United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-80-017
September 1980
Air
Technical Basis for
Developing Control
Strategies for High
Ambient Concentrations
of Nitrogen  Dioxide

-------
TECHNICAL BASIS FOR DEVELOPING CONTROL STRATEGIES FOR

   HIGH AMBIENT CONCENTRATIONS OF NITROGEN  DIOXIDE
        U.S.  Environmental  Protection  Agency
    Office of Air Quality Planning and Standards
        Monitoring and Data Analysis Division
       Research  Triangle Park,  North Carolina
                   September 1980

-------
This document is 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  EPA contractors and
grantees, and nonprofit organizations  -  in limited quantities - from the
Library Services Office (MD-35),  U.S.  Environmental Protection Agency,
Research Triangle Park, North Carolina 27711;  or,  for a fee,  from the
National Technical Information Service,  5285 Port  Royal Road, Springfield,
Virginia 22161.
This document has been reviewed by  the  Office  of Air  Quality Planning and
Standards, U.S. Environmental Protection  Agency, and  approved for publication.
Subject to clarification,  the contents  reflect current  Agency thinking.
                    Publication No.  EPA-450/4-80-017
                                      ii

-------
TABLE OF CONTENTS



                                                                     Page



List of Tables	    v



List of Figures	   vi



Executive Summary  	   ix



 1.0 Introduction  	    1



 2.0 Emissions of NOV   	  	    3
                    X


     2.1  Anthropogenic Emissions   	    3



     2.2  Natural Sources  of NOY	    6
                               A


     2.3  Significance of  Manmade  Versus  Natural  Emissions  	    6



     2.4  Implications 	    7



 3.0 Mechanisms for N02 Formation  in  the  Atmosphere   	   10



     3.1  Photochemical Synthesis  of  Nitrogen  Dioxide   	   11



     3.2  Titration	   13



     3.3  Carryover	   18



     3.4  Implications	   20



 4.0 Monitoring,  Data  Quality and  the Observed Extent of High N02

     Concentrations	  .   22



     4.1  Network Design and Operation  	   22



     4.2  Data Quality Assurance Techniques   	   25



     4.3  N02 Concentrations Observed in  Excess of 0.20 ppm  ....   26



     4.4  Implications	   31



 5.0 Transport of N02   	   34



     5.1  Urban Gradients  of N02	   34



     5.2  Levels  of N02 Reported  in Rural Areas   	   43



     5.3  Chemistry of Transported NO  	   46
                                     /\


     5.4  Implications	   51

-------
Table of Contents (continued)                                          Page:
 6.0 Relation-ships Between N02 and its  Precursors	  .    53
     6.T  ReTatioiTsr Among N02  and Precursors	.  .    53
          6.1.1  Smog Chamber  Experiements	    54
          6.1.2  Empirical Relationships  from Ambient  Data  	    58
          6.1.3  Relationships Derived  from Chemical Kinetics Models    62
     6.2  Implications	    67
 7.0 Impact of Controlling Various Source Categories 	    71
     7.1  Tracer Approaches	    73
     7.2  Use of Peak-to-Mean  Ratios	    76
     7.3  Impact of Point Sources on  Ambient  Levels of N02  	    80
     7.4  Implications	    95
 8.0 Use of Models for N02	  .    99
     8.1  Multisource Urban Models 	   100
          8.1.1  Linear Rollback	100
          8.1.2  OZIPM	108
          8.1.3  Photochemical Box Model  (PBM)  	   110
          8.1.4  Complex Multisource  Models	112
     8.2  Models to Assess the Impact from  Individual  Sources.  ...   112
     8.3  Implications	122
 9.0 References	127
10.0 Acknowledgement 	   131
                                        IV

-------
LIST OF TABLES

Number                                                                Page

 2.1      1977 National  NO  Emissions by Source Category
          (Expressed as  N02)	       5

 2.2      NO  Emissions  in Selected Urban Counties*	       8
            /\

 4.1      Distribution of N02 Concentration >_ 0.20 ppm for  all
          California Monitoring Stations, 1975-1977* 	      30


 5.1      High N02 Values Observed in the Vicinity of Los Angeles,
          1975-1977	      36

 5.2      Spatial Distribution of Hourly N02 Concentrations  in  the
          RAPS Network,  1976	      41

 5.3      N02 Observed at Rural Sites	      44

 5.4      Concentration  of Nitrogen Oxides Observed in Clean
          Remote Areas 	      45

 5.5      Chemical Stability  of NO  Corresponding  to Observed
          Conversation Rates  	      48
 6.1       Summary of Conclusions from Smog  Chamber Experiments  .  .      57

 6.2      Empirically Estimated Dependence  of N02  on  NO   Control  .      60
                                                      A

 6.3      Estimated Impact of Reducing NMOC on Peak N02 Obtained
          with Empirical  Models  	      61

 6.4      Effect of Separate NO  and  NMOC Reductions  on Peak
          N02 Under Various Conclitions*	      63


 8.1       Sensitivity of  Peak N02  to  Changes in NMOC  with Various
          Prevailing NMOC/NO¥ Ratios  	     104
                            A

 8.2      Available Complex Multisource Photochemical  Dispersion
          Models	     113

 8.3      Approaches for  Estimating Maximum Impact on Peak N02
          from Large Individual  Sources  	     117

 8.4      Hypothetical  Example Illustrating Use of Multi  and
          Individual Source Models for Projecting  Future  Short
          Term N02 Concentrations  for Use in SIP's 	     123

-------
LIST OF FIGURES

Number                                                                Page

 3.1      Illustration  of Photochemical and Titration Formation
          of N02	   12

 3.2      Ozone Limiting  Case of Titration 	   16

 3.3      NO Limiting Case of Titration	   17

 3.4      Diurnal  Concentration  Profiles Illustrating Carryover.  .  .   19


 4.1      Seasonal Variation in  Hourly N02 Levels Exceeding
          0.20  ppm at all  non-California Stations Reporting to
          SAROAD during 1975-77  	   27

 4.2      Frequency of Maximum 1-hour N02 Concentrations >_ 0.20 ppm    28


 5.1      Gradients in Maximum hourly N02 Observed in the
          Los Angeles Basin	   37

 5.2      Example  of N02 Transport at El Toro Station on
          January  25, 1975	   39

 5.3      Pollutant plots  for Riverside, October 1-4, 1975 	   40

 5.4      Gradients in Maximum N02 Concentrations Observed During
          the St.  Louis RAPS	   42
 6.1      Nitrogen Dioxide Maximum Concentration versus Initial
         Oxides of Nitrogen  (means of several experiments)
         UNC  Study	   55

 6.2      Dependence  of Nitrogen Dioxide Maximum Concentration
         on Initial  Nitrogen Oxides, Bureau of Mines Study  ....   55

 6.3      Maximum one-hour Nitrogen Dioxide Concentration
         Produced from Irradiation of Multicomponent
         Hydrocarbon/NO  Mixtures, General Motors Study 	   56
                       /\

 6.4      Maximum Nitrogen Dioxide Concentration as a Function
         of Initial  NO  for Two Initial Hydrocarbon Levels,
         UC Riverside Study	   56

 6.5      N02, With 8 hours of Emissions and 3%/hr Dilution  ....   64

 6.6      N02, With 8 hours of Emissions and 10%/hr Dilution ....   65
                                       VI

-------
List of Figures (continued)                                          Page

 6.7      N02, With 10 hours of Emissions  and  3%/hr  Dilution  ....   66
 6.8      N02, NO and 03 Versus Time  for Point A  (Figure 6.7)   ...   68
 6.9      N02, NO and 03 Versus Time  for Point B  (Figure 6.7)   ...   69
7.1
7.2

7.3

7.4

7.5

7.6


7.7

7.8

7.9a

7.9b

7.10a

7.10b

7.11
Pollutant Plots for Lynwood, November 23-25, 1976. . . .
Example of Potential Point Source Impact at Long Beach
on Sunday, November 16, 1975 	 ,
Example of Potential Point Source Impact at Whittier
on Monday, August 25, 1975 	 ,
Dependence of Maximum/mean Ratio on Annual Mean N02
Concentrations 	 ,
Location and Elevation of Clinch River Power Plant
Stations 	 ,
N02/NOV versus NOV at Clinch River on Days with High
NOV .x 	 	 	 ,
X
N02 Ratios Observed at Clinch River During Hours with
High Ambient NO 	 ,
Evidence of NO Aloft Over St. Louis 	
X
Horizontal Plume Traverse 21 km Downwind of Labadie
Stacks on August 14, 1974 	
Horizontal Plume Traverse 21 km Downwind of Labadie
Stacks on August 14, 1974 	
Horizontal Plume Traverse 45 km Downwind of Labadie
Stacks on August 14, 1974 	
Horizontal Plume Traverse 45 km Downwind of Labadie
Stacks on August 14, 1974 	
Diagrams for a Titration Reaction in a Turbulent
. 75

. 77

. 78

81

, 83

, 85


, 86
, 87


, 89

, 90

, 92

92

          Plume	   94
 8.1       ConceptualView of the  Column Model   	109
 8.2       Schematic Diagram of Modeling Domain  for the Photochemical
          Box Model	Ill
                                       VII

-------
List of Figures (continued)                                            Page.


 8.3      Nitrogen Dixoide (N02)  Data from  the N02/03 Sampler
          Siting Study 	    115

 8.4      Procedure for Applying  Rollback in Assessing the  Impact
          of Control  Strategies on N02	    125

-------
EXECUTIVE SUMMARY



     This document has two objectives:  (1) to provide a concise description of  -



the extent and likely causes of high concentrations of nitrogen dioxide (NOpK



and (2) to identify useful procedures for evaluating the effectiveness of strategies



to reduce high concentrations of NOp-  Each chapter reviews one facet of the



subject matter.  The last section in each chapter summarizes direct implications



posed by preceding information.





     According to 1977 estimates, over 95 percent of anthropogenic NO  (i.e.,
                                                                     A


nitric oxide plus nitrogen dioxide) occurs as the result of combustion.  Approxi-



mately 90 percent of these emissions occur as nitric oxide (NO).  Thus, N02 is



primarily a secondary pollutant arising from oxidation of NO in the atmosphere.



As of 1977, there is roughly a 60-40 split in stationary-mobile source emissions



of NO  in the United States.  Natural sources emit about 20 percent of the total
     /\


NO  nationwide.  However, these emissions are diffuse and ambient N0? concentra-
  A                                                                 £.


tions in remote areas are very low.  Hence, natural  NO  emissions need not be
                                                      J\


included in the NO  emission inventories used in SIP's for N0?.
                  A                                          £.




     Conversion of NO to NOp in the atmosphere occurs via two principal mechanisms:



(1) oxidation of NO by organic compounds (.Photochemical synthesis), and (2)



oxidation of NO by ozone (.titration).  Titration occurs rapidly and is likely to



be the primary means by which large individual sources of NO  make their greatest
                                                            A


impact on hourly concentrations of N02«   A third process is the oxidation of NO



by Op (thermal conversion); however, this process is only important in and near



stacks where NO concentrations are very high.   High concentrations of NOp found



at representative monitoring sites 1/2-4 hours travel time from major source



areas of NO  and organic pollutants are likely to be largely the result of photo-
           s\


chemical synthesis.  Observed high concentrations of N02 are most likely to result



from the impact of both synthesis and titration to varying degrees.  Thus, peak  N02  is

-------
approximately proportional to N0v emissions, unless the peak results alimost solely
                                X
fvQ,m. tttira-tfiOR in which the conversiiofl. orff NO to. NOp  is limited by the  amount of
aiv.a]i,Tabl'e-©zone.  Data- from, continuous.!^ operated .monitoring, sites  suggest this
exception, iis very ira-re.  Curves, depicting diurnal patterns of NO, NOo  and  ozone
observed: at a monitoring site can serve as useful means for assessing  whether or
not an observed peak concentration of NOp is entirely attributable  to  ozone-
limited titration or nearly so.

     Proper design of NOp monitoring networks, choice of individual  sites  and
careful data quality assurance are essential prerequisites for formulating an
appropriate strategy for reducing NOp-   At a minimum, the network should incor-
porate (1) a station 1/2-4 hours travel time-(e.g.,  15 km) downwind  of the most
intense area of NO  and organic emissions, and (2) a station within  the most
                  A
intense area of VOC and NO  emissions.   In order to  obtain representative  values,
                          J\
monitors should not be located within 200 meters of  large individual sources..  In
screening the data for validity, it should be noted  that existing data suggest
that hourly NOp concentrations in excess of 0.40 ppm are unlikely.   Values, in
excess of 0.40 ppm should be further scrutinized using statistical  tests and, in
some cases, by reviewing concurrent- diurnal  data- for NO,- NOp, and ozone.

     Long range transport of NOp is not a significant factor leading to
peak hourly concentrations of NO^.   Comparison of urban-suburban data
with rural or remote data reveal concentrations in the urban-suburban
areas are very much higher.  Consequently,, in most cases, there is
little need to consider intercity transport of NOp in estimating local'
control requirements to meet any prospective short term NAAQS for NOp.
A possible exception to this generalization is the case of several  large
cities, aligned by the wind, which are  withi'rr several hours travel  titne*
of one another.  The data base underlying this report is insufficient  to
examine such a case. Eleyated concentrations of NOp sometimes observ.ed'

-------
to persist throughout the night (i.e. the "carryover" phenomenon) most likely



occur as the result of poor dilution in the vicinity of sources of NO .
                                                                     A




     Control strategies for NCL need to be coordinated with strategies for reducing



ozone.  First, experimental evidence and modeling exercises indicate that reducing



organic emissions pursuant to ozone State Implementation Plans (SIP's) may also



reduce peak NOp concentrations attributable to photochemical  synthesis.   Second,



reducing ozone diminishes NOp concentrations,  formed as the result of ozone-



limited titration, proportionally.   Third, long range transport of NCLj  while in



low concentrations, may be sufficient to aggravate violations of the ozone NAAQS



in rural areas.  Unfortunately, the present state of knowledge does not  enable



reliable quantitative estimates of this latter impact to be made.





     Use of SOo and CO data as tracers and use of peak to mean ratios appear to



be useful techniques for qualitatively assessing the impact of NO  source categories
                                                                 A


on observed high hourly concentrations of NOo-   These techniques provide justi-



fication for assumptions incorporated in models to quantitatively estimate the



impact of control programs.  Tracer and peak-to-mean techniques indicate that



observed concentrations of NOo which are greater than 0.20 ppm are almost always



attributable to ubiquitous area sources, most  of which are mobile sources.



Hence, mobile source reductions should ordinarily be more effective on



an area-wide basis than point source reductions in reducing annual



average and hourly peak concentrations of N0~.   The infrequent impact of



elevated point sources is attributable to the  following factors:   (1}



during stable stratification, elevated plumes  remain aloft; (2) corres-



pondence of wind direction and distance of maximum plume impact with the



location of a monitor is infrequent; (.3) NOo monitoring networks are



generally sparse.  In contrast, area sources have low release heights (i.e.,
                                    xi

-------
       thf surface during all meteorological  conditions)  and  are distributed
throughout the urban region in all  directions from the monitors.

     The preceding stattroents do not necessarily mean that  an elevated point
source's impact on hourly concentration is insignificant.   Modeling studies have
shown that areas downwind of major elevated point sources during peak NOp
periods may experience concentrations substantially greater than the general
urban levels associated with area sources, particularly if  ambient levels of
ozone are high.  A control  strategy for a  given urban area  should consider both
the areawide NCp levels and "hot spots" associated with large elevated point
sources.   Control of both mobile and large elevated point sources may be
required to meet a short-term NOp standard.  Control  of NO  emissions
from elevated point sources may also be highly desirable  for  other
reasons (e.g., reduction in fine particulate, acid rain-forming  potential
and impact on downwind ozone formation).

     Models to quantify the impact of control strategies  can  be  categorized as
multisource or individual source models.   Several  multisource models exist.
These range in sophistication from photochemical  air quality  simulation models
(AQSM's)  to proportional rollback.   Although  they have not  been  extensively
tested, photochemical  AQSM's offer the greatest potential for evaluating pro-
spective NOp control strategies.  They appear to be particularly attractive
alternatives in areas  which have already compiled a substantial  data base for the
design of ozone SIP's.

     Key assumptions of the proportional rollback model  also  appear to hold for
use in evaluating broad strategies to reduce  annual  and hourly NOp concentrations.
For example, both peak and mean NOp appear to be proportional  to ambient NO .  In
                                  £                                        X
addition, use of the rollback model can be supplemented with  a crude assessment
of the impact of reducing organic emissions described in  Chapter 8.   The rollback

                                           xii

-------
model represents an acceptable approach in an individual urban area only if some
heed is given to source-receptor relationships.  To reconcile findings concerning
the lack of point source impact on observed high hourly and mean concentrations
of N02 with the assumptions in rollback, the following approach is appropriate.
First, unless tracer data indicate to the contrary, do not include the emissions
from elevated point sources in the emission inventory used in rollback.  This
enables projections of future air quality to be made at existing monitoring sites
which are consistent with present observations.  Second, to assess the impact of
major individual sources on peak hourly NO^, superimpose predictions obtained
with a point source dispersion model upon representative projected values of
future N02 obtained with the rollback model.  It may then be necessary to repeat
this procedure using different (but consistent) sets of meteorological assumptions
in the rollback and point source models.

     There are several available models for estimating the impact of individual
point sources or major roadways on N02.  The following approach represents an
acceptable compromise between sufficient detail and data availability.  Utilize
a Gaussian point source model to estimate maximum concentration of NO  (i.e.,
                                                                     /\
NO + NOp).  Apply the Ozone-Limiting Approach to estimate the peak concentration
of N02.  Special cases in which a source emits N02 (rather than NO) directly into
the atmosphere can be considered using Gaussian models and modifying the inputs
of the Ozone-Limiting Approach.
                                        xiii

-------
1.0  INTRODUCTION



     This document has two objectives.  The first is to provide a concise



description of the extent and likely causes of high ambient concentrations of



nitrogen dioxide (NO^).  The second objective is to identify useful  procedures



for evaluating the effectiveness of strategies to reduce unacceptably high con-



centrations of N02,  The document is intended to be of use to persons having



responsibilities for:





          a)   determining the need for a strategy to reduce short term (e.g.,



hourly) and/or annual concentrations of NO^; and





          b)   designing and evaluating the effectiveness of a strategy for



meeting such an objective.  The information presented in this document is based



on reviews of general references on oxides of nitrogen (such as the  Oxides of



Nitrogen Criteria Document ), the results of laboratory and field studies conducted



by the EPA and others, ambient air quality data reported to the National  Aero-



metric Data Bank (NADB) during 1975-1977 from throughout the nation, and  modeling



studies.





     The document is organized in the following manner.  Chapter 2.0 identifies



sources of emissions of NOV,  The relative importance of various major source
                          X


categories is described.  Since only a small fraction of the oxides  of nitrogen



(NO ). emissions occur as N0?, it is important that the means by which NO   is
   A                       £                                            /»


converted to NO- be identified.  Chapter 3.0 describes the most important pro-




cesses by which N02 is formed in the atmosphere.  Having described the processes




by which high leyels of ambient N02 can exist in the two preceding chapters,



Chapter 4.0 describes the extent of the ambient air quality data base which has



been examined, procedures for identifying data which may be spurious and  existing



ambient levels of NQ2-

-------
              5^-7 dnscuss factcws whnfeffc may/ eantrftiiite to htfgji shcrr-fc term- ejancerin-
trra.tfonB of N%..  Ehasptear 5,0 exanriines tfre rale- of intermediate and; long range
traraisport: iffi co.nftritb;U(tw§ to' observed high levels of ambient N.CL.   Chapter 6:.(T
aidd»?es>ses neJationslrips between, NOp and its two majpr precursors.:   nitric oxide;
(NO) and volatile organic compounds (VOC)..  Laboratory, modeling and empirfcal
studies are drawn upon to surmise existing relationships, between, ambient NQ^ and
precursors which are subject to control programs.  Chapter  7.0 focuses on relation-
ships whfeh appear to exoist between observed high, levels of HQ^ and precursor
emissions froffl several major source categories.  Potential  of  various source
categories for leading to high ambient levels of NO,, is also discussed.

     Chapter 8.0 identifies and briefly discusses modeling  procedures, wh-ieh. may
prpye use.ful in quantifying the extent of controls needed to. attain air quality
goals.  Models which are suitable for urban wide applications, as  well as those
which are useful in estimating the impact of individual sources, are enumerated.
The Tast section in; each chapter summarizes the implications of material presented
in that chapter on the formulation of effective: strategies  concerning NO-.
Hence, these last sections provide a short hand review of key  implications  o.f the
present state of scientific knowledge on developing strategies for reducing
ambient levels of NCL.  The remaining portions of each chapter provide a more
indepth presentation of information leading to the identified  implications  for
strategies.

-------
2.0  EMISSIONS OF  NOY
                    A




2.1  Anthropogenic Emissions



     Oxides of nitrogen (NO ) emissions ordinarily consist of two components:
                           A


nitric oxide  (NO)  and nitrogen dioxide (N02).  Nitrogen dioxide is thus both a



primary and secondary pollutant.  A primary pollutant is one which is emitted



directly to the atmosphere from a source of pollution.  Some sources where most



of the NO  is emitted as N0« are nitric acid manufacturing plants, nitrogen



fertilizer plants  and munitions manufacturing.  The NO  emissions may be up to



100 percent of the total NO  emissions from such sources.  Another source of N0?
                           X                                                   £


emissions is  fuel  combustion.  However, only about 5-10 percent of the NO
                                                                         A


emissions from combustion sources occurs as NOp.  According to the National Air



Quality Monitoring and Emissions Trends Report for 1977, 96 percent of anthro-



pogenic NO  emissions within the U.S. arise from the combustion of fuels.   Due
          A


to the preeminance of combustion sources, the composition of emissions from



combustion sources has important implications on the assessment of ambient NOp.





     Combustion sources are comprised of two major categories:  stationary fuel



combustion (56 percent of the total NO ) and combustion by mobile sources (40 per-
                                      A\


cent).  This  leads to the conclusion that only about 10 percent of the NO
                                                                         A


emissions nationwide occur as NOg.  Because of the importance of combustion



processes as a source of NO , nitrogen dioxide can be thought of chiefly as a
                           A


secondary pollutant.  A secondary pollutant is one which is not emitted directly



by a source of pollution.  Instead, secondary pollutants arise from chemical



reactions in the atmosphere occurring among other primary and secondary pol-



lutants (i.e., nitric oxide, volatile organic compounds [VOC], ozone, etc.)  The



chemical reactions which produce N02 are discussed in Section 3.0.

-------
     Since secondary N02 is chiefly formed through  chemical  reactions  involving
nitric oxide. (NO), it is essential to investigate sources  of both N02  and NO
(i.e., sources of NO ).  For purposes of impact assessment,  manmade sources of
                    /V
NO  can be classified into four types:
  A
          1.  stationary
          2.  mobile
          3.  point
          4.  area
The preceding classifications prove useful  for selecting an  appropriate mix of
controls (mobile versus stationary) and  in modeling applications  (area versus
point).  Table 2.1 presents national  NO   emission estimates  segmented  into several
                                      /\
broad source categories.  Note that several  of the  categories identified in
Table 2.1 include both point and area, as well  as mobile and stationary sources.
The distinctions between "area" and "point," and between "mobile" and  "stationary,"
are as follow.  Individual sources which emit large quantities  of NO  are considered
                                                                    A
explicitly in models and are hence regarded as "point"  sources.   Sources which,
on an individual basis, emit relatively  small  amounts of NO   are  aggregated in
                                                          /\
models and designated as "area" sources.  For example,  a large  isolated electrical
generating power plant is considered to  be a point  source, whereas residential
heating (due to the large number and spread out nature  of  housing) would be
considered as an area source.  The distinction between  mobile and stationary is
just one of ability of a source to relocate in a very short  period of  time.
Thus, an automobile is a mobile source,  whereas a power plant is  a stationary
source.  Because the distinctions between stationary-mobile  and area-point are
not mutually exclusive, it is possible to categorize certain sources as "stationary-
point" etc., as in Table 2.1.

-------
Table 2.1  1977 National NO  Emissions by Source Category
                           /\
                      (Expressed as NOo)
     Source Category

     Fuel Combustion

       Transportation
          (mobile-area)

       Electric Utilities
          (stationary-point)

       Industrial
          (stationary—point/area)

       Residential-Commercial
          (stationary-area)
NO., (XI0° metric tons/yr)
          9.2
          7.1
          5.0
          0.9
Fraction of
Total NO,, Emissions
        X "'""   ' •••--•
     0.40


     0.31


     0.22


     0.04
     Industrial Processes

       Manufacturing
          (stationary-area)

     Miscellaneous
          (area)
          0.7


          0.1
     0.03


       0
                              Total
         23.1
     1.00

-------
2.2  Na&uKafl -Sources -df 'HO
     Thus far, Ihe -locus has iieen *on manmaxte sources of NO .  It is appropriate
                                                          /\


to next examine whether natural sources of NO  are likely to contribute signi-
                                             A


ficantly to ambient concentrations of NOp.  There is much uncertainty about the



extent of natural emissions of NO .  Estimates of xflobal NQV emissions from
                                 A                         X

                                            fi                  -3

natural sources are in the order of 20-90x10  metric tons/year.   These estimates



are arrived at indirectly by determining what they would have to be in order to



balance the global NO  cycle.  Natural emissions primarily result from biological
                     /\


and chemical transformations in the -soil and from lightning.  Relatively small



additional amounts (on a global basis.) of NOV arise from conversion of naturally
                                            X


emitted ammonia (NH3) and from large point -sources (i.e., volcanoes).  Prorating



natural NO  emissions to the United States, on the .basis of land area (i.e., the
          /\


U.S. comprises about 6 percent of the world's land area), gives a maximum estimate



of natural N.OX of about 6x10  metric tons/year emitted within the U.S.  This



.latter figure is roughly Z5 percent of the annual manmade emissions presented in



Table 2.1. and therefore 20 percent of the total (manmade + natural) NO  emitted
                                                                       /\


nationwide.





2.3  Significance of Manmade Versus Natural Emissions



     Although natural emissions constitute as much as 20 percent of the total NO
                                                                                «


.emissions within the U.S., their impact on ambient N02 concentrations is likely



to be much less.  This assertion becomes obvious when one considers emission



density rather than total  emissions.  It is reasonable to assume that natural



emissions are distributed fairly uniformly about the country.  Manmade emissions,



however, are clustered near centers of habitation or in large industrial  complexes.



If the estimated 6x10  metric tons/year of natural emissions  were distributed


                                   o

over the approximately 3,000,000 mi   of the contiguous 48 States, an NO  emission

-------
                                  2

density of two metric tons/year/mi  would result.  Comparing this emission



density with sample urban emission density figures presented in Table 2.2 indicates

                     4

dramatic differences.





     The disparity between NO  emission density attributable to natural sources
                             A


and the emission density in urban areas is reflected by ambient air quality



measurements.  Various monitoring studies conducted in remote areas, and in rural



areas of the United States, have typically reported N0« concentrations of less



than 1 ppb to 10 ppb (i.e., .001-.01 ppm). '   Maximum readings as high as 30-



40 ppb have been observed in rural areas several tens of kilometers downwind from



urban areas.   However, contributions from anthropogenic sources cannot be ruled



out on these occasions.  In contrast to ambient levels of NOp measured in rural



areas, several urban or suburban monitoring sites have reported maximum hourly

                                                   p

N0? concentrations in excess of 200 ppb (0.20 ppm).   Thus, like NO  emission
  £                                                                A


density, differences between N02 concentrations in urban and remote areas as



great as two orders of magnitude have been observed.  Prevailing concentrations



of N02 in urban, rural and remote areas will be described in greater detail  in



Chapters 4 and 5.





2.4  Implications



     Implications of the nature and extent of NOV emissions are summarized below.
                                                X




          1.   The major portion of NO  emissions occur as NO.   Hence, one needs
                                      A


to recognize the role of atmospheric chemistry resulting in the conversion of NO



to N09.  Selection of an appropriate NO  control strategy will, in part, depend
     £*                                 A


on being able to correctly identify factors influencing the conversion of NO to



N0? in the atmosphere.

-------
TABLE 2.^2  NO  Emissions in Selected Urban Counties*
             /\
Urban Area
St. Louis
Chicago
Washington, DC
Minneapolis
Richmond
Houston
County
St. Louis City
St. Louis County
Cook
Washington
Hennepi n
Henrico
Harris
Emissions
tpyt
46,676
80,946
250,644
40,794
54,243
12,980
214,419
County Area
mi2
61
499
954
61
567
232
1,723
Emission Density
tpy/mi 2
765
162
263
669
96
56
124
*  NEDS  data  base
   Area  emissions  as  of  June  6,  1979
   Point emissions  as  of October 2, 1979
 t  tpy  =  tons  per year

-------
          2.   Although most (90 percent) manmade NO  emissions occur as NO,
                                                    x\


there are some manufacturing processes which result in significant amounts of N02



being directly emitted as a primary pollutant.  Such emissions have the potential



for causing locally severe short-term peaks of N02.





          3.   On a national basis, 1977 emission estimates indicate approxi-



mately 60 percent of anthopogenic NO  emissions result from stationary sources,
                                    /\


with the remaining 40 percent arising from mobile sources.  This suggests that



potential for reducing ambient levels of N02 is available through emission reductions



from stationary or mobile sources, or both.





          4.   Although natural NO  emissions are about 25 percent of the manmade
                                  /\


emissions within the U.S., the emission density (i.e., emissions/area) of natural



emissions is very much less than that for manmade emissions in urban areas.



Ambient measurements of NO- in urban and remote areas indicate urban concentra-



tions of N02 are as much as two orders of magnitude greater than concentrations



in remote areas.  This information indicates that natural emissions of NO  would



contribute an insignificant amount to any conceivable NAAQS for N02.  Therefore,



natural emissions of NO  can be ignored in the formulation of strategies for



reducing unacceptably high levels of N02.

-------
3.0  MECHANISMS FOR N02 FORMATION IN  THE  ATMOSPHERE
     Observed concentrations of N0« are chiefly the  result of  chemical  reactions
in the ambient air involving NO and other pollutants.   In order  to  reduce NOp
levels, it is important to know how the NOp  is  formed  in the atmosphere.   Since
there are several  potential  pathways  responsible for this formation,  proper
identification of the significant path leading  to particular violations of NAAQS
for N02 is vital in selecting appropriate controls and in properly  designing
SIP's.

     In subsequent sections, different mechanisms or formation processes  will be
discussed.  Each process will  be outlined and its implications about  the  effec-
tiveness of various control  strategies described.  In  the later  sections, it will
become apparent that the mechanisms presented either involve ozone  directly or
precursors of ozone.   This indicates  that close coordination is  needed  between
N02 control  programs and ozone control programs.

     The three major phenomena with potential for leading to high NO- are:
          1.   photochemical  synthesis;
          2.   the reaction of nitric  oxide with ozone  (titration);
          3.   carryover.

     Each process will now be discussed.   Typical  diurnal patterns  of ambient NO^
concentration are presented, illustrating each  of the  above phenomena.   Next,
findings are discussed which concern:  (1) observed  frequency  with  which  each
phenomenon corresponds with high concentrations of N0?, and (2)  the correspond-
ence (or lack thereof) between the occurrence of high  NOp concentrations  and
physical location of a site.  As described in Chapter  4.0, the quantity of avail-
able N0/N02 data is much greater in California  than  in the rest  of  the  nation.
                                          10

-------
For this reason a distinction is made in the subsequent discussion between
analyses conducted with California data and with other data.

3.1  Photochemical Synthesis of Nitrogen Dioxide
     Photochemical synthesis occurs when chemical reactions are initiated by
radiant energy (sunlight).  More specifically, mixing of nitric oxide (NO) and
non-methane organic compounds (NMOC) in the presence of sunlight leads to chemical
reactions which produce NCL.  This process usually results in maximum NCL con-
centrations occurring a few hours after sunrise.

     Photochemical synthesis of NO^ may occur at any time during the  daytime
hours; however, the speed with which the NCL is formed and destroyed is a function
of sunlight intensity and precursor concentrations.  Thus, the large concentra-
tions of NO and NMOC in the mid morning (sunrise-1100 LOT) will more readily
produce larger amounts of N0~.

     Photochemical synthesis of NO^ occurs throughtout the year.  Monitoring data
described in Chapter 4.0 suggest that the severity of wintertime and summertime
NO,, peaks are comparable.  One difference between seasonal N02 peaks attributable
to synthesis is that during winter peak NOp concentrations may occur later in the
day.  This i;s dye to the -reduced sunlight intensity during the winter months and
the resulting- slower production of organic radicals causing conversion of NO to
N02.

     Figgre 3., 1 provides an illustration of summertime photochemical synthesis.
The data shown in Figure 3.1 were observed by averaging pollutant concentrations
at four sites in the St. Louis RAPS on October 1, 1976,  Typically, photochemical
synthesis begins shortly after sunrise.  The morning traffic peak combines with
                                           11

-------
                                                                                                                     °3
                                                                                                                     Titration
0..

 4
                  S.R.
                        10             12              14


                             Time ol Day (C.S.T.)



Figure 3.1   Illustration of Photochemical and Titration  Formation of
                                                                                          1C
19

-------
adverse meteorological conditions to produce the high concentrations of pre-



cursors (NMOC and NO) needed to sustain the production of high concentrations of



N0?.  Later, emissions are substantially reduced and/or meteorological  conditions



improve (note the diurnal carbon monoxide pattern roughly corresponds to the



decrease in NO  concentrations).
              /\




     A review of diurnal patterns of N02 and ozone using 1975-77 data for California



indicates that photochemical synthesis accounts for 56 percent of the 5400 site

                                                             g

hours with N02 concentrations reported in excess of 0.20 ppm.    A similar review



of data was undertaken for the rest of the United States.    The results indicate



that about 32 percent of the occasions where NO^ exceeded 0.20 ppm can  be unambi-



guously attributed to photochemical synthesis.





     The physical location of a site which primarly experiences high N02 concen-



trations, due to photochemical synthesis, is one within several (e.g.,  1-4 hours)



hours travel time of where heavy NO  and VOC emissions occur.   These areas are
                                   A


usually located in urbanized or larger industrialized areas where primary emissions



of NOY and VOC are both plentiful.
     /\




3.2  Titratjon



     Tvtration is defined as the reaction between ozone and nitric oxide.



          NO + 03 * N.02 + 02                                      (3.1)




This reaction can occur at any time of the day, since it is not dependent on the



presence of sunlight.  In reviewing diurnal patterns of N0? and associated pol-



l.uta.nts, however, the most marked impact of titration in contributing to high



levels of N02 pccurs in the late afternoon or early evening.  Ozone formed in the



late morning or early afternoon, mixes with the fresh afternoon NO  emissions to
                                      13

-------
form NOp.  This reaction occurs very rapidly "in the atmosphere,  unlike the photo-
cheiiiical'ly driven synthesis which proceeds more slowly.
     The spee'd 'of tthe "tftrat'ion reaction is clearly  il'Tustrated !t>y
measurements collected near a section of the San -Diego Freeway  in Los  Angelas.
In this study, NO  monitors were deployed upwind and at varing "intervals downwind
                 A
from the roadway.  Indications are that the NO in the roadway plume  is converted
by the 0, in dilution air..  Th'-is was observed-within 1'0's of feet from the -road-
way, indicating a conversion'time of-a few minutes or less.

     A sample diurnal concentration profile illustrating titration 'is  presented
in the right-hand portion of Fi-gure 3.1 (page 12').  Note that'the amount'of NOp
produced after 150.0  is not greatly different "than the amount of ozone  present at
1400 hours.  As Equation 3.1 states, one part NO plus • one part 0.,-yields one part
NOx.  Howeyer, when  analyzing.ambient data, other processes are taking place
simultaneously.  For -example-, direct-em'ts-s'ion's of NOp, some 'photochemical  .production
of NOp still occurring, and a smoothing effect of a one-hour ^averaging time may
lead to varied amounts of NOp being observed while CL is "declining.  Consequently,
even when titration  is the predominant means by which NOp is formed, an observed
increase in NOp may  not coincide precisely with an observed decrease  in'XL.

     The amount of NOp formed through the titration  reaction is limited either by
the available NO or  by the available 0~,  It is important to be able to identify
which of these two factors limits the NOp, because "of the underlying implications
on what would constitute an effective control strategy.  Clues as to which factor
is limiting may be obtained by examining diurnal, concentration profiles of N0£,
NO and 0.,.  Such an  examination leads to one of two possible outcomes:  (1) the

                                           14

-------
amount of N02 produced is less than the amount of ozone observed, or (2) the
ozone is depleted and excess HO is present.  Either case has the potential  to
produce more N0?.  The first case is termed "NO limiting."  This is where NO
               t.                                                            A
emissions (primarily NO) are not sufficient to mix with the ozone present,  thereby
limiting the amount of N02 produced.  The second case is called "ozone limiting,"
and occurs if insufficient ozone is present to convert available NO to N02.
Figures 3.2 and 3.3 are hypothetical diurnal patterns of 03 limiting and NO
limiting cases of titration respectively.

     In the case where titration formation of N02 is limited by available ozone,
0, control programs should be effective in reducing peak NOp-  In such a case,  a
reduction in ambient ozone levels should result in a corresponding reduction in
N02.  On the other hand, if the amount of N02 formed is limited by the availability
of NO , programs to reduce ozone will affect ambient N0? resulting from titration
     X                                                 £
to a lesser extent until the 0., is reduced to a level where its availability limits
the amount of NO converted to N02.

     Data bases in California and in the remainder of the nation have been
reviewed to ascertain the frequency with which titration reactions appear to be
the predominant cause of peak observed N02.  In California, diurnal  NO, N02
and ozone concentration patterns suggest that titration was primarily responsible
for 40 percent of the hourly N02 values in excess of 0.20 ppm.   The more limited
data base in the remainder of the nation however indicates that in only about
8 percent of the instances in which N02 is in excess of 0.20 ppm does afternoon
titration appear to be a principal contribution to the high observed N02-   An
indication as to what type of site might experience afternoon reaction of 03 and
MO is also available from the California data analysis.  In the Los Angeles  Air
                                       15

-------
a
a.


c~
o
to    .10  —
     .05  —
       12     1     2     3     4
567

    Sunrise
8     9     10
11    12

     Noon
                                                                                                                                   8     9     10   -11    12
                                                                             Time of Day





                                                             Fifiuro 3.2 O/one Liniitinq Case of Titration.

-------
I     r   I~T:
           9    10    11    12    1     23





                      Timf! of Day





          Figurn3.3 NO Limitinq Casnnf Titrntinn.

-------
Basin, sites which are usually impacted  by transported  0_ frequently experience



peak late afternoon N02 peaks in which titration  is  an  important factor.   Presence



of transported.ozone alone is not sufficient  however.   It is also necessary that



the monitor be located near large sources  of  NO .





3.3  Carryover



     Carryover is a condition in which relatively  high  NCL concentrations persist



for several hours (e.g., overnight).  Carryover can  occur at any time of day.>  If



it persists throughout the night, it  is  termed  "nighttime" carryover.  If it



occurs during the late morning to early  afternoon,  it is  termed "afternoon"



carryover.  Carryover, whenever it occurs,  is not  a  formation process.   Rather,



it serves as a base upon which other  mechanisms add  additional  NO^.   Examples of



afternoon and nighttime carryover are illustrated  in Figure 3.4.   Note that in



the example shown the carryover in the afternoon  is  carryover upon which Cu-



limited titration builds.  The carryover which  occurs in  the early morning is



nighttime carryover upon which photochemically  produced N0~ subsequently builds.





     Presence of carryover NOp as a base upon which  NCL peaks formed by synthesis



or titration can build has important  implications  concerning effectiveness of



reducing NO  emissions.  It means that it  is  most  likely  that an N0~ peak is the
           A                                                      £


result of both synthesis and titration.  Thus,  even  when  diurnal  concentration



patterns suggest that a peak is primarily  the result of ozone-limiting titration.,



reduction of NO  emissions should reduce observed  peak  N09.  The reduction in N09
               x                                         c                    f c


would occur as a result of the reduction of the base (i.e., carryover N0~) upon



which the ozone-limiting titration peak  builds.
                                           18

-------
12    1     2     3     4      5 '    6     7
9     10    11    12
                                                                    Time of Day
                                           Figure 3.4  Diurnal Concentration Profiles Illustrating Carryover.

-------
     There is no single explanation for carryover.   In  some cases it may result



from N0? emitted or produced upwind and transported  to  the  monitoring site.  Use



of the term "transport" in this context implies  a movement  of an urban plume of



N0? less than a few tens of kilometers.   In  other cases^ carryover NQp is locally



generated and then trapped in a stagnant or  very slow moving air mass.  A sea



breeze may also cause a recycling of an air  mass, thereby causing the persistance



of high N0~ concentrations.  The underlying  meteorological  basis for carryover is



that poor dispersion will not allow rapid dilution of NOp to occur.   This results



in N0? remaining high for several hours.





3.4  Implications



     1.   Since N0~ is primarily a secondary pollutant,  understanding the means



by which it is formed is important in formulating effective control  strategies.



Two formation mechanisms have been identified:   photochemical synthesis and the



titration reaction with ambient ozone.   A third  phenomenon, "carryover," results



from the persistence of high N0? over several  hours.  However, NOp which is



"carried over" originates from either synthesis, titration  or both.





     2.   If synthesis is identified as the  major cause of  observed violations,



control of NO  emissions should be effective in  reducing ambient N09.  This
             X                                                     c.


subject will be discussed further in Chapter 6.0. If titration is the prime



cause for high N0?, efforts need to be made  to determine whether NO  or ozone
                 £                                                 A


is the limiting factor.  If the limiting factor  is NO ,  reduction of NO  emissions
                                                     X                  X


should prove effective in reducing ambient NO-.   If  ambient N02 is limited by the



amount of available ozone, reducing NO  emissions should nevertheless be effective,
                                      /\


because the carryover N02 upon which a peak  attributable to ozone-limited titration



builds, should be reduced.  Only in a rare case  where diurnal curves suggest
                                          20

-------
ozone limited titration will little or no carryover would control of NO  emissions
                                                                       A


be relatively ineffective.  Even in this unusual case, some reduction of N0~



would result from NO  controls by virtue of some of the NO  being emitted as N09
                    A                                     A                    C-


as a result of instack oxidation reactions.





     3.   Reviewing the diurnal sequence of NO,,, NO and 03 concentrations is a



useful approach for estimating whether an observed N0? peak concentration is



associated with synthesis or titration.  Such information can be effectively



displayed using the graphical procedures illustrated in this chapter.  These



graphs may also be useful in detecting the presence of carryover assessing whether



N09 peaks are formed primarily as the result of titration is limited by available



NO  or by available ozone.
  J\




     4.   Although the graphical techniques described herein are useful in



identifying N0? formation processes, frequently the information they yield may be



ambiguous.  Information on monitor siting, together with the knowledge that the



titration reaction is faster than photochemical synthesis of N02, may help resolve



some ambiquities arising from diurnal concentration profiles.  If a monitor is in



close proximity to a major source of NO  (e.g., within a few hundred meters of a
                                       A>


major highway or in position to be directly impacted by a point source plume),



titration may be the predominant explanation of high observed N09.  If a monitor



is located downwind (e.g., 1/2-4 hours travel time) from major sources of NO  and
                                                                            A


volatile organic compounds (VQC)_, photochemical synthesis may be important in



leading to high concentrations of NO.
                                       21

-------
4.0  MONITORING, DATA QUALITY AND THE OBSERVED EXTENT OF HIGH  N02  CONCENTRATIONS
     In Chapters 2.0 and 3.0, NO  emissions  and processes by which these emissions
                                X
may lead to high concentrations of M02 were  discussed.   This information serves
as important input in the design of appropriate control  strategies.   There is an
even more fundamental question, however.   How does  one  know whether controls  are
required in the first place?  The need for control  programs for  N0? is  largely
determined by reviewing ambient air quality  data.   Not  only that,  but the extent
of required emission controls may depend  on  these data  as well.  Consequently,
the importance of obtaining high quality  data which are representative  of what
one is trying to measure cannot be overemphasized.   Issues pertaining to data
quality can be subdivided as follows:   (1) use of appropriate  instrumentation,
siting criteria and operating procedures,  and (2) review of recorded data to
check for spurious values.

4.1  Network Design and Operation
     Regulations and guidance concerning  appropriate instrumentation and operating
                                 12 13
principles is provided elsewhere.  '    State and Local  Air Monitoring  Stations
(SLAMS) exist to meet the data needs  of individual  State/local agencies.   As
such, it is the SLAMS network which must  be  responsive  to the  needs  for identi-
fying the extent to which a particular city  is out  of compliance with existing
NAAQS and the extent to which particular  sources of NO   may deteriorate ambient
                                                      X
air quality.

     Information and references cited in  Chapter 3.0 can be useful  in designing a
SLAMS network which is likely to detect high N02 concentrations.   Recall  from
Section 3.1 that a review of California data suggests that highest N0_  concentra-
tions attributable to synthesis occur within or immediately downwind of areas
                                        22

-------
with greatest NO  and VOC emissions.  Interpretation of diurnal  NCL patterns
                A                                                 ^*
typifying synthesis suggests that the maximum impact on ambient  NCL occurs in the
mid-morning to early afternoon within about 1-4 hours travel  time of maximum
emissions.  Thus, the location of maximum NCU attributable to synthesis depends
on the prevailing winds.  In order to cover both the case of atmospheric stag-
nation and the case of light, but identifiable, resultant winds, it is recommended
that stations be established within the downtown area and within 15 km of the
downtown area in directions of commonly occurring winds.

     Because of the need to obtain representative data in making attainment/
nonattainment decisions and in designing control strategies,  existing siting
criteria should be adhered to.  The salient features of this  guidance are:

          1)  Inlet probe height located 3-15 meters above ground and at least
one meter away from supporting structures.

          2)  There must be unrestricted air flow in at least three of the four
cardinal wind directions.  It is not desirable to locate an instrument within
20 meters of trees nor is it appropriate to position an inlet probe along a
vertical wall.

          3)  Monitors should not be located within 20 meters of any road nor
within 250 meters of any road with average daily traffic (ADT) in excess of
10,000 vehicles per day.  For roads with ADT of 1000-10,000 vehicles/day, the
monitor should be located no closer than 20-250 meters away.   The exact limita-
tion in this latter case is determined by linear interpolation based on traffic
flow.  For example, if the ADT of a nearby road is 5000 vehicle/day, the moni-
toring probe should not be within 103 meters of the road.
                                        23

-------
     If the purpose of a monitoring station is to document the impact of a
particular -source of emissions, siting specifications for the station may be
somewhat different.  If a source is a primary source of NOp and its emissions
occur near ground level, station location should be as near to the source as one
could reasonably expect the public to be  for a period consistent with that specified
in a NAAQS.  The issue of what is a "reasonable" monitoring location for such
purposes is not a technical issue, but a  political  and philosophical one.  As
such, it is outside the scope of this document.   If a source's NO  emissions
                                                                 /\
occur primarily as NO, its maximum impact will most likely arise as a result of
titration with ambient ozone.  As discussed in Section 3.2, titration of an NO
                                                                              /\
plume occurs very rapidly upon mixing with dilution air.   Hence, siting guidance
should be consistent with that for a primary source of NOp described above.

     The problem of monitor siting for estimating the impact of a large, elevated
point source of NO  is difficult to resolve.   Complications are introduced as a
                  A
result of several factors:

          1)  the probability of observing the impact of a source at any particular
monitoring station at a given time is small;

          2)  there is a tradeoff which occurs between dilution and chemistry
within a plume.  If the plume does not undergo substantial dilution, conversion
of NO to NOp via the titration reaction may be suppressed.  Conversely, if ample
dilution occurs, significant conversion of NO to NO^ may occur, but it may be
difficult to distinguish between the impact of the plume and that of background
air.
                                        24

-------
     The implications introduced by the two preceding observations will be
discussed at greater length in Chapters 7.0 and 8.0.  Maximum impact from an
elevated plume results from a complex combination of meteorology, source operating
characteristics and terrain.  Deployment of a monitoring network to estimate
maximum impact from an elevated source of NOV is best done on a case-by-case
                                            X
basis.  However, generally appropriate monitoring guidance is to orient stations
along the wind axis associated with highest ozone concentrations, paying particular
attention to intermediate distances (e.g., 2-8 km for power plants).  At those
intermediate distances, dispersion is sufficient to bring an elevated plume to
the earth's surface and to mix the plume's NO with ambient ozone.  At greater
distances, dilution may be expected to diminish concentrations.  Even if these
procedures are followed, a network may sometimes have difficulties detecting the
peak impact of a source..  Thus, modeling the impact of an elevated point source
on ambient N0? may sometimes provide a better indication of such a source's
maximum impact than data from available monitoring networks.

4.2  Data Quality Assurance Techniques
     Even assuming monitoring instruments are properly deployed and operated,
there is a need to screen the resulting data for spurious values.  As many as
three types of screening procedures could occur.  First, observed concentrations
of NOp should be compared with existing screened data from other sites.  If
values higher than NOo peaks typically reported elsewhere are observed, grounds
for closer scrutiny exist.  A second screening test which should be routinely
applied is the use of computerized statistical procedures to identify peculiar
gaps or patterns (e.g., spikes) in the data.  These procedures are easy to apply
and have been described at length elsewhere.    In the event the first two stages
                                        25

-------
fail to ."esolve whether a questionable data point  is or is not  valid,  procedures
                           15
described by Richter et.  al    can  be  used  to  provide a more rigorous  test.   These
latter .procedures require one to plot diurnal  patterns of N02,  NO,  ozone and,  if
available, carbon monoxide (CO), and  sulfur dioxide  (S02) concentrations observed
at a site.  Examination of three consecutive  days  (i.e., bracketing the day
recording the suspicious  data)  should allow the  observer to identify  errors which
otherwise would be difficult to detect.  Examples  of such errors  might include
miscoding of N02 as NOX or apparent drifts in the  baseline data recorded by an
instrument.  The exact procedures  are described  in detail elsewhere.
4.3  NOp Concentrations Observed in Excess of 0.20 ppm
     The first data quality check  recommended in Section 4.2  is to  note whether a
high recorded data point is consistent with previously observed data.   To facilitate
such a comparison, this section summarizes findings of a review of  ambient hourly
N02 concentrations of 0.20 ppm or  greater  which  were observed from  1975-1977
inclusive.  During these three years, the  vast majority of hourly N02 observations
were made within the State of California.

     During 1975-77, 161  hourly N0? concentrations of 0.20 ppm  or more were
observed from approximately 40 different sites outside of California.   All  such
values were less than 0.50 ppm, with  values in excess of 0.40 ppm being extremely
rare.  Figure 4.1 illustrates that, although  high  M02 concentrations  appear to
occur more frequently in summertime,  a significant number occur during the other
seasons as well.
     At any given location, diurnal patterns  of  N02  reflect monitor location,  as
well as meteorology and source configuration. Figure 4.2 represents  a composite
                                        26

-------
                           50-

'i'
O-
cx 60
++* H \J ^m
O
CM
O
A|
C

•H
2 30~
4-1
C
01
U
C
O
U
OJ
z 20-
^
3
O
1
CO
ccT 10.
O
O
01
"e
3
•z






37







W
I
N
I
E
R



Nov
Dec
Jan



40









S
P
R
I
N
G



Feb
Mar
Apr



45













S
U
M
M
E
R



May
June
July









_




Fall



A
U
T
U
M
N



Aug
Sept
Oct



                                        SEASONS
Figure 4.1  Seasonal variation in hourly N02 levels exceeding  0.20  ppm at
            all non-California stations reporting to SAROAD during  1975-77.
                                        27

-------
ro
oo
p.
o.
o
CM
d 20-
A|
CO
o
o
•H
oncentrat
en
CM
0
:z
M
3
E§ 10
1
,-H
1
1
1
14-1 b
O
cn
s
fc
M
>-i
sj Q-,
CJ
9





•





3
0











2
1











2
2











^n n~
345










3

6





10






7






R





8

17










9


13









10


13









.11



1?








12










3

13










3

14









5


15









5"


16










3

17


n









18






8





19


\
12








20







7




21




11







22







L
6



23
                                                  HOUR OF DAY



               Figure  4.2   Frequency of tn'aximum 1-hour N02  concentrations >^ Q.20 ppm.

-------
of all monitor stations, outside of California, in which frequency of observing
high N02 is plotted against time of day.  It is clear that high NO,, has been
observed at all times.  However, the distribution in Figure 4.2 has a distinct
bimodal pattern, with a primary peak between 0700-1200 and a secondary peak
between 1800-2300.

     In contrast to the ri02 data base outside of California, the data from that
State are extensive.  Further, screening procedures similar to those described in
                                                      o
Section 4.2 suggest that the data are of high quality.   The California data base
examined for 1975-77 consists of data from 54 stations.  Approximately 50 of
these stations were found to have at least one valid observation of hourly NOp
concentrations >_ 0.20 ppm.  Approximately half (26) of the stations are located
in the Los Angeles area (Los Angeles, Orange, Riverside and San Bernardino Counties),
with nine stations in the San Francisco Bay and six stations in the San Diego
areas.  However, more than 90 percent of the NO  values 0.20 ppm or greater occur
in the Los Angeles area.  Hence, the California data which are of greatest interest
are heavily weighted by observations in the Los Angeles area.

     Table 4.1 presents the distribution of N02 values 21 0.20 ppm observed in
California during 1975-77.  The values in Table 4.1 have been adjusted downward
by 13 percent to account for a 10-17 percent reduction which needs to be made to
             9
the raw data.   The adjustment is necessitated as a result of revisions needed in
the calibration procedure employed during 1975-77 in California.  Hence, the data
in Table 4.1 should be viewed as only approximately correct.  Nevertheless, they
are adequate to provide a basis for screening questionable data, as described in
Section 4.2.  Highlighting some of the information in Table 4.1, one sees that a
full 50 percent of the values reported to be 0.20 ppm or greater are 0.23 ppm or
                                        29

-------
   Table 4.1. Distribution of N02 Concentration >_ 0.20  ppm  for all  California  Monitoring  Stations,  1975-1977*
co
o
Concentration
(ppm)t

.20
.21
.22
.23
.24
.25
.26
.27
.28
.29
.30
.31
.32
.33
.34
.35
.36

Number, of
Hourst

631
604
448
628
252
156
169
92
104
81
119
53
40
36
32
25
15

Cumulative
Percentage
of rdgs
> .20 ppm

17.6
34.5
47.0
64.6
71.6
76.0
80.7
83.3
86.2
88.4
91.8
93.2
94.4
95.4
96.3
97.0
97.4

Approximate
Cumulative
Percentage
of all hours1

99.72




99.92



99.96





99.99


Concentration
(ppm)t

.37
.38
.39
.40
.41
.42
.43
.44
.45
.46
.47
.48
.49
.50
.51
.52
.53
.54
Number of
Hourst

31
9
11
10
6
7
2
5
2
2
2
1
2
1
0
2
0
1
Cumulative
Percentage
of rdgs
> .20 ppm

98.2
98.5
98.8
99.08
99.25
99.44
99.50
99.64
99.69
99.75
99.80
99.83
99.89
99.92
99.92
99.97
99.97
100.00
Approximate
Cumulative
Percentage
All Hours

>99.99

















   *   Where  simultaneous  chemiluminescent  and  colorimetric  measurements were made,  only1
       colorimetric  data  have  been  used  in  this table.
   t   Median of  all  concentrations :>  0.20  =  0.23  ppm
   t.   Total  hours >  0.20 ppm  = 3579

   J   All  hours  estimated by  multiplying (54 stations)(3 years)(365 d/yr)(24 hr/d)(.75 assumed
       fraction of reported values  ^ 1,060,000.

-------
less.  Second, only about 0.3 percent of the estimated total  number of hourly N02
concentrations reported are 0.20 ppm or greater.  Directing attention to the
upper end of the scale, less than 1 percent of the values reported as 0.20 ppm or
greater were _> 0.40 ppm.  These high values constitute less than 0.01 percent of
the estii-nated total number of observations.  Out of a sample of approximately
one million site-hours, there are only four hourly values reported as high as
0.50 ppn or greater.

     Hence, both the California and non-California data suggest that hourly N02
concentrations of 0.40 ppm or more are likely to be extremely rare.  There are
other similarities between the data bases as well.  In California, as elsewhere,
NOp concentrations of 0.20 ppm or more occur throughout the year.   However, in
California, values in excess of 0.20 ppm appear to occur most frequently during
the winter months.  Similarly, high NOp concentrations can occur during any hour
of the day, but are most pronounced during 0600-1300 and 1400-2100 local  time.

4.4  Implications
     1.   Proper design of a N0? monitoring network, use of reference or equivalent
measurement methods, and adherence to careful quality assurance procedures are
essential prerequisites to the design of appropriate control  strategies.   Published
guidance should be followed on the use of appropriate instrumentation and quality
assurance procedures.    Network design needs to recognize the rapidity with
which N0_ is formed, source configuration and prevailing meteorology.  Monitoring
stations in the following locations are suggested.

          a)  Stations should be located one-half to four hours travel  time
downwind from the most intense area of VOC and NO  emissions (e.g., 15 km) in
                                                 A
commonly prevailing daytime wind directions with light winds.
                                        31

-------
          b)  At least one station should be located within the area of most
intense HOY and VOC emissions to account for stagnation  conditions, possible
          f\
carryover and high N02 due to titration.  However,  such  station(s) should be
located at least 200 meters from major sources  so that the station(s) is not
unduly impacted by nearby individual  sources.

          c)  Efforts to assess the impact on N02 of a single source of NO using
monitoring data will most likely require a special  purpose monitoring network.
Such a network is best designed on the basis of case-by-case modeling studies.
However, it is generally most appropriate to orient stations along the wind axis
most closely associated with highest  ozone concentrations.   Maximum concentrations
may generally be expected to occur at intermediate  distances downwind (e.g., 2-
8 km).

     2.   A second essential prerequisite to designing an  appropriate SIP control
strategies is to screen ambient data  for erroneous  values.   Three screening
procedures have been identified:

          a)  Compare values with high N02 concentrations  observed in the city or
in other similar cities elsewhere.  If a value  is outside  the range of existing
observations, it should be subject to further scrutiny.

          b)  Employ routine computerized tests which have been designed to flag
questionable values.

          c)  Plot diurnal concentration patterns for N09,  NO,  0, and, if possible,
                                                        L.        O
CO and S0?.  Curves should be plotted for 72 hour periods,  centered on the day
with questionable N02 readings.  The  diurnal curves can  then be used to identify
inconsistencies among pollutants.
                                        32

-------
Because they are easy to apply, it is recommended that the first two procedures
be applied on a routine basis.  The third procedure should be used to help resolve
uncertainties about individual data points which are raised by either of the
first two tests.

     3.   A review of ambient NO- data from 1975-1977 inclusive suggests the
following:

          a)  Ambient NCL concentrations as great as 0.40 ppm are extremely rare
indeed.  Values in excess of 0.40 ppm should be given close scrutiny prior to
basing a control strategy upon them.

          b)  Hourly NOo concentrations in excess of 0.20 ppm may occur during
any time of the year.  Hence, it is not appropriate to confine monitoring efforts
to certain seasons of the year.
                                        33

-------
5.0  TRANSPORT OF N02
     In discussing "transport"  of a  pollutant,  it  is  important to identify the
geographic seale^of the transport.   Transport of high concentrations  of NOo,
unlike ozone, appears to be limited  to  perhaps  100-200 km,  even under adverse
meteorological conditions.   In  support  of  this  contention,  Section 5.1  describes
N02 gradients observed in the vicinity  of  urban areas.  Section 5.2 then summarizes
recent observations made in rural  areas.   Section  5.3 summarizes available informa-
tion on the chemical stability  of NO (i.e., NO +  N09) in  the atmosphere and  the
                                    /\               £
nature of the reaction products.   This  latter information  helps provide insight
into possible upper limits  to the transport of  high NO  concentrations  and into
                                                      /\
relationships with long range transport of ozone.  Finally,  Section 5.4 identifies
the implications of the foregoing information on the  design  of appropriate control
strategies to reduce unacceptably high  levels of N02  and ozone.

5.1  Urban Gradients of NOo
     The two urban areas with the most  extensive available  NOo monitoring data
within the U.S. are Los Angeles and  St. Louis (as  a result  of the RAPS  data
base).  These two cities represent distinctly different examples for  characterizing
urban N02 gradients.  In the former  case,  the city is surrounded by many miles of
urban development.  Further, the area is frequently characterized by  unfavorable
meteorology (i.e., light winds  and the  presence of a  subsidence inversion restricting
the extent of vertical mixing).  Finally,  the terrain in the Los Angeles Basin
frequently serves to channel the flow of pollutants to downwind areas.   These
conditions combine to maximize  NO^ levels, as well as the  area over which high
N02 can occur.
                                       34

-------
     In contrast, St. Louis is surrounded by less extensive suburbs.  Prevailing
meteorology is more favorable to the rapid dilution of pollutants, and there is
little channelization of windflow.  Hence, a review of St. Louis data should
provide insight into the extent of transport from an urban area under more favor-
able (and, perhaps, more typical) conditions.
     Table 5.1 presents maximum hourly NCL concentrations and the frequency with
which high (> 0.20 ppm) NO^ concentrations occurred at 23 monitoring sites in the
vicinity of Los Angeles during 1975-77.  Figure 5.1 has grouped the data in
Table 5.1 into the following categories:  observations made within 10 miles of
downtown Los Angeles; observations made 10-20 miles away; 30-50 miles away; 50-
70 miles away.  It is clear that, despite the fact that N02 is a secondary pol-
lutant, the MOp problem is likely to be most severe in the vicinity of source-
intensive areas.  For example, in Figure 5.1, the average number of hourly N02
concentrations in excess of 0.20 ppm diminishes from 439 at sites within 10 miles
of downtown Los Angeles to 17 at sites 50-70 miles from Los Angeles.  Similar
reductions in maximum hourly N02 concentrations are experienced as well.  In
Figure 5.1, for example, the typical maximum hourly N02 concentration of 0.50 ppm
observed within 10 miles of downtown Los Angeles is reduced to 0.27 ppm at sites
50-70 miles away.
     It should be remembered that the Los Angeles area is one of sprawling urban
development.  Hence, the frequency and magnitude of recorded high hourly concen-
trations of N02 are not entirely attributable to transport from the most source-
intensive area (50-70 miles upwind).  Methods similar to those described in
Chapter 3.0 can sometimes be used to distinguish whether observed high hourly
concentrations of N02 are attributable to local emissions or to emissions from
                                        35

-------
TABLE 5.1  High N02 Values Observed in the Vicinity of
           Los Angeles, 1975-1977
Distance East
of Downtown
LA, (mi)
    -12
    -10
     -5
     -2
      0
      5
     10
     14
     15
     23
     23
     32
     35
     38
     38
     38
     52
     52
     55
     55
     52
     55
     67
   Site
LA Gault Street
LA Westwood
Lennox
Burbank
LA San Pedro
Lynwood
Pasadena
Temple City
Whittier
La Habra
Azusa
Pomona
Chino
Upland
Upland
Upland
Rubidoux
Rubidoux
Riverside
Riverside
Fontana
San Bernardino
Redlands
No. of
Hourly*
Rdgs > 0.20 ppm
Max. Cone.*
205
628
297
386
633
163
526
264
386
212
166
231
47
13
38
17
8
5
82
31
4
11
9
.35
.62
.40
.48
.62
.39
.46
.47
.62
.47
.38
.41
.38
.34
.30
.27
.28
.26
.35
.30
.24
.25
.29
*  Numbers have not been adjusted downward to account for 10-17% over-
   estimation resulting from calibration method used during 1975-77.
                                   36

-------
                                                                                                        —  r>oo
                                               O  Maximum Hourly Concentration
                                               A  Number of Observations > 0.20 ppm
                 20                  30                  10

                       Miles East or West of Downtown Los Angeles
Fi(|iirf» 5.1  (ir;ulionts in Maximum Hourly MO^ Observed in tho Los Anrjolos Rnsin.

-------
cities which are a few tens of miles upwind.  For example, in Figure  5.2,  an  N02
peak corresponds with increases in CO and 0, during the mid-afternoon, whereas  NO
concentrations remain low throughout the day.  The site (El Toro) is  not one
which is characterized by high local emissions.  The N02 peak* in this case  is
apparently attributable to transport.   In contrast, the N02 peaks illustrated for
October 2-4, 1975 in Riverside, California  in Figure 5.3 suggest local formation
of NOp attributable to synthesis (morning)  and titration (evening).   In the
Los Angeles area data base, instances of local generation of M02 in outlying
areas appear to be much more common than instances of transport.

     Because of the role of local  NO  and VOC emissions in the Los Angeles area,
                                   /\
Figure 5.1 may imply that transport of  NOp  is more important than it  actually is.
For this reason, it is informative to review N02 gradients observed in the St.  Louis
area during the RAPS study.  Table 5.2  and  Figure 5.4 show that, as in Los Angeles,
maximum hourly N02 concentrations  diminish  with increasing distances  from  downtown
St. Louis.  The difference is that the  gradient is somewhat sharper (e.g., within
40 km, maximum, Cgg and C™ concentrations  are less than half those observed
downtown).  A second difference is that the NOp levels in and near St. Louis  are
consistently lower than those in Los Angeles.  This second difference is probably
attributable to more favorable meteorology  and somewhat less extensive emissions
in the St. Louis area. . The Cgg and C,-Q values shown in Figure 5.4 approach
values which are considerably less than 0.10 ppm.  Although maximum N02 concen-
trations 50 km from St. Louis are  about 0.08 ppm, the infrequency with which
* The NOp concentrations shown in  Figure  5.2  need  to be reduced by  10-17  percent,
  as a result of the calibration procedure  utilized.
                                        38

-------
   0.8
   0.7
   0.6
   0.5
   0.4
                                                                                .2
a
 I

z
z
UJ
o

o

0  0.3
   0.2
   0.1
                                                     NO.
                                          12

                                      TIME - hours
                                                      16
20
            24
     Figure 5.2  Example of NOa  transport at  El Tore station on

                  January 25, 1975.   CO has been scaled by  0.01.
                                      .39

-------
    :Q.<8
    0.7
    0.6
2
8:  o.5
P  0.4

-------
TABLE 5.2  Spatial  Distribution of Hourly N02 Concentrations in the
           RAPS Network, 1976
Site
101
102
103
104
105
106
107
110
111
112
108
109
113
114
117
118
119
120
115
116
121
123
124
122
125
Distance
From Center
City (km)
0
1 10
1 10
1 10
1 10
1 10
1 10
£ 10
1 10
1 10
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
20-30
20-30
20-30
30-40
40-50
50-60
50-60
Valid
Hours
6147
6187
5115
5353
4554
5872
6164
7180
6446
6105
6322
6846
5292
7057
6348
6640
6189
6420
7048
6371
5998
6012
6481
6966
6527
Maximum
Cone . (ppm)
.256
.194
.163
.156
.186
.245
.191
.140
.223
.169
.155
.154
.266
.162
.082
.079
.192
.192
.090
.081
.168
.087
.082
.086
.067
C99
(ppm)
.08
.09
.12
.07
.07
.09
.08
.06
.07
.08
.06
.06
.09
.08
.05
.05
.06
.07
.05
.05
.07
.04
.04
.04
.03
CSQ
.026
.031
.020
.024
.026
.029
.029
.015
.022
.025
.014
.010
.022
.012
.008
.009
.015
.016
.009
.009
.008
.005
.003
.002
.001
                                        41

-------
                                                                    A Maximum Hourly Average
                                                                    T Runiji; of Observations
                                                                                      {.]—
                                                                            10
                                      Disunite friini Centor City, km

Fiijiiit! b.4  (jiiiilicnl:, in IVIaxiiiiUnt TOO-, Coiiciiiiliiilions Uhsuivi;d Dtiiniii  Hie Si. Lour.,
.1...
50
                                                                                                           I

-------
transport appears to be the primary cause of high N02 in outlying areas in
Southern California suggests that it is unlikely that high N02 observed in a city
is significantly impacted by N02 transported from a city more than 50-70 miles
away.  This hypothesis is further examined in Section 5.2 where N02 observations
made in rural and remote areas are presented.

5.2  Levels of NOp Reported in Rural Areas
     Concentrations of NOo observed in several recent field studies are summarized
in Table 5.3.  As with the data collected in outlying areas of St. Louis, ambient
N02 concentrations in rural areas are almost always (Cgg) very low.  With rare
exceptions, even the maximum hourly N02 concentrations observed are less than
0.05 ppm.

     Table 5.4 shows mean NO/NO^/NO  concentrations observed in remote areas.
                               t   A
The distinction between "remote" and "rural" is that a remote location is far
from populated areas.  Hence, its pollutant levels are not expected to be influenced
                         3
by anthropogenic sources.   A rural site, on the other hand, is one which is more
likely to be impacted by transport from urban areas.  Rural sites may include
small communities which may generate limited amounts of pollutants, but no major
        3
sources.

     In comparing Table 5.4 with the C™ values in Table 5.3, one sees that there
are, most probably, measurable amounts of residual N02 present at rural  sites.
However, these concentrations are usually an order of magnitude of more below
0.10 ppm.
                                        43

-------
TABLE 5.3  ?NQ2 Observed at Rural Sites
 Site
 Kane, PA
 Coshocton, OH
 Lewisburg, W. VA
 Sampling Period
 June 26-September 30,  1973
 June 26-September 30,  1973
 June 26-September 30,  1973
Approximate
'Max. N02
Cone .  (ppm)
    .045
    .060
    .085
.020
.035
.030
                      .010
                      .015
                      .010
Wilmington, OH
McConnelsville, OH
Wooster, OH
McHenry, MD
Dubois, PA
 June 14-August  31,  1974
 June 14-August  31,  1974
 June 14-August  31,  1974
 June 14-August  31,  1974
 June 14-August  31,  1974
.045
.035
.045
.030
.035
.035
.020
.035
.020
.020
                         .005
                         .005
                         .005
                         .005
                         .005
Bradford, PA
Creston, IA
DeRidder, LA
Wolf Point, MT
Simsbury, CT
June  27-September  30,  1975
June  27-September  30,  1975
June  27-October  31, 1975
June  27-September  30,  1975
July  15-August 21, 1975
.020
.010
.025
.025
.023
.015
.010
.010
.010
 NA
                         .005
                         .005
                         .005
                          ND
                          NA
Montague, MA*
Duncum Falls, OH*
August  1-December 31, 1977
.August  1-December 31, 1977
.073
.043
.016
.018
                         .007
                         .008
Giles County, TN*
Lewisburg, W. VA
August  1-December 31, 1977
August  1-December 31, 1977
.055
.028
.024
.009
                         .011
                         .004
t   Mean daily maximum, approximately equal to Cgg for the sampling period shown
++  Mean rather than C50
ND  Not Detectable
NA  Not Available
 *  SURE data provided courtesy of EPRI
                                       44

-------
        TABLE 5.4  Concentration of Nitrogen Oxides  Observed in Clean Remote Areas
Location
La ramie, Wyoming
Fritz Peak,
Colorado
Fritz Peak,
Colorado
Northern
Michigan
Tropical Areas




Ireland
Concentration (ppb)
:»
X
O.I - 0.4

0.2 - 0.5
0.3 - 0.5






NO
0.01 - 0.05



0.1 - 0.5
0.3 - 0.4
0.3 - 0.7

0.3 - 0.7
SO. 2
N02

<0.1


0.2 - 0.4
0.3 - 0.5
0.3 - 0.6

0.3 - 0.5
0.2 - 1.5
Measurement
Method
Chemi luminescent
Absorption
spectroscopy
Chemi luminescent
Chemi luminescent
Saltzmnn




Chemi luminescent
Date of
Measurement
Summer 1975
Fall 1974;
Summer-
Spring,
1975-1976
Sep 1977
Jun 1977
1965-1966




Jul-Nov
1974
c 	 • — -
Reference
Drummond (1977)
Noxon
(1975, 1978)
Ritter et al.
(1978)
Ritter et al.
(1978)
Lodge and Pate
(1966);
Lodge et al.
(1974)
Lodge and Pate
(1966);
Lodge et al.
(1974)
Lodge and Pate
(1966);
Lodge et al.
(1974)
Lodge and Pate
(1966);
Lodge et al.
(1974)
Cox (1977)
Remarks



10-day average
Under forest
canopy
Above forest
canopy
River bank

Seashore
and maritime
Maritime. Maximum
hourly averages
ranged from 0. 3
to 5.0 ppb.
-P.
en

-------
§,.,3  CMffistipy of;
           ---
                 preaejitedi m the two previous sections indicates that, only very
                   of NsEk/NQ-  are 1 ikely to. be transported over
                        £-   J\


distances.  In this section, observations concerning chemistry of NO  as it is
                                                                    /\


transported are briefly described.  This information is pertinent for two reasons:





          1)  it provides an upper limit estimate of how far high concentrations;



of NOo may be transported under certain conditions;





          2)  it brings, into better focus the interrelationships, between trans-



ported NO  and transported ozone.





In discussing chemical stability, it is more appropriate to consider stability of



NO  (NO + N02) rather than N02.  The reason for this is that within a weTT mixed



urban Djlume subject to photochemical reactions., NO is rapidly converted to N02



and vice versa..  Stability of NOY, however, allows consideration to be given not
                                J\


only to N02,, but to NO available to rapidly replenish any lost N02.





     Recently, a number of efforts have been made- to estimate the chemical



stability of ambient NOV and to determine its reaction products.16' 17»18» 3
                       X


Measurements made downwind of Los Angeles have been used to infer NOV decay rates
                                                                    X


of 2-16 percent/hour, with typical conversion rates being about 10 percent/hour



or slightly less.  However, it has been hypothesized that such estimates may



underestimate the conversion rate of NO .  This, hypothesis follows from the
                                       A


manner in which the conversion rate estimates are made and from the fact that



there is a continuing infusion of fresh NO from cities downwind of Los Angeles.



The hypothesis that a 10 percent/hour NO  conversion rate may be too low has been
                                        A


tested in two sets of experiments.  The first set was made downwind of Phoenixr a.
                                        46,

-------
city with relatively little surrounding development.  Surprisingly, NO  conversion
                                                                      /\

                                                                      18
rates of only about 5 percent/hour were observed in these experiments.    More



recent experiments conducted in the Boston plume, as it is transported over the



ocean, have indicated NO  conversion rates of 14-24 percent/hour, which average
                        A

                      18
about 17 percent/hour.    Corresponding lifetimes for NO  are shown in Table 5.5.
                                                        /\




     All of the chemical lifetimes in Table 5.5 are based on experimental data



obtained on sunny days in photochemically active urban plumes.  It is not possible



to draw definitive conclusions about NO  lifetimes at night, outside of urban
                                       X


plumes or under other meteorological conditions, except to say chemical lifetimes



of NO  may be somewhat longer under conditions which are not conducive to photo-
     A


chemical reactions.  However, data presented in Chapters 3.0 and 4.0 imply that



N02 is at its highest within urban plumes and in sunny weather.   This allows for



some speculation about maximum distances for transport of high NO. concentrations.



If one is willing to assume that daytime NO  conversion rate is  at least 5 percent/
                                           /\


hour and most probably greater than 10 percent/hour, maximum travel distance of



significant N02 is probably limited to 100-300 miles.*  This estimate represents



a maximum one, because the impact of atmospheric dilution is not taken into



account.  It is clear from the minute concentrations of NO  reaction products
                                                          /\


(mostly PAN and nitric acid) that dilution is the primary means  by which ambient



NO  is reduced.  In addition, the information presented in Sections 5.1 and 5.2
  J\


also support the assertion that dilution is the most important factor causing the



distinct differences observed between N02 concentrations in urban and rural



areas.
* This estimate was arrived at by assuming a resultant wind velocity of 10 mph

  prevailing throughout the chemical lifetime of NO  as reported in Table 5.5.

  Nighttime decay rate is assumed to be 0 percent/hour in these calculations.
                                        47

-------
TABLES,. 5  'Chemical  StattyiTfty of ?NQ  Corresponding to Observed
           Gonversafcram
Conversion Rate
(%/hrV
,2
5
10
'14
17
24
Half Life*
(hr.)
35
13
6
5
3.7
2.5
'Lifetime*
(hr.)
-50
1-9
9
7
5.3
3.6
     '*  .Lif-etatme -Is  commonly .as-sumeS 'to be rtfhe 'time required for
        -a .pollutant  :to '.decay -to  V/e or 'to 37% of its original concentration
        Time required -for a  pollutant to '.decay to 50%  of  its original
        concentration
                                   48

-------
     A discussion of transport of NO  is not complete without mention of the
                                    X


interrelationships with transport of ozone.  There are really two issues which



need airing:



          1)  What is the impact of ozone, transported from upwind areas, on N0_?
          2)  What is the impact of transported NO  on downwind concentrations of
ozone?
     The impact of transported ozone on NOp can be seen by referring once again



to Figure 5.3 (page 40).  The dashed lines in the upper set of panels in this



figure indicate that, during the period in question, ozone concentrations on



October 1 and 4 reached Riverside, California from points further west about



4 p.m.  This ozone then reacts with local  NO emissions resulting in N02 concen-



trations approaching 0.20 ppm in the early evening.   This, of course, is an



example of the titration phenomenon discussed in Chapter 3.0.  Control programs



resulting in the reduction of transported ozone should therefore reduce peak N02



concentrations resulting from titration of NO with ozone when ozone is the limiting



factor.





     The impact of transported NO  on downwind ozone is a difficult issue.
                                 /\


Formation of ozone from NO  and NMOC is not the result of a chain-ending reaction
                          A


(i.e., in addition to ozone, NO  and additional organic radicals are also formed
                               /\


as by-products which, in turn, can recombine several more times to form more



ozone).  Therefore, concentrations of NO  well below 0.10 ppm are sufficient to
                                        /\


sustain synthesis of ozone in excess of the NAAQS for ozone.  This observation,



in turn, leads to the following question:   should NOX emissions be controlled



even though they pose no threat in terms of unacceptably high concentrations of



N02?
                                        49

-------
     At this point, control agencies  are confronted  with an  as-yet-to-be-resolved
dilemma.  Model simulations, as well  as  ambient data,  show that  the most immediate
effect of NO emissions is to suppress ambient levels oT ozone wi'thrn "the vicinity
of the emissions.  Since areas with greatest  emissions are usually those in which
the greatest number of people reside, there is a concern that aggressive programs
to curtail NO  emissions in such areas may increase  population exposure to high
             A
concentrations of ozone.  Counterbalancing these arguments are those (briefly
alluded to above) which suggest that  the availability  of NO  may aggravate highest
                                                          A
ozone concentrations in downwind areas and that reduction of available NO  in
                                                                        • A
rural areas may be needed to curb pervasive ozone problems reported in the north-
eastern U.S. and elsewhere.  There is some support for the argument that NO
                                                                           A
emissions aggravate peak ozone concentrations downwind which can be found in the
                       19
Los Angeles trend data.    Arguments  have also been  made that NO  should also be
                                                                A
controlled in order to reduce other oxidants  as well (e.g.,  PAN).

     Both the arguments for and against  aggressive controls  of NO  emissions
                                                                 A
appear plausible.  Unfortunately, the state-of-the-art has not sufficiently
advanced to allow quantitative tradeoffs to be made.  Indeed, much of the research
cited in this section, as well as other  work, has been conducted for the purpose
of developing information needed to make these quantitative  tradeoffs.  It is not
likely, however, that the problem will be satisfactorily resolved within an
immediate time frame.  Therefore, the foregoing information  has  been presented to
alert control officials to the possibility that more stringent NO  controls than
                                                                 A
those implied by any hypothetical NAAQS  for NOo may  be needed.
                                        50

-------
5.4  Implications
     1.   Despite the fact that N02 is a secondary pollutant, it does not appear
that high concentrations are transported over great distances.   Highest concen-
trations of N02 occur near areas with high NO  emission density.  While it is
possible for high NO  to persist throughout the night (i.e.,  the carryover phenomenon),
comparison of urban, suburban and rural data implies that such  persistence results
from poor dispersion in the vicinity of source-intensive areas, rather than as  a
result of transport from afar.

     2.   Chemical decay rates of NO  (i.e., NO + NO )  serve  to provide an upper
                                    A               L-
limit estimate for transport of appreciable concentrations of N02.   Although
there is some spread in the data, decay rates of about  10 percent per hour appear
to be representative of daytime observations within an  urban  plume.   Assuming a
resultant wind speed of 10 mph, an upper limit for transport  of high concentrations
of MOY of about 100-300 miles is estimated.
     /\
     3.   Available urban concentration gradient data,  as well  as rural  data
however, suggest that transport of N02 approaching 0.10 ppm is, in fact, limited
by atmospheric dilution.  In the RAPS network, for example, peak hourly N02 is
reduced from .25 ppm to .08 ppm within 50 km of downtown.  Rural  data indicate  that
N02 concentrations seldom exceed .03 ppm.  Therefore, if one  is primarily concerned
with an hourly N02 concentration of 0.20 ppm or more, impact  of long range trans-
port of NO  can ordinarily be ignored in the design of  control  strategies for
     4.   Use of graphs plotting diurnal  concentrations of NO.,  NO,  03,  CO  and
S0? may be useful in identifying whether an observed high concentration  of  NO.
a downwind area is due to transport of N02-  Using such a procedure, high N02
                                        51

-------
observed in areas downwind of Los Angeles were frequently found to result primarily



from the titration of locally generated NO with ozone which was transported from



t*pv««d Siau-rees.  Heacfr, programs to reduce transported ozone nay also reduce peak



NO- generated in downwind areas under some conditions.





     5.   Principal reaction products of NO  are PAN and nitric acid.  These
                                           A


pollutants exist in minute quantities in the  atmosphere.  Under some conditions,



PAN can regenerate NOp.  In addition, comparison of ruraT with remote data



indicates that low concentrations of residual  NOo can survive over considerable



distances.  Unfortunately, it only requires very small quantities of NOV (well
                                                                       A


below 0.10 ppm) in the presence of NMOC to sustain ozone concentrations above the



NAAQS for ozone.  This suggests that more stringent requirements for controlling



NO  emissions may be needed to meet the ozone  NAAQS than to avoid unhealthy
  A


exposures to MO-..  Unfortunately, quantitative tradeoffs between possible increases



in ozone levels in the vicinity of sources of  NO versus possible decreases in



ozone levels observed downwind of sources of NO cannot be made at the present



time.  While information is being obtained to  develop better tools to quantify



the impact of NO  controls on ozone levels in  rural areas, it is probably a good
                A


idea to minimize any increases in NO  emissions as a part of ozone SIP's.
                                    A
                                        52

-------
6.0  RELATIONSHIPS BETWEEN N02 AND ITS PRECURSORS
     Previous chapters have established that ambient N0_ is primarily a "secondary"
pollutant.  It has also been noted that nitric oxide (NO), non-methane organic
compounds (NMOC),* and ozone (03) play important roles in the formation of N02 in
ambient air.  The purpose of this chapter is to discuss relationships between NO
and these other pollutants in greater detail.  Three means have been used to
examine relationships between N02 and its precursors:  (1) smog chamber experi-
ments, (2) empirical models, and (3)  chemical kinetics models.   Section 6.1
discusses the results obtained using  each of these three approaches.  Section 6.2
then summarizes the implications of these results on the design of control  strategies
to reduce ambient levels of NOp.

6.1  Relations Among NOp and Precursors
     Although it is obvious why peak  and average N0? could be sensitive to NO
                                                   C-                         A
levels, it may not be instinctively clear why N02 should be affected by ambient
levels of NMOC.  Reducing ambient NMOC reduces the concentration of organic
radicals which serve to oxidize NO to NO .  The resulting lower concentrations of
organic pollutants therefore delays the formation of N02.  Such a delay allows
greater dilution of NO to occur prior to oxidation.  The end result is lower con-
centrations of NO  when the NO finally does get oxidized.
* In the discussion in Chapter 6.0, the term NMOC is used rather than VOC
  because the smog chamber, empirical  and kinetics model  evidence all deal  with
  ambient levels of pollutants rather than emissions.  The terms HC and NMHC
  (non-methane hydrocarbon) are sometimes used because they are used in the
  primary references.  In this discussion, NMOC, HC and NMHC may be assumed to be
  synonymous.
                                        53

-------
     6.1.1  Smog Chamber Experiments



          A smog chamber is a transparent container into which known concentrations



of organic pollutants (NMOC) and NO  are injected.   The resulting mixture is then
                                   /\


irradiated with ultraviolet light (using either sunlight or an artificial source).



The resulting concentrations of N(L and  other secondary pollutants can then be



observed.  By varying the initial concentration of  NO  or NMOC and keeping every-
                                                     A,


thing else constant, the sensitivity of  maximum hourly NO  concentrations to



changes in these precursors can be ascertained.





     The results of six sets of smog chamber experiements have been recently


                       20
summarized by Trijonis.    Figures 6.1 - 6.4 summarize the dependence of peak NO



on initial NO  concentrations in these experiments.   As shown in Figures 6.1 -
             A


6.4, despite the wide variety of experimental  conditions, the relationship between



peak N09 and NO  can be approximated by  a straight  line (i.e., is linear).  The
       C.      . X


slope of the line is generally slightly  less than 1:1.  Similar relationships


                                            20
were found between initial  NO  and mean  N09.
                             /\            C—




     Smog chamber data are  more ambiguous with regard to the dependence of NOp on



initial NMOC concentrations.  Most of the data indicate that, within the range of



observations encompassed by the experiment,  there is a weak relationship between



peak N02 and initial NMOC concentrations. For example, the experiments cited



indicate that reducing NMOC by 50 percent results in a reduction in peak NOp



ranging from 0-25 percent.   The experimental evidence reported is inconclusive



regarding the direction of  change in mean N02  corresponding with a reduction in



NMOC.  The results of the smog chamber studies are  summarized in Table 6.1.
                                        54

-------
a
x
re
                .10
 .30        .40       .50




Initial Nitrogen Oxides, ppm
       Figure 6.1  Nitrogen Dioxide Maximum Concentration vs. Initial Oxides of Nitrogen

                           (means of several experiments) UNC Study.
 a.     	
 u

 o
 o

  (N

 O
 E
 3



 'x
 a
                                       Initial NOX, ppm
              Figure 6.2_ Dependence of Nitrogen Dioxide Maximum Concentration

                      on Initial Nitrogen Oxides, Bureau of Mines Study.
                                         55

-------


s
a
a
o
TO
4-1
C
a)
u
c
0
u
3
0
I
1
£
'•x
2



u.tu
0.35



0.30

0.25
0.20


0.15
0.10

0.05

o
iiii L^1 '
	 Baseline Mixture 	



~ ^~-~a ~
^"^ \
0 S ° Varied
_ y " NO _
/ ^ Levels

x - •*
X° .•'l*~^*'"~ °
° ,f0'~^
~ o '*"' ~
0°
1 1 II 1 1
"0 0.1 0.2 0.3 0.4 0.5 0.6
                                Initial NOX Concentration, ppm
   Figure 6.3 Maximum One-hour Nitrogen Dioxide Concentration Produced from Irradiation

           of Multicomponent Hydrocarbon/NO.. Mixtures, General Motors Study.
a
a.


o
I

§
u
 CM
O


E
3
X
re
            • NMHC~2.1 ppmC

            O NMHC ~ 0.7 ppmC
                                     Initial NO  Concentration, ppm
            Figure 6.4 Maximum Nitrogen Dioxide Concentration as a Function of Initial NOX

                        for Two Initial Hydrocarbon Levels, UC Riverside Study.
                                             56
                                                          \

-------
         TABLE  6.1   Summary  of Conclusions  from Smog  Chamber Experiments
tn
                                       MAXIMAL
AVT3AGE NO,
CHAMBER STUDY
University of
finrth Carol ina

Bureau of Mines

i
i
General Motors



UC Riverside !



HEM, Auto
Exhaust i
HEW, Toluene
Dependence
on NOX
Proportional
or si ightly
less than
proportional
Proportions 1



Slightly less
than propor-
tional

Proportional






-------
Dependence
on ilC
50% HC control
reduces maximal
NOg by 10% to
20%
No effect



50% HC control
reduces maximal
N02 by 25%

50% HC control
reduces maximal
N02 by 10% to
15%
-------

-------
Dependence
on NOX
Proportional
or si ightly
less than
proportional
Proportional



Proportional
to slightly
less than pro-
portional
Proportional



Proportional

Proportional
Dc|<-.'.,uf.'ncc
on HC
Uncertain, 50% HC control
may decrease average NQ2
by ?0% or may increase
average NOp
50% 1IC control increases
average N02 ',7 10% to 30%


No effect



No effect



No effect

No consistent ei'.-_-v.t
          SOURCE:   Reference 20

-------
     6.1.2  Empirical Relationships from Ambient Data
          The second approach used to determine relationships between NO  and
precursors has been to examine ambient data for existing statistical  relation-
ships.  Empirical models based on these relationships have then been  constructed.
This approach has been tried at eight locations --  four in the Los Angeles basin,
                                                  20 21
two in Houston and one each in Denver and Chicago.   '    For each site, a step-
wise regression approach was used to derive the coefficients in Equations (6.1)
and (6.2).  For peak NO,, values occurring between 6:00 a.m.  and 4:00  pm:

               =  A •t"Bl[N02]5 + INTNO '  [B2 + B3 [NMHC] + B4 [RATIO]]  (6.1)

     where A, B-j , B2, B3, B^ = coefficients of regression
          [N02]5    = NOo concentration at 4-5:00 a.m.
          [INTNO]   = initial NO input (6-9 a.m.  NOX - [N02]5)
          [NMHC]    = 6-9 a.m. non-methane hydrocarbon concentration
          [RATIO]   = 6-9 a.m. NMHC/NOV ratio
                                      A
For peak N02 values occurring between 4:00 p.m. and 6:00 a.m.,
                 =  A + B1[N02]16 + B2ENITEN°3  + B3 [NITENO][03AFT]   (6.2)
     where
          [NOJ1C  = N09 at 3-4 p.m.
             L- I O      f-
          NITENO   = initial NO input (4-7 p.m.  NOX minus  [N02]1&)
          OoAFT    = ozone at 2-4 p.m.

Analogous equations were also developed for average daytime and nighttime N02.
Data used to derive Equations (6.1) and (6.2) were seasonal Ty disaggregated inta
summer (April -September) and winter (October-March).  As described  subsequently,
                                        58

-------
such disaggregation allows further insight into relationships between NOp and
reduction in NMOC.  Hence, Equations (6.1) and (6.2) were derived and applied
separately to summertime and wintertime data.

     The empirical models derived for six of the eight locations* imply that
reductions in NO  yield slightly less than proportional reductions in peak N09,
                x                                                            c.
providing all other factors are kept constant.  Reductions in average NO^ are
approximately proportional to reductions in NOV.  These results are summarized in
                                              /\
Table 6.2.

     As with smog chamber data, more ambiguous results were obtained when
assessing the impact of reducing NMOC on peak and average NO .   The results are
presented in Table 6.3.  According to the data in Table 6.3, there may be a
slight benefit as measured by reduced peak N0~ accompanying NMOC reduction.  This
effect is practically negligible for mean NOo.  It is interesting to note that
the effect of reducing NMOC appears to be somewhat more pronounced in winter than
in summer.  This is consistent with the hypothesis that since daytime N02 peaks
occur later during winter anyway, additional  efforts to delay N02 formation may
be more effective.

     An important caveat needs to be stated,  before the results shown in Tables 6.2
and 6.3 are taken too far.  Equations (6.1) and (6.2) do not explicitly take
account of meteorological differences.   Inclusion of meteorological  variables in
the relationships may change some of the conclusions, particularly with regard to
*  The two Houston sites were excluded because of poor fit of the models with
   nighttime data.
                                   59

-------
TABLE 6.2  Empirically Estimated Dependence of NCL on NO  Control
                                                 £.      /\
                                     Maximum N0?            Mean N0?
                                     50% NO  Control         50% NO  Control
Location                             No Chafige in NMOC      No Change in NMOC
Downtown LA                               -37%                   -45%
Lennox, CA                                -40%                   -47%
Azusa, CA                                 -45%                   -48%
Pomona, CA                                -31%                   -43%
Denver, CO                                -43%                   -55%
Chicago, IL                               -50%                   -50%

           Average                         41%                    48%
                                      60

-------
Table 6.3.  Estimated Impact of Reducing NMOC  on Peak N02 Obtained With Empirical Models
 Location
Summer Day
% Change in Peak
N02 Accompanying
50% Reduction in
 Summer Day
.•.%";.Change/. ia Ave „
 NC>2 Accompanying
 50% Reduction in
Winter Dav
% Change in Peak
N02 Accompanying
50%vReduction in
Winter Dav
% Change in Ave.
NC-2 Accompanying 50%
Reduction in
Downtown
Los Angeles -19%
Lennox, CA -16%
Azusa, CA 0%
Pomona, CA -17%
Denver, CO - 5%
Chicago, IL 0%
Houston/Mae Dr. 0%
Houston/Aldine 0%
-6%
-4%
+1%
-9%
-7%
0%
0%
0%
-25% -14%
-10% - 5%
-15% - 8%
-20% -19%
- 8% + 8%
-14% - 8%
0% 0%
0% 0%
           Average
        - 7.1%
                                                     -3.1%
                                 -11.5%
         - 5,

-------
NMOC.  Trijonis, for example, indicates  that the apparent  impact of NMOC reduction



on N02 may be overestimated by the failure of the models to  explicitly consider


            21
meteorology.    Therefore, the significance of the numbers in  Table 6.2 and 6.3



is that they agree qualitatively with  smog chamber data, rather than provide



exact numerical values.





     6.1.3  Relationships Derived from Chemical  Kinetics Models



          A chemical kinetics model  is a proposed sequence of  chemical reactions.



The identity of the reactions and the  rates with which they  proceed are based on



experimental (e.g., smog chamber), as  well as theoretical  evidence.  These



models enjoy several advantages over the previously described  smog chamber and



empirical approaches. These include greater flexibility to  test different cases,



use of realistic concentrations, and the ability to vary only  the parameter of



interest (i.e., NMOC or  NOJ.
                          A



                                                                        pp

     The kinetics model  underlying the ozone isopleth model  used in EKMA   has



been modified to express peak hourly NOp (rather than 0,)  as a function of



initial NMOC, NOX and other factors.   The shape  of the resulting N02 isopleths is



more sensitive to assumptions about  dilution rate and emission patterns than is



the shape of ozone isopleths.  It is apparent from Table 6.4 and from Figures 6.5-



6.7 that the effectiveness of NOV and  VOC controls in reducing peak N09 depends
                                x                                     C-


on the prevailing ambient NMOC/NO  ratio and the levels of N0? experienced as
                                 /\                          *•


well.  If 0.30 ppm is assumed to be  a  "typical"  peak N0? value, and the NMOC/NO
                                                       £                        A

                             23
ratio is assumed to be 9.5:1,   results  are similar to those observed in smog



chamber experiments and  empirically.   That is, NOV reductions  result in slightly
                                                 X


less than proportional reductions in peak N02 and NMOC reductions result in ambi-



guous changes in N02 which are usually slightly  beneficial.
                                   62

-------
       Table 6.4.  Effect of Separate NOX and NMOC Reductions on Peak N02 Under Various Conditions*
U>
Run
1
2
3
4
5
6
7
8
9
10
Dilution Rate
hr-1
3%
3%
20%
3%
10% .
3%
3%
3%
3%
10%
Duration
0800-1500
0800-1500
0800-1500
0800-1500
0800-1500
0800-1500
0800-1500
0800-1500
0800-1500
0800-1500
Subsequent
Emissions
1
No
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Duration
-
0600-1600
0600-1600
0600-1400
0600-1400
0600-1600
-
0800-1800
0800-1400
0800-1400
N02 w. 50%
Reduction in NO
X
-47%
-50%
-47%
-50%
-47%
-50%
-47%
-50%
-47%
-50%
N02 w. 50%
Reduction inWH
-10%
0%
-27%
0%
-13%
0%
-10%
0%
+ 3%
-13%
         *  Base  Case Assumptions:   NMOC/NOx Ratio  =  9.5:1
                                    N02  =0.30 ppm

-------
          0.84
          0.72
          0.60  —
          0.48
CTI
      D.

      O.
      o
      z
          0.36
          0.24  -
          0.12
                                        1.0
1.5
                                                 Fi(|iiro 6.5
2.0
3.0
3.5
               2.5



          NMOC, ppmC




with 8 Hours of Emissions and 3%/hour Dilution..
4.0
                                                                                                                                      4.5
5.0

-------
          0.84
          0.72
          0.60 -
          0.48 —
01
Q.
a
 x
O
z
          0.36
          0.24 —
          0.12 —
                     0.5            1.0           1.5           2.0
                                                                                2.5


                                                                            NMOC. ppmC



                                                Fiijure 6.6  NOo with 8 Hours of Emissions and 10%/liour Dilution
3.5          4.0          4.5          5.0

-------
          0.84
          0.72
          0.60
          0.48  -
en
      a.
      a.

       x
      O
      H
         0.24
                                                               ^~~	°'
                                                          /^^	.	0.
         0.36  -
         0.12  -
                                      1.0           1.5
2.0          2.5


       NMOC, ppmC
3.0          3.5          4.0          4.5
                                                                                                                                         5.0
                                            Figure G.7  NC^, with 10 Hours of Emissions and 3%/hour Dilution.

-------
     Examining Figure 6.7 somewhat more closely, one finds the curious result



that in some cases decreasing NMOC may increase the peak NO !   How can this be?



The answer can be found by examining in detail how points A and B on Figure 6.7



were derived.  This is done by plotting concentrations of NO,  N0o» and 03 as a



function of time on Figures 6.8 and 6.9.  Upon doing this, one finds that the



lower NMOC level reflected by Figure 6.8 has delayed the N0« peak to late in the



day, when dilution has been reduced.  The ozone peak is similarly delayed.  NO



emissions occurring late in the afternoon are titrated by the  late-forming ozone,



resulting in high N02 in the early evening.   Hence, the strange shape of the



isopleths in Figure 6.7 is the result of a different phenomenon (i.e., titration



rather than synthesis), causing the highest level of NO^.  These strange looking



isopleths only occur when there are substantial NO emissions late in the afternoon.





     A second factor that is apparent from Figures 6.5-6.7 is  that at very low



NMOC/NOY ratios (less than about 3-4:1) peak N0? is much more  sensitive to NMOC
       A                                       C-


reductions than to NO  reductions.  Thus, under such conditions the generaliza-
                     A


tions which have been drawn from smog chamber experiments and  empirical models



may no longer hold.  Because NMOC/NOV ratios appear to be much higher than this
                                    X


however, previous generalization about the efficacy of NO  controls versus VOC
                                                         /\


controls appear valid using the kinetics model as well.





6.2  Implications



     1.   Smog chamber data, empirical and chemical kinetics models all suggest



that reductions in NO  levels will lead to approximately proportional  reductions
                     A


in both peak and mean concentrations of N02-  An exception to  this may be the



case in which peak N0? occurs solely as the result of titration reactions with



ozone that are limited by the amount of ozone available.
                                        67

-------
cr>
oo
                 .205
                 .175  —
                                         120
  240                    360



Minutes After Beginning of Simulation
480
600
                                                  Figure 6.8  N02, NO and O3 vs. Time for Point A (Figure 6.7).

-------
    .228
a
a


o
o
o
    .195
    .163  -
.130  —
    .091  -
    .065 —
    .033  —
                             120
                                                240                   360




                                            Minutes After Beginning of Simulation




                              Figure 6.9  N02, NO and Og vs. Time for Point B (Figure 6.7).
                                                                                                480

-------
     2.   The proportionality constant between  peak (and mean)  N09 and NO
                                                                 £-       /x


appears to remain approximately the same for a  wide variety of  NMOC/NO  ratios,
                                                                      J\


emission patterns and dilution conditions.   This  implies that the linear rollback



model has the potential to serve as a  useful  means  for  approximating the impact



of NO  controls on ambient N00.
     x                       ^




     3.   The information presented suggests that programs  to reduce NMOC levels



will not have an appreciable impact on mean  N02-   There may, however, be a. small .



reduction in peak [^ accompanying reduction in NMOC under  currently prevailing



NMOC/NO  ratios.  Data examined imply  a reduction in peak N0~ from 0-20 percent
       X                                                    £-


accompanying a 50 percent reduction in NMOC.





     4.   Because the proportionality  constant  between  peak N02 and NMOC reduction



appears to be sensitive to a number of factors, the relationship is probably not



a linear one.





     5.   The preceding conclusions appear valid  under  currently prevailing



NMOC/NO  ratios in urban areas.  However, chemical  kinetics model simulations
       x


suggest that if the prevailing NMOC/NO  ratio is  reduced to very low values (as a
                                      /\


result of ozone SIP's), the relative effectiveness  of NO and VOC controls in
                                                        X


reducing N0? is different.  For example, for ratios of  less than about 2-3:1,



reduction in peak NOp would be more sensitive to  further VOC controls than to



further NO  controls.
          X
                                        70

-------
7.0  IMPACT OF CONTROLLING VARIOUS SOURCE CATEGORIES
     Chapter 6.0 has described relationships between peak and mean N02 concen-
trations and NO  and NMOC precursors.  Chapter 2.0 has identified major categories
               /\
of precursor sources (i.e., stationary, mobile, area, point).  Chapter 7.0 presents
information which is useful for identifying source categories impacting upon
receptor sites.  Thus, the concept of source-receptor relationships is introduced.
One means for surmising source-receptor relationships is through the interpretation
of ambient air quality data.  A second approach is through the use of mathematical
models.

     The first approach is a qualitative one which is useful  in justifying
assumptions which may have to be made in certain models.  The second (modeling)
approach is quantitative in that it can be used to estimate reduction in peak  or
mean N02 accompanying specified reductions in precursors.  Modeling approaches
are discussed in Chapter 8.0.

     Prior to reviewing ambient air quality data to better understand source-
receptor relationships, it is important to remember that there are likely to be
relatively few N02 monitoring sites in a given city.  Hence,  failure to detect an
impact of a source category on observed N02 concentrations cannot necessarily  be
used to infer that the category has a negligible impact on N02 concentrations
throughout the city.  Nevertheless, review of ambient air quality data can be
extremely useful.  Such a review can serve as a means for incorporating source-
receptor, assumptions in various modeling approaches described in Chapter 8.0.
For example, an apparent lack of point source impact at a monitor could justify
elimination of that source category from the emission inventory used in applying
a model without spatial disaggregation (e.g., proportional rollback) to estimate
                                        71

-------
the impact of precursor controls on peak NOp observed  at that site.  Such a
procedure ma'kes it possible for a relatively simple model  to explain why an
observed improvement in air quality has  not accompanied  emission reductions from
certain source categories.  By the same  token,  control  requirements on sources
which d£ impact the monitored air quality data  may be  less severe.

     Review of ambient data collected  in the plumes of point sources also serves
an important function.   The results can  be used to justify modeling approaches
for considering the major impact of these large sources  of NO  which may occur at
                                                             X
locations where there are no monitors.

     A third reason for reviewing ambient air quality  data in detail is that it
serves to emphasize the importance of  proper design of a monitoring network.
For example, further insight is gained  into whether a  monitoring site is a
representative one.  This, in turn, has  implications on  the use to  which the data
are put.  It would not be suitable, for  example,  to devise a citywide emission
control strategy on the basis of monitoring data  which clearly reflected the
impact of a single large nitric acid plant.

     Two procedures have been used to  draw inferences  about contributions of
various source categories to observed  peak NCL  concentrations.  These procedures
entail the use of (1) tracer data and  (2) peak-to-mean ratios.  Results and
inferences of each of these procedures  are described in  Sections 7.1 and 7.2,
respectively.  More detailed observations of point sources and the  resulting
implications are described in Section  7.3.   Finally, the implications of the
findings reported in this chapter to the design of control strategies
are summarized in Section 7.4.
                                           72

-------
7.1  Tracer Approaches
     A tracer approach is one which takes advantage of the correspondence between
emissions of NO  and other pollutants emitted by different source categories.
               /\
For example, suppose emissions from source category A contain pollutant "a,"
while emissions from category B do not contain "a" but contain pollutant "b"
instead.  Further, source category A's emissions do not contain pollutant "b."
In such a case, detection of "a" by a monitor, but not "b," implies that NOp
observed at the site originates from source category A rather than from category B.

     In the case of MOV emissions, carbon monoxide (CO) serves as a reliable
                      A
indicator of NOo originating from mobile sources.  In 1977, nearly all  (i.e.,
96 percent) of manmade NO  was emitted from two source categories:  mobile sources
                         /\
and stationary fuel combustion sources.   Ninety-nine (99) percent of the CO
emanating from these two source categories came from mobile sources.  Reference 1
also indicates that sulfur dioxide (SO ) should serve as a useful indicator of
NOp emanating from stationary combustion sources.  In 1977, 97 percent of S02
from mobile or stationary source combustion sources originated from the latter.
Hence, subject to some caveats which will be identified later, four results are
possible:

          1)  Close correspondence between NOX and CO but not with S02-
Conclusion:  mobile sources are responsible for observed NOp peaks.

          2)  Close correspondence between NOX and S02 but not with CO.
Conclusion:  stationary combustion sources are responsible for observed NO^
peaks.
                                        73

-------
          3)  Close correspondence between  NO   and  both CO and  SO0.
                                            A                   £
Conclusion:  both mobile and stationary combustion  sources contribute signifi-
cantly- to observed NO .

          4)  Lack of correspondence of NO  with  either CO or S09.
                                         A                     £
Conclusion:  a stationary source not characterized  by S02 emissions  (e.g.,
nitric acid plant, power plant burning  natural  gas) may be impacting the moni-
toring site.

     Use of CO and S02 tracers in the analysis  of ambient NOp data proceeds  by
plotting diurnal  NO, N02, CO and S02 concentration  patterns  observed concurrently
at a monitoring station.  This procedure has been extensively followed using
1975-1977 SAROAD data to examine coincident NO, N02, CO and  SOp peaks.  Figure 7.1
exemplifies what was, by far, the most  frequent finding — close correspondence
between diurnal CO and NO or NOp concentrations with little  apparent correspon-
dence with SOp concentration profiles.   This finding is not  too surprising when
one considers that mobile source emissions  of NO  and CO are ubiquitous and  occur
                                               x
near the ground.   Most of the stationary source emissions of S02 and NO  occur as
the result of combustion by electric utilities  or industrial sources.   Emissions
from such sources are generally confined to restricted areas or "points" defined
by the location of the emitting facility.   Frequently, these point source emissions
occur through elevated stacks.  Hence,  the  likelihood of a plume from such a
source impacting at the precise location(s) of  a  monitoring  station(s) for a
better part of an hour or more is slight.

     Despite the fact that the impact of point  sources of NOV emissions is
                                                            A
seldom detected by existing monitoring  networks,  monitoring  data exist which
demonstrate the potential for point sources to  contribute to high hourly
                                        74

-------
     23
0      8     16
Nov 76
                                    32     40
                                   TIME  (HOURS)
                 56
64
    72
25  NOV 76
Figure  7J  Pollutant plots for Lynwood,  November 23-25, 1976.
             Upper panel: 03 (	). CO (+•
             Lower panel: NO (	). N02 (
). and S02 (	); CO concentration scaled by 0.01
                                      75

-------
concentrations of NOp.   In Figures  7.2 and 7.3,  for example,  SO* tracer data
demonstrate that superposition of a point source plume  on  background NOp can
resu.lt in NOp concentrations Above  0.20 .ppm in winter.,  as  well  as in summer.

     Before leaving the subject of  SOp and CO tracers,  several  caveats regarding
the approach need to be identified.  First it should be recalled that the approach
identifies possible causes of high  NOp at the monitoring site only.   Source
categories not having any apparent  impact at existing monitoring sites are not
necessarily unimportant.   Second, there may be stationary  sources of NO  which do
                                                                       /\
not emit SOp (e.g., nitric acid manufacturing, natural  gas pumping stations or
natural gas-fired utilities).  These sources could  impact  on  a monitoring site .
without being identified as stationary sources.   Third, relatively small amounts
of S09 and CO are emitted by stationary area sources of NO such as  residential
     L~                                                    f\
oil burners.  Hence, S02 is not uniquely  identified with large stationary point
sources nor is CO uniquely identified with mobile sources. Finally, relative
amounts of NO  versus CO versus SOp emitted from sources within a given source
             /\                   £-
category vary.  For all these reasons, the approach outlined  in this section can
only be regarded as qualitative.

7.2  Use of Peak-to-Mean Ratios
     Calculating the ratio of peak  hourly NOp concentration to mean  NOp concen-
tration can sometimes be useful
          (a)  in determining whether point source  impact  is  likely, and
          (b)  in assessing whether a short term or annual air quality goal is
more restrictive.
                                        76

-------
   0.8
   0.7
   0.6
E   '
a
a
 I
z
o
<  0.4
cc
z
LU
o
o
°  0.3
   0.2
   0.1
                                                                                                .3
              , NO
                                                 12
                                             TIME - hours
16
               20
                             24
                   Fiqure  7.2  Example of potential point source impact at
                                    Long Beach on Sunday,  16 November 1975.
                                    CO has been scaled by 0.01.
                                               77

-------
   0.8-
   0.7
   0.6
   0.5
             ^       I
                                                                                                    .3
                                                                                                    .2'
O


<  0.4

DC
H
Z
UJ
u
   0.3
   0.2
   0.1
                                                    12


                                              TIME - hours
                                                                   16
                                                                                   20
                                                                                                   24
                  Figure  7.3   Example of potential point source impact at

                                     Whittier on Monday, 25 August  1975.

                                     CO has  been scaled by 0.01.
                                                78

-------
     One would ordinarily expect that a monitoring station which was impacted by



a large, elevated point source of NO  would exhibit higher peak-to-mean ratios
                                    /\


than a site which is impacted by ubiquitous area sources of NOV.  This assertion
                                                              A


is based on the presumption that it is extremely unlikely that a plume from an



elevated point source will impact at a given groundlevel station.  Hence,  the



mean concentration will be low, but the peak concentration could be high,  yielding



a high peak-to-mean ratio.  Conversely, there will always be some impact attri-



butable to ubiquitous groundlevel sources, so that one would expect a lower peak-



to-mean ratio when such sources dominate.





     Peak-to-mean ratios may also be useful in assessing whether it will be more



difficult to comply with an annual  or hourly air quality goal  for N02.  This



information may, in turn, be useful in the design of control strategies since the



relative effectiveness of different strategies could be different for annual



versus short term problems.  To illustrate, suppose the short term air quality



goal were 0.25 ppm NCL not expected to be exceeded more than once per year and



the annual goal were 0.05 ppm.  At sites having peak-to-mean ratios in excess of



5:1  (i.e., .25/.05=5), the short term air quality goal is likely to be of  greatest



concern.





     As suggested in Section 7.1, an extensive review of NO^ data has indicated



that monitoring stations experiencing N02 concentrations in excess of 0.20 ppm



are dominated by mobile (i.e., ubiquitous area) sources of NO .9  Under such
                                                             /\


conditions, Martinez and Mitz found peak-to-mean N0~ ratios varying between 3 and



12 with a median value of 6.3 and about 70 percent of the values between 5 and

    Q

7.6.   These findings are similar to ones found by Trijonis using earlier  data



from 120 urban sites throughout the United States.^'  Trijonis has also found a
                                        79

-------
distinct difference between urban peak/mean  ratios  and  ratios measured  in rural
areas subject to occasional impacts  from elevated point sources.   This  difference
is depicted in Figure 7.4.  Figure 7.4 indicates that if the impact  of  area
sources of NO  (e.g., mobile sources)  or ambient N0« is insignificant compared to
             /\                                     £
that from elevated point sources, the  peak-to-mean  ratio is likely to be  high.
Unfortunately, the converse of this  does not necessarily follow.   That  is,  one
cannot rule out the possibility of a point source significantly impacting at a
monitoring site just because the peak-to-mean ratio is  low.  Thus, although fewer
data are needed to apply the peak-to-mean ratio versus  the tracer  approach, the
latter approach is more useful.  The principal utility  of the peak-to-mean  ratio
approach is to identify cases in which point source impacts occur  and the impact
of area sources is negligible.

7.3  Impact of Point Sources on Ambient Levels of NOp
     Information presented in Sections 7.1 and 7.2  implies that existing  N02
monitoring sites tend to reflect the impact  of area (mobile) sources.   However,
for an analysis of potentially harmful  exposures to N02,  it is necessary  to
consider the impact on ambient N02 of  large  point sources as well.   Air quality
data collected in the vicinity of large point sources give insight into how such
sources should be considered in models to estimate  needed controls.  In this
section, information collected at a  continuously operated groundlevel monitoring
network in the vicinity of the Clinch  River  (TN) power  plant is described.   Next,
evidence of point source plumes observed during the St.  Louis RAPS study  is
described.  Third, air quality measurements  taken within  the plumes  of  several
large point sources using instrumented aircraft are discussed.  Finally,  implica-
tions about appropriate methods for  modeling such sources are identified.
                                       80

-------
20-
15-
10 _
 5 _
              3 Rural/Power Plant Stations
                          Urban Stations
          12345678

              Annual  Mean NO- Concentration, pphm


Figure  7.4.   Dependence  of maximum/mean ratio on  annual  mean
              N02 concentrations.
                              81

-------
     The Clinch River power plant is a 712 MW coal  fired utility located in



complex terrain.  NO, NO  and S0~ monitoring data were  collected for approxi-
                        X       £


mately 1-1/2 years at -six continuously operated monitoring stations shown in



Figure 7.5.  For reference in Figure 7.5,  the top of the plant's stack is located



approximately 2,175 feet above sea level.   Several  observations are apparent.





          1.   The most striking observation is the infrequency with which the



network detected a major impact attributable to the source.   Of approximately



120,000 station-hours of observations, hourly NO values in  excess of
                                                X


0.20 ppm were observed only 59 times or about 0.05  percent of the time.



The highest hourly reading was 0.83 ppm NO .   Only  three N05 concentrations
                                          X                £


in excess of 0.20 ppm were recorded (0.0025 percent of  the time).  These



three readings range from 0.35 - 0.74 ppm  and are suspicious in that the



corresponding NO  data are either missing  or agree  closely with the
                X


recorded N0? value.  It should be pointed  out, however, that observed



instances of N02 in excess of 0.20 ppm would likely have been more



frequent in an area subject to high concentrations  of ozone.  Although



the data are not available, it is unlikely that Clinch  River experiences



hourly ozone concentrations in excess of 0.10 - 0.15 ppm.





          2.   Approximately 50 percent of the NO  values in excess of 0.20 ppm
                                                 }\


occurred between 9 a.m. - 12 noon, giving  credence  to the hypothesis that the



maximum impact from such sources is likely to be during the period in which



morning fumigation occurs.  Twenty-five (25) percent of the time high NO  was
                                                                        X


obseryed between 6 p.m. and 9 a.m.  This is surprising, but it could reflect the



presence of katabatic winds.  Such a finding is not likely in flat terrain.
                                       82

-------
oo
CO
                                                                    Stack Heiyht ~ 200 m



                                           Figure 7.5  Location and Elevation of Clinch River Power Plant Stations.

-------
          3.   During hours in which observed NO  was >_ 0.20 ppm, 90 percent of
                                                X
the NO,/NOV ratios varied between .10 and .50, as shown in Figure 7.6.  There is
      ^   A
a slight negative correlation between the N09/N0  ratio and NO  on these days
                                            L.   X             X
with high NO .  It is possible that observed N09/N0  ratios during periods with
            X                                  c   X
high ground level NO  attributable to a point source may be somewhat higher in
                    X
locations subject to high ozone concentrations.   As in Figure 7.7
suggests, the median N0,,/N0  ratio observed in this sample is about 0.25.  The
                       w   A
three occasions in which hourly NCL concentrations in excess of 0.20 ppm were
observed are all characterized by N09/N0  ratios which lie far outside the range
                                    £   A
observed on every other occasion.

          4.   As expected, there was an excellent correspondence between hours
with high NO  and hours with high S09 observed by the Clinch River network.
            A                       £

     In addition to the Clinch River data, air quality data were measured aloft
on 134 days during the RAPS study in St.  Louis.   A subset of these data have been
                 24
reduced by Evans.    These data are of particular interest, because they were
observed in an urban area.  Evans1  data frequently reveal evidence of NO
                                                                        A
plumes aloft.   In a few cases, such as the one illustrated in Figure 7.8, very
high (> 1.0 ppm) NO  concentrations were observed for periods of a minute or two.
                   A
However, in such cases the plume consisted almost entirely of NO.  The amount of
NO^ present in such plumes is consistent with the depletion in the ambient ozone
levels found within the plume as a  result of the titration reaction.  It is also
of interest to note that despite the presence of these plumes aloft, only six
daily maximum hourly N0? readings were observed  above 0.20 ppm by the 25 station
RAPS network during 1976.  Five of  these readings occurred between 8:00 p.m. and
midnight LCT.   The sixth reading occurred at 9:00 a.m.  Five of the high values
                                          84

-------
        .60
           <0
    .94
IO.08
        .50
        .40
    Of.30
00
en
        .20
       .10
                            \
r- O
             °o
             O
                  O
                     O
D
O
D
               O
            O O      OO

           g>o  o
           CO
           o
        o
       o   o
                        o
                  o     o
                                \
                    O
                                 00
       oco

         o
                                o
                                    o
                                    o
                                 o
                                                 o
                                          o
                                     o
               t0.90
                                        o
                                   o
                                     o
                                        o
                                                                o
                                               o
                                     o
                                                           o
                                                          NO2 > 0.20 ppm
                                                                    o
         0.20
            0.30.
       0.40
0.50
0.60
0.70
0.80
0.90
                                                                     NOX, ppm
1.00
1.10
                                         Ficjure 7.6  NO9/IMOV vs. NOV at Clinch River on Days with High NO,
                                                      £-    J\       J\                                  t

-------
    30
    25
5   20
§   15
"o
w
a

i   10
                                (a)  All Readings > 0.20 ppm
                                                                  N =59

                                                                  Median = .26
            .10
                            .20
   .30



NO2/NOX
                                                         .40
                                                                        .50
    20
w>

5
'^
(0

I   15
O

"o
w
0)


1


z
    10
                             (b) For NOX Readings > 0.40 ppm
                                                                  N = 19

                                                                  Median = .22
             .10
                            .20
.30
                  .40
.50
Figure 7.7  IMO2 Ratios Observed at Clinch River During Hours with High Ambient NO
                                     86

-------
            DaY221,t976      Spiral 31
                                      05:08 to OS: 13
    3000
                            0.050     0.075      0.100     0.125      0.150    ppm
                  10
           12
  14
16
                                                           18
          20     degO
                 Ozone, NO, NOX, and Temperature vs. Altitude, Day 221
            Day 221,1976     Spiral 31
                                         05:08 to 05:13
    3000
    2000
'£   1000
                                                            I
         0




         I
0.25      0.50
 0.75       1.00      1.25      1.50  ppm




_|	I	I	I
                   10
           12
 14
16
18
20   deg 0
                Ozone, NO, NOX, and Temperature vs. Altitude, Day 221
                Figure 7.8  Evidence of NOy Aloft Over St. Louis.
                                       87

-------
occurred on October 1-2, 1976, days with  atmospheric stagnation  in  which several
sites observed above-normal  readings.  The  sixth  high reading  occurred at 10:00 p.m.,
off a- Friday nigtit in" November 1976.  Only one  site  recorded an abnormally high
reading on this occasion. This tends  to  support  the theory of a point source
impact.  However, the time at which the high reading occurred  (10 p.m.) makes it
seem more likely that the data reflect the  impact of a local ground level  source
of are spurious.
     The final set of data which sheds insight on the behavior of point source
plumes of NO  is obtained from aircraft measurements made within plumes.  Studies
            A
of this nature are generally conducted by flying  traverses  at  various  distances
downwind from the source. By noting cross  sectional variations  in  N02, NO,  0,
and S02, at differing distances (travel times), the validity of  various plume
modeling approaches can be checked.
     Airborne measurements have been made of several power  plant plumes.  Perhaps
the most extensively reported set of measurements has been  those made  within the
                                                              or pc
plume of the Labadie power plant as part  of the MISTT project.  '   The Labadie
source is a 2200 MW coal-fired power plant  located  about 50 km WSW  of  downtown
St. Louis.  Figures 7.9 and  7.10 show  cross sectional concentration patterns 21
and 45 km downwind from the  Labadie plant on a particular day.  In  Figures 7.9a
and 7.10a, S0? measurements  are used to trace  the cross sectional dimension  of
the power plant's plume.  Because the  "03 + N02 curve" in these  figures is constant,
it is apparent that the net  formation  of  N02 within the plume  is practically
identical to the net loss of ozone within the  plume.  This  strongly implies  that
N0? formation within the plume occurs  as  the result of the  titration reaction
(i.e., reaction 3.1).  White et al.   '   have  shown that a  simple model which
                                        88

-------
   15
I- /'\S MEASURED  03-I-N02
 Q.
 Q.
    10
cr
t-
2
LU
O
2
O
O
                                  10 km
              MEASURED
                       0
              MEASURED
        ,°0° CALCULATED
                                       !  oQ°ooOO°oooybooo
   Ambient Conditions


T = 20-25 C
RH = 60-70%
Winds: E @ 3-5 m/sec


This profile is measured

about 1 1/2 hours travel
time downwind from source.
              1240
                      1238           1236

                        TIME, CDT
1234
                                                                  76-090
 Figure 7.9a  Horizontal  plume traverse 21km downwind of Labadie stacks on
              August 14,  1974.  Oxidant (03 + N02)  concentrations within
              plume were  consistent with background levels on either side.
              This traverse was flown at 455 m (1500 ft) msl.
                                        89

-------
ol  10
O
o:
h-
-z.
UJ
o
z:
o
o
     0
                                    10  km
                   MEASURED  NO x
           o°o°  CALCULATED NOX
                ^•X,''  MEASURED  NO



                . '„ *  CALCULATED NO
                aoAopxx
               1240
1233           1236

  TIME ,  CDT
1234
                                                                       76-119
       Figure  7.9b  Horizontal plume traverse 21km downwind of Labadie stacks

                   on August 14,  1974.  Most of the NO emitted by the power

                   plant had already been oxidized to NC>2.   Measured NO  profile

                   agreed well with NO  profile scaled from measured S02 concen-

                   trations.  MeasuredTTO profile agreed well with No profile

                   calculated from measured S02 profile and oxidant-preserving

                   photochemistry.   (The same traverse is shown in Figure 7.9a).
                                      90

-------
    10
  CL
  Q.
 2  .
 Id
 O
 O
 O
      MEASURED  03
      MEASURED  Oa*
  \X' MEASURED  S02

O°O° CALCULATED   Oi
                                                  10km
                                     * /   *     «    /  »/
                                     V   Wvy  v
                                    Ambient Conditions

                                     - T = 20-25C
                                     - RH = 60-70%
                                     - Winds: E @ 3-5 m/sec.

                                    ,A ^ - This profile is
                                     N      3-4 hour travel
                                            time downwind from
                                            source.
               V
              1436
                 1434
1432          1430

 TIME,CDT
1428
1426
                                                                                           76-035
Figure 7.10a
    Horizontal plume traverse 45 km downwind of Labadie stacks on 14 August 1974.  Ozone
    concentrations within the plume at this distance were still depressed below background levels
    on either side,  although not as strongly as on traverses  nearer the stacks.  Measured ozone
    and  oxidant profiles  both show large-scale concentration gradient, possible due to urban
    plume from Greater St.  Louis.  The altitude  of this traverse was 610 m (2000 ft) msl.

-------
 E 10
 Q.
 Q.
 O
 cr
 h-
 z:
 ijj
 o
 o
                                                 10 km
          o  o
MEASURED  NOX



CALCULATED NO
                                MEASURED NO



                                CALCULATED NO
                                          f\ A
              436
          1434
1432          1430

TIME,CDT
1428
1426
76-118
Figure t.lOb Horizontal plume traverse 45 km downwind of Labadie stacks on 14 August 1974.  Nitric oxide

            concentrations were near the detection limit of the monitor, and very little excess within the

            plume was observed or predicted.  (The same traverse is shown in Figure 7.10a.)

-------
assumes that N02, NO and ozone concentrations within the Labadie plume can be



estimated using an empirical relationship between SO, and NO  and the photo-
                                                    L.       X
                              OC

stationary state* assumptions.    As shown in Figures 7.9 and 7.10, excellent



agreement between the model and observations is obtained.  On the basis of these



results, White concludes that, within 40 km of the source, N02 and 03 concentra-



tions within the plume are not affected by reactions with organic pollutants.




     Hegg et al. have summarized findings from monitoring studies within the

                                                            07
plumes of two coal-fired utilities and two gas-fired plants.    These observations



are essentially similar to those found in the Labadie plume.  Hegg et al.  conclude



that the conversion of NO to N02 within a power plant plume is limited by the



rate with which this NO comes into contact with ambient ozone.  In short,  conversion



of NO to N02 is limited by the dilution rate.  These results can be interpreted



to mean that the fraction of NO  which is NO- (i.e., NO./NO ) will be greater
                               X            L.          d.   X


near the periphery of a plume than at the center!ine.  Figure 7.11 illustrates



this concept.  Hence, peak N02 concentrations may not occur in the same position



in the plume as peak NO  concentrations (i.e., at the centerline).  Locations
                       J\.

within the plume experiencing high NO  have, necessarily, been exposed to  rela-
                                     /\


tively little dilution air.  Exposure to dilution air is necessary for conversion
* The Photostationary State assumes that an equilibrium between the reactions


               03 + NO -»•  02 + N02           (i.e., the titration reaction)

                       k2
        and             .
                        hv
               NO? + Op -»•  0.5 + NO           (reaction of resulting N09

                 *      ki                    with sunlight)          *


  occurs.  This equilibrium is given by
                                        93

-------
                              (b)
-  Figure  7.11
Diagrams for a titration reaction in a turbulent
plume:  (a) shows a cross-sectional view of the
plume.  The reaction zone is an annulus of Area A
and width L .   The interior of the plume, where the
concentration of NO is high, has area B.  Area C is
the exterior region of ozone.   E is a mixing turbu-
lent eddy of velocity Vp, and scale A; (b) shows
concentrations of the gases across section (a).
   Source:   Reference 27
                             94

-------
of NO to N09 to proceed.  However, dilution reduces NO  concentrations.  Thus,
           £                                          A


there are two counteracting phenomena (dilution and chemical conversion) which



should theoretically limit peak N02 concentrations within the plume of a major



point source.  A result of these competing phenomena is that N09/N0  ratios are
                                                               £   /\


likely to be higher at the plume's periphery than at the centerline.  This is



illustrated in Figure 7.11.  Maximum N02 concentrations are most likely at



intermediate ranges where sufficient dispersion has occurred to bring the plume



to the surface and to bring the reactants together.  Further downwind dilution



will  diminish concentrations.





     Before ending the discussion of NO  point source plumes, one further point
                                       A


needs to be addressed.   When concentrations of NO are very high (as they would be



within a few hundred meters of a source), Reaction (7.1) may result in some



initial  conversion of NO to N02 without substantial dilution.





          2 NO + 02 •> 2 N02                                 (7.1)





The rate with which Reaction (7.1) proceeds is too slow for it to be significant

                                       2

unless NO concentrations are very high.    As a result of Reaction (7.1) and direct



emissions of N09, approximately 5-10 percent of "initial" NO  from combustion
               £                                            A


sources will be usually be NOp.  N02/N0  ratios can, however, be as high 0.25 as

                                 2

the result of the reaction (7.1).   Hence, it is possible to have high N02 concen-



trations at the plume centerline without substantial dilution.  However, the extent



to which such an undiluted plume impacts the ground is probably very limited.





7.4  Implications



     1.    Use of S02 and CO data as tracers is a useful qualitative approach for



assessing whether observed high N02 concentrations are primarily attributable to
                                        95

-------
stationary or mobile sources.  Estimating peak-to-mean  ratios  of N02 serve as



a useful approach (a) for verifying whether a major point source is impacting a



site sporadically, and (b) for assessing whether a short  term  or annual  air



quality goal for NCL is going to be more difficult to meet at  a  given site.  Such



tools provide qualitative insights (a)  on whether to emphasize control  strategies



impacting most heavily on ubiquitous area sources or on large  point sources, and



(b) on whether to stress control strategies which emphasize reducing short term



peaks or long term averages.





     2.   Application of tracer data and peak to mean ratios to  available NO^



data suggest that observed NOp concentrations in excess of 0.20  ppm are  usually



attributable to mobile sources.  This finding could, in part,  reflect a  bias in



the way existing monitors are sited (i.e., frequently near roadways).  However,



the data examined indicate that, even when a monitoring network  is  deployed so as



to maximize the likelihood of detecting the impact of an  elevated point  source,



observed high hourly concentrations of  NO  (and  N0?) resulting from such a source
                                         3\        £


are very infrequent.  The foregoing implies that strategies which stress reduction



of NOV emissions from mobile and area sources will  be most effective in  reducing
     X


observed short term peak N02 concentrations as well as  annual  average values.



Nevertheless, as a safeguard, when possible a case-by-case review of tracer data



is suggested.





     3.   The preceding paragraph should not be  interpreted to mean that major



point sources of NO  necessarily play an insignificant  role in the  N0~/ozone
                   A                                                 *—


problem.  Paragraph 2 referred solely to observed data  collected at continuously



operated ground level monitors.  Such networks are usually quite sparse.  Air-



craft data reveal occasional very high  (e.g., >  1.0 ppm)  concentrations  of NO  as
                                                                             A
                                        96

-------
well as moderate concentrations of NCL within point source plumes.  These N02



concentrations, when superimposed upon the N02 resulting from other sources could



lead to hourly concentrations in excess of concentrations as high as 0.20 or



0.25 ppm.  Because (a) NO  emitted within point source plumes (without substantial
                         J\


amounts of VOC) may be less photochemically active than NO  from well-mixed
                                                          /\


sources, and (b) NO  from elevated sources may be more isolated from the ground
                   A


and other scavengers, NO  from such sources may be relatively more important in
                        /\


the transport of low levels of NOV to rural areas.  As such, they could be poten-
                                 X


tially more important in the formation of ozone in rural areas, formation of fine



particulates over great distances, and as a contributing factor in the acid



deposition phenomenon.






     4.   Data collected within point source plumes suggest that N02 within such



plumes primarily results from titration of NO with ambient ozone.  Some NO  may
                                                                          A


be emitted as NO  as well.  Because of the high concentrations of NO  which may
                L.                                                   X


occur within plumes, the ozone titration reaction is likely to. be limited by the



concentration of ambient ozone.  This suggests that the most effective way of



reducing peak NOp from a point source may be (a)  to limit the amount of NO



emitted as NOp or converted to NO  via the oxidation reaction described in



Reaction (7.1). and/or (b) reduce ambient levels  of ozone as the result of SIP's



for ozone.  However, reduction of NO  emissions in any form may be useful  in
                                    /\


reducing ozone formation in downwwind areas.  Procedures for reducing NO  from
                                                                        /\


large stationary sources are described in the appropriate Control Techniques


         28
Document.
                                        97

-------
     5.   To the extent that St.  Louis  is  a  typical  city,  data  examined  there



suggest that plumes of NO  are frequently  present aloft  in the  early  morning
                         A


hours.  Data from Clinch River tend to  confirm that  the  midmorning  fumigation



phenomenon is the event most likely to  lead  to observed  major impacts from



elevated point sources.  Hence, midmorning may be the most likely time to  observe



a major impact from point sources.   Even with  the fairly dense  monitoring  network



in the RAPS study and the ideally sited network at Clinch  River however, impacts



on ambient NO^ in excess of 0.20  ppm resulting from  the  fumigation  phenomenon



were seldom, if ever, observed.
                                        98

-------
8.0  USE OF MODELS FOR N02
     Previous chapters have discussed sources of NOV, means by which it gets
                                                   /\
converted to N02 in the atmosphere, issues relating to measuring representative
values of N02, prevailing levels of N02 observed by existing monitoring networks,
long range transport of N02, and relationships between N02, its precursors and
principal source categories.  This information is needed in order to make informed
judgments about collecting and interpreting appropriate data for developing
control strategies to reduce unacceptably high levels of ambient NOo.  Questions
that still remain however are:
          (1) "how much control is needed to meet an air quality goal for N02?"
          and
          (2) "what is the impact of an individual source?"
The answers to such questions require quantitative estimates.  The role of models
is to provide means for making such estimates.

     The purpose of Chapter 8.0 is to describe models available for estimating
the impact of precursor emissions on short term and long term N02 concentrations.
The state-of-the-art for N02 modeling is not so far advanced as for other criteria
pollutants.  Although some of the approaches to be described are easy to apply.
such applications cannot be made blindly.  Model  assumptions or estimates which
appear to contradict existing air quality, meteorological  or emissions data need
to be scrutinized and explained if they are to remain credible.  The quantitative
predictions provided by models should be regarded as "best estimates" with some
band of uncertainty about them.  Unfortunately, in most models for N02, this band
of uncertainty is not precisely known.  Such uncertainty is reduced, however,
when use of models is accompanied by high quality input data, and is consistent
                                        99

-------
with qualitative observations described  in  previous  chapters.  Under  these  conditions,
model estimates provide useful quantitative goals  for control  programs.

     The potential for high short term concentration of N0_  is greatest  in  the
following settings:  (1) urban-industrial regions  in which NO  and VOC are  emitted
                                                            A
from a large number of mobile and stationary sources; (2) downwind of sources
with large NO  emission rates in urban regions  or  in non-urban regions which
             /\
periodically experience high ozone levels (e.g., greater than  the MAAQS  for 03);
and (3) downwind of certain industrial sources  that  emit large quantities of N02
directly into the atmosphere.  Accordingly, Sections 8.1 and 8.2 present models
available for estimating the impact of controls in multisource urban  areas  and
from large individual sources of NOV or  N09, respectively.
                                   X     c
8.1  Multisource Urban Models
     Models appropriate for use in multisource  urban areas can be categorized as
simple or complex.  Simple models include linear rollback, OZIPM and  the Photo-
chemical Box Model (PBM).  However, the  only one of  these approaches  widely
available as of September 1980 is rollback.  Consequently, a more lengthy description
of rollback modeling, as applied in assessing the  impact controls on  ambient NO^
is provided herein.  Complex models include several  atmospheric simulation  models.
Such models have been primarily used to  estimate ambient ozone concentrations,
but can also produce estimates of ambient NOp.

     8.1.1  Linear Rollback
          Linear rollback is based .on the assumption that changes in  emissions of
a pollutant will result in proportional  changes in the ambient concentration of
that pollutant.  It was concluded in Chapter 5.0 that background concentrations
(be they natural or manmade) transported into the  city under review may  ordinarily
                                      1QO

-------
be ignored.  This simplifies the rollback model still further.  The rollback
approach proceeds by first estimating the future design value for NCL which would
occur if no additional controls over and above those already planned were implemented.

                   -  i  \  r  future  ->                         to -i\
               Xf  =  (xo)  [ Q	  ]                         (8.1)
                              ^present
     where
          Xf       =  future NOp concentrations, any units
          Xo       =  present design value for NOp, consistent units
          ^future  =  Pr°Jected N0x emissions, any units
          Present =  Present NO  emissions, consistent units
The amount of additional reduction required in NO  emissions to meet either the
                                                 A
annual or a prospective short term NOp is given by Equation (8.2)
                   Xf - X1
               R = —	  x 100                                (8.2)
                      xf
     where
          R  = percent reduction required emissions over and above that already
               projected
          X1 = the appropriate NAAQS for NO^s consistent units.

     Considerable efforts have been made to refine the "Q" terms on the right
hand side of Equation (8.1) in recent years.  These emission terms have been
subdivided into different emission categories (e.g., light-duty motor vehicles,
stationary combustion sources, etc.), each of which can reflect different assumptions
                                           29 30
about growth and future control technology.  '    As a result of these refinements,
the model has come to be known as the "modified rollback" model.  This model has
been widely used in conducting regulatory analyses for prospective N0? air quality
standards and for estimating the impact of various automotive emission standards
                                       101

-------
on future NO  emissions.   The  modified  rollback model computer package  and
            X


accompanying User's Manual will  be available  by 1981.





     Several underlying assumptions  in  the  linear rollback model need to  be



identified.





          1)  Peak and mean N09  concentrations are directly proportional  to  NO
                              £-                                               A


emissions;





          2)  Peak and mean N02  concentrations are independent of ambient NMOC



(or VOC emissions);





          3)  Either all  source  categories  are evenly distributed throughout the



city or the atmosphere is perfectly  mixed so  that the relative contribution  a



source makes to ambient concentrations  is the same as its relative contribution



to total NO  emissions.
           X




     As discussed in Chapter 6.0, the first assumption is generally valid for



both peak and mean concentrations of N02.   The exception occurs when peak N02



concentrations result primarily  from titration of NO with ozone and the amount of



N02 formed is limited by the prevailing 03  concentration.  Even in such a case,



however, it is probably appropriate  to  assume that N02 is proportional  to NOX



emissions, if the N02 peak is  observed  to build upon carryover N02.  Thus, prior



to applying the rollback model,  it is advisable to apply the procedures described



in Sections 3.2 and 3.3 to assure oneself that the observed peak N02 concentration



is not limited by available ozone or that if  it is, some carryover N02  is present.
                                      102

-------
     According to the information presented in Chapter 6.0, average N02 concen-

trations are insensitive to changes in NMOC.  Hence, the second assumption causes

no problem in considering annual mean N02 values.  However, peak NCL concentrations

are slightly sensitive to reductions in NMOC.   This sensitivity depends on the

prevailing ambient NMOC/NO  ratio, diurnal emissions patterns and atmospheric
                          /\
dilution.  Depending on these factors, reductions in peak N02 ranging from 0-

20 percent have been observed to accompany 50  percent reductions of NMOC in smog

chambers or in chemical kinetics models.  It is not possible to consider relation-

ships between peak N0? and NO  and NMOC simultaneously using the rollback approach.
                     £•       /\

However, NOp isopleth diagrams similar to those presented in Figures 6.5-6.7 have

been used for several dilution rates and emission patterns to assess sensitivity

of peak N09 to NMOC at various NMOC/NO  ratios.  Table 8.1 presents the results
          £                           A
obtained for low dilution rates most likely to accompany high urban levels of NO
                                                                                A

and NOp.  Table 8.1 should not be used for dilution rates in excess of 10%/hour.


     The following example illustrates the use of the rollback procedure with the

information in Table 8.1


     Example 1

       Given:
          -Hourly N02 concentration = 0.30 ppm
          -Prevailing NMOC/NO  ratio on days with high hourly N02 = 9:1
          -In the absence of further controls, a projected increase in NO
           emissions of 20% over a defined period (e.g., 1987)
          -An existing ozone SIP requiring 40% reduction in VOC over the same
           defined period (i.e., 1987).

       Find:
       The required reduction in NO  emissions to meet a hypothetical  hourly
       NAAQS for N02 of 0.25 ppm not expected  to be exceeded more than once per
       year.
                                      103

-------
Table 8.1.  Sensitivity of Peak N02 to  Changes  in  NMOC  with Various
            Prevailing NMOC/NO  Ratios
                              J\
                                         % Reduction  in  Peak N02
                                         Accompanying a  50%
               Ratio*                    Reduction in NMOC

               >10:1                              0%
               8-10:1                           10%
               6-8:1                             15%
               4-6:1                             20%
       *  If the prevailing NMOC/NO  ratio  is  not known  on days experiencing
          N02 in excess of the air Duality  control,  assume 0% reduction in
          peak N02 resulting from reductions  in  NMOC.
                                       104

-------
     Solution:

       (1)  Estimate  future NOp design value which would occur  in the absence of
           further controls

                Xf = (0.30)0.20)                                 (8.1)


                Xf = 0.36 ppm*

       (2)  Reduce the future design value consistently with the prevailing NMOC/NO
           ratio.  According to Table 8.1, for a VOC/NO  ratio of 9:1, a 50%
           reduction in NMOC results in a 10% reduction in peak N02,  Since the
           anticipated reduction in VOC emissions is 40% rather than 50%, reduce
           the future design value as shown:
                                              i+O
           Reduction due to VOC Reduction =  (—) (10) = 8%
                                              50

           New N02 Design Value = (0.36) (1.00 - .08)

                                = 0.33 ppm


       (3)  Estimate  the required reduction in projected NO  emissions (Qfuturp)
           needed to attain the NAAQS at the end of the period of interest
           (e.g., 1987).


               R =[°'330"33'25J  x 100                            (8.2)

               R = 24%


     The procedure illustrated in Example 1 is not appropriate in the unusual

case when NOp occurs as the result of ozone-limited titration in the absence of

carryover NOp.  In such a case, it would be necessary to estimate reductions in
                                                               pp 23 31
ambient ozone using models which have been described elsewhere.  ''    In the

absence of better information, the resulting reduction in peak NOp would be

proportional  to the estimated reduction in peak ozone, after making some allowance

for NO  which is emitted directly as NOp.   The procedure is illustrated in Example 2.
*  Note that the calculations should not be carried out to more than two
   significant figures.
                                        105

-------
     Example 2
      Given:
      (1) An N02 Design Value = 0.30  ppm  is observed which  is  primarily attributable
          to ozone-limiting titration occurring in the absence of carryover N02.


      (2) A VOC emission control  program  is estimated to reduce peak  ozone by 30%.


     Find:
      The resulting reduction in  peak NO,,.


     Solution:
      (1) Assume 10% of the N02 results from direct emission of N02

               (.10)(.30 ppm) = .03 ppm of N02 would not be affected  by reducing
               ozone


      (2) This leaves 0.27  ppm N02

               Estimated Future N02 = (0.27)(1-.30) + .03

                                   = 0.22 ppm


     Information presented  in Chapter 7.0 casts serious doubt  concerning the

validity of the rollback model's  assumption that the relative  contribution a

source category makes to ambient  concentrations is the same as its relative

contribution to total NO emissions.   As  discussed in Chapter  7.0, monitoring
                        A

data seldom appear to reflect impacts from major point sources.   This could

result in credibility problems if the rollback model is used as  the basis for

applying controls on a large source which does not impact the  monitoring site.

Failure to observe subsequent improvement in ambient N0? after implementation of

controls could cause criticism of a strategy to reduce emissions  from such sources.

Criticism of this nature may stem solely  from improper use of  the rollback model

rather than from the effectiveness of the strategy per se.  A  similar problem may

occur if the rollback model  is used to assess the impact of a  transportation

control measure which is applied  at one or a few sites which either do not impact
                                      106

-------
monitoring sites or have a disproportionately high impact on a monitor site



(e.g., as may be the case if a roadway near a monitor were turned into a pedes-



trian mall).  Thus, the rollback model is most suitable for estimating the impact



of controls which are applied uniformly to ubiquitous sources.  Examples of such



strategies include the Federal Motor Vehicle Control  Program (FMVCP) and inspection



and maintenance programs.





     The impact of widespread control programs (e.g., inspection and maintenance)



on observed peak and mean concentrations of NCL may be underestimated using the



rollback model unless emissions from large, elevated  point sources are somehow



discounted.  Use of available S02 data (see Chapter 7.0)  as a tracer for N02



originating from large, elevated stationary combustion sources can be used as



justification for discounting such emissions.  If (1) NO   emissions from large
                                                       A


point sources with elevated stacks is accompanied by  SO^  emissions, and (2)



ambient data indicate no correspondence between peak  N02  and observed ambient  SOp



patterns, NO  emissions from large elevated point sources can be eliminated from
            A


the "Q" terms on the right hand side of Equation (8.1).   Because of the unlike-



lihood of observing an impact from an elevated source at  any given point, in



applying the rollback model, NO  emissions can be eliminated from the "Q" terms
                               A


on the right hand side of Equation (8.1) in estimating controls needed to meet



the annual N02 standard regardless of whether there are any S02 data.





     It should be emphasized that the suggestions in  the  previous paragraph are



intended to insure that the rollback model provides estimates which result in



more credible prescriptions for ubiquitously applied  control measures.  The



suggestions do not imply that elevated point sources  of NO  are unimportant.  If
               ~                                           A


elevated point source emissions are eliminated from the rollback model as suggested,
                                       107

-------
their potential impact on peak NCL values needs to be  considered in some other



way.  However, it is first necessary to; describe available  models for estimating.



the impact from point sources.  Therefore,  the discussion of appropriate procedures



for supplementing the rollback model  to consider point sources  will be deferred



until Section 8.2.





     Because the impact of elevated point sources are  not ordinarly expected to



impact appreciably on annual  average NO^ concentrations at  any  giveir site, it is



sufficient to use Equations (8.1)  and (8.2)  to estimated needed NOV emission
                                                                  J\


controls excluding emissions  from  elevated  point sources.   It is not ordinarily



necessary to supplement this  analysis with  applications of  a point source model .



in this case.





     8.1.2  OZIPM



          The OZIPM (Modified Ozone Isopleth Plotting  Package)  is similar


                                                                     22 23 31
conceptually to the OZIPP/EKMA model  described extensively  elsewhere.   '  '    A



well-mixed column of air is assumed to migrate within  an urban  area so that it



encompasses the monitoring site observing the peak N02 concentration at the time



of the observation.  This is  pictured conceptually in  Figure 8.1.  The OZIPM



model provides the user with  a graph of peak hourly N0? (rather than 03) as a



function of initial NMOC and  NO concentrations.  Figure 6.5 (page 62) is an
                               A


example of such a diagram. The positions of the NO isopleths  are also functions



of the NO  and VOC emissions  encountered by  the column along its trajectory,
         /\


diurnal variation in the mixing height, sunlight intensity  and  ozone transported



from upwind.  As with the EKMA procedure, one uses 6-9 a.m.  NMOC/NO  ratios and
                                                                   J\


the peak observed N02 concentration to identify a starting  point on the isopleth



diagram.  Control requirements needed to attain a specified air quality goal are



calculated from the starting  point.




                                       108

-------
    INITIAL
  CONDITIONS
 WIND
              1
 / YYY
                    Mi
                                    Mi
                            Mi
7/7"
                                               DILUTION RATE, x%/hr.
                                                       ETC.
                                KEY

                    ti  =  TIME PERIOD i

                    Mi =  MIXING DEPTH AFTER TIME i
*:
                                                         Qi =  PRECURSORS INJECTED INTO
                                                              COLUMN DURING TIME i
~LJ- = SUNLIGHT INTENSITY DURING
       TIMEi
Figure  8.1  Conceptual view of the column  model.
                                 109

-------
          The OZIPM model has several  potential advantages over  rollback.  First,
the isopleth diagram can be applied to cases  in which NOp results  from either
synthesis or titration.  Second,  the impact of VOC emission  controls  on peak N02
can be more readily determined.   The problem  with uniform distribution of point
sources still  remains however.  Other  problems are that the  OZIPM  model is only
appropriate for short term NCL  concentrations and its applicability to winter NO^
episodes is uncertain.  Finally,  and most importantly, the model needs to be
further evaluated.   Thus, it is not expected  to be widely available for use sin*
control strategy evaluations until late 1981.

     8.1.3  Photochemical Box Model (PBM)
          The PBM model has been  recently developed and is currently  undergoing
further evaluation.  In this model, a  large stationary cell  is superimposed over
                                      32
the urban area as shown in Figure 8.2.    The change in estimated  poTTutant
concentrations with time is dependent  upon the flux of" emissions within the box,
the inflow and outflow of pollutants,  the dilution which occurs  as the mixed
layer rises, entrainment of species from aloft and formation and destruction of
species through chemical reaction.  The model assumes that all emissions are
instantaneously mixed throughout  the box and  ignores horizontal  dispersion.

          The model requires overall urban emissions of CO,  VOC  (divided into
five classes), and  NO  for each hour of the modeling period  (i.e., 0500 to
                     J\
1700).  No spatial  resolution of  emissions is required.  The output gives a
single city-wide prediction for each pollutant for each hour in  the modeling
period.  The box model has been exercised using several days during the summer of
1976.  Air quality  data from twelve monitoring sites were used in  comparing
predicted and observed concentrations.  The comparison showed generally good
                                       110

-------
                   RISIMC MIXED
                    HEIGHT
Figure 8.2  Schematic diagram of modeling domain for
            the Photochemical Box Model.
                        Ill

-------
agreement between predicted and observed concentrations  of CO,  NO,  N02 and 03
averaged among numerous monitoring sites within  a  city.32

          The PBM has many of the same advantages  and  problems  as the OZIPM
model.  An additional advantage is that it is  not  dependent on  the  proper location
of NO/NOX monitors.  One problem with the Box  Model  which  has yet to be solved is
that it predicts the spatially averaged N02 concentration  throughout the box.  A
way must be found to relate this spatial average to  peak N02 concentrations for   -
the model to be directly usable to assess the  impact of  controls to reduce high
short term concentrations of N02>   It is not certain whether or when this latter
difficulty can be overcome.

     8.1.4  Complex Multi source Models
          This section identifies  models which are distinctive  from those described
in Sections 8.1.1 - 8.1.3 in that  they reflect greater attempts to  incorporate
physical/chemical phenomena believed  to occur  in urban areas.   However,  any
resulting increase in credibility  occurs at the  expense  of increased data and
computer requirements for running  the model.   Unless a suitable data base has
already been compiled for ozone modeling, it is  not  likely that such models can
be exercised to assess the effectiveness of control  strategies  to reduce N02
without a lengthy and expensive effort to assemble an adequate  data base.
Table 8.2 identifies the complex multisource photochemical  models discussed in
greater detail in Reference 33.

8.2  Models to Assess the Impact from Individual Sources
     The models identified in Section 8.1  have been  primarily used  to estimate
the impact of numerous rather than individual  sources.   Simple models like roll-
back which have no spatial  resolution are not well suited  for estimating the
                                       112

-------
Table 8.2,  Available Complex Multisource Photochemical Dispersion Models
          Model
          ELSTAR
          IMPACT
          LIRAQ-2
          MADCAP
          TRACE
          URBAN AIRSHED
Type
Trajectory
Grid
Grid
Grid
Trajectory
Grid
Additional
References
    34
    35
    36, 37
    38, 39
    40, 41
    42
                                      113

-------
Impact of an individual source.   Monitoring  data  collected  in  the vicinity of
large sources of NO  suggest that,  at any  given time, the maximum impact of an
                   /v
individual source may be quite localized.  For example,  Figure 7.9a  (page 87)
shows that the plume width from the Labadie  power plant  is  only about 4 km across
                                        pc
as far as 21 km downwind from the plant.     As shown in  Figure 8.3,  Cole et al.
found that the identifiable impact  on NOp  of a major freeway occurred within a
                                 43
few hundred meters of the roadway.     Even the most complex multisource grid
models typically have no greater horizontal  spatial resolution than  2-4 km.  This
limitation is necessary in order to assure computational feasibility.   Conse-
quently, models which focus on individual  sources may be useful  for  new source
review or, if it is desirable, to assess the impact of controlling a major source
in the vicinity of that source.

     Cole and Summerhays have described  several different approaches which have
                                                                44
been used to estimate NOp concentrations in  point source plumes.     In addition,
Cole et al. have applied a relatively simple approach to estimate NOp concentra-
                                                     43
tions downwind from a major freeway with some success.    Only the more simple
approaches will be described herein.   Information concerning a refined reactive
plume model is available in Reference 45.

     The principal distinguishing feature  between simple and complex models for
individual sources of NO  is that simple approaches do not  account for the inter-
                        /\
action of the NO  plume with organic pollutants.   However,  this  limitation is not
                /\
a serious one insofar as predicting peak N02 is concerned.  A  number of field
investigations have indicated that  peak  N02  within an identifiable plume of an
individual source occurs far too rapidly for reactions with organic  pollutants to
                                        pr nc•py  *o
cause a significant part of the problem.   '   '  '   Further,  several  of the
                                     114

-------
   N02  (ppm)

    1.14
    L12
    D.N
    D.06
    0.04
O   ).02
LOW OZONE LEVEL  .

MEDIUM OZONE LEVEL

HIGH OZONE LEVEL
D

O

A
                                                                                •
                                                                               4
                                                                               9
                                                                                                 BACKGROUND NO,
                                                                                                 AVERAGES
                                                                                                 ARBITRARILY PLOTTEl
                                                                                                 At 500 METERS

'
r ).oo


1

0 40


1

80


1

120


1

160


1

200


1

240


1

280


1 1

320 360


1

400


I

440
1
1
1 1 1

480 52
                                    DISTANCE DOWNWIND FROM FREEWAY, METERS
        Figure 8.3  Nitrogen  dioxide (N02) data from the N02/03  sampler siting study.

-------
simple approaches appear to agree quite well  with observed data.   Hence, they
should be adequate for use in evaluating N02  control  strategies.
     Four approaches are identified in Table  8.3.  These approaches are described
                                26 44 45
in detail in several references.  '  '    All of the  approaches require the user
to first estimate maximum hourly NO  concentrations using widely available models
                                   /\
for inert pollutants.  *  '*    In the case of major roadways,  a "line source"
                   47 48
model such as HIWAY  '   would be used for these purposes.  Appropriate point
source models are available on the EPA's UNAMAP system.   '    Of the models
described in Table 8.3, the Ozone-Limiting and the Photostationary State Approaches
are preferable to the others, because they consider the limitations on peak N02
posed by the oxidizing potential of dilution  air.  The Ozone-Limiting Approach is
more practical than the Photostationary State Approach because of the general
lack of rate constant data available to persons likely to be formulating control
strategies.  Preliminary testing of the Ozone-Limiting Approach and several other
N02 models for individual sources has been performed  using aircraft measurements
of power plant plumes in the St. Louis region.  The Ozone-Limiting Method was
found to be reasonably accurate and performed slightly better than a complex
reactive plume model, as well as other simple models.     Consequently, a recom-
mended procedure for applying the Ozone-Limiting Approach is presented in greater
detail.

     The ozone-limiting approach involves two initial  comparisons:  (1) comparison
of the estimated maximum ground level NO  concentration [NO ]    and the maximum
                                        X                  X nlflX
ambient 0, concentration prevailing at the time of day when [NOV]__V is estimated,
         O                                                     X il'clX
and (2) comparison estimated [NOV]    during  midafternoon with prevailing ozone
                                x max
levels at that time.  If 0, concentration is  greater  than [NO L-,v, all of the
                          •3                                  x max
                                        116

-------
 TABLE  8.3   Approaches  for  Estimating Maximum Impact on Peak  N02  from  Large  Individual Sources
  1)  Partial Conversion
     Exponential Decay
                 Hane

 3)  Ozone Limiting Approach
4)  Photostationary State  Approach
                   Description

 1)  Multiply predicted maximum NOX concentration
    by  a  constant  factor to predict peak N02--
    Multiply predicted NOX concentrations  by an
    exponential decay factor

    N02 = Hx


    where [NO ] ax - peak  ground level N0x con-
    centrations estimated  with a model for inert
    pollutants
    k,   « the ratio constant in reaction with
    which N02 photodisassociates in the preset
    of  sunlight:       k
                   NO, £- NO + 0
                                                                                         ice
                                                k,  = rate constant depicting the rate with
                                                which 03 confines with NO to form N02:
                                                               NO
                                                                   03 -  N02
                                                                               02
                                                                       Remarks

                                                        It is difficult to determine what the factor should
                                                        be.  Ambient measurements of NO?/rJOx ratios within
                                                        pluses vary widely.  (See fig.  7.6)   Generally,  the
                                                        greater the dilution,  the hicher the ratio.  As  a
                                                        rule, routinslv collectec .T.on "Storing date  a"E not
                                                        useful for these nurposes,  because ihey  do rot  refie
                                                        the impact of point sources.  Monitoring cata col-
                                                        lected in the vicinity  'vZOOm of  a  major  roed/:ay  may
                                                        be of use for roadway applications.
                                                        Conversion of fJO to N02 appears to be li^:-ri by the
                                                        oxidizing potential of the dilution air.  Also, the
                                                        half  life of the NO is determined by dilution rather
                                                        than  chemistry.  Any value chosen for the half life
                                                        would have to be arbitrary.
                                                                                                                  Remarks
                                                    )  As discussed in Chapter  7.0, the summation of
                                                       (NO + N02  + 03) appears  to be constant across
                                                       point source plurces  (see Figures 7.9 and 7.10).
                                                       This implies that  NO  to NP.2 conversion in plumes
                                                       with high  NO  is limited by prevailing ambient
                                                       ozone.   Fewer arbitrary assumptions are required
                                                       using this method  than methods (1) or (2).   Method
                                                       assumes that complete mixing of the NO  plumes oc-
                                                    	currs at the point where maximum ground level NO
                                                       occurs.  /It also ignores the likelihood of  further
                                                       photodisassociation of NO; upon- formation.   Measure
                                                       rrents or estimates of prevailing ar.bient Oi levels
                                                       at the time (NO )    occurs are  needed.
                                                   4)  White26 and Cole et al.1*3 have found this ap-
                                                      proach  to work well in predicting maximum N0?  down-
                                                      wind  from a major point source  and a major  line
                                                      source.  The approach assumes thorough  mixing  of
                                                      the plumes with anbient air at  the point where
                                                      (N0x)ma  occurs.  Measurements  of the k,.rate  con-
                                                      stant are not typically available.  The approach also
                                                      requires measurement of prevailing 03 concentrations
                                                      at the  time (NO )max occurs.
                                                                  117

-------
NOV is assumed to be NQ9.  If [NO ]_,„ is greater than the prevailing 0, con-
  X                    <-         X fflciX                                 *j


centration, N02 formation is "ozone-limited."  In this latter case, the maximum



N(>2 concentration is assumed to be equal  to the Og concentration plus a correction



factor which accounts for NO  emitted as  N09 and near-stack thermal conversion.
                            X              £-


For combustion sources, the correction factor is assumed to be 10 percent of the



estimated [NO ]   .   The Ozone-Limiting Procedure is outlined in four steps
             X HldX


below.






          1.   A standard dispersion model  for inert pollutants is used to



calculate [N0x]max>






          2.   [N0v]mav is separated into two components:
                  X  niaX





               a)  portion emitted as N02 or thermally converted, equal to



t> ' [N0v]m,v, where  "b" is assumed to be: .10 for combustion sources;
       X maX.





               b)  the remaining NOV fs assumed to be NO subject to conversion by
                                   X


0.,.  This NO is 0-b)[NOvl  v or .90[NOv]mav for combustion sources.
 j                      x max          x  max





          3.   a)   If [0,] > (1-b)  [N0jm  .. then assume that all of the NO is
                         J             X  UlaX
converted to N02 (i.e., [NO^^ = [N02]max).





               b)   If (1-b)[NOx]max > [03],  then
     This procedure should be used to address two situations:   (i) when NO  is
                                                                          /\


maximum, and (ii) when 0, is maximum.  Note that the above procedure can also be



readily applied to assess the impact from other industrial point sources (e.g.,



nitric acid plants, natural gas pumping compressor stations) which emit a larger
                                       118

-------
fraction of NO  as N09.   One would simply increase  the  "b" value, described  in
              A      £

step 2a), to whatever fractional  value is appropriate for these sources.


     The procedure is illustrated with two examples.


     Example 3a
      Given:  A point source dispersion model  for inert pollutants estimates
              [NO ]    = 0.30 ppm to occur in  midmorning as the result of
              fumigation of a power plant's plume.  Existing ozone monitors
              suggest that the highest 03 concentration for this time of day
              is 0.10 ppm.


      Find:  Maximum impact of the point source  on  hourly ambient NO^ levels.


      Solution:
          (1)  Since a power plant is a combustion  source, assume b = 0.10


          (2)  Since (0.90)[NOw]maw = 0.27ppm  >  0.10 ppm = [0,],
                              x max                         3

                          is ozone-limited


                          C°33 +  b '  [NOAax

                        = 0.10 +  (.10)(.30)

                        =0.13 ppm


     Example 3b
      Given:  For the power plant in Example 3a,  [NO ]    for midafternoon is
              estimated to be 0.15 ppm.   Peak  ozone at  tins time is measured to
              be 0.20 ppm.


      Find:  Maximum impact of the point source  on  hourly ambient NO^ levels.


      Solution:
          HT~ b = 0.10

          (2)  (1-b) [N0x]max = (.90)(.15) = 0.14 ppm < 0.20 = [03]

          Therefore, [NO^] max is NO  limiting

                                   - 0.15 ppm
                                       119

-------
     One additional facet of model fug, procedures  for N02 needs to be discussed.
This is the integration of multisource and individual  source techniques described
in Sections 8.1 and1 &..2y respectively.  In the complex multisource models enumerated
in Section 8.1.4, this integration  has already occurred.  That is, these models
already include both area and point sources in their emission inventories.
Because complex models can more accurately simulate source-receptor relationships,
the impact of elevated sources on ground!evel  concentrations have a greater
chance of being correctly estimated.   The only occasion when it may be desirable
to separate a point source from an  inventory used in a complex multisource  model
would be if one were interested in  assessing the  microscale impact of a parti-
cular individual source.  In such a case,  the procedure would be to:

          (1)  eliminate that particular source from the inventory;

          (2)  use the multisource  model  to estimate NO, NOp and ozone concen-
trations attributable to all other  sources;

          (3)  apply an individual  source model (e.g., the Ozone-Limiting Approach)
to simulate the impact of the source of interest, using the multisource model's
estimates as the means for characterizing the oxidizing potential of the dilution
air.

     In order for the rollback model  to be an acceptable approach for estimating
the impact of control strategies on peak short term N02 concentrations, it  is
necessary to account for source-receptor relationships in some rudimentary  way.
As discussed in Section 8.1.1, it will usually be appropriate to exclude elevated
point sources from the inventory in applying the  rollback model  to short term
situations.  If this is done, however, use of the rollback model in individual
                                       120

-------
areas needs to be supplemented by separate consideration of large individual
sources not impacting a monitoring site.  The recommended procedure is as follows.

          (1)  Review available SCL and N02 data on occasions with high NCL using
techniques described in Chapter 7.0.

               a.  If there is a clear correspondence between N02 and S02 peaks,
include emissions from elevated point sources in the emission terms on the right-
hand-side of Equation (8.1).  Estimate required controls using Equation (8.2).

               b.  Otherwise, project future N02 concentrations using Equation  (8.1)
excluding elevated point source emissions from the emission terms.  Proceed to
step (2).

          (2)  Apply the Ozone-Limiting Approach to estimate peak impact of
selected large individual sources on ground level NCU.

          (3)  Repeat (2) for several scenarios as needed (e.g., fumigation case,
afternoon impact during periods with maximum 03> etc.).

          (4)  Repeat (l)(b) using N02 design values which are appropriate for
the times of day assumed for each scenario in (3) and occur nearest to the location
where the peak industrial source impact occurs.  If there are no monitoring data
near this location, use a spatially averaged value for the appropriate time of
day.
          (5)  For each scenario assumed in steps (3) and (4), superimpose the
estimated peak point source impact on the projected future level of N02 obtained
with the rollback model.
                                       121

-------
The foregoing procedure is illustrated numerically in Table  8.4.





8.3  Implications



     T.   Two major categories of models  have  been discussed:   multisource



models and individual source models.   In  order to  provide  comprehensive quanti-



tative estimates of controls needed to attain  a short term air quality goal  for



NOp, both kinds of models need to be utilized.   In designing control  strategies



to meet an annual NOp air quality goal, it  is  generally  sufficient to utilize



multisource models alone.  If rollback is the  multisource  model  utilized,  emissions



from large individual point sources may be  ignored in applying the model  to



estimate the impact of controls in a achieving annual average, air quality goals.





     2.   Although several multisource models  exist, the one that is  most widely



available and easiest to apply is the linear rollback model.   Rollback is  suitable



for use in both annual and short term applications.  Peak  and  mean NCL concen-



trations appear to be proportional to NO  under most circumstances.   Consideration
                                       A


of the impact of reducing NMOC can be made  by  supplementing  the rollback model as



illustrated in Example 8.2 (Section 8.1.1).





     3.   Several relatively simple approaches for estimating  the maximum impact



from large individual sources of NO  (e.g., a  public utility or a major highway)
                                   /\


have been described in Section 8.2.  Of these,  the Ozone-Limiting Approach appears



to represent the best available compromise  between accuracy  and data  requirements.



This approach is also readily adaptable for modeling sources which emit large



proportions of NOX directly as N02-
                                      122

-------
             TABLE  8.4   Hypothetical Examples  Illustrating Use  of  Multi and Individual  Source Models
                           for Projecting Future  Short Term N02 Concentrations for Use in  SIP's
                                 Scenario

              1) Observed nighttime-early morning peaks.   Im-
                 pact from elevated point source assumed  to be
                 negligible due to nocturnal  inversion.
Time of Day

 1900-0800
Estimated Peak Individual
   Source Contribution
  Projected Multi -
Source Contribution

     0.30 ppm
                                              Projected  Peak N02

                                                    0.30 ppm
ro
CO
              2)  Observed mid-morning peaks.  Multisource syn-
                 thesis of N02 superimposed by fumigation im-
                 pact from an elevated point source.
 0900-1100
        0.13 ppm
                                             0.20 ppm
                          0.33 ppm
              3)  Observed mid-afternoon  peaks.  Potential
                 point source impact is  relatively high due
                 to high observed 0^'levels at this time.
                 Projected multisource impact is low due to
                 low N02 concentrations  observed by existing
                 monitors during this time.
 1200-1800
        0.15  ppm
     0.06 ppm
                                                                 0.21 ppm

-------
     4.   Where available, SOp data should be used to verify the impact of large



stationary point sources on high observed hourly concentrations of NO- using the



procedure outlined in Chapter 7.0.   In cases where a review of available evidence



indicates a lac.k of point source impact,  NO  emissions from major elevated-point
                                           /\


sources should be eliminated from the emission inventory when using rollback.



Peak impacts on NO- estimated from large  individual  sources with the Ozone-



Limiting Approach should be superimposed  on projected NO- concentrations thus



obtained with the rollback model.  This procedure is illustrated more completely '



by the flow chart in Figure 8.4.





     5.   The procedure described in paragraph 4 can be repeated for several



scenarios using consistent estimates of impacts from multiple sources and large



individual sources.   An example of several  scenarios is illustrated in Table 8.4.



Such repetition should better insure protection of the public against excessively



high short term exposure to NO-.





     6.   To be implemented, the procedure outlined, in paragraphs 4 and 5 requires



several pieces of data and judgments to be made which may not at first be obvious.





          a.   A certain amount of expertise regarding local  meteorological  conditions



is needed in order to arrive at a realistic set of scenarios.





          b.   Measurements of peak hourly NO- concentrations  and ozone at times



of day which are consistent with selected scenarios  need to be made at monitoring



sites which are likely to represent NO- levels encountered in widespread areas of



the city.





          c.   It is desirable, but not essential,  to have estimates of the



prevailing NMOC/NO  ratio in the city being modeled.
                  ,j\




                                           124

-------
         (5)
Use line source model
& Ozone limiting approach
to estimate the impact
of roadway.   Add pro-
jected rollback cone.
using N02  design value
from upwind site to
estimate future N02.
Exclude major point
sources from  rollback
inventory.
     (4)
Are we estimatin
impact of indiv
roadway?
       No
    (6f
     Do
high N02 and
    coincide
      ?
                                                               (3)              \
                                                           Use  rollback model    \
                                                           without emissions    !
                                                           From elevated point  /
                                                           sources              /'
                              (7)
                          Use  rollback model
                          including emissions
                          from elevated point
                          sources
                                     (8)
                                 Project future N02
                                 using rollback with-
                                 out elevated point
                                 source emissions  in
                                 base or future
                                 oeriods_    	
                                     1
                                               (9)
                                         Likely impact on
                                         future observed N02
                                     (io)               !
                                 Use ozone limiting
                                 approach to estimate
                                 impact of large  point
                                 source of N02  for a
                                 meteorological scenario'
                                 similar to that  in (8)  :
                                     (11)
                                 Add impacts  estimated
                                 in (8)  and (10)
                                          meteorological
                                 scenarios of  interes
                                Jseen considere
                                               (12)
                                          Estimated impact on
                                          peak future N02 for
                                          specified meteorological
                                          scenario
                                     (14)
                                 Consider  next scenario
                                 using appropriate design
                                 valuesin  rollback and
                                 met.  assumptions in the
                                 ozone limiting approach
  Figure  8.4.    Procedure for Applying Rollback  1n Assessing the  Impact  of
                   Control  Strategies  on  N02
                                                   125

-------
          d.  A judgment needs to be made about  future  prevailing  levels  of
ozone.  Since it is anticipated that areas violating  any  prospective  short term
air quality goal for NO^ are also likely  to violate the hourly  MAAQs  for  ozone,
estimates of future peak afternoon ozone  values  should  be available.   Estimates
of future nighttime or morning ozone concentrations should be consistent  with
those made about future transported ozone in ozone SIP's.

     7.   The modeling approaches described in this chapter imply  that-it may be
possible to estimate a prospective violation of  a short term NO^ NAAQS even
though one has not been observed and may  not be  likely  to be observed by  existing
monitoring networks.  For example, if an  hourly  NCL NAAQS of 0.25  ppm were estab-
lished, hypothetical Scenario 2 in Table  8.4 would be a case in point.
                                      126

-------
9.0  REFERENCES

 1.  National Air Quality, Monitoring, and Emissions Trends Report 1977,
     EPA-450/2-78-052, pages 5-11, Office of Air Quality Planning and  Standards,
     U.S. Environmental Protection Agency, Research Triangle Park, North  Carolina,
     December 1978.

 2.  Air Quality Criteria for Oxides of Nitrogen (draft), Chapter 6, revised
     June 1980.

 3.  J. R. Martinez, Survey of the Role of MO  in Nonurban Ozone Formation,
     EPA-450/4-79-035, September 1979.       x

 4.  National Emissions Data System (NEDS), National Air Data Branch,  U.S.
     Environmental Protection Agency, Research Triangle Park, North Carolina.

 5.  Research Triangle Institute (1975), Investigation of Rural  Oxidant Levels
     as Related to Urban Hydrocarbon Control Strategies, EPA-450/3-75-036,
     March 1975.

 6.  C. E. Decker, et al., Formation and Transport of Oxidants Along the
     Gulf Coast and in Northern U.S.. EPA-450/3-76-033, August 1976.

 7.  W. C. Eaton, C. E. Decker, J. B. Tommerdahl, F. E. Dimmock, Study of
     Nature of Ozone, Oxides of Nitrogen and Nonmethane Hydrocarbons in Tulsa,
     Oklahoma, Research Triangle Institute, RTI/1466/000, F.  April  1979.

 8.  SAROAD, 1975-1977, SAROAD Data Summaries, maintained by National  Air
     Data Branch, U.S. Environmental Protection Agency, Research Triangle Park,
     North Carolina.

 9.  J. R. Martinez, K. C. Nitz, Analysis of High NO? in California. 1975-
     1977. EPA-450/4-79-034a, b, c, August 1979.

10.  SAROAD 1975-1977, SAROAD Hourly Concentrations, maintained  by National
     Air Data Branch, U.S. Environmental Protection Agency, Research Triangle
     Park, North Carolina.

11.  D. H. Sennett, personal communication concerning a Brief Analysis of
     LACS Data Base, LACS Study, summer 1978,  Los Angeles, California.

12.  Personal communication of A.  J. Hoffman to J.  D. Sableski,  subject "N02
     Siting Guidance," July 14, 1978.

13.  U.S. Environmental Protection Agency, "Ambient Air Quality  Monitoring,
     Data Reporting and Surveillance Provisions," Federal Register
     (40CRF 51,52,53,58), 44 No. 92 (May 10, 1979), pp. 27558-27604.

14.  T. C. Curran, Screening Procedure for Ambient Air Quality Data,
     EPA-450/2-78-037, U.S. Environmental Protection Agency,  Research  Triangle
     Park, North Carolina, July 1978.
                                       127

-------
15.  H. G. Richter,  E.  L.  Meyer,  and D. H. Sennett,  "A Graphical Procedure
     for Screening and  Analyzing  High N02 Concentrations," Paper 79-12.4,
     presented at 72nd  APCA Meeting, Cincinnati, Ohio, June 1979.

16.  C. W. Spicer, The  Fate of Nitrogen Oxides in the Atmosphere,
     EPA-600/3-76-030,  March 1976.

17.  C. W. Spicer, D. W.  Joseph,  P. R. Sticksel, G.  M. Sverdrup., and G. "F.  Wood,
     Reactions and Transport of Nitrogen Oxides and  Ozone in the Atmosphere.
     draft, EPA Contract  68-02-2439, U.S. Environmental Protection Agency,
     Research Triangle  Park, North Carolina, June 1979.

18.  C. W. Spicer, J. R.  Koetz, G. W. Keigley, G. M. Sverdrup, and .G.  F. Wood,
     A Study of Nitrogen  Oxide Reactions Within Urban Plumes Transported Over
     the Ocean, draft,  EPA Contract No. 68-02-2957,  U.S. Environmental Protection
     Agency, Research Triangle Park, North Carolina, November 1979.

19.  J. C. Trijonis, T. K. Peng,  G. J. McRae, and L. Lees, Emissions and Air
     Quality Trends  in  the South  Coast Air Basin, EQL Memo No.. 16, January  1976,
     California Institute of Technology, Pasadena, California.

20.  J. C. Trijonis, "Dependence  of Ambient M02 on Precursor Control," pap.er
     presented at EPA Technical Symposium on Implication of a .Low NO   Vehicle
     Emission Standard, May 2-4,  1979, Reston, Virginia.

21.  J. C. Trijonis, Empirical  Relationships Between Atmospheric Nitrogen
     Dioxide and Its Precursors,  EPA-600/3-78-018, February 1.978.

22.  Uses, Limitations, .and Technical Basis of Procedures for Quantifying
     Relationships Between Photochemical Oxidants and Precursors,
     EPA-450/2-77-021a, U.S. Environmental Protection Agency, Research Triangle
     Park, North Carolina, November 1977.

23.  Procedures for Quantifying Relationships Between Photochemical Oxidants
     and Precursors: Supporting  Documentation. EPA-450/.2-77-021b. U.S.
     Environmental Protection Agency, Research Triangle Park, North Carolina,
     February 1978.

24.  R. B. Evans, The Contribution of Ozone Aloft to Surface Ozone Maxima,
     PhD Dissertation,  University of North Carolina, Chapel Hill, North Carolina,
     1979.

25.  W. H. White, J. A. Anderson, W. R. Knuth, D. L. Blumenthal, J. C. Hsiung,
     and R. B. Husar, Midwest Interstate Sulfur Transformation and Transport
     Project:  Aerial Measurements of Urban and Power Plant Plumes. Summer  1974,
     EPA-600/3-76-110,  November 1976.

26.  W. H. White, NO -  0$ Photochemistry in Power Plant Plumes:  Comparison
     of Theory with Observation,  Environmental Science and Technology  II (10)
     October 1977, page 995.
                                       128

-------
27.  D. Hegg, P. V. Hobbs, L. F. Radke, and H.  Harrison, "Reactions  of Ozone
     and Nitrogen Oxides in Power Plant Plumes," Atmospheric Environment 11, (6),
     page 521.

28.  Acurex Corporation, Control Techniques for Nitrogen Oxides Emissions from
     Stationary Sources - 2nd Edition, EPA-450/1-78-001, January 1978.

29.  N. deMevers and J. R. Morris, "Rollback Modeling:   Basic and Modified,"
     JAPCA 25, page 943, September 1975.

30.  J. H. Wilson, Jr., "Methodologies for Projecting the Relative Air Quality
     Impacts of Emission Control Strategies," Paper 78-4.1,  presented  at 71st
     Annual Air Pollution Control Association Meeting,  Houston, Texas,  June  1978.

31.  G. Z. Whitten and H. Hogo,  User's Manual for Kinetics Model  and Ozone
     Isopleth Plotting Package.  EPA-600/8-78-014a, July 1978.

32.  K. L. Schere and K. L. Demerjian, "A Photochemical  Box  Model for  Urban
     Air Quality Simulation," Proceedings 4th Joint Conference on Sensing of
     Environmental Pollutants, American Chemical Society, 1978.

33.  R. Y. Wada, M. J. Wong, and E.  Y. Leong, Applicability  of Selected Models
     for Addressing Ozone Control Strategy Issues. EPA-450/4-79-026, December 1979.

34.  F. Lurmann, D. Godden, A. C. Lloyd, and R. A. Nordsieck, A Lagrangian
     Photochemical Air Quality Simulation Model. EPA-600/8-79-015a,  b,  June  1979.

35.  A. Fabrick, R. C. Sklarew and J.  G. Wilson, "Point  Source Model Evaluation
     and Development Study, Appendix C, User Guide to Impact," prepared for
     California Air Resources Board and the California  Energy Resources
     Conservation and Development Commission by Science  Application, Incorporated,
     Westlake Village, California, SAI Contract No.  A5-058087, March 1977.

36.  M. C. MacCracken, et al., "The Livermore Regional Air Quality Model:  I.
     Concept and Development," Journal of Applied Meteorology 1_7, March 1978,
     pages 254-272.

37.  W. H. Duewer, et al., "The  Livermore Regional Air  Quality Model:   II.
     Verification and Sample Application in the San Francisco Bay Area,"
     J. Applied Meteorology 17,  March  1978, pages 273-311.

38.  R. C. Sklarew, M. A. Joncich, and K. T. Tran, "An Operational Air Quality
     Modeling System for Photochemical Smog in  the San Diego Air  Basin," pre-
     sented at the Annual American Meteorological Society Meeting, Reno, Nevada,
     January 1979.

39.  R. C. Sklarew, "Verification of the MADCAP Model of Photochemical  Air
     Pollution in the San Diego  Air Basin," presented at the Annual  American
     Meteorological Society Meeting, Reno, Nevada, January 1979.
                                       129 .

-------
'4*6.  P. J. Orivas, '"TRACE (Trajectory Atmospheric  Chemistry and  Emissions)  User's
     Guide,*' Pacific Efiivrowmantal  Services,  Incorporated, Santa Monica,  California,
     PES Report TPF016, November 1977.

41.  P. J. Drivas and L. G.  Wayne,  "Validation  of  an  Improved  Photochemical
     Air tjaallty Sfmilatitwi  "Model," Pacific  Environmental Services,  Incorporated,
     Santa -Monica, California,  PES  Report TP-014,  March  1977.

42.  S. D. Reynolds, L. E. Reid and T.  W.  Tesche,  "An Introduction to  the
     SAI Airshed Model and Its  Usage,"  Systems  Applications, Incorporated,
     San Rafael, California, SAI Report EF 78-53R, December 1978.

43.  H. S. Cole, N. Mayer, and  H.  Fowler,  "The  Impact of an NO  Diesel  Waiver
     on N02 Concentrations Downwind of  a Major  Line Source," U.S. Environmental
     Protection Agency, Office  of Air Quality Planning and Standards,  internal
     report, May 1979.

44.  H. S. Cole and J. E. Summerhays, "A Review of Techniques  Available for
     Estimating Short Term N02  Concentrations," JAPCA 29 (8),  August 1979,
     pages 812-817.

45.  D. Stewart and M. A. Yocke, User's Guide to the  Reactive  Plume  Model  -
     RPM-II, SAI Report EI-79-93R,  EPA  Contract No. 68-02-2775 (1980)  in
     press.

46.  C. Hosier, "Meteorological  Program of the  Environmental Protection Agency,
     Bull American Meteorological Society  56  (12), December 1975, pages 1261-
     1270.

47.  D. B. Turner, "Atmospheric Dispersion Modeling:   A  Critical Review," JAPCA
     29_ (5), May 1979, pages 502-519.

48.  J. R. Zimmerman and R.  S.  Thompson, User's Guide for HIWAY, A Highway
     Air Pollution Model. EPA-650/4-74-008,  February  1975.

49.  Guidelines for Air Quality Maintenance  Planning  and Analysis, Volume 9
     (revised), Evaluating Indirect Sources,  EPA-450/4-78-001, September  1978.

50.  M. A. Yocke, et. al, Evalution of  Short  Term  N02 Plume Models for
     Point Sources. SAI Report  103-EF-80-91,  EPA Contract No.  68-02-2775,
     (1980), in press.
                                        130

-------
10.  ACKNOWLEDGEMENT
     The principal  authors of this document are Edwin L.  Meyer,  Jr.,
Donald H.  Sennett,  Henry S.  Cole and Harold G.  Richter.   Valuable contributions
have been made by David Iverach.  The authors are also indebted  to
Carole J.  Mask for her very able typing and clerical  support.
                                  131

-------
         (5)
Use line source model
& Ozone limiting approach
to estimate the Impact
of roadway.  Add pro-
jected rollback cone.
using N02 design value
from upwind site to
estimate future H02.
Exclude major point
sources from  rollback
inventory.
     (4)
     e estlmatlri
impact of Indlv
roadway?
      6)
      Do
 high N02 and SOS
     coincide
                                                           Use rollback niodel
                                                           without emissions
                                                           from elevated point
                                                           sources             j
    (7)
Use rollback model
Including emissions
from elevated point
sources
(8)
Project future N02
using rollback with-
out elevated point
source emissions In
base or future
periods
>
s

(9)
Likely Impact on
future observed NO?


                                      (10)
                                  Use  ozone limiting
                                  approach to estimate   )
                                  impact of large point  '
                                  source of N02 for a
                                  meteorological scenario1
                                  similar to that in (8)  I
                                     (ii)
                                 Add Impacts estimated
                                 In (8) and (10)
                                                            r^
                      (12)
                Estimated Impact on
                peak future N02  for
                specified meteorological   /
                scenario                ^J
                                                                                                          A
                                 Have all meteorological
                                 scenarios of 1 nteres
                                  een considered^
                                 Consider next scenario
                                 using appropriate design
                                 values in rollback and
                                 met. assumptions in the
                                 ozone limiting approach
                                          _\___	
                                                                                                           I
    Figure 8.4  Modeling procedures for N02 SIP'S

-------
TABLE 8.3   Approaches  for Estimating Maxima iapact on Peak Ho2 from  Large  Individual Sources
                 Kan,


 1)  Partial Conversion
     Exponential  Decay
                 Name

 3)  Ozone Uniting Approach
4)  Photostationary State Approach
. 1)  Multiply predicted naxlmum NO  concent'ratio
i     by  a  constant  factor to predict peak NO,--
                                            '2)
     Multiply predicted NO, concentrations by an
     exponential decay factor
                                                N02 •
                                               • where t - travel time from source, tine center
                                                T • half 11ft conversion time of HO to N02.
                                                consistent units
                                                b • fraction of NO. emitted as NO,, as smart  to
                                                be 0.10 for combustion sources, could be
                                                higher for Industrial process sources.
                                                (NO,)     • laxlnum estimated ground level
                                                concentrations, of nOK using an Inert pol*
                                                lutant model.
.3)  Method assumes that maximal N02  concentration
    can be Halted by dilution of an NO, plume
    and/or the ozone concentrations  In  the di-
    lution air.  If oione Is the limiting factor.
                                                •here [0,]_ • art lent  ozone concentration* It •
                                                the tine peak NO. occurs.
                                                b • frarttnn of NO  emitted as NO,
                                                [NO J     • Maximui ground level X0f estimated
                                                wltR an inert point source Model .
4)  Approach confines (NO )HI estimated with an
    :1nert pollutant model  nun the  photo*'»f ionary
    state relationship.   As a result, [NO} L,  Is
    estimated as follows:
                                                •here [NO^]^, •  peak  grcund level NOj con.
                                                ctntrations estimated  «1th a •odel for inert
                                                pollutants
                                                k)  ' the ratio constant  In reaction with
                                                •Men N02 photodisassoclates 1n the presence
                                                of sunlight:      n
                                                                   S_
                                                               NO,
                                                                       NO * 0
                                                k,  * rate constant depicting" the rate wit*'
                                                •filch 0, coifclnes «1tti NO to fora NO,:
                                                               «0 » 0, -  NOj
                                                                               02
                                                                       Remarks-

                                                     1)  It 1s  difficult to determine what the factor should
                                                         be.  Artfent measurements of NOi/NO, ratios within
                                                         plumes vary widely.  (See fig. 7.6)   Generally,  the
                                                         greater the dilution, the higher the ratio,  '-l  a
                                                         mle.  routinely collected mon'taring date  irt not
                                                         useful  for these purposes, because they do not refle
                                                         tke impact of point sources.   Monitoring data col-
                                                         lected in the vicinity vJOOm of a najpr roac<.-ay  may
                                                         •« of  use for roadway applications.
                                                        Conversion of NO  to N02 appears to be lirr.fte- by the
                                                        Oxidizing potential of the dilution air.  ~,-,i.  the
                                                        half life of the  no It dettrcined by dilution rather
                                                        than checistry.   Any value chosen for the half life
                                                        Mould have to be  arbitrary.
                                                                      Remarks


                                                   3)  As discussed in Chapter 7.0. the sumption of
                                                       (NO » NO; » Oj) appears to be constant across
                                                       point source plunes  (see Figures 7.9 and 7.10).
                                                       This Implies that  NO to NO? conversion In plums
                                                       with high NO  is limited by prevailing ant lent
                                                       ozone.   Fewer arbitrary assumptions are required
                                                       «1ng this method  than methods (1) or U).  Method
                                                       assumes that complete mixing of the NO  piumes oc-
                                                     ._c«rrs at the point where maximum ground level NO
                                                       •ccurs. /It also Ignores the likelihood of further
                                                       photodisassoclatlon  of NO- upon- formation.  Measure-
                                                       ments or estimates of prevailing ancient Oi  levels
                                                       at the  tine CO,),,, occurs  are  needed.
                                                       Unite" and Cole et al." have  found this ap-
                                                       proach to work well 1n predicting maximynt NO* down-
                                                       wind from a aajor point source and a major line
                                                       source.   The  approach assumes thorough mixing of
                                                       the plumes with anbient air at the point where
                                                       (NO )    occurs.  Measurements of the  k. rate con-
                                                       stant Ire not typically available.  The approach also
                                                       requires  measurement of orevallino 03  concentrations
                                                       at the time (*>„)„,„ occurs.

-------
  T-'"' r P. 3  Ajprr.-icre'i for Estimating Maximum Impact on Peek DC? froa> Large  Individual Sources
;  1)  Partial Conversion
                  Description

1)  Multiply predicted maximum NOX concentration
    by a constant factor to predict peak N02.'
                   Remarks

 1)   It  is difficult  to  determine whejt  the  factor should
     be.  Ambient measurements of N0i/N0x ratio* within
     plumes vary widely.   (See Fig.  7.6)  Generally, the
     greater the dilution,  the higher the ratio.  As a
     rule, routinely  collected monitoring data are not
     useful for these purposes, because they do rvst reflect
     the Impact of point sources.  Monitoring data col-
     lected in the vicinity -v-ZOOm of a major roadway may
     be of use for roadway  applications.
      Exponential  Decay
2)  Multiply predicted NOX concentrations by an
    exponential decay factor

    N02 = (rb)(NOx) (l-e'"jT/T) + (b)(N°x)
  •  where t = travel time from source, time center
    T = half life conversion time of NO to N02,
    consistent units
    b = fraction of N0y emitted as NO,, assumed to
    be 0.10 for combustion sources, c6uld be
    higher for industrial process sources.
    (NO )     = Maximum estimated ground level
       x  m3 x          _    .               _
    concentrations of NO  usmc an Inert pol-
    lutant model.
2)  Conversion of NO to N02 appears to be llrri-.e. by the
    oxidizing potential of the dilution air.  Aiso, the
    half life of the NO 1s determined by dilution rather
    than chemistry.  Any value chosen for the half life
    would have to be arbitrary.

-------
 TABL-  "J..1  Approaches  for  Estimating  Maximum Impact on  Peak  .'IP.  From Larno  Individual  Sources
                 Nane
 3)   Ozone Limiting  Approach
3)
                  Description

    Method assumes that maximum N02 concentration
    can be limited by dilution of an NOx plume
    and/or the ozone concentrations in the di-
    lution air.  If ozone is the limiting factor.
                                                 [NO,]     = TO,]  = [HO ]
                                                 L   ?J max   L  3Jt      xj
                                                                         max
                                                where  [0-,]  = ambient ozone concentrations at
                                                the time peak NO-* occurs.
                                                b = frarHon of NO  emitted as N02
                                                [NO ]     = Maximum ground level NO  estimated
                                                with an inert point source model.
                   Remarks

3)  As discussed in Chapter 7.0,  the  summation  of
    (NO + N02 + 03) appears to be constant  across
    point source plumes (see Figures  7.9  and  7.10).
    This Implies that NO to N02 conversion  in plunes
    with high NO  is limited by prevailing  ambient
    ozone.  Fewer arbitrary assumptions are required
    using this method than methods (1) or (2).   Method
    assumes that complete mixing of the NO  plums oc-
    currs at the point where maximum  ground level NO.
    occurs. /It also ignores the  likelihood of  further
    photodisassodation of N02 upon formation.   Measure-
    ments or estimates of prevailing  ambient  0, levels
    at the time (NO )    occurs are needed.
4)  Photostationary State  Approach
4)  Approach combines (NO )max estimated with an
    -.inert pollutant model witn the photo^fl Honor
    state relationship.   As a result,[NO?]    is
    estimated as follows:                 max
                                                 [NO,]
                                                      max
                                                where [NO  ]    = peak ground level NO  con-
                                                         x max                       x
                                                centrations estimated with a model for inert
                                                pollutants
                                                k]  = the  ratio constant in reaction-with
                                                which N02  photodisassociates in the presence
                                                of sunlight:       n
                                                               N02 i^--  NO + 0

                                                k,  = rate constant depicting" the rate with
                                                which 0-, combines with NO to form N02:

                                                               NO + 03 -*  N02 + 02
                                                                       K 3
                                                    4)  White" an(j cole et al.^ have found this ap-
                                                        proach to work  well  in  predicting maximum NO? down-
                                                        wind from a major  point source and  a major  line
                                                        source.   The  approach assumes thorough mixing of
                                                        the plumes with ambient air at the  point where
                                                        (N0x)max  occurs.   Measurements of the kj rate con-
                                                        stant are not typically available.  The approach also
                                                        requires  measurement of prevailing  03 concentrations
                                                        at the time (W))   v occurs.
                                                                       X HiaX

-------
TABLE 8.4  Hypothetical  Example Illustrating Use of Multi- and Individual  Source Models for Projecting Future Short-Term
               concentrations tor use in
                      Scenario

  1)  Observed  nighttime-early morning  peaks.   Im-
      pact  from elevated  point source assumed  to  be
      negligible due to nocturnal  Inversion.
Time of Day

 1900-0800
Estimated Peak Individual
   Source Contribution
  Projected Multi -
Source Contribution

      0.30 ppm
Projected Peak N0g

      0.30 ppm
  2)   Observed mid-morning peaks.  Mu1t1source syn-
      thesis of N02 superimposed by  fumigation  Im-
      pact  from an elevated point source.
 0900-1100
         0.13 ppm
      0.20  ppm
                                                                       0.33 ppm
  3)   Observed mid-afternoon peaks.  Potential
      point  source  Impact  is relatively high due
      to  high observed 0^'levels at this  time.
      Projected multlsource Impact Is low due to
      low N02 concentrations observed by  existing
      monitors during this time.
 1200-1800
        0.15  ppm
     0.06 ppm
     0.21 ppm

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
i. REPORT NO.
    EPA-450/4-80-017
                              2.
                                                            3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
                                                            5. REPORT DATE
    Technical  Basis for Developing Control  Strategies
    for High Ambient Concentrations of  Nitrogen Dioxide
                                                                          September 1980
              6. PERFORMING ORGANIZATION CODE
7'AUTEHciwin L. Meyer, Donald H. Sennet,  Henry S.  Cole and
    Harold G.  Richter
                                                           8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
    U. S.  Environmental  Protection Agency
    Office  of  Air Quality Planning and  Standards
    Monitoring and Data Analysis Division (MD-14)
    Research triangle Park, North Carolina 27711
                                                            10. PROGRAM ELEMENT NO.
              11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
                                                            13. TYPE OF REPORT AND PERIOD COVERED
    Same
                                                            14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
  Report presents  information which may  prove useful in designing  control  strategies
  to reduce high  short-term and/or annual  concentrations of nitrogen  dioxide (N02).
  Specific implications of findings of the design of control strategies  are identified.
  The following topics are discussed:  (1)  nature and significance  of  manmade and natural
  sources of  NOX;  (2) mechanisms by which  N02 is formed in the atmosphere; (3) monitor-
  ing network design, data quality checks, and the extent of high  N02 concentrations
  observed in the  U.S. from 1975-77;  (4) the  extent to which N02 is transported from
  urban to suburban areas and over longer  distances; (5) derived relationships between
  N02, NO and organic pollutants; (6)  procedures for examining monitoring  data to
  estimate impacts of various source  categories observed on high levels  of N02; and
  (7) modeling procedures for N02 which  are available.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS  C.  COSATI Field/Group
  Nitrogen  Dioxide
  Air Quality  Management
  Photochemical  Pollutants
  Air Pollution  Control Strategies
18. DISTRIBUTION STATEMENT
  Unlimited
19. SECURITY CLASS (This Report)
    Unclassified
21. NO. OF PAGES
     145
                                              20. SECURITY CLASS (Thispage)

                                                  Unclassified
                                                                         22. PRICE
EPA Form 2220-1 (Rev. 4-77)
                      PREVIOUS EDITION IS OBSOLETE

-------