMOC transported in the surface layer
         is reduced from ,46 to .28 (see Section 4.2.3).  The resultant diagram is shown below.


	 ,='.J
". '.' ••.--:~i'
...... . ., .
-i.'--:'-.r.:---::-
-L: -.-.:•::: T. •:: .

i::--:::.:.:.::...::: ;:-!qjll4-ir-rqqL'.
:---:;::tr.-:HF!:;:.::iitiH^.;rH!;
-" •- ljp*"r - : '


                           i—
                            •? •
Figure B-2.  Example emission reduction calculation with the explicit consideration of precursor transport
             in the surface layer,  (continued)

                                               B-ll

-------
         The post-control point is found in the normal  fashion.   First,  the  post-control NOX point is calculated:

                         (NOX)2 = (.253) x (1  - yjg)  =  .202


              The post-control point is then located  on the future case  diagram at  the  inter-section of  the
              .202 NO  coordinate and the 0.12 ppm ozone isopleth (Point 2).   At Point  2,  (NMOC)2 =  .40.
              The VOC emission reduction is calcualted  as

                         % reduction = (1 - y^)  = 74%


              Note that the base point used in this calculation  was that found in Step  6,  not that in Step 3.
Figure B-2.  Example emission reduction calculation with the explicit consideration of precursor  transport
             in the surface layer,   (continued)



                                                    B-12

-------
calculation of the VOC emission reduction 1s carried out according to the pro-
cedure described in Section 4.2.2.  An example 1s presented 1n Figure B-2.

     The above procedure represents an Idealized situation, so extreme care
should be taken in Its application.  First, the procedure should only be applied
to days with similar meteorological conditions (I.e., the wind 1s out of same
general direction).  Second, only days with meteorological conditions conducive
to high ozone formation should be considered.  Third, siting of upwind monitors
is critical.  Any upwind measurements that are affected by local sources will
invalidate the approach.  Also, the upwind measurements must be taken close
enough to the urban area so that further significant dilution of the tranported
precursor concentrations does not occur.  Consequently, the approach can only be
used for contribution factors less than one.  A contribution factor greater than
one indicates the assumptions underlying the approach have been violated, and the
normal OZIPP/EKMA application (i.e., precursor transport 1s not explicitly incor-
porated) must be employed.

B.3  MORE THAN ONE HIGH EMISSION DENSITY AREA
     In some instances, more than one high emission density area may be found
within the region for which a control program(s)  is being developed.  Examples
include the following:
          1)  An urban  core and a second separate,  clearly defined industrallzed
area with high emission density;

          2)  One urbanized area located in two different States, for which a
separate control  program must be devloped for each  area due to jurlsdictional
considerations;
                                        B-13

-------
           3}   Urbanized areas which  consist of two  separate, clearly defined,
 urban  cores  (e.g.,  twin cities).
 Such geographies may present special  problems  in  applying  the OZIPP/EKMA technique.

     Ideally,  the OZIPP/EKMA technique  should  be  applied 1n exactly the same
 manner as  described in Chapter 3,  I.e.,  the column  should  be assumed to originate
 in  the urban core and move directly  towards the site  of peak ozone concentration.
 The urban  core should always be used as  the origin  of column movement because the
 chemical kinetic mechanism is based  upon reactions  of NMOC/NO  mixtures representa-
                                                             J\
 tive of .automotive  exhaust.   Surface wind measurements should be analyzed to
 insure that the peak ozone level  is  located 1n the  general downwind direction of
 the urban  core of Interest.   However, an exact trajectory  cannot usually be
 determined due to uncertainties in deriving the spatial and temporal variations
 in  wind fields.   Consequently, some  assumptions in  estimating trajectories must
 \
'be  made.

     First, consider the case of an  Industrialized  area located near an urban
 area.   In  most instances, the routine OZIPP/EKMA  application will provide the
 best estimate  of overall  regional  control targets.  However, in some cases the
 industrial area may be quite removed from the  urban core.  If the two areas are
 separated  by a region of little or no source activity, the procedures outlined in
 the previous subsection on precursor transport could  be employed in those instances
 on  which the industrial area contributes to the urban area (i.e., the urban area
•is  downwind of the  industrial  region).   However,  estimates of control of the
 industrial area must be arrived at independently  of the OZIPP/EKMA technique.
 Using  the  industrial  area as the origin  for the column movement would be inappro-
 priate considering  the chemistry is  more representative of automotive exhaust
 mixtures.
                                         B-14

-------
     For the case in which an urban area is divided by a jurisdictional  boundary
(e.g., a State), the ideal method is to develop one control estimate for the
entire urban area (i.e., treat the urban area as one entity).   If this cannot be
done because of jurisdictional considerations, each area should be treated as an
entity, and the same procedure followed for each area.  Figure B-3 illustrates
the types of assumptions that could be made for trajectories under two different
situations.

     The case of two separate urban cores in a region may be treated in one of
three ways.  An independent analysis could be performed for each urban core.
Alternatively, the major city could be chosen as the design case, and the results
of that^ analysis applied to each city.  A third alternative is to use the pro-
cedures described in B.2 of this appendix when appropriate.  However, for this
procedure to be applicable, the cities must be separated by a  region of little or
no precursor source activity.
                                        B-15

-------
           7
Figure E-3,  Illustration of the Treatment  of Contiguous  Urban Areas,
                                       B-16

-------
REFERENCES

 1.  Uses, Limitations and Technical Basis of Procedures for Quantifying
     RelationshipsJtetween Photochemical Oxldants and Precursors,
     EPA-450/2-7/-021a, U.S. Environmental Protection Agency, Research Triangle
     Park, North Carolina, November 1977.

 2.  Procedures for Quantifying Relationships Between Photochemical  Oxldants
     and"^Precursors:  Supportlhg Documentation. EPA-450/2-77-021bt U.S.
     Environmental Protection Agency, Research Triangle Park, North Carolina,
     November 1978.

 3.  G. Z. Whltten and H. Hugo, User's Manual for Kinetics Model  and Ozone
     Isopleth Plotting Package, EPA-600/8-78-014a, U.S. Environmental Protection
     Agency, Research Triangle Park, North Carolina, July 1978.

 4.  Ozone Isopleth Plotting Package (OZIPP). EPA-600/8-78-014b,  U.S.
     Environmental Protection Agency, Research Triangle Park, North Carolina,
     July 1978.

 5.  Federal Register. 44 (221) 65669-65670 (November 14, 1979).

 6.  40 CFR 50.9

 7.  Federal Register, 36(84) 8186 (April 30, 1971).

 8.  T. C. Curran, Gu1del1ne_f6r the Interpretation of Ozone Air  Quality Standards,
     EPA-450/4-79-003, U.S. Environmental Protection Agency, Research Triangle
     Park, North Carolina, January 1979.

 9.  J. Trljonls, Verf1catl6n_orthe Isopleth Method for Relating Photochemical
     Oxldant to Precursors. EPA-600/3-78-019, U.S. Environmental  Protection
     Agency, Research Triangle Park, North Carolina, February 1978.

10.  M. C. Dodge, Effect or Selected Parameters on Predictions of a Photochemical
     Model, EPA-600/3-77-048, U.S. Environmental  Protection Agency,  Research
     Triangle Park, North Carolina, June 1977.

11.  J. M. Godowitch, J.  K. Chlng and J. F. Clarke, "Dissipation  of the Nocturnal
     Inversion Layer at an Urban and Rural Site In St.  Louis," Fourth Symposium
     on Turbulence, Diffusion and A1r Pollution,  Reno,  Nevada, 1979.

12.  G. C. Holzworth, Mixing Heights, Wind Speeds, and  Potential  for Urban Air
     Pollution Throughout tne Contiguous United States, AP-101, U.S.  Environmental
     Protection Agency, Research Triangle Park, North Carolina, January 1972.

13.  M. W. Chan, D. W. Allard and I.  Tombach, Ozone and Precursor Transport
     Into an Urban Area - Evaluation of Approaches, EPA-450/4-79-039, U.S.
     Environmental  Protection Agency, Research  TrTangle Park, North  Carolina,
     December 1979.
                                        R-l

-------
14.  C. E. Decker, et. a "I, Ambient Monitoring Aloft of Ozone and Precursors
     Near and Downwind of St. Louis. EPA-450/3-77-009. U.S.  Environmental
     Protection Agency, Research Triangle Park, North Carolina,  January 1977.

15.  G. T. Wolff, et. a!, "Transport of Ozone Associated with A1r Mass," paper
     presented at the 70th Annual Meeting of the A1r Pollution Control  Association
     In Toronto, Canada, June 20-24, 1977.

16.  W. S. Cleveland, et. al , "Photochemical A1r Pollution:   Transport  From the
     New York City Area Into Connecticut and Massachusetts," Science. 191,
     pp 179-181, 1976.

17.  F. L. Ludwig, Assessment of Vertical Distributions of Photochemical
     Pollutants and Meteorological Variables in the Vicinity of Urban ATeas.
     EPA-450/4-79-017, U.S. Environmental Protection Agency, Research Triangle
     Park, North Carolina, August 1979.
18.  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 Non-methane  Hydrocarbons
     In Tulsa, Oklahoma - Volume III:  Data Analysis and Interpretation,
     EPA-450/4-79-008c, U.S. Environmental Protection Agency,  Research Tri angl e
     Park, North Carolina, September 1979.
19.  Procedures for the Preparation of Emission Inventories for Volatile Organic
     Compounds. Volume I. EPA-450/2-77-028, U.S. Environmental  Protection Agency,
     Research Triangle Park, North Carolina, December 1977.  (Second Edition)

20.  Procedures for the Preparation of Emission Inventories for Volatile Organic
     Compounds, Volume II:  Emission Inventory Requirements for PhotocfiemicaT
     TTTr Quality Models, EPA-450/4-79-018. U.S. Environmental  Protection Agency.
     Research Triangle Park, North Carolina, September 1979.

21.  R. C. Wanta and W. P. Lowry, "The Meteorological Setting  for Dispersal  of
     Air Pollutants," Air Pollution. Volume U 3rd edition, A.  C. Stern, ed.,
     Academic Press, 1976, pp 337-352.

22.  G. Z. Whitten and H. Hogo, Mathematical Modeling of Simulated Photochemical
     Smog.  EPA-600/3-77-011, -U.S. Environmental Protection Agency, Research
     Triangle Park, North Carolina, January 1977.

23.  M. C. Dodge, "Combined Use of Modeling Techniques and Smog Chamber Data
     to Derive Ozone-Precursor Relationships."  EPA-600/3-77-001b, ibid., p. 881.

24.  B. Dimitriades, "Effects of Hydrocarbon and Nitrogen Oxides on Photochemical
     Smog Formation."  Environmental Sciences and  Technology 6, 253 (1972).

25.  E. L. Martinez and E. L. Meyer, Jr., "Urban-Nonurban Gradients and Their
     Significance," Proceedings. Ozone/Oxidant Interaction with the Total
     Environment Specialty Conference. Air Pollution Control  AssociationT March
     T57F; -        -
                                        R-2

-------
26.  Report of the Air Monitoring Siting Workshop, U.S. Environmental Protection
Report of the Air Monitoring siting K
Agency, Las Vegas, Nevada, July 19/6.
27.  F. L. Ludwig and E. Shelar, Site Selection for the Monitoring of
     Photochemical A1r Pollutants. EPA-450/3-78-013, U.S. Environmental
     Protection Agency, Research Triangle Park, North Carolina, April 1978.

28.  B. Dimitriades, "An Alternative to the Appendix J Method for Calculating
     Oxidant and NOa Related Control Requirements," International Conference
     on Photochemical Ox1dant_P6llution and Its Control:  Proceedings, Volume II,
     EPA-600/3-77-0016, U.S. Environmental Protection Agency, Research Triangle
     Park, North Carolina, January 1977.

29.  Guidance for Collection of Ambient Non-Methane Organic Compound (NMOC)
     Data for Use in 1982 Ozone SIP Development, and Network Design and Siting
     Criteria for the NMOC and NO  Monitors, EPA-450/4-80-011, U.S. Environmental
     Protection Agency, Research Triangle Park, North Carolina, June 1980.

30.  Technical Assistance Document for the Calibration and Operation of
     Automated Ambient Non-Methane Organic Compound Analyzers, EPA-600/4-81-015,
     U.S. Environmental Protection Agency, Research Triangle Park, North Carolina,
     in preparation.

31.  P. J. Drives, Comparison of Ambient NMHC/NO  Ratios Calculated From
     Emissions Inventories, EPA-450/3-78-026, U.S. Environmental Protection
     Agency, Research Triangle Park, North Carolina, June 1978.

32.  Mobile Source Emission Factors, EPA-400/9-78-005, U.S. Environmental
     Protection Agency, Washington, D.C., March 1978.

33.  User's Guide to MOBILE!:  Mobile Source Emissions Model. EPA-400/9-78-007,
     U.S. Environmental Protection Agency, Washington, D.C., August 1978.

34.  K. H. Lloyd, Cost and Economic Impact Assessment for Alternative Levels
     of the National Ambient Air Quality Standard for Ozone. EPA-450/5-79-002,
     U.S. Environmental Protection Agency, Research Triangle Park, North Carolina,
     February 1979.

35.  W. F. Hunt, Jr., et. al, National Air Quality and Emission Trends Report,
     1977. EPA-450/2-78-052, U.S. Environmental Protection Agency, Research
     Triangle Park,  North Carolina, December 1978.

36.  E. W. Hewson, "Meteorological Measurements," Air Pollution. Volume I,
     3rd edition, A. C. Stern, ed., Academic Press, l9?6, pp 59l::S37":

37.  D. H. Slade, Meteorology and Atomic Energy. NTIS No. TID-24190, NTIS U.S.
     Department of Commerce, Springfield, Virginia  22161, July 1968, pp 33-39.

38.  C. W. Benkley and L. L. Schulman, "Estimating Hourly Mixing Depths from
     Historical  Meteorological Data," Journal  of Applied Meteorology. 18, pp 772-
     780, June 1979.	~

-------
                                    TECHNICAL REPORT DATA
                             (Please read Instructions on the reverse before completing)
1. REPORT NO.
    EPA-450/4-80-027
                                                             3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
    Guideline  for Use of  City-specific EKMA  in
    Preparing  Ozone SIPs
                                                      5. REPORT DATE
                                                           .  ....  March 1981
                                                      6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
    G. L.  Gipson, W. P.  Freas, R. F. Kelly,  E.  L.  Meyer
                                                             8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
                                                             10. PROGRAM ELEMENT NO.
    U.S.  Environmental Protection Agency
    Office  of Air Quality  Planning and Standards
    MDAD, AMTB,  MD-14
    Research  Triangle Park,  North Carolina   27711
                                                      11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS


    Same
                                                      13. TYPE OF REPORT AND PERIOD COVERED
                                                      14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
          Report describes  how to use the city-specific EKMA model  as a means for
    estimating .emission control  requirements  needed to demonstrate attainment of
    the ozone NAAQS in 1982  SIP's.  Topics  addressed include  (a)  use of air quality,
    meteorological and emissions measurements or estimates for  generating ozone,
    isopleth  diagrams; (b) application of the ozone isopleth  diagrams in estimating
    controls  needed to attain the ozone NAAQS at each monitoring  site; and  (c)  esti-
    mation  of the city-wide  control requirement for use in the  SIP.
17.
                                 KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                               b.lDENTIFIERS/OPEN ENDED TERMS
                                                                    c.  COS AT I Field/Group
    ozone
    control  strategies
    photochemical pollutants
    models
    SIPs
    §»»
IPP
18. DISTRIBUTION STATEMENT
    Unlimited
                                               19. SECURITY CLASS (This Report!
                                         20. SECURITY CLASS (This page}
                                                                    21. NO. OF PAGES

                                                                           116
                                                                           22. PRICE
     •m 2220-1 (Rev. 4-77)   PREVIOUS EDITION °s OBSOLETE

-------
                       ADDENDUM TO EKMA GUIDELINE DOCUMENT
                                EPA-450/4-80-027
                                   March 1981
                            SECTION A.1.2 (page A-2)


     The phone number for the National Climatic Center listed in the EKMA

Guideline document is for users who want a whole year's radiosonde data on a

computer tape.  If radiosonde data for selected days are needed, they can be

obtained on hard copy by calling (704) 258-2850, extension 683.


     The cost for computer tapes with a year (or any part of a year) of radio-

sonde data from one station is as follows (each tape has one year's data for one

station):


          --$85 for the first tape

          --S45 for each additional  tape


     The cost for hard copy (from the new number above) is $0.60 per sounding.

At stations with the usual two soundings per day, the cost is $1.20 a day per

station.  Orders for hard copy must total at least $5.00.

-------
 United States
 Environmental Protection
 Agency
Office of Air. Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park NC 27711
Official Business
Penalty for Private Use
8300
Publication No. EPA-450/4-80-027
                                                                                                                               Postage and
                                                                                                                               Fees Paid
                                                                                                                               Environmental
                                                                                                                               Protection
                                                                                                                               Agency
                                                                                                                               EPA 335
                                                                          If your address is incorrect, please change on the above label:
                                                                          tear off; and return to the above address.
                                                                          If you do not desire to continue receiving this technical report
                                                                          series. CHECK HERE D ; tear off label; and return it to the
                                                                          above address.

-------